Enrichment Broth for the Detection of Campylobacter ... - Allen Press

1842 Journal of Food Protection, Vol. 80, No. 11, 2017, Pages 1842–1850 doi:10.4315/0362-028X.JFP-16-450 Copyright Ó, International Association for Food Protection

Research Paper

Enrichment Broth for the Detection of Campylobacter jejuni and Campylobacter coli in Fresh Produce and Poultry YOUMI JO,1†‡ HYE-MIN OH,2† YOHAN YOON,2 SUN-YOUNG LEE,3 JI-HYOUNG HA,1§ WON-IL KIM,1 HWANG-YONG KIM,1 SANGHYUN HAN,1 AND SE-RI KIM1* 1Microbial Safety Team, Department of Agro-Food Safety & Crop Protection, National Institute of Agricultural Sciences, Rural Development Administration, Wanju 55365, Republic of Korea; 2Department of Food and Nutrition, Sookmyung Women’s University, Seoul 04310, Republic of Korea; and 3Department of Food Science and Technology, Chung-Ang University, Anseong 47262, Republic of Korea

MS 16-450: Received 15 October 2016/Accepted 12 April 2017/Published Online 9 October 2017

ABSTRACT Although campylobacteriosis caused by Campylobacter jejuni and Campylobacter coli has been increasingly reported worldwide owing to the consumption of contaminated poultry and fresh produce, the current detection protocols are not selective enough to inhibit unspecific microbes other than these pathogens. Five antibiotics were separately added to Bolton broth, and the survival rates of 18 Campylobacter spp. and 79 non-Campylobacter spp. were evaluated. The survival rate of the nonCampylobacter spp. was the lowest in Bolton broth with rifampin (6.3%), followed by cefsulodin (12.7%), novobiocin (16.5%), and potassium tellurite and sulfamethozaxole (both 17.7%). Also the most effective concentration of rifampin was found to be 12.5 mg/L, which markedly inhibited non-Campylobacter strains while not affecting the survival of Campylobacter strains. After the Campylobacter spp. were enriched in Bolton broth supplemented with 12.5 mg/L rifampin (R-Bolton broth), CampyFood Agar (CFA) was found to be better in selectively isolating the pathogens in the enrichment broth than the International Organization for Standardization method of using modified charcoal cefoperazone deoxycholate agar (mCCDA) for this step. When applied to natural food samples—here, romaine lettuce, pepper, cherry tomato, Korean leek, and chicken—the R-Bolton broth–CFA combination decreased the number of false-positive results by 50.0, 4.2, 20.8, 50.0, and 94.4%, respectively, compared with the International Organization for Standardization method (Bolton broth–mCCDA combination). These results demonstrate that the combination of R-Bolton broth and CFA is more efficient in detecting C. jejuni and C. coli in poultry and fresh produce and thus should replace the Bolton broth–mCCDA combination. Key words: Campylobacter coli; Campylobacter jejuni; Enrichment broth

Campylobacter jejuni and Campylobacter coli cause campylobacteriosis, which generally manifests as gastroenteritis but may be implicated in severe complications such as Guillain-Barré syndrome (19). The number of reported C. jejuni and C. coli outbreaks has been growing worldwide, with a higher incidence than salmonellosis and shigellosis in many regions (2, 11). According to the Centers for Disease Control and Prevention (CDC) annual report (12), C. jejuni and C. coli infections in the United States increased by 13% in 2013 compared with 2006 to 2008, resulting in approximately 845,000 cases of illness. Avian species are the common hosts for C. jejuni and C. coli, and a large proportion of campylobacteriosis cases is associated with poultry consumption (19). Although * Author for correspondence. Tel: (82) 63-238-3395; Fax: (82) 63-2383840; E-mail: [email protected]. † These authors contributed equally to this work. ‡ Present address: Agromaterial Assessment Division, Department of Agro-Food Safety & Crop Protection, National Institute of Agricultural Sciences, Rural Development Administration, Wanju 55365, Republic of Korea. § Present address: Hygienic Safety and Analysis Center, World Institute of Kimchi, Gwangju 61755, Republic of Korea.

consuming poultry that is not adequately cooked is still considered a major risk factor for campylobacteriosis, outbreaks associated with fresh or ready-to-eat produce should not be ignored because of the increasing consumption of fresh produce (28, 31). Fresh fruits and vegetables could become contaminated by coming into contact with sources that contain C. jejuni and C. coli at any stage of production, pre- and postharvest. Chai et al. (6) reported that the average prevalence of C. jejuni and C. coli in fresh vegetables was 29.4 to 67.7%. Potential sources of preharvest contamination include workers’ unwashed and soiled hands, animal feces, irrigation water, and inadequately composted manure (31). Yoo et al. (32) observed that C. jejuni and C. coli infections after consuming fresh fruits and vegetables were more likely when poorly treated poultry manure was used as a fertilizer. These data necessitate finding a method to determine the presence and population levels of C. jejuni and C. coli in fresh produce. The isolation of C. jejuni and C. coli from a raw food commodity is not an easy task because the levels of these pathogens are usually low while the other bacterial

J. Food Prot., Vol. 80, No. 11

competitors of the natural microflora are dominant in that niche (4, 17). To detect C. jejuni and C. coli in food, the International Organization for Standardization (ISO) (21) recommends the use of Bolton broth to enrich the Campylobacter spp. and then modified charcoal cefoperazone deoxycholate agar (mCCDA) as a subsequent selective medium. Both Bolton broth and mCCDA contain antibiotics, including sodium cefoperazone, trimethoprim, vancomycin, cyclohexamide, and amphotericin B, to suppress any unspecific growth of contaminating microbes. Nevertheless, the combined use of Bolton broth and mCCDA is not enough to inhibit some non-Campylobacter strains, such as Acinetobacter spp., Ochrobacterium spp., Pseudomonas spp., and Escherichia coli; this prevents microbiology laboratories from choosing this protocol for the accurate detection of C. jejuni and C. coli (25, 32). Several researchers have worked on modifying the enrichment step to improve the selectivity of the protocol. Chon et al. (9) developed charcoal cefoperazone–polymyxin B–deoxycholate broth, which they determined to be superior to Bolton broth when isolating C. jejuni and C. coli from chicken carcass rinsate. Chon and his colleagues (10) also reported that the supplementation of Bolton broth with triclosan inhibited cefoperazone-resistant microflora, such as extended-spectrum b-lactamase (ESBL)–producing E. coli, in the cefoperazone-based enrichment broth. To date, no studies have been published on optimizing Bolton broth, a selective enrichment broth that could inhibit the unspecific growth of Acinetobacter spp., Ochrobactrum spp., and Pseudomonas spp., or on testing newly introduced enrichment broths for the detection of C. jejuni and C. coli in fresh produce and poultry. Therefore, the purpose of this study was to determine a better detection protocol for C. jejuni and C. coli in fresh produce and poultry by modifying Bolton broth and combining it with a subsequent selective medium.

MATERIALS AND METHODS Bacterial strains and growth conditions. This study was performed using nine C. jejuni strains (CCARM 13111, CCARM 13125, CCARM 13156, CCARM 13324, CCARM 13360, CCARM 13362, KCCM 41772, KCCM 41773, and NCCP 15742) and nine C. coli strains (CCARM 13010, CCARM 13063, CCARM 13188, CCARM 13203, CCARM 13220, CCARM 13249, CCARM 13252, NCCP 11191, and NCCP 11192). Also, used in this study were 79 non-Campylobacter strains (other foodborne pathogens and frequently isolated microflora found in produce); these are listed in Table 1. These strains were obtained from the American Type Culture Collection (ATCC; Manassas, VA), the Korean National Culture Collection for Pathogens (NCCP; Osong, Chungbuk, Korea), the Korean Agricultural Culture Collection (KACC; Wanju, Jeonbuk, Korea), the Korean Culture Center of Microorganisms (KCCM; Seoul, Korea), and the Culture Collection of Antimicrobial Resistant Microbes (CCARM; Seoul, Korea). All strains were stored in 20% glycerol at 808C. Before we conducted each experiment, C. jejuni and C. coli strains grown on mCCDA (Oxoid, Basingstoke, Hampshire, UK) were separately inoculated into 2 mL of tryptic soy broth (TSB; BD Difco Sparks, MD) supplemented with 5% lysed horse blood (Oxoid) in a 24-well microtiter plate (SPL, Gyeonggi, Korea). Then, the cultures were incubated at 428C for

ENRICHMENT BROTH FOR C. JEJUNI AND C. COLI

1843

48 h under microaerobic conditions (5% O2, 10% CO2, and 85% N2) using a CO2 incubator (Galaxy 170 R, Eppendorf, Hamburg, Germany). Some non-Campylobacter strains (e.g., E. coli, Salmonella spp., Cronobacter spp., Staphylococcus aureus, and Shigella spp.) were grown in 2 mL of TSB in a 24-well microtiter plate at 378C or 308C for 24 h. Screening antibiotics added to Bolton broth. To improve the selectivity of the Bolton broth (Oxoid), we tested separately the addition of five antibiotics: rifampin (10 mg/L), potassium tellurite (2.5 mg/L), cefsulodin (10 mg/L), novobiocin (20 mg/L), and sulfamethoxazole (50 mg/L; all from Sigma-Aldrich, St. Louis, MO). Before being added to the Bolton broth, a 1,0003 stock solution of each antibiotic was prepared by dissolving it in dimethyl sulfoxide (Sigma-Aldrich) and filtering the solution through a 0.2-lm syringe filter (MACHEREY-NAGEL, Düren, Germany). Each antibiotic was added to Bolton broth, autoclaved (Apoza, New Taipei City, Taiwan) at 1218C for 15 min, and cooled at room temperature. We added 2 mL of each Bolton broth– antibiotic combination to separate 24-well microtiter plates. Thus, including the control, six different test sets were prepared. We added 10 lL of each strain of the 18 Campylobacter spp. and 79 non-Campylobacter spp. to duplicate wells of the 24-well microtiter plate containing the six Bolton broth types and incubated them under microaerobic conditions at 428C for 48 h in a CO2 incubator. We then streaked a loopful of each enriched culture onto plates containing mCCDA and incubated the plates under microaerobic conditions at 428C for 48 h. The morphology of each colony on the mCCDA plates was observed after incubation. Optimization of the rifampin concentration. To determine the optimal concentration of rifampin for detecting Campylobacter spp. and suppressing non-Campylobacter spp., we prepared nine supplemented Bolton broths, each with 5% lysed horse blood and one of nine final rifampin concentrations: 0, 2.5, 5, 7.5, 10, 12.5, 15, 17.5, and 20 mg/L. We placed 2 mL of each of the Bolton broths into 24-well microtiter plates. We then added 10-lL cultures of the 18 Campylobacter spp. (9 C. jejuni and 9 C. coli strains) and 21 non-Campylobacter spp. (7 Acinetobacter spp., 4 Acinetobacter calcoaceticus, 2 Ochrobacterium anthropi, 3 Ochrobacterium intermedium, 4 Pseudomonas aeruginosa, and 1 multidrugresistant E. coli strains) individually to the 24-well microtiter plates. The inoculated plates were incubated under microaerobic conditions at 428C for 48 h in a CO2 incubator. We then streaked a loopful of each enriched culture onto mCCDA plates and incubated the plates under microaerobic conditions at 428C for 48 h. The morphology of each colony on the mCCDA plates was observed after incubation. Growth of C. jejuni and C. coli cultures in Bolton broth containing various concentrations of rifampin. To evaluate the effects of the different rifampin concentrations on the survivability of C. jejuni and C. coli, we determined their growth rates using three C. jejuni strains (NCCP11211, NCCP15742, and CCARM13085) and three C. coli strains (CCARM13165, CCARM13010, and CCARM13020). Each C. jejuni and C. coli strain was cultured in 10 mL of TSB with 5% lysed horse blood under microaerobic conditions at 428C for 48 h. The cultures were then centrifuged at 2,000 3 g for 10 min at 48C. The cells were washed twice with 10 mL of phosphate-buffered saline (PBS; Oxoid) and resuspended in 10 mL of PBS. Each culture cocktail of C. jejuni and C coli was prepared by blending three strains of each group. These cell suspensions were serially diluted 1:10 until they

1844

JO ET AL.


TABLE 1. Non-Campylobacter spp. tested during the screening of additional antibiotics and selective agars Bacteria tested

Sourcea

Culture temp

Escherichia coli O157:H7 E. coli, nonpathogenic Salmonella Typhimurium Salmonella Enteritidis S. bongori Cronobacter sakazakii Staphylococcus aureus Shigella boydii S. flexneri S. sonnei Listeria monocytogenes Acinetobacter spp. A. calcoaceticus Brevundimonas diminuta Myroides odoratimimus Ochrobacterium anthropic O. intermedium Pseudomonas aeruginosa P. asplenii P. chlororaophis P. cichorii P. corrugata P. fluorescens P. putida P. reactants P. syringae Aeromonas salmonicida Chryseobacterium balustinum Citrobacter freundii Clavibacter michiganesis Dickeya zeae Enterobacter aerogenes E. agglomerans E. cloacae E. faecalis Erwinia mallotivora Hafnia alvei Klebsiella pneumoniae Kocuria kristinae K. rosea Leclercia spp. Morganella morganii Pantoea stewartii P. agglomerans P. ananatis Pectobacterium carotovorum Providencia vermicola P. stuartii Rahnella aguatilis Serratia entomophila S. ficaria S. plymuthica S. rabidaea

ATCC 13890, ATCC 43889, ATCC 43894 Blackberry, lettuce, Korean leek, sprout ATCC 13311 ATCC19946 ATCC 43975 ATCC 29544, barley, fried pepper ATCC 23235, ATCC 25923, ATCC 27664 ATCC 8700 ATCC 12022 ATCC 25931 ATCC19113, ATCC 19116, ATCC19115 Cabbage, green bean sprout, red cabbage, Korean leek (2), lettuce (2) Sprout, Chinese cabbage, chicory (2) Cucumber Green bean sprout Korean leek (2) Lettuce, onion, Korean leek Lettuce, pepper, tomato, KACC 10185 KACC10152 KACC10150 KACC10137 KACC10141 KACC10003 KACC10266 KACC10099 KACC12128 KACC14791 Perilla leaf KACC11404 Lettuce Chicory KACC13732 KCCM41584 KACC10526 KACC11304 KACC10548 CCARM 13 Lettuce Carrot Tomato KACC14327 KACC13822 KACC10540 KACC10054 KACC10059 KACC10057 KACC12350 KACC10188 KACC12325 KACC12108 KACC12322 Perilla leaf Perilla leaf

378C

a

308C

ATCC, American Type Culture Collection; KACC, Korean Agricultural Culture Collection; KCCM, Korean Culture Center of Microorganisms; CCARM, Culture Collection of Antimicrobial Resistant Microbes.

each reached 104. Using 1 mL of each final diluted cell suspension of C. jejuni and C. coli, we inoculated 50 mL of Bolton broth containing 5% lysed horse blood and rifampin to a final concentration of 0, 7.5, 10, or 12.5 mg/L. The inoculated broths

were incubated under microaerobic conditions at 428C for 48 h and withdrawn at 6-h intervals. We serially diluted 1 mL of each broth with 9 mL of 0.1% peptone water. Then, we spread 0.25 mL aliquots of each dilution onto mCCDA plates and incubated the


plates under microaerobic conditions at 428C for 48 h. The colonies on each mCCDA plate were manually counted. Determination of the selective agar for enhancing selectivity. To enhance the accuracy of Campylobacter detection, six selective agars were assessed: mCCDA, Preston agar (Oxoid), Abeyta-Hunt agar (MBcell, Seoul, Korea), CampyFood agar (CFA; Biomerieux, l’Etoile, France), Blaser-Wang agar (Oxoid), and Brilliance CampyCount agar (Oxoid). We placed 2-mL aliquots of each of three enrichment broths, i.e., TSB, Bolton broth containing 5% lysed horse blood, and Bolton broth containing 5% lysed horse blood and 12.5 mg/L rifampin (RBolton broth), into 24-well microtiter plates. We then added 10 lL of each strain of the 18 Campylobacter spp. and 79 nonCampylobacter spp. separately to duplicate wells of the 24-well microtiter plate containing one of the three enrichment broths, and we incubated the plates under microaerobic conditions at 428C for 48 h in a CO2 incubator. Each culture was grown in each of the three different enrichment broths and was then streaked onto the six selective agars. The plates were incubated under microaerobic conditions at 428C for 48 h. The morphology of the colonies on the six selective media was observed after incubation, and the positive colonies on each medium were streaked onto each selective medium to reconfirm them. Validation of the selectivity of the combination of RBolton broth and two selective agars using food. The performance of the combinations of R-Bolton broth and two selective agars, mCCDA and CFA, for detecting C. jejuni and C. coli was assessed using food samples: lettuce, green peppers, cherry tomatoes, Korean leeks, and whole chickens. The vegetable samples were selected on the basis of their different surface morphologies and pH levels. The chickens and fresh produce were purchased from a local supermarket in Wanju, Jeollabuk-do Province (Korea). To validate the combination of R-Bolton broth and selective agars using vegetables, we tested both inoculated and noninoculated vegetables because of the low detection rate of Campylobacter spp. in produce. For the artificially contaminated vegetable samples, lettuce (n ¼ 6), pepper (n ¼ 6), and cherry tomato (n ¼ 6), we used C. jejuni (NCCP 11211, NCCP 15742, and CCARM 13085) and C. coli (CCARM 13165, CCARM 13010, and CCARM 13020). Each strain was inoculated into 10 mL of TSB containing 5% lysed horse blood; the enriched culture was incubated at 428C for 48 h in CO2 incubator. Then the enriched cultures were centrifuged at 2,000 3 g at 48C for 10 min. The cells were washed twice with 10 mL of PBS and resuspended in PBS. Each culture cocktail of C. jejuni and C. coli was prepared by blending three strains of each group. The cell suspensions were serially diluted 1:10 to 103 CFU/mL. We then applied 1 mL of the bacterial cell suspensions to 25 g of the produce using a micropipette. The noninoculated vegetables, including 24 lettuce, 24 Korean leek, 24 green pepper, and 24 cherry tomato samples (n ¼ 96), were used to assess the performance of the combination of R-Bolton broth and each selective agar. In the vegetables experiment, the U.S. Food and Drug Administration Bacteriological Analytical Manual procedure (20) was taken as a basic method, with modifications. We placed 25 g of each vegetable sample in separate sterile plastic bags containing 100 mL of Bolton broth with 5% lysed horse blood and either 0 or 12.5 mg/L rifampin. The Korean leek and lettuce samples were homogenized using a stomacher (Bagmixer 400VW, Interscience, Paris, France) for 2 min at low speed. The pepper and tomato samples were hand shaken for 2 min at 120 times per min to reduce


1845

the release of organic acid or other antimicrobial materials naturally present in plant tissue. After the samples were enriched at 428C for 48 h under microaerobic conditions, the enriched samples were streaked onto mCCDA and CFA plates using a sterile disposable loop. The plates were then incubated under microaerobic conditions at 428C for 48 h, and typical colonies were chosen on the mCCDA and CFA plates. The typical morphology of Campylobacter spp. appears gray or creamy gray on mCCDA and orange-red on CFA. We confirmed suspected colonies using PCR with the Campylobacter Triplex Detection Kit (Kogene Biotech, Seoul, Korea) and 16S rRNA sequencing (Macrogen, Seoul, Korea). In the poultry experiment, the detection of C. jejuni and C. coli was performed in compliance with the U.S. Department of Agriculture, Food Safety and Inspection Service (30) method, with modifications. Briefly, the chicken carcasses were rinsed with 500 mL of buffered peptone water (Oxoid) and gently shaken for 2 min to ensure even distribution. We then combined a 25-mL aliquot of the rinsate with 25 mL of 23 Bolton broth containing 5% lysed horse blood and rifampin to a final concentration of 0, 10, or 12.5 mg/L; the cultures were then incubated under microaerobic conditions at 428C for 48 h. Next, we streaked a loopful of each enrichment culture onto mCCDA and CFA plates. After the plates were incubated under microaerobic conditions at 428C for 48 h, the suspected colonies was confirmed using PCR using the Campylobacter Triplex Detection Kit and 16S rRNA sequencing. Statistical analysis. The growth tests of C. jejuni and C. coli cultures in Bolton broth containing various concentrations of rifampin were repeated twice, with three samples per experiment. Prior to the statistical analysis, we transformed the microbiological data to log CFU per milliliter. We analyzed the data using the analysis of variance (ANOVA) procedure in SPSS ver. 11 (SPSS Inc., Chicago, IL).

RESULTS Comparison of the antibiotics used to supplement Bolton broth. To determine the optimal supplementary antibiotic, we evaluated the effects of five specific antibiotics on the growth of 79 non-Campylobacter spp. and 18 Campylobacter spp. None of Bolton broths supplemented with any of the five antibiotics affected the growth of the Campylobacter spp. Table 2 shows the survival rates of the non-Campylobacter spp. grown in the control and in Bolton broths supplemented with each of the five antibiotics: rifampin, potassium tellurite, cefsulodin, novobiocin, and sulfamethoxazole. The survival rate of the non-Campylobacter spp. was the lowest in the Bolton broth supplemented with rifampin (6.3%), followed by the Bolton broth supplemented with cefsulodin (12.7%), with novobiocin (16.5%), and with potassium tellurite and sulfamethozaxole (both 17.7%). Although all the Acinetobacter spp., except for A. calcoaceticus, were suppressed by all the antibiotics tested, there was little difference in the survival of P. aeruginosa in the control and the media supplemented with potassium tellurite, novobiocin, and sulfamethozaxole. However, on supplemented mCCDA the inhibitory effects of both rifampin and cefsulodin on P. aeruginosa were observed. The growth of O. intermedium was not well inhibited by any of the antibiotics tested, except for

1846

JO ET AL.


TABLE 2. Survival rates of Campylobacter spp. and non-Campylobacter spp. grown in Bolton broth supplemented with five antibioticsa Bolton broth supplemented with:

Strain

Campylobacter spp. Non-Campylobacter spp.

Bolton broth

Rifampin (10 mg/L)

Potassium tellurite (2.5 mg/L)

Cefsulodin (10 mg/L)

Novobiocin (20 mg/L)

Sulfamethoxazole (50 mg/L)

100 (18/18) 29.1 (23/79)

100 (18/18) 6.3 (5/79)

100 (18/18) 17.7 (14/79)

100 (18/18) 12.7 (10/79)

100 (18/18) 16.5 (13/79)

100 (18/18) 17.7 (14/79)

2/4 2/2

4/4 2/2 3/3

4/4 2/2 3/3

4/4 2/2 3/3

4/4 2/2 3/3

1/4

1/4 4/4

1/4

1/4 3/4

1/4 4/4

Non-Campylobacter spp. detected in mCCDA plates for each enrichment broth Acinetobacter spp. (n ¼ 7) 7/7 A. calcoaceticus (n ¼ 4) 4/4 O. anthropic (n ¼ 2) 2/2 O. intermedium (n ¼ 3) 3/3 Brevundimonas diminuta (n ¼ 1) 1/1 E. coli (n ¼ 4) 1/4 P. aeruginosa (n ¼ 4) 4/4 Myroides ordoratimimus (n ¼ 1) 1/1 a

Values are percentage of strains tested and number of strains/total number of strains tested. mCCDA, modified charcoal cefoperazone deoxycholate agar.

rifampin. These results indicate that rifampin is the optimal selective agent for better and precise detection.

Optimization of the rifampin concentration. The growth of the non-Campylobacter strains was obviously inhibited in the Bolton broths supplemented with .12.5 mg/ L rifampin (Table 3). The exception was the E. coli isolated from the sprouts, which were only partially suppressed even at higher rifampin concentrations of 15, 17.5, and 20 mg/L. Unlike the background bacteria, the 18 Campylobacter spp. actively survived concentrations of rifampin up to 20 mg/L

(data not shown). The Campylobacter spp. cultured in Bolton broth supplemented with rifampin concentrations .12.5 mg/L showed relatively slower growth on mCCDA medium compared with those cultured in R-Bolton broth (12.5 mg/L rifampin). Thus, we chose 12.5 mg/L as the lowest concentration of rifampin to add, considering the inhibition of non-Campylobacter spp. growth and possibility of Campylobacter spp. growth.

Campylobacter spp. growth in Bolton broth supplemented with rifampin. Because rifampin effectively

TABLE 3. Growth of 21 non-Campylobacter strains in Bolton broth supplemented to several concentrations of rifampina Rifampin concn (mg/L): Strain

Source

0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

20.0

Acinetobacter spp.

Cabbage Green bean sprout Red cabbage Korean leek Korean leek Lettuce Lettuce Sprout Chinese cabbage Chicory Chicory Korean leek Korean leek Lettuce Onion Korean leek Lettuce Pepper Tomato KACC 10185 Sprout

þ þ þ þ þ þ þ þ þþ þ þ þþþ þþþ þþ þþ þþþ þþ þ þþ þ þþþ

þ þ þ þþ þ þ þþþ þþþ þþ þþ þþþ þ þþ þþþ

þ þ þþþ þþþ 6 6 6 6 6 þþþ

þ þþþ þþ 6 6 6 þþþ

þ þþþ þþþ 6 þþ

þþ

þþ

þþ

þþ

A. calcoaceticus

O. anthropi O. intermedium

P. aeruginosa

E. coli a

þþþ, strong active growth; þþ, active growth; þ, growth; 6, weak growth; , no growth.

1847



TABLE 4. Survival rates of 79 non-Campylobacter spp. grown on selective agars after culturing in one of three enrichment brothsa Selective agar: Enrichment broth

mCCDA

Preston agar

Abeyta-Hunt

Wang agar

CFA

BCA

TSB Bolton broth R-Bolton broth

16.5 (13/79) 13.9 (11/79) 2.5 (2/79)

12.7 (10/79) 5.1 (4/79) 3.8 (3/79)

44.3 (35/79) 27.8 (22/79) 22.8 (18/79)

20.3 (16/79) 10.1 (8/79) 6.3 (5/79)

8.9 (7/79) 5.1 (4/79) 1.3 (1/79)

15.2 (12/79) 10.1 (8/79) 5.1 (4/79)

a

Values are percentages with number of strains detected/total number of strains tested in parentheses. mCCDA, modified charcoal cefoperazone deoxycholate agar; CFA, CampyFood agar; BCA, Brilliance CampyCount agar; TSB, tryptic soy broth; R-Bolton broth, Bolton broth supplemented with 12.5 mg/L rifampin.

inhibited the growth of the background flora, it was necessary to confirm that rifampin did not impede the growth of C. jejuni and C. coli. Therefore, we determined the growth rates of C. jejuni and C. coli in Bolton broths supplemented with four rifampin concentrations: 0, 7.5, 10, and 12.5 mg/L. The ANOVA indicates that no significant deleterious effect of rifampin was observed for C. jejuni and C. coli (P . 0.05; data not shown). These results imply that C. jejuni and C. coli are resistant to rifampin; therefore, it is feasible to use rifampin as an agent to improve the accuracy of detecting these bacteria using Bolton broth supplemented with 12.5 mg/L rifampin.

Efficiency of selective agars after enrichment in RBolton broth. The type of enrichment broth used had a decisive effect on suppressing non-Campylobacter spp. on selective agars. R-Bolton broth improved the inhibition of the competing bacteria on all the selective media we tested (Table 4).

The comparison of several selective agars for the efficient detection of C. jejuni and C. coli is presented in Table 4. CFA showed the lowest survival rate for the background flora of all the tested enrichment broth conditions, indicating the superior suppression ability of CFA over the other selective agars. After the Campylobacter spp. were enriched in R-Bolton broth, the nonCampylobacter spp. showed the lowest survival rate on CFA (1.3%), followed by mCCDA (2.5%) and Preston agar (3.8%).

Validation of the combination of R-Bolton broth and selective agar in food. Table 5 shows the assessment of the combination of R-Bolton broth with each of six selective agars to detect C. jejuni and C. coli from artificially contaminated produce. The isolation of the targeted bacteria was improved by supplementing the enrichment broth with 12.5 mg/L rifampin (i.e., R-Bolton broth).

TABLE 5. Evaluation of a combination of R-Bolton broth and CFA for the detection of C. jejuni and C. coli in artificially contaminated producea Sample

Inoculated with C. jejuni Romaine lettuce (n ¼ 6)

Pepper (n ¼ 6)

Cherry tomato (n ¼ 6)

Inoculated with C. coli Romaine lettuce (n ¼ 6)

Pepper (n ¼ 6)


a

b

Treatment

Presumptive-positive samples

Confirmed positive samplesb

Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA

50 83 83 83 100 100 83 100 100

(3/6) (5/6) (5/6) (5/6) (6/6) (6/6) (5/6) (6/6) (6/6)

0 100 100 100 100 100 100 100 100

(0/3) (5/5) (5/5) (5/5) (6/6) (6/6) (5/5) (6/6) (6/6)

Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA

67 100 100 83 100 100 83 100 100

(4/6) (6/6) (6/6) (5/6) (6/6) (6/6) (5/6) (6/6) (6/6)

50 100 100 100 100 100 80 100 100

(2/4) (6/6) (6/6) (5/5) (6/6) (6/6) (4/5) (6/6) (6/6)

Values are percentages with number of positive samples/total number of samples in parentheses. mCCDA, modified charcoal cefoperazone deoxycholate agar; R-Bolton broth, Bolton broth supplemented with 12.5 mg/L rifampin; CFA, CampyFood agar. Confirmed using PCR and/or 16S sequencing.

1848

JO ET AL.


TABLE 6. Evaluation of a combination of R-Bolton broth and CFA for the detection of C. jejuni and C. coli in noninoculated produce and poultrya Sample

Produce Romaine lettuce (n ¼ 24)

Pepper (n ¼ 24)


Korean leek (n ¼ 24)

Poultry Chicken (n ¼ 36)

a

b

Treatment

Presumptive-positive samples

Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA

75.0 25.0 25.0 4.2 4.2 0 20.8 0 0 70.8 16.7 20.8

Bolton broth mCCDA R-Bolton broth mCCDA R-Bolton broth CFA

75.0 (27/36) 47.2 (17/36) 50 (18/36)

Confirmed positive samplesb

(18/24) (6/24) (6/24) (1/24) (1/24) (0/24) (5/24) (0/24) (0/24) (17/24) (4/24) (5/24)

0 0 0 0 0 0 0 0 0 0 0 0

(0/18) (0/6) (0/6) (0/1) (0/1) (0/0) (0/5) (0/0) (0/0) (0/17) (0/4) (0/5)

0 (0/27) 64.0 (11/17) 94.4 (17/18)

Values are percentages with number of positive samples/total number of samples in parentheses. mCCDA, modified charcoal cefoperazone deoxycholate agar; R-Bolton broth, Bolton broth supplemented with 12.5 mg/L rifampin; CFA, CampyFood agar. Confirmed using PCR and/or 16S sequencing.

When we assessed the Campylobacter spp. using the ISO standard method of Bolton broth and mCCDA, the occurrences of C. jejuni and C. coli were 50% (3 of 6 samples) and 67% (4 of 6), respectively, for romaine lettuce; however, our confirmation tests (PCR and 16S rRNA sequencing) showed that all the presumptive-positive samples of C. jejuni were false positives (0 of 3 samples confirmed) and that only 50% (2 of 4) of the presumptivepositive samples of C. coli were confirmed as C. coli. In contrast, Campylobacter spp. were efficiently recovered from the pepper, cherry tomato, and romaine lettuce when we used R-Bolton broth with mCCDA or CFA. When we isolated Campylobacter spp. from noninoculated vegetable samples using the ISO standard method of Bolton broth and mCCDA, the percentages of presumptive-positive samples were 75.0% for romaine lettuce, 4.2% for pepper, 20.8% for cherry tomato, and 70.8% for Korean leek (Table 6). When we used R-Bolton broth and mCCDA or CFA, the presumptive-positive percentages were 25.0% for romaine lettuce, 0 or 4.2% for pepper, 0% for cherry tomato, and 16.7 or 20.8% for Korean leek. Nevertheless, none of the presumptivepositive samples derived from the three methods (Bolton broth and mCCDA, R-Bolton broth and mCCDA, and RBolton broth and CFA) were confirmed as Campylobacter spp. using PCR and 16S rRNA sequencing. These results show that the combination of R-Bolton broth and either selective agar considerably inhibited the growth of the background microflora in fresh produce. When we isolated Campylobacter spp. from chicken using Bolton broth and mCCDA, the percentage of presumptive-positive samples was 75.0%. However, all these presumptive Campylobacter strains were identified as E. coli at the final stage. When we isolated Campylobacter

spp. from chicken using R-Bolton broth and CFA or mCCDA, the percentage of presumptive-positive samples was 47.2 or 50%, respectively. Moreover, of these presumptive Campylobacter spp., 64% on mCCDA and 94.4% on CFA were confirmed as Campylobacter jejuni using 16S rRNA sequencing and PCR. These results show that the R-Bolton broth–CFA combination is the most efficient method.

DISCUSSION The use of an enrichment broth in the detection of C. jejuni and C. coli in food samples is necessary not only to multiply the target cells but also to suppress competing bacteria. Previous studies (16, 22) demonstrated that Bolton broth is insufficient to detect C. jejuni and C. coli from several food matrices, including chicken and vegetables. The four antibiotics contained in Bolton broth, especially cefoperazone, contribute to its selectivity (27). Cefoperazone is also used in mCCDA, a selective agar recommended by the ISO protocol. However, the use of the same selective agent in both Bolton broth and mCCDA may lead to decreased selectivity (22). To improve the accuracy of detection of Campylobacter spp., various modifications have been suggested, such as changes in the combination of selective agents and the incubation temperatures (4). Although numerous selective media have been developed for the detection of C. jejuni and C. coli, the presence of non-Campylobacter contaminants can still lead to falsepositive results. Thus, it is necessary to establish a better protocol by modifying Bolton broth, supplementing it with specific antibiotics for the detection of C. jejuni and C. coli in vegetables and poultry. In this study, we identified rifampin as a suitable antibiotic to reduce false positives by creating a synergistic


effect on the inhibition of non-Campylobacter spp. In other words, the growth of background bacteria was markedly decreased by the addition of rifampin to Bolton broth (Table 2). Rifampin, one of the selective agents used in other Campylobacter-selective media, acts by binding the bsubunit of RNA polymerase (5). In contrast, cefoperazone (found in both Bolton broth and mCCDA) interferes with cell wall synthesis by suppressing peptidoglycan cross-link formation (26). Because rifampin and cefoperazone have different mechanisms of action, the addition of rifampin to Bolton broth could result in improved selectivity. Furthermore, the differences in the antibacterial spectra of rifampin and cefoperazone would influence the efficiency of Bolton broth supplemented with rifampin. This study determined the optimal amount of rifampin to add as a selective agent. As the level of rifampin concentration increased, fewer background bacteria were found. When the concentration of rifampin reached 12.5 mg/ L, most of the non-Campylobacter spp. we tested were successfully inhibited (Table 3). However, E. coli from the sprouts survived even at 20 mg/L rifampin, showing resistance to the selective material. There was no statistically significant difference among the growths of C. jejuni or C. coli cultured in Bolton broth supplemented with 0, 7.5, 10, and 12.5 mg/L rifampin (P . 0.05; data not shown). Campylobacter spp. are intrinsically resistant to rifampin, and the resistance is mediated by the CmeABC, a multidrug efflux system, unlike other bacteria, in which resistance is conferred by mutation in the DNA-dependent RNA polymerase (15, 24). Other research (14) revealed that spoT plays an important role in resistance to this antibacterial agent by mounting a stringent response. Accordingly, 12.5 mg/L rifampin, as the lowest effective concentration, seems best suited to be incorporated into Bolton broth for the isolation of Campylobacter spp. Several studies have compared the selectivity of selective agars using various food samples. Krause et al. (23) showed that Preston agar and mCCDA have similar selectivity; in contrast, Chon et al. (7, 8) determined that the selectivity of Preston agar is superior to that of mCCDA. Our findings support the results of Ugarte-Ruiz et al. (29), who reported that Preston agar and CFA showed higher selectivity than mCCDA following enrichment of the sample in Bolton broth. In the present study, the inhibition of the growth of nonCampylobacter spp. was significantly improved in the most of the selective agars, except for Preston agar, when they were supplemented with rifampin (Table 4). Because Preston agar includes rifampin, any additional effects in this agar would rarely be displayed. With CFA and mCCDA, the improved selectivity resulted from the distinct antimicrobial agents having different spectra. When R-Bolton broth was used as the enrichment medium, the recovery rate of Campylobacter spp. from deliberately contaminated produce increased. In contrast, when we used the standard ISO method (Bolton broth and mCCDA), all the presumptive positives from the romaine lettuce samples were false positives (Table 5). The low selectivity of the standard method resulted from unidentified competitors, such as Acinetobacter spp., Ochrobacterium


1849

spp. and E. coli. In addition, when we used R-Bolton broth and CFA on food samples (romaine lettuce, chicken, tomato, and Korean leek), the number of false positives decreased compared with the standard method. Many studies (25, 32) investigating the detection of Campylobacter spp. have verified that the background microbial flora contain Pseudomonas spp., E. coli, Acinetobacter spp., and Ochrobacterium spp.. Similarly, these bacteria obstructed the detection of Campylobacter spp. in our study. Acinetobacter spp. and Ochrobacterium spp., in particular, showed up as contaminants in the fresh produce; however, the growth of these contaminants was inhibited by enriching the sample in R-Bolton broth, followed by streaking it onto CFA. In contrast, using the standard method all the presumptive-positive bacteria samples (27 of 36) isolated from chicken were later identified as E. coli; that is, none (0 of 27) were confirmed to be a Campylobacter spp. (Table 6). Moreover, some reports demonstrated the low selectivity of the isolation method using Bolton broth and mCCDA. In one study (3), 41 of 96 vegetable samples tested positive for C. jejuni when the isolation method (Bolton broth and mCCDA) was used, but all 41 were false positives (0% C. jejuni), as confirmed using PCR and 16S rRNA sequencing (3). Also, Ugarte-Ruiz et al. (29) announced that, when their samples were cultured in Bolton broth, 23% were positive and 98.2% of those samples were confirmed as positive according to a qPCR analysis. In this study, most of the false positives for Campylobacter spp. organisms in chicken, when using either the ISO standard method or the R-Bolton broth and mCCDA method, turned out to be E. coli when they were analyzed. The presence of E. coli in the chicken samples is considered to make the analysis of Campylobacter spp. difficult. Recently, it has often been reported (1, 13) that ESBLproducing E. coli, which degrades the cefoperazone in Bolton broth and mCCDA, is highly prevalent in poultry. Also, Hazeleger et al. (18) stated that, when C. jejuni and ESBL-producing E. coli were cocultured in Bolton broth, the C. jejuni showed a severe reduction in growth while the E. coli actively proliferated. Thus, ESBL-producing E. coli is able to overgrow in the Bolton broth and mCCDA plate, complicating the isolation of Campylobacter spp. (16, 22). Therefore, even though the exact mechanism of and interaction between R-Bolton and CFA need additional study, our results suggest that enriching the samples in RBolton broth followed by culturing on CFA can prohibit the growth of ESBL-producing E. coli, inducing advanced selectivity. In summary, the data presented here suggest that the combined use of R-Bolton broth and CFA in detecting Campylobacter spp. demonstrates improved specificity and, in turn, fewer false-positive results when applied to food samples compared with the ISO standard method.

ACKNOWLEDGMENTS This study was carried out with the support of the Research Program for Agricultural Science & Technology Development (Project No. PJ009533) and the National Institution of Agricultural Science, Rural Development Administration, Republic of Korea.

1850

JO ET AL.

REFERENCES 1. Allen, H. K. 2014. Antibiotic resistance gene discovery in foodproducing animals. Curr. Opin. Microbiol. 19:25–29. 2. Altekruse, S. F., N. J. Stern, P. I. Fields, and D. L. Swerdlow. 1999. Campylobacter jejuni—an emerging foodborne pathogen. Emerg. Infect. Dis. 5:28–35. 3. Bae, Y., Y. Hong, D. Kang, S. Heu, and S. Lee. 2011. Microbial and pathogenic contamination of ready-to-eat fresh vegetables in Korea. Korean J. Food Sci. Technol. 43:161–168. 4. Baylis, C. L., S. MacPhee, K. W. Martin, T. J. Humphrey, and R. P. Betts. 2000. Comparison of three enrichment media for the isolation of Campylobacter spp. from foods. J. Appl. Microbiol. 89:884– 891. 5. Campbell, E. A., N. Korzheva, A. Mustaev, K. Murakami, S. Nair, A. Goldfarb, and S. A. Darst. 2001. Structural mechanism for rifampin inhibition of bacterial RNA polymerase. Cell 104:901–912. 6. Chai, L. C., T. Robin, U. M. Ragavan, J. W. Gunsalam, F. A. Bakar, F. M. Ghazali, S. Radu, and M. P. Kumar. 2007. Thermophilic Campylobacter spp. in salad vegetables in Malaysia. Int. J. Food Microbiol. 117:106–111. 7. Chon, J.-W., J.-Y. Hyeon, I.-S. Choi, C.-K. Park, S.-K. Kim, S. Heo, S.-W. Oh, K.-Y. Song, and K.-H. Seo. 2011. Comparison of three selective media and validation of the VIDAS Campylobacter assay for the detection of Campylobacter jejuni in ground beef and fresh-cut vegetables. J. Food Prot. 74:456–460. 8. Chon, J.-W., J.-Y. Hyeon, J.-H. Park, K.-Y. Song, and K.-H. Seo. 2012. Comparison of 2 types of broths and 3 selective agars for the detection of Campylobacter species in whole-chicken carcass-rinse samples. Poult. Sci. 91:2382–2385. 9. Chon, J.-W., H. Kim, J.-H. Yim, J.-H. Park, M.-S. Kim, and K.-H. Seo. 2013. Development of a selective enrichment broth supplemented with bacteriological charcoal and a high concentration of polymyxin B for the detection of Campylobacter jejuni and Campylobacter coli in chicken carcass rinses. Int. J. Food Microbiol. 162:308–310. 10. Chon, J.-W., Y.-J. Kim, H.-S. Kim, D.-H. Kim, H. Kim, K.-Y. Song, and K.-H. Seo. 2014. Supplementation of Bolton broth with triclosan improves detection of Campylobacter jejuni and Campylobacter coli in chicken carcass rinse. Int. J. Food Microbiol. 181:37–39. 11. Cover, K. E., S. A. Ruiz, and A. S. Chapman. 2014. Reported gastrointestinal infections in the U.S. Air Force, 2000–2012. MSMR 21:2–7. 12. Crim, S. M., M. Iwamoto, J. Y. Huang, P. M. Griffin, D. Gilliss, A. B. Cronquist, M. Cartter, M. Tobin-D’Angelo, D. Blythe, K. Smith, S. Lathrop, S. Zansky, P. R. Cieslak, J. Dunn, K. G. Holt, S. Lance, R. Tauxe, O. L. Henao, and the Centers for Disease Control and Prevention (CDC). 2014. Incidence and trends of infection with pathogens transmitted commonly through food—Foodborne Diseases Active Surveillance Network, 10 U.S. sites, 2006–2013. Morb. Mortal. Wkly. Rep. 63:328–332. 13. Gao, L., J. Hu, X. Zhang, L. Wei, S. Li, Z. Miao, and T. Chai. 2015. Application of swine manure on agricultural fields contributes to extended-spectrumb-lactamase producing Escherichia coli spread in Tai’an, China. Front. Microbiol. 6:313. doi:10.3389/fmicb.2015. 00313. 14. Gaynor, E. C., D. H. Wells, J. K. MacKichan, and S. Falkow. 2005. The Campylobacter jejuni stringent response controls specific stress survival and virulence-associated phenotypes. Mol. Microbiol. 56:8– 27. 15. Guo, B., J. Lin, D. L. Reynolds, and Q. Zhang. 2010. Contribution of the multidrug efflux transporter CmeABC to antibiotic resistance in different Campylobacter species. Foodborne Pathog. Dis. 7:77– 83. 16. Habib, I., M. Uyttendaele, and L. De Zutter. 2011. Evaluation of ISO 10272:2006 standard versus alternative enrichment and plating combinations for enumeration and detection of Campylobacter in chicken meat. Food Microbiol. 28:1117–1123.


17. Hayashi, M., S. Kubota-Hayashi, T. Natori, T. Mizuno, M. Miyata, S. Yoshida, J. Zhang, K. Kawamoto, K. Ohkusu, S. Makino, and T. Ezaki. 2013. Use of blood-free enrichment broth in the development of a rapid protocol to detect Campylobacter in twenty-five grams of chicken meat. Int. J. Food Microbiol. 163:41–46. 18. Hazeleger, W. C., W. F. Jacobs-Reitsma, and H. M. W. Den Besten. 2016. Quantification of growth of Campylobacter and extended spectrum b-lactamase producing bacteria sheds light on black box of enrichment procedures. Front Microbiol. 7:1–9. 19. Humphrey, T., S. O’Brien, and M. Madsen. 2007. Campylobacters as zoonotic pathogens: a food production perspective. Int. J. Food Microbiol. 117:237–257. 20. Hunt, J. M., C. Abeyta, and T. Tran. 2001. Campylobacter, chap. 7. In Bacteriological analytical manual (BAM). U.S. Food and Drug Administration, Silver Spring, MD. Available at: https://www.fda. gov/Food/FoodScienceResearch/LaboratoryMethods/ucm072616. htm. 21. International Organization for Standardization. 2006. Microbiology of food and animal feeding stuffs—horizontal method for detection and enumeration of Campylobacter spp.—part 1: Detection method. ISO 10272-1. International Organization for Standardization, Geneva. 22. Jasson, V., I. Sampers, N. Botteldoorn, F. Lo´ pez-Ga´lvez, L. Baert, S. Denayer, A. Rajkovic, I. Habib, L. De Zutter, J. Debevere, and M. Uyttendaele. 2009. Characterization of Escherichia coli from raw poultry in Belgium and impact on the detection of Campylobacter jejuni using Bolton broth. Int. J. Food Microbiol. 135:248– 253. 23. Krause, M., M. H. Josefsen, M. Lund, N. R. Jacobsen, L. Brorsen, M. Moos, A. Stockmarr, and J. Hoorfar. 2006. Comparative, collaborative, and on-site validation of a TaqMan PCR method as a tool for certified production of fresh, Campylobacter-free chickens. Appl. Environ. Microbiol. 72:5463–5468. 24. Lin, J., L. O. Michel, and Q. Zhang. 2002. CmeABC functions as a multidrug efflux system in Campylobacter jejuni. Antimicrob. Agents Chemother. 46:2124–2131. 25. Line, J. E., J. S. Baley, and M. E. Berrang. 2008. Addition of ulfamethoxazole to selective media aids in the recovery of Campylobacter spp. from broiler rinse. J. Rapid Methods Autom. Microbiol. 16:2–12. 26. Matsubara, N., S. Minami, M. Matsuhashi, M. Takaoka, and S. Mitsuhashi. 1980. Affinity of cefoperazone for penicillin-binding proteins. Antimicrob. Agents Chemother. 18:195–199. 27. Moran, L., C. Kelly, M. Cormican, S. McGettrick, and R. H. Madden. 2011. Restoring the selectivity of Bolton broth during enrichment for Campylobacter spp. from raw chicken. Lett. Appl. Microbiol. 52:614– 618. 28. Taylor, E. V., K. M. Herman, E. C. Ailes, C. Fitzgerald, J. S. Yoder, B. E. Mahon, and R. V Tauxe. 2013. Common source outbreaks of Campylobacter infection in the USA, 1997–2008. Epidemiol. Infect. 141:987–996. 29. Ugarte-Ruiz, M., S. Gómez-Barrero, M. C. Porrero, J. Alvarez, M. Garc´ıa, M. C. Comerón, T. M. Wassenaar, and L. Dom´ınguez. 2012. Evaluation of four protocols for the detection and isolation of thermophilic Campylobacter from different matrices. J. Appl. Microbiol. 113:200–208. 30. U.S. Department of Agriculture, Food Safety and Inspection Service (FSIS). 2006. Detection and enumeration method for Campylobacter jejuni/coli from poultry rinses and sponge samples. Available at: https://www.fsis.usda.gov/wps/wcm/connect/0273bc3d-2363-45b3befb-1190c25f3c8b/MLG-41.pdf?MOD¼AJPERES. 31. Verhoeff-Bakkenes, L., H. A. P. M. Jansen, P. H. in ’t Veld, R. R. Beumer, M. H. Zwietering, and F. M. van Leusden. 2011. Consumption of raw vegetables and fruits: a risk factor for Campylobacter infections. Int. J. Food Microbiol. 144:406–412. 32. Yoo, J. H., N. Y. Choi, Y. M. Bae, J. S. Lee, and S. Y. Lee. 2014. Development of a selective agar plate for the detection of Campylobacter spp. in fresh produce. Int. J. Food Microbiol. 189:67–74.