Anal Bioanal Chem (2005) 381: 1145–1151 DOI 10.1007/s00216-004-3010-6
O R I GI N A L P A P E R
O´scar Mun˜iz-Naveiro Æ Raquel Domı´ nguez-Gonza´lez Adela Bermejo-Barrera Æ Jose´ A. Cocho Jose´ M. Fraga Æ Pilar Bermejo-Barrera
Determination of total selenium and selenium distribution in the milk phases in commercial cow’s milk by HG-AAS Received: 27 September 2004 / Revised: 30 November 2004 / Accepted: 7 December 2004 / Published online: 11 March 2005 Springer-Verlag 2005
Abstract A procedure has been developed for determining the selenium in cow’s milk using hydride generation–atomic absorption spectrometry (HG-AAS) following microwave-assisted acid digestion. The selenium distributions in milk whey, fat and micellar casein phases were studied after separating the different phases by ultracentrifugation and determining the selenium in all of them. The detection limits obtained by HG-AAS for the whole milk, milk whey and micellar casein were 0.074, 0.065 and 0.075 lg l1, respectively. The accuracy for the whole milk was checked by using a Certified Reference Material CRM 8435 whole milk powder from NIST, and the analytical recoveries for the milk whey and casein micelles were 100.9 and 96.9%, respectively. A mass balance study of the determination of selenium in the different milk phases was carried out, obtaining values of 95.5–100.8%. The total content of selenium was determined in 37 milk samples from 15 different manufacturers, 19 whole milk samples and 18 skimmed milk samples. The selenium levels found were within the 8.5–21 lg l1 range. The selenium distributions in the different milk phases were studied in 14 whole milk samples, and the highest selenium levels were found in milk whey (47.2–73.6%), while the lowest level was found for the fat phase (4.8–16.2%). A strong correlaO´. Mun˜iz-Naveiro Æ R. Domı´ nguez-Gonza´lez A. Bermejo-Barrera Æ P. Bermejo-Barrera (&) Department of Analytical Chemistry, Nutrition and Bromatology, Faculty of Chemistry, University of Santiago de Compostela, Avda. de las Ciencias s/n, 15782 Santiago de Compostela, Spain E-mail:
[email protected] Tel.: +34-600-942-346 Fax: +34-981-595-012 J. A. Cocho Laboratory of Metabolic and Nutritional Disorders, University Clinical Hospital, 15706 Santiago de Compostela, Spain J. M. Fraga Department of Pediatrics, University Clinical Hospital, 15706 Santiago de Compostela, Spain
tion was found between the selenium levels in whole milk and the selenium levels in the milk components. Keywords Selenium Æ Whole milk Æ Milk whey Æ Fat milk Æ Casein micelles Æ Microwave-assisted acid digestion Æ Hydride generation–atomic absorption spectrometry
Introduction Selenium is a component of selenoproteins, some of which have important enzymatic functions [1]. This essential element has an important role in human health, and its positive effects on human conditions such as infertility [2, 3] and underactive thyroids [4] have been demonstrated [5]. Recent studies on cancer chemo-prevention have attempted to establish how this element affects cancer patients [6–9]. The results from all of these studies showed an inverse relation between the selenium intake and the mortality caused by cancer. These anticarcinogenic properties are known to be an essential part of the enzyme glutathione peroxidase and other reductases [10], which are involved in an antioxidant activity in intracellular reactions. Glutathione peroxidase removes peroxide and other oxidants via the reducing agent glutathione [11]. Selenium is a toxic element at intake levels of more than 400 lg per day [12]. However, this limit is not accepted by all control institutions, some of which suggest higher levels than this as limit values. The recommended daily intake of selenium in adults is about 55 lg [12], but dietary selenium intakes depend on regional selenium availability and the types of food consumed [13]. Nevertheless, twice the levels of the Recommended Dietary Allowance is the suggested intake required to be able to assess the reduction in cancer risk [14]. Selenium is incorporated into organisms through their diet. Seafood and fish are the main sources of selenium, but cereals, meat, nuts, onions, garlic, mush-
1146
rooms and eggs all increase the dietary selenium intake. Foods like fruits, vegetables, milk, milk products and drinks are the poorest selenium sources [15–17]. In areas deficient in selenium, endemic diseases caused by deficiencies in selenium intake were found to occur [18]. For this reason, it is important to analyze the levels of selenium in food. In particular, cow’s milk has a special place in the human diet because it is used from childhood until old age. The selenium content in milk varies substantially depending on the regional conditions, such as the natural selenium level in the soil and water, the levels in cereals, plants, and so on. There are no clear guidelines on what the optimum concentration in milk should be. Aspila [19] has suggested that the selenium concentration in milk for human consumption should be at least 20 lg l1; this amount could potentially supply >10% of our dietary requirement for selenium. However, essential element status is not solely determined by dietary intake; it is also largely influenced by element bioavailability, which depends on the distribution and the chemical form in the food source. Over the past few years, the selenium in cow’s milk has been determined via various analytical techniques, like differential pulse stripping voltammetry [20], double focusing ICP-MS [21], hydride generation– atomic absorption spectrometry (HG-AAS) [22–24], fluorimetry [25, 26], flow injection HG-AAS [27], graphite furnace–atomic absorption spectrometry [28– 30], and hydride generation–atomic fluorescent spectrometry [31–33]. The values reported for selenium in cow’s milk vary significantly; this is due to the the cow’s origin [20, 30], the cow’s feed [27, 30], the soil [29, 30], and the season [26], not the analytical method used. HG-AAS is used in this study to evaluate the selenium levels in cow’s milk from Galicia (in the north-west of Spain) in order to establish a profile of the selenium intake and the differences between whole milk, skimmed milk and milk with different additives; in this paper we report our study of the selenium distribution among the different milk components. Information about the selenium content in cow’s milk (whole milk, skimmed milk and partially skimmed milk) has been reported in literature from Spain before [21, 25, 26, 31–33] as well as in works from other countries, like India [20], Ireland [23], Kuwait [28], Sweden [27], Italy [30], UK [34], Croatia [22] and The Netherlands [35]. The distributions of some trace elements in cow’s milk components have been studied previously [36]. Nevertheless, there is no information available on the selenium distribution and the correlation within the selenium levels in milk components in the literature. Before we can study the bioavailability of selenium in different types of cow’s milk, it is important to know the selenium distribution. Since the food industry is interested in producing new products with extra added nutritional elements, it is
very important to establish simple, precise and accurate analytical methods to evaluate the actual improvements in nutritional value of these products.
Experimental Instrumentation All samples were digested in a 100 ml closed Teflon vessel covered by a safety shield of high-density polypropylene in a segmented mono-block rotor using a microwave Ethos Plus (maximum power 1,000 W, maximum temperature 300 C) with internal temperature control (Milestone S.r.l., Sorisole, BG, Italy). A refrigerated ultracentrifuge L8 Beckman (Alaska, USA) with a SW-40 rotor was used to separate the milk whey, casein micelles and fat. Selenium determinations were carried out using a Perkin Elmer Zeeman 4100 Z atomic absorption spectrometer equipped with a Perkin Elmer FIAS 400 system with a five-port flowinjection valve, manifold and gas–liquid separator used for hydride generation. A selenium electrodeless discharge lamp and an EDL System 2 power supply, both supplied by Perkin Elmer (U¨berlinghen, Germany), were used.
Reagents All reagents were of analytical grade, and ultrapure water of 18 MW cm specific resistivity obtained from a Milli-Q purification system (Millipore Co., Bedford, MA, USA) was used throughout. Glassware, plasticware and polytetrafluoroethylene (PTFE) were soaked in 10% nitric acid for at least 48 h and then rinsed three times with ultrapure water and kept dry ready to use. Nitric acid (70.0% v/v), hydrochloric acid (37% v/v), hydrogen peroxide 35% v/v and urea (all supplied by Panreac, Barcelona, Spain) were used to perform the acid digestion of the samples. Selenium hydride generation was carried out using sodium tetrahydroborate (Aldrich, Milwaukee, WI, USA) and hydrochloric acid (37% v/v) (Panreac). A stock standard solution of selenium, 1,000 g l1 SeO2 in 0.5 M nitric acid (Merck Poole, Dorset, UK) was used to prepare the calibration standards. Reference material 8435 (whole milk powder) from the National Institute of Standards and Technology (NIST) (Gaithersburg, MD, USA) was used to validate the analytical method and the instruments used in the selenium determinations. The cow’s milk samples were bought in Galician supermarkets (Galicia, Spain). These milk products had various compositions, including whole milk, skimmed milk and other products containing various added components like calcium and OMEGA 3.
1147
Fig. 1 The absorbance signal as a function of the volume of urea solution (50%) used
severe signal depression due to their oxidative potential against hydrogen selenide. To avoid this problem, a third step was included in this procedure. The effect of urea concentration on the elimination of nitrogen oxides was studied using 0.05, 0.1, 0.2 and 0.3 ml of urea solution (50%). Although higher levels of urea give a higher absorbance signal, the shape of the signal worsens at the higher levels (Fig. 1). The optimum concentration of urea was chosen as 0.2 ml to remove the excess nitrogen oxides was therefore found to be 0.2 ml. The vessel was then cooled until room temperature and urea was added. Finally, the samples were transferred to volumetric flasks and made up to a final volume of 25 ml with water.
Procedures
Total selenium determination
Sample preparation procedure
The total selenium determination in whole milk and its different components was carried out after pre-treatment of samples by HG-AAS under the optimum conditions summarised in Table 1. The standard addition method was used to determine the selenium concentrations in the samples.
The milk was ultracentrifuged to obtain the three milk phases studied [37]: fat, whey and casein pellets. Milk samples (12 ml) were ultracentrifuged at 160,000·g for 60 min at 4 C with 1 min acceleration and 1 min deceleration times. The different components obtained were the casein micelles and other high relative molecular mass compounds as the precipitate, the whey in the middle of the ultracentrifuge tube, and an upper fat phase; all of them were stored at 20 C until use. The sample pre-treatment was carried out in three steps. In the first step, samples (whole milk and the different components of milk) were digested in triplicate in a microwave oven. The sample (2.5 ml for whole milk and whey milk, and 0.9 g of casein micelles and fat) was placed and weighed in the Teflon vessel. 2.0 ml of nitric acid (70% v/v), 1.0 ml of hydrogen peroxide (35% v/v) and 2.5 ml of ultrapure water were added to each sample to carry out the acid digestion procedure. The samples were heated at 200 C for 10 min in the microwave oven. In the second step, after cooling the vessels down to 50 C, 1.4 ml of hydrochloric acid was added to reduce Se(VI) to Se(IV). Afterwards, the samples were heated at 130 C for 10 min. Nitrite, formed by oxidative decomposition with nitric acid, is an important source of interference at this stage since it forms nitrogen oxides that may cause a
Results and discussion Optimization of significant variables for the selenium hydride generation To optimize the parameters related to the selenium hydride generation, a central composite 23+star orthogonal design with six degrees of freedom, involving 16 experiments [38], was performed. The variables [NaBH4], [HCl] and temperature were optimized for selenium hydride generation. The experimental design matrix and response signal obtained for each experiment using a selenium standard of 5 lg l1 are shown in Table 2. The response surfaces obtained with these optimum values are shown in Fig. 2, indicating that the variables [NaBH4], [HCl] and temperature have clear optimum values in the response surfaces. These correspond, respectively, to a hydrochloric acid concentration of 3.8– 4.2 M, a sodium tetrahydroborate concentration of
Table 1 Selenium hydride generation conditions for HG-AASa Step
1 2
Flow rate (ml min1) Pump 1
Pump 2
Sample
Carrier solution
Reducing solution
10 0
9 9
5 5
Time (s)
Valve
Read
Function
10 20
Fill Inject
*
Sampling Se hydride generation
Wavelength, 196.0 nm; slit width, 0.7 nm; sample loop 500 ll; Ar flow rate 125 ml min1, quartz cell temperature 850 C; measurement mode, peak height; carrier solution HCl, 4 M; reducing solution, NaBH4, 0.2% (m/v)
a
1148 Table 2 Central 23+star orthogonal composite design (n = 11) Run
[NaBH4] (gl1)
[HCl](M)
Temperature (C)
Response signal Abs.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Koa + + + + a* +a* Koa Koa Koa Koa Koa
Kob + + + + Kob Kob a** +a** Kob Kob Kob
Koc + + + + Koc Koc Koc Koc a*** a*** Koc
0.108 0.081 0.051 0.035 0.052 0.036 0.104 0.092 0.087 0.085 0.092 0.047 0.058 0.089 0.067 0.106
The low levels () for [NaBH4], [HCl] and temperature are 0.1%, 2 M and 700C and the high levels (+) are 0.3%, 4 M and 900 C, respectively Ko is the mean value for the variable; ±a is the axial value (high and low) for each variable: Koa=0.2, Kob=3.0, Koc=800, a* = 0.032, a* = 0.368, a** = 2.31, a** = 5.68, a*** = 631.82, a*** = 968.18
0.18–0.22% (w/v), and a temperature of 840–850 C. HCl 4.0 M, NaBH4 0.2% (w/v) and a temperature of 850 C were chosen for selenium hydride generation. Standard calibration and standard addition lines To study the effect of the matrix on the analytical signal, aqueous calibration and standard addition curves were compared (Table 3). There are statistical differences (t-test, 95% confidence level) [38] between the aqueous calibration and standard addition curves obtained with whole milk, milk whey and casein micelles (Table 3), but not between the different slopes from the standard addition curves performed with whole milk, milk whey and casein micelles. Therefore, the matrix effect exists and it is necessary to use the standard addition method to determine the selenium in all of the samples studied.
Fig. 2a–b Estimated response surfaces for a [HCl] and b [NaBH4] for the central composite design
Mass balance study To check the procedure used to obtain the different milk components used in the selenium distribution study, a mass balance study was performed. The selenium levels in the whole milk and the milk whey, casein micelles and fat phases separated by ultracentrifugation were determined in six different cow’s milk samples in triplicate (Table 4). The selenium mass balance for the samples studied was between 95.5 and 100.8%. Because of this, the selenium level in the fat phase was obtained for the other studied samples by calculating the difference between the total selenium level in the whole milk and the sum of the selenium levels in the whey milk and the casein micelles.
Analytical characteristics Sensitivity The detection limit (LOD) and the quantification limit (LOQ), defined as: LOD ¼ 3 SD=m
LOQ ¼ 10 SD=m
where SD is the standard deviation of the blank (n=11) and m is the slope of the standard addition curve, were calculated for the different matrix samples. The values obtained for the LOD were 0.074, 0.065 and 0.075 lg l1, and for LOQ were 0.245, 0.216 and 0.250 lg l1 for whole milk, milk whey and casein micelles, respectively.
Precision The repeatability of the measurements, expressed as the relative standard deviation (RSD), was studied for the whole procedure for each different type of sample. The values obtained were 4.4, 2.7 and 3.3% for the whole milk, the milk whey and the micellar casein phase, respectively.
1149 Table 3 Aqueous calibration and standard addition slopes, t experimental (texp) and t critical (ttheo) after a t-test, 95% confidence level Curve
Slope
Aqueous calibration Standard additionWM Standard additionMWH Standard additionCM
0.0202 0.0139 0.0143 0.0134
Standard additionWM
Standard additionMWH
Standard additionCM
texp
ttheo
texp
ttheo
texp
ttheo
40.54 – 1.92 0.87
2.09 – 2.09 4.30
40.02 1.92 – 1.98
2.09 2.09 – 2.09
43.02 0.87 1.98 –
2.09 4.30 2.09 –
WM, whole milk; MWH, milk whey; CM, casein micelles
Table 4 Mass balance study with cow’s milk samples. The results are expressed as mean±standard deviation (n = 12) Sample
SeWhole
1 2 3 4 5 6
253.9 220.8 241.2 194.4 207.6 218.4
milk
± ± ± ± ± ±
(ng)
6.50 8.30 9.20 9.00 7.00 10.6
SeWhey 133.4 111.5 126.8 103.8 108.2 114.0
milk
± ± ± ± ± ±
(ng)
4.40 3.50 5.40 6.20 4.10 3.80
Accuracy Certified Reference Material 8435 (whole milk powder), with a Selenium Certified content of 0.131±0.014 lg l1, was used to study the accuracy of the procedure. Five replicates of this material were analyzed and the value obtained was 0.129±0.005 lg l1. There was no significant difference between the certified and the found values (t-test, 95% confidence level). Since there are no reference materials for whey milk and casein micelles, the accuracy of the selenium determination in milk whey and casein micelles was estimated using the analytical recovery. The mean values obtained for analytical recovery were 100.9 and 96.9% for milk whey and casein micelles, respectively, within the linear range of concentrations.
SeCasein
micelles
98.2 82.9 93.5 76.3 80.4 85.2
4.90 6.20 6.50 5.20 3.80 3.80
± ± ± ± ± ±
(ng)
19.7 16.0 19.1 16.2 15.6 16.8
99 ± 2 95 ± 2 99 ± 2 101 ± 2 98 ± 1 99 ± 1
± ± ± ± ± ±
2.30 2.50 1.40 2.00 2.80 2.80
Table 5 Selenium concentrations found in 37 commercial milk products—19 whole milks and 18 skimmed milks—from 15 different manufacturers, by HGAAS after microwave-assisted acid digestion. The results are expressed as mean±standard deviation (n=12)
Applications
The selenium concentration was determined in 37 cow’s milk products, 19 whole milk samples and 18 skimmed milk samples by HG-AAS. All samples were digested in quadruplicate and analyzed in triplicate, and the addition method was used during selenium determination. Whole milk and skimmed milk from same manufacturer were analyzed, except for the sample ‘‘Puleva Omega 3’’, that has no commercial skimmed milk product. The results obtained are shown in Table 5. For the whole milk, the Se levels were between 12.2 and 21.0 lg l1, with a mean value of 16.7 lg l1 according to the values found in literature from the last few years [21–31]. For the skimmed milk samples, the selenium range was between 8.5 and 17.2 lg l-1, with a
Mass recovery (%)
mean value of 13.7 lg l1. The differences between whole milk and skimmed milk (for products from the same company) were between 1.1 and 5.2 lg l1, with a mean value of 3.0±1.33 lg l1. In order to prove the existence of a statistically significant difference between the levels of selenium in skimmed milk and whole milk samples from the same company, an one-way ANOVA (95% confidence level) was performed. Statistically significant differences were between all of the samples from the same company, and the same statistical results were
Company/tradename
Total selenium levels in whole milk
SeFat (ng)
Feiraco RAM CLAS Larsa Pascual Lauki Rio Oro del Valle Campobueno Solan President Celta Clesa Champion Puleva Omega 3 RAM Calcio CLAS Calcio Pascual Calcio Lauki Calcio
Selenium concentration (lg l1) Whole milk
Skimmed milk
16.2 18.4 17.3 17.2 20.1 18.2 16.5 16.2 16.9 17.5 12.2 15.6 14.0 21.0 17.2 16.8 13.7 17.7 15.5
14.4 ± 0.7 15.0 ± 0.7 16.2 ± 0.7 16.0 ± 0.5 16.6 ± 0.9 15.1 ± 0.7 14.7 ± 0.5 11.6 ± 0.4 12.5 ± 0.9 15.7 ± 0.6 10.8 ± 0.6 11.2 ± 0.8 12.6 ± 0.6 17.2 ± 0.6 – 12.2 ± 0.9 8.5 ± 0.5 13.8 ± 0.8 12.7 ± 0.6
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.7 0.8 0.6 .7 0.8 0.9 0.5 0.4 0.5 0.4 0.4 0.6 0.6 1.2 0.6 0.9 0.7 0.6 0.4
1150 Table 6 Selenium concentrations in 14 whole milk samples and in different fractions of these samples: milk whey, casein micelles and fat, obtained by HG-AAS after microwave-assisted acid digestion. The results are expressed as mean±standard deviation (n=12) Company/ tradename
Feiraco RAM CLAS Larsa Pascual Lauki Rio Oro del Valle Campobueno Solan President Celta Clesa Champion
Selenium concentration (lg kg1) Whole milk
Milk whey
Casein micelles
Fata
16.2 18.4 17.3 17.2 20.1 18.2 16.5 16.2 16.9 17.5 12.2 15.6 14.0 21.0
8.5 ± 0.4 11.0 ± 0.5 8.1 ± 0.1 9.4 ± 0.2 14.8 ± 0.6 10.8 ± 0.3 8.3 ± 0.3 7.6 ± 0.5 8.4 ± 0.4 9.7 ± 0.9 7.1 ± 0.3 9.0 ± 0.8 8.1 ± 0.5 12.9 ± 0.5
6.0 4.7 7.8 7.0 4.3 4.4 6.1 7.2 7.3 5.9 3.3 5.1 4.3 6.5
1.6 2.7 1.4 0.8 1.0 3.0 2.1 1.3 1.2 1.9 1.8 1.6 1.5 1.6
± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.7 0.8 0.6 0.7 0.8 0.9 0.5 0.4 0.5 0.4 0.4 0.6 0.6 1.2
± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.1 0.1 0.2 0.2 0.1 0.1 0.6 0.2 0.2 0.3 0.1 0.4 0.2 0.1
a The selenium contents in fat samples were calculated as Sefat=Sewhole milkSemilk wheySe casein micelles
found using a Multiple Range Test by using the Fisher test (95% confidence level). The selenium levels in the skimmed milk samples were lower than in the whole milk from the same company. Lower levels of selenium in skimmed milk than in whole milk have also been reported by Martino et al [21] and Cava-Montesinos et al [32, 33]. Selenium distributions in the milk whey, micellar casein and fat phases The selenium distributions in the milk whey, micellar casein and fat phases were studied for 14 whole milk samples. The samples were analyzed in triplicate using a standard addition method. The results obtained are shown in Table 6. The selenium levels in the milk whey were between 7.1 and 14.8 lg l1 (with a mean value of 9.4 lg l1), in the casein micelles they were between 3.3
Fig. 3 Percentage distributions of selenium in the different cow milk components for various milk products
and 7.8 lg l1 (with a mean value of 5.7 lg l1), and in the fat phase they were between 1.2 and 3.2 lg l1 (with a mean value of 1.68 lg l1). The selenium distributions, expressed as percentages, are shown in Fig. 3; the highest selenium levels were found in the whey milk (with selenium percentages between 47.2 and 73.6%), and the lowest percentages (between 4.8 and 16.2%) correspond to the fat phases. Correlations between the selenium levels in whole milk (WH), milk whey (WHM) and casein micelles (CM) were established by applying a linear regression, Z = a+bX+cY, where Z is the selenium content in WM, X is the selenium content in WHM, Y is the selenium content in CM, and a, b and c are constants. The values of a, b and c as well as the correlation coefficient r were obtained by the use of Statgraphics Plus version 5.0 software [38]. The correlation between the different levels of selenium in the milk components and the whole milk is described by the equation: WH ¼ 3:052 þ 0:0978 WHM þ 0:798 CM The correlation coefficient is 0.942 (WM, WHM and CM), which means that there is a good correlation between the selenium contents in whey milk and casein micelles and the selenium content in whole milk.
Conclusions The sample pre-treatment developed in this study was appropriate for determining the Se in whole milk and in different milk components (whey, fat and casein micelles) by HG-AAS. The hydride generation conditions were optimized by experimental design, and the sensitivity of the optimized determination method was adequate for determining Se by HG-AAS in the commercial milk samples and their components. The biggest percentage of Se in cow’s milk samples was found in the whey phase (56.6±6.9%), while the fat component had the lowest Se levels (10.1±3.6%),
1151
according to the results reported in the literature [21]. This is corroborated by our study of whole and skimmed commercial milks, where we found that the selenium level was lower in the skimmed milk than in the whole milk. The differences found between these two types of milk in this study was about 3.0±1.33 lg Se l1, and agree with results found using different analytical methodologies [21, 23, 31, 33]. An important correlation between the selenium levels in the different milk components and the selenium level in the whole milk has been found. Acknowledgements The authors gratefully acknowledge the support of this work by FEIRACO Sociedad Cooperativa Gallega.
References 1. Ko¨hrle J, Brigelius-Flohe R, Bock A, Gartner R, Meyer O, Floe L (2000) Biol Chem 381:849–864 2. Ursini F, Heim S, Kiess M, Maiorino M, Wissing J, Flohe L (1999) Science 285:1393–1396 3. Martin-Romero F, Kryukov G, Lobanov A, Carlson B, Lee B, Gladyshev V, Hatfield L (2001) J Biol Chem 276:29798–29804 4. Schmutzler C, Ko¨hrle J (2000) Eur J Endocrinol 193:15–24 5. Rayman MP (2000) Lancet 356:233–241 6. Clark LC, Coms GF, Turnbull BW, Slate EH, Chalker DK, Chow J, Davis LS, Glover RA, Graham GF, Gross EG, Krongrad A, Lesher JL, Park HK, Sanders BB, Smith CL, Taylor JR (1996) J Am Med Assoc 276:1957–1963 7. Spallholz JE, Palace VP, Reid TW (2004) Biochem Pharmacol 67:547–554 8. Corcoran NM, Najdovska M, Costello AJ (2004) J Urology 171:907–910 9. Whanger PD (2004) Brit J Nutr 91:11–28 10. Fridovich I (1974) Adv Enzymol RAMB 41:35–97 11. Maddipati KR, Marnett LJ (1987) J Biol Chem 262:17398– 17403 12. Monsen ER (2000) J Am Diet Assoc 100:637–640 13. Schrauzer GN, White DA (1978) Bioinorg Chem 8:303–318 14. Combs G (2001) Nutr Cancer 40:6–11
15. Oster O, Prellwitz (1989) Biol Trace Elem Res 20:1–14 16. Matek M, Blanusa M, Grgic J (2000) Eur Food Res Technol 210:155–160 17. Maihara VA, Gonzaga IB, Silva VL, Favaro DIT, Vasconcellos MBA, Cozzolino SMF (2004) J Radioanal Nucl Chem 259:465–468 18. Yang FY, Lin ZH, Li SG, Guo BO, Yin YS (1998) J Trace Elem Elect H 2:157–163 19. Aspila P (1991) J Agr Sci Finland 63:1–74 20. Inam R, Somer G (2000) Food Chem 69:345–350 21. Martino FAR, Sanchez MLF, Sanz-Medel A (2001) Anal Chim Acta 442:191–200 22. Klapeck T, Mandic´ ML, Grgic´ J, Primorac Lj, Perl A, Krstanovic´ V (2004) Food Chem 85:445–452 23. Murphy J, Cashman KD (2001) Food Chem 74:493–498 24. Bermejo P, Barciela J, Pen˜a EM, Bermejo A, Fraga JM, Cocho JA (2001) J Anal Atom Spectrom 16:188–193 25. Rodrı´ guez-Rodrı´ guez EM, Sanz-Alejos M, Diaz-Romero C (1999) J Agr Food Chem 47:1520–1524 26. Rodrı´ guez-Rodrı´ guez EM, Sanz-Alejos M, Diaz-Romero C (2001) J Food Comp Anal 14:419–430 27. Lindmark-Ma˚nsson H, Fonde´n R, Pettersson HE (2003) Int Dairy J 13:409–425 28. Al-Awadi FM, Srikumar TS (2001) J Trace Elem Exp Med 14:57–67 29. Aleixo PC, No´brega JA (2003) Food Chem 83:457–462 30. Licata P, Trombetta D, Cristani M, Giofre´ F, Martino D, Calo M, Naccari F (2004) Environ Int 30:1–6 31. Cava-Montesinos P, Cervera ML, Pastor A, De la Guardia M (2003) Anal Chim Acta 481:291–300 32. Cava-Montesinos P, Cervera ML, Pastor A, De la Guardia M (2004) Talanta 62:175–184 33. Cava-Montesinos P, Cervera ML, Pastor A, De la Guardia M (2004) J Anal Atom Spectrom 19:696–699 34. Givens DJ, Allinson R, Cotrill B, Blake JS (2004) J Sci Food Agric 84:811–817 35. Shen L, Van Dael P, Deelstra H (1993) Z Lebensm Unters For 197:342–345 36. Bermejo P, Domı´ nguez R, Bermejo A (1997) Talanta 45:325– 330 37. Bermejo P, Domı´ nguez R, Bermejo A (1997) Fresen J Anal Chem 357:457–461 38. Manugistics (1992) Reference manual for Statgraphics Plus version 5.0. Manugistics, Inc., Rockville, MD
