lable at ScienceDirect

Food Microbiology 26 (2009) 317–319

Contents lists avai

Food Microbiology

journal homepage: www.elsevier .com/locate/ fm

Comparison of culture media for enrichment and isolation of Salmonellaspp. from frozen Channel catfish and Vietnamese basa fillets

Amit Pal a, Douglas L. Marshall b,*

a Campbell Soup Company, 1 Campbell Place, Camden, NJ 08103, USAb College of Natural and Health Sciences, University of Northern Colorado, Gunter 1000, Box 134, Greeley, CO 80639, USA

a r t i c l e i n f o

Article history:Received 21 February 2008Received in revised form10 December 2008Accepted 11 December 2008Available online 31 December 2008

Keywords:SalmonellaCatfishBasaIsolation methods

* Corresponding author. Tel.: þ1 970 351 2877; faxE-mail address: douglas.marshall@unco.edu (D.L. M

0740-0020/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.fm.2008.12.003

a b s t r a c t

Frozen fillets of Channel catfish and Vietnamese basa fish were used to compare Salmonella spp. recoveryeffectiveness of selective enrichment in Rappaport–Vassiliadis (RV) broth and tetrathionate broth (TT)and selective isolation on Hekteon enteric (HE) agar, xylose lysine deoxycholate (XLD) agar, and bismuthsulfite (BS) agar. Isolate confirmation was through fatty acid methyl ester analysis. Of 60 samplesanalyzed, 25 were found contaminated with Salmonella (42% incidence). Salmonella spp. recovery afterenrichment in RV medium was 35% on HE agar, 30% on XLD agar, and 42% on BS agar. Similarly, afterenrichment in TT broth, HE and XLD agars recovered 22% each and BS agar recovered 37%. No perfor-mance difference (p > 0.05) was observed in the recovery of Salmonella using the combinations of BS, HE,and XLD agars with RV broth and BS agar with TT broth. The combination of selective enrichment in RVand selective isolation on BS gave numerically greatest isolation of Salmonella from Channel catfish andVietnamese basa fish compared to other isolation combinations.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Salmonella has been previously isolated from catfish (Andrewset al., 1977; Wyatt et al., 1979; Hannah and McCaskey, 1995), andChannel catfish (Ictalurus punctatus) has been responsible for onehuman salmonellosis outbreak (CDC, 1991). The primary habitat ofSalmonella is in the intestinal tract of animals, such as birds,reptiles, farm animals, humans, and occasionally insects (Flores-Abuxapqui et al., 2003). Berg and Anderson (1972) determined thatavian fecal material is a primary source of Salmonella in fishproducts. The potential source of Salmonella contamination infarm-raised catfish is likely due to poor water quality, farm runoff,fecal contamination from wild animals or livestock, feed (Ward,1989; Gonzalez-Rodrıguez et al., 2002), processing under poorsanitary conditions (D’Aoust et al., 1992), or poor distribution, retailmarketing, and handling/preparation practices (Zhao et al., 2003).Wyatt et al. (1979) found that high stocking densities and highwater temperature may be responsible for increased Salmonellacontamination of farm-raised catfish.

Salmonella has been isolated not only from domestic catfish butalso from imported fish and fish products, including catfish(D’Aoust et al., 1992; Heinitz et al., 2000; Zhao et al., 2003).Widespread distribution of antibiotic-resistant Salmonella strains

: þ1 970 351 2176.arshall).

All rights reserved.

due to international trade of Salmonella-contaminated seafoodproducts (D’Aoust et al., 1992; Zhao et al., 2003; Ponce et al., 2008)is a public health concern. Survey studies in the USA have showna low incidence (1.5–4.5%) of Salmonella contamination in retailcatfish fillets (Andrews et al., 1977; Hannah and McCaskey, 1995;Heinitz et al., 2000). Low Salmonella incidence was also seen withother U.S. seafood products (1.3% of 768 samples) (Heinitz et al.,2000). In contrast, Salmonella incidence in imported seafoodproducts was greater (7.2% of 11,312 samples). Among all seafoodproducts examined raw fish had the greatest Salmonella incidence,with Vietnamese seafood products having the greatest Salmonellaincidence of all the import products tested (Heinitz et al., 2000).

The U.S. Food and Drug Administration method for Salmonellaisolation from fish and fish products recommends pre-enrichmentwith lactose broth (LB), followed by selective enrichment inRappaport–Vassiliadis (RV) broth and tetrathionate broth (TT), andthen selective isolation of typical and atypical Salmonella colonieson Hekteon enteric (HE) agar, xylose lysine deoxycholate (XLD)agar, and bismuth sulfite (BS) agar (Andrews and Hammack, 2006).Since the above protocol requires multiple enrichment and isola-tion media, the effort of Salmonella isolation from fish could beminimized if media selection is reduced. Therefore, the objective ofthe present study was to use this protocol to assess whether therewere performance differences among the media to isolate Salmo-nella spp. from naturally contaminated U.S. farm-raised Channelcatfish and Vietnamese basa fish.

A. Pal, D.L. Marshall / Food Microbiology 26 (2009) 317–319318

2. Materials and methods

2.1. Fish source and microbial analysis

Frozen farm-raised Channel catfish (Ictalurus punctatus) filletswere purchased from four local retail stores. Frozen basa fish(Pangasius bocourti) fillets were initially purchased from retailoutlets and later, due to unavailability in the market because ofimport restrictions, were purchased from a local seafood distrib-utor. Fillets were transported frozen within retail seafood bags tothe laboratory and stored at �20 �C until day of use. During thecourse of the study, 30 fillets of each fish species from differentretail bags were analyzed. The study was conducted from May 2003through December 2004.

Frozen fish fillets were thawed overnight at 6 �C before micro-bial analysis. Fillets were removed aseptically from bags and 25-gpieces from each fillet were aseptically excised and transferred tosterile stomacher filter bags (ThermoFisher Scientific, Fairlawn, NJ,USA). Fifty milliliters of 0.1% sterile peptone water (BD DiagnosticSystems, Sparks, MD, USA) was added to each bag and blended for2 min in a stomacher (Tekmar Company, Cincinnati, OH, USA).Decimal dilutions from the homogenates were prepared with 0.1%peptone water. One-milliliter aliquots from each dilution were thenplated on duplicate 3M Petrifilm� Aerobic Count plates (3M, St.Paul, MN, USA) that were incubated at 35 �C for 48 h and counted.Similarly, fecal coliform counts were measured using duplicate 3MPetrifilm coliform count plates incubated at 44.5 �C for 24 h. Meancount values were reported as log10 CFU/g.

2.2. Isolation of Salmonella spp.

Twenty-five grams of thawed fish fillet was aseptically placed in225 ml sterile lactose broth (LB; BD Diagnostic Systems) andblended in a stomacher for 2 min. Inoculated LB was asepticallytransferred to capped 250 ml screw-cap jars and kept for 60 min at25 �C, then the jar cap was opened 1/4 turn before incubation at35 �C for 24 h. Following enrichment, jars were vigorously shakenfor 5 s and 0.1 ml of LB was transferred into 10 ml Rappaport–Vassiliadis broth (RV; Sigma Chemical Co., St. Louis, MO, USA) and1 ml of LB was transferred into 10 ml tetrathionate broth (TT; SigmaChemical Co.). RV tubes were incubated at 42 �C and TT tubes at43 �C for 24 h. After mixing each tube, 3 mm loopfuls (10 ml) werestreaked from RV and TT broths onto Hekteon enteric (HE) agar,xylose lysine deoxycholate (XLD) agar, and bismuth sulfite (BS) agarplates (all from BD Diagnostic Systems). All plates were incubatedat 35 �C for 24 h. The presence of both typical and atypical Salmo-nella colonies was examined on each selective agar. Two or moretypical Salmonella colonies per plate were picked and transferredvia stabs to triple sugar iron agar (TSI; BD Diagnostic Systems) andlysine iron agar (LIA; BD Diagnostic Systems) slants, which wereincubated for 24 h at 35 �C. Color changes typical of Salmonellawere noted for each agar slant.

2.3. Confirmation of Salmonella spp.

Isolates that were presumptively identified as Salmonella wereconfirmed by fatty acid methyl ester analysis using gas chroma-tography (Pendergrass, 1998; Yuk and Marshall, 2003; O’Hara,2005). Typical colonies from selective media were streaked ontotriplicate trypticase soy agar plates (TSA; BD Diagnostic Systems).Upon complete determination of purity, approximately 40 mg(15–20 loopfuls) of each isolate was harvested using a steriledisposable loop and placed into a 13 � 100 mm Kimax test tube andsuspended in 1 ml of saponification solution consisting of 45 g ofACS grade NaOH pellets (#S318-500; ThermoFisher Scientific),

150 ml HPLC grade methanol (#A452-4; ThermoFisher Scientific),and 150 ml of deionized distilled water by vortexing for 5 s. Tefloncapped tubes were then heated in a boiling water bath (100 �C) for5 min followed by vortexing for 5 s. The tubes were returned to the100 �C water bath for additional heating for 25 min followed bycooling in cold tap water. This combination of methanolic base andheat lysed cells, which released fatty acids from cell membranesthat were then converted to their sodium salts.

The second step, methylation, used 2 ml of a methylationreagent composed of 325 ml 6.0 N HCl (#LC15370-3; LabChem Inc.,Pittsburgh, PA, USA) and 275 ml HPLC grade methanol (#A452-4;ThermoFisher Scientific). Tubes were recapped and heated for10 min at 80 �C followed by rapid cooling in tap water. The meth-ylation step converted the fatty acids (salt form) to fatty acid methylesters, which increased volatility for gas chromatography analysis.

The third step was extraction, during which the fatty acidmethyl esters were removed from the acidic aqueous phase andtransferred to an organic phase. Extraction was achieved by theaddition of 1.25 ml of the extraction reagent to each tube and thentumbling for 10 min. This reagent consisted of 200 ml HPLC gradehexane (#H302-4; ThermoFisher Scientific) and 200 ml HPLC grademethyl tert-butyl ether (#E127-4; ThermoFisher Scientific). Aftercessation of tumbling, the bottom phase of each sample was dis-carded. The top phase samples were washed with 3.0 ml of a mildbase solution made of 10.8 g ACS grade NaOH pellets (Thermo-Fisher Scientific) and 900 ml deionized distilled water to removefree fatty acids and residual reagents from the organic extract. Afterrecapping, tubes were tumbled for 5 min, then two-thirds of theorganic phase was pipetted into a vial that was capped and readyfor analysis.

A Hewlett–Packard Model 6890 gas chromatograph (Wilming-ton, DE, USA) equipped with a split capillary injector and a flameionization detector was used to analyze fatty acid methyl esters.Separations were obtained using a Hewlett–Packard Ultra 2 cross-linked 5% PHME siloxane column (25 m�0.2 mm�0.33 mm filmthickness) (HP part #19091B-102). Ramping of temperatureprogram was from 170 to 270 �C at 5 �C per minute. Hydrogen wasused as carrier gas with a flow rate of 30 ml/min. MIDI SherlockMicrobial Identification System (MIDI Inc., Newark, DE, USA) wasused for analyzing fatty acid profiles. Each similarity table provideda list of organism names along with a similarity value. Presence ofSalmonella spp. was confirmed if the similarity value was at least0.5 and if there was a minimum of one-tenth separation betweenthe similarity values of other suggested genera.

2.4. Statistical analysis

The Student t-test using the general linear models procedure(SAS� 9.1.2, SAS Institute Inc., Cary, NC, USA) was used to assessdifferences (p � 0.05) between aerobic and fecal coliform counts ofthe two fish species. Paired-wise Fisher exact test (from SAS PROCFREQ) was used to find significant differences (p � 0.05) betweenSalmonella incidence (%) using different enrichment (RV or TT) andisolation media (HE, XLD, and BS).

3. Results and discussion

Mean (�standard deviation) aerobic plate count of the Channelcatfish fillets was 4.4 � 0.7 log10 CFU/g, which was significantlygreater (p < 0.05) than that of basa fillets (3.8 � 0.4 log10 CFU/g)(Fig. 1). The majority of Channel catfish (24 out of 30) and basa fish(27 out of 30) samples had fecal coliform counts below 3 CFU/g. Theremaining samples had mean fecal coliform counts of1.1 � 0.2 log10 CFU/g for 6 Channel catfish samples and 1.0 � 0 log10

CFU/g for 3 basa fish samples, which were not significantly different

0

1

2

3

4

5

6

APC FCCMicrobial counts

lo

g10 C

FU

/g

Channel catfishBasa fish

a

b

aa

Fig. 1. Aerobic (APC) and fecal coliform (FCC) counts of frozen Channel catfish and basafish fillets. Numbers within count type that have the same letter are not significantlydifferent (p > 0.05). N ¼ 30 for each fish type. FCC: 6 positive Channel catfish fillets and3 positive basa fish fillets.

A. Pal, D.L. Marshall / Food Microbiology 26 (2009) 317–319 319

(p > 0.05). Fillets from both fish species were of acceptablemicrobiological quality with total counts below 6 log10 CFU/g(Nychas et al., 2007).

The numbers of Salmonella-positive samples recovered by eachselective isolation media after selective enrichment are summa-rized in Table 1. Twenty-five fish samples out of 60 were Salmonellapositive (42% incidence). The incidence of Salmonella in Channelcatfish and Vietnamese basa was 33% and 50%, respectively.Regardless of the selective isolation media used, the percentages ofpositive samples after selective enrichment in RV were notdifferent (p > 0.05) (Table 1). Salmonella recovery after selectiveenrichment in TT and selective isolation on BS was not different(p > 0.05) than the combination of selective enrichment in RV andselective isolation on BS, HE, or XLD; however, lower recovery(p < 0.05) was seen when samples were selectively enriched in TTand selectively isolated on HE or XLD. Numerically, selectiveenrichment in RV gave more Salmonella-positive samples thanselective enrichment in TT. This result is consistent with otherreports with high count foods, where RV consistently gave greaterrecovery of Salmonella (June et al., 1996; Andrews and Hammack,2006), while foods with low microbial load are better suited toselective enrichment in TT (Hammack et al., 1999). The currentfinding that HE and BS were not significantly different in selectivelyisolating Salmonella when selectively enriched in RV is consistentwith previous work with shrimp (Kumar et al., 2003). The presentstudy found no significant (p > 0.05) isolation differences betweenmedia that were specific to the two fish species.

Using BS for selective isolation resulted in numerically greaterrecovery of Salmonella than the other two selective isolation mediawhen selectively enriched in either RV or TT. This observationcoupled with numerically greater recovery after selective enrich-ment in RV suggests that use of RV for selective enrichment and BSfor selective isolation may be the best combination to determine

Table 1Number of confirmed Salmonella positive fish fillets by each media (n ¼ 30 for eachfish type).

Selective enrichmentmedia

RV TT

Selective isolationmedia

HE XLD BS HE XLD BS

Channel catfish 7 6 10 4 3 8Basa fish 14 12 15 9 10 14Total (60) 21 a 18 a 25 a 13 b 13 b 22 a

Media: Rappaport–Vassiliadis (RV) broth, tetrathionate broth (TT), Hekteon enteric(HE) agar, xylose lysine deoxycholate (XLD) agar, and bismuth sulfite (BS) agar.Numbers with the same letter are not significantly different (p > 0.05).

the presence of Salmonella on these fish. Additional work withlarger sample sizes and further isolate confirmation using poly-valent antisera (O and H) will be needed to confirm this notion. Ifconfirmed, valuable time and resources could be saved by usinga single enrichment and isolation combination.

It is notable that the Salmonella incidence levels reported hereare greater than previous reports with fish (Andrews et al., 1977;Wyatt et al., 1979; Hannah and McCaskey, 1995; Heinitz et al.,2000). Although the reason for this is unknown, a possible expla-nation may be the use of different isolation and identificationmethods among the studies. Additionally, there may be differencesin how fish were produced, processed, and handled over time.Nevertheless, the high incidence of Salmonella contamination ofboth fish species should merit reasonable caution when they areprepared for consumption.

Acknowledgments

This work was supported in part by a USDA-CSREES SpecialGrant (2003-34231-13064).

References

Andrews, W.H., Hammack, T.S., 2006. Salmonella. In Bacteriological AnalyticalManual Online, Chapter 5. Available at http://www.cfsan.fda.gov/webam/bam-5.html. Accessed on 12/25/07.

Andrews, W.H., Wilson, C.R., Poelma, P.L., Romero, A., 1977. Bacteriological survery ofthe channel catfish (Ictalurus punctatus) at the retail level. J. Food Sci. 42, 359–363.

Berg, R.W., Anderson, A.W., 1972. Salmonellae and Edwardsiella tarda in gull feces:a source of contamination in fish processing plants. Appl. Microbiol. 24, 501–503.

CDC. 1991. Annual listing of foodborne disease outbreaks, United States. Available athttp://www.cdc.gov/foodborneoutbreaks/us_outb/fbo1991/fbofinal1991.pdf.Accessed on 01/28/08.

D’Aoust, J., Sewell, A.M., Daley, E., Greco, P., 1992. Antibiotic resistance of agriculturaland foodborne Salmonella isolates in Canada: 1986–1989. J. Food Prot. 55, 428–434.

Flores-Abuxapqui, J.J., Puc-Franco, M.A., Heredia-Navarrete, M.R., de la Vivas-Rosel, M.L., Franco-Monsreal, J., 2003. Comparison between sodium seleniteand sodium tetrathionate broths, incubated at 37�C and 42�C for the isolation ofSalmonella spp. from faeces of carries. Rev. Biomed. 14, 215–220.

Gonzalez-Rodrıguez, M.N., Sanz, J.J., Santos, J.A., Otero, A., Garcıa-Lopez, M.L., 2002.Foodborne pathogenic bacteria in prepackaged fresh retail portions of farmedrainbow trout and salmon stored at 3�C. Int. J. Food Microbiol. 76, 135–141.

Hammack, T.S., Amaguana, R.M., June, G.A., Sherrod, P.S., Andrews, W.H., 1999.Effectiveness of selenite cystine broth, tetrathionate broth, and Rappport-Vassiliadis medium for the recovery of Salmonella spp. from foods with a lowmicrobial load. J. Food Prot. 62, 16–21.

Hannah, T.C., McCaskey, T.A., 1995. Evaluation of the microbial quality and safety ofretail channel catfish fillets. Southern Assoc. Agric. Sci. Abstr 32, 20.

Heinitz, M.L., Ruble, R.D., Wagner, D.E., Tatini, S.R., 2000. Incidence of Salmonella infish and seafood. Int. J. Food Microbiol. 63, 579–592.

June, G.A., Sherrod, P.S., Hammack, T.S., Amaguana, R.M., Andrews, W.H.,1996. Relativeeffectiveness of selenite cystine broth, tetrathionate broth, and Rappport-Vassi-liadis medium for recovery of Salmonella spp. from raw flesh, highly contaminatedfoods, and poultry feed: Collaborative study. J. AOAC Int. 79, 1307–1323.

Kumar, H.S., Sunil, R., Venugopal, M.N., Karunasagar, I., Karunasagar, I., 2003.Detection of Salmonella spp. in tropical seafood by polymerase chain reaction.Int. J. Food Microbiol. 88, 91–95.

Nychas, G.J.E., Marshall, D.L., Sofos, J.N., 2007. Meat, poultry, and seafood. In:Doyle, M.P., Beuchat, L.R. (Eds.), Food Microbiology: Fundamentals and Fron-tiers, third ed. ASM Press, Washington, DC, pp. 105–140.

O’Hara, C.M., 2005. Manual and automated instrumentation for identification of Enter-obacteriaceae and other aerobic Gram-negative bacilli. J. Clin. Microbiol.18,147–162.

Pendergrass, S.M. 1998. Aerobic bacteria by GC-FAME. Method 0801. NIOSH Manualof Analytical Methods, fourth ed. Available at http://www.cdc.gov/niosh/nmam/pdfs/0801.pdf. Accessed on 12/5/08.

Ponce, E., Khan, A.A., Cheng, C.M., Summage-West, C., Cerniglia, C.E., 2008. Preva-lence and characterization of Salmonella enterica serovar Weltevreden fromimported seafood. Food Microbiol. 1, 29–35.

Ward, D.R.,1989. Microbiology of aquacultured products. Food Technol. 43 (11), 82–86.Wyatt, L.E., Nickelson, R., Vanderzant, C., 1979. Occurrence and control of Salmonella

in fresh water catfish. J. Food Sci. 44, 1067–1069. 1073.Yuk, H.G., Marshall, D.L., 2003. Heat adaptation alters Escherichia coli O157:H7

membrane lipid composition and verotoxin production. Appl. Environ. Micro-biol. 69, 5115–5119.

Zhao, S., Datta, A.R., Ayers, S., Friedman, S., Walker, R.D., White, D.G., 2003.Antimicrobial-resistant Salmonella serovars isolated from imported foods. Int.J. Food Microbiol. 84, 87–92.