REGULAR ARTICLES

Isolation, prevalence, and risk factors for infection by shigatoxin-producing Escherichia coli (STEC) in dairy cattle

Marcos Roberto Alves Ferreira & Edismauro Garcia Freitas Filho &

Jefferson Fernando Naves Pinto & Márcia Dias & Cecília Nunes Moreira

Accepted: 9 January 2014# Springer Science+Business Media Dordrecht 2014

Abstract Rectal swabs of 198 Holstein×Gir crossbred beefcattle from 34 milk farms in the central west of Brazil wereanalyzed from August 2010 to February 2011. Strains of shigatoxin-producing Escherichia coli (STEC) were isolated from72.73 % (144/198) of the animals, on over 97 % of thesurveyed properties. The molecular characterization indicatedthe most common toxin gene stx1 in 70.88 % of the animals(202/285), followed by 18.95 % (54/285) stx1/sxt2, and10.18 % (29/285) stx2. The presence of STEC in animalstogether with the probable risk factors based on a questionnairewas evaluated in the owners of the evaluated animals. Resultsshowed that the animal category “calves” and production/technification scale “low” of the farm were related to highSTEC prevalence in cattle. The season did not significantlyaffect the presence of STEC in cattle. The STEC strains areconsidered a major pathogen, causing severe and potentially

lethal diseases in humans such as hemorrhagic colitis andhemolytic uremic syndrome. This high prevalence of STECin dairy cattle poses a significant risk to public health, sincethese microorganisms can contaminate products intended forhuman consumption, e.g., water, raw and pasteurized milk,meat products, dairy products, and/or products of plant origin.

Keywords stx1 . stx2 . Epidemiology . Animal category .

Climate

Introduction

Among the Escherichia coli that produce shiga toxin or shigatoxin-producing E. coli (STEC), the serotype O157:H7 isdescribed as the main pathogen responsible for causing seri-ous and potentially lethal diseases in humans, e.g., hemor-rhagic colitis and hemolytic uremic syndrome (HUS). Thissubgroup of enterohemorrhagic E. coli (EHEC) is usuallyisolated from contaminated animals and food and representsa threat to public health (Karmali et al. 2010; Fernández et al.2012). However, there are over 100 recognized STEC sero-types other than O157 (non-O157), causing human diseases aswell (Coombes et al. 2011).

Several animals are sources of STEC infection; however,the leading role of cattle, which excrete STEC in the feces, inthe spread of this pathogen was confirmed a few years ago(Ojo et al. 2010). Humans become infected with STEC strainsthrough direct contact with infected animals or by ingestion ofcontaminated water, raw and pasteurized milk, meat products,milk, and/or plant-derived products (Pennington 2010;Pacheco and Sperandio 2012).

Worldwide, few studies have been published on the riskfactors associated with the presence of this pathogen in cattle,and for the state of Goiás, this is the first report. Knowledgeabout the risk factors involved in the occurrence of this agent inanimals is essential for public health, to be able to reduce the

M. R. A. Ferreira :C. N. Moreira (*)Departamento de Medicina Veterinária, Campus Jataí, UniversidadeFederal de Goiás, Rodovia BR 364, Km 192 no 3.800 - Pq.Industrial, Caixa Postal 03, CEP 75801-615 Jataí, GO, Brazile-mail: cissanm@yahoo.com.br

M. R. A. Ferreirae-mail: marcosferreiravet@gmail.com

E. G. F. FilhoDepartamento de Biologia Celular e Molecular e BioagentesPatogênicos, Faculdade deMedicina de Ribeirão Preto, Universidadede São Paulo, São Paulo, Brazile-mail: edismauro@gmail.com

J. F. N. PintoDepartamento de Ciências Biológicas, Campus Jataí, UniversidadeFederal de Goiás, Rodovia BR 364, Km 192 no 3.800 - Pq.Industrial, Caixa Postal 03, CEP 75801-615 Jataí, GO, Brazile-mail: jeffernando@ibest.com.br

M. DiasDepartamento de Zootecnia, Campus Jataí, Universidade Federal deGoiás, Rodovia BR 364, Km 192 no 3.800 - Pq. Industrial, CaixaPostal 03, CEP 75801-615 Jataí, GO, Brazile-mail: diasmarcia@yahoo.com.br

Trop Anim Health ProdDOI 10.1007/s11250-014-0541-5

contamination in the supply chain, where the control of thispathogen is vital. In view of the importance of STEC-relatedfoodborne illnesses (FBIs), the purpose of this study was togenotypically characterize E. coli strains isolated from rectalswabs of Holstein×Gir crossbred cattle and analyze them forthe presence of the genes stx1 and stx2, determining the preva-lence of STEC in the animals, and to assess the main risk factorsassociated to the increased prevalence of this pathogen for abetter application of sanitary-hygienic measures involving theentire food production chain from the beginning (living animal).

Materials and methods

Experimental design

The study was conducted from August/September 2010 (winter)to January/February 2011 (summer), surveying the prevalenceand investigating possible risk factors for the occurrence ofSTEC reservoirs in animals on 34 farms located in the munici-pality of Jataí, Goiás, which is the third largest milk producer inBrazil (Mezzadri 2012). The evaluated farms were located in aregion of mesothermal tropical climate (latitude 17° 52′ 53″S,longitude 51° 42′ 52″W), with two well-defined seasons, namelywinter (6.4-mm low rainfall, relative air humidity of 43.7%,mildtemperatures of around 24.7 °C) and summer (heavy rains, 86.7-mm high rainfall, relative air humidity of 94 %, and highertemperatures on average of 26 °C). This number (34 farms)was calculated using the statistical software Epi info 6.04, basedon the expected frequency percentage of farms with bovineSTEC reservoirs in Brazil described previously (Sandrini et al.2007; Stella et al. 2012), based on an expected prevalence of90 %, with an error margin of 10 % and a confidence level of95 %. A systematic random sample was drawn, based on thecalculation of a jump-ratio between the total number of producersintegrated in the cooperatives and the number of farms to besampled, for each production stratum.

To represent the different levels of mechanization, the 34farms were divided into three production strata: 16 farms withan output of up to 100 L/milk/day, with an average productionof 3 L/milk/day per cow, hand milking, and extensive manage-ment; 12 farms with 100 to 400 L/milk/day with an average

production of 10 L/milk/day per cow, milking, and extensivemanagement; and 6 farms that produce over 400 L/milk/daywith an average output of 20 L/milk/day per cow, milking, andsemi-intensive management. Producers were interviewed witha structured questionnaire to identify the risk factors for animalinfection and contamination of the properties.

Each farm was visited and evaluated twice: once in thewinter (dry season) and once in summer (rainy season). On the34 studied properties, a total of 5,804 heads of cattle weredispersed across 2,488 ha. Rectal swabs were collected fromhealthy and nondiarrheic Holstein×Gir crossbred cattle,choosing two calves and one cow from each production unitby simple random sampling. However, in the second evalua-tion, on four farms, it was not possible to collect samples fromtwo cows and four calves due to the absence of animalsavailable for sampling at the time of the visit. Therefore, 198animals were evaluated, namely 34 cows in winter and 32cows in summer (3–8 years old) and 68 calves in winter and64 calves in summer (0–8 months old, with 42 females and 26males in winter, and 42 females and 22males in summer). Thecalves were kept close to their mothers. Whenever possible,the second sampling was performed in the same animals;however, when absent due to death or sale, the animals were(randomly) substituted, once the purpose of the analysis wasto assess the prevalence of STEC in groups under differentconditions rather than the individual prevalence.

The prevalence rate was calculated as the quotient of thenumber of animals with STEC by the total number of animalsobserved, and the same for each category and farm. To eval-uate the effect of the variables, the examined animals weredivided into categories: cows 33.33 % (66/198); calves66.66 % (132/198); and production stratum: low yield 47 %(16/34), average yield 35.3 % (12/34), and high yield 17.7 %(6/34). The information obtained from the questionnaire an-swered by the farm owners with respect to the management,environment, and animals was also assessed to determinepossible risk factors of STEC infection in cattle.

Bacterial strains and determination of virulence factors

Virulence factors of 990 E. coli isolates from rectal swabs of198 healthy calves on 34 farms were determined in the dry and

Table 1 Sequence of primersused to determine virulencefactors

Targetgene Sequence (5’ - 3’) Ampliconsize (bp)

Location withingene

GENBANKaccession

16SrRNA CCCCCTGGACGAAGACTGAC 401 1682–1701 AB035924Control ACCGCTGGCAACAAAGGATA 2082–2063

stx1 TCTCAGTGGGCGTTCTTATG 338 777-796 M17358TACCCCCTCAACTGCTAATA 1114-1095

stx2 GGCACTGTCTGAAACTGCTCC 255 603-623 NC004914TCGCCAGTTATCTGACATTCTG 857-837

Trop Anim Health Prod

rainy period. Each swab was inoculated in Stuart mediumtubes (Difco Laboratories, Detroit, MI, USA), ice-cooled,and analyzed within 24 h. The fecal samples were streakedonto Levine BEM agar (Difco, Detroit, MI, USA) and incu-bated at 37 °C for 24 h. At least five suspect individual E. colicolonies (dark with a greenish metallic sheen) per animal werechosen, and their identity was confirmed by biochemical tests,using citrate and the production of indole, acetoin, and methylred reactive compounds (Koneman et al. 2008).

DNA samples were extracted from the isolates by heat DNAextraction. These samples were analyzed by PCR for the pres-ence of 16SrRNA (internal control) and stx1 and stx2 to charac-terize STEC (for the primers used, see Table 1). The followingamplification protocol was carried out in a MJ Researchthermocycler under the above-described PCR test conditions:stx2 (Paton and Paton 1998) and stx1 (Wang et al. 2002). E. coliO157:H7 and Klebsiella pneumoniae DNAwere used as posi-tive and negative controls, respectively. These control strains,belonging to a collection maintained in the Technology Devel-opment Center/Biotechnology of the Federal University of Pe-lotas, were characterized by genotypic and phenotypic methods.

The amplified products were viewed after molecularweight determination by 2 % agarose gel electrophoresis at70 V for 180 min, in 1× TBE buffer (Tris-HCl of 89 mM,boric acid of 89 mM, EDTA of 2.5 mM, pH 8.0). A DNAladder (φX174/Hae III) was used as molecular weight pattern;

the gels were stained with ethidium bromide (0.5 μg/μL ofethidium bromide solution).

Statistical analyses

In the analysis of possible risk factors for the presence ofSTEC in cattle, all data were analyzed using SAS v.9.0, at5 % probability. The effect of the production stratum, season,and animal category of the farm on STEC occurrence waschecked by logistic regression analysis with dichotomousresponse, using the Wald and odds ratio (OR) tests, consider-ing all effects in the full model, with subsequent modelingdisregarding the nonsignificant effects.

Ethical aspects

The study activities were carried out, and the animals weretreated according to the international standards and in com-pliance with the ethical principles of animal experimentationestablished by Colégio Brasileiro de Experimentação Animal(COBEA) (protocol 206/2009/CEUA/UFG).

Results and discussion

In Brazil, several researchers investigated the prevalence ofSTEC in animals; however, in the state of Goiás, there is nodescription of the prevalence of this pathogen in cattle. Thisstudy is the first to address this topic and one of the few in thecountry to evaluate the effect of STEC prevalence with ex-trinsic (weather, facilities, management, feeding, productionstratum) and intrinsic risk factors (age, health, nutrition) of theanalyzed animals.

The prevalence of dairy cattle with STEC was 72.73 %(144/198), of which 26.39% (38/144) were cows and 73.61%(106/144) were calves, spread across 97 % (33/34) of theproperties studied, with an incidence between 0 and 100 %of the tested animals. The prevalence was higher than that inother studies conducted in Brazil (Sandrini et al. 2007;Andrade et al. 2012) and in other countries such as Argentinaand Korea (Park et al. 2011; Fernández et al. 2012).

Of the E. coli isolated from animals, 28.78 % (285/990)were STEC. These isolates were PCR positive for the16SrRNA, stx1, stx2, or stx1/stx2 virulence factor genes(Fig. 1). In terms of the presence of genes analyzed in STECisolates, 70.88 % (202/285) had only stx1, 18.95 % (54/285)had stx1 and sxt2, and 10.18 % (29/285) had only stx2. Theepidemiological and experimental data showed that gene stx2is more decisive for the development of HUS than stx1(Stearns-kurosawa et al. 2010; Fuller et al. 2011). In this study,the gene with the highest prevalence in STEC was stx1,corroborating other studies analyzing feces of dairy cattle likethat by Stella et al. (2012), which also indicated a higher

Fig. 1 Analysis of the PCR products in a 1.5 % agarose gel for thepresence of virulence genes in the six STEC isolates. Lane M, molecularmass marker (fragment size 900 to 100 bp); lane 1, negative control; lane2, positive control; lanes 3, 4, 7, and 8, STEC isolates positive for16SrRNA (401 bp) and stx1 (338 bp); lane 5, STEC isolate positive for16SrRNA (401 bp) and stx2 (255 bp); lane 6, STEC isolate positive for16SrRNA (401 bp), stx1 (338 bp), and stx2 (255 bp)

Table 2 STEC-carrying animals as related to the animal category, pro-duction stratum of the farm and season according to the to the completemodel of dichotomous multivariate logistic regression

STEC Animal category Production Stratum of the farm Season

Valor-P1 0.0007 0.0102 0.7607

1Wald test.

Trop Anim Health Prod

prevalence of STEC in calves, but in disagreement withstudies of Bosilevac et al. (2011). Most isolates with stx2wereextracted from these animals, confirming the findings ofSandrini et al. (2007), who reported that despite the lowerprevalence of STEC in calves, these animals carried mostisolates of the serogroups that are pathogenic for humans.

The effects of the animal category (adults or young ani-mals), season (summer or winter), and of the productionstratum of the property (low, medium, and high) by logisticregression were analyzed considering as reference the follow-ing: animals tested positive for STEC, animal category (cow),production stratum (low), and dry season (winter), which wasonly nonsignificant for season (p>0.05) (Table 2). Since theSTEC occurrence did not differ statistically between winterand summer, i.e., the prevalence of STEC was not affected byweather factors (temperature, relative humidity, and rainfall).The frequency of animals with STEC in the winter (dryseason) was 50.69 % (73/144) and 49.31 % (71/144) insummer (rainy season). However, Fernández et al. (2012)showed that environmental factors that affected the prevalenceof STEC in cattle were temperature, rainfall, and insect pop-ulation. Similarly, Fernández et al. (2012) and Monaghan et al.(2011) reported that days with higher temperatures provide a

more favorable environment for extraintestinal STEC, resultingin more than one source of infection for humans and animals.

Considering the prevalence of animals with STEC for thesignificant variables (production stratum and animal category)according to the final model of dichotomous multivariate logis-tic regression, the frequency of each variable is shown in Table 3,and the assessment of the probable risk factors, in Table 4.

The variable “animal category” showed that calves were 3.2times more prone to carry STEC than cows (Table 4), asindicated by the confidence interval without 1, corroboratingother studies (Fernández et al. 2012). Previous studies reportedon a higher STEC prevalence in calves after weaning, indicat-ing a decrease in protective immunity conveyed by breast milk,dietary changes, intensive management of animals, overcrowd-ing, and the proper weaning stress as the main causes (Menrathet al. 2010). However, these factors cannot definitively explainthe higher prevalence of STEC in calves. Other risk factorsreported in earlier studies are rumen development, inadequatesanitation of the environment, and stress factors such as trans-port (Jeon et al. 2013). These results differed from the findingsof Sandrini et al. (2007), who observed that calves up to11 months old were less infected by STEC than heifers andadult cows (0.32<OR<0.98, P=0.02).

The analysis of the probable risk factors showed that thevariable “production stratum of the property” influenced theprevalence of STEC, since the probability of STEC-infectedanimals did not differ between farms with low- and medium-production levels—prevalence of 42.6 % (23/54) and 27.8 %(15/54), respectively. The high production properties showeda protective factor, with 32.3 % less chance of having animalswith STEC than the low production units (Table 4). Theseresults are in agreement with those of Sandrini et al. (2007),who observed that small properties (<20 ha) (1.20<OR<3.71,P=0.004) with a lower number of animals (1.25<OR<3.87,P=0.003) had a larger number of animals infected with STEC.

The higher prevalence of STEC in animals on low-production farms can be partly explained by the informationobtained in the questionnaire which revealed a lower mecha-nization level and an inadequate infrastructure for milk pro-duction on these farms, aside from the lower level of knowl-edge of the producers about hygiene conditions and animalmanagement for production. The overall condition of the herd,the health conditions, and physical facilities on the smallfarms were inferior to the properties with medium and highproduction, which may have contributed to the higher preva-lence of cattle with STEC (Sandrini et al. 2007).

The type of material the facilities are made of hampers thebasic hygiene measures, favoring the accumulation of organicmatter. This fact is responsible for the multiplication of STEC,as most of the properties of stratum 1 had poorly structuredcorrals, with dirt and/or stone ground, which probably favoredthe higher prevalence of STEC in animals on these farms(Lahti et al. 2002).

Table 3 Prevalence of STEC considering significant variables (the pro-duction stratum and animal category) according to the final model ofdichotomous multivariate logistic regression

Variable STEC

Positive (n) Negative (n) Total (n)

Animal category

1cows 38 28 66

2 calves 106 26 132

Production stratum

Low 70 23 93

Médium 57 15 72

High 17 16 33

Table 4 Odds ratio of the estimates of the final model of dichotomousmultivariate logistic regression for presence of STEC, considering asreference: animals tested positive for STEC, animal category (cow),production stratum (low), and dry season (winter)

Variable ODDS RATIO IC (95 %)

Animal category (CAT)

Cow vs Calf 3.197 1.632 a 6.265

Productionstratum

MPS vs LPS 1.266 0.591 a 2.710

HPS vs LPS 0.323 0.136 a 0.766

*HPS – High production stratum; MPS – Medium production stratum;LPS – Low production stratum.

Trop Anim Health Prod

Other factors that contributed to increase the risk of STECin animals on low production farms were the lower levels ofanimal nutrition and health management of the herd (periodictesting for brucellosis, tuberculosis, and mastitis) than of theother strata (final model of dichotomous multivariate logisticregression).With regard to dietary factors, studies on confinedanimals in the USA revealed a higher prevalence of STEC inanimals stressed by food restriction, which is a commonsituation on small farms (Franz et al. 2007).

Conclusions

It was shown that the prevalence of STEC in dairy herds in thestate of Goiás, Brazil, is high, representing a considerable riskto public health. The analysis of risk factors showed that theanimal category “calves” and the “low” production stratum ofthe farm were related to a high prevalence of STEC in cattle,while there was no influence of the season. Most STECisolates contained the stx1 gene, although most isolates withstx2were extracted from calves, emphasizing the importanceof the more prevalent strains potentially pathogenic to humansin these animals. An improvement in hygiene conditions andquality of facilities, and an enhanced animal management andnutrition may significantly reduce the bovine STEC carriers,reducing the risk of contamination of food for humanconsumption.

Acknowledgments The authors are indebted to the owners and farmworkers for their generous assistance and cooperation in collection of thesamples. This study was supported by grant 503886/2009-2 from theConselho Nacional de Desenvolvimento Científico e Tecnológico(CNPq).

References

Andrade, G.I., Coura, F.M., Santos, E.L.S., Ferreira, M.G., Galinari,G.C.F., Facury-Filho, E.J., Carvalho, A.U., Lage, A.P., andHeinemann, M.B., 2012. Identification of virulence factors by mul-tiplex PCR in Escherichia coli isolated from calves in Minas Gerais,Brazil. Tropical Animal Health and Production, 44, 1783-1790.

Bosilevac, J. M., and M. Koohmaraie. 2011. Prevalence and characteri-zation of non-O157 Shiga toxin-producing Escherichia coli isolatedfrom commercial ground beef in the United States. Applied andEnvironmental Microbiology, 77, 2103-2112.

Coombes, B.K., Gilmour, M.W. and Goodman, C.D., 2011. The evolu-tion of virulence in non-O157 Shiga toxin-producing Escherichiacoli. Frontiers in Microbiology, 2, 1-3.

Fernández, D., Sanz, M.E., Parma, A.E., and Padola, N.L., 2012.Characterization of Shiga toxin producing Escherichia coli isolatedfrom newborn, milk-fed, and growing calves in Argentina. Journalof Dairy Science, 95, 5340-5343.

Franz, E., Klerks, M.M., De Vos, O.J., Termorshuizen, A.J., and VanBruggen, A.H.C., 2007. Prevalence of Shiga Toxin-ProducingEscherichia coli stx1, stx2, eaeA, and rfbE genes and survival of

E. coliO157:H7 inmanure from organic and low-input conventionaldairy farms. Applied and Environmental Microbiology, 73, 7, 2180-2190.

Fuller, C.A., Pellino, C.A., Flagler, M.J., Strasser, J.E., and Weiss, A.A.,2011. Shiga toxin subtypes display dramatic differences in potency.Infection and Immunity, 79, 1329-1337.

Jeon, S.J., Elzo, M., DiLorenzo, N., Lamb, G.C., and Jeong, K.C., 2013.Evaluation of animal genetic and physiological factors that affect theprevalence of Escherichia coliO157 in cattle. Plos One, 8, 2, e55728.

Karmal i , M.A. , Gannon, V. , and Sargeant , J .M. , 2010.Verocytoxinproducing Escherichia coli (VTEC). VeterinaryMicrobiology, 140, 360-370.

Koneman, E.W., Allen, S.D., Janda, W.M., Schreckenberger, P.C., andWinn, W.C., 2008. Diagnóstico Microbiológico: Texto e AtlasColorido. Guanabara, 6ª edição, 231-246.

Lahti, E., Eklund, M., Ruutu, P., Siitonen, A., Rantala, L., Nuorti, P., andHonkanen-Buzalski, T., 2002. Use of phenotyping and genotypingto verify transmission of Escherichia coli O157:H7 from dairyfarms. European Journal of Clinical Microbiology & InfectiousDisease, 21, 189-195.

Mezzadri, F.P. Análise da Conjuntura Agropecuária (2011-2012) – Leite.http://www.seab.pr.gov.br.Accessed: 20 mar 13, 2013.

Menrath, A., Wieler, L.H., Heidemanns, K., Semmler, T., Fruth, A., andKemper, N., 2010. Shiga toxin producing Escherichia coli: identifi-cation of non-O157:H7-Super-Shedding cows and related risk fac-tors. Gut Pathogens, 2, 7.

Monaghan, A., Byrne, B., Fanning, S., Sweeney, T., McDowell, D., andBolton, D.J., 2011. Serotypes and virulence profiles of non-O157shiga toxin-producing Escherichia coli isolates from bovine farms.Applied and Environmental Microbiology, 77, 24, 8662-8668.

Ojo, O.E., Ajuwape, A.T., Otesile, E.B., Owoade, A.A., Oyekunle, M.A.,and Adetosoye, A.I., 2010. Potentially zoonotic shiga toxin-producing Escherichia coli serogroups in the faeces and meat offood-producing animals in Ibadan, Nigeria. International Journal ofFood Microbiology, 142, 214-221.

Pacheco, A.R., and Sperandio, V., 2012. Shiga toxin in enterohemorrhagicE. coli: regulation and novel anti-virulence strategies. Frontiers inCellular and Infection Microbiology, 2, 81.

Park, J.H., Lim, H.S., Lee, K.,Hong, S., and Park,M.S., 2011. Prevalenceof Shiga toxin-encoding genes and risk factors among dairy farmersin Gyeonggi Province, Korea. Scandinavian Journal of InfectiousDiseases, 43, 4, 275-279.

Paton, A.W., and Paton, J.C., 1998. Detection and characterization ofShigatoxigenic Escherichia coli by using multiplex PCR assays forstx1, stx2, eaeA, enterohemorrhagic E. coli hlyA, rfbO111, andrfbO157. Journal of Clinical Microbiology, 36, 598-602.

Pennington, H., 2010. Escherichia coliO157. Lancet, 376, 1428-1435.Sandrini, C.N.M., Pereira, M.A., Brod, C.S., Carvalhal, J.B., and

Aleixo, J.A.G., 2007. Escherichia coli verotoxigênica: isolamentoe prevalência em 60 propriedades de bovinos de leite da regiãode Pelotas, RS, Brasil. Ciência Rural Santa Maria, 37, 1, 175-182.

Stearns-Kurosawa, D.J., Collins, V., Freeman, S., Tesh, V.L. andKurosawa, S., 2010. Distinct physiologic and inflammatory re-sponses elicited in baboons after challenge with Shiga toxin type 1or 2 from enterohemorrhagic Escherichia coli. Infection andImmunity, 78, 2497-2504.

Stella, A.E., Maluta, R.P., Rigobelo, E.C., Marin, J.M., and Ávila, F.A.,2012. Virulence genes in isolates of Escherichia coli from samplesof milk and feces from dairy cattle. Journal of Food Protection, 75,1698-1700.

Wang, G., Clarke, C. and Rodgers, F., 2002. Detection in Escherichia coliof the genes encoding the major virulence factors, the genes definingthe O157:H7 serotype, and components of the type 2 shiga toxinfamily by multiplex PCR. Journal of Clinical Microbiology, 40,3613-3619.

Trop Anim Health Prod