International Journal of Food Science & Technology

January 2008 - Vol. 43 Issue 1 Page 1-191 Impact of processing treatments and packaging material on some properties of stored dehydrated cauliflower Dattatreya M. Kadam, David V. K. Samuel, Pitam Chandra & Harman S. Sikarwar pages 1–14

A test procedure to characterise the heating performance of domestic microwave ovens Mark Jeffery Swain, Adrien Spinassou & Maria Veronica Lyra Swain pages 15–23

Chemical composition and toxic trace element composition of some Nigerian edible wild mushrooms Olumuyiwa S. Falade, Oluwatoyin O. Adepoju, Olanrewaja Owoyomi & Steve R. Adewusi pages 24–29

Changes in physical and sensory characteristics of marinated broiler drumsticks, treated with nisin and lactoperoxidase system Fa-Jui Tan & Herbert W. Ockerman pages 30–34

Purification and characterisation of antioxidative peptides from unfractionated rice bran protein hydrolysates Abayomi P. Adebiyi, Ayobamitale O. Adebiyi, Tomohisa Ogawa & Koji Muramoto pages 35–43

Effect of heat treatment and refrigerated storage on antioxidant properties of pre-cut celery (Apium graveolens L.) Sonia Z. Viña, Alicia R. Chaves, pages 44–51

Fluorescence spectroscopy in monitoring of extra virgin olive oil during storage Ewa Sikorska, Igor V. Khmelinskii, Marek Sikorski, Francesco Caponio, Maria T. Bilancia, Antonella Pasqualone & Tommaso Gomes pages 52–61

Heterocyclic amine formation during frying of frozen beefburgers Elna Persson, Bea Kovácsné Oroszvári, Eva Tornberg, Ingegerd Sjöholm & Kerstin Skog pages 62–68

Development of an indirect α-actinin-based immunoassay for the evaluation of protein breakdown and quality loss in fish species subjected to different chilling methods

Mónica Carrera, Vanesa Losada, José Manuel Gallardo, Santiago P. Aubourg & Carmen Piñeiro pages 69–75

Effects of a bacteriocin-like substance produced by Leuconostoc mesenteroides subsp. cremoris on spoilage strain Lactobacillus fructivorans and various pathogens Seyhun Yurdugül & Faruk Bozoglu pages 76–81

Soybean variety and storage effects on soymilk flavour and quality Allaoua Achouri, Joyce Irene Boye & Youness Zamani pages 82–90

Physiological and biochemical changes of different fresh-cut mango cultivars stored at 5 °C Gustavo A. Gonzalez-Aguilar, Jorge Celis, Rogelio R. Sotelo-Mundo, Laura A. de la Rosa, Joaquin Rodrigo-Garcia & Emilio Alvarez-Parrilla pages 91–101

Some properties of polyphenol oxidase from lily Ying Yang & Zhang Wang pages 102–107

A rapid direct solvent extraction method for the extraction of cyclobutanones from irradiated chicken and liquid whole egg Ihab Tewfik pages 108–113

Determination of the veterinary drug maduramicin in food by fluorescence polarisation immunoassay

Zhanhui Wang, Suxia Zhang, Nailya R. Murtazina, Sergei A. Eremin, & Jianzhong Shen pages 114–122

Effect of tumbling and multi-needle injection of curing agents on quality characteristics of pastirma

Ahmet Guner, Zafer Gonulalan & Yusuf Dogruer pages 123–129

Measurement of odour after in vitro or in vivo ingestion of raw or heated garlic, using electronic nose, gas chromatography and sensory analysis

Kazuhiko Tamaki, Shigenori Sonoki, Takeshi Tamaki & Katsuo Ehara pages 130–139

A fuzzy comprehensive evaluation for selecting yeast for cider making

Bangzhu Peng, Tianli Yue & Yahong Yuan pages 140–144

Microbiological and biochemical quality of grouper (Epinephelus chlorostigma) stored in dry ice and water ice Geevarethnam Jeyasekaran, Ramasamy Anandaraj, Ponesakki Ganesan, Robinson Jeya Shakila & Duraisamy Sukumar pages 145–153

Antioxidative activities of grape (Vitis vinifera) seed extracts obtained from different varieties grown in Turkey

Oktay Yemis, Emre Bakkalbasi & Nevzat Artik pages 154–159

Possibility of using near infrared spectroscopy for evaluation of bacterial contamination in shredded cabbage

Phunsiri Suthiluk, Sirinnapa Saranwong, Sumio Kawano, Sonthaya Numthuam & Takaaki Satake pages 160–165

Production of a dry sausage from African catfish (Clarias gariepinus, Burchell, 1822): microbial, chemical and sensory evaluations Abdullah Oksuz, Gulsun Akdemir Evrendilek, Muzaffer Seufi Calis & Akif Ozeren pages 166–172

Peroxidase activity, chlorophylls and antioxidant profile of two leaf vegetables (Solanum nigrum L. and Amaranthus cruentus L.) under six pretreatment methods before cooking Odunayo Clement Adebooye, Ram Vijayalakshmi & Vasudeva Singh pages 173–178

The suitability of muscle of Cirrhinus mrigala in the formation of gel: a comparative electrophoretic study of six tropical carp meats Rupsankar Chakrabarti & Badiveddy Madhusudana Rao pages 179–184

Anti-oxidant capacity of dietary polyphenols determined by ABTS assay: a kinetic expression of the results Jara Pérez-Jiménez & Fulgencio Saura-Calixto pages 185–191

International Journal of Food Science and Technology 2008, 43, 1–14

1

Original article Impact of processing treatments and packaging material on some properties of stored dehydrated cauliflower Dattatreya M. Kadam,1* David V. K. Samuel,2 Pitam Chandra3 & Harman S. Sikarwar4 1 Division of Transfer of Technology, Central Institute of Post-Harvest Engineering and Technology (CIPHET), PO: PAU campus, Ludhiana-141004, Punjab, India 2 Division of Post-Harvest Technology, Indian Agricultural Research Institute (IARI), PUSA campus, New Delhi-110012, India 3 Room No 405, KAB II, ICAR, PUSA campus, New Delhi 110012, India 4 Computer Application, Division of Computer Applications, IASRI, PUSA campus, New Delhi 110012, India (Received 4 February 2005; Accepted in revised form 5 January 2006)

Summary

Investigations were carried out to see the impact of blanching time, pretreatment and storage and packaging on the physico-chemical properties of solar dehydrated cauliflower. The processing treatments selected for the study were blanching time of 3, 5, 7 and 9 min, potassium metabisulphite (KMS) pretreatment having 0.5%, 1.0% and 1.5% concentration level and storage in high-density polyethylene, laminated aluminium foil and polypropylene. The cauliflowers were further processed and dehydrated in solar dryer before packing it into different packaging materials. Packed dehydrated cauliflower was stored for 6 months at room temperature. The stored cauliflower samples were tested periodically for their moisture content, rehydration ratio, rehydration coefficient, ascorbic acid and browning. Ranking of blanching time, chemical concentration level and packaging materials were statistically analysed by using SAS package. The samples with 9 min blanching time, followed by dipping in 1.0% KMS solution, and packed in laminated aluminium foil showed better results in comparison with other treatments.

Keywords

Ascorbic acid, blanching time, browning, concentration, moisture, packaging materials, rehydration.

Introduction

Cauliflower (Brassica oleracea var. botrytis) is an important sole cole crop (crop near to the soil) of North India. The estimated postharvest loss of cauliflower per hectare in India is about 49% (Sehgal, 1999). The most serious constraint for shelf-life enhancement is the activity of micro-organisms. Water in food is alleviated to a very low level during dehydration, thus achieving better microbiological preservation and retarding many undesirable reactions during storage (Ibarz & Barbosa-Canovas, 2000), owing to the reduction in water activity. Pretreatments, such as blanching, dipping and sulfiting are common in most drying processes to improve product quality or process efficiency. In recent years, exhaustive efforts have been made for an improvement in the quality retention of dried products by altering processing strategy and/or pretreatment. Blanching results in some degree of chlorophyll degradation with the subsequent formation of pheophytin. The extent of chlorophyll conversion is related to the degree of *Correspondent: e-mail: [email protected] and kadam1k@ rediffmail.com

doi:10.1111/j.1365-2621.2006.01372.x 2007 Institute of Food Science and Technology Trust Fund

blanching. Peroxidase activity is widely used as an index of blanching because peroxidase is the most heat-stable enzyme found in vegetables (Rahman & Perera, 1999). Optimum conditions of blanching, time and temperature, are necessary to achieve the desired quality of dried products (Haas et al., 1974). Permitted levels of sulphur dioxide and other additives (solutes) in dried foods vary from country to country. The allowed limit is 2000 mg SO2 per kg dried food (Anon., 1990). Sulfiting was done by dipping cauliflower for 15 min in a mixture of 1.0% sodium metabisulphite and sodium sulphite (3:1), and better quality and rehydration ratio (RR) of dehydrated cauliflower was found (Srivastava & Sulebele, 1975). For long-term storage of dried fruits or vegetables, sulfuring or sulfite dips are considered to be most promising pretreatments. Hence, standardisation of blanching process for cauliflower with preservative is essential. Dehydration is one of the oldest methods of food preservation and an important food processing stage (Lima et al., 2002). Dehydration of foods is aimed at producing a high-density and high-quality product, which, when adequately packaged has a long shelf life,

2

Properties of stored dehydrated cauliflower D. M. Kadam et al.

after which the food can be rapidly reconstituted without substantial loss of flavour, taste, colour and aroma (Sarsavadia et al., 1999). Rehydration properties of the dehydrated foods are of crucial importance in view of the increased consumer demand for processed products that retain more of their original characteristics. Processing and subsequent storage causes variation in food characteristics. Environmental aspects, like temperature, air humidity or light can trigger reaction mechanisms that lead to the degradation of some quality aspects of the food. During storage, one or more food characteristics can reach an undesirable state, and as a consequence the consumer may reject the product or it can even cause detrimental health. At this moment, it is considered that the food has reached the end of its shelf life (Singh, 1992). Ascorbic acid content of stored fruits and vegetables generally decreases more rapidly at higher storage temperature. Relative humidity has been reported to affect ascorbic acid content and the effect was greater in tissues, which are susceptible to wilting in fruits and vegetables (Ezell & Wilcox, 1952; Watada et al., 1976; Salunkhe et al., 1991). Ascorbic acid losses ranged from 17.1 to 26.5% during cauliflower processing (Raina et al., 1982), loss of 53.5% ascorbic acid on slow cooling as compared with 38% on immediate cooling of blanched cauliflower. Lutz et al. (1987) quoted that the solar drier offers the advantage of protecting the crop during the drying period. To date, several attempts have been reported to optimise the food dehydration process with respect to product quality, productivity and energy cost, and solar drying has been successfully tried for different fruits and vegetables. There is limited literature available on the use of indirect-type tunnel solar dehydrator for cauliflower drying and studies on storage of solar dehydrated cauliflower and their effect. An attempt was made to study the impact of process treatment (optimisation of blanching time and different chemical concentration) and packaging materials on storability of solar dehydrated cauliflower.

of 6 months was investigated. Figure 1 gives the detailed process technology for cauliflower dehydration. Sample preparation

The fresh cauliflower (B. oleracea var. botrytis) was procured from Indo-Israel project, IARI, New Delhi. Cleaning and trimming of procured fresh cauliflower was done manually and edible cauliflower curd was sized to florets. The edible portion of the cauliflower curd was cut into pieces of around 3.0 · 4.5 cm in size by using stainless steel knife after removing nonedible leaves and stem portion. It is easy to cut and reduce the size of the cauliflower when it is fresh. Dried cauliflower is difficult to cut and hence considering these points, it was decided to cut the fresh cauliflower in accordance with ready-touse cooking purposes.

Cauliflower harvesting Transporting to lab in crates Cleaning and trimming Edible curd and sizing Weighing Washing in cold water

Drain out washed cauliflower Blanching for 3,5,7 & 9 min. Cooling in cold water

Drain out cooled cauliflower Dipping in KMS

0.50%

1.00%

1.50%

Materials and methods

The work presented in this section was undertaken to study the impact of processing treatments and packaging materials on some properties of stored dehydrated cauliflower (B. oleracea var. botrytis) and resultant changes in physico-chemical characteristics of dehydrated and rehydrated cauliflower. The experiments were conducted in the Processing and Solar Energy Laboratories, Division of Agricultural Engineering, Indian Agricultural Research Institute, New Delhi, India. The process followed for the preparation of dehydrated cauliflower from fresh cauliflower and its suitability for storage in different packaging materials for a period


Draining out chemical and spreading in trays Loading of tray in trolley for dehydration Dehydrated cauliflower Weighing and packing Polypropylene

HDPE

Laminate aluminum foil

Storage

Figure 1 Process flow chart for cauliflower dehydration.

2007 Institute of Food Science and Technology Trust Fund


Blanching

Boiling water method was used for blanching cauliflower. It was achieved by tying the cauliflower pieces in a muslin cloth and dipping it into boiling water for 3, 5, 7 and 9 min. Samples were then cooled in cold water to avoid over cooking and discolouration. Processing treatments and concentration level

Blanched and cooled cauliflower samples were dipped in a water solution of potassium meta bisulphite (KMS) for 15 min in three different concentration levels of 0.5%, 1.0% and 1.5% and drained cauliflower were spread in trays for drying. Drying conditions

Indirect-type solar dehydrator was used to dry the cauliflower. Around 500 kg per batch capacity solar dryer took 28 sunshine hours to dry the cauliflower from 1157.86% dry basis (d.b.) or 92.05% wet basis (w.b.) to 6.29 ± 0.5% d.b. or 5.91 ± 0.5% w.b. The ambient air temperature and relative humidity ranged from 19.6 to 41.2 C and 11.6% to 55.6%, respectively during March to April. The range of solar radiation during the study was 200 to 980 W m)2. The variation in heated air temperature ranged from 22.2 to 51.1 C and relative humidity ranged from 13.7% to 45.3% at inlet of drying chamber during solar dehydration studies. Packaging material

Availability of packaging material in a local market was the major consideration while selecting appropriate packaging material. Three different packaging materials used are polypropylene (PP, 200 gauge), high-density polyethylene (HDPE, 200 gauge) and laminated aluminium foil (LF), having a dimension of 25 · 18, 21 · 15 and 19 · 13 cm, respectively. Dehydrated samples were packed in these packaging materials and sealed with the help of heat-sealing machine. The samples were stored at room temperature for 6 months and were evaluated for their quality attributes at regular interval, i.e. 1.5 months. Determination of moisture content

The moisture content of dehydrated cauliflower was determined by standard hot air oven method (AOAC, 1970). The temperature of hot air oven ranged from 0 to 250 C with a least count of 0.1 C. The samples were weighed on weighing balance with a sensitivity of 0.01 g. The loss in weight during drying was used to calculate the moisture content of the dried sample.


For determining initial moisture content of sample, 5 g of dried cauliflower samples were kept in hot oven for 4 h at 65 C temperature. The moisture content was determined by using following expression: MCð%d:b:Þ ¼ Weight of water in product 100 Weight of dry matter of product sample

ð1Þ

Rehydration ratio

Rehydration is achieved by keeping the dehydrated product in boiling water for 15-min duration so as to restore the fresh-like appearance of the dehydrated product. It is the ratio of the weight of dehydrated cauliflower to the weight of rehydrated cauliflower. If the weight of the dehydrated sample is A and the drained weight of the rehydrated sample is B, then RR can be written as: RR ¼ A=B ¼ A : B

ð2Þ

Rehydration coefficient

Rehydration coefficient (RC) is defined as the weight of rehydrated cauliflower into dry matter in fresh cauliflower samples taken for dehydration, divided by the weight of dry matter in dehydrated cauliflower samples. RC is calculated as: RC ¼ n

ðDwd ½100 MCi Þ h io f 100 WdtR WdtR MC 100

ð3Þ

where RC is the rehydration coefficient, Dwd is the drained weight of dehydrated sample in grams, MCi is the moisture content of sample before drying (fresh sample) in percent, WdtR is the weight of dried sample taken for rehydration in grams and MCf is the moisture content in the dried sample taken for rehydration in percent. Ascorbic acid

Ascorbic acid (vitamin C) was determined using volumetric method (Sadasivam & Manickam, 1996) and was expressed as mg 100 g)1 sample. Calculation: Amount of ascorbic acid, mg 100 g)1 sample. 0:5 mg V2 mL V1 mL 5 mL 100 mL 100 ð4Þ weight of the sample

Vit.C (mg per 100 g) ¼


3


Blanching time: 3 min Blanching time: 7 min

(a) 6.70


6.65

Moisture content, % db

6.60 6.55 6.50 6.45 6.40 6.35 6.30 6.25 0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

0.0 1.5 3.0 4.5 6.0

1.0% KMS

1.5% KMS

Storage period (months) and KMS concentration level (%)

(b) 6.95

Blanching time: 3 min


6.85



Moisture content, % db

6.75 6.65 6.55 6.45 6.35 6.25 0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

0.0 1.5 3.0 4.5 6.0 1.5% KMS

1.0% KMS



(c) 6.39


6.37 Moisture content, % db

4

6.35 6.33 6.31 6.29 6.27 0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

0.0 1.5 3.0 4.5 6.0

1.0% KMS

1.5% KMS

Storage period (months) and KMS concentration level (%) Figure 2 Effect of packaging materials on moisture content in dehydrated cauliflower during storage. (a) Effect of polypropylene package; (b) effect of high-density polyethylene package; (c) effect of laminated aluminium foil package.




where V1 is the standard dye solution (mL) and V2 is the sample extract solution (mL).

permeability (g m)2 per 24 h) in PP and HDPE packaging material were 4.4 and 3 (N2), 92 and 45 (CO2), and 28 and 10 (O2) respectively. LF had good moisture

Nonenzymatic browning

The increase in the absorbance of the sample extract at 420-nm wavelengths was taken as a measure of nonenzymatic browning (Ranganna, 1986). The extract was prepared by soaking 5 g of cut sample in 100 mL of 60% ethyl alcohol for 12 h and filtered with waterman number one filter paper. The optical density was recorded at 420-nm using a digital spectrophotometer. Statistical analysis

The experiment was conducted using a factorial completely randomised design (CRD) with four blanching times, three preservatives and three preservative concentration levels and three replications each. Physicochemical data were analysed as per procedure factorial CRD, using PROC GLM of SAS. If the factorial effects are found to be significantly different through anova, then these were subjected to multiple comparison procedure using least significant difference at 5% level of significance. The identification of best treatment combinations was done through one way classified anova, followed by multiple comparison procedure. Results and discussion

Moisture content

Dehydrated cauliflower had average moisture content of 6.29% d.b. before they were packed in different packaging materials and stored at room temperature. Samples in all packaging materials absorbed moisture from the atmosphere. It was observed that the samples showed very little increase in moisture content during the first one-and-a-half month of storage; thereafter the gain in moisture content was increased till 4.5 months of storage period, irrespective of blanching time and chemical concentration level (Fig. 2). However, after 6 months of storage, the trend of moisture gain varied from sample to sample. Maximum moisture gain was observed in samples blanched for 3 min, treated with 0.5% KMS, and packed in HDPE package. Samples blanched for 7 and 9 min, treated with KMS and packed in laminated aluminium foil (LF) showed minimum moisture gain after 4.5 months of storage. The increase or decrease in the moisture content of the samples may be attributed to the water vapour transmission rate (WVTR) of the packaging materials. Water and gas permeability of LF is almost nil. PP and HDPE have 5.5 and 6.8 g m)2 per 24 h of WVTR, respectively. Gas


Table 1 Effect of blanching time on moisture content of stored cauliflower in different packaging materials Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

Rank 4

0.0 1.5 3.0 4.5 6.0

0.0000 0.0037 0.0051 0.0097 0.0090

– 3A 7A 9A 3A

– 5B 5B 7B 5B

– 9B 9C 5C 7B

– 7C 3D 3D 9D

(6.3160) (6.3930) (6.4061) (6.4330)

(6.3236) (6.4060) (6.4712) (6.4580)

(6.3233) (6.4150) (6.4900) (6.4581)

(6.3340) (6.4230) (6.6000) (6.4950)

First digit in rank column indicates blanching time in minutes. Same superscripts are not significantly different. (In parenthesis) indicates mean values of MC at particular condition (% dry basis).

Table 2 Effect of chemical concentration level on moisture content of stored cauliflower in different packaging materials Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

0.0000 0.0032 0.0044 0.0084 0.0078

– 0.50A 1.00A 1.00A 1.00A

– 1.00A 1.50B 1.50A 1.50B

– 1.50B 0.50A 0.50B 0.50C

(6.3211) (6.3940) (6.4654) (6.4310)

(6.3211) (6.3982) (6.4670) (6.4633)

(6.3284) (6.4370) (6.5551) (6.4840)

First three digits with decimal indicate chemical concentration level in rank column. Same superscripts are not significantly different. (In parenthesis) indicates mean values of MC at particular condition (% dry basis).

Table 3 Effect of packaging materials on moisture content of stored dehydrated cauliflower Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

0.0000 0.0032 0.0044 0.0084 0.0078

– LFA LFA LFA LFA

– PPB PPB PPB PPB

– HDPEC HDPEC HDPEC HDPEC

(6.2980) (6.3100) (6.3962) (6.3174)

(6.3333) (6.4281) (6.5233) (6.5266)

(6.3391) (6.4845) (6.5640) (6.5411)

LF, laminated aluminium foil; PP, polypropylene, HDPE, high-density polyethylene. Same superscripts are not significantly different. (In parenthesis) indicates mean values at particular condition (% dry basis).


5


(a) 9





8 Rehydration ratio

7 6 5 4 3 2 1 0.0

1.5

3.0

4.5

6.0

0.0

1.5

0.5% KMS

3.0

4.5

6.0

0.0

1.5

1.0% KMS

3.0

4.5

6.0

1.5% KMS


(b) 9 8





Rehydration ratio

7 6 5 4 3 2 1 0.0

1.5

3.0

4.5

6.0

0.0

0.5% KMS

1.5

3.0

4.5

6.0

0.0

1.5

1.0% KMS

3.0

4.5

6.0

1.5% KMS


(c) 9 8





7 Rehydration ratio

6

6 5 4 3 2 1 0.0

1.5

3.0

4.5

6.0

0.0

0.5% KMS

1.5

3.0

4.5

1.0% KMS

6.0

0.0

1.5

3.0

4.5

6.0

1.5% KMS

Storage period (months) and KMS concentration level (%) Figure 3 Effect of packaging materials on rehydration ratio during storage. (a) Effect of polypropylene package; (b) effect of high-density polyethylene package; (c) effect of laminated aluminium foil package.




barrier properties as compared to PP and HDPE packaging materials studied. Blanching time was found to have no significant role in the moisture content of dehydrated cauliflower during the storage period of 6 months (Table 1). Chemical concentration level affected the moisture content during storage of dehydrated cauliflower over a period of 6 months. It was observed that 1.0% KMS treated samples ranked first throughout the storage study, and this was followed by 1.5% and 0.5% KMS-treated samples. On the basis of the observation, it can be concluded that 1.0% KMS-treated samples were best for minimum moisture gain during storage (Table 2). Packaging materials during storage were significantly different from each other in respect to moisture gain. Minimum moisture gain was observed in the LF-packed dehydrated cauliflower, irrespective of blanching time and chemical concentration levels. It may be attributed to almost nil WVTR and gas transmission rate (GTR). Laminated aluminium foil package ranked first, followed by PP and HDPE throughout the 6-month storage (Table 3).

other at 5% CD level throughout the storage period. During zero (initial) days of studies, 1.0% KMS-treated fresh samples have shown better results as compared with 0.5% and 1.5% KMS-treated fresh samples. Better quality and RR of dehydrated cauliflower was obtained from 1.0% sodium metabisulfite and sodium sulfite treatment (Srivastava & Sulebele, 1975). During 6 months of storage period, 0.5% KMS-treated sample was found best, followed by those treated with 1.5% and 1.0% KMS and these were ranked first, second and third, respectively (Table 5). Packaging materials are significantly different from each other and were found to have significant effect on RR at 5% CD level throughout the storage period. LF-packed dehydrated cauliflower ranked first, followed by PP and HDPE, except during 4.5-month interval of storage. The shelf life of the dehydrated cauliflower has increased from 2 to 10 months at ambient temperature, when packed in polyethylene (Pokharkar et al., 1997) and garlic powder in aluminium laminated bags (Ambrose & Sreenarayanan, 1998), whereas HDPEpacked samples have shown higher RR than LF package (Table 6). In general, RR decreased with increase in

Rehydration ratio

It was observed that RR decreased with increase in blanching time irrespective of concentration levels, whereas RR decreased with increase in storage period in all samples. Minimum RR was 3.57 in samples treated with 1.0% KMS, blanched for 5 min and maximum RR was 6.65 for the samples blanched for 7 min, treated with 0.5% KMS and stored in LF for 6 months (Fig. 3). Similar result was found in rehydrated cabbage (Mulay et al., 1994). Blanching time was found to have significant effect on RR of cauliflower at 5% critical difference (CD) level. From Table 4, it is evident that 3-min blanching time ranked first throughout the storage period except 4.5month storage interval, where it was third in position. Concentration levels are significantly different from each Table 4 Effect of blanching time on rehydration ratio of stored

Table 5 Effect of chemical concentration level on rehydration ratio of stored cauliflower in different packaging materials Storage period, months 0.0 1.5 3.0 4.5 6.0

CD0.05 0.0577 0.0477 0.0446 0.0491 0.0446

Rank 1 A

1.00 1.50A 0.50A 0.50A 0.50A

(6.9316) (5.4600) (5.3770) (5.2940 (5.3890)

Rank 2 B

0.50 1.50B 1.50B 1.50B 1.50B

(6.4331) (5.0346) (4.9800) (4.9422) (4.8408)

Rank 3 1.50C 1.00C 1.00C 1.00C 1.00C

(6.0652) (4.9001) (4.8303) (4.7696) (4.6905)

First three digits with decimal indicate chemical concentration level. Same superscripts are not significantly different. (In parenthesis) indicates mean values of RR at particular condition.

Table 6 Effect of packaging materials on rehydration ratio of stored dehydrated cauliflower

cauliflower in different packaging materials Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

Rank 4

0.0 1.5 3.0 4.5 6.0

0.0666 0.0551 0.0515 0.0567 0.0515

3A 3A 3A 9A 3A

5B 5B 5B 5B 7B

7C 7C 7C 3C 5C

9D 9D 9D 7D 9D

(7.0021) (5.9741) (5.7059) (5.1196) (5.1416)

(6.6005) (5.0433) (5.0432) (5.0403) (5.0291)

(6.4012) (4.8701) (4.8701) (4.9619) (4.9113)

(5.8863) (4.6321) (4.6321) (4.8696) (4.5407)

First digit in rank column indicates blanching time in minutes. Same superscripts are not significantly different. (In parenthesis) indicates mean values of RR at particular condition.


Storage period, months

CD0.05

Rank 1

0.0 1.5 3.0 4.5 6.0

– 0.0477 0.0446 0.0491 0.0446

– LFA LFA HDPEA LFA

(5.3001) (5.2704) (5.1496) (5.3600)

Rank 2

Rank 3

– PPB PPB PPB PPB

– HDPEC HDPEC LFC HDPEC

(5.2306) (5.1003) (4.9700) (4.8686)

(4.8500) (4.8216) (4.8786) (4.6863)

LF, laminated aluminium foil; PP, polypropylene; HDPE, high-density polyethylene. Same superscripts are not significantly different. (In parenthesis) indicates mean values at particular condition.


7


Blanching time: 3 min Blanching time: 5 min Blanching time: 7 min Blanching time: 9 min

(a) Rehydration coefficient

0.675 0.600 0.525 0.450 0.375 0.300

0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

1.0% KMS

1.5% KMS


(b)



0.650


0.600


0.550 0.500 0.450 0.400 0.350 0.300

0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

1.0% KMS

1.5% KMS


(c)


8


0.675 0.625 0.575 0.525 0.475 0.425 0.375 0.325 0.275


0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

1.0% KMS

1.5% KMS

Storage period (months) and KMS concentration level (%) Figure 4 Effect of packaging materials on rehydration coefficient of dehydrated cauliflower during storage. (a) Effect of polypropylene package; (b) effect of high-density polyethylene package; (c) effect of laminated aluminium foil package.




storage period, irrespective of blanching time, chemical concentration level and packaging material.

Table 9 Effect of packaging materials on rehydration coefficient of stored dehydrated cauliflower



CD0.05

Rank 1

0.0 1.5 3.0 4.5 6.0

0.0049 0.0040 0.0038 0.0042 0.0038

–

In general, similar results of RR were obtained for RC of stored cauliflower. The results pertaining to the effect of pretreatment on RC are summarised in Fig. 4. RC decreased with increase in blanching time for all three concentration levels. The RC was sout at a minimum (0.486) in 1.5% KMS-treated samples, subjected to 9-min blanching and maximum of 0.647 in 1.0% KMStreated samples and 3-min blanching during 0 days. Raina et al. (1982) found similar range of RC (0.51 to 0.56) in rehydrated cauliflower of different varieties. After 6 months of storage in LF, minimum RC was 0.302 for 5-min blanched cauliflower dipped in 1.0% KMS solution and maximum RC was 0.527 for 3-min blanched cauliflower dipped in 0.5% KMS solution. From Table 7, it is clear that samples blanched for 3 min ranked first throughout the storage period, except for sample stored for 4.5 months. Maximum mean RC was 0.591 and 0.458 during 0 day and 6 months of storage for 3-min blanched cauliflower samples, respectTable 7 Effect of blanching time on rehydration coefficient of stored cauliflower in different packaging materials Storage period, months 0.0 1.5 3.0 4.5 6.0

CD0.05 0.0057 0.0047 0.0044 0.0048 0.0044

Rank 1 A

3 3A 3A 9A 3A

Rank 2

(0.5910) (0.5050) (0.4831) (0.4333) (0.4585)

B

5 5B 5B 5B 7B

Rank 3 C

(0.5576) (0.4261) (0.4261) (0.4261) (0.4244)

7 7C 7C 3C 5C

(0.5410) (0.4111) (0.4122) (0.4200) (0.4161)

Rank 4 9D 9D 9D 7D 9D

(0.4980) (0.3921) (0.3921) (0.4120) (0.3844)

First digit indicates blanching time in minutes. Same superscripts are not significantly different. (In parenthesis) indicates mean values of RC at particular condition.

Table 8 Effect of chemical concentration level on rehydration coefficient of stored cauliflower in different packaging materials

LFA LFA HDPEA LFA

(0.4481) (0.4462) (0.4362) (0.4536)

Rank 2

Rank 3

– PPB PPB PPB PPB

– HDPEC HDPEC LFC HDPEC

(0.4422) (0.4321) (0.4213) (0.4120)

(0.4100) (0.4080) (0.4122) (0.3970)

LF, laminated aluminium foil; PP, polypropylene; HDPE, high-density polyethylene. Same superscripts are not significantly different. (In parenthesis) Indicates mean values at particular condition.

ively. Minimum mean RC was 0.498 and 0.384 during 0 day and 6 months of storage for 9-min blanched cauliflower samples, respectively. Chemical concentration levels were significantly different from each other at 5% CD level throughout the storage period. Zero-day samples treated with 1.0% KMS have shown better result as compared with 0.5% and 1.5% KMS-treated samples. During storage period of 6 months, samples treated with 0.5% KMS were found to be best followed by 1.5% and 1.0% KMS, and these were ranked first, second and third, respectively (Table 8). Maximum mean RC was 0.585 and 0.456 during 0 day and 6 months of storage for samples treated with 1.0% and 0.5% KMS concentration levels and ranked first, respectively. Minimum mean RC was 0.5124 and 0.3970 during 0 day and 6 months of storage for samples treated with 1.5% and 1.0% KMS concentration levels, respectively. Packaging materials were significantly different from each other and affected RC significantly at 5% CD level for all storage intervals. From Table 9, it is clear that LF-packed dehydrated cauliflower ranked first. This may be attributed to low WVTR and GTR, followed by PP and HDPE during storage, except for the samples stored for 4.5 months. In general, RC decreased with increase in storage period, irrespective of blanching time, chemical concentration level and packaging material.


CD0.05

Rank 1

Rank 2

Rank 3

Ascorbic acid

0.0 1.5 3.0 4.5 6.0

0.0049 0.004 0.0038 0.0042 0.0038

1.00A 1.50A 0.50A 0.50A 0.50A

0.50B 1.50B 1.50B 1.50B 1.50B

1.50C 1.00C 1.00C 1.00C 1.00C

Ascorbic acid (vitamin C) present in fresh cauliflower was ranged between 70 and 77.97 mg 100 g)1. Ascorbic acid decreased with increase in blanching time, as vitamin C is sensitive to heat. Cauliflower lost ascorbic acid as drying took place, but dehydrated cauliflower showed higher percentage of ascorbic acid than fresh cauliflower. This is because, as moisture evaporates, the ascorbic acid gets concentrated in dehydrated cauliflower. Freshly dehydrated cauliflower (0 day) has

(0.5856) (0.4610) (0.4549) (0.4483) (0.4561)

(0.5435) (0.4250) (0.4220) (0.4183) (0.4095)

(0.5124) (0.4140) (0.4090) (0.4030) (0.3970)

First three digits with decimal indicate chemical concentration level in rank column. Same superscripts are not significantly different. (In parenthesis) indicates mean values of RC at particular condition.



9



Ascorbic acid (mg 100 g–1)

(a)

575 475 375 275 175 75 0.0

1.5

3.0

4.5

6.0

0.0

1.5

0.50%


3.0

4.5

0.0

6.0

1.5

1.00%

3.0

4.5

6.0

4.5

6.0

1.50%


(b)

575 475 375 275 175 75 0.0

1.5

3.0

4.5

6.0

0.0

1.5

0.50%

3.0

4.5

0.0

6.0

1.5

1.00%

(c)


10

3.0

1.50%


600 550 500 450 400 350 300 250 200 150 100 0.0

1.5

3.0

4.5

6.0

0.0

0.50%

1.5

3.0

4.5

1.00%

6.0

0.0

1.5

3.0

4.5

6.0

1.50%

Figure 5 Effect of packaging materials on ascorbic acid content of dehydrated cauliflower during storage. (a) Effect of polypropylene package; (b) effect of high-density polyethylene package; (c) effect of laminated aluminium foil package.




581.37 mg 100 g)1 of ascorbic acid in samples treated with 0.5, 1.0 and 1.5% of KMS and blanched for 3 min. Ascorbic acid in samples treated with 1.5% KMS after blanching for 9 min had a minimum of 453.92 mg 100 g )1. Figure 5 shows that with increase in storage period, the ascorbic acid content decreased, irrespective of the type of packaging material used. Many researchers found similar results (Watada et al., 1976; Smoot & Nagy, 1980; Raina et al., 1982; Salunkhe et al., 1991; Rogacheva et al., 1995). Drastic loss in ascorbic acid content was observed in all packaging materials during the first 1.5-month storage period. It may be attributed to atmospheric oxygen presence in package, which is an essential element in the loss of vitamin C (Chauhan et al., 1998). However, the losses reduced as storage time progressed, irrespective of packaging material, blanching time and KMS concentrations. Table 10 shows that ascorbic acid decreases with increase in blanching time for fresh sample (0 days) with least significant difference (LSD) of 24.318. Blanching time has significant effect on ascorbic acid of cauliflower. The maximum and minimum ascorbic acid present in dehydrated cauliflower were 581.37 and Table 10 Effect of blanching time on ascorbic acid content of stored cauliflower in different packaging materials Storage period, months CD0.05

Rank 1

Rank 2

Rank 3

Rank 4

0.0 1.5 3.0 4.5 6.0

3A 5A 5A 9A 9A

5B 3A 3A 7A 5B

7C 7B 7A 5B 7C

9D 9C 9B 3B 3D

24.318 09.369 08.599 12.619 07.280

(581.369) (265.041) (235.920) (223.858) (170.186)

(529.090) (257.501) (231.798) (217.122) (159.979)

(486.604) (241.519) (230.900) (198.362) (151.934)

(455.233) (194.710) (207.803) (197.656) (143.888)

First digit indicates blanching time in minutes. Same superscripts are not significantly different. (In parenthesis) indicates mean values of ascorbic acid at particular condition.

Table 11 Effect of chemical concentration level on ascorbic acid content of stored cauliflower in different packaging materials Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

21.060 08.113 07.447 10.928 06.305

0.50A 0.50A 0.50A 1.50A 1.50A

1.00B 1.00B 1.00B 1.00A 0.50A

1.50C 1.50B 1.50B 0.50B 1.00B

(518.142) (261.190) (233.256) (217.911) (161.877)

(492.090) (231.066) (223.283) (212.504) (161.843)

(465.233) (226.820) (223.258) (197.286) (145.773)

First three digits with decimal indicate chemical concentration level. Same superscripts are not significantly different. (In parenthesis) indicates mean values of ascorbic acid at particular condition.


Table 12 Effect of packaging materials on ascorbic acid content of stored dehydrated cauliflower Storage period, months

CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

21.060 08.114 07.447 10.928 06.305

– LFA LFA LFA LFA

–

– HDPEB HDPEB HDPEB HDPEC

(246.756) (229.521) (221.849) (181.601)

PPB PPAB PPA PPB

(237.493) (228.308) (217.450) (155.033)

(234.821) (221.979) (188.446) (132.862)

LF, laminated aluminium foil; PP, polypropylene; HDPE, high-density polyethylene. Same superscripts are not significantly different. (In parenthesis) indicates mean values at particular condition (mg 100 g)1).

455.23 mg 100 g)1 for freshly prepared samples of cauliflower with blanching time of 3 and 9 min, respectively, while nonsignificant effect of blanching time was observed during storage period on the ascorbic acid content, irrespective of packaging materials and chemical concentration levels. By and large, the chemical concentration levels are nonsignificantly different from each other at 5% CD level for freshly prepared dehydrated cauliflower samples (0-day storage). However, in some cases, significant difference was observed during storage studies. Samples treated with 0.5% KMS retained more ascorbic acid (Bajaj et al., 1993) followed by 1.0% and 1.5% KMS treatments during entire length of storage period, except for the samples stored for 4.5 months. The 0.5% KMS-treated samples ranked third with ascorbic acid content of 197.29 mg 100 g)1, whereas 1.5% KMS-treated samples ranked first with mean ascorbic acid content of 217.91 mg 100 g)1 for samples stored for 4.5 months (Table 11). Maximum and minimum mean ascorbic acid difference in 0-day and 6-month-old samples were 71.86% and 68.75%, respectively. Between LF and PP, there was no significant difference during storage except for 6-month samples, in which, it has shown significant difference in all packaging materials (Table 12). In general, ascorbic acid decreased with increase in storage period, irrespective of packaging material used. LF retained more ascorbic acid compared with PP and HDPEpacked samples during 6 months of storages. Nonenzymatic browning

Browning decreased with increase in blanching time and KMS concentration level. Browning in dehydrated cauliflower during storage showed increasing trend (Fig. 6). Sharma et al. (2000) found similar results during storage of dehydrated carrots. Increase in browning of cauliflower was observed in all packaging


11


Optical density

(a)


0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

1.0% KMS

0.70 0.60 Optical density

0.0 1.5 3.0 4.5 6.0 1.5% KMS


(b)

0.50 0.40 0.30 0.20 0.10 0.00 0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

Optical density

12

1.0% KMS

(c) 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00

0.0 1.5 3.0 4.5 6.0 1.5% KMS


0.0 1.5 3.0 4.5 6.0

0.0 1.5 3.0 4.5 6.0

0.5% KMS

1.0% KMS

0.0 1.5 3.0 4.5 6.0 1.5% KMS

Figure 6 Effect of blanching time, potassium metabisulphite concentration levels and laminated aluminium foil packaging materials on browning during storage. (a) Effect of polypropylene package on browning during storage; (b) effect of high-density polyethylene package; (c) effect of laminated aluminium foil package.




Table 13 Effect of blanching time on browning (optical density) of stored cauliflower in different packaging materials

Table 15 Effect of packaging materials on browning (optical density) of stored dehydrated cauliflower


CD0.05

Rank 1

Rank 2

Rank 3

Rank 4


CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

0.0004 0.0006 0.0050 0.0055 0.0027

9A 9A 9A 9A 9A

7B 7B 7B 7B 7B

5C 5B 5C 5C 5C

3D 3D 3B 3D 3D

0.0 1.5 3.0 4.5 6.0

0.0003 0.0005 0.0044 0.0047 0.0023

– LFA PPA LFA LFA

– PPB LFA PPB PPB

– HDPEC HDPEB HDPEC HDPEC

(0.0090) (0.0142) (0.0312) (0.1022) (0.1440)

(0.0168) (0.0254) (0.0565) (0.1518) (0.3600)

(0.0301) (0.0391) (0.0710) (0.2582) (0.3828)

(0.0770) (0.0921) (0.1055) (0.3339) (0.4573)

First digit indicates blanching time in minutes. Same superscripts are not significantly different. (In parenthesis) indicates mean values of browning at particular condition.

materials, irrespective of blanching time, except LF samples of 9-min blanching and treated with 1.5% KMS concentration. HDPE-packed samples showed highest browning level of 0.6637 for samples treated with 0.5% KMS and blanched for 5 min, whereas the minimum browning (0.038) occurred in samples of 9-min blanching, which were 1.5% KMS treated and stored in LF for 6 months. It may be attributed to the WVTR and GTR properties of packaging materials. Browning increased in the dried product with storage period. Blanching time has significant effect on browning of cauliflower at 5% CD level. Table 13 shows that browning (optical density) decreases with increase in blanching time and increases with increase in storage period in all treatment combinations of samples stored. Maximum and minimum browning in dehydrated cauliflower was 0.009 and 0.077, respectively for cauliflower blanched for 9 and 3 min, respectively during 0 days’ storage. Browning increased slowly during initial storage, but drastic increase was noticed after 4.5 months. The 9-min blanching time retains first rank with minimum browning (0.144) and 3-min blanching time retained its last position with maximum (0.4573) Table 14 Effect of chemical concentration level on browning (optical density) of stored cauliflower in different packaging materials

(0.0400) (0.0613) (0.1536) (0.2637)

(0.0409) (0.0625) (0.2173) (0.3423)

(0.0472) (0.0746) (0.2636) (0.4025)

LF, laminated aluminium foil; PP, polypropylene; HDPE, high-density polyethylene. Same superscripts are not significantly different. (In parenthesis) indicates mean values at particular condition.

browning after 6 months storage studies. LSD was 0.0027 at 5% CD level of significance. Chemical concentration levels and packaging materials are significantly different from each other at 5% CD level during storage. Samples treated with higher chemical concentration levels showed lower browning in it and vice versa. Initial samples have shown LSD of 0.0003. Browning increased with increase in storage period in all concentration levels (Table 14). Pawar et al. (1988) obtained similar output during storage of onion flakes. Samples treated with 1.5% KMS show lower optical density (0.2838), whereas 0.5% KMStreated samples show higher optical density (0.3736) after 6 months of storage period. LSD during this period was 0.0023. From Table 15, it is evident that LF-packed dehydrated cauliflower ranked first, followed by PP and HDPE during storage studies. Dehydrated cauliflower had 0.03325 mean optical density prior to packing with 0.0003 LSD and 0.0023 LSD after 6 months of storage at 5% CD level of significance. In general, browning increased with increase in storage period, irrespective of packaging material used. LF-packed samples show minimum browning (0.2637) compared with PP (0.3423) and HDPE (0.4025) packed samples at the end of 6 months of storage. Conclusion


CD0.05

Rank 1

Rank 2

Rank 3

0.0 1.5 3.0 4.5 6.0

0.0003 0.0005 0.0044 0.0047 0.0023

1.50A 1.50A 1.50A 1.50A 1.50A

1.00B 1.00B 1.00A 1.00B 1.00B

0.50C 0.50C 0.50B 0.50C 0.50C

(0.0000) (0.0390) (0.0627) (0.1836) (0.2838)

(0.0000) (0.0422) (0.0655) (0.1919) (0.3510)

(0.0000) (0.0468) (0.0699) (0.2591) (0.3736)

First three digits with decimal indicate chemical concentration level. Same superscripts are not significantly different. (In parenthesis) indicates mean values of browning at particular condition.


The most serious constraint for shelf-life enhancement is the activity of micro-organisms. Drying and dehydration of fresh fruits and vegetables is one of the most energyintensive processes in the food industry and promising method of reducing postharvest losses. For long-term storage of dried fruits or vegetables, sulphuring or using a sulphite dip are the best pretreatments. Blanching process, use of chemical and chemical concentration levels, use of packaging materials to increase storability were optimised individually and developed process technology for dehydration of cauliflower. The samples


13

14


prepared with treatment combination of 3 and 9-min blanching, dipped in 0.5% and 1.0% KMS and packed in LF scored first rank, followed by PP and HDPE. References Ambrose, D.C.P. & Sreenarayanan, V.V. (1998). Studies on the dehydration of garlic. Journal of Food Science and Technology, 35, 242–244. Anon. (1990). Expert panel on food safety and nutrition. Quality of fruits and vegetables, Scientific Status Summary. Food Technology, 6, 99–106. AOAC (1970). Official Methods of Analysis, 11th edn. Washington, DC: Association of official Analysis chemists. Bajaj, M., Aggarwal, P., Minhas, K.S. & Sindhu, J.S. (1993). Effect of blanching treatments on the quality characteristics of dehydrated fenugreek leaves. Journal of Food Science and Technology, 30, 196–198. Chauhan, A.S., Ramteke, R.S. & Eipeson, W.E. (1998). Properties of ascorbic acid and its applications in food processing: a critical appraisal. Journal of Food Science and Technology, 35, 382–392. Ezell, B.P. & Wilcox, M.S. (1952). Influence of storage on carotene, total carotenoids and ascorbic acid contents of sweet potatoes. Plant Physiology, 27, 81–83. Haas, G.J., Prescott, H.E. & Cante, C.J. (1974). On rehydration and respiration of dry and partially dried vegetables. Journal of Food Science, 39, 681. Ibarz, A. & Barbosa-Canovas, G.V. (2000). Unit Operations in Food Engineering. New York, NY: CRC press. Lima, A.G.B., Queiroz, M.R. & Nebra, S.A. (2002). Simultaneous moisture transport and shrinkage during drying of solids with ellipsoidal configuration. Chemical Engineering Journal, 86, 85–93. Lutz, K., Muhlbauer, W., Muller, J. & Reisinger, G. (1987). Development of a multi-purpose solar crop dryer for arid zones. Solar and Wind Technology, 14, 451–470. Mulay, S.V., Pawar, V.N., Throat, S.S., Ghatge, U.M. & Ingale, U.M. (1994). Effect of pre-treatment on quality of dehydrated cabbage. Indian Food Packer, 48, 11–15. Pawar, V.N., Singh, N.I., Dev, D.K., Kulkarni, D.N. & Ingle, U.M. (1988). Solar drying of white onion flakes. Indian Food Packer, Jan– Feb, 15–28. Pokharkar, S.M., Kanawade, L.R. & Mahale, D.M. (1997). Process development for cauliflower dehydration. Journal of Maharastra Agriculture University, 22, 223–226.


Rahman, M.S. & Perera, C.O. (1999). Drying and food preservation. In: Handbook of Food Preservation (edited by M. Shafier Rahman). Pp. 173–216. NewYork, NY: Marcel Dekker, Inc. Raina, B.L., Pruthi, J.S., Kalra, C.L., Teotia, M.S. & Jarnail Singh, J. (1982). Influence of variety on quality of processed cauliflower. Indian Food Packer, 36, 7–15. Ranganna, S. (1986). Colour Measurement. Handbook of analysis and quality control for fruits and vegetable products (2nd edn.). Pp. 497– 528. New Delhi: Tata McGraw Hill publishing company limited. Rogacheva, S.M., Kuntcheva, M.J., Pancheva, I.N. & Obretenov, T.D. (1995). L-ascorbic acid in non-enzymatic reactions. Z. Lebensm Unters Forsch, 200, 52–58. Sadasivam, S. & Manickam, A. (1996). Biochemical Methods (2nd edn.). New Delhi: New Age Internation (P) Ltd. Publisher and TNAU, Coimbatore. Salunkhe, D.K., Bolin, H.R. & Reddy, N.R. (1991). Storage processing and nutritional quality of fruits and vegetables. In: Fresh Fruits and Vegetables (edited by D.K. Salunkhe & S.S. Kadam). Pp. 12–18. Boca Raton, FL: CRC press. Sarsavadia, P.N., Sawhbey, R.L., Pangavhane, D.R. & Singh, S.P. (1999). Drying behaviour of dried onion slices. Journal of Food Engineering, 40, 219–226. Sehgal, S. (1999). Indian Economic Data. New Delhi: Shivam offset press. Sharma, G.K., Semwal, A.D. & Arya, S.S. (2000). Effect of processing treatments on the carotenoids composition of dehydrated carrots. Journal of Food Science and Technology, 37, 196– 200. Singh, R.P. (1992). Scientific principles of shelf life evaluation. In: Shelf Life Evaluation of Foods (edited by C.M.D. Man & A.A. Jones). Pp. 3–24. Glasgow: Blackie Academic and Professional. Smoot, J.M. & Nagy, S. (1980). Effect of storage temperature and duration on total vitamin C content of canned single strength graph fruit juice. Journal of Agricultural Food Chemistry, 28, 417– 421. Srivastava, G.K. & Sulebele, G.A. (1975). Dehydration of cauliflower, effect of pretreatments on dehydration characteristics. Indian Food Packer, 29, 5–10. Watada, A.E., Aulenbach, B.B. & Worthington, J.T. (1976). Vitamins A and C in ripe tomatoes as affected by stage of ripeness at harvest and by supplementary ethylene. Journal of Food Science, 41, 856–858.



15

Original article A test procedure to characterise the heating performance of domestic microwave ovens Mark Jeffery Swain,1* Adrien Spinassou2 & Maria Veronica Lyra Swain1 1 Food Refrigeration and Process Engineering Research Centre (FRPERC), University of Bristol, Churchill Building, Langford, Bristol BS40 5DU, UK 2 Institut Universitaire de Technologie, Zac du Sanital, Rue Nobel, 86100 Chatellerault, France (Received 22 July 2005; Accepted in revised form 31 January 2006)

Summary

A test procedure has been developed and used to characterise the performance of domestic microwave ovens in relation to the heating of chilled ready meals. The procedure uses reproducible test loads, which simulate the heating and weight loss characteristics of a chilled ready meal under consumer use. The temperature distribution after heating from 5 C to a defined minimum temperature of 70 C is measured in the food simulant using a purpose designed multipoint thermocouple ‘hedgehog’ probe. The temperature, weight loss and heating time data from the test are entered into a spreadsheet analysis program, which provides a simple oven performance ‘score’ and/or comprehensive heating performance data.

Keywords

Chilled ready meals, domestic microwave, food simulant, heat distribution, microwave heating, microwave performance, temperature measurement.

Introduction

The number of households in Great Britain using a microwave oven has risen from 62% in 1993 to 89% in 2003 (The Office for National Statistics, 2004). Coupled with this, there has been substantial growth in the chilled ready meals market (70% increase in market value from 1999 to 2003; Chilled Food Association, 2005), and hence there are significant and increasing numbers of chilled ready meals being heated in microwave ovens. The only performance-related information that is provided to the consumer buying a microwave oven is its power output, which is measured using a standard 1000-g water load [International Electrotechnical Commission (IEC), 1999]. This provides little information about how well the oven actually heats a food product, such as a chilled convenience meal. In the United Kingdom, a categorisation letter corresponding to the power output measured into a 350-g water load [UK Ministry of Agriculture, Fisheries and Food (MAFF), 1992] is now used in addition to the usual power output rating. This goes some way towards informing the consumer how much power is available to heat smaller food loads, such as chilled convenience foods, but reveals nothing about how well that power is distributed *Correspondent: Fax: +44 011 79289314; e-mail: [email protected]


within the food. For example, two ovens having an identical power output may heat the same type of food load very differently; one may be prone to produce unacceptable localised ‘cold spots’, whereas the other may achieve a more uniform temperature distribution. A series of investigations funded by MAFF in the early 1990s revealed the considerable variability in the performance of domestic microwave ovens (Burfoot et al., 1991; Swain et al., 1994) and the potential for the survival of Listeria monocytogenes (Walker et al., 1991) when reheating food products. This can result in problems of both food quality and safety. Pasteurisation processes during manufacture are only designed to reduce vegetative pathogenic bacteria to safe levels. A number of factors could potentially lead to the presence of pathogenic bacteria in chilled ready meals, such as insufficient processing, post-pasteurisation contamination and consumer temperature abuse. In conventional heating, much of the bacterial destruction occurs during the time taken for the temperature to rise to its desired level. Microwave heating does not yield the same cumulative temperature–time relationship as conventional cooking methods. Guidelines for verifying microwave reheating instructions prepared by the UK Microwave Working Group (Richardson & Gordon, 1997) state that products must achieve a minimum temperature of 70 C for 2 min or an equivalent time and temperature combination. This reheating schedule is found in a number of advisory documents concerning

16

Heating performance of microwave ovens M. J. Swain et al.

the safe heating of food, and makes use of data on the heat resistance of L. monocytogenes provided by Gaze et al. (1989). The food safety criterion selected for the test described in this paper states that the minimum temperature measured in the test load 1-min postmicrowave heating should be 70 ± 2 C and must not fall below 70 ± 2 C for a further minute. This ensures that all tests reach a common end point allowing simple and meaningful comparisons between ovens. Although there are a number of more sophisticated methods of measuring temperature distributions in food or food simulants during and after microwave heating, such as fibre optic probes, infrared thermography, microwave radiometry and magnetic resonance imaging (MRI) (Knoerzer et al., 2005), the test procedure described employs an array of thermocouples. The main reasons for this were that the instrumentation is relatively low in cost, reliable, simple to use and easy to obtain (Table 1). The main purpose of this paper is to describe the test protocol and provide examples of its use to characterise the performance of seven domestic microwave ovens in relation to the heating of chilled ready meals. The protocol is based on the results of an EU (European Union)-funded collaborative project (Russell et al., 1998) under the Measurement and Testing Programme, which involved five European research, consumer testing and standards laboratories. The procedure was developed with respect to the following criteria: Reproducibility – the test must be highly reproducible, ensuring that the same test results are obtained regardless of the laboratory and/or the tester. Consumer-related – the test must be closely related to consumer use of microwave ovens for heating of chilled ready meals. Food safety – the test procedure must ensure that the temperature/time treatment undergone by the test load is sufficient to achieve the criterion set for microbiological safety. Food quality – the test needs to provide a measure of the consumer acceptability of the ‘food’ after microwave heating. Applicability – the test protocol must be equally applicable to all types of domestic microwave ovens. Materials and methods

Description of the test protocol

The flowchart in Fig. 1 outlines the sequence of events required to carry out the test for the food simulant/ package test load. In order to perform tests that meet all the requirements stated earlier, it was necessary to replace foods having inherent biological variability, with


Prepare food simulant and store at 5 ºC for 12 h minimum

Set microwave oven controls

Remove test load from fridge

Tare balance with empty tray, insulated mat and lid and measure initial weight

30 s

Place test load with lid at the centre of the oven

Heat for time estimated to reach minimum of 70 ºC at 1 min post heating

Remove test load from oven

Measure final weight with lid

30 s

Remove lid and insert multipoint temperature sensor in to test load

Record temperatures for 2 min

Adjust heating time

NO

Minimum temperature = 70 ºC after 1 min? YES Transfer data to spreadsheet

Figure 1 Flowchart outlining the sequence of events required to carry out the test procedure.

a test load having reproducible behaviour. A food simulant with similar microwave heating characteristics to a ‘slow heating’ chilled convenience meal (developed and tested by Swain et al., 2004) was used. The food simulant material was composed of TX151 powder, a hydrophilic polymer (Weatherford, Aberdeen, Scotland), potable water at 20 C and salt in a TX151:NaCl:water ratio of 22.2:0.7:77.1 by weight. The salt was dissolved in the water in an electric blender (Kenwood KM230, Havant, UK) at half speed; the blender speed was then reduced to the minimum setting while the TX151 powder was added. The speed was then set to a maximum of 20 s, just sufficient for the TX151 powder to blend in and 350 ± 1 g of the mixture was immediately poured into a rectangular crystalline poly-



ethylene terephthalate (CPET) tray (Type 1051.11; Etimex UK, Croydon, UK) having external dimensions of 171 · 127 · 35 mm. After allowing the simulant mixture to set, the tray was covered tightly with foil to minimise evaporation and placed in a refrigerator at 5 C for at least 12 h to ensure that it had equalised to 5 ± 1 C throughout. The majority of chilled ready meals are heated with a film lid in place that the consumer is instructed to pierce in several places with a fork. To simulate this in a reproducible manner, a special reusable lid with sixteen holes drilled in set positions (Fig. 2) was manufactured from polytetrafluoroethylene (PTFE). An empty tray and the PTFE lid were placed on the balance and the balance tared. The test load was then removed from the refrigerator, the foil covering removed and then put on to the balance, the PTFE lid placed on and the weight noted. It was ensured that the weight was within ±1 g of the specified weight, and that an identifying mark was made with a permanent marker on a long edge of the tray. The sample was then placed in the centre of the oven turntable with the marked long edge parallel to the oven front and the oven started. The time taken from opening the refrigerator to starting the microwave oven heating had to be 30 ± 1 s. The test load was heated on full power for a time estimated to produce a minimum temperature of 70 C in 1 min after heating. At the end of the heating time, the oven was stopped, the test load removed and the final weight (including lid)

recorded. The test load was then placed into position with the identifying mark at the front prior to the multipoint temperature ‘hedgehog’ sensor being inserted (Fig. 3). It was ensured that the insertion was 30 ± 1 s after the oven had been stopped. The temperatures were then recorded for 2 min using a PC computer-based data logging system (Datascan 7000 series, Measurement Systems Ltd., Newbury, UK). If the minimum temperature recorded 1-min after heating was 70 C (±2 C tolerance), the test was valid, and the data was transferred to the spreadsheet for analysis. If the temperature was outside this tolerance, the heating time was either increased or reduced accordingly and the test repeated with a fresh test load using a cold oven. Having achieved a valid test result, the test was repeated a further five times using the same heating time, a fresh test load and a cold oven. All tests were carried out in an air conditioned test laboratory running at 20 ± 1 C and the power to the ovens was supplied via a constant voltage stabiliser (TS3B RMS; Claude Lyons, Waltham Cross, UK) and a variable transformer (Regavolt 715-G2PE; Claude Lyons, Waltham Cross, UK) providing an input voltage of 230 V ± 1%. Each oven was operated at the maximum power setting (i.e. maximum power available from the source, without boost) with a stopwatch being used to time the oven heating from switching the oven on to switching it off/opening the oven door. Instrumentation

The most important piece of instrumentation used in the test procedure was the postheating temperature measurement system. This was a purpose built jig (‘hedgehog’) that employed an array of thirty-nine single point

Figure 2 Dimensions of polytetrafluoroethylene (PTFE) lid used for heating tests (dimensions in mm).


Figure 3 Photograph of thirty-nine-point temperature sensor ‘hedgehog’ inserted into food simulant test load after microwave heating.


17

18


fine hypodermic (1.2 mm outside diameter) stainless steel, type ‘T’ thermocouples (±0.5 C accuracy). The sensors were arranged in a matrix at four discrete layers (Fig. 4). Mathematical modelling was used to predict the sensor spacing that would, in all cases, determine an unsatisfactory temperature distribution (Russell et al., 1998). However, owing to cost and engineering limitations, it was necessary to reduce the number of probes, the positions being selected to best quantify safety and quality criteria. A data reported in a previous investigation involving replicated heating trials using food in CPET meal trays in 102 domestic microwave ovens (Burfoot et al., 1991), indicated that in over 95% of the ovens, the minimum temperature after heating was recorded in the centre position, 2 mm from the bottom of the tray. Therefore, the bottom layer of probes was biased towards the centre of the assembly to be at a height of 2 mm from the base of the tray when inserted into the food simulant. The next layer of probes (7 mm from base of tray) was arranged in a uniform matrix to provide an indication of the uniformity of heating over a larger area. A greater number of probes were biased towards the two lower layers, as in the majority of previous investigations, these have been found to detect the minimum temperatures. The next layer of probes (12 mm from base of tray) was selected to increase the number of sensors to measure the central region. The corner locations at the top layer of probes (17 mm from base of tray) are typically the hottest locations after heating and consequently experience a high degree of shrinkage and burning. Therefore, it was assumed that

measuring these points would provide an indication of the degree of edge overheating. As these points are arranged symmetrically, they would also provide an indication of the evenness of energy distribution. Sensors were also distributed towards the central region to help provide an indication of top to bottom heating balance. The probes were secured to a flat rigid plate of clear polycarbonate that was able to slide up and down on vertical rods. To measure temperatures after heating, the test load was placed horizontally in a locating fixture, and the horizontal plate with the probes attached (protruding vertically down) was quickly lowered into the food simulant. Analysis spreadsheet

The test protocol includes the use of a Microsoft Excel spreadsheet with user written macros to manipulate and present the data from tests in a user friendly way. Objective input data required for the spreadsheet includes test load heating time(s), initial (before heating) and final (after heating) weight (g) of the food simulant and temperatures (C) recorded using the multipoint thermocouple ‘hedgehog’ sensor inserted into the food load 1 and 2 min after heating and measured ‘MAFF’ power output (W). In addition, the oven make, model, IEC power output (W) and unique test reference code are entered. Once all input data are entered, the user starts the data analysis macro from the tools menu to automatically produce a four-page report, each page being in a separate Microsoft Excel worksheet. Each

12 24 4 1 .5

12 24 36 48 63.5

KEY:

Probe 2 mm from base Probe 7 mm from base Probe 12 mm from base Probe 17 mm from base


Figure 4 Location and height of temperature sensors in the food simulant contained within the polyethylene terephthalate (CPET) tray (dimensions in mm).



analysis page can be reviewed on screen by selecting the appropriate worksheet tab named ‘Layer Data (table)’, ‘Layer Chart’, ‘Analysis Table’ or ‘Oven Scores’. Figures 5–8 show samples of the data output from the analysis spreadsheet. The first two pages provide visual representations of the temperature distribution within the test load separated into four layers, corresponding to the four depths of the ‘hedgehog’ probe positions.

KEY:

Temperature < 70 °C 70–74.9 °C 75–84.9 °C 85–94.9 °C >95 °C

Layer Data (table)

The first page is the numerical temperature data 60 s after heating, rearranged into a format representing the probe positions in the food simulant. The temperature data are separated into four ‘layers’ representing the four different depths at which the probes are positioned in the food simulant. ‘Top’, ‘Up Mid’, ‘Lo Mid’ and ‘Bottom’ refer to the layers of probes 17, 12, 7 and 2 mm from the base of the food stimulant, respectively.

Top Layer 1

Up mid Layer 2

Layer Chart

The second page is a ‘three-dimensional’ representation of the temperature data to help visualise the temperature distribution within the test load 60 s after heating. The temperatures have been colour coded in bands of 5 C.

Lo mid Layer 3

Analysis Table

The third page contains a table with a detailed analysis of the test, including temperature distribution and

Bottom Layer 4

Top 1 1 86.398 2 3 4 5 6 7 87.641

1

2

3

1

2

3

1 2 3 4 5 6 7

2

3

80.339

6

7

76.333

75.628

76.161

79.452

8

4

Up mid 5 6

75.308

76.015

74.667

72.19

70.529

75.154

4

83.592

2

5

78.745 75.059

1

5

9

10

11 88.48

93.72

1 2 3 4 5 6 7

1 2 3 4 5 6 7

4

4

79.756 80.44

Bottom 6

81.984 81.201

9

10

11

9

10

11

78.821

80.22 69.606

80.621

7

8

82.397 81.634

80.703

8

76.879 81.282

5

8

7

77.41 74.38

80.591

3

Lo mid 6

7

83.889

9

10

11

76.397 80.815

83.382

76.168

Figure 5 Sample of the first report page of the analysis spreadsheet – layer data (table).


Figure 6 Sample of the second report page of the analysis spreadsheet – layer chart.

weight loss information. The oven make, model, serial number and output power are printed at the top of the page with the test reference code for identification. The minimum, maximum, range, mean and standard deviation of temperatures recorded 60 s after microwave heating are tabulated for each of the following regions within the food simulant; the four layers labelled ‘Top’, ‘Up Mid’, ‘Lo Mid’ and ‘Bottom’ as defined earlier; all the probes in the simulant, labelled ‘Overall’; all probes behind the centre line parallel to the long sides of the tray, with the marked side at the front, labelled ‘Back’; all probes forward of the centre line, labelled ‘Front’; all probes to the left of the centre line normal to the long sides of the tray, labelled ‘Left’; all probes to the right of the centre line, labelled ‘Right’; all probes in the top two layers labelled ‘Top L1 & L2’; all probes in the bottom two layers, labelled ‘Bottom L3 & L4’; all probes in the outer two cells in the ‘Layer chart’, labelled ‘Outside’ and all remaining inner probes, labelled ‘Inside’. The difference between the mean temperature of each pair of zones is also shown to help identify any asymmetrical heating effect or unusual energy deposition, and can be


19

20


Oven make Oven model Oven serial number Output power (IEC 705, 1000 g) (W) Output power (350 g water load) (W)

= = = = =

Oven E

Test Reference Code

= Er5 TX

Target minimum temperature

=

0 1065.5 905.64

70

Temperature data

Top Up mid Lo mid Bottom

L1 L2 L3 L4

Overall L1–L4

Min (°C) 75.6 70.5 69.6 76.2

Max Range Mean (°C) (°C) (°C) 93.7 18.1 83.0 76.0 5.5 74.0 83.9 14.3 78.7 83.4 7.2 80.4

SD (°C) 6.93 2.14 3.82 2.28

69.6

93.7

4.86 Min Temp < 70 °C at 60 s Min Temp = 71.3 °C at 120 s 3.97 49.67 5.72 50.33 –1.0

24.1

79.3

Split T Diff (%) (°C) 26.25 23.41 24.89 25.45

Back Front

74.7 70.5

88.5 93.7

13.8 23.2

79.3 80.3

Left Right

72.2 69.6

87.6 93.7

15.5 24.1

79.5 79.8

4.20 5.65

49.90 50.10

–0.3

Top L1 & L2 Bottom L3 & L4

70.5 69.6

93.7 83.9

23.2 14.3

79.1 79.5

7.00 3.30

49.90 50.10

–0.3

50.78 49.22

2.5

51.34 48.66

4.3

Top L1 Bottom L4 Outside Outer 2 Inside

75.6 69.6

93.7 83.4

18.1 13.8

81.6 77.4

4.85 4.02

Heating time (s) = 240.00 Heating rate (°C s–1) = 0.31 Power x Heating time (kJ) = 217.35 –1 Power x Heating time/Temp rise (kJ °C ) = 2.92 Weight loss (g) = 10.90 Weight loss (%) = 3.11

Score Speed Energy Quality Evenness Temperature Overall

7.20 6.11 8.44 6.76 5.33 6.77

(min 0%, max 20%) (min SD = 0, max SD = 15)

Figure 7 Sample of the third report page of the analysis spreadsheet – analysis table.

used to determine different indicators or ‘measures’ of temperature distribution, including top to bottom, side to side and front to back heating balance and edge overheating. Also on this page are the following data. Heating time (s) – the time used in the test to heat the food simulant. Heating rate (C s)1) – the mean temperature rise (C) of the food simulant divided by the heating time (s). Power · heating time (kJ) – output power (W), measured using UK MAFF recommended 350-g water load method, multiplied by the heating time (s) – an estimation of the energy absorbed by the oven load during microwave heating. Power · heating time/temp. rise (kJC)1) – as earlier, but divided by the mean temperature rise of the food simulant – an estimation of the energy absorbed by the load per degree of mean temperature rise. Weight loss (g) – the difference between the initial weight of the food simulant before heating and the final weight of the food simulant after heating, in the microwave oven being tested. Weight loss (%) – as earlier, but stated as a percentage of the initial weight before heating. Owing to the large amount of information that was generated, an overall ‘scoring’ system was devised. This combined results from each section of the analysis and provided a simplified score on a ten-point scale, with 0 representing the minimum and 10 the maximum rating. This has the advantage that it was easier to rank the ovens relative to each other using a single ‘score’ rather than comparing arrays of figures. It also provides an indication of how much scope is available for modifying the characteristics of the oven. Five performance attributes were scored on a ten-point scale; ‘speed’, ‘energy’, ‘quality’, ‘evenness’ and ‘temperature’, with the average of these providing an ‘overall’ score. Speed – a score indicating the relative speed of heating of the oven based on the ‘heating time’ defined earlier. The higher the number, the faster the oven heats food. Speed ¼ 10 (heating time min. htg. timeÞ ð10=½max. htg. time min. htg. timeÞ

Figure 8 Sample of the fourth report page of the analysis spreadsheet – oven scores.


where min. htg. time is set to the shortest heating time (s) expected of any oven (currently set to 100 s), and would result in a maximum score of 10 (fastest oven). Max. htg. time is set to the maximum heating time (s) beyond which the heating time (currently set at 600 s) would be considered ‘unacceptable’ for a chilled convenience meal and would score 0 (slow heating). Energy – a score based on the power output (measured into a 350-g water load) of the oven and the heating time used to heat the food simulant. The lesser the



‘energy’ used, the higher the score up to a maximum of 10. Energy ¼10 ðPH min. PHÞ ð10=½max. PH min. PHÞ where PH is ‘power · heating time’ as defined earlier. Min. PH is currently set at 50 kJ and Max. PH is set at 480 kJ. Quality – a score based on the weight loss (%) from the food simulant during microwave heating. The higher the weight loss the lower the ‘quality’ score. Quality ¼ 10 ðwt. loss min. wt. lossÞ ð10=½max. wt. loss min. wt. lossÞ where wt. loss is ‘weight loss (%)’ as defined earlier. Min. wt. loss is currently set at 0% and max. wt. loss is set at 15%. Evenness – a score based on the standard deviation (SD) of the mean of all thirty-nine temperature measurements made at 60 s after heating. The smaller the SD, the more even the temperature. Evenness ¼10 ðSD min. SDÞ ð10=½max. SD min. SDÞ where min. SD is the value of SD (currently set at 0) that will produce an ‘Evenness’ score of 10 (most even temperature) and max. is the value of SD (currently set at 15) that will produce an ‘Evenness’ score of 10 (most uneven temperature). Temperature – a score based on the mean temperature; the greater the mean temperature is above that needed to ensure meeting the microbial safety criteria of 70 C at 60 s after heating, the lower the score. A score of 10 would be achieved if all temperatures just reach 70 C, with the score reducing towards 0, the closer the mean temperature approaches 95 C. Temperature ¼ 10 ðmean temp. 70Þ ð10=½95 70Þ Overall – this is the mean of all five scores. The minimum temperature recorded 60 s after heating is checked to see whether it is 70 ± 2 C. If it is not within this range, a warning is printed in the Analysis Table 1 Details of domestic microwave ovens selected for the tests

Table stating that it is below or above range. The minimum temperature position is also checked at 120 s after heating, and a statement printed to indicate whether or not the temperature has remained above 70 ± 2 C. If the temperature has fallen below 68 C, then there is a possibility of a ‘cold spot’ that has not been detected directly by the probe. The macro also determined if the minimum temperature requirement had been satisfied. If the minimum temperature 1 min after heating was not equal to 70 C (±2 C), then it was indicated that the test was not valid; the minimum temperature was either too low or too high, the heating time was incremented or decremented accordingly and another test was carried out. However, if the test was a replicate test, it was not necessary to satisfy the minimum temperature requirement. The correct heating time had already been determined and remained the same for all tests. In addition, the minimum temperature was determined for the following for 1 min to ensure that it did not fall below 68 C. Oven scores

The fourth page is a bar chart indicating the scores achieved by the oven (given numerically in the Analysis Table). Each bar represents a different characteristic of the oven’s performance (labelled speed, energy, quality, evenness, temperature and overall), having met the set criteria of the test protocol. Results and discussion

Table 2 shows the mean data obtained from six replicated tests for the seven microwave ovens used to derive the heating performance scores. The mean measured power output into a 350-g water load (MAFF power output) ranged from 743 (oven A) to 919 W (oven C). Only one oven (oven A) had a power output that fell within the specified range for a category E microwave (741–800 W), and was the same as that claimed by the manufacturer. The six other ovens were all above 800 W, and therefore above category E, as claimed by the manufacturer.

Oven code

Manufacturer’s stated IEC power output (W)

Manufacturer’s stated MAFF heating category

Cavity volume (L)

Type of oven

Type of turntable

A B C D E F G

800 800 900 800 1000 800 850

E E E E E E E

14.8 20.0 32.0 18.3 30.2 20.0 23.0

Combination Microwave + grill Combination Microwave + grill Combination Microwave solo Microwave solo

Metal Glass Glass Glass Glass Glass Glass

IEC, International Electrotechnical Commission; MAFF, Ministry of Agriculture, Fisheries and Food.



21

22


Oven Measured MAFF Measured MAFF Heating SD of simulant Mean simulant Weight code power output (W) heating category time (s) temperatures (°C) temperature (°C) loss (%) A B C D E F G

743 879 919 848 906 909 828

E >E >E >E >E >E >E

540 310 320 285 240 240 315

12.7 6.0 7.2 5.0 4.7 5.5 6.8

76.2 82.7 84.0 84.6 79.7 82.6 84.1

Heating times determined to reach the minimum temperature required for a valid test (70 ± 2 C) ranged from 240–320 s for the six ovens with valid test results. Oven A was unable to heat the test load to the required minimum temperature before exceeding the maximum weight loss (15%) acceptable for a valid test result. The mean weight loss of 22% resulted in a high degree of shrinkage of the food simulant test load. Consequently, some of the temperature measurement probes were either too close to the surface of the simulant or in air, yielding unreliable results. Therefore, oven A was judged to have an unacceptable heating performance as defined by the criteria of the test procedure. The percentage weight loss of the food simulant during cooking was used as an indicator of product quality (the greater the weight loss, the greater the likelihood of irreversible degradation of product quality). The standard deviation (SD) of the thirty-nine temperatures in each test load measured 1 min after heating ranged from 4.7 (oven E) to 7.2 C (oven C) for the six ovens with valid heating performance test results. These values were subsequently used as indicators of how evenly each microwave oven heated the test load. The mean of the thirty-nine measured temperatures in each test load measured 1 min after heating ranged from 79.7 (oven E) to 84.6 C (oven D) for the six ovens with valid heating performance test results. Mean weight loss during heating ranged from 3.4% (oven E) to 10.8% (oven C) for the six ovens with valid heating performance test results. Figure 9 shows a plot of the mean and SD of six replicates of the heating performance scores (speed, energy, quality, evenness, temperature and overall), derived using the analysis spreadsheet for the six ovens that met the test criteria. As oven A failed to heat the test load to the required minimum temperature before exceeding the maximum weight loss limit (15%) required for a valid test, it was rated unsatisfactory. Out of the other six ovens, oven C had the lowest overall score (4.5) and oven E the highest (6.7). The analysis also provided extra information about the distribution of temperatures throughout the food simulant test load allowing the tester to better understand the behaviour of each individual oven.


22.6 10.2 10.8 7.2 3.4 5.1 8.9

0

1

Table 2 Measured Ministry of Agriculture, Fisheries and Food (MAFF) power output (W) and heating category, simulant heating time (s), SD of thirty-nine measured simulant temperatures (C), mean of thirty-nine measured simulant temperatures and weight loss during microwave heating (%). All values are means of six replicates

2

3

4

5

6

7

8

9

10

Speed Energy

B C D E F G

Quality Evenness Temperature Overall

Figure 9 Plot showing mean performance scores of six microwave ovens (oven codes B to G). Oven code A is omitted as it failed to reach the requirements of a valid test. Error bars are ±1 SD.

Figures 5–8 show sample report pages of the data output from the analysis spreadsheet, which allow the tester to easily assess the performance of each oven. Conclusion

A test procedure to characterise the heating performance of domestic microwave ovens in relation to chilled convenience meals has been developed and successfully used to test seven recent models. The results show that the test procedure can clearly identify microwave ovens with unsatisfactory heating performance. The test can also be used to identify the relative performance of ovens with acceptable heating performances. The uncertainty of using foods having inherent biological variability as the test material has been overcome by employing a food simulant with reproducible behaviour as the test load. Acknowledgments

The authors would like to acknowledge the support of the EU and input from the collaborating partners that helped to develop the original microwave oven test procedure (Contract no. MATI-CT 940014) and Which?, Consumer’s Association, London, UK for providing microwave ovens and support for the investigations.



References Burfoot, D., Foster, A.M., Self, K.P., Wilkins, T.J. & Phillips, I.C. (1991). Reheating in Domestic Microwave Ovens: Testing Uniformity and Reproducibility. UK Ministry of Agriculture, Fisheries and Food (MAFF) Microwave Science Series Report No. 3. Chilled Food Association (2005). Chilled foods market data spreadsheet. Available at: http://www.chilledfood.org/downloads/ UK%20Chilled%20Food%20Market%201999–2003.xls (accessed 1 March 2005). Gaze, J.E., Brown, G.D., Gaskell, D.E. & Bank, J.G. (1989). Heat resistance of Listeria moncytgenes in homogenates of chicken, beef steak and carrot. Food Microbiology, 6, 251–259. International Electrotechnical Commission (1999). IEC Publication 60705:1999. Household Microwave Ovens – Methods for Measuring Performance. Geneva: International Electrotechnical Commission. Knoerzer, K., Regier, M. & Schubert, H. (2005). Measuring temperature distributions during microwave processing. In: The Microwave Processing of Foods (edited by H. Schubert & M. Regier). Pp. 243–263. Cambridge: Woodhead Publishing Limited. ISBN-13: 978-1-85573–964-2. Richardson, P.S. & Gordon, K.H. (eds.) (1997). Guidelines on the Verification of Reheating Instructions for Microwaveable Foods – Guideline No 14. Campden and Chorleywood Food Research Association, Gloucestershire. ISBN: 090594206X.


Russell, S.L., Swain, M.J., Fernicola, V. et al. (1998). Determination of Unsatisfactory Temperature Distributions within Foods Heated in Microwave Ovens, MATI-CT 940014 Programme. EUR 18545 EN Report. Pp. 8–79. ISSN 1018–5593. Swain, M.J., Foster, A.M., Phillips, I.C. & James, S.J. (1994). Variation of Uniformity of Microwave Reheating within Commercial Chilled Convenience Meals. UK Ministry of Agriculture, Fisheries and Food (MAFF) Microwave Science Series Report No. 10. Swain, M.V.L., Russell, S.L., Clarke, R.N. & Swain, M.J. (2004). The development of food simulants for microwave oven testing. International Journal of Food Science and Technology, 39, 623– 630. The Office for National Statistics (2004). Consumer durables, central heating and cars: 1972 to 2003: General Household Survey 2003. Available at: http://www.statistics.gov.uk/StatBase/Expodata/ Spreadsheets/D8744.xls (accessed 1 March 2005). UK Ministry of Agriculture, Fisheries and Food reference document (MAFF) (1992). A voluntary system for the categorisation of domestic microwave ovens for heating small food loads and for the consequent labelling of domestic microwave ovens and microwaveable foods: Guidance note (revised October 1992). Walker, S.J., Bows, J., Richardson, P. & Banks, J.G. (1991). Listeria survival in chilled retail products: effects of recommended microwave cooking. Microwave Science Series Report No. 1. London: MAFF Publications.


23


24

Original article Chemical composition and toxic trace element composition of some Nigerian edible wild mushrooms Olumuyiwa S. Falade,* Oluwatoyin O. Adepoju, Olanrewaja Owoyomi & Steve R. Adewusi Department of Chemistry, Obafemi Awolowo University, Ile-Ife, Nigeria (Received 11 October 2005; Accepted in revised form 11 July 2006)

Summary

Two essential amino acids (methionine and tryptophan); anti-nutritional factors (tannin and trypsin inhibitor) and toxic elements (Pb, Cd, Ni, As, Hg and Cr) were determined spectrophotometrically from five edible wild mushrooms. The tryptophan content was between 1.00 and 1.82 g (100 g))1 but methionine was low at 0.26–1.38 g (100 g))1. Tannin content was high (30.3–40.0 mg g)1) but trypsin inhibitor was low (22.0–39.5 TIU g)1). Trace elements analysis reviled Pb (0.34–5.06 mg kg)1) to be the highest of all the trace elements. Cd was (0.06–1.70 mg kg)1), Ni (0.26–2.08 mg kg)1), As (0.17–0.92 mg kg)1), Hg (0.01– 0.05 mg kg)1) and Cr (0.04–0.22 mg kg)1). These mushrooms are nutritious but must be well processed to eliminate or at least reduce the levels of tannin and Pb to improve their nutritional values.

Keywords

Methionine, tannin, trace elements, trypsin inhibitor, tryptophan, wild mushroom.

Introduction

Mushrooms are vast becoming popular food items in several parts of the world because of better information about their nutritional values and for their appetising flavour. For example, mushrooms have been reported to grow on nutritionally valueless substances such as wood stump and sawdust and thus have the capacity to convert these waste products into high protein food (Chang & Hayes, 1978). In addition to nutritional and economic benefits, mushrooms are also considered for their medicinal and religious medicinal importance. The aqueous extract of Coriolus versicolour, for instance, has been reported to display a wide range of biological activities including stimulatory effects on different immune cells and inhibition of cancer growth (Chu et al., 2002). Pleurotus ostreatus was observed to lower cholesterolemia by almost 40% while some mushrooms are useful for diabetic and heart patients (Bobek et al., 1991; Fukushima et al., 2000). Furthermore, the power to ward off the evil spirit has been ascribed to Phalius aurantioca (Oso, 1977). Adewusi et al. (1993) investigated the nutritional values of some Nigerian edible wild mushrooms and observed a net protein retention (NPR) of 3.8 for Chlorophyllum molybditis and as low as 0.8 for Termitomyces robustus. The former performed better than casein while the latter could hardly maintain the protein *Correspondent: E-mail: [email protected]

status of the animals. Although, mushrooms are rich in protein and some other important nutrients (Silva et al., 2002), the low utilisation of three of the six samples investigated by Adewusi et al. (1993) was presumed to be due to the presence of anti nutritional and toxic compounds. Plants and animals used by man for food have been reported to contain chemical substances known to be toxic or which may reduce the bioavailability of other nutrients (Adewusi & Falade, 1996; Chung et al., 1998). Tannin and Trypsin inhibitors are known to inhibit the digestive enzymes and hence reduce feed conversion and growth (Chung et al., 1998). Trace elements, whether essential or non-essential, above some threshold concentration levels can cause morphological abnormalities, reduce growth and increase mortality and mutagenic effects in humans (Pier & Bang, 1980). Most of the chemical analyses on Nigeria wild mushrooms have been limited to proximate composition. Hence the present study was designed to look at the chemical composition of some selected mushrooms with the aim of explaining the differences in biological values, weight lost and toxicity observed earlier in the animals fed mushroom diets in our laboratory. Materials and methods

Materials

Leutinum subnudus and C. molybditis were harvested inside the zoological garden of Obafemi Awolowo

doi:10.1111/j.1365-2621.2006.01375.x 2007 The Authors. Journal compilation 2007 Institute of Food Science and Technology Trust Fund

Chemical composition of some Nigerian wild mushrooms O. S. Falade et al.

University, Ile-Ife Nigeria, while Psathyrella atroumbonata (Pegler), Termitomyces robustus (Beeli) Heim and Termitomyces eurrhius were bought in a local market in Ile-Ife. Mushroom samples were cleaned in the laboratory and dried at 50 C to a constant weight in a Gallenkamp oven (Model SA 9059 B, Loughborough, UK). The dried samples were milled in a ball grinder (Siebtechnik TS 250, Molheim-Ruhr, Germany) in to fine flour (mesh size number 72) and stored in plastic containers inside a refrigerator at about 4 C until used. Certified Reference Materials NIES no. 9 (Sagasso) and NIES no. 10 (rice flour unpolished) were obtained from National Institute for Environmental Studies (NIES), Japan Environment Agency, Onogawa 16-2, Tsukuba, Ibaraki, 305 Japan. Analytical procedure

Moisture content was determined according to the AOAC (1984) method. The modified vanillin-hydrochloric acid (MV-HCl) method of Price et al. (1978) was used for tannin determination using catechin as the standard. Trypsin inhibitor was determined by the method of Kakade et al. (1974) as modified by Adewusi & Osuntogun (1991). Each sample (0.2 g) was extracted with 10 mL 0.01 m NaOH for 1 h, centrifuged at 4000 ·g for 30 min and diluted tenfold. Different aliquots of this extract were added to 5 mL benzoyl-DL-arginine-pnitroanilide hydrochloride (BAPNA) solution for hydrolysis with 2 mL of 0.2 mg mL)1 trypsin (Sigma Type II, St. Louis, MO, USA) in 0.0001 m HCl serving as the standard. The reaction was stopped after 10 min with 1.0 mL of 30% acetic acid and the absorbance read at 410 nm. Methionine was determined by the method of Gehrke & Neuner (1974) with some modifications while tryptophan was determined by the procedure developed by De Vries et al. (1980) also with modifications as described below.

added followed by 0.33 mL of 0.5% (w/v) sodium nitroprusside and the solution mixed properly and allowed to stand for 6 min at room temperature. Finally, 1.67 mL of 8.0 m H3PO4 was added and mixed thoroughly and the absorbance of the resultant coloured solution was read at 526 nm and compared with a calibrated curve from the standard solutions. Tryptophan determination

The tryptophan content of the hydrolysates and standard tryptophan solutions was determined spectrophotometrically as described earlier (De Vries et al., 1980). p-dimethylaminobenzaldehyde (DAB) reagent (8.0 mL) was pipetted into test tubes 2.0 mL of each hydrolysates obtained from above was added. The mixture was mixed thoroughly, stoppered and placed in the dark for 5 h; 0.048% (w/v) sodium nitrite (0.1 mL) was added and mixed thoroughly. The mixture was left for 30 min for colour development. The absorbance of the samples, standards and blank (mixture of 2.0 mL of hydrolysate and 8.0 mL of 10.8 m H2SO4) was read at 590 nm. Trace element content was determined from 0.5 g of each samples as well as 0.3 g of NIES certified reference materials [Rice flour (NIES no. 10) and Sargasso (NIES no. 9)] weighed in triplicate. 10 mL concentrated HNO3 was added to each sample in a digestion flask and allowed to stand overnight. The samples were heated carefully until the production of brown nitrogen (IV) oxide fume has ceased. The flasks were cooled and 2–4 mL of 70% perchloric acid was added. Heating was continued until the solutions turned colourless. The solutions were transferred into 50 mL standard flasks and diluted with distilled water. The trace element content was then analysed by Alpha 4 model Atomic Absorption Spectrophotometer (Fisons Chem-Tech, Analytical, Kempston, Bedford, UK). Statistical analysis

Enzyme hydrolysis

Samples that contained 25 mg protein were weighed into 10 mL standard flasks. Two millilitre of each standard solution (0–100 and 0–70 mg L)1) for tryptophan and methionine respectively was also pipetted into 10 mL volumetric flask. One millilitre of pronase (4.0 mg mL)1) was later added to the sample and the standard as well as a flask containing 2 mL of water to be used as a pronase blank. The enzyme hydrolysis was continued as described earlier (De Vries et al., 1980) and the hydrolysates made up to 10 mL with 0.03 m phosphate buffer, pH 7.5. Methionine determination

Each of the hydrolysates (2.0 mL) obtained above was pipetted into test-tubes, 0.88 mL of 3.0 m NaOH was

The results were expressed as a mean of three determinations. Data were subjected to one way analysis of variance and the correlation analysis by using GraphPad InStat version 3.06 (Graphpad Software Inc., EL Camino, San Diego, USA) for Windows 2003. Results and discussion

Moisture content of the mushrooms used in this study ranged between 81.4% and 93.2% (Table 1). The high moisture content of these mushrooms could account for their low shelf-life. Mushrooms are very difficult to preserve at room temperature. The technique used locally for preserving mushroom involves the reduction of the moisture content by sun drying or smoking.

2007 The Authors. Journal compilation 2007 Institute of Food Science and Technology Trust Fund

International Journal of Food Science and Technology 2008

25

26


Sample

Moisture* (%)

P. Atroumbonata T. robustus L. subnudus C. molybditis T. eurrhius

92.3 93.2 81.4 88.1 89.4

± ± ± ± ±

0.7a 2.1a 0.9b 2.0c 0.1c

Tannin* (mg g)1) 33.3 40.0 28.3 32.0 30.3

± ± ± ± ±

0.6a 2.7b 0.6c 0.9ad 0.8d

Trypsin Inhibitor† (TIU g)1)

Tryptophan* [g (100 g))1 sample]

22.0 35.5 24.4 39.0 39.5

1.07 1.82 1.75 1.05 1.00

± ± ± ± ±

0.15a 0.01b 0.16b 0.06a 0.11a

Methionine* [g (100 g))1 sample] 0.36 0.26 1.38 0.45 0.60

± ± ± ± ±

Table 1 Chemical composition of mushrooms expressed on dry weight basis

0.03ab 0.06a 0.11d 0.08bc 0.06c

Values in the same column with the same superscripts are not significantly different at the 5% probability level. *Mean ± SD of triplicate analysis. † Mean of two determinations.

Anti-nutritional factors

Tannin content of these mushrooms varied between 28.3 and 40.0 mg g)1 (Table 1). The range of tannin obtained for these samples was relatively high compared with a range of 1.4–4.6 mg g)1 reported for vegetables (Falade et al., 2004) and 1.6–29.7 mg g)1 reported for young cassava leaves (Awoyinka et al., 1995). Tannin could account for the lost in weight observed earlier for T. robustus (Table 2; Adewusi et al., 1993). Incidentally, this mushroom recorded the highest tannin content. Tannins are known to retard growth through reduced digestion and/or absorption (Laurena et al., 1984). Adewusi et al. (1993) had earlier suspected the presence of tannin in this sample and presumed it to be responsible for the observed index of protein quality (Table 2). This suspicion seemed confirmed with the highly significant negative correlations obtained between tannin and the indexes of protein quality obtained earlier (protein efficiency ratio (PER) ¼ )0.97, NPR ¼ )0.96 and in vitro digestibility ¼ )0.47 at P < 0.05). Trypsin inhibitor was between 22.0 and 39.5 TIU g)1. This factor differs significantly among the samples with the exception of C. molybditis and T. eurrhius with 39.0 and 39.5 TIU g)1, respectively. Trypsin inhibitor obtained for these samples was low when compared with a range of 15 000–23 000 TIU g)1 and 6700– 23 300 TIU g)1 reported for Phaseolus vulgaris and cowpea, respectively (Elias et al., 1979; Adewusi & Osuntogun, 1991). This means that trypsin inhibitor is not likely to pose any problem in the utilisation of these mushrooms as food especially when trypsin inhibitor Table 2 Result obtained from animal fed mushroom diets

Sample

N consumed

PER

NPR

% in vitro digestibility

P. atroumbonata T. robustus C. molybditis

5.72 7.67 8.52

1.5 ± 0.3 )0.2 ± 0.4 2.6 ± 0.9

2.5 ± 0.4 0.8 ± 0.4 3.8 ± 0.9

86.5 87.6 92.1

Source: Adewusi et al. (1993).


can be inactivated by heat treatment such as steaming and cooking (Liener, 1994). The fact that this factor may not be relevant in determining the protein quality of these samples was evident from the highest PER and NPR reported for C. molybditis (Adewusi et al., 1993) despite the highest trypsin inhibitor (39.0 TIU g)1) obtained for it. The low correlation factor between trypsin inhibitor and PER (r ¼ 0.07) and NPR (r ¼ 0.14) seemed to confirm the irrelevance of trypsin inhibitor in the utilisation of mushrooms as food. Tryptophan and methionine

The results of these two essential amino acids are also presented in Table 1. Tryptophan ranged between 1.00 and 1.82 g (100 g))1 sample. Tryptophan content of P. atroumbonata, C. molybditis and T. eurrhius were not significantly different (P < 0.05) but significantly different from those of T. robustus and L. subnudus. Tryptophan is one of the essential amino acids often limiting in the diets of developing countries. This amino acid becomes more important when its sparing role for niacin is taken into consideration. Tryptophan can be converted to niacin. Tryptophan content of the mushroom samples was higher than 0.4 g (100 g))1 reported for Phycomyces blakesleeanus (Shlomai et al., 1992). They were also higher than the [0.214–0.298 g (100 g))1 sample] values reported for Vicia faba, cowpea and chick peas (FAO, 1982). These mushrooms will provide more than the recommended tryptophan requirement for preschool-age child which is 11 mg g)1 crude protein (FAO/WHO/UNU, 1985). This shows that these mushrooms are probably superior to many legumes currently consumed in Nigeria and most developing countries. Inclusion of edible mushrooms into human diets would likely improve the tryptophan status of the diets and hence improve the protein quality of the diets. The methionine range was between 0.26 and 1.38 g (100 g))1. Methionine is the limiting essential amino acid in legumes a major component of the diet of developing countries. A major stress is furthermore placed on the little quantity of methionine available in



ND 0.07 0.22 0.08 0.12 (0.08) ND 0.004 0.003 0.005 0.004 (0.001)


0.17 0.35 0.29 0.27 (0.00) (0.001) (0.02) (0.95)

ND 0.19 0.39 0.30 0.29

(0.03) (0.04) (0.03) (0.10)

VB CRM

Ni

1.30 (0.06)

ND 0.023 (0.003) 0.32 (0.02) 1.82 (0.06) 0.72 (0.96) 1.35 (0.05) ND ND ND 1.35 (0.05) NIES no. 9 NIES no. 10a NIES no. 10b NIES no. 10c Mean ± SD

1.30 (0.06)

CRM CRM Reference materials

Pb

VB

Cd

VB

0.02 0.03 1.79 0.70

)1

Table 3 Trace element content of certified reference materials (lg g )*

The knowledge of the levels of toxic trace elements in plant foods is necessary because of their effects on human health. Mushrooms have been reported to accumulate several trace elements at greatly exceeding contents than in other plant foods (Kalac et al., 2004). In addition, ‘tree killers’ containing toxic trace elements could have been used on the tree trunks on which some mushrooms grew. The reliability of the method used for trace element determination was estimated by taking the reference materials through the same procedure used for the samples. The results obtained for each element are in good agreement with the values reported for these reference materials (Table 3). The trace elements content of the samples presented in Table 4 indicated that Pb ranged between 0.34 and 5.06 mg kg)1. Cd was between 0.06 and 1.70 mg kg)1, Ni (0.26–2.08 mg kg)1), As (0.17–0.92 mg kg)1); Hg (0.01–0.05 mg kg)1) and Cr (0.04–0.22 mg kg)1). Of all the trace elements in these samples, Hg was the lowest while Pb was the highest. The Pb content of the samples analysed was within the range of 2.7–6.5 mg kg)1 reported earlier for 24 species of mushrooms (Yilmaz et al., 2003). Lead has been reported to cause irreversible damage to the central nervous system and permanent mental retardation (Chisolm, 1965). All the samples with the exception of C. molybditis were higher in Pb compared with 2.35 mg kg)1 obtained for Agaricus bitorquis (Quel.) Sacc. (Tuzen et al., 1998).

(0.01) (0.03) (0.04) (0.09)

Trace element

Values in the same column with the same superscripts are not significantly different at the 5% probability level. The value in parenthesis is standard deviation. *Mean ± SD of triplicate analysis. CRM, certified reference materials; VB, value obtained.

120 (5) 0.24 (0.08) 0.15 (0.03) 0.17 (0.04) 30.1 (59.9) 115 (9) 0.17 0.11 0.15 28.9 (57.4)

CRM CRM

As

VB

Hg

VB

0.007 0.007 0.007 0.007

(0.002) (0.003) (0.002) (0.000)

CRM

Cr

VB

0.05 0.25 0.06 0.12

(0.005) (0.04) (0.02) (0.11)

the diet because of its utilisation in the detoxification of cyanide consumed as part of cassava. The methionine content of these samples was lower than 1.7 g (100 g))1 reported for P. blakesleeanus (Shlomai et al., 1992) with the exception of L. Subnudus which was similar to it. The levels of this essential amino acid in the analysed samples with the exception of L. subnudus agreed with 0.46 g (100 g))1 obtained for P. ostreatus – an edible mushroom (Hadar & Cohen-Arazi, 1986). All the samples will provide at least half of the recommended requirement for methionine/cystine of 25 mg g)1 crude protein for preschool-age child (FAO/WHO/UNU, 1985). Of the three mushroom used earlier for nutritional study (Adewusi et al., 1993) excerpted in Table 2; C. molybditis with the highest PER and NPR was incidentally the sample with the highest methionine. This could account in part for the better performance of the rats on this mushroom. O. S. Falade (2004, unpublished PhD thesis) observed the highest PER of 2.0 for rats fed Acacia colei supplemented with 0.4% methionine followed by 0.4% cystine supplemented (PER ¼ 1.2). This shows that mushroom could complement other legume plant foods.


27

28


Sample

Pb

P. atroumbonata T. robustus L. subnudus C. molybditis T. eurrhius

4.05 2.53 5.06 0.34 4.72

Cd ± ± ± ± ±

0.01a 0.24c 0.43ab 0.01d 0.02b

1.70 0.69 0.87 0.06 0.66

Ni ± ± ± ± ±

0.13a 0.21ab 0.30b 0.01c 0.08b

0.52 2.08 1.91 0.78 0.26

As ± ± ± ± ±

0.01a 0.17b 0.09b 0.01c 0.01d

0.64 0.92 0.92 0.60 0.17

Hg ± ± ± ± ±

0.04a 0.20a 0.20a 0.01a 0.01b

0.04 0.30 0.05 0.01 0.02

Table 4 Trace element Content of mushrooms (mg g)1) expressed on dry weight basis*

Cr ± ± ± ± ±

0.02a 0.01a 0.02a 0.01a 0.01a

0.22 0.11 0.04 0.10 0.20

± ± ± ± ±

0.03a 0.02b 0.01c 0.03bc 0.02a

Values in the same column with the same superscripts are not significantly different at the 5% probability level. *Mean ± SD of triplicate analysis.

The acceptable daily intake of Pb for adults was (0.21–0.25 mg day)1; FAO/WHO, 1993). This shows that too much consumption of these mushrooms with the exception of C. molybditis could lead to Pb body burden and this could be a cause for concern. Soaking of mushrooms in 0.3% table salt and subsequent boiling for 15 min was observed to reduce Pb, Hg and Cd by as mush as 74.1%, 85.1% and 65.6%, respectively (Svoboda et al., 2002). This is similar to the technique used locally for the processing of mushroom before consumption, but the aim, here, is to remove maggots and other microbial contaminants. Cadmium has been associated with renal damage; cancer and childhood aggression [International Agency for Research on cancer (IARC), 1993]. The Cd content of these mushrooms was lower than the range of 0.26–2.0 mg kg)1 reported earlier for 24 species of mushrooms (Yilmaz et al., 2003) and 2.5–5.5 mg kg)1 reported for Armillaria mellea – another mushroom (Kalac et al., 2004). The consumption of these mushrooms is not likely to cause Cd body burden judging from the acceptable daily intake of 0.06–0.07 mg day)1 (FAO/WHO, 1993). Cadmium poisoning will occur only if about 41 g on dry weight basis of the sample with the highest Cd level is consumed. Soaking of these mushrooms in salt water will also help in lowering the Cd content (Svoboda et al., 2002). Nickel has been linked to lung cancer (Yen, 1999). The range of Ni obtained in this study was lower than 0.05–5 mg kg)1 reported for plant foods (National Academy of Sciences, 1975) and 0.87–19.0 mg kg)1 reported earlier for 24 micro fungi species (Yilmaz et al., 2003). Although, one may be tempted to indicate that the samples are safe from Ni point of view, the tolerable upper intake level of 1 mg day)1 reported for this toxic element [Food and Nutrition Board (FNB), 2001] is also low. Termitomyces robustus (Beeli) Hein and L. subnudus may therefore cause unsafe accumulation of Ni in human body. Arsenic (As) has been reported to be carcinogenic (IARC, 1987). It produces its toxicity by binding with tissue sulphydryl groups and also to enzymes of the Krebs cycle, hence interferes with oxidative phosphorylation (Greaf, 1994). The toxic level of As was put at


0.35 mg day)1 (Underwood, 1971). This implies that about 380 g on dry weight basis of the sample with the highest As content have to be consumed before the toxic effects of As can be apparent. The implication of this is that none of these samples is likely to cause As poisoning. Mercury (Hg) poison was reported to cause foetal malformation, liver and kidney damage (Weiss & Landrigan, 2000). The range reported in this study for Hg was lower than 0.014–14.0 mg kg)1 obtained earlier for thirteen species of wild mushrooms (Falandysz et al., 2002). The range was also lower than 1.0–20.0 mg kg)1 reported for different species of Agaricus campestris and Agaricus arvensis (Kalac et al., 2004). The joint conference of UNEP/ILO/WHO (1991) set a tolerable level of 0.029 mg day)1 for Hg. To exceed this guideline, 580 g on dry weight basis of mushroom with the highest Hg must be consumed which is not likely. Hence it may be safe to say that none of these mushrooms is likely to cause Hg poisoning. Chromium (Cr), unlike the other elements analysed, is considered essential to man because of its ability to increase glucose tolerance in type-2 diabetes mellitus patients (Anderson, 2000). The recommended dietary intake for Cr is 0.035 mg day)1 for male and 0.025 mg day)1 for the female (FNB, 2001). The levels of Cr obtained for these samples with the exception of L. Subnudus revealed that mushroom is a good source of this essential element. In conclusion, these mushrooms are rich in the two essential amino acids analysed and also in Cr. Tannin and Pb are the two main factors that could limit their utilisation for food but the problem can be overcome by adequate processing such as soaking in salt water and proper cooking before consumption. References Adewusi, S.R.A. & Falade, O.S. (1996). The effects of cooking on extractable tannin, phytate, sugars and mineral solubility in some improved Nigerian legume seeds. Food Science and Technology International, 2, 231–240. Adewusi, S.R.A. & Osuntogun, B.A. (1991). Effect of cooking on tannin content, trypsin inhibitor activity and in vitro digestibility of some legume seeds in Nigeria. Nigerian Food Journal, 9, 139–145.



Adewusi, S.R.A., Alofe, F.V., Odeyemi, O., Afolabi, O.A. & Oke, O.L. (1993). Studies on some edible wild mushrooms from Nigeria: 1. Nutritional, teratogenic and toxic considerations. Plant Foods for Human Nutrition, 43, 115–121. Anderson, R.A. (2000). Chromium in the prevention and control of diabetes. Diabetes and Metabolism, 26, 22–27. AOAC (1984). Official Methods of Analysis. 14th edn. Arlington, VA: AOAC. Awoyinka, A.F., Abegunde, V.O. & Adewusi, S.R.A. (1995). Nutrient Content of young cassava leaves and assessment of their acceptance as a green vegetable in Nigeria. Plant Foods for Human Nutrition, 47, 21–28. Bobek, P., Ginter, E., Jurcovicova, M. & Kuniak, L. (1991). Cholesterol-lowering effect of the Mushroom Pleurotus ostreatus in hereditary hypercholesterolemic rats. Annals of Nutrition and Metabolism, 35, 191–195. Chang, S.T. & Hayes, W.A. (1978). The Biology and Cultivations of Edible Mushrooms. London: Academic Press, INC. Chisolm, J.J. (1965). Chronic lead intoxication in Children. Developmental Medicine and Child Neurology, 7, 529–536. Chu, K.K., Ho, S.S. & Chow, A.H. (2002). Coriolus versicolour: a medicinal mushroom with promising immunotherapeutic values. Journal of Clinical Pharmacology, 42, 976–984. Chung, K., Wong, T.Y., Wei, C., Huang, Y. & Lin, Y. (1998). Tannins and human health. Critical Reviews in Food Sciences and Nutrition, 38, 421–464. De Vries, J.W., Koski, C.M., David, C.E. & Paul, A.L. (1980). Comparison between a spectrophotometric and a high-pressure liquid chromatography method for determining tryptophan in food products. Journal Agricultural and Food Chemistry, 28, 896–898. Elias, L.G., De Fernandez, D.G. & Bressani, R. (1979). Possible effects of seed coat polyphenolics on the nutritional quality of bean protein. Journal of Food Science, 44, 524–526. Falade, O.S., Dare, A.F., Bello, M.O., Osuntogun, B.O. & Adewusi, S.R.A. (2004). Varietal changes in proximate composition and the effect of processing on the ascorbic acid content of some Nigerian vegetables. Journal of Food Technology, 2, 103–108. Falandysz, J., Gucia, M., Skwarzec, B., Frankowska, A. & Klawikowska, K. (2002). Total mercury in mushrooms and underlying soil substrate from the Borecka Forest, northeastern Poland. Archives of Environmental. Contamination and Toxicology, 42, 145–154. FAO (1982). Food Composition Tables for the Near East. Rome: Food and Agriculture Organization of the United Nations. FAO/WHO (1993). Evaluation of Certain Food Additives and Contaminants. WHO Technical Report series 837. Geneva: FAO/WHO. FAO/WHO/UNU (1985). Energy and Protein Requirements. Report of a joint FAO/WHO/UNU Expert Consultation. WHO Tech. Rep. Ser. No. 724. Geneva: WHO. Food and Nutrition Board (FNB) (2001). Dietary Reference Intakes for vitamin A, vitamin K, Arsenic, Boron, chromium, copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Pp. 1–28. Washington, DC: Institute of Medicine, National Academy Press. Fukushima, M., Masuo, N., Yasuko, M., Tetsu, O., Yukiko, F. & Kei, S. (2000). Hepatic LDL receptor mRNA in rats is increased by dietary mushroom (Agaricus bisporus) fibre and sugar beet fibre. Journal of Nutrition, 130, 2151–2156. Gehrke, C.W. & Neuner, T.E. (1974). Automated chemical determination of methionine. Journal Association of Official Analytical Chemists, 57, 682–688. Greaf, J.W. (1994). Heavy metal poisoning. In: Harrison’s Principle of Internal Medicine (edited by K.J. Isselbacher, J.D. Wilson,

E. Braunwald, J.B. Martin, A.C. Fauci & M.D. Kaspen). 13th edn. Pp. 2461–2466. New York, NY: McGraw-Hill. Hadar, Y. & Cohen-Arazi, E. (1986). Chemical composition of the edible mushroom Pleurotus ostreatus produced by fermentation. Applied and Environmental Microbiology, 51, 1352–1354. International Agency for Research on cancer (IARC) (1987). Arsenic and its compounds. IARC Monogram, 7, 100–106. International Agency for Research on cancer (IARC) (1993). IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol. 58. Beryllium, Cadmium, Mercury, and Exposures in the Glass Manufacturing Industry. Lyon: International Agency for Research on Cancer. Kakade, M.L., Rackis, J.J., Mcghee, J.E. & Puski, G. (1974). Determination of trypsin inhibitor activity of soy products: a collaborative analysis of an improved procedure. Cereal Chemistry, 51, 376–383. Kalac, P., Svoboda, L. & Havlickova, B. (2004). Content of cadmium and mercury in edible mushrooms. Journal of Applied Biomedicine, 2, 15–20. Laurena, A.C., Van Den, T. & Menoza, E.M.T. (1984). Effects of condensed tannins on in vitro digestibility of cowpea (Vigna unguiculata (L.) Walp.). Journal of Agricultural and Food Chemistry, 32, 1045–1048. Liener, I.E. (1994). Implications of antinutritional components in soybean foods. Critical Reviews in Food Science and Nutrition, 34, 31–67. National Academy of Sciences (1975). Nickel. p. 227. Washington, DC: National Academy of Sciences. Oso, B.A. (1977). Pletorus tuber-regium from Nigeria. Mycologia, 69, 271–279. Pier, S.M. & Bang, M.K. (1980). Environment and health. In: Ann Arbor Science (edited by N.M. Trieff). p. 367. Michigan, MI: The Butherworth Group. Price, M.L., Van Scoyoc, S. & Butler, L.G. (1978). A critical evaluation of the vanillin reaction as an assay for tannin in sorghum grains. Journal of Agricultural and Food Chemistry, 26, 1214–1216. Shlomai, P., Margalith, P. & Mokady, S. (1992). Nutritional evaluation of the fungus Phycomyces blakesleeanus as a protein source. Journal of Science of the Food and Agriculture, 58, 125–128. Silva, S.O., Costa, S.M.G. & Clement, E.E. (2002). Chemical composition of Pleurotus pulmonarius (Fr.) Quel., substrates and residue after cultivation. Brazilian Archives of Biology and Technology, 45, 531–535. Svoboda, L., Kalac, P., Spicka, K. & Janouskova, D. (2002). Leaching of cadmium, lead and mercury from fresh and differently preserved edible mushroom, Xerocomus badius, during soaking and boiling. Food chemistry, 79, 41–45. Tuzen, M., Ozdemir, M. & Demirbas, A. (1998). Study of heavy metals in some cultivated and uncultivated mushrooms of Turkish origin. Food Chemistry, 63, 247–251. Underwood, E.J. (1971). Trace Elements in Human and Animal Nutrition, 3rd edn. New York, NY: Academic Press. UNEP/ILO/WHO (1991). International Programme on Chemical Safety. Geneva: Environmental Health Criteria 118 ‘Inorganic Mercury’, WHO. Weiss, B. & Landrigan, P.J. (2000). The developing brain and the environment. An introduction. Environmental Health Perspectives, 108, 373–374. Yen, T.F. (1999). Environmental Chemistry: Essentials of Chemistry for Engineering Practice. Vol. 4. New Jersey, NJ: Prentice Hall Inc. Yilmaz, F., Isiloglu, M. & Merdivan, M. (2003). Heavy metal levels in some macrofungi. Turkish Journal of Botany, 27, 45–56.



29


30

Original article Changes in physical and sensory characteristics of marinated broiler drumsticks, treated with nisin and lactoperoxidase system Fa-Jui Tan1* & Herbert W. Ockerman2 1 Department of Animal Science, National Pingtung University of Science and Technology, Pingtung, 91201, Taiwan, ROC 2 Department of Animal Sciences, The Ohio State University, Columbus, OH 43210, USA (Received 12 April 2005; Accepted in revised form 15 June 2006)

Summary

The lactoperoxidase system (LPS) and nisin have been shown to inactivate some micro-organisms in foods. However, further studies were needed to evaluate whether these treatments had any influence on the physical and sensory characteristics of broiler drumsticks. In this study, a solution that contained 1% acetic acid and 3% salt with pH adjusted to 4 was developed as a standard marinade. The LPS consisted of 1-lg mL)1 lactoperoxidase, 5.9-mm potassium thiocyanate (KSCN), and 2.5-mm H2O2 was added to the marinade, followed by nisin at 100 IU mL)1. The results showed that the physical characteristics, including raw and cooked drumstick pH, percentage incorporation of marinade solution, cooking loss, skin and muscle L*, a*, b* values, and sensory characteristics, including skin and muscle sensory colour, aroma and flavour, off aroma, off flavour, juiciness and tenderness of the broiler drumsticks, treated with 100-IU mL)1 nisin and LPS were not impaired.

Keywords

Lactoperoxidase system, marination, nisin, physical characteristics, sensory characteristics.

Introduction

Marination is a procedure of treating meat with an aqueous mixture of vinegar, salt and spices before cooking. Marination of poultry is practised to improve the product’s physical and sensory attributes, such as tenderness, water-holding capacity, and flavour (Young & Lyon, 1997a,b; Hashim et al., 1999a,b; Lemos et al., 1999; Xiong & Kupski, 1999; Young & Buhr, 2000; Zheng et al., 2000; Young & Smith, 2004). Nisin, which is considered a ‘biopreservative’, is a natural, non-toxic, and heat-stable polypeptide, produced by Lactococcus lactis, and has been shown to inhibit gram-positive micro-organisms (Ray, 2001) and gram-negative bacteria, when combined with chelating agents, such as EDTA (Stevens et al., 1992). Even though several studies involving nisin in poultry products have been reported, only a few ones have evaluated the effects of nisin on the physical and sensory characteristics of meat (Usborne et al., 1986; Rozbeh et al., 1993). The lactoperoxidase system (LPS), which consists of lactoperoxidase (LP), thiocyanate (SCN)), and hydrogen peroxide (H2O2), is an inhibitory system that is present naturally in bovine milk. This system has been shown to be inhibitory against some pathogenic and spoilage microorganisms, and has been studied mainly for the appli*Correspondent: Fax: +886-8-774-0148; e-mail: [email protected]

cation in milk and dairy products (Reiter & Harnulv, 1984). Even though LPS has been reported to have little effect on the sensory and physical characteristics of treated milk and dairy products (Martinez et al., 1988; Ridley & Shalo, 1990), limited studies involving the effect of LPS on the qualities of treated meat and poultry products have been reported. The objective of this study was to investigate the effects of adding nisin and LPS on some physical and sensory characteristics of marinated broiler drumsticks. Materials and methods

Development of the marinade solutions

A solution containing 1% acetic acid, 3% salt, and 20mm disodium EDTA (Fisher Scientific Co., Kansas City, MO, USA) was prepared as the standard marinade in this study. Nisin, prepared from commercial nisin powder (Sigma Chemical Co., St. Louis, MO, USA) at 100 IU mL)1, was added to the standard marinade. An LPS consisting of 1-lg mL)1 lactoperoxidase (EC 1.11.1.7; purity index 0.82; Sigma Chemical Co.), 5.9-mm KSCN (Fisher Scientific Co., Pittsburgh, PA, USA), and 2.5-mm H2O2 (30%, Fisher Scientific Co.) was added to the marinade for the LPS-added treatment. The pH of the marinade solutions was then adjusted to 4. For the sensory evaluation, 0.3% black pepper and 0.15% garlic powder were added to the

doi:10.1111/j.1365-2621.2006.01378.x Ó 2007 Institute of Food Science and Technology Trust Fund

Drumsticks’ qualities on treatment with nisin and LPS F. J. Tan and H. W. Ockerman

standard marinade solution. No flavouring agents were added for the physical evaluation. Sample preparation

A total of ninety-six drumsticks were purchased from a local supermarket; placed in an insulated container; transported to the laboratory within 30 min; and then stored in a 4°C walk-in cooler until the experiment was conducted. Drumsticks were marinated in plastic containers with marinade solution, so that all the drumsticks could be covered completely by the marinade solution and marinated at 4°C for 18 h. pH value measurement of drumsticks

Muscles of the whole drumsticks without bone and skin were blended for 30 s, using a Waring Lab Blender (Model 31BL91; Dynamics Corporation of America, New Hartford, CT, USA). Ten grams of muscle sample was mixed with 100 mL of distilled water in a polyethylene bag for 1 min, using a Seward Stomacher (Model 400; Tekmar Company, Cincinnati, OH, USA), and the pH of the mixtures was measured. Percentage incorporation of marinade solution

After marinating, the drumsticks were removed from the marinade solutions, and drained for 5 min at 4°C. Percentage incorporation of marinade solution was calculated as follows. Percentage incorporation of marinade solution ¼ ½ðWtafter marinating and draining Wtbefore marinating Þ= Wtbefore marinating 100

then the cooking loss measurement and sensory evaluation were accomplished. Cooking loss was calculated as follows. Cooking loss ¼ ½ðWtbeforecooking Wtaftercooking Þ= Wtbeforecooking 100 Sensory evaluation

Descriptive analysis was conducted by a seven-member trained panel to evaluate the intensities of the sensory characteristics of the raw and cooked samples. Analysis included the evaluation of skin colour, muscle colour, marinated broiler aroma (combined raw marinated broiler aroma consisted of broiler aroma, acidic, peppery, and garlic aromas), and off aroma (any aroma which was not expected from the raw marinated broiler) for the raw samples. After cooking, the samples were cooled to room temperature (approximately 25°C), and served to the sensory panel. The same characteristics, as for raw broilers, were also evaluated for cooked samples. In addition, juiciness and tenderness were evaluated. The sensory evaluation was conducted using a 1 to 9 scale, with 1 representing the lowest intensity and 9 the highest intensity, for all attributes except colour (1 ¼ light colour; 9 ¼ dark colour), juiciness (1 ¼ not juicy; 9 ¼ very juicy), and tenderness (1 ¼ not tender; 9 ¼ very tender). Statistical analyses

Three trials were conducted in this study. There were four samples for each treatment. Data were analysed using the general linear model (GLM) of Statistical Analysis System’s Procedures (SAS Institute Inc., Cary, NC, USA) with a 5% level of significance. Means were separated using Duncan’s multiple range test.

Instrumental colour measurement

The Hunter ‘L*’ (lightness), ‘a*’ (redness), and ‘b*’ (yellowness) values of drumstick were measured with a Chroma Meter (model CR-300; Minolta Co., Ltd. Ramsey, NJ, USA), on the surfaces of the skin, and on the muscle without skin covering, before and after marinating, and after cooking. Three measurements were made at different locations on the drumsticks and averaged. A standard plate (number 18433010) with ‘Y’ ¼ 92.60, ‘x’ ¼ 0.3140, and ‘y’ ¼ 0.3206 was used as a reference. Cooking loss measurement

Preweighed drumsticks were cooked in a 177°C preheated conventional oven (Type EF111; The G.S. Blodgett Co. Inc., Burlington, VT, USA) to an internal temperature of 70°C, equilibrated to room temperature, and

Ó 2007 Institute of Food Science and Technology Trust Fund

Results and discussion

Physical evaluation of marinated broiler drumsticks with or without the addition of nisin and/or LPS is shown in Table 1. After marinating for 18 h at 4°C, the pH values of raw drumsticks ranged from 5.23 to 5.43, without significant difference. After cooking, the pH values of the samples increased without significant difference. There were no significant (P < 0.05) differences for the percentage incorporation of marinade solution, and cooking loss among the marinated treatments 1 through 4. Table 2 illustrates the colour evaluation of the marinated broiler drumsticks, with or without the addition of nisin and with or without LPS. The skin and muscle L* values increased after marinating, and decreased after cooking. After marinating, samples with LPS added (treatments 2 and 4) had significantly


31

32


Table 1 Physical evaluations of marinated* broiler drumsticks with or without the addition of nisin and/or lactoperoxidase system (LPS)

Table 2 Skin and muscle colour evaluation* results of marinated† broiler drumsticks with or without the addition of nisin and/or lactoperoxidase system (LPS)

Treatment No. Treatment No. Parameter Nisin addition (100 IU mL)1) LPS addition† Raw drumstick pHà (after marinating) Cooked drumstick pHà Percentage incorporation of marinade solution (%) Cooking loss (%)

1

2

3

4 Parameter

) ) 5.33a 5.64a 1.05a

) + 5.28a 5.51a 0.86a

+ ) 5.23a 5.53a 1.48a

+ + 5.43a 5.74a 0.48a

27.05a

27.21a

25.69a

25.29a

a, b Means, within a row, without the same superscript are significantly different (P < 0.05). *Marination: broiler drumsticks were marinated with a solution which contained 1% acetic acid, 3% salt, and was adjusted to pH 4. †LPS: 1-lg mL)1 lactoperoxidase, 5.9-mM KSCN, and 2.5-mM H2O2. àDrumstick pH was measured for muscle without skin.

(P < 0.05) lower skin and muscle a* values, when compared with the samples without LPS (treatments 1 and 3). Similarly, after cooking, the skin and muscle a* values of treatments 2 and 4 were significantly lower than the values of treatments 1 and 3. This reduction of a* values of the LPS-treated samples was probably the result of the addition of H2O2, which is one of the components of the LPS treatment, and is a strong oxidising agent that is occasionally used as a bleaching agent in the food industry (Daeschel & Penner, 1992). The significant changes of skin a* values of the LPStreated samples (treatments 2 and 4) in this study did not agree with the report by Wolfson (1992), where immersing broiler legs and thighs in a 50°C bath containing the LPS of 1-lg mL)1 lactoperoxidase, 5.9-mm KSCN, and 2.5-mm H2O2 for 5 min, and then storing at 4°C for 24 and 48 h did not significantly (P > 0.01) affect the Hunterlab colour values (L, a, b) of the treated samples. The possible reason of this disagreement between the two studies is probably that the LPS-treated time (18 h in this study) was much longer than the time of 5 min in Wolfson’s study. In addition, under the experimental condition of this study (4°C), the enzyme activity was comparatively lower and the amount of H2O2 remnant was higher, which probably favoured the oxidative action and colour changes encountered when comparing the condition of 50°C in the Wolfson’s study. No significant differences were observed for the skin and muscle b* values, either after marinating or after cooking. Based on a 1–9 scale, in which 1 and 9 represent the lighter and darker colour respectively, the marinated samples in the current study had sensory raw skin colour scores between 2.0 and 3.0, and raw muscle colour between 3.7 and 4.7, without significant differences (Table 3). The results of sensory colour evaluation of


1

Nisin addition (100 IU mL)1) LPS additionà Skin L* value Before marinating After marinating After cooking a* value Before marinating After marinating After cooking b* value Before marinating After marinating After cooking Muscle L* value Before marinating After marinating After cooking a* value Before marinating After marinating After cooking b* value Before marinating After marinating After cooking

2

) )

3 ) +

4

+ )

+ +

72.63ab 84.13a 69.66a

70.25bc 81.21a 65.97a

71.16bc 81.28a 65.93a

70.02c 84.39a 66.23a

5.76a 0.80a 1.32a

4.83a )1.65b 0.42b

5.03a 0.32a 1.18a

5.15a )1.86b 0.40b

8.67a 11.49a 23.24a

9.49a 12.36a 24.41a

9.58a 10.78a 23.43a

8.49a 12.96a 24.57a

62.09a 79.32ab 75.70a

60.19a 77.07b 75.55a

60.37a 80.95ab 76.08a

58.00a 88.20a 77.21a

12.42a 3.69a 2.34a

13.58a )0.80b 0.35b

10.54a 3.17a 2.06a

10.38a )2.71b 0.33b

12.42a 12.42a 17.19a

13.58a 14.51a 19.64a

10.54a 11.61a 17.99a

10.38a 15.1a 19.25a

a, b, c Means, within a row, without the same superscript are significantly different (P < 0.05). *Colour evaluation: L* ¼ light and dark, a* ¼ red and green, and b* ¼ yellow and blue, measured with a Chroma Meter (model CR-300; Minolta Co., Ltd. Ramsey, NJ, USA). †Marination: broiler drumsticks were marinated with a solution which contained 1% acetic acid, 3% salt, and was adjusted to pH 4. àLPS: 1-lg mL)1 lactoperoxidase, 5.9-mM KSCN, and 2.5-mM H2O2.

treated samples owing to the addition of nisin agreed with previous studies by other researchers. Rozbeh et al. (1993) pointed out that refrigerated vacuum-packaged beef, treated with CO2, N2, and 500-IU g)1 nisin had similar sensory colour scores when compared with the samples without nisin during storage at 3°C for 8 weeks. Usborne et al. (1986) also reported that there was no significant (P > 0.05) difference for colour and preference ratings between the pork bellies, either cured with nitrite only or cured with nitrite and 2000 or 4000IU mL)1 nisin. In the current study, no significant difference of the marinated broiler aroma among treatments was observed. Low (1.1 to 1.6, based on a 1–9 scale) and



Table 3 Sensory characteristic intensities* of marinated† broiler drumsticks with or without the addition of nisin and lactoperoxidase system (LPS) Treatment No. Parameter

1

2

3

4

Nisin addition (100 IU mL)1) LPS additionà Raw samples Skin colour Muscle colour Marinated broiler aroma§ Off-aroma– Cooked samples Skin colour Muscle colour Marinated broiler flavour** Off flavour†† Juiciness Tenderness

) )

) +

+ )

+ +

2.7a 3.0a 5.6a 1.6a

3.0a 2.4a 5.3a 1.1a

2.0a 2.4a 4.9a 1.4a

2.6a 2.1a 6.1a 1.4a

4.7a 3.9a 5.4a 1.3a 4.3a 4.7a

4.1a 4.1a 4.0b 1.3a 3.3a 4.1a

4.0a 3.0a 5.0ab 1.1a 4.3a 4.4a

3.7a 2.9a 5.7a 1.3a 3.9a 3.9a

a, b

Means, within a row, without the same superscript are significantly different (P < 0.05). *The sensory evaluation was conducted using a 1–9 scale, with 1 representing the lowest intensity and 9 the highest intensity, for all attributes except for colour (1 ¼ light, 9 ¼ dark), juiciness (1 ¼ not juicy, 9 ¼ very juicy), and tenderness (1 ¼ not tender, 9 ¼ very tender). †Marination: broiler drumsticks were marinated with a solution which contained 1% acetic acid, 3% salt, and was adjusted to pH 4. àLPS: 1-lg mL)1 lactoperoxidase, 5.9-mM KSCN, and 2.5-mM H2O2. §Marinated broiler aroma: a combination aroma consisted of broiler, acidic, salty, peppery, and garlic aromas. –Off-aroma: any aroma that was not expected from the raw marinated broiler aroma. **Marinated broiler flavour: a combination flavour consisted of broiler, acidic, salty, peppery, and garlic flavours. ††Off flavour: any flavour that was not expected from the cooked marinated broiler flavour.

non-significant difference for the sensory off-aroma scores for the raw samples were obtained. After cooking, there was no significant difference for the sensory cooked skin and muscle colour scores of all treatments. Similarly, there was no significant difference for the marinated broiler flavours among the cooked samples. Low (1.1–1.6, based on a 1–9 scale) and non-significant difference for the sensory off-flavour scores of the cooked samples for all treatments were obtained in the current study. No significant differences for the sensory juiciness and tenderness scores were detected for the marinated samples. The results of sensory evaluation of treated samples, owing to the addition of nisin, agreed with previous studies by other researchers. Several studies involving the effect of adding nisin on the sensory quality of treated fish products have been reported. Nilsson et al. (1997) reported that a treatment which consisted of 1000-IU g)1 nisin and packed in 60%


CO2 and 40% N2 did not negatively influence the sensory characteristics of cold-smoked salmon samples. After evaluating the intensities of several sensory attributes, including smoked odour, fishy odour, fishy flavour, sweet taste, salty taste, and elasticity of the samples, Nykanen et al. (2000) found that injecting 4000–6000-IU mL)1 nisin did not alter the sensory characteristics of the treated smoked rainbow trout samples. No negative influences on the sensory characteristics of cold-smoked fish samples, owing to the addition of nisin were also reported (Nykanen et al., 1998, 1999). Conclusion

It was concluded that the treatment consisting of 100-IU mL)1 nisin and LPS (1-lg mL)1 LP, 5.9-mm KSCN, and 2.5-mm H2O2), did not impair the physical characteristics, which include raw and cooked drumstick pH, percentage incorporation of marinade solution, cooking loss, skin and muscle L*, a*, b* values, and sensory characteristics, including skin and muscle sensory colour, aroma and flavour, off aroma, off flavour, juiciness and tenderness of the marinated broiler drumsticks. References Daeschel, M.A. & Penner, M.H. (1992). Hydrogen peroxide, lactoperoxidase systems, and reuterin. In: Food Biopreservatives of Microbial Origin (edited by B. Ray & M. Daeschel ). Pp. 155–175. Boca Raton, FL: CRC press, Inc. Hashim, I.B., Mcwatters, K.H. & Hung, Y.C. (1999a). Quality enhancement of chicken baked without skin using honey marinades. Poultry Science, 78, 1790–1795. Hashim, I.B., Mcwatters, K.H. & Hung, Y.C. (1999b). Marination method and honey level affect physical and sensory characteristics of roasted chicken. Journal of Food Science, 64, 163–166. Lemos, A.L.S.C., Nunes, D.R.M. & Viana, A.G. (1999). Optimization of the still-marinating process of chicken parts. Meat Science, 52, 227–234. Martinez, C.E., Mendoza, P.G., Alacron, F.J. & Garcia, H.S. (1988). Reactivation of the lactoperoxidase system during raw milk storage and its effect on the characteristics of pasteurized milk. Journal of Food Protection, 51, 558–561. Nilsson, L., Huss, H.H. & Gram, L. (1997). Inhibition of Listeria monocytogenes on cold-smoked salmon by nisin and carbon dioxide atmosphere. International Journal of Food Microbiology, 38, 217–227. Nykanen, A., Lapvetelainen, A., Hietanen, R.M. & Kallio, H. (1998). The effect of lactic acid, nisin whey permeate, sodium chloride and related combinations on aerobic plate count and the sensory characteristics of rainbow trout. Lebensmittel Wissenschaft und Technologie, 31, 286–290. Nykanen, A., Lapvetelainen, A., Hietanen, R.M. & Kallio, H. (1999). Applicability of lactic acid and nisin to improve the microbiological quality of cold-smoked rainbow trout. Zeitschrift-fuer-LebensmittelUntersuchung-und-Forschung-A, 208, 116–120. Nykanen, A., Weckman, K. & Lapvetelainen, A. (2000). Synergistic inhibition of Listeria monocytogenes on cold-smoked rainbow trout by nisin and sodium lactate. International Journal of Food Microbiology, 61, 63–72.


33

34


Ray, B. (2001). Food biopreservatives of microbial origin. In: Fundamental Food Microbiology, 2nd edn (edited by B. Ray ). Pp. 205–209. Boca Raton, FL: CRC Press. Reiter, B. & Harnulv, G. (1984). Lactoperoxidase antibacterial system: natural occurrence, biological functions and practical applications. Journal of Food Protection, 47, 724–732. Ridley, S.C. & Shalo, P.L. (1990). Farm application of lactoperoxidase treatment and evaporative cooling for the intermediate preservation of unprocessed milk in Kenya. Journal of Food Protection, 53, 592–597. Rozbeh, M., Kalchayanand, N., Field, R.A., Johnson, M.C. & Ray, B. (1993). The influence of biopreservatives on the bacterial level of refrigerated vacuum packaged beef. Journal of Food Safety, 13, 99–111. Stevens, K.A., Sheldon, B.W., Klapes, N.A. & Klaenhammer, T.R. (1992). Effect of treatment conditions on nisin inactivation of Gramnegative bacteria. Journal of Food Protection, 55, 763–766. Usborne, W.R., Collins-Thompson, D.L. & Wood, D.S. (1986). Sensory evaluation of nisin-treated bacon. Canadian Institute of Food Science and Technology Journal, 19, 38–40. Wolfson, L.M. (1992). Inhibition of Salmonella typhimurium by the lactoperoxidase system in a broth system and on poultry. Master thesis. Pp. 122–150. Lincoln: University of Nebraska.


Xiong, Y.L. & Kupski, D.R. (1999). Time-dependent marinade absorption and retention, cooking yield, and palatability of chicken filets marinated in various phosphate solutions. Poultry Science, 78, 1053–1059. Young, L.L. & Buhr, R.J. (2000). Effect of electrical stimulation and polyphosphate marination on drip from early-harvested, individually quick-frozen chicken breast fillets. Poultry Science, 79, 925–927. Young, L.L. & Lyon, C.E. (1997a). Effect of postchill aging and sodium tripolyphosphate on moisture binding properties, color, and Warner-Bratzler shear values of chicken breast meat. Poultry Science, 76, 1587–1590. Young, L.L. & Lyon, C.E. (1997b). Effect of calcium marination on biochemical and textural properties of pre-rigor chicken breast meat. Poultry Science, 76, 197–201. Young, L.L. & Smith, D.P. (2004). Effect of vacuum on moisture absorption and retention by marination broiler fillets. Poultry Science, 83, 129–131. Zheng, M., Detienne, N.A., Barnes, B.W. & Wicker, L. (2000). Tenderness and yields of poultry breast are influenced by phosphate type and concentration of marinade. Journal of the Science of Food and Agriculture, 81, 82–87.



35

Original article Purification and characterisation of antioxidative peptides from unfractionated rice bran protein hydrolysates Abayomi P. Adebiyi,1* Ayobamitale O. Adebiyi,2 Tomohisa Ogawa3 & Koji Muramoto3 1 Department of Food Science and Engineering, Ladoke Akintola University of Technology, PMB 4000, Ogbomoso, Oyo State, Nigeria 2 Department of Pure and Applied Biology, Ladoke Akintola University of Technology, PMB 4000, Ogbomoso, Oyo State, Nigeria 3 Department of Biomolecular Science, Graduate School of Life Sciences, Tohoku University, Sendai 981-8555, Japan (Received 12 April 2006; Accepted in revised form 2 June 2006)

Summary

Crude rice bran protein (CRBP) was prepared by alkaline extraction and then treated with 0.6 m HCl to remove phytic acid. The phytate-free rice bran protein (PFRBP) was hydrolysed with proteases M, N, S, P and pepsin under optimal conditions. Hydrolysates obtained from various hydrolysis periods were subjected to analysis for their degree of hydrolysis (DH) and functional properties. The hydrolysates were fractionated by reversed-phase column chromatography on Kaseigel ODS resin (120–140 lm) using a stepwise gradient of aqueous ethanol, and their activities were measured. The 40% ethanol fraction of protease P 4 h-hydrolysate was separated by successive reversed-phase high-performance liquid chromatography and the amino acid sequences of isolated antioxidative peptides were determined by a protein sequencer and matrix-assisted laser desorption ionisation-time of flight mass spectrometry. Crude rice bran protein had higher antioxidative activity than PFRBP, due to the presence of phytic acid. Phytate contents of rice bran, CRBP and PFRBP were 2.5%, 1.42% and 0%, respectively. The activity of PFRBP increased upon protease digestion. Protease M hydrolysates showed the highest DH, but the lowest antioxidative activity. Hydrolysates with DH below 10% had higher antioxidative activity than those above 20%. This result indicates that the antioxidative activity of the hydrolysates is inherent to their characteristics amino acid sequences of peptides depending on the protease specificities.

Keywords

Antioxidative activity, phytic acid, purification, peptides, rice bran protein.

Introduction

Protein hydrolysates have been employed to provide nutrients for individuals who experience difficulties in the digestion of intact protein (Mannheim & Cheryan, 1990). Oligopeptides, which are produced during food processing or hydrolysis of proteins, are also known to possess highly potent biological activities. Numerous beneficial biological activities for human health, such as moderation of hypertension, bacteriolysis, antioxidation of fat and enhancement of calcium adsorption in digestive tract, have been found in peptides derived from food proteins (Akahoshi et al., 2000). The functional and immunological properties of proteins can be improved by partial hydrolysis and thus the hydrolysed proteins can be used in food systems as additives for beverage and infant formulae, as food texture enhancers or as pharmaceutical ingredients *Correspondent: Fax: +234-803-9401181; e-mail: [email protected]


(Mannheim & Cheryan, 1990). Compared with acid or alkali hydrolysis, enzymatic hydrolysis of proteins provides milder process conditions and little or no undesirable side reactions or products. In addition, the final hydrolysates, after neutralisation, contain less salt and the functionality of the final product can be controlled by selection of specific enzymes and reaction conditions (Deslie & Cheryon, 1988). Free radicals and singlet oxygen are recognised as major factors causing various diseases, such as cancer, cardiovascular disorders and diabetes (Lin & Chang, 2004). Therefore, the health maintenance function of antioxidant components in various foods has received much attention in recent years. In most cases, the majority of the antioxidant activities may be from compounds, such as flavonoids, isoflavone, flavones, anthocyanin and catechins rather than from vitamins C, E and b-carotene (Kahkonen et al., 1999). Several studies have analysed the antioxidant potentials of a wide variety of vegetables, herbs and spices, teas, etc. (Cuvelier et al., 1994; Maruta et al., 1995; Chuda et al.,

36

Antioxidative peptides from rice bran protein A. P. Adebiyi et al.

1996; Tada et al., 1996; Azuma et al., 1999). The antioxidant activity of polyphenols is mainly because of their redox properties, which allow them to act as reducing agents, hydrogen donors, singlet oxygen quenchers and metal chelators (Osman et al., 2004). The emergence of natural extracts possessing antioxidant properties will help in reducing the current dependency on synthetic antioxidants in food applications. Lipid peroxidation is the biological damage caused by free radicals that are formed under oxidative stress. In lipid-containing model systems, the antioxidative effect can be explained in part by their radical scavenging activity because the hydrogen-donating ability of samples does not necessarily indicate activity in lipid model systems (Ahn et al., 2004). It can be assumed that several factors, such as metal chelating properties, interactions with an emulsifier or proteins, and the distribution between the oil and water phases are more complex in lipid systems and may be important for the antioxidant action (Schwarz et al., 2000). The antioxidative activity of natural substances is due to the presence of active compound present in them. The antioxidant present in the sample may have different functional properties such as reactive oxygen species scavenging, inhibition of the generation of free radicals, chain breaking activity and metal chelation. The antioxidative activities either of amino acids or peptides have been investigated to gain insight into the antioxidative mechanism of protein hydrolysates. Structure-activity relationship of dipeptides consisting of alanine, tyrosine, histidine and methionine at the N-terminus on linoleic acid were investigated (Chen et al., 1995). The dipeptides showed higher antioxidative activities than the constituent amino acid mixtures in an aqueous system. Histidine-containing peptides have been known to be antioxidative. For example, the antioxidative properties of carnosine, which is b-alanyll-histidine found in animal skeletal muscle, have been extensively studied (Chen et al., 1998). The antioxidant mechanism of the peptide has been postulated to be metal chelation or free radical scavenging. The objectives of this investigation are to assess the antioxidative activity of enzymatic hydrolysates derived from rice bran proteins and characterise the purified peptides. Materials and methods

Preparation of rice bran protein

Full-fat rice bran (Hitomebore, Japanese rice variety) was purchased from a local rice store in Sendai, Japan. Rice bran was defatted by extracting twice with three volumes of hexane. Defatted rice bran (200 g) was stirred in 2 L of 0.05 m NaOH for 1 h. The slurry was centrifuged at 4000 · g, 4 C for 15 min. The sediment


was discarded and the supernatant was collected. The pH of the supernatant was adjusted to 4.0 and allowed to rest for 1 h at 4 C. The precipitate was washed twice with distilled water by centrifuging at 4000 · g, 4 C for 10 min. The pH of the suspension was adjusted to 7.0 before freeze-drying. The resulted crude rice bran protein (CRBP) was stored at 5 C. Crude rice bran protein (10 g) was dispersed in 500 mL of 0.6 m HCl for 30 min, and then centrifuged at 4000 · g, 4 C for 10 min. The supernatant was discarded and the sediment was washed twice with distilled water by recentrifuging. The resulted phytatefree rice bran protein (PFRBP) was neutralised, lyophilised and stored at 5 C until use. The antioxidative activities of PFRBP and CRBP were examined. Determination of phytic acid in rice bran proteins

Phytate was determined by the method of Latta & Erskin (1980) with a slight modification. Sample (0.2 g) was extracted with 4 mL of 0.6 m HCl for 2 h with stirring. The solution was centrifuged at 3000 · g for 30 min and diluted five times with distilled water. Phytic acid was separated by applying 1 mL of the diluted solution through a Dowex-1 ion-exchange column (1.5 · 10 cm) previously washed with 15 mL of 0.7 m NaCl. The column was washed out with 15 mL of 0.1 m NaCl. The retained phytic acid was eluted with 15 mL of 0.7 m NaCl and 3 mL of the obtained elute was mixed with 1 mL of Wade reagent (0.03 g ferric chloride 6 H2O and 0.3 g of sulphosalicilic acid in 100 mL of distilled water). The absorbance of each solution was read by a spectrophotometer at 500 nm. The calibration curve was prepared with standard solution of phytic acid. Enzymatic hydrolysis of rice bran protein

Phytate-free rice bran protein (2 g) was dissolved in 200 mL of distilled water and hydrolysed with 20 mg of a protease (protease M, N, S, P and pepsin) under optimal condition for each protease (Chen et al., 1995). Hydrolysed samples were collected at 1, 2, 4, 8 and 24 h. Enzymes were inactivated by heating the samples in boiling water for 3 min. The hydrolysates were neutralised and centrifuged at 2000 · g for 10 min. The supernatants were lyophilised and stored at room temperature until use. The degree of hydrolysis (DH) of hydrolysates was determined using the o-phthaldialdehyde (OPA) method of Church et al. (1983). The OPA reagent was prepared by combining the following reagents and diluting to a final volume of 50 mL with distilled water: 25 mL of 100 mm sodium tetraborate, 2.5 mL of 20% SDS (w/w) and 40 mg of OPA (dissolved in 1 mL of methanol) and 100 lL of b-mercaptoethanol. This reagent was prepared daily. A small aliquot of hydrolysates (usually 10–50 lL) was



added directly to 2 mL of OPA reagent in a cuvette, the solution was mixed briefly and incubated for 2 min at ambient temperature. The absorbance at 340 nm was measured by a spectrometer. The percentage of DH was calculated by using the equation:

defined as the induction period. The induction period refers to the relative antioxidative activity of the samples. Fractionation of active peptides

where Mw is the averaged molecular weight of amino acids, 120, D340nm the absorbance at 340 nm, d the dilution factor, 1/41, e the constant value, 6000 m)1 cm)1 and P is the mg mL)1, 0.1%.

Protein hydrolysates (20 g each) were dissolved in 400 mL of distilled water and centrifuged (12 000 · g, 4 C, 30 min). The supernatant was subjected to a column (4.0 · 16.5 cm) of Kaseigel ODS (120–140 lm) and eluted with a stepwise gradient of ethanol solution (0–80%). Yield and activity of each fraction were measured.

Antioxidative activity

Separation of antioxidative peptides

Antioxidative activity of protein hydrolysates was determined using the ferric thiocyanate method as described by Chen et al. (1995). Samples dissolved in 1.5 mL of 0.1 m sodium phosphate buffer (pH 7.0) and 1.0 mL of 50 mm linoleic acid in ethanol (99.5%) were mixed in test tubes (5 mL volume). The tubes were sealed tightly with silicon rubber caps and kept at 60 C in the dark. At regular intervals, aliquots of the reaction mixtures were withdrawn with a microsyringe for measurement of the oxidation using the ferric thiocyanate method with a slight modification. To 50 lL of the reaction mixture 2.35 mL of 75% ethanol, 50 lL of 30% ammonium thiocyanate and 50 lL of 20 mm ferrous chloride solution were added in 3.5% HCl. After 3 min, the absorbance of the reaction mixture at 500 nm was measured in a 1-cm cuvette by a DU 530 spectrophotometer (Becman, Kyoto, Japan). The number of days taken to attain an absorbance of 0.3 was

Protein hydrolysates (200 mg) were separated by reversed-phase high-performance liquid chromatography (RP-HPLC) on a preparative column ODS 10 (20 · 250 mm) using a linear gradient of acetonitrile

ðMwD340nm Þ=ðd e PÞ 100;

Figure 1 Dose-dependent antioxidative activity of rice bran protein (RBP). : Crude rice bran protein, h: Phytate-free RBP. The activity was measured by the ferric thiocyanate method. The results are shown as relative activities by adjusting the control to 1.0 and are the average of two experiments.


Figure 2 Hydrolysis and antioxidative activity of phytate-free rice bran protein (PFRBP). (a) Degree of hydrolysis (DH) of PFRBP. : protease M at pH 3, 50 C, n: protease N at pH 7, 55 C, : protease S at pH 8, 70 C, ): protease P at pH 8, 45 C, : pepsin at pH 2, 37 C. Sample dilution factor was 41 by dissolving 50 lL of sample in 2 mL of OPA reagents. (b) Antioxidative activity was determined by the ferric-thiocyanate method. The results are shown as relative activities by adjusting the control to 1.0, and are the averages of two experiments. Sample concentration was 0.1%. The arrows, PF and CR, indicate the activity of unhydrolysed phytate-free (PFRBP) and phytate-containing rice bran protein samples.


37


Figure 5 Antioxidative activity of P4-40 hydrolysate fractions. Antioxidative activity was determined by the ferric thiocyanate method. Sample concentration was 0.008%. Relative antioxidative activity was calculated by adjusting the activity of the control to 1.0. Samples with relative antioxidative activity higher than 1.0 were subjected to RPHPLC on a TSK ODS 120 T column. Figure 3 (a) Yields and antioxidative activity of protease P4 hydrolysates. The P4 hydrolysate (20 g) was fractionated by reversed-phase column chromatography on a column (4.0 · 16.5 cm) of Kaisegel resin. Elution was carried out with stepwise gradients of ethanol 0–80% (200 mL each). (b) Antioxidative activity of P4 hydrolysates. Antioxidative activity was determined by the ferric thiocyanate method. Sample concentration was 0.04%.

22 23

17 18 19 21 20

27 30

25

37 35 43 36 40 33 41 32 34 42 38 39

28 24 29 31

14 13 12

80 45

44

47

49 46 48 51

16

60

50

11 10 9

52

40

Acetonitrile (%)

26

15

1.28 AUFS at 250 nm

38

8 6 12

3 4

Amino acid sequence analysis

7

20

5

0 0

20

40

monitored at 250 nm. Each peak was collected and concentrated using a centrifugal evaporator and 0.2 mg of dried sample were used to determine the antioxidative activity. The peaks having relative antioxidative activity above 1.0 were rechromatographed by RP-HPLC on a TSK gel ODS 120T column (7.5 · 250 mm; Tosoh, Tokyo, Japan) using a linear gradient of acetonitrile in 6 mm HCl at a flow rate of 1 mL min)1. Antioxidative activity was determined using 0.2 mg of each sample. Further purification was carried out for eight samples having high antioxidating activity by RP-HPLC using a C4Wakosil column with linear gradient of acetonitrile in 0.1% TFA. Purified peptides were used for further analysis.

60

80 Time (min)

100

120

140

Figure 4 RP-HPLC chromatogram of P4-40 fraction. Separation was carried out on a preparative column ODS 10 (20 · 250 mm) using a linear gradient of acetonitrile (10–40% in 120 min) in 0.1% TFA at a flow rate of 3 mL min)1. The sample (200 mg 10 mL)1) was injected into the column.

(10–40% in 120 min) in 0.1% trifluoroacetic acid (TFA). The separation was accomplished at 40 C at a flow rate of 3 mL min)1, and elution peaks were


Amino acid sequences of purified peptides were analysed using a gas-phase sequencer (Model PPSQ-10; Shimadzu, Kyoto, Japan) with an online HPLC analyser for phenylthiohydantoin (PHT) amino acid derivatives. The peptide were covalently attached to arylamine-derivatised poly(vinylidene difluoride) membranes (Millipore, Milford, MA, USA) before analysis to prevent washing out of samples during Edman degradation. Mass spectrometry

The matrix-assisted laser desorption ionisation-time of flight (MALDI-TOF) mass spectrometry analysis of purified peptide samples was performed on a KOMPACT MALDI mass spectrometer (Shimadzu).



0.6 m HCl for 30 min resulted in PFRBP. Higher antioxidative activity of CRBP was due to the presence of phytic acid (Fig. 1). The antioxidative activity was found to be dose-dependent and no prooxidative activity was observed. Five different proteases, proteases M, N, P, S and pepsin, originated from different sources and with distinct substrate specificities, were used for the hydrolysis of PFRBP. DH and antioxidative activity of the hydrolysates were monitored for a 24-h period for each protease (Fig. 2). The DH was estimated by quantification of cleaved peptide bonds as assessed by the OPA spectophotometric assay. The a-amino groups released by hydrolysis react with OPA and b-mercaptoethanol to form an adduct that has strong absorbancy at 340 nm. In every experiments, the antioxidative activity of PFRBP increased upon hydrolysis (Fig. 2).

Amino acid analysis

Peptide samples were hydrolysed in 6 m HCl containing 3% phenol at 145 C for 4 h in 100 lL microcapillary tubes. The hydrolysates were derived with 4-dimethylaminoazobenzene-4-sulphonyl chloride, and were analysed on a Capcelpak C18 column (5 lm, 4.6 · 150 mm; Shiseido, Tokyo, Japan). Results

Effects of hydrolysis on antioxidative activity of rice bran protein

Phytate contents of rice bran, CRBP and PFRBP were 2.5%, 1.42% and 0%, respectively. Protein extraction manipulation significantly decreased the phytate content in the CRBP and the extraction with

Peak 11

1

Peak 15

3 4

2 2 3 4

Peak 17

5 1

5

3

Peak 20

4 3 2

1

2 1

4

6 5

Figure 6 RP-HPLC chromatograms and antioxidative activity of peaks 11, 15, 17 and 20 derived from P4 hydrolysate. Rechromatograph was carried out on a TSK ODS 120T column (7.5 · 250 mm) using a linear gradient of acetonitrile (10–30% in 35 min) in 6 mm HCl at a flow rate of 1.5 mL min)1 and 40 C. Antioxidative activity was measured using the ferric thiocyanate method.



39

40


3 4

Peak 22

1

5

2

2 3 45

Peak 23 6

6

1

7

3

Peak 25 2 3 1

4

2 1

Peak 26 4 5 6


The antioxidative activity did not increase with increasing DH. Protease M showed the highest DH values but poor activity. The samples of DH below 10% had higher activity than those above 20%. Hamada (1999) reported that 74–92% of rice bran proteins could be solubilised with protease treatment of 2–10% DH values. Protease P yielded the most active hydrolysate, the activity reached the maximum at 4 h (P4, protease P of 4 h-hydrolysis) and decreased gradually. The activities were similar with protease S and N but protease M and pepsin showed similar changes in their activity. This result indicates that the antioxidative activity of the hydrolysates is inherent to their characteristics amino acid sequences of peptides depending on protease specificities.


Purification of antioxidative peptides in P4

The protease P hydrolysate of PFRBP (P4) was fractionated by reversed-phase chromatography on a Kaseigel resin column by a stepwise gradient of ethanol (0–80%). It has the advantage that a large quantity of sample can be loaded onto the column at a time and separated into fractions of increasing hydrophobicity. High percentages of hydrophilic peptide fraction and insoluble fraction constitute the bulk of P4 hydrolysate (Fig. 3). The 40% ethanol fraction of P4 hydrolysate (P4-40) showed the highest antioxidative activity and the insoluble fraction had the lowest antioxidative activity. The P4-40 fraction was then subjected to RP-HPLC for further purification.



The P4-40 was fractionated into 52 fractions (Fig. 4). Each collected peak was tested for antioxidative activity by the ferric thiocyanate method (Fig. 5). Fifteen antioxidative peptide fractions having activity higher than 1.0 were further purified on a TSK ODS 120 T column with a different solvent system (Figs 6–9). Final purification step was carried out with a C4 column for eight active peptides obtained with a TSK ODS 120T column (Fig. 10). The amino acid sequences of isolated antioxidative peptides in this study were summarised in Table 1. Their structures were confirmed by molecular mass and amino acid composition. The sequences of the isolated peptides were subjected to the BLAST protein database system. They have 100% homologies with the putative globulinlike Oryza sativa japonica cultivar group. The N-termini of the peptides were either alanine or valine or tyrosine,

while all the isolated peptides had arginine and valine in the internal sequences, and phenylalanine residue at the C-terminus. Discussion and conclusions

Rice bran has very short shelf life because of the rapid deterioration of the lipid fraction in the presence of lipase and lipoxygenase (Goffman & Bergman, 2003), hence fresh rice bran was defatted and stored at 4 C until use in this study. Phytic acid (myo-inositol hexaphosphate, IP6) is widely found in cereals, it is historically considered to be an antinutrient. It contains phosphate groups, which bind mineral ions, such as calcium, iron and zinc ions, causing a decrease of their bioavailability in human and animals (Ahn et al., 2004). Higher antioxidative activity of CRBP was due to the

3

6

5 2 4 1

7

Peak 34 2

Peak 32

3

4

2 1

Peak 33

5 6

3

Peak 36

4

3 1

5 2 1

6




41

42


Peak 41

Peak 42

2

4

3 2

3

4

5

1

1

2

Figure 9 RP-HPLC chromatograms and antioxidative activity of peaks 41, 42 and 43 derived from P4 hydrolysate. Rechromatograph was carried out on a TSK ODS 120T column (7.5 · 250 mm) using a linear gradient of acetonitrile (20–50% in 35 min) in 6 mm HCl at a flow rate of 1.5 mL min)1 and 40 C. Antioxidative activity was measured using the ferric thiocyanate method.

Peak 43 3

1

4

5

1

2 3

Peak 23-2

Peak 32-3

1 Figure 10 RP-HPLC chromatograms of peaks 23-2 and 32-3 on a Wakosil C4 column at a flow rate of 1 mL min)1 using a linear gradient of acetonitrile (15–30% in 30 min) in 0.1% TFA.

presence of phytic acid. This observation is similar to previous reports (Hix et al., 1997; Emmons & Peterson, 1999; Ahn et al., 2004). In this study, the ferric thiocyanate method was routinely used for the measurement of antioxidative activity. The auto-oxidation process of linoleic acid in the ethanol – 0.1 m sodium phosphate buffer (pH 7.0) system was examined by the ferric thiocyanate method. The oxidation of linoleic acid generates linoleic acid hydroperoxide, which decomposed to many secondary oxidation products. The oxidation is measured by coloured products formed ferrous by the oxidation of ferric ions. Low absorbance values indicate high levels of antioxidative activity. The


presence of phosphate in the assay is known to inhibit the auto-oxidation by chelating ions, thus enhancing antioxidative effects and lower the pro-oxidative effect with relatively high amino acid concentrations (Chen et al., 1995). Chen et al. (1996) observed that the pH and the concentration of sodium phosphate buffer influenced the auto-oxidation of linoleic acid. Therefore, the pH and concentration of the buffer were carefully controlled for the measurement of antioxidative activity. Several antioxidative peptides were isolated and their structures were determined. Hydrophobic amino acid residues existing in the peptides may play a role in



Table 1 Amino acid sequences of antioxidative peptides derived from rice bran proteinsa Amino Peak acid Sample number sequence Sample Peak 1 23-2

Peak 3

Sample Peak 1 32-3

a

AIRQGDVF

VLEANPRSF

Measured Calculated Amino molecular molecular acid mass mass composition 907.53

905.3

1038.09

1032.2

YFPVGGDRPESF 1373.6

1370.8

D(0.8)Q(0.9)G (0.5)A(0.8)V (0.8)I(0.6)F(0.8) R(0.8) N(0.4)S(0.8)E (0.7) P(0.8)A(0.8)V (0.5)L(0.8)F (0.8)R(0.9) D(0.7)S(0.6)G (1.6)P(1.5)E (0.8)V(0.8)Y (0.9)F(1.7)R(0.8)

The molecular mass was measured by MALDI-TOF-MS.

increasing the interaction with fatty acids, and the antioxidative activity of peptides depends on the constituents amino acids (Chen et al., 1995). No particular structural motif could be detected among these peptides. Acknowledgments

This work was supported by the funds of the Minsitry of Agriculture, Forestry and Fisheries of the Japanese Government. APA gratefully acknowledges the Japanese Ministry of Education, Culture, Science and Technology for granting him scholarship to undergo his study in Japan. References Ahn, H.-J., Kim, J.-H., Jo, C., Kim, M.-J. & Byun, M.-W. (2004). Comparison of irradiated phytic acid and other antioxidants for antioxidant activity. Food Chemistry, 88, 173–178. Akahoshi, A., Sato, K., Nawa, Y., Nakamura, Y. & Ohtsuki, K. (2000). Novel approach for large scale, biocompatible, and low-cost fractionation of peptides in proteolytic digest of food protein based on the amphoteric nature of peptides. Journal of Agricultural and Food Chemistry, 48, 1955–1959. Azuma, K., Nakayama, M., Koshioka, M. et al. (1999). Phenolic antioxidants from the leaves of Corchorus olitorius L. Journal of Agricultural and Food Chemistry, 47, 3963–3966. Chen, H.-M., Muramoto, K. & Yamauchi, F. (1995). Structural analysis of antioxidative peptides from soybean b-conglycinin. Journal of Agricultural and Food Chemistry, 43, 574–578.


Chen, H.M., Muramoto, K., Yamauchi, F. & Nokihara, K. (1996). Antioxidant activity of designed peptides based on the antioxidant peptide isolated from digests of a soybean protein. Journal of Agricultural and Food Chemistry, 44, 2619–2623. Chen, H.M., Muramoto, K., Yamauchi, F. & Nokihara, K. (1998). Antioxidant properties of histidine-containing peptides designed from peptide fragments found in the digests of a soybean protein. Journal of Agricultural and Food Chemistry, 46, 49–53. Chuda, Y., Ono, H., Ohnishiki-Kameyama, M., Nagata, T. & Tsushida, T. (1996). Structural identification of two antioxidant quinic acid derivatives from garland (Chrysanthenum coronarium L.). Journal of Agricultural and Food Chemistry, 44, 2037–2039. Church, F.C., Swaisgood, H.E., Porter, D.H. & Catignani, G.L. (1983). Spectrophotometric assay using o-Phthaldialdehyde for determination of proteosis in milk and isolated milk proteins. Journal of Dairy Science, 66, 1219–1227. Cuvelier, M.E., Berset, C. & Richard, H. (1994). Antioxidant constituents in sage (Salvia officinalis). Journal of Agricultural and Food Chemistry, 42, 665–669. Deslie, W.D. & Cheryon, M. (1988). Functional properties of soy protein hydrolysates from a continuous ultrafiltration reactor. Journal of Agricultural and Food Chemistry, 36, 26–31. Emmons, C.L. & Peterson, D.M. (1999). Antioxidant activity and phenolic contents of oat groats and hulls. Cereal Chemistry, 76, 902–906. Goffman, F.D. & Bergman, C. (2003). Hydrolytic degradation of triacylglycerols and changes in fatty acid composition in rice bran during storage. Cereal Chemistry, 80, 459–461. Hamada, J.S. (1999). Use of proteases to enhance solubilization of rice bran proteins. Journal of Food Biochemistry, 23, 307–321. Hix, D.K., Klopfenstein, C.F. & Walker, C.E. (1997). Physical and chemical attributes and consumer acceptance of sugar-snap cookies containing naturally occurring antioxidants. Cereal Chemistry, 74, 281–283. Kahkonen, M.P., Hopia, A.I. & Vuorela, H.J. (1999). Antioxidant activity of plant extracts containing phenolic compounds. Journal of Agricultural and Food Chemistry, 47, 3954–3962. Latta, M. & Erskin, M.A. (1980). Simple and rapid colorimetric method for phytate determination. Journal of Agricultural and Food Chemistry, 28, 1313–1315. Lin, C.-H. & Chang, C.-Y. (2004). Textural change and antioxidant properties of broccoli under different cooking treatments. Food Chemistry, 90, 9–15. Mannheim, A. & Cheryan, M. (1990). Continuous hydrolysis of milk protein in a membrane reactor. Journal of Food Science, 55, 381–385. Maruta, Y., Kawabata, J. & Niki, R. (1995). Antioxidative caffeoylquinic acid derivatives in the roots of burdock (Arctium lappa L.). Journal of Agricultural and Food Chemistry, 43, 2592–2595. Osman, H., Nasarudin, R. & Lee, S.L. (2004). Extracts of cocoa (Theobroma cacao L.) leaves and their oxidation potentials. Food Chemistry, 86, 41–46. Schwarz, K., Huang, S.W., German, J.B., Tiersch, B., Hartmann, J. & Frankel, E.N. (2000). Activities of antioxidant are affected by colloidal properties of oil-in-water and water-in-oil emulsions and bulk oils. Journal of Agricultural and Food Chemistry, 48, 4874–4882. Tada, M., Matsumoto, R., Yamaguchi, H. & Chiba, K. (1996). Novel antioxidants isolated from Perilla frutescens Britton var. crispa (Thunb). Bioscience Biotechnology and Biochemistry, 60, 1093–1095.


43


44

Original article Effect of heat treatment and refrigerated storage on antioxidant properties of pre-cut celery (Apium graveolens L.) Sonia Z. Vinã1 & Alicia R. Chaves2* 1 CIDCA (Centro de Investigacioń y Desarrollo en Criotecnologı´ a de Alimentos), Facultad de Ciencias Exactas, Universidad Nacional de La Plata (UNLP) 2 Consejo Nacional de Investigaciones Cientı´ ficas y Tećnicas (CONICET), Calle 47 y 116, S/N, La Plata (B1900AJJ), Buenos Aires, Argentina (Received 28 April 2006; Accepted in revised form 16 June 2006)

Summary

This work studies the effect of two types of heat treatment, dry air at 48 C for 1 h and water immersion at 50 C for 90 s, and of storage time at 0 C on a number of quality parameters for pre-cut celery: browning potential, soluble phenols content, total flavonoids, chlorogenic acid, ascorbic acid and antioxidant capacity. Pre-cut celery was placed in crystal polyethylene terephthalate trays covered with polyvinyl chloride film. Samples were taken after 0, 1, 7, 14 and 21 storage days. Treatments reduced browning potential and chlorogenic acid content and, in addition, allowed ascorbic acid concentration to be retained for a longer time. For this reason, the application of heat treatments in minimally processed celery would be beneficial.

Keywords

Antioxidant activity, Apium graveolens L., controlled stress, fresh cut vegetables, phenolics.

Introduction

The objectives of heat treatments, developed primarily for fruits, are to achieve insect disinfestation, to control diseases, to modify tissue response to other types of stress and to maintain product quality during storage (Paull & Jung Chen, 2000). These aims are the same as those set for the post-harvest of vegetables. To put them into practice, several time–temperature combinations must be tested in advance, taking into account that the botanical origin of edible parts (fruits, stems, petioles, leaves, buds, inflorescences and the like) is, in vegetables, considerably broader. Moreover, the optimum time and temperature combination chosen to extend fresh product quality during storage depends on cultivars, maturity stage, size and growing conditions (Fallik, 2004). In addition, selection of treatment type among heating in dry air, steam or water, may depend on product characteristics. All biological systems are known to respond to heat treatment, inducing transcription and selective translation of a determined gene group. The induced synthesis of heat shock proteins (HSP) correlates with an improved tolerance to a number of different abiotic stress factors (Loaiza-Velarde et al., 1997). HSP synthesis is noticeably favoured when compared with the Correspondent: Fax: +54 221 425 4853; e-mail: [email protected]

production of most other proteins (Vierling, 1991), so the application of a controlled thermal stress would condition vegetable tissue to withstand other stress types, amongst them those caused by cutting or senescence processes during subsequent storage. Application of heat treatments in minimally processed products would comprise additional objectives, namely to reduce physiological alterations in the plant, induced by mechanical and oxidative damage, as well as to lessen the responses linked to cicatrisation or wounding protection (Saltveit, 2000). Modifications to phenolic metabolism and tissue antioxidant capacity constitute very common defence mechanisms in plants. Diverse phenolic compounds can be induced by biotic or abiotic stress factors, such as high light intensity, UV radiation, pathogen attack, nutritional deficiency, low temperature and mechanical damage (Dixon & Paiva, 1995). Derivatives of cinnamic acid such as caffeic, p-coumaric and ferulic acids, referred to collectively as hydroxycinnamic acids, and chlorogenic acid are mostly synthesised by phenylalanine ammonia lyase catalysed conversion of l-phenylalanine to trans-cinnamic acid, which is stimulated in response to wounding or physiological stress (Wen et al., 2003). Several studies have shown that chlorogenic and iso-chlorogenic acids, derived from phenylpropanoid metabolism, accumulate in cut iceberg lettuce tissue (Ke & Saltveit, 1989; Toma´s-Barberań et al., 1997; Fukumoto et al., 2002). Moreover, phytochemicals such as flavonoids and other phenolics have


Antioxidants in heat-treated pre-cut celery S. Z. Vinã and A. R. Chaves

antioxidant activity and may help to protect cells against the oxidative damage caused by free radicals (Wada & Ou, 2002). Celery adapts easily to minimal processing but the main detrimental factors for its quality are vascular browning at the ends of cut petioles, flaring of the cut ends and development of pithiness (i.e. the formation of aerenchyma in the pith) (Saltveit & Mangrich, 1996; Loaiza-Velarde et al., 2003). It has been shown that a heat-shock treatment can diminish wound-induced physiological changes leading to reduced quality (i.e. tissue browning) and shortened shelf life (Loaiza-Velarde et al., 2003). The objective of the present work was to analyse the influence of two types of heat treatment and of refrigerated storage on several chemical components contributing to the antioxidant power of pre-cut celery. Materials and methods

Plant material and processing

Celery plants (Apium graveolens L.) cv Golden Boy, grown in greenhouse, were received from a La Plata grower (Province of Buenos Aires, Argentina). This is a white or self-whitening variety, widely cultivated in the zone. Once the plants reached the commercial size (after about 2 months of being transplanted), they were harvested early in the morning, brought to the laboratory and processed immediately. Leaves and basal segments of the rosettes were eliminated to obtain unbranched petioles. They were washed in running drinking water to remove any soil residues, and subsequently cut with a sharpened knife in 4-cm long sticks. These were disinfected by immersion in chlorinated water (100 ppm active chlorine, pH 6–6.5, 8 C) for 3 min and blotted dry. Selection of treatments

For immersion heat treatments, the following preliminary temperature–time combinations were tested: 45 C–120 s, 50 C–90 s and 55 C–60 s, the storage period being of 6 days at 20 C to speed up the manifestation of damage. Based on these results, a second stage of testing comprised immersion at 50 C– 90 s and 55 C–30 s, with a storage time at 0 C of 28 days. Immersion treatments were carried in heated distilled water using a thermostatic bath with permanent stirring. Celery cuts were placed in a plastic basket, and dipped during the selected times. Samples were subsequently immersed in chlorinated water with ice (100 ppm of active chlorine, pH 6–6.5) for 3 min for cooling and disinfestation. The product was packaged after eliminating the excess water by draining on absorbent paper.

With regard to air heat treatments, they were carried out after applying chlorinated water to samples as mentioned earlier. Once disinfested, the product was treated in a heated air oven for combinations of 48 C– 60 min and 50 C–20 min, followed by storage at 0 C for 28 days. Once air treatments were finished, samples were allowed to cool at room temperature before packaging. To evaluate and select the diverse treatments, their effect on sensorial attributes and damage development was considered (specially rot, yellowing and softening). Selected treatments and storage conditions

The selected treatments to be studied in this work are: (i) control sample (C), i.e. not exposed to heat treatment; (ii) thermally treated product by immersion in water at 50 C for 90 s (I, immersion); (iii) thermally treated product in dry-heated air (oven) at 48 C for 1 h (HA, hot air). In all tests, trays 17-cm long, 13-cm wide and 5-cm deep were used, which were made of crystal polyethylene terephthalate (PET) covered with self-adhering polyvinyl chloride (PVC) film (thickness, 10 lm; O2 permeability, 11 232 cm3 m)2 atm)1 day)1; CO2 permeability, 48 552 cm3 m)2 atm)1 day)1; water vapour permeability, 40 g m)2 day)1). The trays contained 175 g of product and were kept for 3 weeks in a cold store at 0 C with a relative humidity of 85%. Samples were taken for analysis at 0, 1, 7, 14, and 21 days. The whole experiment was repeated twice. Chemical analysis

For each sampling point, the material coming from three trays was combined and homogenised. Immediately before the analysis, part of the pool was frozen in liquid N2 and crushed in a laboratory mill (Janke & Kunkel Ika Labortechnik A10, Staufen, Germany). From this material, subsamples of exact weight were taken to carry out the corresponding determinations. Browning potential

Extraction was performed with ethanol 96 and absorbance (320 nm) of the solutions was measured (LoaizaVelarde et al., 1997). Extractions and determinations were carried out in duplicate and final results were expressed as absorbance units (AU) per gram of fresh tissue. Total phenols content

Aliquots (20 mL) of the alcoholic extracts were concentrated at reduced pressure (30 mm Hg, 40 C) in a rotary evaporator R-124 (Bu¨chi Labortechnik AG, Flawil, Switzerland), until dryness. Residues were resuspended in doubly distilled water. Total phenols were quantified employing the Folin-Ciocalteu reagent (Swain & Hillis,



45

46


1959), whereas absorbance readings were taken at 760 nm. Catechin was used as standard in the 3.75– 12.75 lg mL)1 concentration range. Duplicate extractions and determinations were conducted and final results were expressed as lmol g)1 of fresh tissue. Chlorogenic acid concentration

This determination was carried out as reported in a previous work (Vinã & Chaves, 2006). Samples were extracted and concentrated as mentioned earlier. Here, residues were resuspended in 1 mL high-performance liquid chromatograph (HPLC) grade methanol, and analysed in a Waters Model 6000A (Milford, MA, USA) HPLC, fitted with UV–VIS detector. A C18 column was employed (particle diameter 5 lm; internal diameter 4.6 mm; length 25 cm), using an 85:10:5 mixture of water:methanol:formic acid as running solvent. A flow rate of 1 mL min)1 was used. Detection was conducted at 320 nm. A standard solution of chlorogenic acid with a concentration of 0.87 lg mL)1 was used both to identify and quantify this compound. The UV–VIS spectrum of the fraction resulting from chromatographic runs was compared with the standard solution to confirm identification. Extractions and determinations were conducted in duplicate and results were expressed as nmol g)1 of fresh tissue. Total flavonoids content

It was determined by the technique described by Kim et al. (2003), with modifications. Samples were extracted, concentrated and resuspended in doubly distilled water as described earlier. To prepare reaction mixtures, a test tube was added with 1500 lL of doubly distilled water and 500 lL of the concentrated samples. Other compounds were added sequentially: initially (zero time) a volume of 150 lL of 5% NaNO2; after 5 min, 150 lL of 10% AlCl3 and finally, after additional 6 min, 500 lL of 1 m NaOH. Solutions were mixed by stirring in a vortex and then absorbance at 510 nm was measured. A standard curve was constructed based on catechin concentrations in the range of 7.5–36.6 lg mL)1. Extractions and determinations were conducted in duplicate. Total flavonoid levels in the samples were expressed as nmol g)1 of fresh tissue. Ascorbic acid content

A modified version of the method proposed by Wimalasiri & Wills (1983) was used. Samples were taken from the homogenised-frozen-crushed material, each weighed accurately to 3 g and extracted with 5 mL of aqueous solution of 3% citric acid. After 10 min, they were centrifuged at 11 500 g for 5 min at 5 C. Aliquots of 1 mL from each extract were centrifuged again in an Eppendorf 5415C equipment for 2 min at 14 000 r.p.m. The same HPLC equipment described in ‘Chlorogenic acid concentration’ section was used, with the same


column. In this determination, however, the mobile phase was a 70:30 mixture of acetonitrile:water with 0.01 m NH4H2PO4 and pH adjusted to 4.3 with orthophosphoric acid. Flow rate was 2 mL min)1, detection being carried out at 254 nm. For identification and quantification, a standard ascorbic acid solution of 35 lg mL)1 was employed. Extractions and determinations were carried out in duplicate and final results were expressed as milligram of ascorbic acid per 100 g of fresh tissue. Antioxidant power

Samples previously frozen in N2 and crushed were treated with 5 mL of methanol. The antioxidant power (AP) of the extracts was determined by reaction with the stable radical 2,2-diphenyl-1-picrylhydrazyl (DPPH•) in a methanol solution, using a modified version of the method proposed by Brand-Williams et al. (1995). Concentration of the extracts was varied in the reaction mixtures adding 0, 200, 400, 600, 800 or 1000 lL of each of them to a 3.9 mL methanol solution of DPPH• (25 ppm), completing a final volume of 4.9 mL with methanol. The reaction was allowed to progress and absorbance was measured at 515 nm after a constant value was reached. Then, DPPH• was calculated through a calibration straight line obtained in a range of concentrations of this substance. Finally, the remaining DPPH• concentration was plotted as a function of the extract volume in the reaction mixture, to calculate EC50 (effective mean concentration) for each sampling point. EC50 was defined as the mass (grams) of tissue required to reduce DPPH• concentration to half its initial value. Extractions and determinations were carried out in duplicate. Final results were expressed as AP, defined as the reciprocal of EC50 (AP ¼ 1/EC50). Statistical analysis

All data were treated by analysis of variance (anova). Sources of variation were time (five levels) and treatment (three levels). Means were compared using Fisher’s least significant difference (LSD) test. Differences at P < 0.05 were considered significant. Results and discussion

Selection of treatments

Concerning immersion treatments, the time–temperature combination 50 C–90 s led to lower damage by pathogens and a good retention of surface colour in samples stored at 20 C (data not shown). By using 55 C–60 s, pieces retained their green colour, but rot and softening were noticeable. Therefore, to proceed further with conservation tests at 0 C, the treatment at



0.16

Initial values averaged 0.08 AU per gram of fresh tissue. A significant increase in browning potential was observed for control samples over the first week of storage (P < 0.05), reaching a maximum at day 7, of almost twice the initial value. From then on, there was a tendency for browning to decrease, though nevertheless the browning potential in the untreated product stored at 0 C for 21 days was significantly higher (P < 0.05) than its initial level. In immersion-treated samples, browning potential was mostly constant (P > 0.05) over the first two weeks in cold store, to reduce towards day 21, to values slightly below the initial value. Samples treated in hot air experienced a significant increase (P < 0.05) of browning potential up to 14 days of storage. At this time, a maximum of 1.4 times of the initial value was reached. In our experiments, the initial browning potential values were 2.5 times as low as those found by LoaizaVelarde et al. (2003). These authors published browning potentials of 0.2 AU per gram of fresh tissue for petiole pieces 5 mm in length, increasing up to 0.6 AU per gram fresh tissue after 5 weeks at 0 C. Besides, for immersion-treated samples at 50 C for 90 s, the increase in browning potential was delayed by 3 weeks compared with the controls (Loaiza-Velarde et al., 2003), showing a similar trend as that observed in our experiments, where the immersion treatment hindered browning. Our results showed, in turn, that the hot air treatment was less effective. In spite of the preceding facts, enzymatic browning in the variety under study here (Golden Boy) did not produce severe damage because, although there was some development in the stored product (control and treated), manifestations were highly localised (brown-orange spots coincident with exposed vascular strands).

0.14

Total phenols

Table 1 Incidence of damage in pre-cut celery treated by water immersion or hot air after 28 storage days at 0 C

Treatment Control Immersion Hot air

Temperature and time combination

Soft rot incidence (%)

Yellowing

Softening

– 50 55 48 50

3 4 8 4 7

++ – – + ++

– – + – +

C–90 s C–30 s C–1 h C–20 min

–, unaffected product; +, moderately affected product; ++, seriously affected product. Percentage of soft rot incidence was calculated by the number of pieces affected related to the total number of inspected pieces in each sampling point.

50 C for 90 s was chosen; besides the exposure time of the 55 C treatment was reduced from 60 to 30 s. Again, the treatment with the best results was 50 C–90 s (Table 1). With respect to heated air treatments, the 48 C–1 h combination caused lower yellowing and minimised firmness losses (Table 1). For these reasons, the treatments selected to study their effect on the antioxidant properties of pre-cut celery were immersion in water at 50 C for 90 s and heated air at 48 C for 1 h. Browning potential

Browning potential (AU g–1 fresh tissue)

Figure 1 shows the results in this topic. No significant differences of browning were observed between the controls and the thermally treated pieces immediately after heat stress by immersion or dry air application.

0.12 0.10 0.08 0.06 Control

0.04

Immersion 0.02

Hot air

0.00 0

5

10

15 Time (days)

20

25

Figure 1 Browning potential of pre-cut celery exposed to heat treatments and stored for 21 days at 0 C (LSD0.05 ¼ 6 · 10)3).

Figure 2 shows total phenol concentration in cut celery as a function of time. Total phenol content in samples, measured immediately after applying immersion and hot air treatments, was slightly higher than that for the control, though not different enough to be statistically significant (P > 0.05). Such content averaged 0.20 lmol g)1 of fresh tissue at the beginning of storage. Instead, no significant differences were found between measured values during storage, nor with respect to their initial values. The same behaviour was found in treated samples. In the three cases, a tendency of total phenol content to increase was observed up to day 14. Loaiza-Velarde et al. (1997) have analysed results from immersion heat treatments applied to lettuce, and found combinations of 50 C–90 s and of 55 C–60 s to slow down the increase of phenolic compounds concentration with respect to the control, within 72 h of storage at 10 C.



47


Total Phenols (µmol g–1 fresh tissue)

0.30 0.25 0.20 0.15 0.10 Control Immersion Hot air

0.05 0.00 0

5

10 15 Time (days)

20

25

Figure 2 Total phenols in pre-cut celery experiencing heat treatment and stored for 21 days at 0 C (LSD0.05 ¼ 0.06).

Chlorogenic acid

Figure 3 shows that chlorogenic acid levels in cut celery immediately after both thermal treatments was somewhat higher than in controls though the variations were not significant (P > 0.05). Initial values were within 14.1 and 21.2 nmol g)1 of fresh tissue and the controls experienced a rapid increase in chlorogenic acid, reaching twice the initial concentration after 24 h at 0 C. Maximum concentration was observed at day 7, being 2.2 times as high as the initial value. Then, concentration decreased reaching values slightly below the initial

at day 21. In immersion-treated samples, a significant increase was also observed for chlorogenic acid concentration after 24 h in cold store, but this elevation was comparatively smaller, as the maximum level attained was 1.5 times as high as the initial value. Then it decreased, reaching values alike the initial. Concerning the samples treated in heated air, a significant concentration decrease was observed for this compound, considering the first 14 days of storage at 0 C. From then and until day 21, levels remained mostly constant. Our results agree with those found for lettuce by LoaizaVelarde et al. (1997), who have observed a chlorogenic acid build-up in control samples after 3 days of storage at 10 C. They also found that exposure at 50 C for 90 s retained initial levels of chlorogenic acid over the 72 h of the experiment. If a hierarchy of tissue response to different types of abiotic stress would exist (Saltveit, 2000), then it would be possible to redirect protein synthesis in minimally processed products from enzyme production related to damage response towards HSP (Saltveit, 2000). Thus, the enzyme synthesis participating in phenylpropanoid metabolism, which would be induced by mechanical injury, may be partially or totally repressed to favour HSP generation (Saltveit, 2000). A consequence of this would be the lower increase of chlorogenic acid observed in our experiments with minimally processed celery, for immersion- and hot air-treated samples. In this matter, efficiency of hot air treatment seems to be higher compared with immersion, i.e. to regulate the increase of chlorogenic acid that is observed in the untreated control. Total flavonoids

35.0 Control Chlorogenic acid (nmol g–1 fresh tissue)

48

30.0

Immersion Hot air

25.0 20.0 15.0 10.0 5.0 0.0 0

5

10 15 Time (days)

20

25

Figure 3 Chlorogenic acid content in heat-treated pre-cut celery, stored for 21 days at 0 C (LSD0.05 ¼ 7.3).


Studying the main group of phenolic compounds, the total flavonoid content was analysed. Results are exhibited in Fig. 4, and they indicate that the application of heat treatments, both by immersion and dry air, induced a sudden decrease in total flavonoids compared with the controls. In fact, initial flavonoid concentration in untreated samples was 1.3 and 2.5 times as high as those resulting from immersion and hot air treatments, respectively. This would indicate that thermal stress promoted a decrease of total flavonoids that was more marked after hot air treatment. Control samples experienced a decrease of 38% in total flavonoids, with respect to the initial value, after 24 h. Values kept decreasing but more slowly up to day 14, and remained almost constant up to day 21, where the concentration was of 53% of the starting level. In immersion-treated samples, flavonoid evolution was similar to that for controls though, in contrast, flavonoids increased slightly in those samples receiving hot air treatment (P < 0.05) over the first week in cold store and then remained nearly constant



Table 2 Ascorbic acid content (mg/100 g fresh tissue) of heat-treated pre-cut celery stored for 21 days at 0 C (LSD0.05 ¼ 0.4)

50.0 Control

Total flavonoids (nmol g–1 fresh tissue)

45.0

Ascorbic acid (mg/100 g fresh tissue)

Immersion 40.0

Hot air

35.0 30.0 25.0

Time (days)

C

I

HA

0 7 14 21

3.6 4.9 3.0 2.9

3.1 3.6 3.0 3.1

3.4 4.2 3.5 3.3

C, control; I, immersion thermal treatment (50 C, 90 s); HA, hot air thermal treatment (48 C, 1 h).

20.0 15.0 10.0 5.0 0.0 0

5

10 15 Time (days)

20

25

Figure 4 Total flavonoids content in heat-treated pre-cut celery, stored for 21 days at 0 C (LSD0.05 ¼ 4.5).

up to the end of storage. No differences were observed from the seventh day of storage between treated and control samples (P > 0.05). Ewald et al. (1999) have analysed the effect of various thermal treatments and processing methods on the flavonoids content in onion. For steam-blanched products, losses of 39% for quercetin and of 64% in kampferol were observed. Although the effect observed for heat-treated celery petioles was also a loss of total flavonoids, the decrease produced by water immersion at 50 C for 90 s was of 26% with respect to the initial content in the control in our experiments. Despite hot air treatments being much less drastic than conventional blanching, the use of hot air in cut celery led to a 60% decrease in total flavonoids with respect to untreated product at the beginning of the experiments. Ascorbic acid

Table 2 shows results from ascorbic acid (AA) determinations. Just after applying immersion treatment (day 0), there was a slight though significant decrease in AA content, compared with the control. In turn, hot air did not produce changes (P > 0.05). With regard to the controls, a significant increase by 36% of AA content was observed over the first week of storage at 0 C, with respect to the initial level. From then on AA content decreased significantly by 20% on day 21, compared with the initial value. In immersion-treated samples, the behaviour was very similar, though variations were less important. For instance, at day 7 the increase reached

16% of the initial value, while the subsequent reduction led to concentrations similar to the initial, at day 21. In the hot air treatment, a similar trend was presented. After 7 days of storage at 0 C, AA concentration increased significantly, by 24%, compared with the initial value. AA levels decreased from then on, returning to practically the initial values for 21 days. Such treatment would induce higher AA retention, as measured levels were higher than in controls and samples exposed to immersion treatment. The best known function of AA in plant cells is chloroplast protection against oxidative damage. Inactivation of O 2 and H2O2 produced by such organelles is catalysed by superoxide dismutase and ascorbate peroxidase. The monodehydroascorbate formed upon the action of this enzyme can be reduced directly by ferredoxin. Moreover, a second enzymatic system known as ascorbate-gluthathion system is important to regenerate AA from dehydroascorbic acid (Horemans et al., 2000). McCarthy & Matthews (1994) have indicated that minimal processing of fruits and vegetables would reduce AA content in tissues. However, ascorbate synthesis was also found to increase under stress conditions and that variation of its concentration would be a good indicator of the extent of damage experienced by the plant tissue (Stegmann et al., 1991). Evolution of AA content in cut celery, either heattreated or not, was in agreement with our previous findings for the same cultivar (Vinã & Chaves, 2006), where a slight increase of AA concentration was reported over the first week in refrigerated storage. More recently, Go´mez & Arte´s (2005) working with a green celery cultivar have found that AA content decreased after 15 days at 4 C, when pieces were stored in macroperforated polypropylene bags. Referring to our results, if synthesis or ‘recovery’ of AA content occurs, this fact would be a response to processing damage or to gradual senescence, which comes out even during storage at low temperatures. In view that in our experiments, such elevations were observed after about 7 days in store, it is likely that they would be linked with processes occurring before the



49


start of tissue senescence. Heat treatments would attenuate this response to a greater or lesser extent, affecting in some way the normal ageing processes of cut celery. Antioxidant power

The evolution of AP in cut celery exposed to heat treatments is shown in Fig. 5. At the beginning of the tests, AP of the control and hot air-treated samples were not significantly different (P > 0.05), averaging 4.4 g)1 of fresh tissue. Immediately after application of immersion treatment, AP decreased significantly by 41% compared with the control. This observation for the initial point of storage, agrees with the results for AA, where immersion treatment also induced a concentration decrease with respect to the control. Concerning controls, AP exhibited a significant decrease which, after 7 days of storage, reached 50% of the initial value. From then on, AP showed an increase, which first recovered the initial level, and then exceeded it by 30%, remaining constant up to the end of the conservation period. This behaviour was similar to that found when studying antioxidant capacity of cut celery stored at several temperatures (Vinã & Chaves, 2006). In immersion-treated samples, AP increased up to 14 days, to regain values alike those of controls. On applying hot air, AP evolved as in the controls. In this series of experiments on the effect of thermal treatments on AP of cut celery, no differences between control and treated samples were observed after 7 days of storage. 9.0 8.0 7.0 Antioxidant power (g–1)

50

6.0 5.0 4.0 3.0 Control

2.0

Immersion 1.0

Hot air

0.0 0

5

10 15 Time (days)

20

25

Figure 5 Antioxidant power of heat-treated pre-cut celery, stored for 21 days at 0 C (LSD0.05 ¼ 0.6).


Conclusions

The results of the present study indicate that heat treatments were beneficial to lessen the increases in browning potential verified in the controls over the first week at 0 C. In the same period, the hot air treatment inhibited the concentration increase of chlorogenic acid. This last compound can act as substrate for enzymatic browning reactions. In this regard, immersion treatment was not as effective. Immediately after application, treatments caused losses in total flavonoids. Concerning AA retention, hot air treatment was more beneficial than immersion. Acknowledgments

This work was financially supported by the Agencia Nacional de Promocioń Cientı´ fica y Tecnolo´gica (ANPCyT, PICT 1998-7088). Dr Sonia Z. Vinã wishes to thank the CONICET for granting her a scholarship. References Brand-Williams, W., Cuvelier, M.E. & Berset, C. (1995). Use of a free radical method to evaluate antioxidant activity. LebensmittelWissenschaft und-Technologie, 28, 25–30. Dixon, R.A. & Paiva, N.L. (1995). Stress-induced phenylpropanoid metabolism. The Plant Cell, 7, 1085–1097. Ewald, C., Fjelkner-Modig, S., Johansson, K., Sjo¨holm, I. & Akesson, B. (1999). Effect of processing on major flavonoids in processed onions, green beans and peas. Food Chemistry, 64, 231–235. Fallik, E. (2004). Prestorage hot water treatments (immersion, rinsing and brushing). Postharvest Biology and Technology, 32, 125–134. Fukumoto, L.R., Toivonen, P.M.A. & Delaquis, P.J. (2002). Effect of wash water temperature and chlorination on phenolic metabolism and browning of stored iceberg lettuce photosynthetic and vascular tissues. Journal of Agricultural and Food Chemistry, 50, 4503–4511. Go´mez, P.A. & Arte´s, F. (2005). Improved keeping quality of minimally fresh processed celery sticks by modified atmosphere packaging. Lebensmittel-Wissenschaft und-Technologie, 38, 323–329. Horemans, N., Foyer, C., Potters, G. & Asard, H. (2000). Ascorbate function and associated transport systems in plants. Plant Physiology and Biochemistry, 38, 531–540. Ke, D. & Saltveit, M.E. (1989). Wound-induced ethylene production, phenolic metabolism and susceptibility to russet spotting in iceberg lettuce. Physiologia Plantarum, 76, 412–418. Kim, D-O., Jeong, S.W. & Lee, C.Y. (2003). Antioxidant capacity of phenolic phytochemicals from various cultivars of plums. Food Chemistry, 81, 321–326. Loaiza-Velarde, J.G., Toma´s-Barbera´, F.A. & Saltveit, M.E. (1997). Effect of intensity and duration of heat-shock treatments on woundinduced phenolic metabolism in iceberg lettuce. Journal of the American Society for Horticultural Science, 122, 873–877. Loaiza-Velarde, J.G., Mangrich, M., Campos-Vargas, R. & Saltveit, M. (2003). Heat shock reduces browning of fresh-cut celery petioles. Postharvest Biology and Technology, 27, 305–311. McCarthy, M.A. & Matthews, R.H. (1994). Nutritional quality of fruits and vegetables subject to minimal processes. In: Minimally Processed Refrigerated Fruits & Vegetables (edited by Robert C. Wiley). pp. 313–326. London: Chapman & Hall Inc. Paull, R.E. & Jung Chen, N.J. (2000). Heat treatment and fruit ripening. Postharvest Biology and Technology, 21, 21–37.



Saltveit, M.E. (2000). Wound induced changes in phenolic metabolism and tissue browning are altered by heat shock. Postharvest Biology and Technology, 21, 61–69. Saltveit, M.E. & Mangrich, M.E. (1996). Using density measurements to study the effect of excision, storage, abscisic acid and ethylene on phitiness in celery petioles. Journal of the American Society for Horticultural Science, 121, 137–141. Stegmann, H.B., Schuler, P., Ruff, H.J., Knollmu¨ller, M. & Loreth, W. (1991). Ascorbic acid as an indicator of damage to forest. A correlation with air quality. Zeitschrift fur Naturforschung C: Journal of Biosciences, 46C, 67–70. Swain, T. & Hillis, W.E. (1959). The phenolic constituents of Prunus domestica I. The quantitative analysis of phenolic constituents. Journal of the Science of Food and Agriculture, 10, 63–68. Toma´s-Barberań, F.A., Loaiza-Velarde, J., Bonfanti, A. & Saltveit, M.E. (1997). Early wound- and ethylene-induced changes in phenylpropanoid metabolism in harvested lettuce. Journal of the American Society for Horticultural Science, 122, 399–404. Vierling, E. (1991). The roles of heat shock proteins in plants. Annual Review of Plant Physiology and Plant Molecular Biology, 42, 579–

620. Cited by Loaiza-Velarde, J.G., Toma´s-Barbera´, F.A. & Saltveit, M.E. (1997). Effect of intensity and duration of heat-shock treatments on wound-induced phenolic metabolism in iceberg lettuce. Journal of the American Society for Horticultural Science, 122, 873–877. Vinã, S.Z. & Chaves, A.R. (2006). Antioxidant responses in minimally processed celery during refrigerated storage. Food Chemistry, 94, 68–74. Wada, L. & Ou, B. (2002). Antioxidant activity and phenolic content of Oregon caneberries. Journal of Agricultural and Food Chemistry, 50, 3495–3500. Wen, A., Delaquis, P., Stanich, K. & Tivonen, P. (2003). Antilisterial activity of selected phenolic acids. Food Microbiology, 20, 305–311. Wimalasiri, P. & Wills, R.B.H. (1983). Simultaneous analysis of ascorbic acid and dehydroascorbic acid in fruit and vegetables by high-performance liquid chromatography. Journal of Chromatography A, 256, 368–371.



51


52

Original article Fluorescence spectroscopy in monitoring of extra virgin olive oil during storage Ewa Sikorska,1* Igor V. Khmelinskii,2 Marek Sikorski,3 Francesco Caponio,4 Maria T. Bilancia,4 Antonella Pasqualone4 & Tommaso Gomes4 1 2 3 4

Faculty of Commodity Science, Poznan´ University of Economics, al. Niepodlegosći 10, 60-967 Poznan´, Poland Universidade do Algarve, FCT, Campus de Gambelas, Faro 8005-139, Portugal Faculty of Chemistry, A. Mickiewicz University, Grunwaldzka 6, 60-780 Poznan´, Poland University of Bari, PROGESA Dept, Agro-Industry Section, via Amendola 165/a, I-70126 Bari, Italy (Received 06 March 2006; Accepted in revised form 20 March 2006)

Summary

The present study demonstrates the use of fluorescence spectroscopy for monitoring changes in virgin olive oil during storage. Total luminescence and synchronous scanning fluorescence spectroscopy techniques were tested with the purpose to check their ability to monitor changes occurring in olive oil during storage in different conditions: in clear and green glass bottles exposed to light, and in darkness. Total luminescence spectra of the initial oil samples in n-hexane solutions exhibited intense peaks, one with a maximum appearing at 320 nm in emission and 290 nm in excitation, attributed to tocopherols, and another appearing at 670 nm in emission and 405 nm in excitation, belonging to the pigments of the chlorophyll group. The intensity of these emissions decreased during storage depending on the storage conditions. Additional bands appeared in oils exposed to light in the intermediate range of excitation and emission wavelengths, arising from unidentified compounds. Bands attributed to tocopherols, chlorophylls and those tentatively ascribed to phenolic compounds were observed in the synchronous scanning fluorescence spectra, allowing monitoring of the storage effects on these constituents and their quantitative assessment after appropriate calibration. The results presented confirm the capability of the fluorescence techniques to monitor the quality of oil products.

Keywords

Chlorophylls, fluorescence spectroscopy, olive oil, principle component analysis, tocopherols.

Introduction

Virgin olive oil, obtained from the fruits of Olea europaea L. by mechanical means without any extra treatment, is one of the few vegetable oils consumed in its natural state (Rahmani & Csallany, 1998). It has excellent organoleptic and nutritional properties, especially appreciated by consumers. However, like other vegetable oils, it is susceptible to oxidation, which has been recognised as the predominant cause of oil deterioration during storage (Morello et al., 2004). The oxidative reactions also cause a partial loss of the minor constituents of olive oil, considered primarily responsible for its beneficial health effects (Chen & Huang, 1998; Manna et al., 1997; Visoli & Galli, 1998). Tocopherols are the most effective lipid-soluble antioxidants that protect cell membranes from peroxyl radicals and mutagenic nitrogen oxide species (Chow, *Correspondent: Fax: +48 61 854 3993; e-mail: [email protected]

1991). Phenolic components of the olive oils have been reported to be beneficial to health too, because of their antioxidant activity. Numerous epidemiological studies have linked the consumption of olive oil to a reduction of cardiovascular disease and various tumours (Keys, 1995; Tavani & La Vecchia, 1995). Increasing effort has been devoted to the development of methods capable of detecting and quantifying the oil oxidation. For this purpose several analytical techniques, including gas chromatography, high-performance size-exclusion chromatography (HPSEC) and u.v. spectrophotometry have been used (Manna et al., 1997; Wheatley, 2000; Caponio et al., 2003; Gomes et al., 2003). Fluorescence spectroscopy is a rapid, nondestructive analytical technique with high specificity and sensitivity. The importance of this technique for food analysis has increased in recent years. Among the benefits of fluorescence spectroscopy is its high sensitivity to a wide array of potential analytes, and in general, avoidance of consumable reagents and extensive sample

doi:10.1111/j.1365-2621.2006.01384.x Ó 2007 The Authors. Journal compilation Ó 2007 Institute of Food Science and Technology Trust Fund

Monitoring of olive oils by fluorescence E. Sikorska et al.

pretreatment (Oldham et al., 2000). Conventional fluorescence techniques, relying on the measurements of single emission or excitation spectra, are often insufficient in the analysis of complex systems. In such cases, total luminescence or synchronous scanning fluorescence (SSF) techniques are applied, improving the analytic potential of the fluorescence (Gutierrez et al., 1987; Ndou & Warner, 1991; Patra & Mishra, 2002). The total luminescence spectroscopy (TLS) involves simultaneous acquisition at multiple excitation and emission wavelengths in order to increase the method selectivity. The resulting emission–excitation data matrix (EEM) provides the total intensity profile of the sample over the range of excitation and emission wavelengths scanned. Wider application of these techniques in recent years was prompted by technical developments in spectrofluorometers, which enable rapid measurements of the complete EEM. The SSF technique offers an alternative to the TLS. The synchronous fluorescence spectrometry takes advantage of the hardware capability to vary both the excitation and the emission wavelengths during the analysis. In this method the excitation and the emission monochromators are scanned simultaneously, synchronised so that a constant wavelength difference is maintained between the two. Such measurements are simpler and faster than EEM collection and can be conducted on a majority of conventional spectrofluorometers. Although the synchronous fluorescence spectra contain less information than the EEM, they are potentially more informative than single excitation and emission spectra. Moreover, by choosing an appropriate offset between excitation and emission wavelengths, the resolution and/or the intensity of a particular fluorescent component may be enhanced. The potential of all the fluorescence techniques can be improved by applying multivariate statistical methods in the analysis of the fluorescence results. Several papers have discussed the potential of the fluorescence in the direct analysis of edible oils. Fluorescence spectroscopy was used for characterising of various types of oil (Kyriakidis & Skarkalis, 2000; Zandomeneghi et al., 2005), monitoring changes in frying oil (Engelsen, 1997), determination of chlorophylls and pheophytins in olive oil (Diaz et al., 2003), discrimination between different kinds of edible oils (Engelsen, 1997; Kyriakidis & Skarkalis, 2000; Giungato et al., 2002) and between differently processed olive oils (Guimet et al., 2004). We have recently applied the total fluorescence spectroscopy and the synchronous fluorescence spectroscopy for characterisation and discrimination of various kinds of edible oils (Sikorska et al., 2004, 2005b). This article continues our pursuit in exploring the possibility of application of the fluorescence methods in the analysis of vegetable oils. The total fluorescence

and the SSF techniques have been used for monitoring changes in fluorescent components of extra virgin olive oils during storage under different conditions. Experimental

Materials and sampling

Extra virgin olive oil from the Coratina cultivar, obtained in the crop season 2001–2002 on a traditional plant (a stone mill and a hydraulic press) was used for our experimental tests. Once in the laboratory, the oil was transferred into clear and green glass bottles, 150 mL each. The bottles were hermetically sealed. The bottling procedure was the same as that used in the oil mills. The clear bottles were divided in two groups. The first group was placed in a carton and stored in darkness, while the second group of clear bottles and the green bottles were stored under diffuse lighting, simulating the conditions of a supermarket shelf: the average temperatures during winter and summer were 15 ° and 25 °C, respectively, with the light intensity of 1000 lux. The samples were withdrawn from storage at fixed times: initially (0 months) and after 1, 2, 4, 6, 8, 10 and 12 months from bottling and storage in the conditions described. One bottle was retrieved at each of the storage times, and each analysis was performed with two replicates. Tocopherol (97%), and n-hexane, acetone and ethanol, all HPLC grade, were purchased from Aldrich (Steinbaum, Germany), oleuropein was obtained from Extrasynthese (Genay, France). The absorption spectra of olive oil and transmission spectra of bottles were measured using a Cary 5E spectrophotometer (Varian). Analytical determinations

The chlorophyll fraction was evaluated from the absorption spectrum of each oil sample dissolved in hexane, according to the AOCS method (1993). The chlorophyll content was expressed as milligram of pheophytin-a per kilogram of oil. The phenolic compounds were extracted and purified according to the method described by Cortesi et al. (1981), while the quantitative determinations were carried out as reported by Favier et al. (1994), using Folin-Ciocalteau as a reagent and measuring the absorption at 765 nm. The results obtained were expressed as milligram of gallic acid per kilogram of oil. Fluorescence measurements

The fluorescence spectra were obtained on a Fluorolog 3-11 Spex-Jobin Yvon spectrofluorometer. A xenon lamp source was used for excitation. The excitation and

Ó 2007 The Authors. Journal compilation Ó 2007 Institute of Food Science and Technology Trust Fund


53

54


emission slit widths were 2 nm. The acquisition interval and the integration time were maintained at 1 nm and 0.1 s, respectively. A reference photodiode detector at the excitation monochromator stage compensated for the source intensity fluctuations. The individual spectra were corrected for the wavelength-dependent response of the system. Right-angle geometry was used for the oil samples diluted in n-hexane (1% v/v) in a 10 mm fusedquartz cuvette. The low concentration of the samples was chosen to avoid spectral distortions, which may occur in more concentrated solutions. The three-dimensional spectra were obtained by measuring the emission spectra in the range from 290 to 700 nm repeatedly, at excitation wavelengths from 250 to 550 nm, spaced by 5-nm intervals in the excitation domain. The fully corrected spectra were then concatenated into the EEM. The synchronous fluorescence spectra were collected by simultaneously scanning the excitation and the emission monochromators in the 250–700 nm range, with a constant wavelength difference Dk between them. The fluorescence intensities were plotted in function of the excitation wavelength. Three spectra were recorded for each of the samples, with Dk of 10, 60, and 80 nm.

is the predicted concentration value for a where ypred i sample in the cross-validation procedure, yref i the reference value and N the number of samples. Principal component analysis (PCA) was used for the multivariate evaluation of the fluorescence data (Wold et al., 1987). PCA modelling was performed on the synchronous fluorescence spectra measured at Dk ¼ 10 nm. An average of the three spectra obtained for each of the samples was used in the analysis. The spectral structure and distribution of samples was evaluated on the basis of the scores and the loading plots. The score plots visualise the relationship between the oil samples for each of the principal components (PC), while the loading plots were used for interpretation of the corresponding spectral variations. The synchronous fluorescence spectra measured at Dk ¼ 10 nm were used for PLS and PCA analysis because of relatively high intensity and the most effective band separation. The data analysis by PCA and PLS was performed using Unscrambler 9.0 software (CAMO, Oslo, Norway).

Fluorescence data analysis

Total luminescence spectra

Partial least squares regression (PLS) was used for testing correlation between the synchronous fluorescence spectra (X) and the components (chlorophylls and total phenols) concentration (Y) in the olive oil samples. PLS is a method for relating the variations in one or several response variables (Y-variables) to the variations of several predictors (X-variables), with explanatory or predictive purposes (Brereton, 2000). PLS models both the X- and Y-matrices simultaneously to find such latent variables in X that will best predict the latent variables in Y. PLS modelling was performed on the synchronous fluorescence spectra measured at Dk ¼ 10 nm. An average of three recorded spectra per sample was used in the analysis. Data pretreatment involved mean centering. Full cross-validation was applied for all regression models. Cross-validation is a strategy for validating calibration models based on systematically leaving out groups of samples in the modeling, and testing the left out samples in a model based on the remaining samples. The regression models were evaluated using the correlation coefficient (r), and the validation parameter, Root Mean Square Error of Cross-Validation (RMSECV), as a term indicating the prediction error of the model. The RMSECV is defined by the equation: sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi PN pred 2 yref i Þ i¼1 ðyi RMSECV ¼ N

Figure 1 shows the absorption spectra of a virgin olive oil in n-hexane and the transmission spectra of the clear and green glass, used as the bottle material for the oil storage. The virgin olive oil absorption in the 450– 520 nm range is ascribed to carotenoid pigments. The carotenoid band overlaps with the chlorophyll absorption at 380–450 nm, while the characteristic band at 650–700 nm originates exclusively from the absorption of chlorophylls and pheophytins. The clear glass transmits light in the region characteristic for both carotenoids and chlorophylls absorption. In contrast, green bottles transmit almost nothing below 500 nm, protecting carotenoids and the short-wavelength absorption of chlorophylls, while transmitting only about 20% in the region characteristic to the long-wavelength absorption of chlorophylls. Figure 2 shows the three-dimensional TLS of the same virgin oil and their changes owing to storage in different conditions. The spectra shown in Fig. 2 exhibit general characteristics similar to those obtained previously for various kinds of edible oils. The assignment of the emission bands to specific chemical components, based on comparison of the three-dimensional and single excitation and emission spectra with those of the respective reference compounds, was described in detail previously (Sikorska et al., 2004, 2005b). The relatively intense band, observed in the starting oil sample, with the excitation in the range of ca. 270–





(a) 0.8

Absorbance

0.6

0.4

0.2

0.0 300

400

500 600 Wavelength (nm)

700

800

(b) 80

a

Transmittance (%)

60

40

b

20

0 300

400


700

800

Figure 1 (a) Absorption spectrum of nonstored virgin olive oil in n-hexane. (b) Transmission spectra of clear (a) and green (b) glass of the bottles used for the oil storage.

310 nm and emission in the range of c. 300–350 nm has been ascribed to the tocopherols. The olive oils mainly contain a-tocopherol in the average amount of ca. 5 mg (100 g))1 (deMan, 1999). It was recently shown that phenolic compounds may also contribute to the emission of virgin olive oil, which have fluorescence parameters similar to those of tocopherol (Zandomeneghi et al., 2005). In fact, the slight shift of maxima of the short-wavelength emission in function of the excitation wavelength indicates that the emission in this region originates from more than one species. The long-wavelength band, at 350–420 nm in excitation and 660–700 nm in emission, is characteristic of the fluorescence of pigments of the chlorophyll group, which includes chlorophylls a and b and pheophytins a and b (Diaz et al., 2003). The pigments of the chlorophyll group mainly occur in crude oils, obtained directly by the extraction from the oilseed, and are removed during

subsequent processing. Virgin olive oils are known to contain considerable amounts of pigments of this group (Rahmani & Csallany, 1998), pheophytin a being the most important (Psomiadou & Tsimidou, 2001). The intensity of the tocopherol and chlorophyll bands decreased in all of the stored samples, to the extent dependent on the storage conditions. The chlorophyll bands disappeared completely in oil stored in clear glass bottles under light. The oil stored under light in green bottles also exhibited a reduced intensity of the chlorophyll pigments after 10 months of the experiment, with a very small reduction in the oil stored in darkness. Interestingly, a new fluorescence band appeared in the oil stored under light both in green and clear glass bottles, with a maximum at ca. 300 nm in excitation and 400 nm in emission. A similar band had been noted by us previously in a variety of edible oils, which revealed an emission band at ca. 400–450 nm. The shape and intensity of such emissions varied for different oils (Sikorska et al., 2004, 2005b). The total luminescence characteristics of the olive oil in our study are similar to those reported by Giungato et al. (2002). They found for a virgin olive oil in isooctane an emission band with the excitation maximum at 285 nm and the emission maximum at about 315 nm, attributed to tocopherols, and a second band with the excitation maximum at 410 nm and the emission maximum at 669 nm, ascribed to the chlorophylls. After refining, the tocopherol band was observed with excitation and emission maxima at 295 and 331 nm respectively, while the chlorophyll emission was not detectable any more. Another emission was observed for this oil with two excitation maxima at 300 and 315 nm and an emission maximum at 406 nm, the origin of this emission having remained uncertain. This emission should correspond to the emission band observed for our virgin olive oil exposed to light in both clear and green glass bottles. Synchronous fluorescence spectra

Figure 3 shows changes in the synchronous fluorescence spectra of the virgin olive oil stored in different conditions recorded with the emission–excitation offset (Dk) of 10 and 60 nm, respectively. The synchronous fluorescence spectra obtained with smaller Dk usually show an effective bandwidth reduction, resulting in spectral simplification and enhanced resolution of the overlapping fluorescence of different components. The spectra of the virgin olive oil recorded at Dk ¼ 10 nm show three major bands with their maxima at around 284, 301 and 666 nm. As follows from previous studies by the present and other authors (Sikorska et al., 2004, 2005b), the main identified fluorescent components of oils are the compounds of the tocopherol and the chlorophyll groups.



55


(a)

(b)

25E + 06

25E + 06

20E + 06

20E + 06

l (cps) 15E + 06

l (cps) 15E + 06

10E + 06

0

(c)

10E + 06

300 400 EX (nm) 500

50E + 05 300

0

(d)

25E + 06

20E + 06

20E + 06

l (cps) 15E + 06

l (cps) 15E + 06 300 400 EX (nm) 500

50E + 05 300

600 700 400 500 EM wavelength (nm)

300

25E + 06

0

300 400 EX (nm) 500

50E + 05


10E + 06

10E + 06

300 400 EX (nm) 500

50E + 05 0


300


Figure 2 Total luminescence spectra of virgin olive oil: (a) a nonstored sample, and stored for 10 months: (b) in darkness, and exposed to light: (c) in green glass bottles, (d) in clear glass bottles.

(a) 1 x 107 Fluorescence intensity (cps)

Fluorescence intensity (cps)

1 8 x 106 2 6 x 106

1 2

4 x 106 3 4

6

2 x 10

3

300

400

500

600

700

4 x 106 1 3 x 106

2 4 3

2 x 106


Wavelength (nm)

(b)

3 4

Pheophytins Chlorophylls

2.4 ×107 a 7

1.8 ×10

1.2 ×107 6.0 ×106

b

300


600

700

2 1 0 300

320 340 360 380 Wavelength (nm)

400

1 x 106

0 300

Oleuropein Tocopherols

Figure 4 Synchronous florescence spectra measured at Dk ¼ 10 nm, of oleuropein in n-hexane-ethanol (10:1, v/v) mixture, a-tocopherol in n-hexane, and bacteriopheophytin c in acetone (a ) Dk ¼ 10 nm, b ) Dk ¼ 60 nm). All spectra were normalised to the same maximum fluorescence intensity.

1x106

5x105

3.0 ×107

0.0 200

4 0


56

400

500

600

700

Wavelength (nm) Figure 3 Synchronous fluorescence spectra of nonstored virgin olive oil (1) and their changes owing to storage in darkness (2), and under light in green glass (3) and clear glass (4) bottles for 10 months: a ) Dk ¼ 10 nm, b ) Dk ¼ 60 nm.


Thus, the bands with the maxima at 301 and 666 nm are respectively ascribed to tocopherols and chlorophylls. This conclusion is supported by comparison of the spectra of the olive oil with those of the respective standards, see Fig. 4. The intense band at 284 nm has not been observed in the spectra of the previously studied oils (Sikorska et al., 2005b). This band could originate from the phenolic compounds such as phenolic aglycons, based on the molecules of tyrosol and hydroxytyrosol, deriving from the phenolic glycosides present in the olives. Such phenolic compounds exist in virgin olive oils, in contrast to refined vegetable oils. Oleuropein is the main glycoside present in high amounts in unprocessed olive fruit



and leaves. Its degradation leads to the formation of hydroxytyrosol and related compounds. The spectrum of pure oleuropein, shown in Fig. 4, exhibits a maximum at about 289 nm and is shifted by about 5 nm as compared with the fluorescence band observed in the virgin olive oil. This shift may results from the solvent: the phenolic compounds are poorly soluble in nonpolar solvents, thus the spectrum was recorded in the nhexane–ethanol mixture. On the other hand, the emission observed in the oil may originate from oleuropein derivatives, with slightly different emission properties. The virgin olive oil from the Coratina cultivar was found to contain a variety of phenolic constituents, some of which may contribute to the emission discussed (Caponio et al., 2001). Papadopoulos et al. (2003) has recently reported the total phenol contents of the aqueous extracts of olive and seed oils in the range of 55.3–125.9 mg caffeic acid equivalents per kilogram of olive oils, compared with only 0.5–3.2 and 0.6–2.2 mg caffeic acid equivalents per kilogram of sunflower and corn oils, respectively. The authors state that higher total phenol contents were accompanied by higher values of absorbance and fluorescence intensities, at 280 and 314 nm respectively, of the corresponding extracts. However, the absorbance and fluorescence intensity ratios of the olive oils to the seed oils were much higher than the respective ratio of the antioxidant activities, indicating that compounds other then phenols might have contributed to the absorption and emission observed. Note that it is hard to detect the two overlapping bands in the short-wavelength region relying on visual analysis of the three-dimensional TLS only, which, however, is readily achieved using the synchronous scanning techniques. There is no evidence of formation of new fluorescent compounds in the SSF spectra recorded at Dk ¼ 10 nm, whereas they are readily obvious in the TLS. However, in the SSF at Dk ¼ 60 nm (Fig. 3b) and 80 nm (not shown) a new band appears with a low intensity, corresponding to the new fluorescent component. The spectra recorded at Dk ¼ 60 nm exhibit a broad shortwavelength band originating from the tocopherol emission; the chlorophyll emission appears at 600–700 nm and is split into two bands, with their maxima at 611 and 663 nm, respectively, see Fig. 4 for comparison. An additional low-intensity band with the maximum at 316 nm appears in the spectra of oils stored under light in both clear and green glass bottles for 10 months, corresponding to the spectral characteristics of the new fluorescent component observed in the TLS experiments. This example illustrates the flexibility of the synchronous fluorescence technique which enables monitoring different fluorophores by an appropriate choice of the Dk values. The new fluorescent components

were only detected in the oil stored in clear and green bottles under light, and are tentatively identified as photodecomposition/photo-oxidation products of chlorophyll or other pigments. Virgin olive oils also contain pigments of the carotenoid group. These compounds are usually nonfluorescent, whereas their photodecomposition products could contribute to the new fluorescence. On the other hand, the new fluorescence band may indicate formation of other photooxidation products. It was shown that the oils kept in the dark mainly contain products of the primary oxidation, while those exposed to light contain products of the secondary oxidation (Caponio et al., 2005). These products may also react with other oil components and form fluorescent compounds (Kikugawa & Beppu, 1987; Liang, 1999). However, these hypotheses should be verified, and the respective fluorescent products identified, which could be a subject of a future study. At this stage we may conclude with some certainty that the formation of these products is associated with either the direct chlorophyll decomposition, or their photosensitising action. The comparison of the absorption spectrum of the virgin olive oil with the transmission spectra of the clear and green glass, Fig. 1, leads to a conclusion that the chlorophyll pigments are the main species that absorb light in the green bottles. However, we were unable to quantitatively correlate the decrease in the chlorophyll fluorescence intensity with the increase of the new product fluorescence. One reason for this may be the very low intensity of this product emission. Contributions from the direct photolysis of lipids by the u.v. radiation may be excluded based on the transmission spectra of the glass bottles, Fig. 1, used for the oil storage. Figure 5 shows the changes during storage in the SSF spectra of the olive oil, recorded at Dk ¼ 10 and 60 nm. The fluorescence intensity of tocopherols and chlorophylls decreases with the storage time, with the rates depending on the storage conditions. The intensity of these emissions decreases very slowly in the oil stored in darkness, being reduced by no more than 30% of the starting values during the experiment. For the oil exposed to light in both clear and green glass bottles, the tocopherol intensity decreases continuously, with a higher rate in clear bottles. The principal differences are observed in the chlorophyll contents. The decomposition of the chlorophylls is very fast in the oil stored in clear bottles, being almost complete after the first two months. It is slower in green bottles, although still much faster if compared with the samples stored in darkness, where a reduction by only about 23% of the starting intensity is observed. The photodegradation of chlorophylls in green bottles is reduced because of the absorption of green glass in the region where the intense short-wavelength absorption



57


(b)

80 60 40

Dark Green glass Clear glass

20

0

(c)

Fluorescence intensity at 284 nm (%)

100

2

4 6 8 Storage time (months)

10

(d)

100 80 Dark Green glass Clear glass

60 40 20 0

100 80 60 40

Dark Green glass Clear glass

20 0

12



(a)


58

0

2


10

12

10

12

600 500 Dark Green glass Clear glass

400 300 200 100

0

2


10

12

0

2


Figure 5 Changes in the fluorescence of virgin olive oil stored in darkness and under light in green and clear glass bottles; monitored at: (a) 284 nm (emission tentatively ascribed to the phenolic compounds); (b) 301 nm (tocopherols); (c) 666 nm (chlorophylls, pheophytins), all data from synchronous scans at Dk ¼ 10 nm; (d) at 316 nm, data from synchronous scans at Dk ¼ 60 nm. Fluorescence intensity normalised to 100% for the starting sample.

of the pigment occurs. In contrast to tocopherol and chlorophyll, the changes of the emission intensity of the 284 nm band were less pronounced and seemingly independent on the storage conditions. These observations are in agreement with previous studies concerning the effect of storage conditions on the Coratina cultivar virgin olive oil (Caponio et al., 2005). Previously, it had been shown that the oils exposed to light showed tocopherol, chlorophyll and carotenoid contents lower than the same oils kept in darkness, while the decrease in phenolic substances was much less pronounced in all the storage conditions. The fact that the emission at 284 nm behaves in the way similar to that observed for the phenolic compounds supports the tentative attribution of this fluorescence band. The new band with the maximum at 316 nm in the SSF spectra recorded at Dk ¼ 60 nm only appears in samples exposed to light during the first months of the experiment.


Quantitative analysis

To test the quantitative relation between fluorescence ascribed to distinct fluorophores and real content of respective compounds in the olive oil samples, univariate and multivariate regression analysis was performed between analytically determined chlorophylls and total phenol concentration, and spectral data. We have recently shown for a series of vegetable oils, using the same regression methods, that in diluted samples the total tocopherol content correlates well with the fluorescence intensity at kex ¼ 301 nm/kem ¼ 311 nm (in univariate approach) and synchronous fluorescence spectra at Dk ¼ 10 nm (in multivariate approach) (Sikorska et al., 2005a). Linear regression analysis of the fluorescence intensity at 666 nm vs. the chlorophyll concentration yielded the regression coefficient of 0.987. This indicates that the fluorescence band with the maximum at 666 nm in



synchronous spectra can be ascribed mainly to chlorophyll, confirming our previous hypothesis. A much worse correlation, with the regression coefficient of 0.646, was found between the fluorescence intensity at 284 nm and the total phenol concentration. PLS modelling was performed for synchronous fluorescence spectra. The synchronous fluorescence spectra at Dk ¼ 10 nm were used as the x-variable, while the concentration of chlorophylls and total phenols determined by spectrophotometric methods was used as the y-variable. The results are summarised in Table 1. The results obtained using the entire spectra are similar to those of univariate analysis, where fluorescence intensity at a single selected wavelength was considered. PLS regression gave good results in modelling of the chlorophyll content, with high regression coefficient for predicted vs. measured concentration, and RMSECV value below 10% of the average chlorophyll content. Multivariate analysis for the total phenols yielded a generally poor model with lower regression coefficient and RMSEC exceeding 10% of the average total phenols content. Moreover, in this case the model explained only 50% of the total Y variance. Both the previously reported (Sikorska et al., 2005a) and the present results show that synchronous fluorescence spectra of diluted oil samples can be used for quantification of tocopherols and chlorophylls after an appropriate calibration. Poor results for total phenolic compounds are probably because of the fact that olive oils contain a variety of these compounds, with only some of them fluorescent. Therefore, to build an optimal model, the fluorescent phenolic fraction should be correlated to the fluorescence spectra. Quantification of these compounds requires further studies. Qualitative analysis

In the preceding section we have shown that the analysis of the TLS and SSF spectra enables monitoring of particular fluorescent compounds present in an olive oil. This feature can be useful for studies of the mechanism of deterioration of olive oils during storage. On the other hand, the fluorescence techniques can be used as a

Table 1 Results of the partial least squares (PLS) regression for quantification of chlorophylls and total phenols using the synchronous fluorescence spectra at Dk ¼ 10 nm

rapid tool for quality monitoring. In this application, the overall fluorescence changes might be more relevant than the analysis of particular components. Such overall changes may be conveniently and quantitatively evaluated using multivariate analysis of the fluorescence spectra. In particular, we performed PCA on the synchronous fluorescence spectra measured at Dk ¼ 10 nm. Figure 6 shows the score plot of the PC1 vs. PC2 principal components of the PCA model performed on the olive oil samples, and the loading plots, which display the spectral structure of the first two PC in the form of a one-vector plot. The PC1 and PC2 principal components represent 97% and 1% of the total variation, respectively. The distribution of the samples in the score plot shows that the PC1 mainly explains variations caused by the light exposure and the transparency of the bottles. The samples stored in darkness and in green glass bottles are grouped with positive PC1 values, while those exposed to light in clear glass bottles exhibit negative PC1 values. The importance of the contribution into the PC1 of the emission bands with the maxima at ca. 666 nm, and to a lower extent at 301 nm, is evident from the corresponding loadings plot. These wavelengths correspond to the chlorophyll and tocopherol emissions, indicating that the major variation in fluorescence is because of degradation of these components, primarily chlorophyll. The PC2, on the other hand, seems to classify the samples according to the storage time. The PC2 decreases with the storage time for all the storage conditions. This may indicate that PC2 describes processes that are not induced by the light exposure. The second loading shows high contributions of the emission band at 301 nm, and a negative contribution of the chlorophyll band, demonstrating that this component is dominated by the degradation of the tocopherols with possible concurrent stabilisation of the chlorophylls. The PCA results are consistent with the photosensitising properties of the chlorophylls and the antioxidative properties of tocopherols. The overall trend from the PCA is that a significant decrease in tocopherol and chlorophyll fluorescence is observed because of the light

PLS model

Mean compound concentration (standard deviation) (mg kg)1)

Number of latent variables

Chlorophylls Total phenols

17 (11)* 297 (47)**

2 3

r

RMSECV (mg kg)1)

RMSECV (%)

Explained variance of Y (%)

0.991 0.688

1.4* 34.5**

8.2 11.6

96.9 50.7

*Milligram of pheophytin-a per kilogram of oil. **Milligram of gallic acid per kilogram of oil. RMSECV, Root Mean Square Error of Cross-Validation.



59


(a) CG2

2.0 × 106

CG1

Start S D1 D2 D4 D8 DG1 D6 D10

CG4 0.0

CG10 CG6 CG12

PC2 (1%)

60

D12DG2 DG4

–2.0 × 106

DG10 6

–4.0 × 10 –2 × 107

0

–1 × 107

1 × 107

PC1 (97%)

(b) 0.4

0.2 PC1 loading 0.0 0.4

the formation of hydroperoxides, which on further decomposition primes the radical oxidation reaction (Kiritsakis & Dugan, 1984, 1985). It is important to note that green glass bottles only partially prevent the decrease in the chlorophyll contents. Tocopherols, which absorb only in the u.v. region, cannot be destroyed by light directly; however, they do act both as electron donors, slowing down the oxidation reaction, and as electron acceptors, quenching or scavenging the singlet oxygen, with consequent inhibition of the photo-oxidation of lipids (Morello et al., 2004). Photo-oxidation is much faster than auto-oxidation, and implies a faster consumption of tocopherols in the samples exposed to light. Thus, we may infer that in the presence of light the oils are mainly protected from oxidation by tocopherols and carotenoids, and that phenolic substances should have a secondary role. It is known that the phenolic substances act by giving away an electron, thus they can interrupt the radical chain reaction of oxidation (Cuppett et al., 1997). The main reaction occurring in darkness is the auto-oxidation; here the phenolic substances seem to be involved more than other antioxidants in the protection of oils from oxidation. Conclusions

0.2 PC2 loading 0.0 300

400

500

600

700

Wavelength (nm) Figure 6 (a) Scores plot for the two most significant principal components, PC1 vs. PC2, of a principal component analysis of the synchronous scanning fluorescence (SSF) spectra (Dk ¼ 10 nm) of virgin olive oil samples. A nonstored sample (Start), and samples stored in different conditions: in darkness (D), in green glass bottles (DG) and in clear glass bottles (CG). The samples are numbered according to the months of storage. The values in brackets describe the fraction of the total variation explained by each of the PC. Each point represents an average of the spectra obtained from three independent samples. (b) One-vector loading plots for the PC1 and PC2 principal components.

exposure throughout the storage period. The observed sample grouping in the scores plot is in agreement with our conclusions arising from the analysis of the particular fluorophores in SSF, and the chemical nature of chlorophylls, tocopherols and phenolic compounds. The decrease of chlorophyll content in the samples exposed to light is in agreement with the results by other Authors (Psomiadou & Tsimidou, 1998), and is because of their photosensitising properties. Photo-oxidation of oils in the presence of chlorophylls leads to the formation of the highly reactive singlet oxygen that tends to react with the unsaturated fatty acids provoking


The present study demonstrates the potential of the fluorescence spectroscopy in the analysis of olive oils. Rapid fluorescence measurements can be used directly for monitoring the degradation processes in virgin olive oils during storage. TLS allows obtaining a comprehensive view of changes in fluorescent components during storage. SSF techniques enable monitoring of various compounds affecting the oil oxidation (including tocopherols, chlorophylls and possibly phenolic compounds) in a single scan, and their quantitative assessment after appropriate calibration. Moreover, PCA of the synchronous fluorescence spectra revealed systematic changes in the overall emission characteristics dependent on the storage conditions, such as exposure to light, and packaging. We conclude that fluorescence spectroscopy seems to be a promising alternative or at least a technique complementary to the analytical methods currently used in the evaluation of oil quality. Acknowledgment

A grant from the Polish State Committee for Scientific Research No. 2P06T 112 26 (2004–2006) is gratefully acknowledged. References AOCS (1993). Official methods and recommended practices of the American Oil Chemistry Society, Method Cc 13i-96, 4th edn. Washington, DC: AOCS.



Brereton, R.G. (2000). Introduction to multivariate calibration in analytical chemistry. Analyst, 125, 2125–2154. Caponio, F., Gomes, T. & Pasqualone, A. (2001). Phenolic compounds in virgin olive oils: influence on the degree of olive ripeness on organoleptic characteristic and shelf-life. European Food Research and Technology, 212, 329–333. Caponio, F., Gomes, T. & Summo, C. (2003). Assessment of the oxidative and hydrolytic degradation of oils used as liquid medium of in-oil preserved vegetables. Journal of Food Science, 68, 147–151. Caponio, F., Bilancia, M.T., Pasqualone, A., Sikorska, E. & Gomes, T. (2005). Influence of the exposure to light on extra virgin olive oil quality during storage. European Food Research and Technology, 221, 92–98. Chen, B.H. & Huang, J.H. (1998). Degradation and isomerization of chlorophyll a and beta-carotene as affected by various heating and illumination treatments. Food Chemistry, 62, 299–307. Chow, C.K. (1991). Vitamin E and oxidative stress. Free Radical Biology and Medicine, 11, 215–232. Cortesi, N., Ponziani, A. & Fedeli, E. (1981). Caratterizzazione degli oli vergini e raffinati mediante HPLC dei componenti polari. Nota preliminare. (Characterization of virgin and refined olive oils by HPLC of polar components: Note 1.) Riv Ital Sostanze Grasse, 58, 108–114. Cuppett, S., Schnepf, M. & Hall, C. (1997). Natural antioxidants – are they a reality? In: Natural Antioxidants: Chemistry, Health Effects, and Applications (edited by F. Shahidi). Pp. 12–24. Champaign, IL: American Oil Chemist’s Society for monograph. deMan, J.M. (1999). Principles of Food Chemistry, 3rd edn. New York, NY: Kluwer Academic/Plenum Publishers. Diaz, T.G., Meras, I.D., Correa, C.A., Roldan, B. & Caceres, M.I.R. (2003). Simultaneous fluorometric determination of chlorophylls a and b and pheophytins a and b in olive oil by partial least-squares calibration. Journal of Agricultural and Food Chemistry, 51, 6934– 6940. Engelsen, S.B. (1997). Explorative spectrometric evaluations of frying oil deterioration. Journal of the American Oil Chemists Society, 12, 1495–1508. Favier, F., Caporale, G. & Bertuccioli, M. (1994). Rapid determination of phenol content in extra virgin olive oil. Grasas y Aceites, 45, 68–70. Giungato, P., Notarnicola, L. & Colucci, L. (2002). Evaluation of fluorescence spectroscopy potential in edible oil analysis. In: Current Trends in Commodity Science, vol. 2 (edited by R. Zielinski). Pp. 513–518. Poznan: Poznan University of Economics Press. Gomes, T., Caponio, F. & Delcuratolo, D. (2003). Non-conventional parameters for quality evaluation of refined oils with special reference to commercial class olive oil. Food Chemistry, 83, 403–408. Guimet, F., Ferre, J., Boque, R. & Rius, F.X. (2004). Application of unfold principal component analysis and parallel factor analysis to the exploratory analysis of olive oils by means of excitation-emission matrix fluorescence spectroscopy. Analytica Chimica Acta, 515, 75–85. Gutierrez, M.C., Rubio, S., Gomezhens, A. & Valcarcel, M. (1987). Simultaneous determination of histidine and histamine by 2nd derivative synchronous fluorescence spectrometry. Talanta, 34, 325– 329. Keys, A. (1995). Mediterranean diet and public health: personal reflections. American Journal of Clinical Nutrition, 61, 1321–1323. Kikugawa, K. & Beppu, M. (1987). Involvement of lipid oxidation products in the formation of fluorescent and cross-linked proteins. Chemistry and Physics of Lipids, 44, 277–296. Kiritsakis, A. & Dugan, L.R. (1984). Effect of selected storage conditions and packaging materials on olive oil quality. Journal of the American Oil Chemists Society, 61, 1868–1870.

Kiritsakis, A. & Dugan, L.R. (1985). Studies in photooxidation of olive oil. Journal of the American Oil Chemists Society, 62, 892–896. Kyriakidis, N.B. & Skarkalis, P. (2000). Fluorescence spectra measurement of olive oil and other vegetable oils. Journal of AOAC International, 83, 1435–1439. Liang, J.H. (1999). Fluorescence due to interactions of oxidizing soybean oil and soy proteins. Food Chemistry, 66, 103–108. Manna, C., Galletti, P., Cucciolla, V., Moltedo, O., Leone, A. & Zappia, V. (1997). The protective effect of the olive oil polyphenol (3,4-Dihydroxyphenyl)Ethanol counteracts reactive oxygen metabolite-induced cytotoxicity in caco-2 cells. Journal of Nutrition, 127, 286–292. Morello, J.R., Motilva, M.J., Tovar, M.J. & Romero, M.P. (2004). Changes in commercial virgin olive oil (cv Arbequina) during storage, with special emphasis on the phenolic fraction. Food Chemistry, 85, 357–364. Ndou, T. & Warner, I.M. (1991). Applications of multidimensional absorption and luminescence spectroscopies in analytical chemistry. Chemical Reviews, 91, 493–507. Oldham, P.B., McCarroll, M.E., McGown, L.B. & Warner, I.M. (2000). Molecular fluorescence, phosphorescence, and chemiluminescence spectrometry. Analytical Chemistry, 72, 197R–209R. Papadopoulos, K., Triantis, T., Yannakopoulou, E., Nikokavoura, A. & Dimotikali, D. (2003). Comparative studies on the antioxidant activity of aqueous extracts of olive oils and seed oils using chemiluminescence. Analytica Chimica Acta, 494, 41–47. Patra, D. & Mishra, A.K. (2002). Recent developments in multicomponent synchronous fluorescence scan analysis. Trends in Analytical Chemistry, 12, 787–798. Psomiadou, E. & Tsimidou, M. (1998). Simultaneous HPLC determination of tocopherols, carotenoids, and chlorophylls for monitoring their effect on virgin olive oil oxidation. Journal of Agricultural and Food Chemistry, 46, 5132–5138. Psomiadou, E. & Tsimidou, M. (2001). Pigments in Greek virgin olive oils: occurrence and levels. Journal of the Science of Food and Agriculture, 81, 640–647. Rahmani, M. & Csallany, A.S. (1998). Role of minor constituents in the photooxidation of virgin olive oil. Journal of the American Oil Chemists Society, 75, 837–843. Sikorska, E., Gliszczynska-Swiglo, A., Khmelinskii, I.V. & Sikorski, M. (2005a). Synchronous fluorescence spectroscopy of edible vegetable oils. Quantification of tocopherols. Journal of Agricultural and Food Chemistry, 53, 6988–6994. Sikorska, E., Gorecki, T., Khmelinskii, I.V., Sikorski, M. & Koziol, J. (2005b). Classification of edible oils using synchronous scanning fluorescence spectroscopy. Food Chemistry, 89, 217–225. Sikorska, E., Romaniuk, A., Khmelinskii, I.V. et al. (2004). Characterization of edible oils using total luminescence spectroscopy. Journal of Fluorescence, 14, 25–35. Tavani, A. & La Vecchia, C. (1995). Fruit and vegetable consumption and cancer risk in a Mediterranean population. American Journal of Clinical Nutrition, 61, 1374–1377. Visoli, F. & Galli, C. (1998). Olive oil phenols and their potential effects on human health. Journal of Agricultural and Food Chemistry, 46, 4292–4296. Wheatley, R.A. (2000). Some recent trends in the analytical chemistry of lipid peroxidation. Trends in Analytical Chemistry, 19, 617–628. Wold, S., Esbensen, K. & Geladi, P. (1987). Principal component analysis. Chemometrics and Intelligent Laboratory Systems, 2, 37–52. Zandomeneghi, M., Carbonaro, L. & Caffarata, C. (2005). Fluorescence of vegetable oils: Olive oils. Journal of Agricultural and Food Chemistry, 53, 759–766.



61


62

Original article Heterocyclic amine formation during frying of frozen beefburgers Elna Persson1, Bea Kovaćsne´ Oroszva´ri2, Eva Tornberg2, Ingegerd Sjo¨holm2 & Kerstin Skog1* 1 Applied Nutrition and Food Chemistry, Department of Food Technology, Engineering and Nutrition, Lund Institute of Technology, Lund University, PO Box 124, SE-221 00 Lund, Sweden 2 Food Engineering, Department of Food Technology, Engineering and Nutrition, Lund Institute of Technology, Lund University, PO Box 124, SE-221 00 Lund, Sweden (Received 09 February 2006; Accepted in revised form 23 June 2006)

Summary

The formation of heterocyclic amines (HCAs) was studied during frying of beefburgers with different fat contents (6.7%, 16.1% and 39%). Beefburgers were fried from the frozen state for 60, 90, 120 s, and until the centre temperature had reached 72 C (approximately 150 s) in a double-sided pan fryer. The beefburgers were analysed for HCAs with solid-phase extraction and LC/MS detection, and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (4,8-DiMeIQx), 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), 1-methyl-9H-pyrido[3,4-b]indole (Harman) and 9H-pyrido[3,4-b]indole (Norharman) were detected in all samples. The concentrations of HCAs ranged between 0 and 2.3 ng g)1. The HCAs concentrations were fitted to a first-order reaction model. The amounts of HCAs in beefburgers fried from the frozen state were in the same range as in beefburgers that have reached room temperature before frying. Furthermore, it was found that the formation of HCAs is not only concentration-controlled but also mass transport-controlled and that kinetic models stated in earlier studies fit relatively well our analysed values on HCAs in fried beefburgers.

Keywords

Beefburger, fat content, frying, frozen, Harman, heterocyclic amine formation, kinetic, MeIQx, Norharman, PhIP.

Introduction

The trend of eating readymade food products is increasing (Borgen & Skog, 2004), and consumers who rely on industrially prepared foods, have the right to demand safe and nutritious food. Fried beefburgers is one of the most common meat dishes in the Western diet (Voskuil et al., 1999; Busquets et al., 2004). Heterocyclic amines (HCAs) are mutagenic/carcinogenic compounds that are formed at ppb levels during cooking of meat (Sugimura, 2000). Frying produces higher levels of HCAs than other cooking methods and fried meat is responsible for a large proportion of our HCAs intake (Keating & Bogen, 2001). HCAs are formed from free amino acids, glucose and creatine or creatinine; compounds that are normally present in muscle tissue (Ja¨gerstad et al., 1983). The precursors are water-soluble and are transported to the surface during cooking. HCAs are mainly formed in the crust owing to the higher surface temperature and drier conditions. More than twenty different HCAs are known, the most common ones being 2-amino-3,8*Correspondent: Fax: +46 46 222 4532; e-mail: [email protected]

dimethylimidazo[4,5-f]quinoxaline (MeIQx), 2-amino3,4,8-trimethylimidazo[4,5-f]quinoxaline (4,8-DiMeIQx) and 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) (Skog et al., 1998). Harman (1-methyl-9H-pyrido[3,4-b]indole) and Norharman (9H-pyrido[3,4-b]indole) are two other common HCAs that are co-mutagenic and they may also act as neurotoxins (de Meester, 1995; Kuhn et al., 1996). Based on animal studies, the International Agency for Research on Cancer (IARC) has classified one HCAs as a probable (2A) and eight HCAs as possible (2B) human carcinogens (IARC, 1993). Several factors are known to influence the formation of HCAs; cooking time and temperature being the most important (Skog et al., 1998). The choice of cooking method affects the concentration of HCAs; frying and grilling producing most HCAs (Sinha et al., 1995, 1998a,b; Skog et al., 1997; Solyakov & Skog, 2002). The amounts and relative concentrations of precursors are important. In fried beef patties prepared from mixtures of different beef tissues, and thus with different precursor concentrations, the creatine content was shown to explain most of the mutagenic activity (Laser Reuterswa¨rd et al., 1987b). High natural levels of glucose have been shown to reduce the concentration of PhIP in both pork and chicken (Olsson et al., 2002;


HCAs formation during frying frozen beefburgers E. Persson et al.

Solyakov & Skog, 2002). Free amino acids are important for the formation of HCAs, but in meat cooking experiments, no single amino acid has been shown to act as a key precursor (Laser Reuterswa¨rd et al., 1987b; Olsson et al., 2002). The influence of fat content on the formation of cooked food mutagens has not yet been clearly elucidated. A higher fat content means a lower content of muscle meat, which may lead to lower amounts of creatine, glucose and free amino acids. Literature data on mutagenic activity in relation to fat content are not consistent. One study reported the mutagenic activity to be almost independent of fat content in fried beef patties (Bjeldanes et al., 1983). In another study, beef patties containing 0–35% added fat, a peak in mutagenic activity was found at 10% fat (Spingarn et al., 1980). In a study of fried beef, containing 8%, 15% and 30% fat, the beefburger with 15% fat was found to have the highest mutagenic activity (Knize et al., 1985). A decrease in mutagenic activity with increasing fat content (2.8–16.6%) was found in oven roasted meat loaf (Holtz et al., 1985). There are few reports on the formation of individual HCAs in relation to fat content in fried meat. The concentration of 2-amino-3-methylimidazo[4,5-f]quinoline (IQ) was found to be lower in low-fat beef patties than in high-fat ones (Barnes & Weisburger, 1983), and IQ formation was shown to increase with fat content in fried meat (Johansson & Ja¨gerstad, 1994). On the contrary, beefburgers containing 5% fat were found to have higher levels of HCAs than those containing 15% fat and this result was explained by the dilution of precursors (Abdulkarim & Scott Smith, 1998). Thus, it is difficult to distinguish between the physical (more efficient heat transfer) and chemical effects of fat on the formation of HCAs. To assess the health risk of HCAs, estimation of HCAs intake is necessary and it is important to know how cooking conditions, meat composition and fat content affect the concentrations. Reliable data on HCAs concentrations in meat, together with data on dietary habits may allow accurate assessment of HCAs intake. This information can be used in epidemiological research to examine the relation between HCAs and human cancer. The kinetics of HCAs formation has been studied previously in chemical and meat juice model systems, although not in real meat cooking experiments. Arvidsson et al. (1997) showed the formation of HCAs in a model system to fit first-order reaction kinetics, suggesting a monomolecular or bimolecular reaction in which one component is in large excess (Arvidsson et al., 1997, 1999). In another kinetic study, increasing fat content was shown to reduce the concentration of HCAs as well as the activation energy in heated meat emulsions (Hwang & Ngadi, 2002).


Industrially prepared beefburgers are generally taken out of the freezer just before frying, i.e. the patties have not reached room temperature, but so far no data have been published on the formation of HCAs fried from the frozen state. The objectives of the present study were to examine the formation of HCAs during frying of frozen minced beef patties in relation to different fat contents and to estimate kinetic parameters. Materials and methods

Chemicals

Solvents and chemicals were of HPLC or analytical grade. Water was passed through a Milli-Q water purification system (Millipore, Bedford, MA, USA). The following HCAs were used as reference compounds: IQ, MeIQ, MeIQx, 4,8-DiMeIQx, PhIP, Harman and were purchased from Toronto Research Chemicals (Toronto, Canada). Norharman was purchased from Aldrich (Steinhem, Germany). The chemical purity of the synthetic references was higher than 99%, according to the manufacturers. This was confirmed using HPLC with UV detection for each of the reference compounds. A mixture of the different HCAs in MeOH (2 ng of each compound per microlitre) was used as a spiking mixture. Chemicals used for the analysis of creatine and creatinine were purchased from Roche Diagnostics Scand AB (Bromma, Sweden). Caffeine was obtained from Sigma (Stockholm, Sweden). The following materials were used for solid-phase extraction: diatomaceous earth (Isolute), obtained from Sorbent AB (Va¨stra Fro¨lunda, Sweden) and PRS (propylsulfonic acid silica) and C18 columns (Varian), from Scantech Lab (Partille, Sweden). Beefburgers

The sample material was a subsample from a larger study designed to investigate the characteristics of heat and mass transfer during frying (Kovaćsne´ et al., 2005). Three types of meat muscle (shank, rib, brisket) were used to prepare low-, medium- and high-fat beefburgers. The meat was first minced through both an 8-mm orifice plate and afterwards a 3-mm grinder plate was used. The brisket was minced with tallow to increase the fat content. The final fat contents of shank, rib and brisket were 6.7%, 16.1% and 39%, respectively. After grinding, 1.5% of salt was added to the batches, which were vacuum-packed and stored at )45 C until required. The minced meat was thawed and beefburgers were prepared using three plastic moulds having a diameter of 10 cm and different heights, 2, 3 and 5 mm, respectively. They were placed on top of each other with K-type thermocouples in between. In this way it was possible to ensure the exact position of the thermocouples in the


63

64


centre and 2 mm from the beefburger surface (for further details, see Kova´scne´ et al., 2005). The beefburgers were stored in a freezer to reach )20 C before frying. A double-sided, thermostat-controlled, Teflon-coated square frying pan, 235 · 235 mm, was used for the experiments and the temperature was recorded by thermocouples (Persson et al., 2003). The frozen beefburgers were fried simultaneously on both sides at 175 C for 60, 90 and 120 s, and until the centre temperature reached 72 C (approximately 150 s). No frying fat was used. The distance between the top and bottom plates was kept constant by a 10-mm thick Teflon frame. Extraction and analysis of HCAs

HCAs were extracted from the beefburgers using the solid-phase extraction method developed by Gross & Gru¨ter (1992), with slight modification (Torbio et al., 1999). Briefly, samples were homogenised in 1 m NaOH and mixed with diatomaceous earth. The HCAs were extracted with ethyl acetate and adsorbed onto cartridges containing PRS, a strong cation exchanger. This method gave two extracts, one containing PhIP and the IQx compounds, and one containing Norharman and Harman. Extraction recovery rates for the different HCAs were determined by the addition of 100 lL spiking mixture to one sample, extracted in parallel with three unspiked samples, and the samples were analysed in triplicate. The extracts were evaporated to dryness under nitrogen and thereafter dissolved in 100 lL of caffeine solution to correct for variations in injection volume. The samples were analysed using a liquid chromatography/mass spectroscopy (LC/MS) system as described previously (Ba˚ng et al., 2002). Liquid chromatography was performed with a Zorbax SB-C8 (150 · 4.6 mm, i.d. 5 lm; Agilent Technologies, Palo Alto, CA, USA) column. The eluent phase was a combination of water (adjusted to pH 3.5 with acetic acid) and acetonitrile. An ion-trap mass detector, (LCQDECA from Thermo Finnigan, San Jose´, CA, USA) with an electrospray ion source (ESI) was used, and single-ion monitoring (SIM) was performed. The HCAs were quantified by determining the peak areas and the results were corrected for incomplete recovery. Analysis of fat, creatine and glucose

The raw meat was analysed for glucose according to the glucose oxidase–peroxidase method, and creatine and creatinine by an enzymatic method (Boehringer Mannheimer) as described earlier (Arvidsson et al., 1997). All values were determined in duplicate. The absorbance was measured using a UV spectrophotometer (Perkin


Elmer, Lambda 10). The fat content in the raw beefburgers was determined (Kovaćsne´ et al., 2005). Kinetics of HCAs formation

Based on results from model experiments, the HCAs formation can be described as a first-order reaction (Arvidsson et al., 1997, 1999): C ¼ Að1 ekt Þ where C is the concentration of the HCAs (ng g)1 raw meat), A is the estimated maximum concentration of the HCAs (ng g)1 raw meat), k is the rate constant for formation of the HCAs (s)1) and t is the cooking time (s). Values of the maximum concentration (A) and the rate constant for formation (k) were obtained by using nonlinear regression with Matlab 6.5. Data on HCAs formation from a meat model study (Arvidsson et al., 1999) were used to estimate HCAs concentrations at various time points. Results

Analysis of HCAs

In the fried beefburgers from shank, rib and brisket, the concentrations of the mutagenic HCA MeIQx, 4,8DiMeIQx and PhIP ranged from 0.1 to 1.6 ng g)1, and the co-mutagens Harman and Norharman ranged from 0.1 to 2.3 ng g)1. The data are presented in Fig. 1. IQ and MeIQ were not detected in any samples. The concentrations of MeIQx, 4,8-DiMeIQx and PhIP were practically the same in the beefburgers from rib and brisket. The concentration of HCAs generally increased with increasing frying time, although a decrease was observed in some of the HCAs at the longest frying time, and this decrease was most pronounced for PhIP. The concentrations of HCAs were corrected for incomplete recovery. The recovery rates of MeIQx, 4,8-DiMeIQx, PhIP, Norharman and Harman in the solid-phase extraction method were 66 ± 12%, 63 ± 12%, 48 ± 13%, 51 ± 14% and 44 ± 9%, respectively. Estimated HCAs concentrations (Arvidsson et al., 1999) are shown in Fig. 1. Kinetics

The kinetics of HCAs formation in fried beefburgers was described using the first-order kinetic model. The kinetic parameters are presented in Table 1. Generally, the rate of formation (k) was highest in beefburgers from brisket and lowest in beefburgers from rib, except for PhIP and 4,8-DiMeIQx. The highest values of maximum concentration (A) were obtained for beefburgers from rib, except for Norharman. The lowest values of maximum concentration were obtained for


ng/g–1 raw meat


1.0

MeIQx

0.8 0.6 0.4 0.2

MeIQx

0.0 50

60

70

80

90

100 110 120 130 140 150 160

Time (s) ng/g–1 raw meat

Table 1 Calculated kinetic parameters for formation of heterocyclic amines (HCAs) in fried frozen beefburgers

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 50

4,8-DiMeIQx

4.8-DiMeIQx PhIP

Harman 60

70

80

90

100 110 120 130 140 150 160

ng/g–1 raw meat

Time (s) PhIP

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 50

60

70

80

90

100 110 120 130 140 150 160

ng/g–1 raw meat

Time (s) Harman

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

A (ng g)1)

k (s)1)

R2

Shank Rib Brisket Shank Rib Brisket Shank Rib Brisket Shank Rib Brisket Shank Rib Brisket

0.013 0.064 0.006 0.016 0.084 0.013 0.008 0.084 0.025 0.045 0.183 0.015 0.438 0.426 0.121

0.026 0.009 0.028 0.025 0.003 0.018 0.042 0.033 0.016 0.020 0.006 0.022 0.012 0.011 0.014

0.80 0.84 0.92 0.87 0.34 0.86 0.94 0.92 0.65 0.74 0.92 0.99 0.83 0.71 0.67

Abbreviations: MeIQx, 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline; 4,8-DiMeIQx, 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline; PhIP, 2amino-1-methyl-6-phenylimidazo[4,5-b]pyridine; Harman, 1-methyl-9Hpyrido[3,4-b]indole; Norharman, 9H-pyrido[3,4-b]indole.

the calculations for PhIP in beefburgers from shank and rib. Concentration of glucose and creatine in the raw beefburgers

50

60

70

80

90

100 110 120 130 140 150 160

Time (s) Norharman

2.5

ng/g–1 rawemeat

Norharman

Beefburger

2.0 1.5

The concentration of glucose in the raw beefburgers from shank, rib and brisket was 4.0 ± 0.4, 3.8 ± 0.3 and 4.0 ± 0.3 lmol g)1 meat wet weight, respectively. The concentration of creatine in shank, rib and brisket was 22.3 ± 0.1, 22.9 ± 0.1 and 19.8 ± 0.1 lmol g)1 meat wet weight. The amount of creatinine was below the detection limit (0.1 lmol g)1 meat wet weight).

1.0

Weight loss during cooking

0.5 0.0 50

60

70

80

90

100 110 120 130 140 150 160

Time (s) Figure 1 Concentration of MeIQx (2-amino-3,8-dimethylimidazo[4,5f]quinoxaline), 4,8-DiMeIQx (2-amino-3,4,8-trimethylimidazo[4,5f]quinoxaline), PhIP (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine), Harman (1-methyl-9H-pyrido[3,4-b]indole) and Norharman (9Hpyrido[3,4-b]indole) in beefburgers (ng g)1 raw meat) fried at 175 C for 60–150 s. Concentrations are shown for beefburgers from shank, 6.7% fat (filled diamond); rib, 16.1% fat (filled square) and brisket, 39% fat (filled triangle). Estimated concentration data from Arvidsson et al. (1999), marked with (X), are presented for MeIQx, 4,8-DiMeIQx and PhIP. Because of interfering peaks, Harman was not quantified for the longest frying time.

beefburgers from brisket, except for PhIP. The model used did not take into account any degradation; thus, the frying time around 150 s was not included in


The total weight loss (fat + water) for the shank, rib and brisket beefburgers was 7.2–25.0%, 9.9–37.0% and 13.9–47.6%, respectively. The water loss based on the initial water content was the same for all three types of beefburgers and was not influenced by the fat content (Kovaćsne´ et al., 2006a). Discussion

Commercially prepared beefburgers are often fried straight from the freezer, but there are no reports explicitly considering the concentrations of HCAs in beefburgers fried from the frozen state. The concentrations of HCAs found in this investigation are in the same range as in beefburgers that have reached room temperature before frying (Skog et al., 1998; Persson et al., 2003). Of the two co-mutagens, Norharman was


65

66


generally formed in higher amounts than Harman. This is in agreement with results from other studies (Johansson et al., 1995; Totsuka et al., 1999). The recovery of HCAs during the solid-phase extraction in this study was comparable with other reports (Johansson et al., 1995; Abdulkarim & Scott Smith, 1998; Borgen et al., 2001; Klassen et al., 2002). The recovery was similar for all samples and was not affected by the fat content. Variations in HCAs recoveries have been observed depending on the meat analysed (Busquets et al., 2004). We wanted to study the development of HCAs concentration with frying time and the beefburgers cooked for 120 s or less were not ready to eat. Interestingly HCAs were detected already after 60 s. In the samples fried for 60 and 90 s, low amounts of MeIQx, 4,8-DiMeIQx and PhIP (£0.2 ng g)1), were detected, but after 120 s, higher amounts (up to 1.6 ng g)1) were found. When the beefburgers are placed in the pan, the pan temperature initially falls (Persson et al., 2002), which may explain the low amounts of HCAs at the shorter cooking times. Kovaćsne´ et al. (2006a,b) have shown that up to about 90 s of frying, most of the mass transfer of water, containing HCAs precursors, is directed inwards towards the centre of the beefburger. This phenomenon could be one of the reasons for the low amounts of HCAs formed during that period of frying. After crust formation has started the mass transport of water turns towards the frying surface from the centre and the higher the water content. Thus more precursors are transferred to the surface for HCAs formation. This phenomenon might explain the observed but vague fat dependence on HCAs formation. From these observations, it can be suggested that HCAs formation is not only concentration-controlled but also mass transport-controlled, which agrees with earlier findings that the drip loss controls HCAs formation (Persson et al., 2003). In some cases, the beefburgers fried for about 150 s contained smaller amounts of PhIP and Norharman than the beefburgers fried for 120 s. Other authors have also observed a decrease in HCAs concentration with time, although at higher temperatures or longer cooking times than in this study. Decreased HCAs formation has been observed with longer cooking time in pan-broiled salmon (Gross & Gru¨ter, 1992) and with increased temperature in pan-fried herring, chicken breast and pan residue of lamb chops (Skog et al., 1997). Decreased HCAs concentrations have been observed in model systems (Jackson & Hargraves, 1995; Arvidsson et al., 1997; Pais et al., 2000; Bordas et al., 2004). The phenomenon of decreasing amounts of HCAs may be because of the degradation of HCAs and further reactions forming other products. Generally, the amounts of HCAs were highest in the beefburgers with the lowest fat content, but the


variations between the different beefburgers were not significant. In other studies, it has been shown that creatine or creatinine is essential for the formation of HCAs (Laser Reuterswa¨rd et al., 1987a) and that the amount of creatine in the meat may affect the HCAs formation (Olsson et al., 2002; Solyakov & Skog, 2002). In the present study, the concentrations of creatine and glucose in the three beef muscles were similar and in accordance with results reported by other authors (Laser Reuterswa¨rd et al., 1987b; Pais et al., 1999). As the concentrations of creatine and glucose were the same for all beefburgers, their different HCAs content could not be attributed to differences in precursor concentration in the total of the beefburger. The analysed data on concentrations of HCAs in ready-to-eat beefburgers were compared with data from a kinetic study using a meat model system (Arvidsson et al., 1999). The concentration data for MeIQx, 4,8DiMeIQx and PhIP from that study were recalculated and expressed as ng g)1 raw meat. As can be seen in Fig. 1, the theoretical concentrations calculated from the kinetic model are in the same range as the experimental concentrations found in the present study. This is despite the fact that in the present much shorter cooking times were used. Furthermore, on crust formation in a beefburger, much drier conditions occur than in the meat model system and dry conditions enhance formation of, e.g. PhIP. The general conclusions from this study are that the amounts of HCAs in beefburgers fried from frozen are in the same range as in beefburgers that have reached room temperature before frying. There was a tendency that low-fat beefburgers contained higher amounts of HCAs, especially PhIP and Norharman. Furthermore, it was found that the formation of HCAs is not only concentration-controlled but also mass transport-controlled (Kovaćsne´ et al., 2006b) and that kinetic models stated in earlier studies (Arvidsson et al., 1997, 1999) fit relatively well with our analysed values on HCAs in fried beefburgers. Acknowledgments

This work was carried out with financial support from the Commission of the European Communities, specific RTD program ‘Quality of Life and Management of Living Resources’, QLK1-CT99-001197, ‘Heterocyclic Amines in Cooked Foods – Role in Human Health’. It does not necessarily reflect its views and in no way anticipates the Commission’s future policy in this area. References Abdulkarim, B.G. & Scott Smith, J. (1998). Heterocyclic amines in fresh and processed meat products. Journal of Agricultural and Food Chemistry, 46, 4680–4687.



Arvidsson, P., van Boekel, M.A.J.S., Skog, K., Solyakov, A. & Ja¨gerstad, M. (1997). Kinetics of formation of polar heterocyclic amines in a meat model system. Journal of Food Science, 62, 911– 916. Arvidsson, P., van Boekel, M.A.J.S., Skog, K., Solyakov, A. & Ja¨gerstad, M. (1999). Formation of heterocyclic amines in a meat juice model system. Journal of Food Science, 64, 216–221. Ba˚ng, J., Nukaya, H. & Skog, K. (2002). Blue chitin columns for the extraction of heterocyclic amines from cooked meat. Journal of Chromatography A, 977, 97–105. Barnes, W.S. & Weisburger, J.H. (1983). Lipid content and mutagen formation in the cooking of beef. Proceedings of AACR, 95. Bjeldanes, L.F., Morris, M.M., Timourian H. & Hatch, F.T. (1983). Effects of meat composition and cooking conditions on mutagen formation in fried ground beef. Journal of Agricultural and Food Chemistry, 31, 18–21. Bordas, M., Moyano, E., Puignou, L. & Galceran, M.T. (2004). Formation and stability of heterocyclic amines in a meat flavour model system - Effect of temperature, time and precursors. Journal of Chromatography B, 802, 11–17. Borgen E. & Skog K. (2004). Heterocyclic amines in hamburgers from restaurants and fast-food chains. Molecular Nutrition and Food Research, 48, 292–298. Borgen, E., Solyakov, A. & Skog, K. (2001). Effects of precursor composition and water on the formation of heterocyclic amines in meat model systems. Food Chemistry, 74, 11–19. Busquets, R., Bordas, M., Toribio, F., Puignou, L., Galceran, M.T. (2004). Occurrence of heterocyclic amines in several home-cooked meat dishes of the Spanish diet. Journal of Chromatography B, 802, 79–86. Gross, G.A. & Gru¨ter, A. (1992). Quantification of mutagenic/ carcinogenic heterocylic aromatic amines in food products. Journal of Chromatography, 592, 271–278. Holtz, E., Skjo¨ldebrand, C., Ja¨gerstad, M., Laser Reutersva¨rd, A. & Isberg, P.E. (1985). Effect of recipes on crust formation and mutagenicity in meat products during baking. Journal of Food Technology, 20, 57–66. Hwang, D.K. & Ngadi, M. (2002). Kinetics of heterocyclic amines formation in meat emulsion at different fat content. Lebensmittel Wissenschaft und Technologie, 35, 600–606. International Agency for Research on Cancer (1993). Monographs on the evaluation of carcinogenic risk to humans. Lyon, 163–242. Jackson, L.S. & Hargraves, W.A. (1995). Effects of time and temperature on the formation of MeIQx and DiMeIQx in a model system containing threonine, glucose and creatine. Journal of Agricultural and Food Chemistry, 43, 1678–1684. Ja¨gerstad, M., Laser Reutersva¨rd, A., et al. O¨ste, R. (1983). Creatinine and Maillard reaction products as precursors of mutagenic compounds formed in fried beef. In: The Maillard reaction in Foods and Nutrition ACS Symposium Series (edited by G. Waller & M. Feather). Pp. 507–519. Washington, DC: Amer Chemical Society. Johansson, M.A. & Ja¨gerstad, M. (1994). Occurrence of mutagenic/ carcinogenic heterocyclic amines in meat and fish products, including pan residues, prepared under domestic conditions. Carcinogenesis, 15, 1511–1518. Johansson, M.A., Fredholm, L., Bjerne, I. & Ja¨gerstad, M. (1995). Influence of frying fat on the formation of heterocyclic amines in fried beefburgers and pan residues. Food and Chemical Toxicology, 33, 993–1004. Keating, G.A. & Bogen, K.T. (2001). Methods for estimating heterocyclic amine concentrations in cooked meats in the US diet. Food and Chemical Toxicology, 39, 29–43. Klassen, R.D., Lewis, D., Lau, B.P.-Y. & Sen, N.P. (2002). Heterocyclic aromatic amines in cooked hamburgers and chicken obtained from local fast food outlets in the Ottawa region. Food Research International, 35, 837–847. Knize, M.G., Andresen, B.D., Healy, S.K. et al. (1985). Effects of temperature, patty thickness and fat content on the production of


mutagens in fried ground beef. Food and Chemical Toxicology, 23, 1035–1040. Kovaćsne´, O.B., Sjo¨holm, I. & Tornberg, E. (2005). The mechanisms controlling heat and mass transfer on frying on beefburgers I. The influence of the composition and comminution of meat raw material. Journal of Food Engineering, 67, 499–506. Kovaćsne´, O.B., Tejedor, E.B., Sjo¨holm, I. & Tornberg, E. (2006a). Permeability and mass transfer as a function of the cooking temperature during the frying of beefburgers. Journal of Food Engineering, 74, 1–12. Kovaćsne´ O.B., Bayod, E., Sjo¨holm, I. & Tornberg, E. (2006b) The mechanisms controlling heat and mass transfer on frying of beefburgers. III. Mass transfer evolution during frying. Journal of Food Engineering, 76, 169–178. Kuhn, W., Muller, T., Grosse, H. & Rommelspacher, H. (1996). Elevated levels of harman and norharman in cerebrospinal fluid of parkinsonian patients. Journal of Neural Transmission, 103, 1435– 1440. Laser Reuterswa¨rd, A., Skog, K. & Ja¨gerstad, M. (1987a). Effects of creatine and creatinine content on the mutagenic activity of meat extracts, bouillons and gravies from different sources. Food and Chemical Toxicology, 25, 747–754. Laser Reuterswa¨rd, A., Skog, K. & Ja¨gerstad, M. (1987b). Mutagenicity of pan-fried bovine tissues in relation to their content of creatine, creatinine, monosaccharides and free amino acids. Food and Chemical Toxicology, 25, 755–762. de Meester, C. (1995). Genotoxic potential of beta-carbolines, a review. Mutation Research, 339, 139–153. Olsson, V., Solyakov, A., Skog, K., Lundstro¨m, K. & Ja¨gerstad, M. (2002). Natural variations of precursors in pig meat affect the yield of heterocyclic amines – effects of RN genotype, feeding regime, and sex. Journal of Agricultural and Food Chemistry, 50, 2962–2969. Pais, P., Salmon, C.P., Knize, M.G. & Felton, J.S. (1999). Formation of mutagenic/carcinogenic heterocyclic amines in dry-heated model systems, meats, and meat drippings. Journal of Agricultural and Food Chemistry, 47, 1098–1108. Pais, P., Tanga, M.J., Salmon, C.P. & Knize, M.G. (2000). Formation of the mutagen IFP in model systems and detection in restaurant meats. Journal of Agricultural Food Chemistry, 48, 1721–1726. Persson, E., Sjo¨holm, I. & Skog, K. (2002). Heat and mass transfer in chicken breasts, effect on PhIP formation. Zeitschrift fu¨r Lebensmittel-Untersuchung und Forschung, 214, 455–459. Persson, E., Sjo¨holm, I. & Skog, K. (2003). Effect of high waterholding capacity on the formation of heterocyclic amines in fried beefburgers. Journal of Agricultural and Food Chemistry, 51, 4472– 4477. Sinha, R., Rothman, N., Brown, E.D. et al. (1995). High concentrations of the carcinogen 2-amino-1-methyl-6-phenylimidazo-[4,5b]pyridine (PhIP) occur in chicken but are dependent on the cooking method. Cancer Research, 55, 4516–4519. Sinha, R., Knize, M.G., Salmon, C.P. et al. (1998a). Heterocyclic amine content of pork products cooked by different methods and to varying degrees of doneness. Food and Chemical Toxicology, 36, 289–297. Sinha, R., Rothman, N., Salmon, C.P. et al. (1998b). Heterocyclic amine content in beef cooked by different methods to varying degrees of doneness and gravy made from meat drippings. Food and Chemical Toxicology, 36, 279–287. Skog, K., Augustsson, K., Steinbeck, G., Stenberg M. & Ja¨gerstad, M. (1997). Polar and non-polar heterocyclic amines in cooked fish and meat products and their corresponding pan residues. Food and Chemical Toxicology, 35, 555–565. Skog, K.I., Johansson, M.A. & Ja¨gerstad, M.I. (1998). Carcinogenic heterocyclic amines in model systems and cooked foods, a review on formation, occurrence and intake. Food and Chemical Toxicology, 36, 879–896.


67

68


Solyakov, A. & Skog, K. (2002). Screening for heterocyclic amines in chicken cooked in various ways. Food and Chemical Toxicology, 40, 1205–1211. Spingarn, N.E., Kasai, H., Vuolo, L.L. et al. (1980). Formation of mutagens in cooked foods. III. Isolation of a potent mutagen from beef. Cancer Letters, 9, 177–183. Sugimura, T. (2000). Nutrition and dietary carcinogens. Carcinogenesis, 21, 387–395. Torbio, F., Puignou, L. & Galceran, M.T. (1999). Evaluation of different clean-up procedures for the analysis of heterocyclic


aromatic amines in a lyophilized meat extract. Journal of Chromatography A, 836, 223–233. Totsuka, Y., Ushiyama, H., Ishihara, J. et al. (1999). Quantification of the co-mutagenic beta-carbolines, norharman and harman, in cigarette smoke condensates and cooked foods. Cancer Letters, 143, 139–143. Voskuil, D.W., Augustsson, K., Dickman, P.W., van’t Veer, P. & Steineck, G. (1999). Assessing the human intake of heterocyclic amines, limited loss of information using reduced sets of questions. Cancer Epidemiology Biomarkers and Prevention, 8, 809–814.



69

Original article Development of an indirect a-actinin-based immunoassay for the evaluation of protein breakdown and quality loss in fish species subjected to different chilling methods Mońica Carrera, Vanesa Losada, Jose´ Manuel Gallardo, Santiago P. Aubourg & Carmen Pin˜eiro* Instituto de Investigaciones Marinas (IIM-CSIC), c/Eduardo Cabello 6. 36208 Vigo, Spain (Received 14 July 2005; Accepted in revised form 24 July 2006)

Summary

a-Actinin release from the myofibrillar protein fraction to the sarcoplasm can be considered as an accurate proteolysis index in seafood muscle. The main objective of the present study was to develop a specific enzyme-linked immunosorbent assay (ELISA), based on the use of a monoclonal antibody against a-actinin to evaluate the degree of proteolysis in two different chilled fish species – European hake (Merluccius merluccius) and turbot (Psetta maxima) – kept under two different storage systems: flake ice and slurry ice. Comparison with sensory assessment, K-value and sarcoplasmic protein profiles was carried out. A different degree of proteolysis could be observed in both fish species; thus, the immunoassay was shown to be useful in monitoring the protein degradation events in hake muscle especially under flake ice storage. In the case of turbot, as very low proteolysis development could be obtained, the assay was not suitable for assessing quality changes. A different break point of immunoassay values for each fish species is suggested.

Keywords

a-Actinin, ELISA, hake, liquid ice, quality, refrigeration, turbot.

Introduction

The postmortem tenderization of fish muscle is one of the major problems related to the loss of freshness and quality in chilled seafood products. One of the main causes of such tenderization is the breakdown of Z-line structures of the muscle fibre (Ando et al., 1991). The principal protein of Z-line is a-actinin. In fish muscle, aactinin (100 kDa, pI 5.6) (Papa et al., 1995), represents 2% of total myofibrillar protein content (Takahashi & Hatori, 1992). This myofibrillar protein has been reported to be involved in the anchorage of end-to-end actin filaments with opposite polarity to the Z-line between adjacent sarcomeres. In addition, this protein is bound to elastic titin filaments in the Z-line by the Nterminal part and M-line-associated proteins in the middle of the sarcomere by the C-terminal end of the same actinin molecule (Trinick, 1991; Small et al., 1992). a-Actinin has also been reported to play a key role in postmortem changes affecting the Z-line structure (Astier et al., 1991; Seki & Tsuchiya, 1991; Papa et al., 1996). Previous reports have proposed that the release of a-actinin from the myofibrillar protein fraction depends on different proteolytic mechanisms, such as the activity of endogenous proteases like calpains and cathepsins *Correspondent: Fax: +34 986 292762; e-mail: [email protected]

(Goll et al., 1992; Taylor et al., 1995; Aoki et al., 2000; Lamare et al., 2002; Verrez-Bagnis et al., 2002; Ladrat et al., 2003; Delbarre-Ladrat et al., 2004a,b). With a view to slowing down such fish deterioration, different methods during the chilling storage (Ashie et al., 1996) such as traditional flake ice (Nunes et al., 1992), refrigerated sea water (Kraus, 1992) and the use of chemical additives (Ponce de Leoń et al., 1993; Hwang & Regenstein, 1995) have been applied. Recently, the use of slurry ice (an ice-water suspension at subzero temperature) – also known as fluid ice, flow ice, slush ice or liquid ice – has been reported to be a promising technique for the preservation of aquatic food products (Chapman, 1990; Harada, 1991; Huidobro et al., 2002). In previous reports, our group has studied the quality loss of different seafood species when refrigerated under flake and slurry ice conditions (Losada et al., 2004; Pin˜eiro et al., 2004; Aubourg et al., 2005; Rodrı´ guez et al., 2006). Currently, the methods employed for monitoring changes associated with quality loss in fish can be classified as sensory, physical, physico-chemical, chemical and microbiological (Olafsdo´ttir et al., 1997). Among these, the study of the release of a-actinin into the sarcoplasmic fraction has been considered as a useful procedure for evaluating the rate of proteolysis in seafood, this protein being proposed as a potential biomarker of quality and freshness in chilled fish


70

a-actinin assessment in chilled fish by specific ELISA M. Carrera et al.

(Bandman & Zdanis, 1988; Tsuchiya et al., 1992; Papa et al., 1995, 1996; Verrez-Bagnis et al., 1999; Morzel et al., 2000; Delbarre-Ladrat et al., 2004b) even in the earliest postmortem steps. To our knowledge, no previous method based on the detection of a-actinin by means of a specific enzyme-linked immunosorbent assay (ELISA) has been developed up to now with the purpose of evaluating the rate of proteolysis in chilled fish species. In a first attempt (Carrera et al., 2004), our group applied this immunoassay procedure to evaluate the a-actinin release in three different fish species (hake, turbot and horse mackerel) during chilled storage under flake ice conditions. As a result, species showing a low fat content (hake and turbot) provided promising correlation values between the optical densities derived from ELISA and the traditional quality-loss indices. The present study focused on both lean species, taking into account the effect that differences such as the nature of the fish species and size may have on autolytic degradation and quality loss (Huss, 1999; Dalgaard, 2000). The main objective of this study was to assess the evolution of the loss of freshness in hake and turbot during the chilled storage under two different systems (flake ice and slurry ice) by the above-mentioned ELISA. Comparison with results obtained from sarcoplasmic protein electrophoretic profiles, nucleotide degradation and sensory assessment was carried out. Materials and methods

Fish material, processing and sampling

European hake (Merluccius merluccius) specimens were caught off the Atlantic coast of north-western Spain and kept in flake ice until they were brought to our laboratory (6 h later). Farmed turbot (Psetta maxima) specimens were obtained from Stolt Sea Farm, S.A. (Carnota, Spain) and kept on flake ice for 6 h until they were brought to our laboratory. For both fish species, individual specimens were divided into two batches: one of them reserved to the flake ice treatment, and the other to the slurry ice process. Fish specimens (not headed, not gutted) under both treatments were placed inside an isothermal room at 2 C. Fish samples from both treatments were taken for analysis on days 2, 5, 8, 12, 15 and 19 for hake, and on days 2, 5, 9, 14, 19, 22, 26, 29, 33, 36 and 40 for turbot. For each species and each chilling system, three different sets (n ¼ 3) were studied separately during the whole experimental period. Refrigeration systems

In this study, a slurry ice prototype (FLO-ICETM; Kinarca S.A.U., Vigo, Spain) was used. The composition of the slurry ice binary mixture, prepared from filtered sea water, was 40% ice/60% water, the


temperature being adjusted to )1.5 C. Flake ice was prepared with an Icematic F100 Compact device (CASTELMAC SPA, Castelfranco, Italy). The temperature of the flake ice was )0.5 C. The fish specimens were surrounded by flake ice or slurry ice at a fish/ice ratio of 1/1. When required, both ice were replenished. Sensory analysis

Sensory analysis was conducted by a panel consisting of five experienced judges, who based appraisals according to guidelines concerning fresh and refrigerated fish (Council Directive 91/493/EEC 1991; Rodrı´ guez et al., 2003). Four categories were ranked: highest quality (E), good quality (A), fair quality (B) and unacceptable quality (C). Sensory assessment of the fish included the examination of the following parameters: skin, external odour, gills, consistency and flesh odour. Once fish specimens had been subjected to sensory analyses, the white muscle was separated and used for biochemical analyses. Nucleotide degradation analysis

Analysis of the autolytic nucleotide degradation in fish muscle was carried out by HPLC according to the method of Ryder (1985). The K-value was calculated according to the following concentration ratio: K-value (%) ¼ 100 · (hypoxanthine + inosine)/(adenosine triphosphate + adenosine diphosphate + adenosine monophosphate + inosine monophosphate + inosine + hypoxanthine). Solubilization of sarcoplasmic protein fraction

Sarcoplasmic protein extracts from fish muscle were prepared in a low-ionic-strength buffer composed of 10 mm Tris–HCl (pH 7.2) + 50 mm pentamethyl sulphonic acid (PMSF) as previously described (Pin˜eiro et al., 1998). All extracts were maintained at )80 C until analysis. Protein concentrations in the extracts were determined by means of the protein microassay method (Bio-Rad Laboratories Inc., Hercules, CA, USA). A standard curve constructed from bovine serum albumin (BSA) was used as the reference. SDS-PAGE

Electrophoretic analyses of the sarcoplasmic protein fraction from fish muscle were carried out in commercial horizontal sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gels (245 · 110 · 1 mm Excel-Gel SDS Homogeneous 15%; Amersham Biosciences, Uppsala, Sweden). Protein bands were visualised by silver staining as previously described (Pin˜eiro et al., 1998).



ELISA

a-Actinin content in the sarcoplasmic protein extracts from fish muscle was determined by an indirect ELISA. Briefly, 96-well high-binding plates (Costar Corning Inc, New York, USA) were coated with aliquots from each sarcoplasmic protein extract (20 lg mL)1/50 lL), incubated at 4 C overnight, blocked by the addition of PBS/1% BSA solution for 2 h at 37 C and washed three times with PBS/0.1% Tween 20. Then, wells were incubated with 50 lL of mouse monoclonal anti-aactinin sarcomeric antibody (MAb) (diluted 1:500 in PBS/0.1% BSA) (Sigma, St Louis, MO, USA) for 1.5 h at 37 C and washed again. Bound antibodies were detected using 50 lL per well of horseradish peroxidase (HRP)-labelled goat anti-mouse Igs (diluted 1:1000 in PBS/0.1% BSA) (Sigma). After washing, the colorimetric reaction was developed with the addition of the substrate o-phenylene-diamine (OPD, Sigma) according to manufacturer’s instruction and stopped by the addition of 3 m sulphuric acid. Finally, optical density (OD) at 492 nm was measured by means of an ELISA Microplate Reader (Labsystems iEMS Reader MF; Molecular Devices Co, Sunnyvale, CA, USA). Three replicate wells per sample and a negative control without primary antibody were considered in all cases. The OD value obtained for each extract was corrected by subtracting the OD value determined in negative samples. The optimised method was validated using calibration standards at 1, 5, 10, 40, 60, 80 and 100 lg mL)1 per well into PBS/0.1% BSA of a commercial and purified a-actinin from chicken gizzard (Sigma). In the absence of a-actinin purified from the fish species, the specificity of sarcomeric anti-a-actinin antibody was tested previously by Western blot assay (Rybicki & von Wechmar, 1982). Sarcoplasmic protein electrophoretic profiles from chilled hake and turbot were studied by SDS-PAGE. For this, homogeneous vertical gels (10%T and 3%C) including 0.1 m Tris– 0.1 m Tricine–0.1% SDS (pH 8.25), and 0.2 m Tris (pH 8.9) as cathode and anode solutions, respectively, were employed. When the electrophoresis development was accomplished, gels were transferred to Hybond-P PVDF membranes (Amerham Biosciences, Uppsala, Sweden) in a Mini-Trans-Blot-Cell (Bio-Rad Laboratories, Inc., Hercules, CA, USA), incubated and stained according to Rybicki & von Wechmar (1982). Results

Fish freshness as determined by sensory analysis

European hake specimens stored in slurry ice maintained good quality up to day 8, while hake specimens stored in flake ice exhibited good quality only up to day

2 (Table 1). In the case of turbot, the specimens chilled in slurry ice showed good quality up to day 22, while their counterparts stored in flake ice did so only until day 14. A higher shelf-life was obtained by employing slurry ice conditions for both hake and turbot when compared with their counterparts stored under flake ice. Comparison between the sensory acceptance scores of both species showed a markedly faster quality loss for hake than for turbot. Evaluation of protein degradation by monitoring the electrophoretic profiles of sarcoplasmic proteins

In the electrophoretic protein profiles obtained for hake (Fig. 1i), a marked increase in the concentration of two protein bands of 23 and 24.5 kDa, respectively, was observed after 5 days of storage for the batch chilled in flake ice; these bands were subsequently degraded by day 15. However, for the hake batch stored in slurry ice, the presence of such protein bands increased only after day 12. In turbot stored in slurry ice (Fig. 1ii), the sarcoplasmic protein profiles did not reveal changes in protein bands corresponding to molecular weights in the range of 87–94 kDa. However, when the batch chilled under flake ice is considered, such a band range was found to be markedly modified as a result of fish damage, being visualised with great difficulty. Previous reports have proposed certain soluble polypeptides as qualitative biomarkers of fish spoilage or freshness (Papa et al., 1996; Verrez-Bagnis et al., 1999; Morzel et al., 2000). These reports agree with the above-mentioned changes in protein bands as shown in Fig. 1. Evaluation of fish freshness by nucleotide analysis

Assessment of nucleotide degradation was carried out by means of the K-value (Fig. 2, bars). In global terms, both fish species provided a different evolution. Thus, hake showed a progressive increase throughout the experiment, in all cases the K-index being under 60% value.

Table 1 Summary of sensory acceptance during chilled storage of both fish species Hake a

Turbot b

Aa

Cb

22 14

29 19

Ice treatment

A

C

Slurry ice Flake ice

8 2

15 8

Freshness categories: A (good) and C (unacceptable). a Last time (days) that ‘A’ quality was observed. b First time (days) that ‘C’ quality was observed.



71

72


Figure 1 Electrophoretic profiles obtained in 15% ExcelGel homogeneous SDS-PAGE from sarcoplasmic proteins of (i) hake and (ii) turbot. The different lines include a lowmolecular-weight standard (st: 14–94 kDa) and fish samples corresponding to the different storage times (days) for both chilling conditions. Arrows indicate the most remarkable changes detected in certain protein bands during storage.

However, turbot stored under flake ice provided a logarithmic pattern with time, so that a sharp increase was observed in the 0- to 14-day period, that was followed by no changes; values were above 60% in the 14- to 40-day period in all cases. Turbot stored under slurry ice showed a sharp increase until the end of the storage, when scores attained the 60% value. As expected from the sensory results, and according to our previous results (Rodrı´ guez et al., 2003; Losada et al., 2004), storage in flake ice implied significantly higher (P < 0.05) K-values than storage in slurry ice. In the case of European hake specimens, day-to-day comparison revealed significantly (P < 0.05) higher K-values on days 12, 15 and 19 for the batch stored in traditional flake ice than for the specimens stored in slurry ice. Moreover, turbot specimens stored in flake ice showed a higher (P < 0.05) K-value development than their counterparts stored under slurry ice. From the actual results it can be concluded that the application of slurry ice slowed down the nucleotide degradation in hake and turbot although the K-value showed a very different evolution in both species, so that


this quality index cannot be defined as a general parameter to evaluate the freshness in fish species. Development of an indirect ELISA assay based on a-actinin assessment to evaluate fish freshness

An indirect ELISA assay employing a commercial a-actinin mouse MAb, was chosen in order to estimate the degree of proteolysis concerning the a-actinin release in both chilled hake and turbot species. The specificity of the commercial a-actinin mouse MAb, was proved by the Western blot procedure described in the Materials and methods section. Figure 3 shows the single stained band obtained from a-actinin control and turbot sarcoplasmic profiles after 36 days of storage under flake and slurry ice; profile comparison showed the presence, in both chilled turbots, of a single fragment with similar molecular weight (100 kDa) than in the a-actinin control. The a-actinin immunoassay was validated by employing calibration standards at 1, 5, 10, 40, 60, 80 and 100 lg mL)1, prepared using a commercial a-actinin



Figure 4 a-Actinin ELISA standard curve. Absorbance values of triplicate determinations (n ¼ 3) are shown, the standard deviations being expressed by bars.

Figure 2 Comparison between the K-value and the quantitative immunoassay based on a-actinin assessment during chilled storage of: (i) hake and (ii) turbot. Bars indicate the K-values (Y1 axis) obtained for the slurry ice batch (white bars) and flake ice batch (black bars). Lines indicate the OD values obtained with the ELISA method for slurry ice (discontinuous line) or flake ice (solid line) conditions. Determination of a-actinin content in the sarcoplasmic protein extracts was carried out in terms of optical density (OD) at 492 nm (Y2 axis).

standard from the chicken gizzard. Figure 4 shows the standard dose-dependent curve for pure a-actinin at different concentrations and its corresponding regression equation. The mean regression coefficient for analysis was r2 ¼ 0.9659. Besides, the non-specific binding of the peroxidase-labelled second antibody was tested by means of a negative control without primary antibody. Figure 2 (lines) shows the results obtained by means of the referred ELISA aimed at detecting a-actinin

Figure 3 Western blot using commercial specific mouse a-actinin MAb against chicken a-actinin as control (A), and sarcoplasmic protein extracts from turbot chilled to day 36 on slurry and flake ice.

release in the sarcoplasmic protein extracts obtained from the muscle of fish specimens (hake, Fig. 2i; turbot, Fig. 2ii) stored either in flake ice or slurry ice. As can be seen, a tendency for a fair correlation between the proposed ELISA (lines) and the K-value (bars) during storage was observed for both species in the case of flake ice storage, showing a progressive increase for hake and a logarithmic one for turbot, according to evolution of the K-value. However, as slurry ice conditions did not provide (P > 0.05) changes in the OD values throughout the experiment, a good agreement with the K-value was not obtained. The OD values obtained were found to be highly species-dependent, being markedly bigger in the case of hake (Fig. 2i lines). Thus, low OD values (7.5) and temperature (40–50 C) ranges (Toiguchi et al., 1989). The colour of soymilk can also be affected by protein aggregates, droplet size concentration and lipid content

(Chantrapornchai et al., 1998; Chanamai & McClements, 2001). Soymilk samples prepared from S08-80, Vinton 81, FG1 and S20-20 soybean cultivars all showed particle size distribution around 0.8 lm, except for S03W4 with particle size distribution of 0.4 lm. The lower protein and fat contents of soymilk prepared from this cultivar (Table 2) may have resulted in smaller particles and its whiter colour (lowest DE). As colour is an important parameter for consumer acceptability of soymilk product, the soymilk products were compared with commercially available soymilk and the results obtained are shown in Table 4.

Table 4 Hunterlab colour values of soymilk made from five different cultivars

Solid-phase microextraction analysis of volatile compounds in soymilk samples

Colour values Product Standard Cow milk (Parmalat 3.25%) Commercial soymilk Product A Product B Product C Product D Soymilk made from cultivars S08-80 Vinton 81 FG1 S20-20 S03W4

L

a

b

96.06

)2.31

8.07

72.98 78.60 84.30 84.50

1.09 )1.29 )1.20 )0.46

13.45 10.00 15.31 15.45

23.94 17.60 13.85 14.44

± ± ± ±

1.0 0.7 0.4 0.1

83.07 87.12 85.17 83.55 86.38

)0.93 )1.28 )0.44 )0.72 )0.44

13.97 15.86 14.84 16.33 14.44

14.33 11.90 12.96 15.08 11.74

± ± ± ± ±

0.2 0.5 0.1 0.3 0.2

Table 5 shows the relative percentages of the major compounds identified in soymilk samples prepared from the five soybean cultivars. Identities of the soymilk volatile compounds were confirmed on the basis of the retention times of pure standards and MS spectral libraries. The main constituents identified were hexanal, hexanol, pentanol, 1-octen-3-ol, 2-pentyl furan, octanone, octanal and nonanal. Other compounds present, but at lower concentrations, included heptanal, benzaldehyde, heptanol, 2-heptanone, 5-ethyl cyclopentene carboxaldehyde and pentadecanol. Among the five soybean cultivars, only soymilk made from S20-20 variety showed significant (P < 0.05) difference (highest) in its total volatiles compared with other soymilk (Fig. 1). Soymilk sample with the lowest level of total volatiles was from S03W4 variety, followed by soymilk samples from FG1, S08-80, Vinton 81 and S20-20 cultivars. It is widely accepted that the principal

DE

0

The ‘L’ scale denotes lightness-to-darkness in 100–0 units. The ‘a’ scale represents redness (+a) vs. greenness ()a) and the ‘b’ scale represents yellowness (+b) vs. blueness ()b). The L, a, b values are referred to as tristimulus values for the Hunter L, a, b solid.

Total volatile (%) Volatiles

S08-80

Vinton 81

FG1

S20-20

S03W4

MS identificationa

Pentanol Hexanal Hexanol 2-Heptanone Heptanal Benzaldehyde Heptanol Octen-3-ol Octanone 2-Pentylfuran Octanal Nonanal 5-Ethylcyclopent-1-ene carboxaldehyde Pentadecanol

5.2 49.7 22.6 0.6 1.2 0.9 1.0 7.5 2.8 4.5 0.7 1.4 1.2

6.5 61.2 13.6 0.5 1.0 1.4 1.0 4.2 2.0 4.6 1.1 1.8 1.2

8.1 66.0 14.6 0.6 0.9 1.3 0.7 3.1 2.1 5.4 3.2 1.9 1.2

7.3 58.1 17.8 0.5 0.9 1.1 0.4 2.6 1.8 5.7 0.8 1.8 1.1

7.2 58.5 10.5 1.4 2.4 2.2 0 2.2 1.8 8.5 1.5 2.1 1.7

Standard Standard Standard Standard Standard Standard Standard Standard Standard Standard Standard Standard Library

0.6

0

0

0

0

Table 5 Identification of volatile compounds in soymilk prepared from different cultivars using SPME–GC–MS

Library

a

The volatiles were either positively identified using MS of pure standards or identified using the GC–MS spectra library.



Total volatiles

S 08-80

Hexanal

Vinton 81 FG 1 S 20-20 Soybeans varieties

S 03W4

Figure 1 (h) Total volatiles recovery (MS response-peak areas) and ( ) percent ratio of hexanal to total volatiles recovery of soymilk prepared from different soybean cultivars.

contributors to off-flavours in soymilk are the volatile carbonyl compounds, particularly hexanal (MacLeod & Ames, 1988). Figure 1 also shows the percentage of hexanal relative to total volatiles in the soymilk prepared from different cultivars. Significant (P < 0.05) differences in hexanal contents were observed among the different soymilk samples. Soymilk made from S08-80 variety had the lowest hexanal followed by soymilk made from S03W4, FG1, Vinton 81 and S20-20 cultivars. Determination of lipoxygenase activity

Lipoxygenase isozyme activities were assayed using linoleic acid as substrate at pH 6.3 for LOX-II and LOX-III and at pH 9.0 for LOX-I. The results as presented in Table 6 showed that the total LOX activity was significantly (P < 0.001) different among the different soybean cultivars, except between Vinton 81 and S20-20 cultivars. The LOX-I activities for all cultivars were higher than the activities of LOX-II and LOX-III. Similar results were reported by Marczy et al. (1995) for other soybean varieties. Other authors have reported that LOX isozyme activities in soybean seeds are influenced significantly by the growing locations

Table 6 Lipoxygenase (LOX) isozymes activitiesa (units per mg protein) in various soybean cultivars

Total LOX

1600

Total volatiles

45

1400

40

1200

35 30

1000

25

800

20

600

15

400

10

200

5 0

0 S 08-80 Vinton 81

FG 1

Total volatiles (Area counts x 106)

45 40 35 30 25 20 15 10 5 0

Total LOX (units mg–1 protein)

Peak area count x 106

Storage effects on soymilk flavour and quality A. Achouri et al.

S 20-20 S 03W4

Soybean cultivars Figure 2 LOX activity relationship to total volatiles compounds in different soybean cultivars.

(Kumar et al., 2003) and by both cultivar and climatic conditions (Marczy et al., 1995). It is well established that the beany flavour associated with various soy-based foods results from the formation of volatile carbonyl compounds during catalysis of polyunsaturated fatty acids by LOX (Rackis et al., 1979). Moreover, the isozymes (LOX-II and LOX-III) catalyse the formation of volatile compounds much more intensively than does LOX-I (Marczy et al., 1995). The data reported in Fig. 2 showed that soybean cultivars (Vinton 81 and S20-20) had the highest LOX activity that corresponded to a greater content of total volatiles, whereas with S03W4, the low enzymatic activity resulted in lower total volatiles. It appears, therefore, that there is a positive correlation (correlation coefficient ¼ 0.82) between the total volatile compounds observed and the LOX activity, which confirms the involvement of this enzyme in the biogenesis of the volatile compounds (Ridolfi et al., 2002). Effect of soybean storage

Because of its higher protein, lower fat contents and better colour characteristics, which are highly recom-

Cultivars

Protein (mg)

LOX-I

S08-80 Vinton 81 FG1 S20-20 S03W4 Stored Vinton 81

2.28 2.36 2.32 2.26 2.25 1.97

1012.8 1203.4 996.7 1232.2 732.1 939.2

± ± ± ± ± ±

12 8 8.5 4 4 6

LOX-II + LOX-III

Total LOX activity

66.3 242.7 127.8 186.3 90.2 211.9

1079.1 1446.1 1124.5 1418.5 822.3 1161.1

± ± ± ± ± ±

0.5 2.5 4.6 1.0 0.7 1.0

± ± ± ± ± ±

12.4a 11.3b 1.9c 5.2b 5.2d 7.9

Mean values within the same column followed by different superscript letters are significantly different (P < 0.05). a Values are average of duplicates ± SD.



87


7

Delta E (color difference)

16

6.8 Soymilk pH

6.6 6.4 6.2 6 1

2

3

4 5 6 7 8 Storage time (month)

9

10

14 13 12 11 0

Figure 3 Soybean storage effect on the pH of soymilk over a period of 10 months.

mended for soymilk production, the Vinton 81 variety was selected for follow-up studies on the effect of soybean storage (10 months) and on physical properties and flavour characteristics of soymilk. In an earlier study, which compared the quality and yield of soymilk and tofu made from thirteen soybean varieties, Cai et al. (1997) also reported that Vinton was one of the most popular commercial varieties used in the soymilk and tofu industry. No differences were observed in WUF of soybeans during the 10 months of storage under the applied conditions. The amount of water imbibed by soybeans during soaking is important in the processing of soybeans into soymilk and tofu, as it reduces the amount of energy required to grind them and increase the subsequent rate of nutrient extraction (Shurtleff & Aoyagi, 1990). Other authors have reported that water uptake decreased when seeds were stored at higher temperatures and higher relative humidities (Ohta et al., 1979; Lambrecht et al., 1995). Storage of the soybeans did not also have a significant (P < 0.05) impact on the pH of the prepared soymilk (Fig. 3). The pH values of the soymilk remained between 6.4 and 6.6. However, the colour of the soymilk prepared from the stored soybeans became increasingly darker over the 10-month storage period. This can be seen by the increase in DE values from 11.9 to 14.9 over the period of storage (Fig. 4). In their study, Narayan et al. (1988a,b) reported earlier that the colour of stored soybeans changed from creamy yellow to brown with the increase in storage period and the intensity of the colour increased at ambient temperature during storage. The effect of soybean storage on the flavour profile of soymilk is represented in Fig. 5. Storage resulted in a significant (P < 0.05) decrease in the total volatiles recovery over time. The sharpest decrease in total volatiles was observed during the initial months of storage (88% loss after 3 months). Beyond the third month, total volatiles recovery remained fairly constant.


15

10

11

1

2

3

4 5 6 7 8 Storage time (month)

9

10

11

Figure 4 Soybean storage effect on soymilk colour difference (DE) over a period of 10 months.

Peak area count x 106

88

40 35 30 25 20 15 10 5 0

Total volatiles

0

1

2

3

Hexanal

4 5 6 7 Storage time (month)

8

9

10

Figure 5 Soybean storage effect on the total volatiles (h) recovery (MS response-peak areas) and per cent ratio of hexanal ( ) to total volatiles recovery of soymilk over a period of 10 months.

This decrease in total volatile corresponds with the 20% decrease in the total LOX activity observed during storage. Similar findings were also reported by several authors (Saio et al., 1980; Narayan et al., 1988a; Thomas et al., 1989; Lambrecht et al., 1995). This finding suggests that soymilk prepared from stored soybeans may result in better soymilk flavour and quality. The higher level of total volatiles in the soymilk samples for the control and during the first month of storage could have been influenced by the intrinsic properties of soybeans at the time of their harvesting (i.e. moisture content, chlorophyll (pigment) concentration, LOX activity and chemical reactions between soybean constituents). Conclusions

The work reported in this article has shown that in addition to variety, storage affects the quality of the soybeans used for making soymilk, particularly its flavour. Some researchers have also reported a decrease



in protein extractability as a result of storage (Saio et al., 1980; Thomas et al., 1989). The LOX activity of the different soybean cultivars was well correlated with the formation of volatile compounds. Further studies on flavour, quality and yield using a wider variety of soybean stored under different conditions (temperature, relative humidity, light/dark) for varying periods will be useful to determine the optimal conditions of storage of soybeans for soymilk production. References Al-Kishtaini, S.F. (1971). Methods of preparation and properties of water extracts of soybeans. Ph.D. Thesis. Urbana, IL: University of Illinois. AACC (1983). Approved Methods of the American Association of Cereal Chemistry, Method 44-15A, 8th edn. St. Paul: AACC. AACC (2003). Approved Methods of the American Association of Cereal Chemistry, Methods 08-03 and 30-25, 8th edn. St. Paul, MN: AACC. Arthur, C.L. & Pawliszyn, J. (1990). Solid-phase microextraction with thermal desorption using fused silica optical fibers. Analytical Chemistry, 62, 2145–2148. Association of Official Analytical Chemists (1995). Official Methods of Analysis, Method 992, 15th edn. Washington, DC: AOAC. Badenhop, A.F. & Hacker, L.R. (1970). Effects of soaking soybeans in sodium hydroxide solution as pretreatment for soymilk production. Cereal Science Today, 15, 84–88. Bradford, M. (1976). A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry, 72, 248–254. Cai, T.D., Chang, K.C., Shih, M.C., Hou, H.J. & Ji, M. (1997). Comparison of bench and production scale methods for making soymilk and tofu from 13 soybean varieties. Food Research International, 30, 659–668. Carrao-Panizzi, M.C., Beleia, A.D., Prudencio-Ferreira, S.H., Oliveira, M.C.N. & Kitamura, K. (1999). Effect of isoflavones on beany flavor and astringency of soymilk and cooked whole soybean grains. Pesquisas Agropecua´ria Brasileira, Brasilia, 34, 1045–1052. Chanamai, R. & McClements, D.J. (2001). Prediction of emulsion color from droplet characteristics: dilute monodisperse oil-in-water emulsions. Food Hydrocolloids, 15, 83–91. Chantrapornchai, W., Clydesdale, F.M. & McClement, D.J. (1998). Influence of droplet size and concentration on the color of oil-inwater emulsions. Journal of Agricultural and Food Chemistry, 46, 2914–2920. Coulibaly, K. & Jeon, I.J. (1992). Solid-phase extraction of less volatile flavor compounds from ultrahigh-temperature processed milk. Journal of Agricultural and Food Chemistry, 40, 612–616. Flurkey, W.H., Young, L.W. & Jen, J.J. (1978). Separation of soybean lipoxygenase and peroxidase by hydrophobic chromatography. Journal of Agricultural and Food Chemistry, 26, 1474–1476. Garrote, R.L., Bertone, R.A., Silva, E.R. & Avalle, A. (2001). Comparison of two rapid methods of lipoxygenase assay in blanched green peas, green beans and potatoes. Food Science and Technology International, 7, 171–175. Katayama, O. & Tajima, M. (2003). Shokuhin to iro [Color of Food], Change in Color of Food. Pp. 71–117. Tokyo: Korin (in Japanese). Khatib, K.A., Aramouni, F.M., Herald, T.J. & Boyer, J.E. (2002). Physicochemical characteristics of soft tofu formulated from selected soybean varieties. Journal of Food Quality, 25, 289–303. Kon, S., Wagner, J.R., Guadagni, D.G. & Horvat, R.J. (1970). pH adjustment control of oxidative off-flavors during grinding of raw legume seeds. Journal of Food Science, 35, 343–345.

Kumar, V., Rani, A., Tindwani, C. & Jain, M. (2003). Lipoxygenase isozymes and trypsine inhibitor activities in soybean as influenced by growing location. Food Chemistry, 83, 79–83. Lambrecht, H.S. (1991). Soybean storage and its effect on tofu and soymilk production. M.S. Thesis. pp. 35. West Lafayette, IN: Purdue University. Lambrecht, H.S., Neilsen, S.S., Liska, B.J. & Neilsen, N.C. (1995). Effect of soybean storage on tofu and soymilk production. Journal of Food Quality, 19, 189–202. MacLeod, G. & Ames, J. (1988). Soy flavor and its improvement. CRC Critical Reviews in Food Science and Nutrition, 27, 219–401. Marczy, J.Sz., Simon, M.L., Mozsik, L. & Szajani, B. (1995). Comparative study on the lipoxygenase activities of some soybean cultivars. Journal of Agricultural and Food Chemistry, 43, 313–315. Marsili, R.T. (1999a). Comparison of solid-phase microextraction and dynamic headspace methods for the GC-MS analysis of lightinduced lipid oxidation products in milk. Journal of Chromatographic Science, 17, 17–23. Marsili, R.T. (1999b). SPME-MS-MVA as an electronic nose for the study of off-flavors in milk. Journal of Agricultural and Food Chemistry, 47, 648–654. Monma, M., Terao, J., Ito, M., Saito, M. & Chikuni, K. (1994). Carotenoid components in soybean seeds varying with seed color and maturation stage. Bioscience, Biotechnology and Biochemistry, 58, 926–930. Nakabayashi, T. (1968). Tannin of fruits and vegetables. III. Polyphenolic compounds and phenol-oxidising enzymes of edible burdock. Nippon Shokuhin Kogyo Gakkaishi, 15, 199– 206. Narayan, R., Chauhan, G.S. & Verma, N.S. (1988a). Changes in the quality of soybean during storage. Part 1-effect of storage on some physico-chemical properties of soybean. Food Chemistry, 27, 13–23. Narayan, R., Chauhan, G.S. & Verma, N.S. (1988b). Changes in the quality of soybean during storage. Part 2-effect of soybean storage on the sensory qualities of the products made therefrom. Food Chemistry, 30, 181–190. Ohta, T., Takano, K., Nikkuni, I., Hashizume, K. & Saio, K. (1979). Soybean quality change and its influence on food processing. Shokuhin Sogo Kenkyujo Kenkyu Hkoku Report of National Food Research Institute, 35, 56–70. Okubo, K., Iijima, M., Kobayashi, Y., Yoshikoshi, M., Uchida, T. & Kudou, S. (1992). Components responsible for the undesirable taste of soybean seeds. Bioscience, Biotechnology and Biochemistry, 56, 99–103. Pereira, S., Chaves, L. & Hartman, L. (2002). Economic evaluation of soymilk production: comparative study of two processes. Alimentaria, 97, 107–113. Rackis, J.J., Sessa, D.J. & Honig, D.H. (1979). Flavor problems of vegetable food proteins. Journal of the American Oil Chemists’ Society, 56, 262–271. Reddy, P.V. & Mital, B.K. (1992). Physical and chemical characteristics of soymilk. Journal of Food Science and Technology, 29, 193– 194. Ridolfi, M., Terenziani, S., Patumi, M. & Fontanazza, G. (2002). Characterization of the lipoxygenases in some olive cultivars and determination of their role in volatile compounds formation. Journal of Agricultural and Food Chemistry, 50, 835– 839. Saio, K.I., Nikkuni, Y., Ando, M., Otsuru, M., Terauchi, Y. & Kito, M. (1980). Soybean quality changes during model storage studies. Cereal Chemistry, 57, 77–82. Saito, M., Kudo, H., Mandarino, J.M.G. & Benassi, V.T. (2004). Effects of variety and cultivating region on the color of soymilk and other soybean processing foods in Brazil. Japan Agricultural Research Quarterly, 38, 179–183.



89

90


Saxena, S. & Singh, G. (1997). Suitability of new soybean cultivars in the production of soymilk. Journal of Food Science and Technology, 34, 150–152. Shurtleff, W. & Aoyagi, A. (1990). Tofu and Soymilk Production: The Book of Tofu, Vol. 2, p. 448. Lafayette, CA: Soyfoods Center. Taira, H., Toriu, H. & Saito, M. (1985). Quality of soybean seeds grown in Japan: varietal differences in total carotenoids content and color of soybean seeds grown in Nagano Chushin agricultural experiment station. Report of National Food Research Institute (Japan), 47, 92–104. Thomas, R., DeMan, J.M. & DeMan, L. (1989). Soymilk and tofu properties as influenced by soybean storage conditions. Journal of the American Oil Chemists’ Society, 66, 777–782.


Toiguchi, S., Hayashi, K., Adachi, Y., Motoki, M. & Haraguchi, K. (1989). Purification and characterization of soybean oxidase. Nippon Shokuhin Kogyo Gakkaishi, 36, 597–602. Wang, Z.H., Dou, J., Macura, D., Durance, T.D. & Nakai, S. (1998). Solid-phase extraction for GC analysis of beany flavours in soymilk. Food Research International, 30, 503–511. Wilkens, W.F., Mattick, L.R. & Hand, D.B. (1967). Effect of processing method on oxidative off-flavors of soybean milk. Food Technology, 21, 1630–1633.



91

Original article Physiological and biochemical changes of different fresh-cut mango cultivars stored at 5 C Gustavo A. Gonzalez-Aguilar,1* Jorge Celis,1 Rogelio R. Sotelo-Mundo,1 Laura A. de la Rosa,2 Joaquin Rodrigo-Garcia2 & Emilio Alvarez-Parrilla2 1 Centro de Investigacioń en Alimentacioń y Desarrollo, A.C. (CIAD, AC), Direccioń de Tecnologı´ a de Alimentos de Origen Vegetal, Carretera a la Victoria Km 0.6 La Victoria, Hermosillo, Sonora CP83000, Mexico 2 Departamento de Ciencias Ba´sicas, Instituto de Ciencias Biome´dicas, Universidad Autońoma de Ciudad Jua´rez, Anillo Envolvente del PRONAF y Estocolmo s/n, Ciudad Jua´rez, Chihuahua CP32310, Mexico (Received 21 September 2005; Accepted in revised form 15 June 2006)

Summary

Treatments to inhibit browning, decay and to extend shelf life of ‘Keitt’, ‘Kent’ and ‘Ataulfo’ mango cultivars as a fresh-cut produce were investigated. Combinations of calcium chloride (CaCl2), antioxidants [ascorbic acid (AA), citric acid (CA)] and two commercial film coatings resulted in a reduction of browning and deterioration of fresh-cut mangoes stored at 5 C, especially for the Ataulfo cultivar. The use of CaCl2 + AA + CA significantly reduced colour deterioration, loss of firmness and did not affect sensory characteristics of fresh-cut mango, with a larger effect in the Ataulfo cultivar. In general, these treatments prevented loss of sugar and vitamin C of cubes during storage at 5 C. Shelf life of this cultivar was 21 days, while that of Keitt and Kent was only 9 and 12 days, respectively. There is a correlation between carotene and vitamin C content of Ataulfo mango and its longer shelf life compared with the other cultivars.

Keywords

Calcium chloride, b-carotene, edible coating, Mangifera indica, shelf life, sugars, vitamin C.

Introduction

Postharvest losses of tropical fruits are a serious problem because of rapid deterioration during handling, transport and storage (Yahia, 1998). Although fresh-cut produces are more accepted and demanded by the market, the process itself may decrease the shelf life because of wounding, increased metabolic activities and loss of enzymes and substrates. Other disadvantages of fresh-cut include browning, softening, decay and offflavour development (Watada et al., 1990; Varoquax & Wiley, 1994). Fresh-cut vegetables deteriorate faster than intact produce (Cantwell, 1995). This is a direct result of the wounding associated with processing, which leads to a number of physical and physiological changes affecting the viability and quality of the produce (Saltveit, 1997). Moreover, tropical fresh-cut fruit increases the rates of respiration and ethylene production within minutes of cutting (Abe & Watada, 1991) and has a reduced shelf life of 1–3 days at optimal temperatures compared with 1–2 weeks for the whole fruit (Ahvenainen, 1996). The *Correspondent: Fax: +52-662-2800055; e-mail: [email protected]


visual symptoms of fresh-cut produce deterioration include loss of firmness, changes in colour (especially increased oxidative browning at the cut surface) and microbial contamination (Varoquax & Wiley, 1994; Brecht, 1995). Fresh-cut tropical fruits are more perishable than those from temperate climate, such as apples, nectarines, peaches and pears. For this reason, it is important to control the variables related to the deteriorative processes of these fresh-cut produces, in order to maintain quality attributes for a time period long enough to allow marketing. Fresh-cut fruits and vegetables are generally packaged in film bags or containers over-wrapped with film, which create a modified atmosphere within the package (MAP). Low storage temperature and MAP are commonly used to extend the shelf life of many whole and fresh-cut fruit and vegetable products, as they reduce the respiration rate, surface damage and browning (Gorny, 1997; Thompson, 1998). More recently, it has reported that film coating can extend shelf life in fresh-cut produces. Edible coatings are thin films that improve produce quality and can be safely eaten as part of the product and do not add unfavourable properties to the foodstuff (Baldwin, 1994; Ahvenainen, 1996). Edible coatings

92

Physiological changes of fresh-cut mango G. A. Gonzalez-Aguilar et al.

provide a barrier against external elements and therefore increase shelf life (Guilbert et al., 1996) by reducing gas exchange, loss of water, flavours and aroma and solute migration towards the cuticle (Saltveit, 2001). The first kind of edible coatings were water–wax microemulsions, used since the 1930s to increase brightness and colour in fruits, as well as fungicide carriers. Water loss is another problem that can be controlled with edible wax coatings (Debeaufort et al., 1998). Edible waxes can also offer protection against cold damage under storage (Nussinovitch & Lurie, 1995). Nowadays, an edible coating is made of polysaccharides, proteins and lipids (Guilbert et al., 1996) and resins as well (Baldwin et al., 1995). Lipid-based coatings are excellent for preventing dehydratation, and add brightness to the epidermis. Different kind of components and concentrations lead to different levels of gas exchange (O2 and CO2). One commercially available edible coating is SemperFreshTM (AgriCoat Industries Ltd, Berkshire, UK) which is made of saccharose fatty acid ester on a carboxymethyl cellulose base, containing fatty acid mono- and diglycerides (Bayindirli et al., 1995). The saccharose polyester component of edible films is known as a major barrier for moisture loss (Park et al., 1994). These components slowdown ripening and increase shelf life of produces (James & McGregor, 2000). A factor that must be considered when selecting an edible coat is that products which provide adequate gas exchange are not good water barriers. Furthermore, those that prevent water losses led to anaerobic conditions in the MAP (Baldwin, 1994). The impact of this anaerobic environment on the biochemistry of fresh-cut fruit has not been addressed and should not be underestimated in the formation of unpleasant flavours and aromas. However, information about the effects of film coating on fresh-cut mangoes is very scarce and the effect of minimal processing on different mango cultivars has not been investigated yet. Losses in vitamin C, sugars, b-carotene during storage of fresh-cut fruits and vegetables are very important and used as primary quantitative parameters of quality (Gorny, 2001). The content of ascorbic acid (AA) is an important parameter, as it is an antioxidant itself and increases the nutritional composition of the fresh-cut fruit. For example, its presence in red fruits is known to protect anthocyanins from oxidation, leading to less colour losses (Va´mos-Vigya´zo´, 1981). A previous report in fresh-cut mango indicates that AA reduces browning and deterioration (Gonza´lez-Aguilar et al., 2000). One main effect of edible coatings and antioxidants is prevention of nutrient loss during cold storage of fresh-cut mangoes. Wounding and cutting increases the rate of vitamin loss (Klein, 1987), although little is known about the effect of treatments in fresh-cut produce during cold storage. Therefore, the objective of this article was to evaluate the effect of AA, citric acid


(CA), CaCl2 and film coating on the quality of three cultivars of fresh-cut mango stored at 5 C. Materials and methods

Plant materials

Mango fruits (Mangifera indica L.) cultivars Ataulfo, Kent and Keitt were obtained from Rosario, Sinaloa, Mexico during June and July 2002. The samples were randomly selected and subjected to hydrothermal treatment because of fruit fly quarantine requirements for export. Mangoes were stored at 15 C until use. Transportation conditions were identical to those used for commercial handling. Selected mature fruit had an initial firmness of 18–27 N. Fruits were sorted to eliminate damaged or defective material, washed for 3 min with chlorinated water (250 ppm), dried and randomly divided into three to four lots for each variety. Mangoes were manually cut and pealed into 8 cm3 cubes, washed with chlorinated water (150 ppm) and drained before treatment. Experimental treatments were (1) control (nontreated); (2) antioxidant treatment (1% CaCl2 + 2% AA + 2% CA; pH 2.3); (3) Gustec film (1.5% v/v) and (4) SemperFresh film (1.5% v/v). Mango cubes (14–16 cubes per tray) were dipped for 3 min in test solutions, drained and dried using paper towels and placed in a 250 mL polystyrene plastic tray covered with lid. Thirty trays per treatment for each variety were stored at 5 C for up to 25 days. Respiration rate was measured by determination of oxygen, carbon dioxide and ethylene after storage in a gastight chamber, using gas chromatography (Varian Star 3400CX, Ottawa, ON, Canada) with ionisation flame detector and thermal conductivity detection. A 2 m length HAYESEP N (Valco Instruments Co. Inc., Houston, TX, USA) column (1/8¢¢, 80/100) was used with a filament temperature of 205 C, injector temperature of 70 C and detector temperature of 170 C. Gas standards were 25% O2, 5% CO2 and 1 ppm C2H4. At 3-day intervals, samples were taken, in order to determine colour (three trays per treatment), total solids, firmness, ethanol, acetaldehyde, AA, sugars and b-carotene (three trays per treatment). Physical measurements

Colour was measured with a Minolta colorimeter CR300 (Minolta Corp., Ramsey, NJ, USA) using the Hunter scale L*, a*, b*, where L* was a darkening index and b* the yellow. Two colour measurements per mango cube were taken. Total solids were determined using a digital refractometer (Atago PR-101; Atago Co. Ltd., Tokyo, Japan). Firmness was measured using a firmness tester (Texture Analyser TA-XT2; Arrow Scientific,



Lane Cove, NSW, Australia) with a stainless steel spherical probe of 1/4 ¢¢ diameter and a speed of 10 mm s)1. Texture was reported as force in Newton (N) to penetrate 10 mm. Decay and browning index

Acceptability, measured as the extent of decay and darkening, was determined in ten trays per treatment at the end of the shelf life (19 days for Ataulfo and 12 days for Kent and Keitt), through a subjective scale. Subjective evaluation was done by an untrained panel of ten people (six men and four women) using a hedonic scale (1 ¼ no darkening or damage, 3 ¼ minimal, 5 ¼ moderated, 7 ¼ severe and 9 ¼ extreme). Data were analysed as percentage of distribution of scores using statistical analysis with chi-square statistics. Biochemical measurements

and a mobile phase of acetonitrile/1 m KH2PO4 (75:25 v/v) with a flow rate of 1.5 mL min)1. AA was detected by ultraviolet absorption at 268 nm and concentrations were calculated using a standard curve, and expressed in terms of fresh weight. Sugars

Fructose, glucose and sucrose were determined as described by Smith et al. (1986). Fruit tissue (10 g) was homogenised with 50 mL of water, filtered and placed in a 100 C water bath (Series 180; Precision Scientific, Chicago, IL, USA) for 15 min. The homogenate was centrifuged for 15 min at 10 000 rpm, filtered through a 0.22-lm filter and analysed by HPLC using a lBondapak/carbohydrate analytical column (3.9 · 300 mm, 10 lm) and a mobile phase of acetonitrile/water (80:20 v/v) with a flow rate of 1.5 mL min)1. Sugars were detected and quantified at 192 nm using standard curves, and expressed in terms of fresh weight.

b-Carotene

Ethanol and acetaldehyde content

Carotene was measured as previously described by Mejia et al. (1988). Fresh tissue (1 g) was homogenised (Ultra Turrax T-25 Basic S1; IKA Werke, USA [Staufen, Germany]) at 13 500 rpm for 2.5–3 min with 15 mL of tetrahydrofuran, containing 0.4% Butylated Hydroxytoluene (BHT). The mixture was centrifuged for 15 min at 10 000 g (Allegra 64R Centrifuge; Beckman Coulter, Fullerton, CA, USA), filtered through a 0.22-lm filter and analysed by HPLC using a Microsorb RP-C18, 3 lm (4.6 mm · 10 cm) column with a 3 cm guard column (Supelco, Sigma Aldrich Co., Bellefonte State, PA, USA) with acetonitrile/methanol/tetrahydrofuran (58:35:7) as the mobile phase at a flow rate of 1.0 mL min)1. b-Carotene was detected by ultraviolet absorption at 460 nm and identified by retention time comparison with standards and by standard addition or ‘spiking’ (Johnson & Stevenson, 1978). Quantification was done by use of external standards and expressed in terms of fresh weight.

The determinations were based on the protocol by Davis & Chace (1969). Briefly, 5 g of tissue was placed in 20 mL capacity amber-coloured tubes and incubated at 65 C in a water bath for 15 min. Headspace samples of 1 mL were injected into a gas chromatograph equipped with 2 m · 1/8 in. chromosorb stainless steel column packed with 80/100 lm mesh Porapak 101. Retention times and standard curves of ethanol and acetaldehyde in water solutions were used for peak identification and quantification. Oven temperature was 100 C, injector temperature 110 C and detector temperature 180 C (flame ionization detector) and N2 was used as carrier gas. Results and discussion

Colour changes

Ascorbic acid, sugars and b-carotene were analysed using a Varian Solvent Delivery System pump Model 9012 and a Rheodyne Model 7125 injector (Rheodyne Inc., Cotati, CA, USA) fitted with a 10 lL loop and a Varian Model 9020 UV-VIS absorbance detector. AA was determined according to Doner & Hickts (1981). Fruit tissue (10 g) was homogenised for 2 min with 50 mL of an aqueous solution containing 30 g L)1 metaphosphoric acid and 80 mL L)1 acetic acid. The homogenate was filtered and centrifuged for 15 min at 10 000 rpm. The supernatant was filtered through 0.22-lm filter, and analysed by HPLC with a waters– NH2 type lBondapak (Waters Corporation, Milford, MA, USA) analytical column (3.9 · 300 mm, 10 lm),

All three varieties of fresh-cut mangoes presented changes in colour (decreases in L* and b* values) during storage at 5 C (Fig. 1), in agreement with Rattanapanone et al. (2001), who observed a reduction in the L* value in ‘Tommy Atkins’ and Kent mango cubes stored in air during 8 days. The initial L* values for the three cultivars (Ataulfo, Kent and Keitt) were very similar between them (64–66) and decreased during storage, evident by the loss of brightness, a parameter used as an indicator of browning in fresh-cut fruits (Gonza´lezAguilar et al., 2004). Figure 1 shows the decrease of the L* value, during different periods of time, depending on the shelf life of each mango variety. In all cases, lightness of control samples showed a decrease of approximately 12–16 units. When the different treatments are compared, it is clear that the mixture of antioxidants (AA + CA) have the best protection



Ascorbic acid

93


Control Antioxidants Gustec SemperFresh

Ataulfo

Kent

Lightness (L*)

70

Keitt

a

b

c

70

65

65

60

60

55

55

50

50 LSD0.05 = 3.40

LSD0.05 = 3.48

LSD0.05 = 3.50

45

45

0 40

b*

94

d

e

f

0 40

38

38

36

36

34

34

32

32

30

30

28

28

26 24

26 LSD0.05 = 2.14

LSD0.05 = 2.10

LSD0.05 = 2.27

24

0

0 0

3

6

9

12

15

18

21

0

3

6

9

12

0

3

6

9

12

Days at 5 °C Figure 1 Colour variation (decrease in L* and b* values) of fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C. For all figures, data points are mean of fifteen replicates and LSD at 5% level for time are shown (antioxidants: 2% AA + 2% CA, edible films: 2% GustecTM or SemperFreshTM). Data from SemperFresh are not shown in panels b, c, and f, as those were not significantly different from control.

against browning. This was more evident for Ataulfo and Kent cultivars, where a reduction of only four units was observed at the end of the experimental period (Fig. 1a and b), compared with Keitt cultivar, where a reduction of more than 10 units was observed. In the case of the commercial film coating treatments, Fig. 1a shows that for Ataulfo cubes, there was a slight protection against colour change. However, for Kent and Keitt cubes, no protection against colour change is evident (Fig. 1b and c). SemperFresh coating had no significative effect on colour in Kent and Keitt cubes. Decreases of b* values during storage, which were reduced by treatment by antioxidants in the three cultivars, are shown in Fig. 1d–f. Both protective edible films had a minimal effect on Ataulfo mangoes with respect to antioxidant treatment, but better than con-

trols. Having in mind that a decrease in b* indicates a loss of yellowness, which is an important parameter in mangoes, these results suggest that antioxidants preserved the original colour of the cubes, reducing its browning. Antioxidant applications significantly maintained constant b* values of fresh-cut mangoes, Ataulfo and Kent, after 3 days. However, b* values in Keitt fresh-cut cubes decreased in higher extent with similar storage times (Fig. 1f). The different protective effect observed on fresh-cut mangoes treated with antioxidants, could be attributed to the individual characteristics of the evaluated cultivars. It has been observed in different studies that edible films prevented deteriorative processes in whole fruit, but in this particular case did not prevent significantly deterioration and browning of fresh-cut mangoes.




Control Antioxidants Gustec SemperFresh Ataulfo

Kent

Firmness (N)

8

Keitt

a

b

c

8

7

7

6

6

5

5

4

4

3

3

2

2

1

1

0

LSD0.05 = 1.05 0

3

LSD0.05 = 0.36 6

9

12

15

18

21

0

0

LSD0.05 = 0.33

3

6

9

12

0

3

6

9

12

Days at 5 °C Figure 2 Firmness of fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C. For all figures, data points are mean of six replicates and LSD at 5% level for time are shown (antioxidants: 2% AA + 2% CA, edible films: 2% Gustec or SemperFresh). Data from SemperFresh are not shown in panels b and c, as those were not significantly different from control.

Decay and acceptability of the fresh-cut mangoes were measured subjectively by ten untrained panellists. As one of the leading problems with fresh-cut products is their enzymatic browning, acceptability was measured

considering the browning index. The decay and browning indexes for fresh-cut Ataulfo mango are shown in Fig. 3a and d, where the scores given by all ten panellists were summarised as percentage. Considering that browning values above seven represent severe or extreme damages, a score of 7 was considered as the maximum acceptability level. According to these criteria, after 21 days, 90% of the control cubes were not accepted by the panellists. When the treatments were analysed, it was evident that the use of antioxidants increased the shelf life of the product, as only 33% of the cubes were rejected after 21 days. None of the edible films presented good protection against decay (77% and 80% rejection for Gustec and SemperFresh, respectively). Similar results were obtained for the decay analysis. These results are in agreement with those obtained by Gonza´lez-Aguilar et al. (2000), which observed that the use of antioxidants reduced browning of Kent cubes stored at 10 C. In contrast, in the present study, decay and browning indexes of Kent and Keitt cultivars stored at 5 C were not affected by any of the treatments at 9 and 12 days of storage, respectively. For Kent cubes, the rejection scores were 65%, 55% and 60% for the control, antioxidant and Gustec film, respectively. Keitt cubes rejected values were 60%, 65% and 75% for the control, antioxidant and Gustec film, respectively. These results suggest that antioxidant treatment increases shelf life of Ataulfo fresh-cut mango, while no significant



Firmness loss

It has been reported (Chantanawarangoon, 2000; Gonza´lez-Aguilar et al. 2000) that there is a decrease in the firmness of fresh-cut mangoes during storage at 5 C, because of the release of water and other compounds as a consequence of the cutting process. The texture variation, expressed as firmness loss, of the fresh-cut mangoes studied is shown in Fig. 2. The initial firmness for Ataulfo (7–8 N) was higher than that for Kent and Keitt (5–6 N). In agreement with previous studies, during the storage period, there was a reduction in the firmness of the three mango varieties. Kent and Keitt cubes also showed a sharp reduction in their firmness during the first 3 days of storage at 5 C, while Ataulfo cubes had a less dramatic firmness decrease. All three cultivars maintained a significantly higher degree of firmness throughout the storage period; when compared with untreated samples (control). No significant protection of firmness loss was observed in cubes treated with edible films, in particular with SemperFresh coating applied on Kent and Keitt cubes. Decay and browning index

95


Ataulfo 60

40

Keitt

Kent 50

50

1 3 5 7 9

P < 0.05

P > 0.05

P > 0.05

40

40

30

30

20

20

10

10

0

0

Decay

% Distribution

50

30 20 10 0 60 50

50

P < 0.05

50

P > 0.05

P > 0.05

40

40

30

30

20

20

10

10

40 30 20 10

0

0 Con

Gus

SF

Treatments

Ant

Browning

% Distribution

96

0 Con

Gus

Ant

Treatments

Con

Gus

Ant

Treatments

Figure 3 Decay and browning of fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C, at the end of the conservation period (21, 9 and 12 days for Ataulfo, Kent and Keitt, respectively). For all figures, data points are mean of fifteen replicates (antioxidants: 2% AA + 2% CA, edible films: 2% Gustec or SemperFresh). Data from SemperFresh are not shown for Kent and Keitt mangoes, as those were not significantly different from control.

conservation effect was observed for Kent or Keitt varieties stored at 5 C. SemperFresh coating had no significative effect in decay and browning index of Kent and Keitt cubes. According to these results, we concluded that shelf life of fresh-cut mangoes differ with the cultivar. Shelf life of Ataulfo mango was 21 days, while that of Keitt and Kent was only 9 and 12 days, respectively. The lower rejection percentages obtained for the Ataulfo cultivar could be related to the lower levels of colour change and firmness loss described previously, because the rapid loss of firmness is related to less visual appearance, deterioration and browning as observed in fresh-cut pineapples (Gonza´lez-Aguilar et al., 2004). It has been observed that methoxy-pectinbased edible coating maintained freshness of ‘Arumanis’ fresh-cut mango for up to 5 days at 5 C.

b-Carotene occurs naturally in fruits on the lipid fraction or membranes. Initial b-carotene concentration was higher for Ataulfo mangoes, than for Kent and Keitt cultivars (Fig. 4a–c). It was also observed that b-carotene showed a rapid increase during the first 3 days of storage, being greater for antioxidant-treated Ataulfo mangoes (Fig. 4a–c). In the case of Ataulfo, the initial increment of

b-carotene was followed by a sudden fall and a slight recovery for up to 21 days of storage. After the initial increase in b-carotene concentrations, Kent and Keitt cultivars showed a gradual reduction of the vitamin concentration. Mercadante & Rodriguez-Amaya (1998) found similar increases in b-carotene content during ripening in Tommy-Atkins and Keitt mangoes, with final values of approximately 55 mg g)1 fresh tissue. Mango ripening produces an increase in b-carotene content, which is more significant at room temperature, and may be due to an increase in mevalonic acid and geraniol syntheses, which lead to higher levels of total carotenes (Mitra & Baldwin, 1997). It is worthwhile to mention that initial values of b-carotene were higher for the Ataulfo variety. Moreover, after 21 days of storage, b-carotene levels remain almost the same in control samples, and were slightly higher in samples treated with protective films. The final b-carotene concentration in Kent and Keitt cultivars, after 12 days of storage, was higher than the initial levels; however, no effect of treatments was observed. However, no significant effect on b-carotene content was observed in cubes treated with SemperFresh coating applied to Kent and Keitt cubes. According to these results, it is suggested that the fresh-cut processing does not inhibit the increase in b-carotene because of mango ripening and more



b-Carotene



Kent a

350

β−Carotene (mg 100g–1 fw)

Keitt b

c

300

300

250

250

200

200

150

150

100

100

50

50 LSD0.05 = 146.21

0 140

Ascorbic acid (mg 100g–1 fw)

350

LSD0.05 = 81.71

d

LSD0.05 = 31.24

e LSD0.05 = 4.42

f

0 140

LSD0.05 = 1.61

120

120

100

100

80

80

60

60

40

40

20

20 LSD0.05 = 17.58

0

0 0

3

6

9

12

15

18

21

0

3

6

9

12

0

3

6

9

12

Days at 5 °C Figure 4 Evolution of b-carotene and vitamin C concentration in fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C. For all figures, data points are mean of six replicates and LSD at 5% level for time are shown (antioxidants: 2% AA + 2% CA, edible films: 2% Gustec or SemperFresh). Data from SemperFresh are not shown in panels b, c, e and f, as those were not significantly different from control.

important, even at the end of shelf life, all three cultivars retain their carotene content, providing an important source of this scarce nutrient to the diet.

Fruits are naturally recognised by their vitamin C contribution to diet, and it is known that processing and ripening decreases its levels (Lee & Kader, 2000). As the oxidative processes occur more rapidly in fresh-cut products, they are expected to have more losses compared with the whole fruits (Allong et al., 2000). This behaviour was found in this work (Fig. 4d–f). The antioxidant treatment prevented vitamin C degradation, and furthermore, even enriched the mango cubes of the

three varieties, because of the immersion of the cubes into a vitamin-C-containing solution. Initial values of this vitamin were much higher for Ataulfo mangoes (115 mg g)1) compared with 15 mg g)1 for Kent and Keitt cultivars. These initial differences consequently led to final values of 80, 10 and 12 mg g)1 for Ataulfo, Kent and Keitt varieties, respectively. Oxidative reactions seem to be the main cause of vitamin C deterioration, and therefore the antioxidant treatment may prevent such losses, leading to a fresh-cut product with a final AA content as high as the fresh fruits for Kent and Keitt mangoes (Fig. 4e and f). Figure 4d shows a slight protection against vitamin C degradation after 9 days when compared with control, which is more evident for Gustec compared with SemperFresh film. No significant



Ascorbic acid

97



Kent

Keitt

3.5

Fructose (mg 100g–1 fw)

a

Glucose (mg 100g–1 fw)

b

c

3.0

2.5

2.0

1.5 0.0

LSD0.05 = 0.74

LSD0.05 = 0.38

LSD0.05 = 1.00

d

2.0

e

f

1.5

LSD0.05 = 1.03

1.0

0.5 LSD0.05 = 0.33

LSD0.05 = 0.21

0.0 13

Sucrose (mg 100g–1 fw)

98

g

h

12

i

11 10 9 8 7 6 5 0

LSD0.05 = 1.51

0

3

6

LSD0.05 = 1.33

LSD0.05 = 1.24

9

12

15

18

21

0

3

6

9

12

0

3

6

9

12

Days at 5 °C Figure 5 Concentration of glucose, fructose and sucrose of fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C. For all figures, data points are mean of six replicates and LSD at 5% level for time are shown (antioxidants: 2% AA + 2% CA, edible films: 2% Gustec or SemperFresh). Data from SemperFresh are not shown for Kent and Keitt mangoes, as those were not significantly different from control.




protection against vitamin C degradation was observed when edible films were used with Kent and Keitt cultivars, as observed in Fig. 4e and f. Sugars

Sugars are a key component of the organoleptic properties of the mango fruit. Variations in sugar content during storage and treatment effects are shown in Fig. 5. In general, these values show great variability, and they have a major importance in the flavour and acceptance of the fruit. Each cultivar has its own characteristic variation in sugars upon ripening (Selvaraj et al., 1989). Several authors indicate that glucose and fructose are decreased, while sucrose increases (Vazquez-Salinas & Lakshminarayana, 1985). Others

suggest that starch is hydrolysed and sugars are increased during ripening stages (Mitra & Baldwin, 1997). In general, qualitative and quantitative changes in sugars observed during storage in fresh-cut products do not follow a characteristic pattern, as observed in this study. Previous results demonstrate that changes in individual monosaccharides (fructose and glucose) do not correlate with levels of sucrose observed in the tissue after storage (Soliva-Fortuny et al., 2004). It is of common knowledge that sucrose is the main sugar in a ripe mango (Gonza´lez-Aguilar et al., 2000). In this study, initial sucrose levels found in Ataulfo were sixfold higher (12 mg g)1) compared with those observed in Kent and Keitt cultivars (Fig. 5g–i). Glucose levels of approximately 1 mg g)1 were found in Ataulfo and Kent, while non-treated Keitt cubes showed

Control Antioxidants Gustec SemperFresh Kent

Ataulfo

Keitt

2.0

2.0

Ethanol (µL 1g–1 fw)

a 1.5

1.5

1.0

1.0

0.5

0.5

0.0

0.0

d

Acetaldehyde (µL 1g–1 fw)

c

b

e

f

0.12

0.12

0.09

0.09

0.06

0.06

0.03

0.03

0.00

0.00 0

3

6

9

12

15

18

21

0

3

6

9

12

0

3

6

9

12

Days at 5 °C Figure 6 Changes in ethanol and acetaldehyde concentration of fresh-cut (a) Ataulfo, (b) Kent and (c) Keitt mangoes treated with antioxidants or edible films during storage at 5 C. For all figures, data points are mean of four replicates (antioxidants: 2% AA + 2% CA, edible films: 2% Gustec or SemperFresh). Data from SemperFresh are not shown in panels b, c, e and f, as those were not significantly different from control.



99

100


undetectable glucose levels (Fig. 5d–f). Fructose content was very similar among the three cultivars, ranging from 2 to 3 mg g)1. Keitt mangoes treated with antioxidants showed slightly higher levels of glucose and fructose (Fig. 5a–c). Ethanol and acetaldehyde

unwanted by-products, leading to a higher acceptability of the fresh-cut Ataulfo mangoes. According to the results obtained in this study, we conclude that Ataulfo mango is an excellent variety for use in the fresh-cut industry, with better shelf life and nutritional content when treated with the antioxidant treatment of AA and CA. In addition, the long shelf life of Ataulfo fresh-cut mango compared with the other cultivars offers to producers a new alternative to marketing this cultivar as fresh-cut produce to distant markets.

Ethanol and acetaldehyde are by-products of fruit fermentation, and their presence is associated with offflavours and odours, which reduce quality of produce (Ke et al., 1991). Previous reports on the use of edible coatings and modified atmospheres indicated that their inappropriate use could lead to formation of odours and flavours unpleasant to the consumer (Baldwin et al., 1995; Guilbert et al., 1996). In this study, antioxidant or edible coatings prevented ethanol formation in all three cultivars, during storage at 5 C (Fig. 6a–c). It is noteworthy to mention that Ataulfo is the cultivar that developed higher amounts of ethanol during senescence (1.7 lL g)1 fresh weight), and this resulted in higher levels of acetaldehyde as well (Fig. 6d–f), while Keitt showed undetectable levels of ethanol during the 12 days of storage. Ethanol production was significantly reduced by all treatments in the same extent when compared with control samples. Acetaldehyde was synthesised in untreated mangoes to final concentrations of 0.16 and 0.08 lL g)1 in Ataulfo and Kent cultivars, respectively. In agreement with the undetectable ethanol levels observed in Keitt mangoes, acetaldehyde production was also the lowest of the three cultivars (0.03 lL g)1). It is worthwhile to indicate that antioxidants and protective films inhibited ethanol and acetaldehyde production of Ataulfo during the 21 days of storage (Fig. 6a and d). To date, there is no data indicating the threshold (perception) level for identification of ethanol or acetaldehyde volatiles in fresh-cut mangoes by panellists (consumers). Lack of production of ethanol and acetaldehyde is a good indicator of the favourable effect of edible coatings and antioxidants on the freshcut fruits. In apple slices, anaerobic fermentation byproducts were easily eliminated by use of a modified atmosphere of 21% O2 and 0.03% CO2 (Gil et al., 1998). Moreover, it is known that CO2 has an inhibitory effect on ethanol and acetaldehyde accumulation in apple slices. One important conclusion of this work is the excellent nutritional quality of the Ataulfo cultivar. We have shown that its vitamin C and b-carotene content are high for the fruit and are well kept during the processing as a fresh-cut product. Moreover, in agreement with previous studies, antioxidant, and to a lesser extent, protective films reduced colour changes, firmness loss and decay and browning indexes, while inhibiting

Abe, K. & Watada, A.E. (1991). Ethylene absorbent to maintain quality of lightly processed fruits and vegetables. Journal of Food Science, 56, 1589–1592. Ahvenainen, R. (1996). New approaches in improving the shelf life of minimally processed fruit and vegetables. Trends in Food Science and Technology, 7, 179–187. Allong, R., Wickham, L.D. & Mohammed, M. (2000). The effect of cultivar, fruit ripeness, storage temperature and duration on quality of fresh-cut mango. Acta Horticulturae, 509, 487–494. Baldwin, E.A. (1994). Edible coatings for fresh fruits and vegetables: past, present, and future. In: Edible Coatings and Films to Improve Food Quality (edited by J.M. Krochta, E.A. Baldwin & M. Nisperos-Carriedo). Pp. 25–63. Lancaster, PA: Technomic Publishers Co. Baldwin, E.A., Nisperos-Carriedo, M.O. & Baker, R.A. (1995). Use of edible coatings to preserve quality of lightly (and slightly) processed products. Critical Reviews in Food Science and Nutrition, 35, 509–524. Bayindirli, L., Su¨mnu¨, G. & Kamadan, K. (1995). Effects of SemperFresh and JONFRESH fruit coatings on poststorage quality of ‘‘Satsuma’’ mandarins. Journal of Food Processing and Preservation, 19, 399–407. Brecht, J.K. (1995). Physiology of lightly processed fruits and vegetables. Hortscience, 30, 18–22. Cantwell, M. (1995). Postharvest management of fruits and vegetables stems. In: Agro-ecology, Cultivation, and Uses of Cactus Pear. FAO: Plant Production and Protection Paper (edited by G. Barbera, P. Inglese & E. Pimienta-Barrios). Pp. 120–143. Rome: FAO publications. Chantanawarangoon, S. (2000). Quality maintenance of fresh-cut mango slices. Master Thesis. p. 79. Davis, CA: University of California. Davis, P.L. & Chace, W.R. (1969). Determination of alcohol in citrus juice by gas chromatographic analysis of headspace. Hortscience, 4, 117–119. Debeaufort, F., Quezada-Gallo, J.A. & Voilley, A. (1998). Edible films and coatings: tomorrow’s packaging: a review. Critical Reviews in Food Science and Nutrition, 38, 299–313. Doner, L.W. & Hickts, K.B. (1981). High-performance liquid chromatographic separation of ascorbic acid, erythorbic acid,



Acknowledgments

The authors thank the financial support from CIAD, CONACYT-SAGARPA grant 12510 (Me´xico) and CYTED Project XI.22 ‘Desarrollo de Tecnologı´ as para la Conservacioń de Productos Vegetales Frescos Cortados’. This work is part of the M.Sc. Thesis of the author Jorge Celis. References


dehydroascorbic acid, dehydroerythorbic acid, dikelogulonic acid, and dikeloglunoconic acid. Analytical Biochemistry, 115, 225–230. Gil, M.I., Gorny, J.R. & Kader, A.A. (1998). Responses of Fuji apple slices to ascorbic acid treatments and low-oxygen atmospheres. Hortscience, 33, 305–309. Gonza´lez-Aguilar, G.A., Wang, C.Y. & Buta, J.G. (2000). Maintaining quality of fresh-cut mangoes using antibrowning agents and modified atmosphere packaging. Journal of. Agricultural and Food Chemistry, 48, 4204–4208. Gonza´lez-Aguilar, G.A., Ruiz-Cruz, S., Cruz-Valenzuela, R., Rodriguez-Fe¨lix, A. & Wang, C.Y. (2004). Physiological and quality changes of fresh-cut pineapple treated with antibrowning agents. Food Science and Technology-Lebensmittel-Wissenschaft and Technologie, 37, 369–376. Gorny, J.R. (1997). Modified atmosphere packing and the fresh-cut revolution. Perishables Handing Newsletter, 90, 4–5. Gorny, J.R. (2001). Food Safety Guidelines for the Fresh-Cut Produce Industry, 4th edn. 216 p. Arlington, VA: International Fresh-cut Produce Association. Guilbert, S., Gontard, N. & Gorris, L.G.M. (1996). Review article. Prolongation of the shelf life of perishable food products using biodegradable films and coatings. Food Science and TechnologyLebensmittel-Wissenschaft and Technologie, 29, 10–17. James, K. & McGregor, A.M. (2000). Surface coatings for fresh fruit and freshly prepared fruits and vegetables. Postharvest News Information, 11, 45N–47N. Johnson, E. & Stevenson, R. (1978). Quantitative Analysis In Basic Liquid Chromatography. 223 p. Palo Alto, CA: Varian Associates. Ke, D., Rodriguez-Sinobas, L. & Kader, A.A. (1991). Physiology and prediction of fruit tolerance to low-oxygen atmospheres. Journal of the American Society for Horticultural Science, 116, 253–260. Klein, B.P. (1987). Nutritional consequences of minimally processing of fruits and vegetables. Journal of Food Quality, 10, 179–193. Lee, S.K. & Kader, A.A. (2000). Preharvest and postharvest factors influencing vitamin C content of horticultural crops. Postharvest Biology and Technology, 20, 207–220. Mejia, L.A., Hudson, E., Gonzales de Mejia, E. & Vazquez, F. (1988). Carotenoid content and vitamin A activity of some common cultivars of mexican peppers (Capsicum annuum) as determined by HPLC. Journal of Food Science, 53, 1448–1451. Mercadante, A.Z. & Rodriguez-Amaya, D.B. (1998). Effects of ripening, cultivar differences and processing of carotenoid composition of mango. Journal of Agricultural and Food Chemistry, 46, 128–130. Mitra, S. & Baldwin, E.A. (1997). Mango. In: Postharvest physiology and storage of tropical and subtropical fruits (edited by S. Mitra). Pp. 85–122. New York, NY: CAB International.

Nussinovitch, A. & Lurie, S. (1995). Edible coatings for fruits and vegetables. Postharvest News Information, 6, 53N–57N. Park, H.J., Bunn, J.M., Vergano, P.J. & Testin, R.F. (1994). Gas permeation and thickness of the sucrose polyesters, SemperFresh coatings on apples. Journal of Food Processing and Preservation, 18, 349–358. Rattanapanone, N., Lee, Y., Wu, T. & Watada, A.E. (2001). Quality and microbial changes of fresh-cut mango cubes held in controlled atmosphere. Hortscience, 36, 1091–1095. Saltveit, M.E. (1997). Physical and physiological changes in minimally processed fruits and vegetables. In: Phytochemistry of Fruit and Vegetables (edited by A. Toma´s-Barberań & R.J. Robins). Pp. 205– 220. Oxford: Clarendon Press. Saltveit, M.E. (2001). Fresh-cut product biology. In: Fresh-Cut Products: Maintaining Quality and Safety. Davis, CA: University of California. Selvaraj, Y., Kumar, R. & Pal, D.K. (1989). Changes in sugar, organic acids, amino acids, lipids constituents and aroma characteristics of ripening mango fruit. Journal of Food Science and Technology, 26, 308–313. Smith, J.S., Villalobos, M.C & Kottemann, C.M. (1986). A research note. Quantitative determination of sugars in various food products. Journal of Food Science, 51, 1373–1375. Soliva-Fortuny, R.C., Elez-Martı´ nez, P. & Martı´ n-Belloso, O. (2004). Microbiological and biochemical stability of fresh-cut apples preserved by modified atmosphere packaging. Innovative Food Science and Emerging Technologies, 5, 215–224. Thompson, A.K. (1998). Modified atmosphere packing. In: Controlled Atmosphere Storage of Fruit and Vegetables (edited by A.K. Thompson). Pp. 95–116. Wallingford: CAB International. Va´mos-Vigya´zo´, L. (1981). Polyphenol oxidase and peroxidase in fruits and vegetables. Critical Reviews in Food Science and Nutrition, 15, 49–127. Varoquax, P. & Wiley, R. (1994). Biological and biochemical changes in minimally processed refrigerated fruits and vegetables. In: Minimally Processed Fruits and Vegetables (edited by R.C. Wiley). Pp. 226–268. London: Chapman & Hall. Vazquez-Salinas, C. & Lakshminarayana, S. (1985). Compositional changes in mango fruit during ripening at different storage temperatures. Journal of Food Science, 50, 1646–1649. Watada, A.E., Abe, K. & Yamauchi, N. (1990). Physiological activities of partially processed fruits and vegetables. Food Technology, 44, 16–122. Yahia, E. (1998). Modified and controlled atmospheres for tropical fruits. Horticultural Reviews, 22, 123–183.



101


102

Original article Some properties of polyphenol oxidase from lily Ying Yang & Zhang Wang* Department of Food Science and Engineering, Southern Yangtze University, 170 Huihe Road, Wuxi, Jiangsu, 214036, China (Received 25 September 2005; Accepted in revised form 26 June 2006)

Summary

A study of crude polyphenol oxidase (PPO) from lily bulbs was carried out to provide information useful for guiding food processing operations. Optimum pH for the enzyme activity in the presence of catechol, were 4.0 and 7.0 at room temperature(approximately 20 C) and the enzyme was stable in the pH range from 5.0 to 6.5 at 4 C for 10 h. Its optimum temperature was 40 C and the heat inactivation of the enzyme followed first-order kinetics. Lily PPO possessed a diphenolase activity toward catechol, catechin and gallic acid; catechin was the best substrate for the enzyme considering the Vmax/Km ratio. The most effective enzyme inhibitor was sodium sulphite, although ascorbic acid, l-cysteine and thiourea were also effective inhibitors at high concentration. But NaCl and citric acid were poor inhibitors of the enzyme. Data generated by this study might help to better prevent lily bulbs browning.

Keywords

Browning, enzymology property, lily, polyphenol oxidase.

Introduction

Polyphenol oxidase (PPO) is widely distributed in vegetables and fruits. It is a copper-containing enzyme, which catalyses the hydroxylation of monophenols to o-diphenols and the oxidation of o-dihydroxyphenols to o-quinones, utilising molecular oxygen. The corresponding o-quinones subsequently polymerise to brown pigments. The action of PPO leads to the change in the colour and flavour of fruits and vegetables during harvest, storage and processing, which reduces the commercial value of the fruits, vegetables and their products. Lily belongs to the genus Lilium of the family Liliaceae, which contains about 80 species, native to temperate areas of northern hemisphere. According to the plant morphology and the planting situation of China, the species of lily used for food and medicine include Lilium lancifolium Thunb, Lilium brownie F.E. Brown var. viridulum Baker, Lilium pumilum DC and so on. The lilies used in this research which were grown in Yixing, Wuxi city belong to L. lancifolium Thunb (Liu & Huang, 2001). Its bulbs are used in traditional Chinese medicine to treat coughs, sore throat, tuberculosis, bronchitis, haematemesis, neurasthenia and restlessness and they can also improve the immunity function of the body. Lily bulbs contain not only starch and protein, but also steroidal saponins and alkaloids. In China lily bulbs can also be used to make many kinds of delicious dishes, soups and snacks. For these reasons fresh lily bulbs are

*Correspondent: E-mail: [email protected]

one kind of important vegetable in China. Beijing, Shanghai and other big cities consume 95% of the lily bulbs yielded in Lanzhou city, and 20% of these lily bulbs are exported to Japan and Southeast Asia regions. However, lily bulbs and their corresponding products can easily turn brown during transportation, storage and processing, leading to a great wastage. In some places people had to reduce the growth of lilies because it is hard to preservation. Therefore, the properties of PPO from lily bulbs need to be investigated to help people control the enzymatic browning. PPO has been widely studied in various plants such as pear (Halim & Montgomery, 1978), blueberry (Farid et al., 1997), plum (Siddig et al., 1992), grape (Cash et al., 1976), medlar (Barbaros et al., 2002), banana (Yang et al., 2000), tea leaf (Halder et al., 1998) and apple (Shannon & Pratt, 1967), but little has been known about lily PPO. In order to improve the utilisation of lily bulbs, a study of crude PPO from lily bulbs was carried out to provide information useful for guiding food processing operations. The objective of this study was to achieve a better understanding of the enzymological properties of lily PPO that catalyses the browning reaction during lily bulbs storage, processing and transportation. Materials and methods

Plant materials

Lily bulbs were purchased from local market in Wuxi, China and stored approximately 1 month at )20 C until used in the experiment.


Some properties of polyphenol oxidase from lily Y. Yang and Z. Wang

Crude enzyme preparation

Lily bulbs (50 g, )20 C) were homogenised with 400 mL cold acetone ()20 C) for 1 min. Acetone and water are miscible. If acetone was not completely excluded, the residual acetone could lead to enzyme denaturation. So the homogenate was filtered quickly with vacuum and the precipitate was dried by blowing cold wind until all acetone was evaporated. The remainder was dissolved in 150 mL of 0.1 m cold sodium phosphate buffer (pH 6.5). After being stirred for 30 min at 0 C, the suspension was centrifuged at 10 000 · g for 30 min at 4 C. The supernatant was used as crude PPO. Assay of PPO activity

Polyphenol oxidase activity was determined by measuring the increase in absorbance at 410 nm with a spectrophotometer. The sample cuvette contained 2.0 mL of 20 mm catechol (prepared in 0.2 m sodium acetate buffer, pH 4.0), 0.9 mL of 0.2 m sodium acetate buffer pH 4.0 and 0.1 mL of enzyme solution. Each sample was assayed in triplicate. Reference cuvette (blank) contained 2.0 mL of the same substrate solution and 1.0 mL of 0.2 m sodium acetate buffer. The initial velocity was calculated from the slope of the absorbance-time curve. One unit of PPO activity was defined as the change in absorbance of 0.001 min)1.

90 C, prior to the addition of 0.1 mL of enzyme solution. The relative activity of PPO at a specific temperature was determined spectrophotometrically by addition of enzyme to the mixture as rapidly as possible. Residual PPO activity was determined in the form of per cent residual PPO activity at the optimum temperature. To determine the effect of temperature on PPO stability, 0.2 mL of crude enzyme solution were injected into capillary tubes (i.d. 2.0 mm, o.d. 3.0 mm), and incubated at various temperatures for different time in a water-bath. After heat treatment, the enzyme samples were cooled in an ice bath immediately. The sample cuvette contained 2.0 mL of 20 mm catechol (prepared in 0.2 m sodium acetate buffer, pH 4.0), 0.9 mL of 0.2 m sodium acetate buffer, pH 4.0 and 0.1 mL of heated enzyme solution. The percentage residual PPO activity was calculated by comparison with unheated enzyme. Based on the Arrhenius equation, activation energy (Ea) was determined, by measuring the maximal initial rate at different temperatures and plotting the logarithmic value of Vmax vs. 1/T (Segel, 1976). The inactivation rate constant k values at different temperatures were determined by linear regression of the logarithm of the activity changes (A/A0) vs. time. Other activation parameters such as the free energy change (DG), enthalpy (DH) and entropy (DS) were determined as reported by Galani & Owusu Apenten, (1997).

Effect of pH on PPO activity and stability

Two kinds of buffer solutions were used for this study: 0.2 m sodium acetate buffer for the pH range of 3.6– 5.6, and 0.2 m sodium phosphate buffer for pH 6.0– 8.0. Catechol (20 mm), catechin (10 mm), DL-3,4-dihydroxyphenyalanine (DL-DOPA 20 mm), gallic acid (20 mm), l-tyrosine (5 mm) were used to determine the effect of pH on PPO activity. To determine the effect of pH on PPO stability, 0.1 mL of enzyme solution was incubated in 0.9 mL of various buffer solutions (pH 3.6–8.0) for 10 h at 4 C, and the residual activity was measured at 2 and 10 h, respectively. The enzyme activity was measured according to the method described above. Residual PPO activity was determined in the form of per cent residual PPO activity at the optimum pH. Effect of temperature on PPO activity and stability

The PPO activity was determined at various temperatures controlled by a water-bath. The mixtures of 0.9 mL of 0.2 m sodium acetate buffer (pH 4.0) and 2.0 mL of catechol solution (20 mm, prepared in 0.2 m sodium acetate buffer, pH 4.0) were incubated for 5 min at various temperatures over the range of 10–

KB T DG exp K¼ h RT

ð1Þ

DH ¼ Ea RT

ð2Þ

DH DG T

ð3Þ

DS ¼

where R (8.3145 J mol)1 K)1) is the universal gas constant, KB (1.3806 · 10)23 J K)1) is the Boltzman constant and h (6.6261 10)34 J s) is the Planck constant. Substrate specificity

Michaelis-Menten constant (Km) and maximum reaction velocity (Vmax) were determined using four substrates (catechol, catechin, gallic acid and l-tyrosine). They were assayed in different concentrations, and at the optimum pH and wavelength for each substrate: 0.3, 0.6, 1.2 and 2.4 mm for catachin and ltyrosine; 0.6, 1.2, 2.4 and 4.8 mm for catechol and gallic acid. Data were plotted as 1/activity and 1/ substrate concentration according to the method of Lineweaver & Burk (1934). Substrate specificity (Vmax/ Km) was calculated by using the data obtained on a Lineweaver-Burk plot.



103


Effect of inhibitors

The reaction mixture contained 2.0 mL of 20 mm catechol in 0.2 m sodium acetate buffer (pH 4.0), 0.4 mL of 0.2 m sodium acetate buffer (pH 4.0), 0.5 mL of inhibitor solution (0.1–10 mm) in 0.2 m acetate buffer (pH 4.0) and 0.1 mL of enzyme solution. Each of the inhibitors was assayed at various concentrations(0.1 mm, 1 mm and 10 mm). Relative enzyme activity was calculated as a percentage of the activity without any inhibitor. Results and discussion

In this study, all data were derived from a study of the crude enzyme preparations. Therefore, the properties described in this paper were only apparent quantities. However, the properties of a crude enzyme prepared in this study can be as relevant to the food industry as those of the purified or isolated enzyme (Duangmal & Owusu Apenten, 1999). Effect of pH on PPO activity and stability

120 100 80 60 40 20 0

this study, the pH optima of lily PPO with other substrates were also determined. The pH optima of the enzyme with gallic acid, DL-DOPA and catechin were all around 4.0 and 7.0, and the activity at pH 4.0 was higher than that at pH 7.0 except for catechin. These phenomena may be resulted from the presence of isoenzymes, in the case of which sweet cherry (Pifferi & Cultrera, 1974) had two predominant PPOs with two optimal pH values in the presence of pyrocatechol at pH 4.2 and 6.5, respectively, the exact reason should be detected by a better purification. As shown in Fig. 2, lily PPO was stable between pH 5.0 and 6.5, the activity decreased below pH 4.6 or above pH 7.0. There was no change in activity of lily PPO between pH 4.6 and pH 7.0 after 2 h and the enzyme retained more than 90% of its original activity at 4 C after 10 h. Peppermint PPO was stable between pH 6.0 and 7.0 (Kavrayan & Aydemir, 2001), pear PPO was unstable below pH 3.5 (Rivas & Whitaker, 1973) and medlar PPO was unstable below pH 4.02 or above pH 9.0 (Barbaros et al., 2002). These results indicated that PPO was unstable at acidic media and relatively stable near neutral pH. So treating vegetables and fruits with acid for some time could inhibit enzymatic browning. Effect of temperature on PPO activity and stability

The optimum temperature of lily PPO was 40 C (Fig. 3). It was higher than the optimum temperatures of PPO from plum [25 C (Siddig et al., 1992)] and medlar [35 C (Barbaros et al., 2002)]. The activity of lily PPO at 10 C was nearly 70% of that at 40 C. When the temperature was higher than 40 C, the effect of denaturation on the reaction rate was much greater, thus the activity of lily PPO decreased. According to the Arrhenius equation, the activation energy Ea in the

120

Relative activity (%)

Two pH optima were observed for lily PPO with catechol as substrate (Fig. 1): one at pH 4.0 and the other at pH 7.0. The peak at pH 4.0 was a little higher than that at pH 7.0. The activity decreased quickly between pH 4.0 and pH 5.0, and increased slowly between pH 5.0 and pH 7.0. In most cases, PPO enzymes from different plants have only one pH optimum, such as those from potato 5.0 (Balasingam & Ferdinand, 1970), medlar 6.5 (Barbaros et al., 2002), longan fruit 7.0 (Jiang, 1999), but some ones have two pH optima, such as those from apple (Shannon & Pratt, 1967) and sweet cherry (Pifferi & Cultrera, 1974). Aylward and Haisman (Aylward & Haisman, 1969) reported that the optimum pH for maximum PPO activity in plants varied from approximately 4.0 to 7.0, depending on the extraction methods, substrate used for assay and localisation of the enzyme in the plant cell. In


104

100 80 60 40

2h 10 h

20

3

4

5

6

7

8

9

PH Figure 1 The pH optima of lily PPO with catechol as substrate. The maximum activity was considered as 100%.


0 3

4

5

6 pH

7

8

9

Figure 2 pH stability of lily PPO with catechol as substrate. The maximum activity was considered as 100%.



Table 1 Thermodynamic parameters for thermal inactivation of lily

120


PPO

100 80 60 40 20 0

0

20

40

60

80

Figure 3 The optimum temperature of lily PPO with catechol as substrate. The maximum activity was considered as 100%.

reaction during conversion of reactant to product was calculated to be 11.2 kJ mol)1. The enzyme showed no loss in activity after 30 min at 20 C and 30 C (Fig. 4). At higher temperatures, it was inactivated. The rate of heat inactivation increased with increasing temperature and could be described by a firstorder decay process. When the temperature was higher than 60 C, the rate of enzyme denaturation greatly increased. Fifty per cent of lily PPO activity was lost after heating 27.2, 4.5 and 1.3 min at 60, 70 and 80 C, respectively, and the enzyme could be completely inactivated after 15 s at 100 C. Compared with lily PPO, PPOs from plum (Siddig et al., 1992) and banana (Yang et al., 2000) were more stable, they could maintain their activities for 30 min without any loss at 70 C. The thermodynamic parameters for thermal inactivation of lily PPO are shown in Table 1. The k-values were calculated from the slope of the lines in Fig. 4, Ea for crude lily PPO heat-inactivation was calculated from 0.2 0

Ig (relative activity)

–0.2 –0.4 20 °C, 30 °C 40 °C

–0.8 –1

50 °C 60 °C 70 °C 80 °C

–1.2 –1.4

k · 104 (s)1)

DH (kJ mol)1)

DG (kJ mol)1)

DS (J mol)1 K)1)

40 50 60 70 80

1.3 2.3 4.2 29.8 90.9

98.2 98.1 98.0 98.0 97.9

52.1 52.4 52.4 48.5 46.7

147.2 141.5 137.0 144.3 145.0

100

Temperature (°C)

–0.6

Temperature (°C)

the plot of lnk vs. 1/T. From 40 to 80 C Ea for crude lily PPO heat-inactivation was calculated to be 100.8 kJ mol)1, which was much higher than that of catalysis (11.2 kJ mol)1), so the effect of the temperature on the denaturation rate was much greater than that on catalysis. The average DH and DS were 98.02 kJ mol)1 and 143.0 J mol)1 K)1, respectively. The DH value of lily PPO was similar to that of potato PPO (Duangmal & Owusu Apenten, 1999), implying that the number of noncovalent bonds broken in forming a transition state for enzyme inactivation was similar. However, PPO from lily had a smaller free energy change than that from potato. The DS value of lily PPO was larger than that of potato PPO, hence the net change in disorder of enzyme and solvent accompanying the transition state formation was larger and benefit to reduce the DG value. Substrate specificity

Lily PPO showed different activities towards different substrates. Table 2 indicates that lily PPO had different optimum pH, Km, and Vmax values. Km is a characteristic constant of an enzyme for each particular substrate, and shows the affinity between enzyme and substrate. Vmax, the maximum velocity of an enzyme-catalysed reaction when the concentration of enzyme does not change, shows the catalytic efficiency. The best substrate for each enzyme depends on two factors: strong substrate binding, as expressed by a low Km, and high catalytic efficiency, as expressed by a high Vmax value. Therefore, the criterion for the evaluation of the best substrate is based on the ratio of Vmax/Km (Palmer, Table 2 Kinetic parameters for the oxidation of various substrates by lily PPO

–1.6 –1.8 –2 0

10

20 30 Time (min)

40

Figure 4 Heat inactivation of lily PPO at various temperatures with catechol as substrate.

Substrate

pH optimum

Vmax (UmL)1 min)1)

Km (mM)

Vmax/Km

Catechol Catechin Gallic acid L-tyrosine

4.0 7.0 4.0 nd

1429 769.2 496.2 nd

3.4 2.2 2.1 nd

416.5 667.7 232.3 nd

nd, not detected.



105

106


1995). The highest substrate binding and catalytic efficiency of lily PPO were with gallic acid and catechol, respectively. Although both the substrate binding and catalytic efficiency of lily PPO with catechin were not the highest, catechin was the best substrate for the enzyme considering the Vmax/Km ratio. Lily PPO showed no activity toward l-tyrosine, which suggested that lily PPO had no monophenolase activity. Effect of inhibitors

The effects of various compounds on the activity of lily PPO are shown in Table 3. Sodium sulphite was the most effective inhibitor. Sodium sulphite is a reducing agent and can react with o-quinones to form a colourless complex, thus reduces the activity of PPO on phenols (Wang, 1991). It has been suggested that sulphite reacts with disulfide bonds within PPO, leading to changes in the tertiary structure of the enzyme and inactivation (Duangmal & Owusu Apenten, 1999). Although sodium sulphite was effective, it is harmful to human health because of a number of reports that some humans, especially asthmatics, may be sensitive to it (Sapers, 1993). Therefore, it is not recommended to use this reagent to prevent browning in food. Thiourea, lcysteine and ascorbic acid at high concentration could greatly inhibit PPO activity. During the experiment, a lag period for oxidation was observed when they were used, suggesting that there was no quinone accumulation in the reaction mixture during this period, although Table 3 Effect of various compounds on the activity of lily PPO

Inhibitor Thiourea

L-cysteine

Ascorbic acid

Benzoic acid

Sodium sulphite

NaCl

Citric acid

Concentration (mM)

Relative activity

10.0 1.0 0.1 10.0 1.0 0.1 10.0 1.0 0.1 10.0 1.0 0.1 10.0 1.0 0.1 10.0 1.0 0.1 10.0 1.0 0.1

0.0 3.1 62.1 3.5 6.9 58.6 0.0 3.5 55.2 13.8 55.2 75.9 0.0 0.0 0.0 34.5 69.0 86.2 65.5 75.9 82.8


the mechanisms of their inhibition effects were not the same. Thiourea and l-cysteine can react with the copper of the enzyme to inhibit PPO activity (Kahn & Andrawis, 1985; Lerch, 1987). l-cysteine can also easily form complex with o-quinones and PPO is inhibited by the formation of additional products (Janovitz-Klapp et al., 1990). There are three main reasons for the antibrowning effect of ascorbic acid: (i) it is a reducing reagent of quinones; (ii) it can form a complex with Cu2+ in PPO to inactivate the enzyme; (iii) it can be oxidised directly by PPO. PPO catalyses the oxidation of phenolic substrates to o-quinones whilst ascorbic acid converts o-quinones back to phenolic compounds (Wang, 1991). Changes in the absorbance (410 nm) can not be observed at the beginning of the reaction, thus creating a lag period (Duangmal & Owusu Apenten, 1999). If the concentration of ascorbic acid was low, when it was used up, PPO recovered its activity. Golan et al. (1984) had reported that ascorbic acid caused the irreversible inhibition of mushroom PPO. Therefore, high ascorbic acid concentration can have a satisfactory effect. The activity of PPO decreased as the concentrations of NaCl and citric acid increased, but the effect was not as good as the inhibitors mentioned above. NaCl can exclude the O2 of the media, and interact with the copper at the active centre of the enzyme (Martinez & Whitaker, 1995). In this study the concentration of NaCl might have been low, resulting in unsatisfactory effect. The buffer used in this study had a high buffer capacity, thus the inhibition effect of citric acid by reducing the pH of the system was not obvious. In a word, use of appropriate concentration of l-cysteine or ascorbic acid is a good method to inhibit lily PPO activity. Conclusion

As a kind of nourishing tonic vegetable, lily bulbs can be used as food or medicine. However, browning reactions in lily bulbs are a serious problem, which can greatly decrease their commercial value. So a study on the properties of lily PPO was carried out to protect the lily bulbs from browning. Data on pH, temperature, specific substrates and inhibition profiles of lily PPO were similar to those of other plant PPOs enzymes. Data generated by this study can help people provide appropriate conditions or methods for transportation, storage of lily bulbs as well as prevention from enzymatic browning during processing. This will greatly improve the commercial utilisation of lily bulbs. References Aylward, F. & Haisman, D.R. (1969). Oxidation systems in fruits and vegetables–their relation to the quality of preserved products. Advances in Food Research, 17, 71–76.



Balasingam, K. & Ferdinand, W. (1970). The purification and properties of a ribonucloenzyme o-diphenol oxidase from potatoes. Biochemistry Journal, 118, 15–23. Barbaros, D., Ahmet, C., Nese, A., Asim, K. & Saadettin, G. (2002). Characterization of polyphenoloxidase from medlar fruits (Mespilus germanica L., Rosaceae). Food Chemistry, 77, 1–7. Cash, J.N., Sistrunk, W.A. & Stutte, C.A. (1976). Characteristics of Concord grape polyphenol oxidase involved in juice color loss. Journal of Food Science, 41, 1398–1402. Duangmal, K. & Owusu Apenten, R.K. (1999). A comparative study of polyphenol oxidases from taro (Colocasia esculenta) and potato (Solanum tuberosum var. Romano). Food Chemistry, 64, 351–359. Farid, K., Bernard, R., Michel, G. & Maurice, M. (1997). Mechanism of browning in fresh highbush blueberry fruit (Vaccinium corymbosum L): partial purification and characterization of blueberry polyphenol oxidase. Journal of Agriculture and Food Chemistry, 73, 513–516. Galani, D. & Owusu Apenten, R.K. (1997). The comparative heat stability of bovine b-lactoglobulin in buffer and complex media. Journal of the Science of Food and Agriculture, 74, 89–98. Golan, A., Goldhirsh, A. & Whitaker, J.R. (1984). Effect of ascorbic acid, sodium bisulfite and thiol compounds on mushroom polyphenol oxidase. Journal of Agricultural Food Chemistry, 32, 1003– 1009. Halder, J., Tamuli, P. & Phaduri, A.N. (1998). Isolation and characterization of polyphenol oxidase from Indian tea leaf (Camellia sinensis). Journal of Nutritional Biochemistry, 9, 75–80. Halim, D.H. & Montgomery, M.W. (1978). Polyphenol oxidase of d’Anjou pears (Pyrus communis L.). Journal of Food Science, 43, 603–608. Janovitz-Klapp, A.H., Richard, F.C. & Nicolas, J.J. (1990). Inhibition studies on apple polyphenol oxidase. Journal of Agricultural and Food Chemistry, 38, 926–931. Jiang, Y.M. (1999). Purification and some properties of polyphenol oxidase of longan fruit. Food Chemistry, 66, 75–79. Kahn, V. & Andrawis, A. (1985). Inhibition of mushroom tyrosinase by tropolone. Phytochemistry, 24, 905–908.

Kavrayan, D. & Aydemir, T. (2001). Partial purification and characterization of polyphenoloxidase from peppermint (Mentha piperita). Food Chemistry, 74, 147–154. Lerch, K. (1987). Molecular and active site structure of tyrosinase. Life Chemistry Report, 5, 221–234. Lineweaver, H. & Burk, D. (1934). The determination of enzyme dissociation constant. Journal of American Chemistry Society, 56, 658–661. Liu, D.J. & Huang, Y.J. (2001). Lilies. Pp. 8–12. Beijing, China: Traditional Chinese Medicine Press. Martinez, M. & Whitaker, J.R. (1995). The biochemistry and control of enzymatic browning. Trends in Food Science and Technology, 6, 195–200. Palmer, T. (1995). Kinetics of single-substrate enzyme catalysed reactions. In: Understanding Enzymes, 4th edn. Pp. 107–127. Hertfordshire: Prentice Hall/Ellis Horwood. Pifferi, P.G. & Cultrera, R. (1974). Enzymatic degradation of anthocyanins: the role of sweet cherry polyphenol oxidase. Journal of Food Science, 39, 786–791. Rivas, N.J. & Whitaker, J.R. (1973). Purification and some properties of two polyphenol oxidase from Bartlett pears. Plant Physiology, 52, 501–507. Sapers, G.M. (1993). Browning of foods: control by sulfites, antioxidants and other means. Food Technology, 47, 75–84. Segel, I.H. (1976). Biochemical Calculations. Pp. 273. New York, NY: John Wiley Sons. Shannon, C.T. & Pratt, D.E. (1967). Apple polyphenol oxidase activity in relation to various phenolic compounds. Journal of Food Science, 32, 479–483. Siddig, M., Sinha, N.K. & Cash, Y.N. (1992). Characterization of a polyphenol oxidase from Stanley plums. Journal of Food Science, 57, 1177–1179. Wang, Z. (1991). Food Enzymology. Pp. 268. Beijing, China: China Light Industry Press. Yang, C.P., Fujita, S., Ashrafuzzaman, M.A., Nakamura, N. & Hayashi, N. (2000). Purification and chracterization of polyphenol oxidase from banana (Musa sapientum L.) pulp. Journal of Agricultural Food Chemistry, 48, 2732–2735.



107


108

Original article A rapid direct solvent extraction method for the extraction of cyclobutanones from irradiated chicken and liquid whole egg Ihab Tewfik School of Biosciences, University of Westminster, 115, New Cavendish Street, London W1W 6UW, UK (Received 31 March 2006; Accepted in revised form 26 June 2006)

2-Alkylcyclobutanones (cyclobutanones) are accepted as radiolytic markers in lipid-containing foods (LCF). To date, cyclobutanones have not been detected in non-irradiated foods. Their identification is the basis of a standardised detection test for irradiated LCF (BS EN 1785, 2003). This paper reports work to develop and refine a new rapid method for the analysis of cyclobutanones in irradiated LCF (e.g. chicken meat and liquid whole egg). Direct solvent extraction (DSE) enables the efficient screening of large numbers of food samples and is not as resource intensive as the BS EN 1785 (2003) method. The new DSE appears to be a promising, rapid, simple and robust method for the analysis of irradiated lipid-containing foods.

Summary

Keywords

Food processing chemistry, food safety, gas chromatography–mass spectroscopy, irradiated food, irradiation.

Introduction

Foodstuffs are irradiated primarily to limit microbial deterioration and so improve shelf life (Rahman et al., 1996a; Tewfik et al., 1998; Delinceé, 2002a). Food irradiation technique dates back to 1905 [e.g. British Patent No. 1609 (Appleby & Banks, 1905)], but it has found limited commercial use or public acceptance. Currently, two ionising irradiation sources are used commercially, which include gamma radiation [derived from 60Co (cobalt-60) sources] and electron irradiation (derived from electron accelerators) (WHO, 1994, 1999; Rahman et al., 1995; FSIS, 1999; Tewfik et al., 1999; WHO, 2000; Diehl, 2002). Upon irradiation, several changes can be induced in food that could be utilised for detection purposes. Nevertheless, with the exception of cyclobutanones, decades of research have failed to find chemical markers that are unique to irradiated food (Letellier & Nawar, 1972; Nawar, 1986; Boyd et al., 1991; Stevenson, 1992). It is generally assumed that gamma irradiation and electron irradiation are comparable techniques, both giving rise to cyclobutanones as radiolytic markers (BS EN 1785, 2003; WHO, 1994; Tewfik et al., 1996, 1998, 1999; Delinceé, 2002a,b). 2-Dodecylcyclobutanone (2-DCB) arises from the palmitic acid (n-hexadecanoic acid), while 2-tetradecylcyclobutanone (2-TCB) arises from stearic acid (n-octadecanoic acid) naturally present in foodstuffs (Nawar, 1986; Boyd et al., 1991; Stevenson, Correspondent: Fax +44 (0) 207 742 8189; e-mail: i.tewfi[email protected]

1992). Other products of the irradiation process include hydrocarbons, lipid hydroperoxides and free radicals (Nawar, 1986; Rahman et al., 1995; Delinceé, 1998; Verniest et al., 2004). Although the use of gas chromatography–mass spectroscopy (GC–MS) to detect cyclobutanones is relatively straightforward (Boyd et al., 1991; Stevenson, 1992), their extraction from complex food matrices remains a problem. The BS EN 1785 (2003) has been shown to be applicable to a wide range of high lipidcontaining food types and irradiation levels. It is, however, a complex procedure involving a long Soxhlet extraction (6 h) followed by a slow column chromatography stage utilising large volumes of expensive solvents. The BS EN 1785 (2003) method is therefore time consuming, complex, expensive and not suited as a quick screening tool. A need, therefore, exists to develop a rapid method capable of screening a variety of irradiated foods (Rahman et al., 1996a,b; Delinceé, 1998; Del Mastro, 1999; Tewfik et al., 1999; Diehl, 2002). The overall objective of this study was to develop and validate a rapid, cheap, low technology alternative to the BS EN 1785 (2003) extraction method. As a minimum, the method would be capable of qualitatively identifying one or more markers in foodstuffs that have been irradiated. At best, the methodology would also be capable of quantifying the concentration of one or more markers and therefore the irradiation dose applied to the food. A simple and rapid method [direct solvent extraction (DSE)–GC–MS] for the detection of irradiated chicken

doi:10.1111/j.1365-2621.2006.01399.x 2007 The Author. Journal compilation 2007 Institute of Food Science and Technology Trust Fund

DSE method for extraction of cyclobutanones I. Tewfik

and liquid whole egg samples at various doses is presently described for the first time and compared with the existing method (i.e. BS EN 1785, 2003). Materials and methods

Samples

A fresh British minced chicken and liquid whole egg samples were purchased from various supermarkets in London, UK. The minced chicken samples were homogenised in a blender and equally divided into two portions (for irradiation and control). The liquid whole egg, after de-shelling, were thoroughly mixed and equally divided into two portions (for irradiation and control) prior to sampling. All samples were stored ()20 C) until required and to reduce the chance of microbial contamination (control samples). Chemicals and reagents 1 Hexane – AR from Sigma Chemical Co., Dorset, UK. 2 Heptane – AR from Sigma Chemical Co. 3 Florisil – mesh 60–100 PR grade from Sigma Chemical Co. 4 Sodium sulphate – AR from Sigma Chemical Co. 5 2-Dodecylcyclobutanone, purity >99%, from QuChem, Queen’s University of Belfast, UK. 6 Isodrin – from Sigma Chemical Co. Irradiation of samples

Chicken and egg samples were irradiated to different doses at )20 C to guarantee same irradiation conditions as control samples. Fresh homogenised chicken samples were irradiated at 1.0, 2.0, 3.0, 4.0 and 5.0 kGy while liquid whole egg samples were irradiated at 1.0, 1.5 and 2.0 kGy. Irradiation was carried out using a 60Co as gamma ray source, at the National Physical Laboratory (NPL), Teddington, UK. Irradiator type was Gammacell 220, Nordion, Canada (dose rate was 1.0 kGy h)1). Amber Perspex dosimeters (Type 3402B; AEA Technology, Harwell, Oxfordshire, UK) were used for the measurement of the applied irradiation dose. The absorbance of the dosimeters was measured spectrophotometrically at 603 nm. The corresponding dose was obtained from calibration graphs by NPL. Methods The direct solvent extraction method Direct solvent extraction sample size. The extraction of

cyclobutanones using the DSE method is essentially identical for all lipid-containing foodstuffs. The only

variation in the procedure relates to the sample size; the exact sample size is dependent on the fat content of the sample. The small florisil column used in the clean-up step of DSE method is capable of removing limited amount of lipid materials from the extract. In the case of large sample size, or high fat content, triglycerides may also appear in the final cyclobutanone extract. Ideal sample size for chicken and egg samples was 2.0 g and 3.0 g, respectively. Direct solvent extraction reagent preparation. Extraction

solvent system: hexane:heptane, 9:1 v:v. Deactivation of florisil: florisil was first roasted at 550 C for 5 h, then cooled to 100 C and then transferred to a desiccator to cool to ambient temperature, it was then deactivated by adding 20 g of distilled water to 100 g of florisil while shaking thoroughly. The deactivated florisil was stored in an airtight jar and allowed to stand for 3 h before use. Sodium sulphate (AR) was roasted at 450 C for 4 h, cooled to 100 C then stored in a desiccator over silica gel. Internal standard solution: 10 ppm (mg L)1) isodrin in hexane, stored at 4 C. This solution is stable for at least 4 months. 2-Dodecylcyclobutanone standard: stock solutions at 1000 and 10 ppm (mg L)1) in AR hexane. Store at 4 C. In current study, it was envisaged to include more cyclobutanone standards; however, only very few were commercially available, this may be remedied by the synthesis of more 2-alkylcyclobutanone standards such as 2-TCB, 2-tetradecenyl- and 2-tetradecadienyl-cyclobutanones (Verniest et al., 2004). Direct solvent extraction glassware. No special glassware

is required, except 250 mL separating funnels fitted with Teflon stopcocks. Direct solvent extraction methodology. In a mortar and

pestle, grind suitable sample size (i.e. 2.0 g of chicken and 3.0 g of egg samples) with sufficient roasted anhydrous sodium sulphate, to form a fine tilth, typically 10–20 g of sodium sulphate. Transfer the tilth to a 250 mL separating funnel, fitted with a Teflon stopcock, and add 35 mL of the extraction solvent system. Shake the contents vigorously for 10 min. Allow the mixture to stand for 10 min then shake for a further 5 min. In dropwise manner, drain the extract through a 5 g bed of 20% deactivated florisil in a glass-woolplugged 250 mL funnel. Add a further 30 mL of the extraction solvent system to the mixture retained in the separating funnel, shake for a further 5 min and pass the extract through the same florisil bed. Evaporate the combined extracts using a water bath at 40 C and gentle stream of dry nitrogen. When the volume has been reduced to 0.05) in the two different curings (Table 3). Cheng & Ockerman (2003) reported that tumbling significantly increased the pH values of roasted beef. There were significant differences (P < 0.05) in the moisture contents between dry-cured and brine-injected samples (Table 3). Our results were at variance with the findings of Gurbuz (1994) who determined the moisture contents of dry-cured and brine-injected pastirma as 42.65% and 45.16% respectively. Although tumbling resulted in a significant decrease (P < 0.05) in moisture contents of the brine-injected samples, it caused significant increases (P < 0.05) in moisture contents of the dry-cured samples. Significant increase in the moisture contents of the dry-cured samples is, therefore, becoming evident that tumbling resulted in increase in water-holding capacity as reported by some researchers (Dzudie & Okubanjo, 1999; Pietrasik & Shand, 2003; Hullberg & Lundstro¨m, 2004; Pietrasik & Shand, 2004). But an explanation of those results that tumbling caused a decrease in the moisture contents of brine-injected samples may not be possible. Pietrasik & Shand (2003) stated that while quantity of brine had a significant effect on the moisture content of cooked beef rolls, tumbling was not of significant importance. On the other hand, Dzudie & Okubanjo (1999) determined that moisture contents of the tumbled raw ham prepared using pre-rigour meat were significantly higher than the post-rigour meat. Traditional curing has caused higher salt contents in the samples when compared with the brine-injected samples. These differences were significant (P < 0.05) at final product (pastirma) (Table 3). Similar result was reported by Gurbuz (1994) who determined the lowest salt content in brine-injected pastirma when compared with pastirma samples dry cured and submerged in brine. In the present study, salt content of the tumbled samples in dry-cured group was lower than the non-tumbled group. But the effect of tumbling on salt content of the samples was not significant (P > 0.05). While Dzudie & Okubanjo (1999) stated that increased tumbling time caused significant increase in the salt contents of goat hams, Hullberg et al. (2005) stated a similar result of our study that salt contents of the cured-smoked tumbled pork loin was lower than the non-tumbled samples. Microbiological analysis

After curing process for 36 h, coliform bacteria decreased from 2.48 to 3.50 log10 CFU g)1 to 1.71– 2.44 log10 CFU g)1, and it was not observed after drying about 14 days. This result might be due to the effects of drying and curing agents such as nitrate (Cassens, 1994). Similar results are reported by other researchers (Kotzekidou & Lazarides, 1991; Dogruer,


1992; Dogruer et al., 2003; Gurbuz et al., 2003). Salama & Khalafalla (1987) reported that high concentration of NaNO2 was quite sufficient to get rid of coliform group as one of the major groups reflecting the sanitary quality of food products within 24 h. Kotzekidou & Lazarides (1991) stated that low aw and pH values and the competitive flora (lactic acid bacteria) probably contribute to the inhibition of Enterobacteriaceae. Effect of different curing techniques on Staphylococcus/Micrococcus number was found significant (P < 0.05) after curing. Gurbuz (1994) found Staphylococcus/Micrococcus number in dry-cured and brineinjected pastirma as 9.2 · 105 and 3.8 · 106 respectively. In the present study tumbling processes caused significant increases in Staphylococcus/Micrococcus number especially in brine-injected pastirma samples. Hunt & Kropf (1984) stated that the mechanical disruption of the meat facilitates spreading and contamination of micro-organism. On the other hand, Kotzekidou & Lazarides (1991) reported that Staphylococcus aureus was not detected in any of commercial pastirma samples, that if the product is contaminated with S. aureus this micro-organisms will not grow but they will survive for at least 60 days of refrigerated storage. Differences seen in total mesophilic aerobic counts in two applications (curing techniques and tumbling process) were not significant. Similarly Cheng & Ockerman (2003) reported that there was no significant difference for mesophile numbers between non-tumbled and tumbled roast beef of 0%, 0.25%, 0.4% phosphate level. Gurbuz (1994) determined the total mesophilic aerobic counts in dry-cured and brine-injected pastirma samples as 6.6 · 106 and 7.5 · 106 respectively. In the present study, halophilic bacteria numbers of the dry-cured and the brine-injected pastirma were significantly different (P < 0.05). These differences may have resulted from salt contents of the pastirma samples. In addition, halophilic bacteria were significantly higher (P < 0.05) in the tumbled samples than the non-tumbled samples. The lactic acid bacteria counts were significantly higher in the brine-injected pastirma than the dry-cured pastirma. These differences might be related to the pH value, moisture and salt contents of the samples. Significant increases in the lactic acid bacteria of the tumbled samples after curing and third drying were seen. It may have resulted from lactic acid bacteria contamination caused by mechanical disruption of the meat by tumbling process. Lactic acid bacteria, playing an important role in the forming of sensory attributes of pastirma (Aksu & Kaya, 2002a), contaminate pastirma during manufacture. Higher lactic acid bacteria contents may be related to the moisture content of the samples. As a matter of fact, moisture content was found higher in non-tumbled samples than tumbled samples. On the other hand, Yetim et al. (1996) reported that tumbling


New technique in pastirma production A. Guner et al.

)1

Table 4 Effect of tumbling and curing techniques on some microbiological characteristics of pastirma (log10 CFU g )

Before curing Coliform bacteria Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling Staphylococcus/Micrococcus Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling Total mesophilic aerobic Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling Halophilic bacteria Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling Lactic acid bacteria Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling Yeast and mould Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling

After curing (with or/not tumbling)

After third drying

Final product (pastirma)

2.48 2.47 3.50 3.49

± ± ± ±

0.035L 0.025 0.021K 0.043

1.71 2.11 1.92 2.44

± ± ± ±

0.137Y 0.408X 0.309N 0.026M

3.56 3.57 3.66 3.69

± ± ± ±

0.083 0.057 0.114 0.025

3.53 4.04 3.12 4.30

± ± ± ±

0.014Y, K 0.108X 0.022N, L 0.056M

3.76 4.36 4.28 3.34

± ± ± ±

0.232 0.120 0.174M 0.035N

4.25 4.18 3.58 6.20

± ± ± ±

0.156 0.495 0.303N 0.349M

5.44 5.45 5.54 5.54

± ± ± ±

0.026 0.074 0.016 0.036

4.75 4.92 4.67 4.92

± ± ± ±

0.116 0.091 0.240 0.149

5.74 4.83 5.82 4.94

± ± ± ±

0.983 0.440 0.629 0.319

8.34 7.42 8.17 8.54

± ± ± ±

0.769 1.260 0.207 0.215

3.07 3.13 3.08 3.07

± ± ± ±

0.005 0.066 0.040 0.054

3.15 3.55 3.30 3.72

± ± ± ±

0.054Y 0.025X 0.129N 0.042M

2.89 4.72 3.23 3.50

± ± ± ±

0.102Y 0.423X 0.525 0.210

4.25 ± 0.211K 3.88 ± 0.202 2.91±0.232N, L 6.12 ± 0.311M

3.51 3.51 3.46 3.44

± ± ± ±

0.033 0.057 0.060 0.015

2.95 3.03 2.47 3.10

± ± ± ±

0.033K 0.044 0.130N, 0.062M

2.78 4.48 5.21 4.17

± ± ± ±

0.133Y, L 0.325X 0.440 K 0.875

5.52 6.87 7.63 7.66

± ± ± ±

0.708L 0.797 0.049K 0.116

3.30 3.30 3.25 3.27

± ± ± ±

0.060 0.042 0.019 0.071

2.76 2.33 2.36 2.78

± ± ± ±

0.089X 0.070Y 0.343 0.235

1.67 1.53 2.31 1.46

± ± ± ±

0.142 0.119 0.507M 0.087N

1.20 1.30 2.15 2.02

± ± ± ±

0.100 0.001 0.579 0.276

n.d. n.d. n.d. n.d.

L

n.d. n.d. n.d. n.d.

Values after ‘±’ represent standard error. n.d., Not determined. X, Y: Different uppercase letters in a column show significant differences between the samples salted dry. M, N: Different uppercase letters in a column show significant differences between the samples salted with brine injection. K, L: Different uppercase letters in a column show significant differences between the non-tumbled samples.

process caused a slight decrease in the counts of aerobic plate count (APC), Pseudomonas, lipolytic and proteolytic bacteria. The data presented in Table 4 state that yeast and mould number found between 3.25 and 3.30 log10 CFU g)1 in the fresh meat decreased to between 2.15 and 1.20 log10 CFU g)1 in the final product. Dogruer (1992) reported that no significant difference was observed in the yeast and mould numbers of pastirma produced by using different curing times and press weights. Gurbuz (1994) determined no significant changes in yeast and mould number of fresh meat and pastirma. On the other hand, Kotzekidou & Lazarides (1991) reported that yeast and mould were not detected in commercial pastirma samples. Effect of tumbling on yeast and mould was significantly important (P < 0.05) in the dry-cured samples after curing and in the brineinjected samples after third drying. When differences were seen in yeast and mould numbers, the salt contents

of the samples which have higher yeast and mould counts were found lower than the samples having low yeast and mould. Sensory evaluation

As could be followed from the Table 5, dry-cured pastirma (traditional group) received higher colour and appearance score while brine-injected pastirma had higher flavour and texture score. But it was significant (P < 0.05) only for colour attribute. These results might be due to the salt contents of pastirma samples seen in Table 3. Just as Heikal et al. (1972) reported that the amount of sodium chloride had a great effect on the quality of pastirma, but excessive amount of sodium chloride decreased the sensory properties, especially texture attributes. They found the salt contents of pastirma between 10.67% and 14.02%. On the other hand, tumbling process affected the flavour,



127

128


Dry curing Dry curing + tumbling Brine injection Brine injection + tumbling

Flavour

Texture

Colour

6.00 7.22 6.55 6.44

6.44 7.00 7.33 6.33

7.77 7.66 6.00 6.88

± ± ± ±

0.623 0.464 0.689 0.689

± ± ± ±

0.503 0.527 0.552 0.500

± ± ± ±

Appearance 0.464K 0.527 0.645L 0.351

7.22 7.77 6.00 7.22

± ± ± ±

0.464 0.618 0.745 0.618

Table 5 Effect of tumbling and curing techniques on some sensory attributes of pastirma

±: Values after ‘±’ represent standard error. K, L: Different uppercase letters in a column show significant differences between the non-tumbled samples.

texture and appearance positively in the dry-cured pastirma and the color and appearance in the brineinjected pastirma. Differences in sensory scores of pastirma samples might also be related to the lactic acid bacteria effective on sensory properties of pastirma (Aksu & Kaya, 2002a). As could be seen from Table 4, lactic acid bacteria was higher in the tumbled pastirma than the non-tumbled pastirma. Aktas et al. (2005) have used starter culture combination to degrade myofibrillar protein for improving sensory attributes, especially tenderness and juiciness, of pastirma. Conclusions

This research indicated that dry curing had considerable positive effect on pastirma. Using brine injection of curing agents instead of dry curing had no large effects on the quality characteristics of pastirma. Tumbling caused important increase in the microbial count especially in brine injection group. Although tumbling and injection curing did not interact on sensory attributes, tumbling increased the sensory properties of dry-cured pastirma. References Aksu, M.I. & Kaya, M. (2001). Pastirma uretiminde starter kultur kullaniminin son urun ozellikleri uzerine etkisi. Turkish Journal of Veterinary Animal Science, 25, 847–854 (in Turkish). Aksu, M.I. & Kaya, M. (2002a). Potasyum nitrat ve starter kultur kullanilarak uretilen pastirmalarin bazi mikrobiyolojik ve kimyasal ozellikleri. Turkish Journal of Veterinary Animal Science, 26, 125– 132 (in Turkish). Aksu, M.I. & Kaya, M. (2002b). Farkli kurleme yontemleri ve starter kultur kullanilarak pastirma uretimi. Turkish Journal of Veterinary Animal Science, 26, 909–916 (in Turkish). Aksu, M.I. & Kaya, M. (2002c). Ticari starter kultur preparatlarinin pastirma uretiminde kullanim imkanlari. Turkish Journal of Veterinary Animal Science, 26, 917–923 (in Turkish). Aksu, M.I., Kaya, M. & Ockerman, H.W. (2004). Effect of modified atmosphere packaging, storage period and storage temperature on the residual nitrate of sliced-pastirma, dry meat product, produced from fresh meat and frozen/thawed meat. Food Chemistry, 93, 237–242. Aktas, N., Aksu, M.I. & Kaya, M. (2005). Changes in myofibrillar proteins during processing of pastirma (Turkish dry meat product) produced with commercial starter cultures. Food Chemistry, 90, 649– 654.


AOAC (1990). Official Methods of Analysis, 15th edn. Washington, DC: Association of Official Analytical Chemists. BAM (1998). Bacteriological Analytical Manual, 8th edn. Gaithersburg, MD: BAM. Cassens, R.G. (1994). Meat Preservation, Preventing, Loses and Assuring Safety. Pp. 79–84. Westport, CT: Food & Nutrition Press, Inc. Cheng, J.H. & Ockerman, H.W. (2003). Effect of phosphate with tumbling on lipid oxidation of precooked roast beef. Meat Science, 65, 1353–1359. Dogruer, Y. (1992). Farkli tuzlama su¨releri ve baskilama agirliklarinin pastirma kalitesine etkileri uzerine aras¸tırmalar. PhD Thesis, Selcuk Universitesi Saglik Bilimleri Enstitusu, Konya, Turkey (in Turkish). Dogruer, Y., Guner, A., Gurbuz, U. & Ucar, G. (2003). Sodyum ve potasyum nitratin uretim periyodu suresince pastirmanin kalitesine etkisi. Turkish Journal of Veterinary Animal Science, 27, 805–811 (in Turkish). Dzudie, T. & Okubanjo, A. (1999). Effects of rigor state and tumbling time on quality of goat hams. Journal of Food Engineering, 42, 103– 107. Gokalp, H.Y., Kaya, M. & Zorba, O. (1995). Et urunleri isleme muhendisligi. Pp. 309–341. Erzurum: Ataturk Universitesi (in Turkish). Gurbuz, U. (1994). Pastirma uretiminde degisik tuzlama tekniklerinin uygulanmasi ve kaliteye etkileri. PhD Thesis, Selcuk Universitesi Saglik Bilimleri Enstitusu, Konya (in Turkish). Gurbuz, U., Dogruer, Y., Yalcın, S., Nizamlioglu, M. & Guner, A. (2003). Pastirma yapim teknolojisinin gelistirilmesinde sicak dumanlama uygulanmasi ve kaliteye etkisi. Veteriner Bilimleri Dergisi, 19, 57–66 (in Turkish). Heikal, H., Dashlouty, M.S. & Saied, S.Z. (1972). The quality of pastirma as affected by autolysis of the camel meat. Agricultural Research Review, 50, 235–242. Hullberg, A. & Lundstro¨m, K. (2004). The effects of RN genotype and tumbling on processing yield in cured-smoked pork loins. Meat Science, 67, 409–419. Hullberg, A., Johansson, L. & Lundstro¨m, K. (2005). Effect of tumbling and RN genotype on sensory perception of cured-smoked pork loin. Meat Science, 69, 721–732. Hunt, M.C. & Kropf, D.H. (1984). Color and appearance. In: Advances in Meat Research, Vol. 3 (edited by A.M. Pearson & R.T. Dutson). Pp. 125–159. New York, NY: Van Nostrand Reinhold Company Inc. Isıklı, N.D. & Karababa, E. (2004). Rheological characterization of fenugreek paste (cemen). Journal of Food Engineering, 69, 185–190. Kotzekidou, P. & Lazarides, H.N. (1991). Microbial stability and survival of pathogens in an intermediate moisture meat product. Lebensmittel-Wissenschaft und Technologie, 24, 419–423. Pietrasik, Z. & Shand, P.J. (2003). The effect of quantity and timing of brine addition on water binding and textural characteristics of cooked beef rolls. Meat Science, 65, 771–778. Pietrasik, Z. & Shand, P.J. (2004). Effect of blade tenderization and tumbling time on the processing characteristics and tenderness of injected cooked roast beef. Meat Science, 66, 871–879.



Salama, N.A. & Khalafalla, G.M. (1987). Microbiological and chemical studies during badterma cured meat processing. Archiv fur Lebensmittel hygiene, 38, 33–68. Steel, R.G.D. & Torrie, J.H. (1981). Principles and Procedures of Statistics, 2nd edn. Tokyo: McGraw-Hill Int. Book Co.. Stone, H. & Sidel, J.L. (1985). Sensory Evaluations Practices. London: Academic Press Inc.

Tekinsen, O.C. & Dogruer, Y. (2000). Pastirma All About. Konya: Selcuk Universitesi Basimevi. Yetim, H., Ockerman, H.W. & Yousef, A.E. (1996). Bacteriological evaluations of tumbled catfish with egg white. Food Microbiology, 13, 365–368.



129


130

Original article Measurement of odour after in vitro or in vivo ingestion of raw or heated garlic, using electronic nose, gas chromatography and sensory analysis Kazuhiko Tamaki,1* Shigenori Sonoki,2 Takeshi Tamaki3 & Katsuo Ehara4 1 2 3 4

Department of Food Science & Technology, University of California Davis, Davis, CA 95616, USA Azabu University, Kanagawa 229-8501, Japan Tokyo Bunka Junior College, Tokyo 164-8638, Japan Tokyo Institute of Technology, Tokyo 152-8550, Japan (Received 15 June 2005; Accepted in revised form 14 June 2006)

Summary

The possibility of characterising the garlic odour in in vitro and in vivo was demonstrated using the newly developed electronic nose, based on an array of metal oxide semiconductor sensors. Two grams of raw and heat-treated garlic, and breath odour after eating 2 g of raw garlic and heat-treated garlic were analysed with an electronic nose. Furthermore, calculation of F-value (odour quality) and S-value (odour strength) demonstrated distinct odour differences between the samples, and that the electronic nose could differentiate between the various garlic associated odours corresponding to the different origins (in vivo or in vitro), or to the different processing (raw or heat-treated). The correlation between gas chromatography and sensory analysis was also discussed in order to identify the volatile compounds in the sample, and to investigate the association with the response of human perception to the samples. Results showed that odour sensor data were easier to obtain and were well correlated with both types of instrument.

Keywords

Electronic nose, gas chromatography and sensory analysis, in vitro and in vivo study, oral malodour, raw garlic and heated garlic, volatile sulphide component.

Introduction

Garlic (Allium sativum L.) is widely used in various food recipes for its unique flavour and nutrition. However, garlic is also the cause of malodourous smell after ingestion. Studies have shown that garlic’s various sulphur substances are the main cause of the malodour (Block, 1985; Laakso et al., 1989; McDowell & Kassebaum, 1993; Scully et al., 1997). The principal components of the malodour detected by the human nose are hydrogen sulphide, methyl sulphides, and dimethyl sulphide and allyl sulphide compounds. In particular, allyl sulphide compounds are detected in breath after garlic ingestion (Tonzetich, 1977). These mercaptan compounds have a strong disagreeable smell in the human nose but it is experientially known that heattreated garlic odour is not as strong as raw garlic odour and breath after eating heat-treated garlic is not as strong as after eating raw garlic. Studies have also shown that there is a difference in the amount of volatile compounds in heat-treated garlic and fresh garlic. In this study, we analysed the change of the odour

*Correspondent: E-mail: [email protected]

characteristic of garlic using both instrumental analysis and human evaluation. Recently, as the mechanism of the human sense of smell has been clarified, chemical sensor technology has been developed which can perceive odours in a similar way to a human nose. Depending on what types of elements are applied, many sensors have been developed: calorimetric sensors, metal oxide semiconductors (Lundstrom et al., 1992) and quartz sensors (Hauptman et al., 1993; Faccio et al., 1994). Many electronic noses are also manufactured by commercial companies and can be applied in a wide range of fields. These companies include UMA Airsense (Schwerin, Germany), Alpha M.O.S. (Tovlouse, France) and Cyrano Sciences Inc. (Danbury, CT, USA). Using the technology previously developed by Ehara (1989, 1993a,b) we started to develop a new odour sensor. We developed a semiconductor-based electronic nose which incorporated multiple metal oxide elements that can respond to different odour molecules with a wide range of sensitivity (Gardner & Bartlet, 1994; Di Natale et al., 1994). Metal oxide semiconductors have an advantage over other sensors in terms of the following properties: easily constructed and sensitive to a low vapour concentration, resistance to change in


Measurement of garlic odour K. Tamaki et al.

humidity and corrosive acid vapours (Shurmer et al., 1991). Metal oxide sensors also have a wide range of selectivity because of the abundance of electrons on the surface of the semiconductor. By combining the different semiconductor materials, the developed electronic nose can respond sensitively to a wide range of odour molecules (Ehara, 1989, 1993a,b). This sensor is composed of six different types of semiconductors, doped with catalytic metals, with different conductibility. Patterns of resistance change can be generated to create a specific profile for the volatile compounds in the sample. In comparison with existing electronic noses the materials in this new sensor have a higher sensitivity to a broad range of compounds, and so the odour sensor is able to record odour compounds at very low ppb concentrations, without the extra concentration step that has been previously required. This is advantageous in capturing the same volatile compounds emitted from the samples that are detected by the human nose. This sensor property allowed us to construct a portable sensor and measure the sample odour in a very convenient form, which closely reflected the ability of the human nose to catch volatile compounds, and had the same profile as other more expensive commercial sensors. Although several brief studies have been conducted using this new odour sensor to measure the organic vapours of several food products (Ehara, 1990; Murata et al., 1995), no studies have investigated how the different response patterns can be observed on testing the odour of garlic products or human breath. Thus, in comparison with other instrumental and psychophysical methods, proper measurement of the strength of the garlic causing odour must be investigated. In this study, first samples of raw garlic and heattreated garlic for in vivo and in vitro studies were evaluated using the discrimination method by human subjects. Next, volatiles released from the in vivo and in vitro samples were identified by GC and GC-MS with time-release analysis. Then, using the newly developed sensor, which consists of six arrays of metal conductors, the character and the strength of the garlic odour were analysed for the in vivo and in vitro samples. Correlation between the sensory evaluation, GC analysis and sensor analysis was assessed. The results showed that the sensor could successfully evaluate the odour character and strength, and the possibility of applying the odour sensor to the field of the food processing was indicated.

sensitivity by first distinguishing the odours of b-phenylethyl alcohol, methyl cyclopentenolone 30 ppm, isovaleric acid 10 ppm, c-undecalactone 30 ppm and skatole 10 ppm. All subjects signed a written consent form.

Materials and methods

GC analysis

Sensory evaluation Assessors

Thirty-five assessors (all Japanese females, age ranged from 19 to 25 years) took part in the experiment. Assessors were required to demonstrate normal odour

In vitro test

Two grams of raw garlic was grated, while another 2 g of garlic was heated for 1 min in a microwave oven and then grated. The grated garlic was transferred into a Petri dish, and then sealed with the plastic lid until the experiment began. A pair comparison test was chosen as the sensory method. This method was used as the most sensitive method to determine the difference between two samples for the specific product attributes. Before the experiment, each of the 35 assessors was asked to smell the odour of raw grated garlic, as a warm up, and told to use this smell as a standard. The assessors were blindfolded to prevent them seeing the appearance of the samples presented on a tray, and then asked to choose the sample with the stronger garlic odour. The sample order was randomised and counterbalanced across all the assessors. After the experiment, they were asked to comment on any special attribute they noticed when they chose their sample. In vivo test

One gram of raw garlic was grated, while another 2 g of garlic was heated for 1 min in a microwave oven and then grated. One gram of grated garlic from each sample was transferred into a plastic glass, and then the garlic was ingested by chewing for 30 s. One female subject (aged 18 years) ingested a whole raw garlic, and the next day the same subject ingested a whole heated garlic. Breath samples were collected in a commercial plastic bag which was manufactured by OMI Odoair Service Co. Ltd (Tokyo, Japan). The subject held her breath for 10 s, and then exhaled in to a bag sealed tightly around her mouth. As described in the procedure for the in vivo test, 35 assessors were then presented with pairs of the plastic bags, one with raw garlic rinsed breath and another with heated garlic rinsed breath. The assessors were again blindfolded and asked to choose the stronger odour between two bags. Sample order was randomised and counterbalanced across assessors. After the experiment, they were asked to comment on any special attributes they noticed when they chose the sample.

Analytical conditions

The identities of the sulphur-containing gases for in vivo and in vitro samples were established by using GC and GC-MS spectrometry. As a standard reagent for GC analysis, the following chemicals were used: methanethiol, dimethyl sulphide and dimethyl disulphide (Wako



131

132


Pure Chemical Industries, Osaka, Japan), allylthiol and methyl propyl disulphide (Aldrich Chemical Co. Inc., St. Louis, MO, USA), allyl methyl sulphide and diallyl disulphide (Tokyo Kasei Organic Chemicals, Tokyo, Japan). The gas chromatograph (Shimadzu GC, 14B; Shimadzu, Kyoto, Japan) conditions were as follows: polyphenyl ether (PPE) 5 ring 10%, 3.2 mm · 3.1 m glass column, temperature programmed from 65 C for 3 min to 170 C at 30 C min)1, the detector was a flame photometric detector (FPD) (140 C) and the carrier gas was N2 (55 mL min)1). The GC-MS conditions were as follows: the GC-MS (Shimadzu GC-MS QP 1000) had a fused-silica capillary column (Supelco SPB-1), 0.32 mm · 30 m glass, temperature programmed from 60 to 170 C at 30 C min)1 and held at 170 C for 2 min, the carrier gas was He (2 mL min)1), the electron ionization (EI) mode was at 70 eV, the source temperature was 180 C and the GCMS interface temperature was 250 C. In vitro test

Three grams respectively of raw and heated-garlic were crushed by applying pressure with a spoon. The heated garlic was prepared by the same process as described above. Samples (0.2 g) were transferred into a 125-mL vial which was sealed tightly, then kept in a container temperature maintained at 23 C and then 1 mL of head-space gas was removed from the vial for analysis at 0, 30, 60, 90 and 120 min.

Table 1 Material and properties of six sensors in the electronic nose Sensor

Material

Selectivity

CH-1 CH-2 CH-3 CH-4 CH-5 CH-6

SnO2, sintered metal oxide SnO2, thin-film metal oxide SnO2, thin-film metal oxide SnO2, thin-film metal oxide SnO2, sintered-metal oxide ZnO, thin-film metal oxide

Nitrogen Nitrogen Sulphur Hydrocarbon Alcohol, aromatic Sulphur

In vivo test

One healthy subject (aged 19 years) with no malodour took part in this study. The subject had not ingested any garlic for 24 h before the study. The basic study consisted of two treatment periods; on one day the subject chewed and then swallowed 1 g of raw garlic and measurements were carried out over the next 2 h. The next day the subject’s breath was measured over a 2-h period after ingesting 1 g of heated garlic. The subject’s breath was collected in syringes following the method described by Aoki (1970). Samples were collected immediately before garlic consumption and then after 0, 30, 60, 90 and 120 min. Electronic nose analysis Odour sensor preparation

Two different types of elements (sintered metal oxide and thin-film oxide) were combined. The former sensor element reacts sensitively to odourant compounds with low molecular weights, perceived as light odourants, while the latter sensor type is more sensitive to odour compounds with relatively large molecular weights, such as an unsaturated aromatic hydrocarbon group compound like toluene or methanethiol, perceived as heavier odourants. The production of the sensor and its properties has been previously described (Ehara, 1993a,b). A total of six semiconductor sensors were combined in this electronic nose to cover a broad range of chemical properties. Each sensor’s material name and properties are shown in Table 1. The sensing principles of the metal oxide semiconductor are not still clear but it can be explained as follows: metal oxide is normally produced in a slightly reduced state, but oxygen in the air is adsorbed on the surface and forms an electron depletion layer due to a high electron affinity of oxygen. This increases the resistance of the semiconductor. With the existence of reducing gas, an oxidation reaction proceeds against the adsorbed oxygen and electrons captured by oxygen are released. As a result, the resistance of the semiconductor decreases and voltage in the circuit drops. The reaction details also change depending on

Figure 1 Simple diagram of the electric circuit; the sensor element consists of a parallel resistance circuit of the semiconductor and heater coil. Combined resistance (R) is converted into a voltage in the bridge circuit, which is connected to the counter.




Figure 2 Schematic diagram of the electronic nose structure.

the condition of the coil heating temperature and sensing materials. Figure 1 shows the structure of the sensor circuit. The coil and the semiconductor compose a parallel resistance circuit (Rs and Rh). Rs is the sensing element whose resistance varies according to injected gas composition and heater temperature, which is controlled by the bridge voltage that is imposed across the sensing element. Changes in combined resistance are measured when an odour gas is adsorbed onto the surface of the sensing element, which causes electron conductivity changes to decrease the Rs. Changes in R can be measured as a change in voltage as a sensing element is incorporated into the bridge circuit. A simple diagram of the electronic nose design is shown in Fig. 2. The sensor was mounted inside a 30mL gas chamber, in which a sample plate was mounted to hold known amounts of the tested compounds. The chamber has an open window, which can be sealed tightly after a sample was placed on the plate. Output from the bridge circuit was connected to a data collector to measure changes in resistance. The data collector was linked to an external computer which controlled data management and analysis. Changes in voltage transmitted through the data collector were displayed on a CRT display at regular time intervals and the data was processed and spider graphs were generated to display the sensor response at any given time.

released inside the chamber, the odour detector started collecting the odour signal and the output value was plotted against time. When the output value reached saturation level after 1 min, a spider graph was also generated to see the instant response pattern of the six sensors. The data collection process lasted about 5 min. In vitro test

Before measurement of the garlic, the odour of multiple food products was investigated as a control to determine the sensor’s response to real food products with different odours. Lemon, kusaya and natto were purchased from a commercial market to represent pleasant and unpleasant odours. Fifty grams of each product was placed in turn on the glass plate for analysis, with the lemon first having been sliced. Then the next day the garlic sample measurements were conducted, using the same as in the sensory experiment. In vivo study

The subject’s odour was measured using the same sample obtained for the sensory experiment. Results and discussion

Sensory evaluation of the garlic odour In vitro and In vivo tests

Data collection procedure

Data collection for each sample was conducted as follows. The platinum coil wire upon which sinteredoxide and tin-film oxide metal was mounted was heated up to 300 C. After the odour compounds were

A significant majority of the judges indicated that raw garlic had a stronger odour (Binomial test, 33/35, P < 0.001) and also that it produced a stronger breath after garlic ingestion (Binomial test, 27/35, P < 0.001). These results indicated that raw garlic odour is stronger



133

134


than the heated-garlic, for both in vitro and in vivo samples. Interestingly, the result of the in vivo study showed a less significant difference than for the in vitro study. The assessors commented that the garlic odour of the in vivo study was clearly different from that of the in vitro study and also very unpleasant. They also commented that the odour of heat-treated garlic and the breath rinsed with heated-garlic became weaker. Considering these results, it can be concluded that judges perceived clear differences in the odours of these four samples from the in vitro and in vivo experiments. GC analysis of garlic In vitro test

The identities of the sulphur-containing gases arising from grated raw garlic and heated garlic were analysed by GC and GC-MS (Fig. 3). Eight peaks were found and these were identified as methanethiol, dimethyl sulphide, allylthiol, allyl methyl sulphide, dimethyl disulphide, methyl propyl sulphide, diallyl disulphide and 3-(allylthio) propionic acid. The amount of each compound was calculated at different time intervals, except for allyl methyl sulphide and 3-(allylthio) propionic acid (Table 2). Methanethiol, dimethyl sulphide and allylthiol are low-molecular sulphur compounds (LMSC) and were present in a slightly larger quantity than allyl methyl sulphide and dimethyl disulphide, which are relatively high-molecular sulphur compounds (HMSC) in both raw garlic and heated garlic. Although there was little difference in the amount of LMSC between the raw and the heated garlic, there was a clear difference in HMSC. Methyl propyl disulphide and diallyl disulphide, which is the main decomposition product from allicin, were found in both the raw and the

heated garlic, but the concentrations were higher in the raw garlic. The reason may be because the formation of these sulphur compounds would be stopped by thermal inactivation of the enzyme alliinase. This assumption was supported by the fact that these compounds could not be detected when the heating time was prolonged to enhance thermal inactivation (data not shown). GC data for the change in the amount of sulphur compounds followed the typical enzyme alliinase reaction passage. In raw garlic alliinase is activated by grating or cutting of garlic. Allylcystein sulphoxide or alliin is rapidly converted into allicin by its enzymatic reaction. Allicin is further converted into higher molecular disulphide compounds, which are attributed to the unique pungent odour of garlic. In the GC data, allylthiol was detected at 0 min but soon began to decrease. This shows that grating induced a rapid conversion of allicin to allylthiol but the conversion decreased significantly as the reaction progressed. The amount of the other LMSCs showed little change, although there was a slight increase in dimethyl sulphide after 60 min. For HMSCs, diallyl disulphide increased continuously to reach a maximum after 60 min. This fact confirmed that when raw garlic was grated the change of alliin into allicin occurred immediately, followed by conversion into HMSCs as the enzymatic reaction proceeded. The heated garlic contained higher concentrations of LMSC such as methanethiol, dimethyl sulphide and allylthiol than the raw garlic. The amounts of methanethiol and allylthiol from the heated garlic increased over time. Also, the rate of increase of the thiol compound was considerably greater for heated garlic than for raw garlic. This fact suggests that production of allylthiol can be accelerated more by physical conditions such as heating or the accompanying reactions, rather than by the enzymatic reaction itself. For HMSCs, the

Figure 3 Gas chromatograms of the headspace vapours sampled just after grating raw garlic (left) and heat-treated garlic (right). Peak components were as follows: 1, methanethiol; 2, dimethyl sulphide; 3, allylthiol; 4, allyl methyl sulphide; 5, dimethyl disulphide; 6, allyl methyl disulphide; 7, methyl propyl disulphide; 8, diallyl disulphide; 9, 3-(allylthio) propionic acid.




Table 2 Time course of volatile sulphide component formation in raw and heat-treated garlic (in the in vitro test)

Time (min) Compound

0

30

Raw garlic 0.0280 0.0280 0.0200 0.0190 0.0467 0.0359 0.0040 0.0030 0.0091 0.0112 0.0845 0.1140 0.6390 1.1900 Heat-treated garlic 0.0420 0.0755 0.0451 0.0800 0.0597 0.1311 0.0080 0.0064 0.0079 0.0070 tr. tr. 0.0758 0.1290

Methanethiol Dimethyl sulphide Allylthiol Allyl methyl sulphide Dimethyl disulphide Methyl propyl disulphide Diallyl disulphide Methanethiol Dimethyl sulphide Allylthiol Allyl methyl sulphide Dimethyl disulphide Methyl propyl disulphide Diallyl disulphide

60

90

120

0.0270 0.0174 0.0281 0.0040 0.0118 0.1120 1.3800

0.0270 0.0275 0.0232 0.0056 0.0122 0.0930 1.1750

0.0260 0.0311 0.0209 0.0048 0.0126 0.0830 0.9200

0.0862 0.0767 0.1565 0.0064 0.0071 tr. 0.1060

0.0957 0.0764 0.1667 0.0064 0.0066 tr. 0.0920

0.1068 0.0737 0.1764 0.0064 0.0050 tr. 0.0658

These volatile compounds were characterised and quantified by GC-MS under the analytical conditions described in the text. Values are expressed in ppm; tr. represents ‘trace’.

amount of allyl methyl sulphide and dimethyl disulfide increased for 30 min and then decreased gradually, although methyl propyl disulfide was not detected significantly. In vivo test

Gas chromatograms of raw and heated garlic in the in vivo test are shown in Fig. 4 and Table 3 shows the changes in the amounts of the volatile sulphide components over time in the in vivo test. Immediately after ingesting the raw garlic, the concentrations of LMSCs were greatest especially methyl sulphide and allyl

sulphide, but they soon started decreasing. In contrast in the heat-treated garlic eating, lower concentrations of LMSCs (methanethiol, allylthiol and allyl methyl sulphide) were observed with little detected after 30 min. Therefore these results indicate that the mouth normally contains a small concentration of methanethiol and dimethyl sulphide (Fig. 4) but immediately after garlic eating, larger concentrations of methanethiol and allylthiol and smaller concentrations of allyl methyl sulphide, allyl methyl disulphide, and diallyl disulphide are produced. These higher concentrations of LMSCs, especially allyl compounds, were uniquely observed in

Figure 4 Gas chromatograms of the breath odour of one subject (a) before, (b) after eating 1 g of grated raw garlic, and (c) after eating 1 g of grated heat-treated garlic. Peak components are described in Fig. 3.



135

136


Time (min) Compound

Methanethiol Dimethyl sulphide Allylthiol Allyl methyl sulphide Dimethyl disulphide Methyl propyl disulphide Diallyl disulphide Methanethiol Dimethyl sulphide Allylthiol Allyl methyl sulphide Dimethyl disulphide Methyl propyl disulphide Diallyl disulphide

0

30

Raw garlic 2.3569 0.3382 n.d. n.d. 2.3000 0.4488 0.0589 0.0406 n.d. n.d. n.d. n.d. 0.1014 n.d. Heat-treated garlic 0.0499 0.0083 n.d. n.d. 0.0532 0.0024 0.0289 0.0058 n.d. n.d. n.d. n.d. n.d. n.d.

60

90

120

0.1400 n.d. 0.1788 0.0470 n.d. n.d. n.d.

0.1232 n.d. 0.0909 0.0339 n.d. n.d. n.d.

0.0598 n.d. 0.0455 0.0317 n.d. n.d. n.d.

0.0080 n.d. 0.0024 0.0058 n.d. n.d. n.d.

0.0075 n.d. n.d. n.d. n.d. n.d. n.d.

n.d. n.d. n.d. n.d. n.d. n.d. n.d.

Table 3 Time course of volatile sulphide component formation in raw and heat-treated garlic (in vivo-test)

These volatile compounds were characterised and quantified by GC-MS under the analytical conditions described in the text. Values are expressed in ppm; n,d, represents ‘not detected’.

the breath after garlic eating, which is different from pathological breath (Yasuno et al., 1989). Thus, the malodourous smell after eating garlic is thought to be attributed to LMSCs. The data also showed that the LMSC concentration in the breath was greater after ingesting the raw garlic solution than the heat-treated garlic. This supports the results of the sensory analysis where the assessors indicated a difference in odour quality difference for the breath after eating heat-treated garlic compared with raw garlic. GC results also showed that no HMSCs were detected after either raw or heattreated garlic ingestion, except for diallyl disulphide which was found immediately after eating raw garlic. The small amounts of HMSCs suggest that the alliinase enzyme reaction and the following reaction could be affected by the biochemical conditions in the mouth and that HMSCs would not be produced. Electronic nose analysis

spear-shaped radar graph, such as that shown for lemon (Fig. 5). On the other hand, for unpleasant smelling substances the radar graph shape was very different. For instance, natto1 and kusaya2 are known for their unpleasant smells, in comparison with the fruity citrus smell of lemon. In contrast to the display for lemon, the responses for natto and kusaya had a similar irregular shape due to the small response of CH6 and CH3 (Fig. 5). Other examples of pleasant spear-shaped response included coffee and lavender, and unpleasant odours such as trimethylamine (TMA) and blue cheese had irregular shapes (data not shown). These results indicate that different odours can be identified by the responses of the six sensors in terms of the strength and the quality. The strength of the odour can be defined by the S-value, which is defined as an integration of the area under the spider graph. The quality of the odour can be defined by the F-valve, which is defined as (CH3+CH6) ⁄ (CH2+CH4).

Pleasant and unpleasant odour measurement

Garlic odour measurement

Resistance values from the six sensors were fed into a processor and analysed by computer program to produce responses displayed in spider graphs. Several hundred kinds of food were tested and the sensor responses observed. In general, certain regular response patterns were found: responses from sensors 2 and 4 were small, and those from sensors 3 and 6 were greatly enlarged for pleasant smells. This produced a typical

Garlic odour is different from lemon in terms of the quality of the odour. The garlic odour in the in vitro and in vivo studies was measured with the newly developed sensors.

1

When raw garlic is cut or grated, it produces specific pungent odour. However, when heat-treated garlic is cut 2

Natto is a popular food in Japan made by fermenting cooked soybeans with Bacillus subtilis (natto), and has a characteristic aroma and stickiness. It contains many nutrients originating from both soybeans as well as from intact cells and metabolites of B. subtilis.


In vitro test

Kusaya is a specially brined and dried fish produced in the Izu Islands south of Tokyo. It is famous for its malodour and it is often the subject of taste controversies, much like British marmite and French blue cheese. It is popular among the Japanese for its unique flavour and because it can be preserved for a long time.



Kusaya: F = 1.068, S = 18016

Natto: F = 0.898, S = 4913.77

Lemon: F = 4.0282, S = 6736

CH1

CH6

126.6

CH2 55.5

31.3

CH1

CH1

41.7

CH6

CH6

CH2 86.3

54.9

78

CH2

39.9 22.6

26.9 44.8 CH5

90.3 CH5

51.2CH3

55.9 CH4

55.7

34.1

96.8 CH3

150.4 CH3

CH5

CH4

CH4

Figure 5 Spider graphs of 1 g samples of lemon, natto and kusaya showing the resistance values (mV) from the six sensors outputs (CH1–CH6) in the electronic nose; the F-values and S-values are shown.

or grated, it has a very weakly pungent odour which is different from that of raw garlic. In the in vitro test the S-value of the raw garlic (S ¼ 39 995) was much greater than for the heat-treated garlic (S ¼ 16 889) (Fig. 6), although the graph does not have a spear shape. The Fvalue of the raw garlic was 1.955 compared with 4.302 for the heat-treated garlic. These results are comparable with the results of the sensory analysis showing that heating moderated the garlic’s pungent odour. In vivo test

with after ingesting heat-treated garlic. Figure 7 shows the results of the in vivo experiment, after ingesting garlic. The intensity of odour after raw garlic was ingested was stronger (S ¼ 35 819) than for the heattreated garlic (S ¼ 15 230). Both values were much greater than that of the breath odour before any garlic was ingested (S ¼ 1463). The F-value for the heattreated garlic (0.9805) was greater than for the raw garlic (0.8640), indicating that the heat-treated garlic had a better quality odour. However, the shape of the spider graph results for the raw garlic and heat-treated

Typically people’s breath after ingesting raw garlic is much more malodourous and unpleasant compared

Water: F = 1.056, S = 1463 Raw garlic: F = 0.8640, S = 35819 Heat-treated garlic: F = 0.9805, S = 15230

Raw garlic: F = 1.955, S = 39995 Heat-treated garlic: F = 4.302, S = 16889

CH1

CH1

CH6

CH2

CH5

CH3

CH6 CH2

212.6 26.7 22.1

229.4

95.7 71.6

40 74.4

30.3 48.2

CH3 225.8 247.5

CH5

CH4 CH4 Figure 6 Spider graphs showing the resistance values (mV) from the six

sensors outputs (CH1–CH6) of the electronic nose analysis of the in vitro study samples of 0.2 g mL)1 grated raw garlic (—–) and heattreated garlic solution (– – –); the F-values and S-values are also shown.

Figure 7 Spider graphs showing resistance values (mV) from the six sensors outputs (CH1–CH6) of the electronic nose analysis of the in vivo study samples of a subject’s breath after rinsing with 1 mL of pure water (- - - -), 1 g mL)1 grated raw garlic solution (—–) and grated heattreated garlic solution (– – –); the F-values and S-values are also shown. The values of six sensors outputs are shown in the table.



137

138


garlic in the in vivo study was fairly similar, compared with the different shapes produced in the in vitro study. The difference between the S-values for raw garlic eating and heat-treated garlic eating in the in vivo test (S ¼ 20 589) was smaller than in the in vitro test (S ¼ 23 106). Also, the difference between F-values was smaller in the in vivo test (F ¼ 0.1165) than in the in vitro test (F ¼ 2.347). This indicates biochemical interactions in the in vivo test reduced the difference in garlic odour between the raw garlic and the heated garlic. In the sensory test the assessors mentioned that mouth odour after eating garlic was fairly unpleasant but breath odour after eating heat-treated garlic was not so unpleasant. These comments agreed with the results of the sensor analysis. Conclusions

Food flavour is a mixture of many different types of chemical compounds and these compounds are correlated to create a unique sensation. For instance, ethyl alcohol odour is not very pleasant, but people enjoy its smell in alcoholic drinks, mixed with the odour of the other substances created during fermentation. Thus, it is challenging to develop the method to evaluate the odour quality conveniently. The newly developed electronic nose consists of a semiconductor of combined sintered and thin film metal oxides. The former sensor type reacts sensitively to light odourants (compounds with low molecular weights), while the latter sensor type is more sensitive to heavier odourants (compounds with relatively large molecular weights) such as unsaturated aromatic hydrocarbon group compounds like toluene or methanethiol. In this study, we developed a sensor analysis method to measure the changes in odour quality caused by different treatments of garlic. By using the electronic nose, we verified that an electronic nose could numerically define the characteristics and strength of the odour in terms of F-value and S-value. The electronic nose’s F and S values showed that garlic odour characteristics and strength were different in in vitro and in vivo. For the in vitro study, the F-value was lower and S-value was higher for raw garlic than for heated garlic and this result matched the GC and sensory analysis. In the in vivo study, S-value and F-values of both breath samples were lower than those of the in vitro study. This suggests that raw garlic has a strong odour and its odour changes into unpleasant odour in in vivo. On the other hand, garlic odour became moderate with heating and so when heated garlic was ingested, it did not produce a strong unpleasant odour. Again this was supported by sensory and GC analyses. The sensory results showed that the strength of


garlic odour was stronger for raw garlic than for heattreated garlic for both the in vivo and in vitro studies. The GC analysis showed higher values of HMSC in raw garlic immediately after grating, and little difference in LMSC between raw and heated garlic. HMSCs such as allyl methyl disulphide, methyl propyl disulphide and diallyl disulphide largely contributed to the specific odour and sharp taste of the garlic, thus the GC result confirmed the sensory result. On the other hand, in the in vivo test, LMSC values were higher after eating raw garlic, but HMSCs were hardly detectable for either raw garlic or heat-treated garlic samples. When garlic is ingested, various LMSC increase and these LMSCs are considered to be the major cause of malodourous breath after garlic ingestion, which was detected in the sensory analysis. This study proves that the developed electronic nose has a feature for evaluating product odour intensity and quality in terms of pleasantness. In the rapid process of food product manufacturing, there are many attributes to be monitored, but the most important criterion is to keep the product quality favourable for consumers. Usually both sensory and GC analysis have been combined to measure the quality of odours. Thus, this new odour sensor analysis has the advantage of being a convenient method. The S- and F-values are useful concepts, and as the formula is derived experimentally it is clearly applicable to odours of many kinds of food products. Also, other than pleasantness and unpleasantness, it could also be possible to express odour characteristics by multiple criteria. Thus, in future experiments, the odour of the various food products should be investigated to find how sensor output and the results of the analysis of odour compounds are related. Acknowledgments

The authors wish to thank Mr Shiozawa Hiroaki, New Cosmos Electric Co. Ltd and The Bank of Okinawa, Ltd for technical contribution and financial support. References Aoki, H. (1970). Analysis of organic volatiles in oral cavity by gas chromatography. Journal of the Japanese Organization for Research of Periodontology, 11, 3–19. Block, E. (1985). The chemistry of garlic and onions. Scientific American, 252, 114–119. Di Natale, C., d’Amico, A., Davide, F., Faglia, G., Nelli, P. & Sberveglieri, G. (1994). Performance evaluation of a SnO2-based sensor array for the quantitative measurement of mixtures of H2S and NO2. Sensors and Actuators, B Chemical, 20, 217–224. Ehara, K. (1989). Odor sensor. Sensor Technology, 9, 59–63. Ehara, K. (1990). Odor measuring method by metal oxide semiconductor. Odor Sensor, PPM The Nippon Kogyo Shinbun, Pp. 22–29. Ehara, K. (1993a). Development of odor sensor apparatus corresponding to sensory evaluation. Function and Materials, 13, 14–19.



Ehara, K. (1993b). Semiconductor sensor odor detection. Ouyou Butsuri, 62, 261–264. Faccio, M., Ferri, G., Mancini, F. & Di Rosa, P. (1994). Resonating quartz sensors. In: Proceedings of the First European School on Sensors-ESS ‘94. Pp. 71–86. Gardner, J.W. & Bartlet, P.N. (1994). A brief history of electronic noses. Sensors and Actuators, B Chemical, 18–19, 211–219. Hauptman, P., Lucklum, R., Hartmann, J., Auge, J. & Adler, B. (1993). Using the quartz microbalance principle for sensing mass changes and damping properties. Sensors and Actuators, A Physical, 37–38, 309–316. Laakso, I., Seppanen-Laakso, T., Hiltunen, R., Muller, B., Jansen, H. & Knobloch, K. (1989). Volatile garlic odor components: gas phases and adsorbed exhaled air analyzed by headspace gas chromatography-mass spectrometry. Planta Medica, 55, 257–261. Lundstrom, I., Hedborg, E., Spetz, A. et al. (1992). Sensors and Sensory Systems for an Electronic Nose. Pp. 303–319. Nato ASI Series, E 212. Boston: Kluwer Academic Publishers.

McDowell, J.D. & Kassebaum, D. K. (1993). Diagnosing and treating halitosis. Journal of the American Dental Association, 124, 55–64. Murata, H., Oomori, M., Hasegawa, A. & Ehara, K. (1995). Application of the semiconductor sensors for the measurement of oral malodor II. Bull Nippon Dent Univ, Gen ed, 24. Scully, C., el-Maaytah, M., Porter, S.R. & Greenman, J. (1997). Breath odor; etiopathogenesis, assessment and management. European Journal of Oral Sciences, 105, 287–293. Shurmer, H.V., Corcoran, P. & Gardner, J.W. (1991). Integrated tin oxide odor sensor. Sensors and Actuators, B Chemical, 4, 117–121. Tonzetich, J. (1977). Production and origin of oral malodor: a review of mechanisms and methods of analysis. Journal of Periodontology, 48, 13–20. Yasuno, Y., Iwakura, M. & Shimada, Y. (1989). Relation between volatile sulfur compounds in mouth air and some symptoms in patients complaining of bad breath. Journal of Dental Health, 39, 663–674.



139


140

Original article A fuzzy comprehensive evaluation for selecting yeast for cider making Bangzhu Peng, Tianli Yue* & Yahong Yuan College of Food Science and Engineering, Northwest A and F University, 712100 Yangling, Shaanxi, China (Received 11 August 2005; Accepted in revised form 2 June 2006)

Summary

In this article, we present a new approach in solving the problem of selecting proper and excellent cider yeast. Eight different yeast strains commonly used in the cider industry were selected for this experiment. Different ciders were fermented by the yeasts. Sensory evaluation was conducted first. Then, aroma components were identified by the headspace-solid phase microextraction–gas chromatography–mass spectrometry. Their contents were determined using quantitative standards with known amounts of target components and 2-octanol as internal standards. A model of fuzzy comprehensive evaluation was established and used to assess and rank the cider-making attributes of all eight yeast strains. Saccharomyces cerevisiae F14 was selected out of the eight yeasts for the best cider. The results suggest that the method described in this article is more accurate and objective than conventional methods.

Keywords

Aroma compounds, cider, fuzzy comprehensive evaluation, sensory evaluation, yeast.

Introduction

Yeasts play an important role in the fermentation industry, especially in the making of fruit wine. The performance of yeast has a great effect on the quality of the fermented product. Some metabolic differences exist among Saccharomyces cerevisiae strains (Cavazza et al., 1989; Ubeda et al., 1998), and for this reason, there is a large range of commercially available yeasts that can be used in the manufacture of different types of cider. Yeast can convert the sugar in apple juice into alcohol. Alcoholic fermentation is the single most important aspect of cider making. It is at this stage that cider starts developing its flavour, aroma, body and structure (Barnet et al., 1990). The end result of brewing may be a cider that does not reflect a desired style or that has off-flavours if the yeast is not appropriate. Therefore, cider makers and cider researchers must take care in selecting the yeast type. The selection of proper yeast for the production of different types of cider is an important problem facing the cider industry. For years, the method of sensory evaluation has been very popular with researchers as well as cider makers (Heymann & Noble, 1987; Arrhenius et al., 1996; McCloskey et al., 1996; De La Presa Owens & Noble, 1997; Fischer et al., 1999; Vannier et al., 1999). *Correspondent: Fax: +86-29-87091717; e-mail: [email protected]

This method was shown to be flawed, because judges had different perceptions about the intensity of aroma components, their overlapping effect and duration. Furthermore, their knowledge and skills differed as well. The ability to critically evaluate a fruit wine’s sensory properties, make a judgement about the quality and convey these impressions in a clear, unambiguous way are all essential skills for judges. But these skills are not always consistent between members of a judging panel. Sensory evaluation relies heavily on human perceptions, and it is the human brain that combines all of a food’s sensory aspects into a specific psychological perception (Stone & Sidell, 1993). Many factors influence the complex process by which the brain decides whether or not it likes a certain taste. Therefore, the same sample does not always tastes the same to different judges. Most of the commonly used sensory evaluation methods for cider have been shown to be flawed on selecting cider yeast. We must develop other analytical methods and sensory evaluation that could lead to a new approach in selecting proper yeast for cider. Chemical analysis is an effective method for analysing the chemical components in the aroma of fruit wine. Chemical analysis indicates a correlation between the flavour quality and the intensity of chemical components (Noble & Sahnnon, 1987). Higher quality fruit wines (evaluated by sensory evaluation to be more intense) also had higher flavour intensity as determined by chemical analysis (Vannier et al., 1999). In this article, we combined chemical


Fuzzy comprehensive evaluation for selecting yeast B. Peng et al.

analysis with sensory evaluation to create a new method of evaluating the quality of ciders based on fuzzy comprehensive evaluation (FCE). In contrast to a single sensory evaluation, the method takes advantage of more objective information provided by the technique of quantitative chemical analysis. Thus, the evaluation result is more objective and precise than the conventionally used single sensory evaluation. Materials and methods

fermentation, then sensory evaluation was performed by a group of ten expert panellists. All panellists were experienced in cider sensory analysis. Ciders were presented in random order at 10 C in coded standard wine-tasting glasses according to ISO standard 3591 (ISO, 1977) and covered with a watch glass to minimise volatile compounds from escaping. It is a 100-point method that requires detailed evaluation of each cider in a fully equipped tasting room according to ISO standard 8589 (ISO, 1988). The panellists gave scores for appearance 0–20, aroma 0–30, taste 0–40 and style 0–10.

Apple juice and yeast strains

Apple juice was extracted from fresh fuji apples using a vacuum pressing and filtration system and stored for 24 h at 4 C prior to fermentation. The eight yeast strains are showed in Table 1.

Aroma components analysis

There were three replicates of each fermentation-yeast strain, and fermentations were all carried out in twentyfour spherical flasks with flat bottoms with 1500 mL of adjusted apple juice (the pH modulated by malic acid was 3.4 and the sugar content modulated by sucrose was 20 Bx before fermentation. In addition, the apple juice was sulphited to a final SO2 concentration of 60 ppm). After the different yeasts were incubated on a shaker for 48 h in apple juice medium (AJM; with the pH of 3.4, the sugar content of 15 Bx, the CaCl2 concentration of 0.01 m L)1 and the MgCl2 concentration of 0.01 m L)1) at 28 C, the AJM was added into adjusted apple juice to ferment by 8% (v/v). Fermentations were incubated at 22 C for 8 days, and completed fermentations were clarified by centrifugation (4000 r.p.m. for 10 min) and supernatants were stored at 20 C in the dark for 30 days, then sensory evaluation and analysis of aroma components were performed.

The aroma components were collected using 100 lm PDMS fibre (Supelco Co., Bellefonte, PA, USA), headspace sampling, for 30 min absorption at 30 C (Mangas et al., 1996). Then, analysis of the major aroma components was performed using a Thermo Finnigan Trace gas chromatograph (GC) coupled to a Trace DSQ mass spectrometer (MS) with helium (1.0 mL min)1) as the carrier gas. The fibre was directly desorbed in 2 min and the injector temperature was 200 C (Mangas et al., 1996). The GC column was a 30 m · 0.25 mm i.d. (0.25 lm film thickness) HP-5MS capillary column. The GC oven program was as follows: isothermal at 110 C for 2 min, 8 C min)1 to 200 C for 5 min (Escudero et al., 2000). The GC to MS transfer line was kept at 230 C. The MS was operated in the negative chemical ionisation mode using methane (2 mL min)1) as the reagent gas (Mamede et al., 2005). Recorded spectra were compared with NIST (National Institute for Standardisation, USA) and Wiley (6th edn –220.000 spectra) mass spectral libraries. Compound identification was confirmed by their retention time indices. The response factors for major aroma components were determined using quantitative standards with known amounts of target components and 2-octanol as internal standards.

Sensory evaluation

Method for selecting yeast for cider making

The young dry cider samples were sealed and stabilised at 20 C in the dark for 3 months after completion of

The results of the chemical analysis and sensory evaluation were processed by FCE to select excellent yeast based on cider quality.

Fermentations and stabilisation

Table 1 List of the experimental strains Strains

Source

Results

Saccharomyces cerevisiae 1750 S. cerevisiae F4 S. cerevisiae F6 S. cerevisiae F8 S. cerevisiae F11 S. cerevisiae F14 S. cerevisiae F15 S. cerevisiae F16

China Center of Industrial Culture Collection Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF Laboratory of Bioreactor in NWSUAF

Results of sensory evaluation and aroma component quantitative analysis

The results of sensory evaluation on different ciders are presented in Table 2. The total score is 100 points, and the score for each cider is equal to the average value given by ten experts. According to previous research work related to cider aroma (Stone & Sidell, 1993; Mangas et al., 1996; Vidrih & Hribar, 1999), we



141

142


Table 2 Sensory quality analysis of different ciders Samples a

Sensory scores a

F4

F6

F8

F11

F14

F15

F16

1750

79.5

81.5

87

89

90.5

88

81

83

Mean values.

summed up the principal aroma components of cider composed of eight kinds of higher alcohols (no. 1–8, Table 3) and six kinds of esters (no. 9–13, Table 3). Identification of those components was made using GC– MS. Results from the quantitative analysis of these aroma components from different ciders are listed in Table 3. Results of fuzzy comprehensive evaluation Establishing the evaluation sets

The evaluation sets are established on the basis of the different ciders fermented with the selection of yeast strains U ¼fu1 ; u2 ; . . . ; un g ¼ fF4; F6; F8; F11; F14; F15; F16; 1750g Establishing the factor sets

Sensory evaluation scores indicated the subjective and intuitive evaluation of the panellists on the cider. The sensory evaluation equates the applied language variation and level (agreement) variation to achieve a common discursive extension (Gawel, 1997). Therefore, the sensory evaluation score is one main factor (x1). The volatile components, especially the esters and higher alcohols produced during alcoholic fermentation determine the quality of each wine (Valero et al.,

2002). On the basis of the varieties of aroma components in cider and other researchers’ experiences correlative with cider (Mangas et al., 1996; Vidrih & Hribar, 1999), we proposed that total higher alcohols (x2) and total esters (x3) are two additional factors. Thus, the factor sets are X ¼ {x1,x2,…,xn} ¼ {x1,x2,x3}. Establishing the weight sets

The main steps for establishing the weight sets in this research can be described as follows (Williams & Webb, 1994): 1 Selection of one or more panels to participate in the investigation. The panellists are experts in the area of cider. 2 Development of the Delphi questionnaire. 3 Transmission of the questionnaires to the panellists 4 Analysis of the responses. The experts were asked to provide their estimates on the weights of sensory evaluation score, esters and higher alcohols, then we established the weight sets by applying fuzzy Delphi method to integrate the experts’ opinions. Table 4 presents three factors and their weight values, respectively. That is A ¼ {a1,a2,…,an} ¼ {a1,a2,a3} ¼ n P {0.35,0.30,0.35} satisfying, ai ¼ 1ði ¼ 1; 2; . . . ; nÞ. i¼1

where A is a fuzzy subset of X and indicates the weight assignments for the three factors. Establishing membership function for a single factor

The total score of sensory evaluation is 100 points. If the score of a cider sample is under fifty points, the corresponding yeast strain is directly eliminated from )1 a

Table 3 Results from the quantitative analysis of the aroma components in different ciders (mg L ) Ciders No.

Aroma components

Formula

M

F4

F6

F8

F11

F14

F15

F16

1750

1 2 3 4 5 6 7 8 9 10 11 12 13

3-Methyl-1-butanol 2-Methyl-1-butanol 2,3-Butanediol 1-Butanol 2-Phenylethyl alcohol 2,3-Octanediol 4-Hydroxy-benzene ethanol Ethyl acetate 2-Hydroxy-propanoic acid ethyl ester Ethyl caprylate Ethyl decanoate Hexanoic acid ethyl ester Ethyl hydrogensuccinate

C5H12O C5H12O C4H10O2 C4H10O C8H10O C8H10O2 C8H18O2 C4H10O2 C5H10O3 C10H2002 C12H24O2 C8H1602 C6H10O4

88 88 90 74 122 138 146 88 118 172 200 144 146

20.46 ND 11.99 2.50 4.51 1.27 0.54 25.23 1.08 10.24 14.05 20.18 2.23

35.08 ND 18.27 8.10 17.18 10.52 1.73 20.12 1.91 12.46 17.29 13.47 7.48

35.66 ND 11.3 8.79 18.17 0.74 ND 13.24 3.61 14.55 21.34 21.25 3.06

30.50 ND 14.48 5.31 15.62 4.92 1.78 25.50 2.52 14.86 18.30 17.69 10.62

ND 44.56 18.01 8.97 25.31 11.26 2.57 34.56 4.55 13.28 13.56 18.32 7.21

ND 38.12 8.56 5.30 21.21 5.54 3.00 26.38 2.90 16.88 22.15 20.75 5.39

ND 20.30 6.38 3.26 10.06 1.36 1.16 15.21 2.63 10.08 17.4 22.28 10.21

ND 32.21 4.83 1.44 20.63 6.34 1.59 26.34 4.81 9.18 11.92 24.87 1.16

ND, not detected. Mean values of three replicates.

a




Table 4 Factors and their weight values

B ¼ A R ¼ ½0:090

Factors

Sensory score

Total higher alcohols

Total esters

Weights

0.35

0.30

0.35

the evaluation set. Therefore, the membership function model we constructed for the sensory evaluation score on the basis of Table 2 is Uðx1 Þ ¼ x1 =100; ð50 x1 < 100Þ

8 < 1; Uðx3 Þ ¼ ðx3 20Þ=70; : 0;

x3 90 70 x3 < 90 x3 < 70

Obviously, there is a quantifiable relation existing between the factor sets X and the membership function sets U(xi), which may be expressed as follows rij 2 ½ 0

1 ði ¼ 1; 2; . . . ; m;

j ¼ 1; 2; . . . ; nÞ

R ¼ (rij)m·n is a fuzzy relation matrix, also called an evaluation matrix of FCE, and is expressed as follows

2

r11 6 r21 6 R ¼ 6 .. 4 .

r12 r22 .. .

rm1

rm2

r1n r2n .. . rmn

2 0:795 7 7 4 7 ¼ 0:000 5 0:757

0:815 0:834 0:753

The calculation of FCE is as follows: 0:751

0:809

0:967

where the fuzzy transformation operator is (d,¯). " # m m X X bj ¼ ðai rij Þ ¼ min 1; ðai rij Þ ðj ¼ 1; 2; . . . ; nÞ i¼1

i¼1

Normalised by bj b0j ¼ Pn

j¼1

bj

ðj ¼ 1; 2; . . . ; nÞ

0:134

0:161 0:142

0:100 0:116 The comprehensive evaluation result showed that the quality of samples decreased in the order of F14 > F15 > F11 > F6 > F8 > F16 > 1750> F4. From the result, we can see that F14 was the best stain for cider.

Chemical analysis is an effective method for analysing the chemical component in the aroma of fruit wine, and modern analytical equipment provides a powerful tool to quantitatively measure these components. It has been used extensively to classify wines based on their aroma components (Noble & Sahnnon, 1987; Vannier et al., 1999). In this article, we used chemical analysis as a tool to profile the principal aroma components in cider. These volatile components showed an important contribution to cider aroma. The results of the sensory analysis could be related to the results of the chemical analysis, although the latter did not take into account minor volatile compounds. The low scores of these samples in sensory evaluation, as shown in Table 2, were probably because of the low intensity of the aroma compounds produced by the yeast. The methodologies for yeast evaluation can be classified as either subjective or objective, depending on how the decision information is obtained and used. The sensory evaluation method for selecting yeast for cider

3

Result of fuzzy comprehensive evaluation

B ¼ A R ¼ ½0:543 0:799 0:858 0:597 0:702;

0:125

Discussion

As total higher alcohols and total esters can improve the quality of cider within a limited range (Fischer et al., 1999), the membership function models of the total alcohols and total esters were based on Table 3 as follows 8 x2 100 < 1; Uðx2 Þ ¼ ðx2 45Þ=55; 45 x2 < 100 : 0; x2 < 45 and

0:132

0:870 0:539 0:815

0:890 0:502 0:992

0:905 1:00 1:00

0:880 0:810 0:668 0:00 1:00 0:897

3 0:830 0:400 5 0:832

making usually have something to do with the decision makers’ recognition of fuzziness and uncertainty; therefore, its objectivity and fairness are questioned in some extent because of the personal partiality and bias. FCE is applied here as a method for solving the problem of imprecise subjective judgements and incomplete objective information. FCE has huge potential in selecting yeast for cider, because it can integrate the subjective and objective approaches, and the results in this article also showed that FCE has certain advantages over conventional methods. For example, it does not rely heavily on subjective human perceptions, thus it is more objective and fair. The use of the single sensory evaluation methods can easily result in distortion,



143

144


especially when evaluating samples with a similar score. In contrast, FCE can detect tiny differences between the samples. It should be utilised by the fruit wine industry around the world because of its practicability and superiority. Acknowledgments

We gratefully acknowledge the food quality supervision and testing centres (Yang Ling), Ministry of Agriculture, China, for providing the chemical measurements of cider samples. This work is part of a project funded by NSFC of China (project no. 2001BA901A19). References Arrhenius, S.P., McCloskey, L.P. & Sylvan, M. (1996). Chemical markers for aroma of Vitis vinifera var. Chardonnay regional wines. Journal of Agricultural and Food Chemistry, 44, 1085–1090. Barnet, J.A., Payne, R.W. & Yarrow, D. (1990). Yeasts, Characteristics and Identification, 2nd edn. Pp. 443–469. Cambridge: Cambridge University Press. Cavazza, A., Versini, G., Dalla Serra, A. & Romano, F. (1989). Characterization of six Saccharomyces cerevisiae strains on the basis of their volatile compounds production, as found in wines of different aroma profiles. In: Proceedings of the Seventh International Symposium on Yeasts (edited by A. Martini & E. Herbert). Pp. 163– 169. New York, NY: Wiley. De La Presa Owens, C. & Noble, A.C. (1997). Effect of storage at elevated temperatures on aroma of Chardonnay wines. American Journal of Enology and Viticulture, 48, 310–316. Escudero, A., Cacho, J. & Ferreira, V. (2000). Isolation and identification of odorants generated in wine during its oxidation: a gas chromatography-olfactometric study. European Food Research and Technology, 211, 105–110. Fischer, U., Roth, D. & Christmann, M. (1999). The impact of geographic origin, vintage and wine estate on sensory properties of Vitis vinifera cv. Riesling wines. Food Quality and Preference, 10, 281–288.


Gawel, R. (1997). The use of language by trained and untrained experienced wine tasters. Journal of Sensory Studies, 12, 267–284. Heymann, H. & Noble, A.C. (1987). Descriptive analysis of commercial cabernet sauvignon wines from California. American Journal of Enology and Viticulture, 38, 41–44. ISO (1977). Sensory Analysis – Apparatus – Wine-Tasting Glass. Pp. 3– 7, ISO 3591:1977, Group B. Geneva: ISO. ISO. (1988). Sensory Analysis. Guide for the Installation of a Chamber for Sensory Analysis. Pp. 98–110. ISO 8589:1988, Group E. Geneva: ISO. Mamede, M.E.O., Cardellob, H.M.A.B. & Pastore, G.M. (2005). Evaluation of an aroma similar to that of sparkling wine: sensory and gas chromatography analyses of fermented grape musts. Food Chemistry, 89, 63–68. Mangas, J.J., Gonzalez, M.P. & Rodriguez, R. (1996). Solid-phase extraction and determination of trace aroma and flavour components in cider by GC-MS. Chromatographia, 42, 101–105. McCloskey, L.P., Sylvan, M. & Arrhenius, S.P. (1996). Descriptive analysis for wine quality experts determining appellations by Chardonnay wine aroma. Journal of Sensory Studies, 11, 49–67. Noble, A.C. & Sahnnon, M. (1987). Profiling Zinfandel wines by sensory and chemical analyses. American Journal of Enology and Viticulture, 38, 1–5. Stone, H. & Sidell, J. (1993). Sensory Evaluation Practices. Pp. 279– 298. San Diego, CA: Academic Press Inc. Ubeda, J.F., Briones, A.I. & Izquierdo, P.M. (1998). Study of the oenological characteristics and enzymatic activities of wine yeasts. Food Microbiology, 15, 399–406. Valero, E., Moyano, L., Millan, M.C., Medina, M. & Ortega, J.M. (2002). Higher alcohols and esters production by Saccharomyces cervisiae influence of the initial oxygenation of the grape must. Food Chemistry, 78, 57–61. Vannier, A., Brun, O.X. & Feinberg, M.H. (1999). Application of sensory analysis to champagne wine characterization and discrimination. Food Quality and Preference, 10, 101–107. Vidrih, R. & Hribar, J. (1999). Synthesis of higher alcohols during cider processing. Food Chemistry, 67, 287–294. Williams, P.L. & Webb, C. (1994). The Delphi technique: a methodological discussion. Journal of Advanced Nursing, 19, 180–186.



145

Original article Microbiological and biochemical quality of grouper (Epinephelus chlorostigma) stored in dry ice and water ice Geevarethnam Jeyasekaran,* Ramasamy Anandaraj, Ponesakki Ganesan, Robinson Jeya Shakila & Duraisamy Sukumar Department of Fish Processing Technology, Fisheries College and Research Institute, Tamil Nadu Veterinary and Animal Sciences University, Tuticorin 628 008, India (Received 18 May 2005; Accepted in revised form 16 May 2006)

Summary

The changes in the microbiological and biochemical quality of grouper (Epinephelus chlorostigma) stored in dry ice (solid carbon dioxide) were investigated. Fresh fish stored in dry ice at the ratio of 1:1 (wt/wt) were found to be organoleptically suitable for consumption when they were stored for 30 h without re-icing. Fish stored in water ice (as control) at the ratio of 1:1 (wt/wt) and in a combination of dry ice and water ice at the ratio of 1:0.2:0.5 (wt/wt/wt) were acceptable up to 18 and 24 h, respectively. Total bacterial load ranged from 105 to 107 CFU g)1, while total psychrophiles from 103 to 106 CFU g)1. Total lactics were found in the levels of 102–104 CFU g)1. Total volatile basic nitrogen contents were within the limit of acceptability in all the three treatments, whereas trimethylamine nitrogen content exceeded the limit (15 mg%) and hypoxanthine content was 11.11 mg (100 g))1 on the 30 h in grouper stored only in dry ice. Lowest temperature of )5.8 °C was recorded in grouper stored only in dry ice. Hundred percent CO2 environment within the package was found in grouper that were stored in dry ice and combination of dry ice and water ice.

Keywords

Dry ice, grouper, quality, shelf life, water ice.

Introduction

The rate of deterioration in fish is highly temperature dependent and can be inhibited by the use of low storage temperature (Sivertsvik et al., 2002). Chilling slows down the deterioration of stored seafoods. When the atmospheric surrounding of the product is modified to reduce oxygen concentration, the shelf life is increased considerably because of further reduction in the rate of chemical oxidation by oxygen and in the growth of aerobic microorganism (Stiles, 1991). Most of the exporters use only crushed ice for chilling and transporting fresh fish at the ratio of 1:1 (fish:ice), and sometimes it is even higher in tropical condition (Lima dos Santos et al., 1981), which results in exorbitant transportation cost and leakage problems because of melting of ice. The other disadvantages of using water ice are more drip loss, textural toughness, nutrient loss, and protein extractability (Putro, 1989). Application of dry ice (solid carbon dioxide) for the preservation of seafoods has recently gained popularity in India. Carbon dioxide is the most important gas used in modified atmospheric packaging of fish. Because of *Correspondent: Fax: +91-461-2340574/2340401; e-mail: [email protected]

its bacteriostatic and fungistatic effect, it inhibits growth of many spoilage bacteria (Sivertsvik et al., 2002). Carbon dioxide-enriched atmospheres have been increasingly used in the last few years for the distribution of seafoods (De la Hoz et al., 2000). Dry ice acts as coolant in the present trends of shipping of fresh seafoods (Schoemaker, 1990) and has certain advantages viz. bacteriostatic effect; and it acts as insulant enveloping the fish upon evaporation (Putro, 1989). Recently, dry ice is often mixed with water ice to save shipping weight, cost and extend the cooling energy of water ice. Several exporters use dry ice in combination with water ice blindly without any scientific basis for fresh fish transportation. However, it has been earlier reported by our laboratory that combination of dry ice and water ice at the ratio of 20:50 wt/wt is efficient for keeping fresh fish in good quality with better shelf life (Sasi et al., 2000). It has also been reported that dry ice can be used as a novel-chilling medium for the rapid transportation of fresh fish by air and found that the fish, Emperor breams stored in the said dry ice and water ice combination was superior when compared with storage in only water ice (Jeyasekaran et al., 2004a). As fish, grouper (Epinephelus chlorostigma), is one of the commercially important fish varieties that is chilled and exported from India by air and the

doi:10.1111/j.1365-2621.2006.01408.x Ó 2007 The Authors. Journal compilation Ó 2007 Institute of Food Science and Technology Trust Fund

146

Quality changes in dry ice-stored grouper G. Jeyasekaran et al.

knowledge of specific spoilage organisms of different fish from various aquatic environments of tropical region and under different packaging condition is still limited, the present study was undertaken to investigate the effect of chilling with dry ice and water ice, as a shortterm method of fish preservation, on the quality of grouper (Epinephelus chlorostigma) in relation to their storage life. Materials and methods

and dry ice. In order to avoid direct contact between the ice and fish, dry ice was wrapped in kraft paper pouch and water ice in polythene bags. Packages were wrapped in polythene bags, placed in conical-shaped styrofoam boxes and sealed airtight with cellophane tape. The boxes were stored at room temperature (29 ± 2 °C). The packs were not re-iced during the entire process of study. One pack from each lot (about 28 fish) was opened and periodically analysed in triplicate for sensory, bacteriological and physical quality until sensory rejection.

Raw material

Fish, grouper (Epinephelus chlorostigma), were purchased from a fish-landing centre of Tuticorin, which is situated 1 km far away from the laboratory. Time interval between harvesting and arrival of fish at the fish-landing centre was 12 h and during this period they were iced by directly mixing with crushed ice. Fish were immediately brought to the laboratory in insulated containers and washed with potable water. Whole fish had an average weight of 710 g and length of 37 cm. Fish were beheaded and washed. They were divided into three lots of 14 kg each. First lot was packed with dry ice from a ThermoSafe Dry Ice Machine (Thermosafe, Houston, TX, USA) at the ratio of 1:1 (fish:dry ice), second lot with a standardised combination of dry ice and water ice (Sasi et al., 2000) at the ratio of 1:0.2:0.5 (fish:dry ice:water ice) and the third lot with water ice (Ziegra Flake ice Maker, Germany) at the ratio of 1:1 (fish:water ice), as normally practiced in tropical conditions (Lima dos Santos et al., 1981), which served as control. They were designated as packages I, II and III. Gloves were worn during handling of ice and fish. Care was taken to avoid direct contact of fish with water ice

Sensory evaluation

Sensory characteristics and overall acceptability of grouper (Epinephelus chlorostigma) were assessed by a panel of six experienced panelists belonging to the Faculty of Fisheries College and Research Institute on the basis of a ten-point scale on each sampling. Sensory characteristics studied included general appearance, odour and texture of fish. Scale employed for evaluating sensory quality of chilled grouper was developed based on the guidelines given by Lima dos Santos et al. (1981). Scale employed for evaluating sensory quality of chilled grouper is given in Table 1. The scores were given in the decreasing order scale with 10–9 for excellent, 8–7 for good, 6–5 for fair and acceptable, 4–3 for poor and 2–1 for very poor. The mean of the scores given by the panel represented the overall sensory quality (Huss, 1988). A score less than 4 indicates that the fish was rejected. Bacteriological analysis

Bacteriological analysis carried out in this study included total bacterial load (TPC), total psychrophiles, total

Table 1 Scale employed for sensory evaluation of grouper (Epinephelus chlorostigma) stored in dry ice and water ice General appearance

Texture

Odour

Score

Very fresh, shiny appearance, reddish white meat

Very firm, elastic to finger touch, muscle not yet in rigor Moderately firm, elastic, muscle in pre-rigor Firm, moderately elastic, muscle in rigor Slightly firm, slightly elastic, muscle in rigor

Fresh seaweedy odour

10

Loss of fresh seaweedy odour No odours, neutral odour Slightly musty, mousy, garlic odour

9 8 7

Slight soft, muscle passing out of rigor Moderately soft, muscle in post rigor Soft, slightly loose flesh

Bready, malty, yeasty odour Lactic acid, sour milk odour Butyric or acetic acid, or chloroform odour Stale cabbage water, phosphine-like odour Ammoniacal odour

6 5 4

Fresh, shiny, meat in reddish white in colour Fresh, slight loss in shiny appearance Slight loss in freshness, slight change in reddish white meat colour Loss in freshness, pale white meat Meat colour in light white, slightly bleached Complete loss in freshness, meat in milky white colour, slightly bleached Completely bleached, meat yellowish white in colour Discoloured, pale yellowish meat Completely discoloured, yellowish meat

Very soft, loosened flesh Very soft and flabby, slight retaining of finger indentation, flesh easily torn Extremely soft and flabby, strong retention of marks, flesh very easily torn


Faecal, H2S, strong ammoniacal and putrid odours

3 2 1



Biochemical analysis

Biochemical quality parameters studied included total volatile basic nitrogen (TVB-N), trimethylamine nitrogen (TMA-N) and hypoxanthine (Hx). The total volatile base nitrogen and TMA-N contents were determined by the Conway microdiffusion method (Cobb et al., 1973). Hx content was determined using the method proposed by Burt (1976).

T852 temperature probe (Consort, Turnhout, Belgium). Gas composition of different packaging environment was measured by a MAP check 9000 gas analyser (PBI Dansensor Ringsted, Denmark). Statistical analysis

Analysis of variance (anova) was performed by using statistical package (spss 10.0) to examine whether any significant difference exists between different treatments, with respect to the different fish quality characteristics at 5% level. Results and discussion

Overall sensory scores of grouper (Epinephelus chlorostigma) stored in dry ice are shown in Fig. 1. Raw fish exhibited fresh seaweedy odour, shiny appearance, bright eyes, very firm texture and reddish white meat with a sensory score of 9.9. No remarkable change was observed in package III, while the dry ice contacted fish surface became frozen in package II on the 1 h, whereas package I was completely frozen because of a drastic fall in temperature. As the fish exporters from India have already been using the dry ice blindly for the export of chilled fish to various developed countries by air and no reliable data are available on the quantity of dry ice used for chilling fish, in our study we have used the dry ice at the ratio of 1:1 as one of the packages along with an already standardised optimum level of combination of dry ice and water ice, in comparison with water ice. On the 6 h, package II retained slight seaweedy odour and shiny appearance, while no remarkable odour was observed in package III. On the 12 h of storage, bleaching of meat colour at the exposed meat area of the beheaded fish and firm texture was observed in package III. No noticeable odour was observed in package II, but the meat colour was better than package III on the 12 h. On the 18 h, package III exhibited slight 12 10 Sensory score

lactics, total coliforms and total anaerobic sulphite reducers. Media used in this study were obtained from Hi-Media Laboratories, Mumbai, India and chemicals from S. D. Fine Chemicals Limited, Mumbai, India. Fish muscle was cut into very small pieces using sterile knife and forceps and pooled together. Then, 25 g was taken from this pool and homogenised using 225 mL sterile physiological saline (0.85%) and serial decimal dilutions of each homogenate were carried out with the same diluent for the respective bacteriological analysis (APHA, 2001). Appropriate dilutions were spread plated onto Trypticase soya agar (TSA) for the enumeration of TPC and total psychrophilic count. The plates were incubated at 37 °C for 24 h for the enumeration of TPC, whereas they were incubated under refrigerated condition (5 °C) for 7 days for the enumeration of psychrophiles. The bacterial flora isolated from packages I, II and III were identified by various biochemical tests (LeChevallier et al., 1980; Balows et al., 1992). Doublelayer pour plate technique was followed for the enumeration of total lactics using Lactobacillus MRS agar (USFDA, 2001). Inoculated plates were incubated at the room temperature for 72 h. After appropriate incubation, number of suspected colonies developed on the plates were counted and expressed as CFU g)1. Most probable number (MPN) technique was followed for the enumeration of total coliforms and total anaerobic sulphite reducers using Lauryl sulphate tryptose broth and differential reinforced clostridial medium (DRCM), respectively. The inoculated tubes were incubated at 37 °C for 24 h for the enumeration of total coliforms. The tubes showing acid and gas productions were counted as positive and expressed as MPN g)1, whereas, for total anaerobic sulphite reducers, the tubes were incubated in a water bath at 37 °C for 96 h. The tubes exhibiting black precipitate were counted as positive and expressed as MPN g)1.

8 6 4

Physical analysis

2

Physical parameters studied were pH, temperature and gas composition. pH was determined by using a Digisun pH meter 707, (Digisun Electronics, Hyderabad, India) by taking 10 g of homogenised sample in 100 mL distilled water. Changes in temperature of all the packages were recorded by using an Ultrafreezer Model

0 0

1

6

12

18

24

30

36

Storage period (h) 100% Dry ice

20% Dry ice + 50% Water ice

100% Water ice

Figure 1 Changes in sensory quality of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.



147


ammoniacal odour and soft texture with a sensory score of 5.2, whereas package II exhibited the similar sensory characteristics only on the 24 h of storage. On the contrary, package I was partially thawed on the 6 h of storage. On the 12 h, the fish became normal in condition and exhibited fresh meat colour with slightly firm texture with a score of 8.8. On the 24 h, package I exhibited slight ammoniacal odour and slightly lost its texture with a score of 6.3. On the 30 h, package I had strong ammoniacal odour and soft texture with a score of 4.9. Sensory results of salmon (Salmo salar) steaks showed shelf life extension of 6 days under 20% CO2enriched atmosphere, when compared with air (De la Hoz et al., 2000). Scott et al. (1986) observed the shelf life extension of 9 days in snapper (Chrysophrys auratus) when stored in 100% carbon dioxide atmosphere at )1 °C when compared with normal atmospheric storage. Sasi et al. (2003) reported that combination of dry ice along with water ice extended the shelf life of seerfish when compared with simple water ice alone. The changes in sensory quality of fish stored in different packages of ice were statistically significant (P < 0.05). The changes in total bacterial count of grouper stored in dry ice are shown in Fig. 2. Fresh fish exhibited an initial total bacterial population of 106 CFU g)1, which reduced by a log on the 1 h in all the three packages and the same was maintained up to 6 h. Application of CO2 gaseous environment and exposure to low temperature inhibits bacterial growth during the lag phase (Clark & Lentz, 1969; Jay, 1987). This might be the reason for reducing total bacterial population on the 1 and 6 h of storage. On the 12 h, the total bacterial population increased by a log in all the three packages, which maintained up to 24 h in packages I and II. At the end of the storage period, packages I and III had a population of 107 CFU g)1. Papadopoulos et al. (2003) reported that a mesophilic count of 107 CFU g)1 was the maximum level for the acceptability of marine fish, which correlates with the results of the present

study. The same population of 107 CFU g)1 was also observed by Jeyasekaran et al. (2004a) when emperor breams (Lethrinus miniatus) were stored in a combination of dry ice and water ice. Fresh fish, grouper used in this study carried a bacterial flora of Moraxella, Staphylococcus, Micrococcus, Aeromonas, Alteromonas and Streptococcus (Fig. 3). Moraxella constituted about 47% of the flora. Liston (1980) reported that the flora on tropical fish often carries a slightly higher load of Gram-positive bacteria compared with fish from colder waters. But, in the present investigation, 45% flora in raw fish was from Gram-positive group. Moraxella was the dominant flora in package I, which constituted about 25% followed by Aeromonas, Pseudomonas, Flavobacterium, Alcaligenes, Micrococcus, Plesiomonas, Bacillus, Staphylococcus, Corynebacterium and Streptococcus (Fig. 4). However, Pseudomonas was the dominant microflora in packages II and III with varying percentage. Pseudomonas and H2S producing bacterial population have been reported to be the specific spoilage bacteria in fish from temperate and tropical waters (Lima dos Santos et al., 1981; Gram & Huss, 1996). In package II, Pseudomonas constituted about 33% followed by Alcaligenes, Moraxella, Micrococcus, Aeromonas, Flavobacterium, Streptococcus, Staphylococcus and Bacillus (Fig. 5) while, in package III it constituted 32% followed by Moraxella, Staphylococcus, Alcaligenes, Micrococcus, Aeromonas, Flavobacterium, Streptococcus and Bacillus (Fig. 6). Callow (1932) suggested that displacement of oxygen by CO2 inhibits the growth of aerobic microorganisms and results in shift of dominant flora. The present investigation also observed the shift of dominant flora between air held and dry ice stored steaks. It was observed by De la Hoz et al. (2000) that 4%

11% 2%

9 8 Bacterial count (log 10)

148

7

47%

6

33%

5 4 3 2

3%

1 0

0

1 100% Dry ice

12 18 6 Storage period (h) 20% Dry ice + 50% Water ice

24

30

100% Water ice

)1 Figure 2 Changes in total bacterial population (CFU g ) of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.


Micrococcus Alteromonas Staphylococcus

Streptococcus Moraxella Aeromonas

Figure 3 Bacterial flora associated with raw grouper (Epinephelus chlorostigma).



9%

3%

2%

8%

1%

1%

5%

4%

10%

18%

11%

1% 33%

0 14%

23%

2%

24% 3% Micrococcus Staphylococcus Bacillus Alcaligenes Moraxella

28%

Pseudomonas Flavobacterium Aeromonas Plesiomonas Corynebacterium

Figure 4 Bacterial flora of grouper (Epinephelus chlorostigma) stored in 100% dry ice.

9%

Micrococcus Alcaligenes Streptococcus Flavobacterium

Moraxella Bacillus

Pseudomonas Aeromonas Staphylococcus

Figure 6 Bacterial flora of grouper (Epinephelus chlorostigma) stored in 100% water ice.

14%

8

5% 6%

22%

Bacterial count (log 10)

7 6 5 4 3 2 1

18%

0 0

1

6

12

18

24

30



100% Water ice )1

Figure 7 Changes in total psychrophiles (CFU g ) of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.

3% 7%

Micrococcus Flavobacterium Streptococcus

16%

Pseudomonas Bacillus Staphylococcus

Moraxella Alcaligenes Aeromonas

Figure 5 Bacterial flora of grouper (Epinephelus chlorostigma) stored in a combination of dry ice and water ice.

Pseudomonas and Shewanella were the dominant spoilage flora responsible for the spoilage of salmon steaks stored in air.

Changes in total psychrophiles of grouper stored in dry ice are presented in Fig. 7. Initial psychrophilic population was 105 CFU g)1, which reduced by a log on the 1 h in all the packages. On the 6 h, the population further reduced by a log in package II, whereas the same was maintained in packages I and III. Sasi et al. (2003) reported that the lower psychrophilic count was mainly due to the effect of dry ice, which creates CO2 enriched environment. On the 12 h, all the packages attained a psychrophilic population of 105 CFU g)1, which maintained on the 18 h in package I, while it increased by a



149


6 5 Bacterial count (log 10)

150

4 3 2 1 0 0

1

6

12

18

24

30



100% Water ice

)1

Figure 8 Changes in total lactics (CFU g ) of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.

log in packages II and III and the same count (106 CFU g)1) was maintained until the end of the storage period in all the packages. Similar psychrophilic count was also observed at the end of storage period by Jeyasekaran et al. (2004b) in ice-stored barracudas (Sphyraena barracuda). Changes in total lactics population of grouper stored in dry ice are shown in Fig. 8. Initial lactics population was 104 CFU g)1, which reduced to 102 CFU g)1 in all the packages on the 1 h and this count was maintained up to 12 h. De la Hoz et al. (2000) observed that the growth of lactic acid bacteria was not promoted by the enrichment of the atmosphere with 20% CO2, although it resulted in a greater presence of this group in the microbiota responsible for the spoilage. On the 18 h of storage, lactic acid bacterial population increased from 102 to 103 CFU g)1 in all the packages. The population further increased by a log on the subsequent storage period and maintained until the end in package I. A significant difference (P < 0.05) in lactics count was observed between different packages of ice-stored grouper fish. It has been earlier reported that dry ice

provides a favourable environment for the growth of lactics (Sasi et al., 2003). Changes in total coliforms and anaerobic sulphite reducers of grouper stored in dry ice are presented in Table 2. Initial coliform count was 175 MPN g)1. On further storage, a decreasing trend in coliforms count was observed up to 6 h of storage in package III, while up to 1 h in packages I and II, which might be due to fall in temperature. On continued storage, the population gradually increased and reached 1600 MPN g)1 at the end of the storage in all the three packages. Lilabati & Vishwanath (1998) also reported the presence of coliforms throughout the storage of fish, Notopterus chitala in water ice. Initial total anaerobic sulphite reducer count was 1.5 MPN g)1. During the storage period, the population did not exhibit any consistent trend as it varied from 2.5 to 110 MPN g)1 in package I, 0.7 to 9.5 MPN g)1 in package II and from 2.5 to 30 MPN g)1 in package III. This inconsistent trend was also reported by Jeyasekaran et al. (2004a) and suggested that the partial aerobic and anaerobic conditions prevailing during storage might be the reason to cause such differences. Changes in the TVB-N contents of grouper stored in dry ice are shown in Fig. 9. Fresh fish exhibited an initial TVB-N content of 16.67 mg%. On the 1 h, the TVB-N content decreased to 9.16 and 14.39 mg% in packages I and II, respectively, while it increased to 17.43 mg% in package III. Sasi et al. (2000) reported that the formation of TVB-N and TMA-N was comparatively less in the 20% dry ice-stored sardine along with 50% water ice. On the 6 h, the content further decreased in packages I and II. On further storage, the content continuously increased in all the three packages and reached a level of 28.99, 22.18 and 24.42 mg% in packages I, II and III, respectively, at the end of the storage period. During storage period, TVB-N content did not exceed the limit of acceptability. Dalgaard (2000) reported that the rejection limit of 30–35 mg% for TVB-N varied with species and processing condition.

Total anaerobic sulphite reducers (MPN g)1)

Total coliforms (MPN g)1) Storage period (h)

Package I

Package II

Package III

Package I

Package II

Package III

0 1 6 12 18 24 30

175 40 48 426 1600 1600 1600

175 41 542 125 345 1600 DC

175 49 6.0 542 1600 DC DC

1.5 4.5 2.5 110 20 9.5 110

1.5 0.7 2.5 0.9 9.5 9.5 DC

1.5 2.5 2.5 9.5 30 DC DC

Table 2 Changes in total coliforms and total anaerobic sulphite reducers of grouper (Epinephelus chlorostigma) stored in dry ice and water ice

DC, discontinued. Package I, 100% dry ice; Package II, 20% dry ice + 50% water ice; Package III, 100% water ice.



35

12

30

10 Hx content [mg (100g)–1]

TVB-N content (mg %)


25 20 15 10

6 4 2

5

0

0 0

1 100% Dry ice

6

12 18 Storage period (h)


24

0

30

1 100% Dry ice

100% Water ice

6

12 18 Storage period (h)


24

30

100% Water ice

Figure 11 Changes in Hx content of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.

Figure 9 Changes in TVB-N contents of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.

7

18 16

6.8

14

6.6

12 10

pH

TMA-N content (mg%)

8

6.4

8

6.2

6 4

6

2

5.8

0 0

1

6

12

18

24

0

30

1


12

18

24

30

Storage period (h)


6

100% Dry ice

100% Water ice

20% Dry ice + 50% Waterice

100% Waterice

chlorostigma) stored in dry ice and water ice.

Figure 12 Changes in pH of grouper (Epinephelus chlorostigma) stored in dry ice and water ice.

The changes in TVB-N contents were statistically significant (P < 0.05) between different packages of grouper fish stored in ice. It has also been observed in the present study that the changes in TVB-N contents were almost similar to the changes in the total bacterial count of grouper. Changes in TMA-N contents of grouper stored in dry ice are given in Fig. 10. Initially, the TMA-N content was 9.72 mg%, which was in decreasing trend up to 12 h in package II and 6 h in package III, while the content decreased to 1.31 mg% on the 1 h in package I. On further storage, the content increased gradually in all the three packages and exceeded the limit of acceptability (15 mg%) in package I at the end of storage period, while it was within the limit in the packages II and III. In marine fish, the rate of increase in TMA-N varies considerably from species to species (Amu & Disney, 1973; Huss, 1988). The differences observed in the rate of accumulation of TMA-N contents could be due to the differences in the growth of bacteria capable of reducing trimethylamine-oxide (TMAO) and in the content of

TMAO in fish (Hebard et al., 1982). Oehlenschlager (2002) suggested that the formation of TMA is much less in ungutted and gill cut/bled fish because of a much reduced microbial action. Changes in Hx contents of grouper stored in dry ice are given in Fig. 11. Raw fish exhibited an initial Hx content of 3.91 mg (100 g))1, which reduced to 2.93 mg (100 g))1 in package III on the 1 h, while it reduced to 1.33 and 1.89 mg (100 g))1 in packages I and II on the 12 h, respectively. On further storage, the content increased and reached to 11.11, 6.15 and 8.61 mg (100 g))1 at the end of storage period in packages I, II and III, respectively. Perez-Villarreal & Pozo (1990) observed a linear increase in Hx content at a slow rate reaching 20 mg (100 g))1 at the end of storage period of ice-stored albacore. Changes in pH of grouper stored in dry ice are shown in Fig. 12. Fresh fish had a pH of 6.84. During storage period, the pH did not show any clear trend. pH observed in package I ranged from 6.23 to 6.88, 6.30 to 6.66 in package II and 6.20 to 6.88 in package III. De la

Figure 10 Changes in TMA-N contents of grouper (Epinephelus



151

152


Table 3 Changes in temperature profile of grouper (Epinephelus

chlorostigma) stored in dry ice and water ice Temperature profile (°C) Storage period (h)

Package I

Package II

Package III

Storage room

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

13 )5.0 )5.8 )4.5 )4.0 )3.1 )2.9 )1.9 )1.5 )0.9 )0.4 1.2 2.1 3.9 5.8 8.1 9.8 12.7 14.1 16.6 19.7 20.9 22.7 24.4 25.2 26.3 27.3 27.8 28.4 28.8 29.3

21.7 7.2 4.4 3.5 4.2 4.6 5.3 6.5 7.6 8.7 9.8 11.5 12.4 14.9 16.4 17.6 18.7 20.2 21.1 22.2 23.5 24.3 25.3 26.1 26.9 DC

17.9 9.1 7.5 6.3 6.8 7.6 9.7 12.1 14.0 16.2 17.8 19.7 21.1 22.5 23.6 24.5 25.3 26.3 26.8 27.4 28.0 28.4 28.9 29.3 DC

30.4 29.9 29.9 30.5 31.0 30.8 30.9 31.3 31.2 31.4 31.4 31.4 31.3 31.2 31.0 30.1 30.7 30.7 30.4 30.3 30.1 30.0 30.0 30.1 30.2 30.3 30.4 30.5 30.7 30.9 31.0 31.3


Hoz et al. (2000) observed that the rise in pH was lower in salmon (Salmo salar) steaks when stored in CO2enriched atmospheres than in the air atmosphere, which Package I

Package II

was attributed to the growth and metabolism of gramnegative oxidase-positive organisms. Temperature profile of grouper stored in dry ice is presented in Table 3. Immediately after packaging, the temperatures recorded in packages I, II and III were 13.0, 21.7 and 17.9 °C, respectively. At that time, room temperature was 30.4 °C. A lowest temperature of )5.8 °C was observed in package I on the 2 h and the sub-zero temperature was maintained up to 10 h of storage. However, LeBlanc & LeBlanc (1992) observed a lowest temperature of )1 °C when the haddock fillets were packed with 25% CO2 snow. But, in packages II and III, the lowest temperatures recorded were 3.5 and 6.3 °C on the 3 h of storage, respectively. After that, it gradually increased throughout the storage period. The results of present study confirmed the fact that the rate of deterioration/spoilage is highly temperature dependent (Sivertsvik et al., 2002). Changes in gas composition of grouper stored in dry ice are presented in Table 4. The atmospheric gas composition during this study was oxygen at 20.5%, carbon dioxide 0.6% and nitrogen 78.9%. After 1 h of packing, the oxygen, carbon dioxide and nitrogen contents were 0.028%, 100% and 0% in package I, 0.02%, 100% and 0% in package II and 20.4%, and 4.8% and 74.8% in package III, respectively. Highest level of 100% CO2 was noticed on the 1 h in the packages I and II, which was due to the packing of fish with dry ice. On the contrary, slight variation was observed in package III throughout the storage period. Earlier studies reported that the longer shelf life obtained in dry ice-packed fish might be due to high content of CO2 in such packages (Villemure et al., 1986; Dhananjaya & Stroud, 1994; Randell et al., 1999). Conclusion

Based on the results of microbiological, biochemical, physical and sensory quality tests, it may be concluded that the shelf life of grouper stored in a combination package of dry ice and water ice extended up to 25% in relation to their storage only in water ice. The

Package III

Storage period (h)

O2 (%)

CO2 (%)

N2 (%)

O2 (%)

CO2 (%)

N2 (%)

O2 (%)

CO2 (%)

N2 (%)

0 1 6 12 18 24 30

20.5 0.028 2.08 7.42 16.2 18.3 18.2

0.6 100 90.2 63.0 26.7 9.1 8.6

78.9 0 7.7 29.6 65.4 72.6 73.2

20.5 0.022 3.23 8.78 17.5 16.7 DC

0.6 100 84.4 56.3 15.0 14.1 DC

78.9 0 12.4 34.9 67.5 69.2 DC

20.5 20.4 20.8 19.9 18.3 DC

0.6 4.8 2.6 2.6 3.0 DC

78.9 78.8 76.5 77.5 78.7 DC

Table 4 Changes in gas composition of grouper (Epinephelus chlorostigma) stored in dry ice and water ice





combination package was found to be economical for the export of chilled fish by air, as the reduction in the cost of air transport was found to be 15–25%, when compared with the traditional method of packing and exporting fish with only water ice. Acknowledgments

The authors thank the Indian Council of Agricultural Research (ICAR), New Delhi, Government of India for the financial assistance extended for carrying out this study. The support and facilities provided by the Dean, Fisheries College and Research Institute, Tamilnadu Veterinary and Animal Sciences University, Tuticorin, India is also gratefully acknowledged. References Amu, L. & Disney, J.G. (1973). Quality changes in West African marine fish during iced storage. Tropical Science, 15, 125–138. APHA. (2001). Compendium of Methods for the Microbiological Examination of Foods. Pp. 25–35. Washington, DC: American Public Health Association. Balows, A., Truper, H.G., Dworkin, M., Harder, W. & Schleifer, K.H. (1992). The Prokaryotes – A Handbook on the Biology of Bacteria: Ecophysiology, Isolation, Identification and Application. 2nd edn. New York, NY: Springer-Verlag. Burt, J.R. (1976). Hypoxanthine: a biochemical index of fish quality. Aberdeen: Torry Research Station. Callow, E.H. (1932). Gas storage of pork and bacon. Part I-Preliminary experiments. Journal of the Society of Chemical Industry, 51, 116–119. Clark, D.S. & Lentz, C.P. (1969). The effect of carbon dioxide and growth of slime producing bacteria on fresh beef. Canadian Institute of Food Technology Journal, 2, 72–75. Cobb, F., Alanoz, I. & Thompson, C. (1973). Biochemical and microbial studies on shrimp: volatile nitrogen and amine nitrogen analysis. Journal of Food Science, 38, 431–436. Dalgaard, P. (2000). Freshness, quality and safety in seafoods. FLAIRFLOW, Europe Technical Manual F-FE 380A/00. Dublin: The National Food Centre. De la Hoz, L., Lopez-Galvez, D.E., Fernandez, M., Hierro, E. & Ordonez, J.A. (2000). Use of carbon dioxide enriched atmospheres in the refrigerated storage (2 °C) of salmon (Salmo salar) steaks. European Food Research Technology, 210, 179–188. Dhananjaya, S. & Stroud, G.D. (1994). Chemical and sensory changes in haddock and herring stored under modified atmosphere. International Journal of Food Science and Technology, 29, 575–583. Gram, L. & Huss, H. (1996). Microbiological spoilage of fish and fish products. International Journal of Food Microbiology, 33, 589–595. Hebard, C.E., Flick, G.J. & Martin, R.E. (1982). Occurrence and significance of trimethylamine oxide and its derivatives in fish and shellfish. In: Chemistry and Biochemistry of Marine Food Products (edited by R.E. Martin, G.J. Flick & D.R. Ward). Pp. 149–272. Westport, CT: AVI. Huss, H.H. (1988). Fresh Fish: Quality and Quality Changes. Rome: FAO. Jay, J. (1987). Modern Food Microbiology, 1st edn. p. 642. New Delhi: CBS Publishers and Distributors. Jeyasekaran, G., Ganesan, P., Jeya Shakila, R., Maheswari, K. & Sukumar, D. (2004a). Dry ice as a novel chilling medium along with water ice for short-term preservation of fish emperor breams,

lethrinus (Lethrinus miniatus). Innovative Food Science and Emerging Technologies, 5, 485–493. Jeyasekaran, G., Ganesan, P., Maheswari, K., Jeya Shakila, R. & Sukumar, D. (2004b). Effect of delayed icing on the microbiological quality of tropical fish: barracudas (Sphyraena barracuda). Journal of Food Science, 69, 197–200. LeBlanc, R.J. & LeBlanc, E.L. (1992). The effect of superchilling with CO2 snow on the quality of commercially processed haddock (Melanogrammus anglefinus) fillets. In: Seafood Science and Technology (edited by H.H. Huss, M. Jakobsen & J. Liston). Pp. 247–257. London: Fishing News Books. LeChevallier, M.W., Seidler, R.J. & Evan, T.M. (1980). Enumeration and characterization of standard plate count bacteria in chlorinated raw water supplies. Applied and Environmental Microbiology, 40, 922–930. Lilabati, H. & Vishwanath, W. (1998). Biochemical, nutritional and microbiological quality of ice stored Notopterus chitala of Imphal market, Manipur. Indian Journal of Fisheries, 45, 441–446. Lima dos Santos, C.A.M., James, D. & Teutscher, F. (1981). Guidelines for chilled fish storage experiments. FAO Fisheries Technical Paper, 210, 17–21. Liston, J. (1980). Microbiology in fisheries science. In: Advances in Fish Science and Technology (edited by J.J. Connell). Pp. 138–157. Farnham: Fishing News Books. Oehlenschlager, J. (2002). Influence of different pre-storage treatments on resulting shelf life of ice-stored Barents-sea cod. Journal of Aquatic Food Product Technology, 11, 187–200. Papadopoulos, V., Chouliara, I., Badeka, A., Savvaidis, I.N. & Kontominas, M.G. (2003). Effect of gutting on microbiological, chemical, and sensory properties of aquacultured sea bass (Dicentrarchus labrax) stored in ice. Food Microbiology, 20, 411–420. Perez- Villarreal, B. & Pozo, R. (1990). Chemical composition and ice spoilage of albacore (Thunnus alalunga). Journal of Food Science, 55, 678–682. Putro, S. (1989). Dry ice – possible uses in fresh and live fish handling. INFOFISH International, 4, 24–25. Randell, K., Hattula, T., Skytta, E., Sivertsvik, M. & Bergslien, H. (1999). Quality of filleted salmon in various retail packages. Journal of Food Quality, 22, 483–497. Sasi, M., Jeyasekaran, G., Shanmugam, S.A. & Jeya Shakila, R. (2000). Chilling fresh fish in dry and wet ice. Asian Fisheries Science, 13, 375–382. Sasi, M., Jeyasekaran, G., Shanmugam, S.A. & Jeya Shakila, R. (2003). Evaluation of the quality of seerfish (Scomberomorus commersonii) stored in dry ice (solid carbon dioxide). Journal of Aquatic Food Product Technology, 12, 61–72. Schoemaker, R. (1990). Shipping of fresh seafoods. INFOFISH International, 4, 27–30. Scott, D.N., Fletcher, G.C. & Hogg, M.G. (1986). Storage of snapper fillets in modified atmospheres at )1 °C. Food Technology in Australia, 38, 234–238. Sivertsvik, M., Jeksrud, W.K. & Rosnes, J.T. (2002). A review of modified atmosphere packaging of fish and fishery products – significance of microbial growth, activities and safety. International Journal Food Science and Technology, 37, 107–127. Stiles, M.E. (1991). Scientific principles of controlled/modified atmosphere packaging. In: Modified Atmosphere Packaging of Food (edited by B. Ooraikul & M.E. Stiles). Pp. 18–25. London: Ellis Horwood. USFDA. (2001). Bacteriological Analytical Manual, 8th edn. Revision A, Gaithersburg, MD: AOAC International. Villemure, G., Simaro, R.E. & Picard, G. (1986). Bulk storage of cod fillets and gutted cod (Gadus morhua) under carbon dioxide atmosphere. Journal of Food Science, 51, 317–320.



153


154

Original article Antioxidative activities of grape (Vitis vinifera) seed extracts obtained from different varieties grown in Turkey Oktay Yemis,* Emre Bakkalbasi & Nevzat Artik Department of Food Engineering, Faculty of Engineering, Ankara University, Diskapi Campus, Diskapi 06110 Ankara, Turkey (Received 8 May 2006; Accepted in revised form 25 July 2006)

Summary

In this study, total phenolic contents and antioxidant activities of grape seed extracts obtained from twelve different grape seeds from common varieties grown in Turkey were determined. Grape seeds were extracted with 70% acetone and extraction yield of grape seed were calculated. The total phenolic content of grape seed extracts were determined by the Folin-Ciocalteu procedure and ranged from 33 945 to 58 730 mg per 100 g extract as gallic acid equivalent. Antioxidant activities of grape seed extracts with two different free radical scavenging methods, ABTS [2,2/-azinobis (3-ethylbenzothiazoneline-6-sulfonic acid)] and DPPH (2,2-diphenyl-picrylhydrazyl) assays, using Trolox equivalent as standards, were investigated. Grape seed extracts exhibited antioxidant activities 2.46–4.14 and 3.55–5.76 [Trolox equivalent antioxidant capacities (TEAC) mg)1 extracts] in ABTS and DPPH assays, respectively. Compared with varieties, Muskule extracts exhibited the lowest total phenolic content, TEACABTS and TEACDPPH value while Narince extracts had the highest total phenolic content and TEACDPPH value, and Alphonse Lavallee´ had the highest TEACABTS value. Total phenolic content showed that there is a significant correlation with TEACDPPH (r ¼ 0.7974, P £ 0.001) and TEACABTS values (r ¼ 0.4860, P £ 0.05).

Keywords

ABTS, antioxidant activity, DPPH, grape seed, total phenolic content.

Introduction

Grape (Vitis vinifera) is one of the world’s largest fruit crops with an annual production of 65 486 235 metric tonnes (Mt). Turkey is the sixth largest producer with an estimated production of 3 650 000 Mt of grape (Faostat, 2004). Most of the production has been processed to different products such as raisin, wine, vinegar, grape juices and different traditional products like grape concentrate (pekmez) in the food industry (Simsek et al., 2004). There are about 133 different varieties, of which twenty-two types are domestic origins of wine grape in Turkey (Celik, 2002). Grape seed is an important by-product of grape processing. Grape seeds are known to be a rich source of a number of phenolic compounds. Polyphenols in grape seeds are mainly flavonoids, including the monomeric flavan-3-ols catechin, epicatechin, gallocatechin, epigallocatechin and epicatechin 3-O-gallate, and procyanidin dimers, trimers and more highly polymerised procyanidins (Fuleki & Ricardo-da-Silva, 1997). Grape seed extract is known as a powerful antioxidant that *Correspondent: Fax: +90 312 317 8711; e-mail: [email protected]

protects the body from premature aging, disease and decay. The pharmacological and nutraceutical benefits derived from grape seed polyphenols are because of their free radical scavenging capability. There are a number of studies reported that grape seed polyphenols reduce the risk of cancer and heart disease by inhibiting the oxidation of low-density lipoprotein (LDL) (Shi et al., 2003). Teissedre et al. (1996) indicated that phenolics in grapes and wines inhibited human LDL oxidation. Antioxidant activity is the common assay used and widely accepted by researchers as an anticancer indicator (Tsai et al., 2005). The antioxidant activity of grape seed extract has been tested by using b-carotene linoleate and linoleic acid peroxidation (Jayaprakasha et al., 2001) and 2,2-diphenyl-picrylhydrazyl (DPPH) methods (Jayaprakasha et al., 2003). The radical scavenging activity of acetone:water:acetic acid (90:9.5:0.5) and methanol:water:acetic acid (90:9.5:0.5) extracts of grape seed was shown to be 45.6% and 41.3%, respectively at 25 ppm concentration (Jayaprakasha et al., 2003). The addition of antioxidants is a method of increasing shelf life, especially of lipids and lipid-containing foods. Nowadays, there has been a growing interest to use natural antioxidants especially of plant origin rather


Antioxidative activities of grape seed O. Yemis et al.

than synthetic antioxidants, such as butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT). The usages of these synthetic antioxidants have been restricted because they are suspected to have toxicological effects (Mavadi & Salunkhe, 1995). Recently, the antioxidant power and health benefits of grape seed extract have been demonstrated by scientific studies and this leads to their use as a dietary supplement and food additive (Nakamura et al., 2003). The aim of this work is to evaluate grape seeds as a source of natural antioxidant polyphenols for their possible use as dietary supplement or food antioxidants. To this purpose, the extract yield, total phenolic content and antioxidant capacity of twelve common varieties, grown in Turkey were investigated. Materials and methods

Materials

Mature grapes from twelve different varieties of V. vinifera were harvested in different locations of the Middle Anatolia in the 2004 growing seasons. General information about the varieties including colour, consumption style and the number of seeds are given in Table 1. Grape seeds were removed manually from the berries. The seeds were washed with tap water to separate pomace residue, air-dried and stored at )20 C until analysed. The seeds were ground with a laboratory-type grinder (Falling number AB KT-30, Stockholm, Sweden) before extraction. ABTS [2,2/-azinobis (3-ethylbenzothiazoneline-6-sulfonic acid)] diammounium salt was obtained from Fluka (Steinheim, Germany). Methanol, ethanol, n-hexane, acetone (analytic grade) for extraction and Folin-Ciocalteu reagents were purchased from Merck (Darmstadt, Germany). Standards of gallic acid, trolox, potassium persulfate and 2,2-diphenyl-pic-

Table 1 Colour, consumption style and number of seeds about the analysed grape cultivars

Grape cultivars White cultivars Muskule Razaki Emir Hasandede Narince Coloured cultivars Karadimrit Muscat of Hamburg Alphonse Lavallee´ Okuzgozu Kalecik karasi Alicante Boushet Papaz karasi

rylhydrazyl (DPPH) stable radical were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). Methods Extraction

The extraction procedure of phenolic compounds in grape seeds was based on the modified methods of Baker et al. (1995) and Yilmaz & Toledo (2006). Grape seed powder was deoiled with n-hexane (one part powder to ten parts of hexane, w/v) and n-hexane removed in a vacuum oven (40 C). Nearly 0.5 g of the deoiled grape seed powder was put into a centrifuge tube and extracted by shaking with 9.5 mL of 70% (v/v) acetone in water for 1 h, in the dark and at room temperature. After this, the mixture in the tube was centrifuged at 8750 g (Sigma, Osterode am Harz, Germany) for 15 min at 20 C and the supernatant was transferred into an amber bottle. The procedure was repeated four times and the supernatants were combined. Pooled supernatants were evaporated in a rotary vacuum evaporator (Buchi B114; Flawil, Switzerland) (T £ 45 C). Residue was dissolved in methanol and the volume was adjusted to 100 mL with methanol. Two replicate extractions of each sample were performed. Dry matter content

Dry matter content of grape seed were determined by drying duplicate samples in the vacuum oven at 60 C (AOAC, 1995). Evaluation of antioxidant activity Test for ABTS radical-scavenging

activity. Radicalscavenging activity of grape seed extract was determined spectrophotometrically (Shimadzu 1601 UV-VIS spectrophotometer) with ABTS decolourisation assay (Re et al., 1999). ABTS was used as the free radical provider and generated by reacting with

Colour

Consumption style

Number of seeds

Green-yellow Gren-pinky yellow Yellowish-green Green-yellow Yellow

Table Table Wine Wine Wine

1–3 2–4 2–3 2–3 2–3

Reddish purple Purplish black Purplish black Black with gray bloom Black with blue bloom Heavy blue-gray bloom Black with gray bloom

Raisin Table Table Wine Wine Wine Wine

1–2 2–3 3–4 2–3 1–2 1–2 2–3



155

156


potassium persulfate (2.45 mm) for 16 h. The solution was diluted to obtain an absorbance of 0.7 (±0.02) with ethanol (98%) before use. ABTS+ cation solutions (1 mL) were transferred to the 1-cm path length disposable microcuvette and the initial absorbance of the solution in the microcuvette was recorded at 734 nm. The extracts under investigation were diluted to have it fit in the concentration range of the standard. The four different concentrations of extracts were added into the cuvette containing the ABTS+ solution and the mixtures were shaken. The decrease of absorption was recorded at the end of 6 min at 25–30 C. Trolox was used as a standard and the trolox equivalent antioxidant capacity (TEAC) of each extract was calculated as detailed in the later section and expressed as millimoles of trolox per milligram of extract. Each extract was analysed in triplicate. DPPH radical-scavenging assay. Radical-scavenging activity of grape seed extracts were measured using the slightly modified method of Brand-Williams et al. (1995). An aliquot of grape seed extract (0.025 mL) at four different concentrations was added to 0.975 mL of DPPH solution (0.025 g L)1 in methanol). The absorbance (Shimadzu 1601 UV-VIS spectrophotometer) was measured immediately at 515 nm (ATo) and then, the mixture was left for 120 min in dark at room temperature until the reaction reached a plateau (AT120). The decrease in absorption was used for calculation of TEAC value as in eqns 1 and 2. Trolox was used as a standard. Methanol was used instead of the sample for blank measurements. All determinations were performed in triplicate. Trolox equivalent antioxidant capacity. Results of ABTS and DPPH assay were given as TEAC values. TEAC value is an indicator of the antioxidant capacity of the sample relative to trolox on a molar basis. Calibration curves for both antioxidant activity assays were prepared with different final concentrations of trolox and the TEAC values of antioxidant assays were calculated as follows (Van den Berg et al., 1999; Oomah et al., 2005; Madhujith & Shahidi, 2005). DA ¼ ðAðToÞ AðTÞ Þ DAblank

ð1Þ

TEACsample ¼ ½DAsample =m d

ð2Þ

where A(To) is the absorbance value of the sample or standard at 0 min; A(T) is the absorbance value of the sample or standard at 6 min for ABTS+• and 120 min for DPPH•; DAsample is the reduction in corrected absorbance; m is the slope of the standard curve and d is the dilution factor. Determination of total phenolic content

Total soluble phenolics in extracts were determined with Folin-Ciocalteu’s phenol reagent according to an


improved version of the procedure explained by Singleton & Rossi (1965). A 10-lL grape seed extract [0.5 g (100 mL))1] was transferred to a test tube and distilled water (765 lL) was added on the extract. Exactly 75 lL of Folin-Ciocalteu’s phenol reagent and 750 lL of sodium carbonate (60 g L)1) were added and slightly mixed. The mixture in the tube was held for 90 min at room temperature and then the mixture was transferred to disposable microcuvette. The absorbance of the developed colour was read at 725 nm by the spectrophotometer. The total content of phenolic compounds in each grape seed extracts was determined by a standard curve prepared with gallic acid and expressed as mg per 100 g of extract. Each extract was analysed in triplicate. Statistical analysis

Statistical analysis of the result was performed using SAS/STAT (SAS. Ins. Inc. Cri. NCI) software. PROC GLM with Duncan’s multiple comparison test (a ¼ 0.05) and PROC CORR with correlation coefficients was performed. Results and discussion

Different solvents such as 70% methanol/water (Yilmaz & Toledo, 2004), methanol (Santos-Buelga et al., 1995), ethyl acetate (Guendez et al., 2005), 10% water ⁄ ethyl acetate (Pekic et al., 1998), 70% acetone/water (Oszmianski & Sapis, 1989) have been used in extraction of phenolic compound from grape seeds. Baker et al. (1995) investigated the effect of different solvents on the extraction of phenolic compounds and reported that 70% acetone was the best mixture solvent for yielding procyanidins and total phenols. In this context, the preliminary experiments revealed that 70% acetone was the best solvent for extraction of phenolic compounds from grape seeds as it afforded a maximum yield of phenolics. The amounts of dry matter and extraction yield of grape seed are shown in Table 2. The extraction yields of grape seed ranged from 10.76 to 23.03% (w/w), being highest for Papaz karasi and lowest for Okuzgozu. Narince variety was found to be the lowest extraction yield (15.90%) among white cultivars. Baydar et al. (2004) reported that acetone:water:acetic acid (90:9.5:0.5) mixture resulted in 15.31% (w/w) extraction yield of grape seed of Narince variety. Our values of extraction yield obtained for Narince variety are quiet similar with the results of Baydar et al. (2004). Moreover, it was shown that the extraction yield of Okuzgozu and Karadimrit varieties have lower than both the coloured and white varieties. The content of phenolic compound of grape seed extracts varieties varied between 33 945 and 58 730 (mg gallic acid eq. per 100 g extract) (Table 3). The total phenolic content of Narince extract was the highest



Table 2 Dry matter and extraction yield of dried grape seed


Dry matter* (g per 100 g of dried grape seed)

Extraction yield (percentage of dry grape seed)

92.69e 95.40a 93.73cd 93.68cd 93.40ed

19.06bcd 16.57cd 19.83abc 18.72bcd 15.90d

94.37bc 93.71cd 93.24ed 94.60ab 93.34ed 93.43ed 93.94bcd

11.90e 18.22bcd 16.42cd 10.76e 19.51bc 20.42ab 23.03a

*Values followed by different superscript letter(s) within each column are significantly different at P £ 0.05.

Table 3 Free radical-scavenging activity of the grape seed extracts for ABTS [2,2/-azinobis (3-ethylbenzothiazoneline-6-sulfonic acid)] and DPPH (2,2-diphenyl-picrylhydrazyl) assays (expressed as mm trolox equivalents)


Total phenolic content*†

TEAC*‡ ABTS

TEAC*‡ DPPH

33 945d 52 216ab 50 150abc 51 532ab 58 730a

2.46e 4.00ab 2.67de 2.81de 3.69abcd

3.55e 4.87bcd 4.95bcd 4.71cd 5.76a

627dc 947ab 137ab 121bc 179ab 434ab 431ab

2.71de 3.89abc 4.14a 3.57abcd 2.97cde 2.99bcde 2.81de

3.65e 4.63d 5.03bcd 4.66cd 5.49ab 4.43d 5.37abc

39 53 54 45 51 53 55

*Values followed by different superscript letter(s) within each column are significantly different at P £ 0.05. † Milligram of gallic acid equivalents per 100 g of extracts. à Millimoles trolox per mg of extracts.

among all cultivars whereas Muskule extract had the lowest. The order of total phenolic content in seeds of twelve different varieties was Narince > Papaz karasi > Alphonse Lavallee´ > Muscat of Hamburg > Alicante Boushet > Razaki > Hasandede > Kalecik karasi > Emir > Okuzgozu > Karadimrit > Muskule. Baydar et al. (2004) reported that the total phenolic content of seed extract of Narince

variety was found to be 66.7%. This result is a higher value. It is interesting to note that in spite of the extraction yield of Narince variety being the lowest among white cultivars, the total phenolic content of Narince variety was found to be the highest. Total phenolic content of Bangalore variety was reported by Jayaprakasha et al. (2003) to be 46% (w/w) (catechin eq.). Many similar studies have been in which the cultivars and environmental conditions affect the phenolic content of grape seeds/fruits (Silva et al., 1991; Prieur et al., 1994; Fuleki & Ricardo-da-Silva, 1997; Revilla et al., 1997). Two in vitro antioxidant assays, DPPH and ABTS, were used as routine ways to assess the potential antioxidant capacity of grape seed extracts. The DPPH and ABTS assays are useful to evaluate the free radical scavenging of water- and non-water-soluble compounds (Llorach et al., 2004). The grape seed extracts showed a high capacity for scavenging both DPPH and ABTS assays. Table 3 shows the scavenging activities of grape seed extract obtained by the DPPH and ABTS methods and the TEAC values. As seen in Table 3, Narince extracts had the highest, and Muskule extracts had the lowest TEACDPPH value. TEACDPPH values of grape seed extracts decreased in the following order: Narince > Kalecik karasi > Papaz karasi > Alphonse Lavallee´ > Emir > Razaki > Hasandede > Okuzgozu > Muscat of Hamburg > Alicante Boushet > Karadimrit > Muskule. This order of the TEACDPPH values was similar to the order of the total phenolic content. Total phenolic content was significantly correlated with TEACDPPH value (r ¼ 0.7974, P £ 0.001). The order of TEACABTS value of grape seed extracts was: Alphonse Lavallee´ > Razaki > Muscat of Hamburg > Narince > Okuzgozu > Alicante Boushet > Kalecik karasi > Papaz karasi ¼ Hasandede > Karadimrit > Emir > Muskule. According to TEA CABTS results, Alphonse Lavallee´ extracts had the highest while Muskule extracts had the lowest antioxidant activity. Moreover, both the total phenolic content and TEACDPPH value of Muskule extracts was the lowest. Although TEACABTS value was significantly correlated with the total phenolic content (P £ 0.05), this correlation is very poor (r ¼ 0.4860). It means that we cannot use the measure of total phenolic content to predict the antioxidant activity expressed as TEACABTS. Many studies on correlations between antioxidant activity and phenolic content have been reported. As a general rule the antioxidant capacity has been positively correlated with phenolic content (Madhujith & Shahidi, 2005; Tsai et al., 2005; Sun & Ho, 2005). In this study, which compares the results of two radical-scavenging tests, on the contrary to Miliauskas et al. (2004), TEACDPPH value was higher than TEACABTS value.



157

158


Also no significant correlation was found between TEACDPPH and TEACABTS values. Madhujith & Shahidi (2005) reported that antioxidant effectiveness must be studied by other methods because antioxidant activity in foods depends on various factors such as polarity, solubility and metal-chelating capacity. For all analyses, difference between white and coloured cultivars was insignificant while difference among cultivars was significant (P < 0.05). Interestingly, the antioxidant activity expressed as TEACABTS of Razaki extracts was found to be higher than Hasandede extracts in spite of the difference between the total phenolic content of Razaki and Hasandede extracts being insignificant among white cultivars. However, the difference between the antioxidant activity expressed as TEACDPPH of Razaki and Hasandede extracts was insignificant while the difference between the antioxidant activity expressed as TEACABTS of Razaki and Hasandede extracts was significant. A similar state was also observed among coloured cultivars. The difference between values of TEACABTS of Alphonse Lavallee´, Kalecik karasi, Alicante Boushet and Papaz karasi extracts was shown to be significant while the difference between the total phenolic content of Alphonse Lavallee´, Kalecik karasi, Alicante Boushet and Papaz karasi extracts was insignificant. Similarly, the difference between the total phenolic content of Kalecik karasi and Alicante Boushet extracts was insignificant while the difference between the antioxidant activity expressed as TEACDPPH of Kalecik karasi and Alicante Boushet extracts was found to be significant. Conclusion

Antioxidative potentials of twelve different grape seed extracts obtained from common varieties grown in Turkey were investigated and a considerable difference was found between the cultivars in terms of the total phenolic content and antioxidant activity. This study also showed that there is a correlation between results of the total phenolic content and two different antioxidant activity methods. High total phenolic content and antioxidant activity were determined in some varieties. Especially, Narince, Alphonse Lavallee´, Papaz karasi, Kalecik karasi and Razaki extracts can be used as supplement and antioxidant additive after studies of stability and interferences. These varieties may be proposed for industrial purposes as important sources of antioxidant phenolics. Acknowledgments

This work was supported by ‘Ankara University Research Fund’ (grant 2005 074 50005 HPD), Turkey. The authors thank Hakan Karaca, MSc, for reviewing this manuscript.


References AOAC (1995). Official Methods of Analysis, 16th edn. Arlington, VA: AOAC International. Baker, J., Bridle, P., Garciaviguera, C. & Kallithraka, S. (1995). Survey of solvents for the extraction of grape seed phenolics. Phytochemical Analysis, 6, 265–267. Baydar, N.G., Ozkan, G. & Sagdic, O. (2004). Total phenolic contents and antibacterial activities of grape (Vitis vinifera L.) extracts. Food Control, 15, 335–339. Brand-Williams, W., Cuvelier, M.E. & Berset, C. (1995). Use of a free radical method to evaluate antioxidant activity. LebensmittelWissenschaft und Technologie/Food Science and Technology, 28, 25–30. Celik, H. (2002). Grape cultivar catalogue. Sunfidan AS Mesleki Kitaplar Serisi. 2. Ankara. Faostat (2004). Agriculture Data. Available at: http://faostat.fao.org/ default.aspx (accessed 20 December 2004). Fuleki, T. & Ricardo-da-Silva, J.M. (1997). Catechin and procyanidin composition of seeds of grape cultivars grown in Ontario. Journal of Agriculture and Food Chemistry, 45, 1156–1160. Guendez, R., Kallithraka, S., Makris, D.P. & Kefalas, P. (2005). Determination of low molecular weight polyhenolic constituents in grape (Vitis vinifera sp.) seed extracts: correlation with antiradical activity. Food Chemistry, 89, 1–9. Jayaprakasha, G.K., Singh, R.P. & Sakariah, K.K. (2001). Antioxidant activity of grape seed (Vitis vinifera) extracts on peroxidation models in vitro. Food Chemistry, 73, 285–290. Jayaprakasha, G.K., Selvi, T. & Sakariah, K.K. (2003). Antibacterial and antioxidant activities of grape (Vitis vinifera) seed extracts. Food Research International, 36, 117–122. Llorach, R., Tomas-Barberan, F.A. & Ferreres, F. (2004). Lettuce and chicory by products as a source of antioxidant phenolic extract. Journal of Agriculture and Food Chemistry, 52, 5109–5116. Madhujith, T. & Shahidi, F. (2005). Antioxidant potential of pea beans (Phaseolus vulgaris L.). Journal of Food Science, 70, S85– S90. Mavadi, D.L. & Salunkhe, D.K. (1995). Toxicological aspects of food antioxidants. In: Food Antioxidants (edited by D.L. Mavadi, S.S. Deshpande & D.K. Salunkhe). p. 267. New York, NY: Marcel Dekker Inc. Miliauskas, G., Venskutonis, P.R. & Van Beek, T.A. (2004). Screening of radical scavenging activity of some medicinal and aromatic plant extracts. Food Chemistry, 85, 231–237. Nakamura, Y., Tsuji, S. & Tonogai, Y. (2003). Analysis of proanthocyanidins in grape seed extracts, health foods and grape seed oils. Journal of Health Science, 49, 45–54. Oomah, B.D., Cardador-Martinez, A. & Loarca-Pina, G. (2005). Phenolics and antioxidative activities in common beans (Phaseolus vulgaris L). Journal of the Science of Food and Agriculture, 85, 935–942. Oszmianski, J. & Sapis, J.C. (1989). Fractionation and identification of some low molecular weight grape seed phenolics. Journal of Agriculture and Food Chemistry, 37, 1293–1297. Pekic, B., Kovac, V., Alonso, E. & Revilla, E. (1998). Study of the extraction of proanthocyanidins from grape seeds. Food Chemistry, 61, 201–206. Prieur, C., Rigaud, J., Cheynier, V. & Moutounet, M. (1994). Oligomeric and polymeric procyanidins from grape seeds. Phytochemistry, 36, 781–784. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M. & RiceEvans, C. (1999). Antioxidant activity applying an improved ABTS radical cation decolourization assay. Free Radical Biology and Medicine, 26, 1231–1237. Revilla, E., Alonso, E. & Kovac, V. (1997). The content of catechins and procyanidins in grapes and wines as affected by agroecological factors and technological practices. Pp. 69–80. Washington, DC: American Chemical Society.



Santos-Buelga, C., Francia-Aricha, E.M. & Escribano-Bailon, M.T. (1995). Comparative flavan-3-ol composition of seeds from different grape varieties. Food Chemistry, 53, 197–201. Shi, J., Yu, J., Pohorly, J.E. & Kakuda, Y. (2003). Polyphenolics in grape seeds –biochemistry and functionality. Journal of Medicinal Foods, 6, 291–299. Silva, R.C., Rigaud, J., Cheynier, V. & Chemina, A. (1991). Procyanidin dimmers and trimers from grape seeds. Phytochemistry, 30, 1259–1264. Simsek, A., Artik, N. & Baspinar, E. (2004). Detection raisin concentrates (pekmez) adulteration by regression analysis method. Journal of Food Composition and Analysis, 17, 155– 163. Singleton, V.L. & Rossi, J.A. (1965). Colorimetry of total phenolics with phosphomolybdic-phosphotungstic acid reagents. American Journal of Enology and Viticulture, 16, 144–158. Sun, T. & Ho, C.T. (2005). Antioxidant activities of buckwheat extracts. Food Chemistry, 90, 743–749.

Teissedre, P.L., Frankel, E.N., Waterhouse, A.L., Peleg, H. & German, J.B. (1996). Inhibition of in vitro human LDL oxidation by phenolic antioxidants from grape and wines. Journal of the Science of Food and Agriculture, 70, 55–61. Tsai, T.H., Tsai, P.J. & Ho, S.C. (2005). Antioxidant and antiinflammatory activities of several commonly used spices. Journal of Food Science, 70, C93–C97. Van den Berg, R., Haenen, G.R.M.M., Van den Berg, H. & Bast, A. (1999). Applicability of an improved trolox equivalent antioxidant capacity (TEAC) assay for evaluation of antioxidant capacity measurements of mixtures. Food Chemistry, 66, 511–517. Yilmaz, Y. & Toledo, R.T. (2004). Major flavonoids in grape seeds: antioxidant capacity of catechin, epicatechin and gallic acid. Journal of Agriculture and Food Chemistry, 52, 255–260. Yilmaz, Y. & Toledo, R.T. (2006). Oxygen radical absorbance capacities of grape/wine industry byproducts and effect of solvent type on extraction of grape seed polyphenols. Journal of Food Composition and Analysis, 19, 41–48.



159


160

Original article Possibility of using near infrared spectroscopy for evaluation of bacterial contamination in shredded cabbage Phunsiri Suthiluk,1 Sirinnapa Saranwong,2 Sumio Kawano,2 Sonthaya Numthuam1 & Takaaki Satake1* 1 Graduate School of Life and Environmental Sciences, University of Tsukuba, Tennodai 1-1-1, Tsukuba, Ibaraki, 305-8572, Japan 2 National Food Research Institute, Kannondai 2-1-12, Tsukuba, Ibaraki, 305-8642, Japan (Received 22 May 2006; Accepted in revised form 25 July 2006)

Summary

The possibility of near infrared (NIR) spectroscopy to measure the amount of bacterial contamination in shredded cabbage was investigated. NIR measurements in the short wavelength region from 700 to 1100 nm were done using two types of saline solutions: one was used to stomach with the samples as the conventional method and the other was used to wash the outer surface of the sample to examine the possibility of a nondestructive method. Partial least squares regression (PLS) was used to develop the equations for bacterial amount. Spectra from the stomacher solution and the washing solution produced similar results. Sufficiently accurate results could be obtained with the bias-corrected standard error of prediction (SEPs) of 0.46 log CFU g)1 for the stomacher solution and 0.44 log CFU g)1 for the washing solution. NIR spectroscopy was clarified to be a rapid and non-destructive method for prediction of bacterial contamination in shredded cabbage.

Keywords

Food safety, fresh-cut vegetables, NIR, non-destructive, total plate count.

Introduction

Production and consumption of fresh-cut vegetables have been growing rapidly in recent years. Commercially available fresh-cut vegetables have become increasingly popular because of their high sensory quality and convenience (Lund, 1989). However, one of the most crucial problems of this product is its short shelf life because of the rapid deterioration caused by micro-organisms. The process of fresh-cut produce such as shredding, cutting or trimming affects the microorganism in several ways. First, it exposes nutritious internal tissue fluids to the micro-organism and thereby accelerates their growth and the product’s spoilage (Bracket, 1992). Second, it provides more surface area on which the micro-organisms can grow. Micro-organism will affect overall product quality consequently; the amount of micro-organism contamination is a very important food safety indicator for ready-to-eat freshcut vegetables. For determination of micro-organisms, conventional microbiological methods such as total plate count and coliform count have been employed. However, these techniques have a major disadvantage as the time needed to obtain the results is 48 h or more. *Correspondent: Fax: +81 29 853 6130; e-mail: [email protected]

Therefore, the system for facilitating a rapid and nondestructive evaluation of micro-organism contamination in fresh-cut products should be developed for precise and real-time quality control and monitoring. Near infrared (NIR) spectroscopy has been widely used in quality evaluation of many agricultural and food products. In place of conventional chemical analyses that usually are destructive, expensive, time-consuming and create a high level of waste, NIR quality evaluation can be non-destructive, less cost, rapid, repeatable and chemical-free. After the establishment of an analytical system using NIR, the analysis is very simple and can be performed by unskilled personnel (Osborne et al., 1993). Screening for micro-organisms by NIR spectroscopy is still in the beginning stages because of the small size of their concentrations. Davies et al. (1987) first attempted to determine the excessive mould content in tomatoes. After that, Sørensen & Jepson (1997) used NIR spectroscopy for detection of Clostridium tyrobutyricum in cheese. Rodriguez-Saona et al. (2001) tried to identify bacterial strains by applying the DESIR method. These works demonstrated the promise of NIR for microbial qualitative analysis. In addition, Lin et al. (2004) also used NIR for the quantitative analysis of determination of microbial contamination in chicken meat. However, in order to develop a practical and stable system, more sophisticated work must be done. The objective of this


Possibility of NIRS for bacterial evaluation P. Suthiluk et al.

study was to clarify the possibility of NIR for rapid and non-destructive evaluation of the amount of bacteria (log TPC value) in shredded cabbage as a representative of fresh-cut vegetables in order to develop a practical and stable system for food safety.

Sample 30 g + sterilized 0.85% saline solution 30 mL

Washed sample

Drained solution

+ sterilized 0.85% saline solution 240 mL and stomach 2 min

Extract B

Materials and methods Liquid portion

12.5 mL

12.5 mL

Samples

Heads of cabbage (Brassica olercea var. capitata) purchased at a local supermarket were used as samples. The heads were transported to a laboratory at the National Food Research Institute and kept in a cold storage room at 5 C until the experiment on the next day. Before shredding, three layers of wrapper leaves were discarded, the heads were divided into halves, and the core portions were removed. A commercially available slicer was used to shred the cabbage into 2-mm slices. To reduce surface contamination and imitate the preparation of fresh-cut vegetables, the thoroughly mixed shredded cabbage was washed under running tap water for 10 min. A colander was used for draining. Thirty-gram samples of the shredded cabbage were placed in gamma-ray-sterilised polyethylene bags and kept in a storage room at 10 oC. Sampling was done at 0, 2, 4 and 6 day(s) after storage. Sample extraction

To measure the NIR spectra, two types of extract solution samples were used. The first one (called ‘Extract A’ hereafter) obtained followed the conventional method (will be described in Extraction procedure) and the second one was obtained from a sterilised 0.85% saline solution that was used to wash only the surface of the sample (called ‘Extract B’ hereafter) to examine the possibility of a non-destructive NIR method. Extraction procedure

The extraction procedure is illustrated in Fig. 1. First, 30 mL of sterilised 0.85% saline solution was added to the shredded cabbage sample that had been kept for a certain duration in a polyethylene bag. The shredded cabbage was mixed with the solution by hand shaking vertically for 1 min and drained. Then, the drained solution (Extract B) was taken and divided into halves (12.5 mL each). One was kept on crushed ice for NIR measurement and the other was kept for a further procedure. For the Extract A experiment, 240 mL of sterilised 0.85% saline solution was added to the shredded cabbage washed as described earlier. This mixture was mixed with a stomacher for 2 min. The liquid portion was taken and divided into halves. Half of this liquid was mixed with the drained solution from surface washing described earlier. Therefore, this

120 mL

+ 12.5 mL

NIR measurement

mixing Extract A

Microbiological analysis

NIR measurement

Figure 1 Sample extraction procedure.

mixture (Extract A) represented the microbiological composition of the cabbage for both the surface and inside. Extract A solution was used for both NIR measurement and microbiological analysis. Spectral acquisition

Ninety of the samples were measured. NIR transmittance spectra were collected in the short wavelength range, from 700 to 1100 nm. NIRSystems6500 with a fiber optic ‘Transmittance Probe’ (FOSS NIRSystems, Silver Springs, MD, USA) and an aluminum block test tube holder were used for the measurement. To acquire the NIR spectra, individual Pyrex-glass test tubes [15(ø) · 105(h) mm] were used as sample cells. Prior to the NIR measurement, sample temperature was controlled by dipping the test tubes containing the samples into a water bath at 25 oC for exactly 20 min. Microbiological analysis

The extracted solutions were serially diluted with sterilised 0.85% saline solution. Inoculation was done onto the Petriflim ‘Aerobic Count Plate’ (3MTM, USA; AOAC Official Methods of Analysis, method 990.12; 3M Microbiology). To allow for the growth of bacteria and numeration, the films were incubated at 35 C for 48 h. The number of bacteria were expressed as log CFU g)1 and used as constituent values for data analysis. Data analysis

The samples were divided to calibration set and validation set according to their constituent values with an odd–even method. The statistical characteristics of the data sets are



161


Table 1 Statistical characteristics for log (total plate count) value of calibration and validation sets

(a)

0.25

Calibration Validation

60 30

2.85 2.85

7.08 6.95

5.56 5.52

1.15 1.14

log CFU g)1

log (1/T)

0.20

Standard Sample set No. Minimum Maximum Mean deviation Unit

Original spectra

0.15

6 days

0.10

4 days

0.05

2 days

0.00

0 day

–0.05 –0.10

(b)

0.04 0.03 0.02 0.01 0.00 –0.01 –0.02 –0.03 –0.04 –0.05 –0.06 –0.07 700

d2log (1/T)

shown in Table 1. The pretreatments of multiplicative scattering correction (MSC), second derivative or normalisation were used. Partial least squares (PLS) regression was used for calculation. The Unscrambler program (CAMO, Oslo, Norway) was used for MSC, second derivative and PLS calculations. The Near Infrared Spectral Analysis Software (NSAS) program (FOSS NIRSystems) was used for normalising.

–0.15 700


NIR spectra

Original [log (1/T)] values of the Extract A from shredded cabbage stored at 10 C for 0, 2, 4 and 6 days are shown in Fig. 2a. The spectra shifted upward with the increase in storage period. Visually, the extracted solutions of shredded cabbage stored for longer period became less transparent because of an increase in suspended particles. The particles possibly leaked from the cabbage cell because of cell wall or cell membrane damage, or might have been caused by the microorganisms that multiplied in the samples. The transparency of the solution altered the light-scattering property in the same way as fat globules in milk (Chen et al., 2002). As the storage period increased, the number of suspended particles increased, and the scattering effect became more severe. To remove the scattering effect, MSC and second derivative (segment size ¼ 10 nm, gap size ¼ 0 nm) were used. Figure 2b shows that the baseline shift was successfully removed with this pretreatment. A strong absorption band because of water, the major component of the extracted solution, could be observed at 962 nm. Flat portions noticed at both edges were caused by the calculating algorithm of The Unscrambler program. Even when the baseline shift was removed, the spectra were still not in a proper condition. As the test tubes used in this experiment were not the same, slight differences in diameter and glass thickness could have caused the differences in spectra measured (Chen et al., 2002). To remove the sample size effect, the MSCsecond derivative spectra were normalised with the d2log (1/T) value of the water band at 962 nm (Kawano et al., 1993). The normalised spectra used for calibration are shown in Fig. 2c.


(c)

1.20 1.00 0.80 0.60 0.40 0.20 0.00 –0.20 –0.40 –0.60 –0.80 700

d2log (1/T)/d2log (1/T 962)

162

750

800

850

900

950

1000

1050

MSC + second derivative spectra

962 750

800

850

900

950

Normalized spectra

750

800

850

1000

1050

962

900

950

1000

1050

Wave length (nm) Figure 2 (a) Original, (b) multiplicative scattering correction (MSC) + second derivative and (c) normalised spectra of the ‘Extract A’ solution from shredded cabbage stored at 10 C for 0, 2, 4 and 6 day(s).

Calibration

The PLS calibrations for the log (TPC) value were carried out using the spectra acquired from the Extract A solutions. To obtain the best result, wavelength regions used for the calibration varied in the same way as those described by Saranwong et al. (2001). Noted from the lowest standard error of calibration (SEC) and bias-corrected standard error of prediction (SEP) values, the full wavelength region from 700 to 1100 nm provided the best result (Table 2), suggesting that the information needed for the log (TPC) determination might be spread over the whole region. The scatter plots between actual and predicted log (TPC) values using the calibration equation developed from the full wavelength region are shown in Fig. 3. By the paired t-test, there



Table 2 The best partial least square calibration and validation results of log (total plate count) value

150

Spectra

Wavelength region (nm)

F

R

SEC

SEP

Bias

Extract A Extract B

700–1100 700–1050

10 10

0.95 0.92

0.36 0.46

0.46 0.44

0.16 )0.03

F, number of factors used in the calibration equation; R, multiple correlation coefficient; SEC, standard error of calibration; SEP, biascorrected standard error of prediction; bias, the average of difference between actual and near infrared values.

Regression coefficient (K)

874

50 0 –50 –100 –150

778 1062

760

–200 700

750

800

8

850 900 950 Wavelength (nm)

1000

1050

Figure 4 Partial least squares (PLS) regression coefficient plots of PLS calibration for predicting log (total plate count) from near infrared spectra of the ‘Extract A’ solution, using the entire wavelength region (700–1100 nm).

7

6

(a)

5

0.30 Original spectra

0.25

6 days

0.20

4 F

log (1/T)

NIR-predicted TPC (log CFU g–1)

968

100

= 10

SEP = 0.46

3

Bias = 0.16

4 days

0.15

2 days

0.10

0 day

0.05 0.00 –0.05

2 2

3

4 5 6 Actual TPC (log CFU g–1)

7

8

–0.10 700

Figure 3 Scatter plots between actual total plate count (TPC) and near infrared spectroscopy-predicted TPC of validation set developed from the ‘Extract A’ solution spectra.

were no significant differences between the actual and the NIR-predicted values, indicating that the developed system was sufficiently accurate. To clarify the calibration structure, the PLS regression coefficient plots, which are the regression coefficients plotted against wavelength, are shown in Fig. 4. Many strong peaks could be observed, indicating the complexity of the calibration model. Water seems to have an important role in the model, as strong coefficients could be noticed at 760 and 968 nm (Osborne et al., 1993). The band assignments of other strong peaks were difficult, as the actual concentration of bacteria in the extract solution was very small. It is possible that the bacterial colonies floating in the sample may have affected the scattering condition of the sample, or the bacteria itself or its by-product(s) may have had some interactions with the water molecules of the extract solution, and thereby caused the differences in the spectra. Consequently, detail experiments should be done to clarify the physicochemical basis of the calibration developed in the future experiment.

d2log (1/T)

(b)

750

800

850

900

950

1000

1050

0.0006 0.0004

MSC + second derivative

0.0002

spectra

0.0000 –0.0002 –0.0004 –0.0006 964

–0.0008 –0.0010 700

750

800

850

900

950

1000

1050

Wavelength (nm) Figure 5 (a) Original and (b) multiplicative scattering correction (MSC) + second derivative spectra of the ‘Extract B’ solution from shredded cabbage stored at 10 C for 0, 2, 4 and 6 day(s). [Correction added after online publication 23 November 2007: changes made to legend for Figure 5].

To expand the application of the NIR from the rapid method to non-destructive evaluation, the spectra of Extract B was used to develop a calibration equation for the log (TPC) value. In this case, unlike that of Extract A, the MSC pretreatment could only work to reduce the interferences (Fig. 5). Extract B did not contain as many suspended particles as Extract A because the stomacher process was not included. The scatter plots for the best calibration



163


result using the wavelength region of 700–1050 nm are shown in Fig. 6. The results obtained were nearly the same as those for the spectra of Extract A, both in terms of wavelength region and accuracy. For the calibration structure, the regression coefficient plots of Extract B eqn are shown in Fig. 7. The calibration structure of Extract B was quite similar to that of Extract A. Dominant peaks could be observed at almost the same positions, including the water bands (768 and 972 nm). By the 95% statistical test (Fearn, 1996), there was no significant difference between the SEPs of the best equations obtained from Extracts A and B, indicating that the system is applicable for

NIR-predicted TPC (log CFU g–1)

8 7 6

non-destructive bacterial determination of shredded cabbage. Conclusions

The NIR measuring system developed in the short wavelength region was capable of prediction of bacterial contamination in shredded cabbage. By the PLS calibration equations that were developed, either from spectra of the saline solution that was used to wash and stomach with the sample or from spectra of the saline solution that was used to wash only the outer surface of the sample, sufficiently accurate SEPs of 0.46 and 0.44 log (CFU g)1) could be obtained. The similarity in the calibration structure of the two equations could be observed by the wavelength region and the importance of water in the equation. Water seemed to have an important role in the calibrations. NIR spectroscopy has great potential as the rapid detection and non-destructive method and this study should be continuing toward a practical and stable system for the evaluation of other fresh-cut vegetables.

5

Acknowledgment

The authors would like to acknowledge and express deep appreciation to the Monbukagakusho scholarship for providing the opportunity to perform this experiment.

4 F

3

= 10

SEP = 0.44 Bias = –0.03

2 3

2

6 5 4 Actual TPC (log CFU g–1)

7

References

8

Figure 6 Scatter plots between actual total plate count (TPC) and

near infrared spectroscopy-predicted TPC of validation set developed from the ‘Extract B’ solution spectra.

4000

Regression coefficient (K)

164

924 2000

0

–2000

776 952

–4000 700

750

800

850

900

950

1000

1050

Wavelength (nm) Figure 7 Partial least squares (PLS) regression coefficient plots of PLS calibration for predicting log (total plate count) from near infrared spectra of the ‘Extract B’ solution, using wavelength region of 700– 1050 nm.


Bracket, R.E. (1992). Shelf stability and safety of fresh produce as influenced by sanitation and disinfections. Journal of Food Protection, 55, 808–814. Chen, J.Y., Iyo, C., Terada, F. & Kawano, S. (2002). Effect of multiplicative scatter correction on wavelength selection for near infrared calibration to determine fat content in raw milk. Journal of Near Infrared Spectroscopy, 10, 301–307. Davies, A.M.C., Dennis, C., Grant, A., Hall, M.N. & Robertsob, A. (1987). Screening of tomato puree for excessive mould content by near infrared spectroscopy: a preliminary evaluation. Journal of the Science of Food and Agriculture, 39, 349–355. Fearn, T. (1996). Comparing standard deviations. NIR News, 7, 5–6. Kawano, S., Fujiwara, T. & Iwamoto, M. (1993). Nondestructive determination of sugar content in Satsuma mandarin using near infrared (NIR) Transmittance. Journal of Japanese Society of Horticultural Science, 62, 465–470. Lin, M., Al-Holy, M., Mousavi-Hesary, M., Al-Qadiri, H., Cavinato, A.G. & Rasco, B.A. (2004). Rapid and quantitative detection of the microbial spoilage in chicken meat by diffuse reflectance spectroscopy (600–1100 nm). Letter of Applied Microbiology, 39, 148–155. Lund, D. (1989). Food processing: from art to engineering. Food Technology, 43, 242–247. Osborne, B.G., Fearn, T. & Hindle, P.H. (1993). Practical NIR Spectroscopy: With Applications in Food and Beverage Analysis, 2nd edn. Essex: Longman Scientific & Technical. Rodriguez-Saona, L.E., Khambaty, F.M., Fry, F.S. & Calvey, E.M. (2001). Rapid detection and identification of bacterial strains by



Fourier transform near-infrared spectroscopy. Journal of Agricultural and Food Chemistry, 49, 574–579. Saranwong, S., Sornsrivichai, J. & Kawano, S. (2001). Improvement of PLS calibration for brix value and dry matter of mango using information from MLR calibration. Journal of Near Infrared Spectroscopy, 9, 287–295.

Sørensen, L.K. & Jepson, R. (1997). Detection of cheese batches exposed to Clostridium tyrobutyricum spoilage by near infrared spectroscopy. Journal of Near Infrared Spectroscopy, 5, 91– 97. 3M Microbiology. Available at: http://www.3m.com/microbiology/ home/products/petrifilm/pp_over.html (accessed on 1 May 2006).



165


166

Original article Production of a dry sausage from African catfish (Clarias gariepinus, Burchell, 1822): microbial, chemical and sensory evaluations Abdullah Oksuz,1 Gulsun Akdemir Evrendilek,2* Muzaffer Seufi Calis1 & Akif Ozeren1 1 Department of Fisheries and Fish Processing, Faculty of Fisheries, Mustafa Kemal University, Tayfur Sokmen Campus, 31034 Alahan-Hatay, Turkey 2 Department of Food Engineering, Faculty of Agriculture, Mustafa Kemal University, Tayfur Sokmen Campus, 31034 Alahan-Hatay, Turkey (Received 31 March 2006; Accepted in revised form 24 August 2006)

Summary

Production of a dry sausage from African catfish and determination of its microbial, chemical and sensory properties during a 70-day storage at both 4 and 22 C were prompted. pH of the samples at 4 and 22 C did not significantly change during the storage (P > 0.05). Moisture content of the samples was 74%, and reduced to 45% at 4 C and to 22% at 22 C. Protein content of the samples was 20.71%, and increased to 42.5% at 4 C and to 57.99% at 22 C. Total lipid content was 4.5%, and increased to 10.98% at 4 C and to 15.68% at 22 C (P < 0.05). Microbial analyses showed that there was a significant reduction in total aerobic mesophilic and psychrotrophilic bacteria, total mould and yeast, total lactic acid bacteria, total Enterobacteriaceae and Staphylococcus aureus counts at both 4 and 22 C (P < 0.05). Samples stored at 4 C had significantly higher sensory ratings than that of the samples stored at 22 C (P < 0.05).

Keywords

African catfish, fish sausage, microbial quality, sensory evaluation, shelf life.

Introduction

With an increase in health concerns, consumption of red meat is diminished and sea foods are getting more popular. Fish meat is a good source of protein (15–20%), vitamins, minerals, carbohydrates and water-soluble components (Londahl, 1981). Because of lack of variability in different food products produced from fish meat and eating habits for fresh fish, fish consumption is less than that of other meat sources. Therefore, the potential of fish meat in the development of food products needs to be explored. Several attempts have been made to produce new food products from various fish meats but African catfish. For example, production of fish ball from croaker fish meat was found to be a successful alternative to red meat fish ball. Sensory analyses of both fried and boiled fish ball received good sensory scores (Shenoy et al., 1975). Bigueras et al. (1985) produced fish ball and sausages from sprat and stored them with and without vacuum packaging. Both products received high sensory acceptance by a selected panel. El Sahn et al. (1990) prepared minced meat from big scale sand smelt *Correspondent: Fax: +90-326-245–5832; e-mail: [email protected]; [email protected]

Atherina boyeri and used them to prepare meat balls, fish flour and pasta which received a high sensory acceptance. Production of African catfish is gaining popularity in some countries (Mu¨ler, 2004) including Turkey. Compared to the amount of production, African catfish is one of the less consumed and less preferred fish as a fresh, grilled or cooked meat source. In order to increase catfish consumption, alternative food products formulated with catfish need to be developed. However, little research has been carried out to the full utilisation of African catfish and published data on value-added African catfish products are limited. Therefore, the objectives of this study were to produce a dry sausage from African catfish and determine its microbial, chemical and sensory properties during a 70-day storage at both 4 and 22 C. Materials and methods

Meat source and ingredients

Fresh African catfish (Clarias gariepinus, Burchell, 1822) from Go¨lbas¸ı Dam (Hatay, Turkey) was transported in ice to laboratory. Fish was cut, gutted, skinned and filleted. Fish fillets were minced with a mechanical mincer (hole size 1.2 mm diameter).


Dry sausage production from African catfish A. Oksuz et al.

Garlic, salt, red pepper, black pepper, cinnamon, allspice, carnation, cumin, ginger, coconut and olive oil were purchased from local markets in Hatay (Turkey). Salicylic acid was purchased from Merck (Darmstadt, Germany). Manufacturing of fish sausage

The formulation used in the production of sausage is given in Table 1. All spices, olive oil and salicylic acid were mixed with minced fish fillet and homogenous batter was obtained. After that, the batter was covered and kept overnight at 4 C. Next day, the batter was stuffed into natural casing (cow intestine) which was softened by soaking in hot water (weight of the coil was 300–400 g) (Cosansu & Ayhan, 2000). Half of the sausage samples were stored at 22 ± 2 C and relative humidity of 85–90% and the other half at 4 ± 2 C and relative humidity of 50–55%. End of the storage time (STi) was decided based on the sensory panel and all the samples were stored for 70 days. The procedure was repeated twice. Preparation of sausage samples for analyses

Three samples for each of the storage temperatures (STa) were taken immediately after stuffing on the 3, 7, 14, 21, 28, 35, 42, 49, 56, 63 and 70th days for analyses of pH, moisture, sensory analysis and microbial count. Total lipid, total protein and ash analyses were performed at the start and end of the shelf-life studies. Chemical analyses of fish sausage samples

Approximately 10 g of the African catfish meat and sausage samples were homogenised with distilled water Table 1 Formulation of fish sausage Amount Ingredients

g

%

Minced catfish meat Garlic Salt Red pepper Black pepper Cinnamon All spice Carnation Cumin Ginger Coconut spice Olive oil Salicylic acid

5500 170 114 110 78.6 78.6 78.6 78.6 78.6 78.6 78.6 45 11

84.61 2.61 1.75 1.70 1.20 1.20 1.20 1.20 1.20 1.20 1.20 0.70 0.17

Total

6500

100


in the ratio of 1:10, and pH value was measured by Orion perpHect logR meter (Model 370, Fisher Scientific, Pittsburgh, PA, USA). Total moisture content of the fish meat and sausage samples were determined by the AOAC sand pan technique (AOAC, 1990). Ash content was performed by the method described by Harbers (1994). The crude protein was estimated according to the Kjeldahl method (AOAC, 1990; Chang, 1994). Bligh and Dyer method (Hanson & Olley, 1963) was used to extract the total lipid content of the fish samples. Microbiological analyses

Twenty-five gram of fish and sausage samples were aseptically sampled, cut into small pieces, mixed with 225 mL of 0.1% peptone water and stomached (Interscience, St. Nom-La-Breteche, France) for 90 s. Total viable mesophilic bacteria count (TMC), total viable psychrotrophilic bacteria count (TPC), total mould and yeast count (TMYC), total viable lactic acid bacteria (LAB) counts, total viable Enterobacteriaceae count (EC) and total viable Staphylococcus aureus count (SA) were performed. Serial dilutions were made with 0.1% peptone water, and corresponded dilutions were plated onto plate count agar (PCA) for total TMC and TPC, potato dextrose agar (PDA) acidified with 10% tartaric acid for TMYC, violet red bile dextrose agar (VRBA) for EC and Baird Parker (BP) agar with egg yolk tellurite enrichment for S. aureus and MRS agar for LAB count. PCA plates for TMC count and VRBA and BP plates were incubated at 35 ± 2 C for 48 h, whereas PDA plates were incubated at room temperature (22 ± 2 C) for 5 days, PCA plates for TPC were incubated at 7 ± 2 C for 10 days and MRS plates were incubated at 30 ± 2 C anaerobically for 72 h. LAB count was performed only at the beginning and end of the study. All analyses were performed in triplicate. PCA, WRBA, PDA, BP and MRS agars, egg yolk tellurite enrichment, peptone and tartaric acid were purchased from Merck. Sensory analysis

Fifty trained sensory panelists from Mustafa Kemal University participated in the sensory evaluation. Before cooking, the sausage samples were analysed for appearance, flavour, vacuum condition (whether or not there is a shrinkage, swelling or burst in casing), filling condition (air bubbles), amount of water leaks to casing, texture, colour and appearance of crosscut. The samples were cooked in a microwave oven (Arcelik, Istanbul, Turkey) for 1 min at 800 W power. Cooked samples were analysed for flavour, colour, texture, homogeneity of meat particles, taste, spices and aftertaste on a 9-point hedonic scale with ‘one’


167


being undesirable and ‘nine’ desirable. For the amount of water leaks into casing, the scale was marked as ‘one’ was no water leak to casing and ‘nine’ was excess amount of water leaks to the casing. Data analysis

Data analysis was performed by Minitab software version 13.20 (Minitab Inc., State College, PA, USA). Log10 transformations were performed on microbial data. Effects of the STa and duration on chemical, physical and sensory properties of the samples were determined by one-way anova and Tukey’s multiple comparison tests at the significance level of 0.05. Samples stored at 4 and 22 C were also compared with each other on the chemical, physical and sensory properties based on two sample t-tests. Multiple linear regression (MLR) models performed to account for variations in 13 sensory attributes of dry sausage as a function of STa, STi, pH, total viable EC, moisture content (MC), total viable bacteria count (PCA), TMYC and SA for the product revealed following equation: Y ¼ b b1 STa þ b2 STi þ b3 pH þ b4 EC þ b5 MC þ b6 PCA þ b7 TMYC þ b8 SA where Y is response variable, b1–b8 are individual coefficients, STe, STi, pH, EC, MC, PCA, TMYC and SA are the explanatory indicators. Results and discussions

Production of African catfish dry sausage was inspired by traditional type Turkish soudjouck. The type of spices used, filling of the samples into natural casing and the shape of the end product were adopted from Turkish soudjouck (Cosansu & Ayhan, 2000). In order to determine the final formulation, initial tests were performed to determine the best-liked formula in sausage production. For these reasons, six different formulas with different amount of spices were tested. In addition, another batch of sausage without addition of any spice but salt was produced to be served to the sensory panel. Based on the sensory analyses, the formula indicated in Table 1 was used to produce the final sausage. The pH value of the African catfish meat was in the range of 6.7–6.8. Because of the addition of salicylic acid, pH value of the batter was reduced to 5.5. During the shelf–life studies at 4 and 22 C, pH of the sausage samples was approximately 5.5. There was no significant change in the pH values of the samples (P < 0.05). In order to observe changes in the pH values of the samples during the storage period, no lactic acid bacteria was added after the batter preparation. Therefore, no


significant change in the pH of the samples during the storage period was observed (P > 0.05). The pH range of the fish sausage was closer to traditional Turkish soudjouck made from sheep or cow meat. In Turkish-type soudjouck, the batter was left for fermentation for 5 days, with the aim of lowering pH of the batter closer to 4.5 (usually 4.8–5) (Kolsarıcı et al., 1993). However, it is possible that after fermentation, pH of the soudjouck samples can increase above 5.0 (Cosansu & Ayhan, 2000). It was reported that the initial pH of the fermented Turkish soudjouck ranged between 6.22 and 6.29, and pH on the 21st day varied from 5.53 to 5.82 (Kayaardı & Gok, 2004). The pH range of the fish sausage samples was closer to that of salchichoń (a Spanish dry cured sausage) which was claimed to be microbiologically safe (Lizaso et al., 1999). Ambrosiadis et al. (2004) reported that traditional Greek sausages had a pH range of 4.67–6.09. The pH range of the fish sausage samples was comparable to that of the traditional Turkish soudjouck, salchichoń and Greek sausages. Moisture content of the samples at both 4 and 22 C were shown in Fig. 1. African catfish meat had a moisture content of 74.45–71.99%. At the start of the shelf-life studies, the sausage samples had a moisture content of 74%. After 7 days of storage, the samples stored at both 4 and 22 C had a moisture content of about 50% (P < 0.05). At the end of the 70th day, the samples stored at 4 C had a moisture content of 45%, and the samples stored at 22 C had a moisture content of 22% (P < 0.05). Closer to the end of the storage period at 22 C, the samples lost moisture and became dry. However, the amount of moisture after the 70-day storage at 4 C was still higher than that of Turkish soudjouck made from sheep meat (Cosansu & Ayhan, 2000). Compared with the soudjouck samples produced from ground beef (Ceylan & Fung, 2000), the fish 80

22C

4C 70

% moisture content

168

60 50 40 30 20 10 0

7

14

21

28

35

42

49

56

63

70

Days Figure 1 Moisture content of the fish sausage samples at 4 and 22 C stored for 70 days.




3.5 TMC-22C TPC-4C

EC-22C EC-4C

TMYC-22C TMYC-4C

SA-22C SA-4C

3.0

Survival log10 cfu g-1

sausage samples stored at 4 C had a similar moisture content; however, the samples stored at 22 C had lower moisture content. Ash content of the African catfish meat was approximately 1–1.5%. Ash content of the sausage samples was 1.45% at both 4 and 22 C. After the 70-day storage, ash content of the samples was 1.43 at 4 C and 1.57 at 22 C, respectively (P > 0.05). The crude protein content of the African catfish meat changed from 15.38% to 18.08%. At the start of the shelf-life studies, protein content of the samples was 20.71% at both 4 and 22 C. After the 70-day storage, these values increased to 42.5% for 4 C and 57.99% for 22 C (P < 0.05). As the moisture content of the samples decreased during the shelf-life studies, there appeared to be an increase in protein content. Total lipid content of the samples at the start of the shelf-life studies for both 4 and 22 C was 4.5%. After the 70-day of storage, total lipid content of the samples was 10.98% at 4 C and 15.68% at 22 C (P < 0.05). Total lipid content of the African catfish muscles changed from 2.31% to 5.25%. As a result of addition of olive oil and spices, lipid content of the samples was higher than that of the catfish muscle. African catfish meat had an initial TPC, TMYC, TEC and S. aureus count of 2.00, 1.98, 2.00 and 0.3 log10 CFU g)1. After batter preparation, these numbers were found to be slightly higher than that of the catfish muscle (P > 0.05). It is possible that, addition of other ingredients during batter preparation had an effect on the increase of the microbial count. During the storage, initial viable bacteria count of the samples, 2.45 log10 CFU g)1, was reduced to 1.2 and 0.9 log10 CFU g)1 at 4 and 22 C by the end of the 70-day storage (P < 0.05), respectively. At the beginning of the shelf-life studies, the samples had TMYC of 2.48 log10 CFU g)1. By the end of the 70 days, this number was diminished to 1.7 log10 CFU g)1 at 4 C. The samples stored at 22 C had the same initial level of TMYC; however, after the 70-day storage, this number was reduced to 0.25 log10 CFU g)1 (Fig. 2). There was a significant reduction in the TMYC of the sausage samples stored at both 4 and 22 C (P < 0.05). The samples stored at both 4 C and 22 C had the same initial level of 2.49 log10 CFU g of EC, and by the end of the 70 day storage study, this number was reduced to 0.15 log10 CFU g (P < 0.05). As for S. aureus, the samples stored at both 4 and 22 C had an initial count of 0.95 log10 CFU g)1. By the end of the 56th day, these numbers were reduced to an undetectable level (P < 0.05). Initial LAB count of the samples at the beginning of the shelf-life study was 1.28 log10 CFU g)1 at both 4 and 22 C, and this number was reduced to 0.6 log10 CFU g)1 at 4 C and an undetectable

2.5 2.0 1.5 1.0 0.5 0.0 0

7

14

21

28

35

42

49

56

63

70

Days Figure 2 Microbial analyses of the fish sausage samples at 4 and 22 C stored for 70. ( ) Total viable Enterobacteriaceae count at 22 C; (*) total viable EC count at 4 C; ( ) total viable mesophilic bacteria count at 22 C; (d) total viable psychrotrophilic bacteria count at 4 C; ( ) total mould and yeast count (TMYC) at 22 C; (|) TMYC at 4 C; (·) total viable Staphylococcus aureus count at 22 C and (_) total viable SA at 4 C.

level at 22 C by the end of the 70-day storage period (P < 0.05). Moisture content of the samples at both 4 and 22 C significantly reduced during the storage period. Similarly, microbial load of the samples also decreased. There was no significant reduction in pH of the samples at both 4 and 22 C. It is possible that reduced moisture content had a positive effect on microbial reduction. Hilmarsdottir & Karmas (1984) stated that intermediate fish products (moisture content of 31.4–49%) are safer in terms of microbial quality and that S. aureus and total viable aerobic bacteria counts of the fish products were greatly reduced by storage at 26 C and decreased moisture content. Fish sausage developed in this study had a similar microbial reduction. The highest microbial count of the samples detected in this study was 2.5 log10 CFU g)1. This number was much lower than that of the threshold value (5 · 106 CFU g)1) used to determine expiry of microbiological shelf life of vacuum-packaged sausages (von Holy et al., 1991; Dykes et al., 1996). At the start of the storage period, the samples stored at both 4 and 22 C had higher sensory values. With an increase in the STi, there was a significant decrease in the sensory values of the samples (P < 0.05). In terms of appearance, flavour, vacuum, filling condition, texture and appearance of crosscut, the raw soudjouck samples had higher sensory values at both 4 and 22 C. The significant reduction in the sensory properties of the soudjouck samples stored at both 4 and 22 C was observed after 14 days of storage. During the 70-day


169

170


storage, in general, the samples at 4 C had higher sensory values than those at 22 C (P < 0.05). There was no significant change in the amount of water leaks to casing during the storage period at both 4 and 22 C (P > 0.05) (Tables 2 and 3). The sensory properties of the cooked samples were similar in that flavour, colour, texture, homogeneity of meat particles, taste, spices and aftertaste ratings for the samples at both 4 and 22 C were very high at the start

of the shelf-life study. With an increase in the STi, there was a significant reduction in the ratings of these sensory properties (P < 0.05). For all the selected attributes, ratings of the cooked samples at 4 C were higher than those of the samples stored at 22 C (P < 0.05) (Tables 2 and 3). Sensory properties of the samples changed with the STi and STa, and the STa was the determining factor at the same period of STi. As the samples stored at 22 C

Table 2 Sensory analyses (mean ± SD) of fish sausage stored at 4 C Days Material

Sensory properties

0

Raw soudjouck

Appearance Flavour Vacuum Filling conditions Water leaks to casing Texture Appearance of crosscut Flavour Colour Texture Homogeneity of meat particles Taste Spices Aftertaste

8.00 8.20 8.80 8.80 1.00

Cooked soudjouck

3 ± ± ± ± ±

0.00a 0.44a 0.44a 0.44a 0.00a

8.40 ± 0.89a 8.60 ± 0.54a 8.80 8.80 9.00 9.00

± ± ± ±

0.44a 0.44a 0.30a 0.00a

8.80 ± 0.44a 9.00 ± 0.00a 9.00 ± 0.00a

7

8.00 8.00 8.20 8.20 1.00

± ± ± ± ±

0.00a 0.70a 0.83ab 0.83a 0.00a

7.20 ± 1.92a 7.80 ± 0.78a 7.80 7.80 8.40 8.80

± ± ± ±

0.64a 0.83a 0.39a 0.44ab

7.40 ± 0.89b 7.80 ± 0.44b 8.00 ± 0.83b

14

8.00 7.80 8.20 8.20 1.00

± ± ± ± ±

0.70a 0.44a 0.44ab 0.44a 0.00a

7.40 ± 0.54a 7.40 ± 1.34a 7.40 7.80 7.40 8.20

± ± ± ±

0.89a 0.83a 1.51ab 0.44b

7.40 ± 0.54b 8.00 ± 0.70b 7.80 ± 0.83b

6.90 7.40 7.40 7.50 1.00

28 ± ± ± ± ±

0.22b 0.54a 0.41bc 0.50b 0.00a

7.40 ± 0.54a 7.50 ± 0.50b 7.60 7.40 6.40 6.80

± ± ± ±

0.54a 0.54b 0.54b 0.44c

7.20 ± 0.44b 7.20 ± 0.44bc 7.20 ± 0.44b

6.37 7.00 7.00 6.62 1.00

42 0.47b 0.00b 0.00c 0.25b 0.00a

± ± ± ± ±

6.50 ± 0.40b 6.50 ± 0.57b 6.50 6.50 6.67 5.87

0.57b 0.57bc 0.50b 0.25d

± ± ± ±

7.25 ± 0.50b 6.62 ± 0.47c 7.00 ± 0.81bc

6.50 6.50 7.00 6.50 1.00

56 ± ± ± ± ±

0.40b 0.57b 0.30c 0.57bc 0.00a

5.75 ± 0.50bc 5.00 ± 1.00bc 5.75 6.50 5.87 5.50

± ± ± ±

0.50b 0.57bc 0.75bc 0.40d

6.75 ± 0.95b 7.00 ± 0.81bc 6.00 ± 0.50c

70

5.37 6.50 7.00 6.00 1.00

± ± ± ± ±

0.47c 0.57b 0.50c 0.00c 0.00a

4.75 ± 0.50c 5.00 ± 0.81c 5.50 5.50 5.35 5.75

± ± ± ±

0.57b 0.50c 0.25c 0.28d

6.00 ± 0.81b 5.25 ± 0.50d 5.00 ± 0.00d

4.75 6.25 6.37 6.37 1.00

± ± ± ± ±

0.28c 0.50b 1.10c 0.49c 0.00a

2.25 ± 0.95d 4.75 ± 0.95c 5.25 5.00 2.75 5.25

± ± ± ±

0.55b 1.15c 0.95d 0.95d

4.25 ± 0.50c 4.00 ± 0.41e 4.50 ± 0.57d

Data in the same row with different superscript alphabets are significantly different (P 6 0.05).

Table 3 Sensory analyses (mean ± SD) of fish sausage stored at 22 C Days Material

Sensory properties 0

Raw soudjouck Appearance Flavour Vacuum Filling conditions Water leaks to casing Texture Appearance of crosscut Cooked Flavour Colour soudjouck Texture Homogeneity of meat particles Taste Spices Aftertaste

8.00 8.20 8.60 8.80 1.00

3 ± ± ± ± ±

0.00a 0.44a 0.54a 0.44a 0.00a

7.20 7.20 7.60 8.00 1.00

7 ± ± ± ± ±

1.30ab 1.09ab 0.51a 1.22ab 0.00a

8.40 ± 0.89a 6.50 ± 0.54b 8.60 ± 0.54a 6.80 ± 0.44b 8.00 8.00 7.80 8.50

± ± ± ±

0.00a 0.00a 0.44a 0.35a

7.50 7.40 7.40 7.90

± ± ± ±

0.35b 0.22b 0.22ab 0.22b

8.82 ± 0.24a 8.10 ± 0.22a 8.80 ± 0.44a 8.50 ± 0.00a 9.00 ± 0.00a 8.50 ± 0.00b

7.10 7.70 8.00 8.40 1.00

14 ± ± ± ± ±

1.02ab 0.27a 1.00a 0.89ab 0.00a

7.40 6.80 7.80 8.40 1.00

28 ± ± ± ± ±

0.41b 0.27b 0.44a 0.22ab 0.00a

6.00 6.75 6.87 7.62 1.00

42 ± ± ± ± ±

0.00c 0.28b 0.25b 0.47bc 0.00a

5.60 ± 0.89b 4.00 ± 0.70c 3.25 ± 0.50c 6.20 ± 0.83bc 6.10 ± 0.41bc 5.75 ± 0.28c 6.90 7.20 7.00 7.60

± ± ± ±

0.22c 0.44bc 0.40ab 0.22b

7.60 ± 0.22b 8.10 ± 0.40a 7.90 ± 0.42c

6.30 6.70 7.20 7.00

± ± ± ±

0.27d 0.27c 0.27ab 0.00c

6.75 6.25 6.87 7.25

± ± ± ±

0.50cd 0.50c 0.25b 0.28cb

7.20 ± 0.27bc 7.00 ± 0.00c 7.50 ± 0.00b 7.50 ± 0.00b 7.70 ± 0.27c 7.50 ± 0.30c

6.00 6.12 7.12 7.37 1.00

56 ± ± ± ± ±

0.81c 0.25c 0.25b 0.25bc 0.00a

3.25 ± 0.50c 5.00 ± 0.00d 6.62 6.12 6.37 6.62

± ± ± ±

0.47cd 0.25c 0.25b 0.47cb

6.62 ± 0.47c 7.37 ± 0.25b 7.37 ± 0.25c

4.87 5.25 6.75 7.25 1.00

70 ± ± ± ± ±

0.25d 0.95c 0.50b 0.50bc 0.00a

4.00 3.33 6.33 7.33 1.00

± ± ± ± ±

0.50d 0.28d 0.57b 0.57bc 0.00a

3.00 ± 0.81c 1.33 ± 0.57d 5.25 ± 0.50cd 2.66 ± 0.57e 5.50 6.50 4.75 6.62

± ± ± ±

0.57e 0.57cb 1.25c 0.47cb

4.75 ± 0.50d 6.75 ± 0.50c 5.75 ± 0.50d

4.33 6.00 1.66 4.33

± ± ± ±

0.57f 0.00c 0.57d 1.52d

3.33 ± 1.15e 4.33 ± 2.08d 3.00 ± 1.00e

Data in the same row with different superscript alphabets are significantly different (P 6 0.05).




Table 4 Multiple linear regression models of measured fish sausage attributes Response variables

Explanatory variables

r2

P

Appearance Flavour Vacuum condition Filling condition Texture Appearance of crosscut Cooked flavour Cooked colour Cooked texture Homogeneity of meat particles Cooked taste Cooked spices Cooked aftertaste

8.55 ) 1.45STa ) 0.76STi + 0.52pH ) 0.865EC ) 0.0085MC ) 0.603PCA + 1.11 TMYC ) 1.16SA 5.93 + 0.29STa ) 0.26STi + 0.41pH + 0.032EC + 0.0057MC ) 0.052PCA ) 0.188TMYC + 0.281SA 1.39 ) 0.128 STa–0.235 STI + 1.12pH + 0.245EC + 0.0223MC + 0.004PCA + 0.034TMYC-1.10SA 12.7 + 2.11STa ) 0.238STi ) 0.79pH + 0.458EC ) 0.0165MC + 0.520PCA ) 1.43TMYC + 3.06SA 6.3 ) 2.42STa ) 1.46STi + 1.49pH ) 0.144EC ) 0.0971MC + 1.11PCA + 1.97TMYC ) 4.17SA 5.3 + 4.12Ta ) 0.428STi ) 0.41pH + 1.79EC ) 0.103MC + 2.79PCA ) 2.52TMYC + 4.31SA

96.8 98.1 95.1 95.3 97.3 93.1