Version of Record: https://www.sciencedirect.

com/science/article/pii/S0168160518307694
Manuscript_a222012405f5b4f75d590c1d04f8855c

1 Phylogroups, pathotypes, biofilm formation and antimicrobial resistance

2 of Escherichia coli isolates in farms and packing facilities of tomato,

3 jalapeño pepper and cantaloupe from Northern Mexico

5 Hesperia Andrea Corzo-Ariyama1, Alam García-Heredia1, Norma Heredia1, Santos

6 García1, Juan León2, LeeAnn Jaykus3 and Luisa Solís-Soto1*.

8 1Departamento de Microbiología e Inmunología, Facultad de Ciencias Biológicas,

9 Universidad Autónoma de Nuevo León, Apdo. Postal 124-F, Ciudad Universitaria,

10 San Nicolás de los Garza, Nuevo León 66455, México, 2Hubert Department of

11 Global Health, Rollins School of Public Health, Emory University, Atlanta, Georgia,

12 USA, 3Department of Food, Bioprocessing and Nutrition Sciences, North Carolina

13 State University, Raleigh, North Carolina, USA.

14

15

16

17 *Corresponding author. E-mail: [email protected]

18

19

20

© 2018 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
21 ABSTRACT

22 The most commonly used indicator of fecal contamination in fresh produce

23 production and packing is Escherichia coli. In depth analysis of the prevalence and

24 characteristics of naturally occurring E. coli strains in these environments is

25 important because it can (1) serve as an indicator of sources of fecal

26 contamination; and (2) provide information on strain pathogenicity, persistence,

27 and other defining characteristics such as multidrug resistance. In this study, we

28 analyzed 341 E. coli strains isolated from the jalapeño pepper, tomato and

29 cantaloupe farm environments, in Northeast Mexico. Strains were isolated from

30 produce, farmworkers’ hands, soil and water. Pathotypes, genotypes, biofilm

31 formation and antibiotic resistance were characterized. Phylogenetic subgroups

32 and identification of diahrreagenic E. coli were determined by PCR; biofilm

33 formation was quantified using a plate-based colorimetric method. Antibiotic

34 resistance was analyzed by the Kirby Bauer diffusion disc method. Most isolates

35 (N=293, 86%) belonged to phylogenetic group A. Only four isolates (1.2%) were

36 diarrheagenic: EPEC (N=3) and ETEC (N=1). Antibiotic resistance to tetracycline

37 (23.2%) and ampicillin (19.9%) was high, and only 3.5% of the strains presented

38 resistance to more than 5 antibiotics. Biofilms were produced by most strains

39 (76%), among which 34.4% were categorized as high producers. The presence of

40 antibiotic resistant E. coli strains that may contain gene markers for pathogenicity

41 and which can form biofilms suggests potential health risks for consumers.

42

43 Keywords: Phylogroup, pathotypes, biofilm, antibiotic resistance, E. coli

44

2
45 1. Introduction.

46 Escherichia coli is a normal inhabitant of the digestive tract of warm blooded animals,

47 including humans (CDC 2014). E. coli is one of the dominant enteric species in

48 human feces and it has been used as an indicator of fecal contamination for close to

49 a century. E. coli is often regarded as harmless; however, there are E. coli groups

50 which have acquired virulence factors and have the ability to cause diarrheal disease

51 in healthy humans (Kaper et al. 2004). A recent systematic review reported a low

52 prevalence of pathogenic E. coli on farms (range 0-1.6%) and packing facilities (0-

53 10%) from fourteen studies testing produce for pathogens. Only in the US, at least

54 nine documented outbreaks of pathogenic E. coli have been linked to the

55 consumption of fresh produce such as lettuce, spinach and sprouts from 2010 to

56 2017 (CDC 2018). Produce can become contaminated in the field or from improper

57 sanitation, handling or processing. The use of compost, sewage contaminated

58 water for irrigation and droppings from wild and domesticated animals and birds

59 are all considered sources of pathogen contamination on fresh produce (Liu et al.

60 2013).

61 The most commonly used indicator of fecal contamination in fresh produce

62 production and packing is Escherichia coli. In depth analysis of the prevalence and

63 characteristics of naturally occurring E. coli strains in these environments is

64 important because it can (1) serve as an indicator of sources of fecal contamination

65 through microbial source tracking (Carlos et al., 2010); (2) identify potentially

66 pathogenic strains; (3) provide information about antimicrobial resistance profiles

67 that can be used to understand strain emergence and clinical treatment of disease

68 (Boehme et al. 2004); and (4) allow us to characterize of the propensity for biofilm

3
69 formation to predict environmental persistence of this organism (Balcazar et al.

70 2015).

71 For microbial source tracking, a classification system has been developed

72 based on phylogenetic cluster characteristics of this bacterium (Carlos et al. 2010;

73 Lee 2011). In this classification, A, B1, B2 and D constitute the main phylogroups,

74 and the subgroups A0, A1, B1, B22, B23, D1 and D2, have been proposed to

75 increase discrimination of E. coli strains (Carlos et al. 2010). Recently, Clermont et

76 al. (2013) proposed a refined classification, adding four more phylogroups: C, E, F,

77 and Escherichia cryptic clade I. The strains of all phylogroups differ in phenotypic

78 and genotypic characteristics (Carlos et al. 2010; Gordon et al. 2008) such as their

79 sugar metabolism, antibiotic resistance profiles, growth temperature ranges,

80 ecological niches, and the presence and/or absence of select virulence factors

81 (Carlos et al. 2010; Gordon et al. 2008). The ability to identify phylogroups has

82 been useful in predicting human health risks. For example, diarrheal disease-

83 causing E. coli are more likely of the B1 and E phylogroups and extraintestinal

84 infection-causing E. coli strains are more likely of the B2 and E phylogroups

85 (Nowrouzian et al. 2006).

86 E. coli, though part of the normal intestinal biota of animals and humans,

87 has the potential to pose human health risks though acquired pathogenic virulence

88 factors that induce diarrhea. These diarrheagenic E. coli (DEC) strains are

89 classified into six different pathogenic types also known as pathotypes that include:

90 enterotoxigenic E. coli (ETEC), enterohaemorrhagic E. coli (EHEC), enteroinvasive

91 E. coli (EIEC), enteropathogenic E. coli (EPEC), enteroaggregative E. coli (EAEC)

92 and diffuse adherent E. coli (DAEC) (Kaper et al. 2004; Russo and Johnson 2000).

4
 93 DEC strains have been associated with outbreaks of severe disease, i.e., bloody

94 diarrhea and hemolytic uremic syndrome (HUS) as well as travelers' diarrhea in

95 association with consumption of contaminated food and water (FDA 2012). In

96 Mexico for instance, the presence of EPEC, ETEC, STEC (E. coli producer of

97 shiga toxin) and ETEC has been reported in ready-to-eat cooked vegetable salads

98 (León et al. 2013), dairy, meat products, seafood, fish and prepared foods

99 (Canizalez-Roman et al. 2013).

100 E. coli can also be an indicator of human health risks if strains are resistant

101 to one or more antibiotics, and studies have identified isolates from agricultural

102 foodstuff and vegetables in which strains were resistant to more than five

103 antibiotics (Boehme et al. 2004; Schwaiger et al. 2011). The antibiotic resistance

104 and susceptibility profiles of E. coli strains vary depending on geographic location,

105 time of exposure to the antimicrobial compound, and other environmental factors

106 (Dombek et al. 2000). Finally, the ability of E. coli to form biofilms may represent a

107 strategy for strains to persist in produce and the production environment. Biofilms

108 can protect bacteria from sanitizers, predation, desiccation and UV radiation

109 (Costerton et al. 1995), providing an advantage for bacterial survival and

110 persistence.

111 Mexico is one of the top producers of fresh produce and a significant trading

112 partner with the U.S. and other countries. Eleven outbreaks have occurred in the

113 US due to the consumption of contaminated produce originating from Mexico in the

114 period 2006 to 2017 (CDC, 2018). The associated contaminated produce included

115 sprouts, leafy greens, spinach and lettuce (FDA, 2012). Hence, there is a need for

116 better understanding of the characteristics of naturally occurring E. coli isolates

5
117 from this region of the world. The purpose of this study was to identify the

118 phylogroups, pathotypes, antibiotic resistance profiles and biofilm formation ability

119 of E. coli strains previously isolated and archived (Heredia et al. 2016) from a

120 longitudinal study along the production chain of jalapeño pepper, tomato and

121 cantaloupe from Nuevo Leon and Coahuila, Mexico. This information provides

122 knowledge about the E. coli strains circulating in the farm environment, the

123 possible sources of contamination in production, and the likelihood of pathogenic

124 and antibiotic resistant strains that could provide a risk to public health.

125

126 2. Material and methods

127 2.1 Bacterial strains

128 In a previous study, 341 E. coli isolates (117 from jalapeño pepper farms, 154 from

129 tomato farms and 70 from cantaloupe farms) were obtained from the production

130 chain in the Nuevo Leon and Coahuila states in Mexico (Heredia et al. 2016).

131 Strains were isolated from water [the intake before the irrigation hose (called

132 source) (24.9%) and the in-field irrigation hose (19.9%)]; from farmworkers hands

133 [during harvest (15.2%), at the distribution point (8.8%), or at packaging (2.6%)];

134 produce [before harvest (9.4%), during harvest (4.7%), at distribution (4.1%), and

135 during packaging (4.7%)]; and from soil (5.6) (Table 1). Control strains included E.

136 coli O157:H7 ATCC 43895 (EHEC, kindly provided by Dr. Lynne McLandsborough,

137 University of Massachusetts, Amherst, MA, USA); E. coli ATCC 25922 (non-

138 pathogenic, donated by Becton Dickinson Co., Mexico); and E. coli O111:NM

139 ATCC 43887 (EPEC), E. coli O78:H11 ATCC 35401 (ETEC) (both commercially

6
140 acquired); and E. coli 042 (EAEC) (donated by Dr. Fernando Navarro-García,

141 CINVESTAV, Mexico).

142 All strains were maintained in brain heart infusion broth (BHI, BD, Franklin

143 Lakes, NJ, USA) with glycerol (20% v/v, Sigma) at -80°C. Fresh cultures were

144 made by inoculating an aliquot (10 µl) into 5 ml tryptone soy broth (TSB, BD,

145 Franklin Lakes, NJ, USA), followed by incubation for 18-24h at 37°C. Working

146 cultures were prepared by streaking an aliquot onto BHI agar tubes and incubating

147 for 24h at 37°C. After the incubation, strains were kept at 4°C until use. New

148 cultures were prepared every three months.

149

150 2.2 Determination of E. coli phylogroups

151 A triplex PCR was used as described by Clermont et al. (2000) to determine

152 phylogroups. Briefly, E. coli strains were grown in TSB and incubated overnight at

153 37°C. An aliquot (500µl) from the culture was used for DNA extraction (Wang et al.

154 1997), homogenized with 1 ml of phosphate buffer solution (PBS 0.05mol/l, pH 7.4)

155 and centrifuged at 9000 x g for 3 min (Eppendorf Microcentrifuge 5415C). Pelleted

156 bacteria were resuspended in 50µl of ultrapure water (miliQ), diluted 1:10 with

157 Triton X-100 (1% v/v, Sigma Aldrich Mexico) and heated (95-100°C) for 10 min.

158 Samples were cooled by immersion in ice water after centrifugation. An aliquot of

159 3μl of DNA-containing supernatant was used as a template for PCR amplification.

160 PCR was performed in a ThermoHybaid (Model HBPX110) thermocycler. Primers

161 were used to amplify TSPE4.C2, chuA, and yjaA genes (Clermont et al. 2000). The

162 final volume of the PCR mixture was 20μl containing 2.5U Taq polymerase

7
163 (BIOLINE, Mexico), MgCl2 1.5mM, 0.2mM dNTP (BIOLINE, Mexico), 2μl of 10x

164 reaction buffer, 3μl of DNA template, and 1μM of each TSPE4.C2, chuA and yjaA

165 oligonucleotides. The amplification conditions included an initial denaturation step

166 of 4 min at 94° C, followed by 30 cycles comprised 94°C for 5s, 59°C for 10s and a

167 final extension of 5 min was done at 72°C.

168 Amplicons (10μl) were separated in a 2% agarose gel and visualized after

169 staining with ethidium bromide (50μg/ml) under UV light (Gel Logic 200 Imaging

170 System, Kodak). E. coli ATCC 25922 (non-pathogenic) was used as a control

171 because it contains all three genes analyzed, does not produce biofilms, and

172 belongs to phylogroup B2. Isolates belonging to phylogroup A0 were confirmed

173 according to the PCR protocol of Higgins et al. (2007), in which a 365 bp fragment

174 of the beta-galactosidase gene (lac-Z) was amplified. Amplicons were separated

175 and visualized as described above.

176

177 2.3 Identification of pathotypes of E. coli (DEC)

178 Overnight TSB cultures were processed for DNA extraction as described by Wang

179 et al. (1997). This was followed by application of the multiplex PCR of Vidal et al.

180 (2005), using probes that amplified the genes stx 1, stx 2, eae (EHEC), eae, bfp

181 (EPEC), stII, lt (ETEC), virF, ipaH (EIEC), aafII (EAEC) and daaE (DAEC). The

182 amplification reaction consisted of 10U Taq polymerase, 1.5 mM MgCl2, 1mM

183 dNTP, 5µl of reaction buffer 10 X, 3µl of template DNA, and 0.5µM of each set of

184 primers (in the case of EHEC and EPEC primers, the concentration used was

185 0.1µM)) in a final volume of 50µl. The amplification conditions consisted of 35

8
186 cycles of 94°C for 1.5 min, 60°C for 1.5 min, and 72°C for 1.5 min. The

187 amplification products were visualized as described above.

188

189 2.4 Biofilm formation assays

190 Biofilm formation assays were performed according to Naves et al. (2008), with

191 some modifications (media used and temperature of incubation). Briefly, strains

192 were grown in tubes containing 5 ml of Mueller Hinton (MH) broth (BD, Franklin

193 Lakes, NJ, USA), and incubated overnight (18 h) at 37°C. Aliquots (20μl) were

194 inoculated into 96 well flat-bottomed polystyrene microtiter plates (Nunc™, Thermo

195 Scientific™, UK) containing 180µl of Luria Bertani (LB) broth (EMD Millipore

196 Corporation, Germany) supplemented with sodium citrate (1%) and glucose

197 (0.2%). Plates were incubated overnight at 37°C, and optical densities (ODs) were

198 read at 630 nm (Bio-Tek Epoch multi-volume spectrophotometer, Bio-Tek

199 Instrument, Winooski, VT, USA). Broth was removed and wells were rinsed once

200 with 200μl of distilled water, air dried, and stained with 200μl of 0.1% crystal violet

201 (CV, Sigma Aldrich, Mexico) for 15 min. The colorant was discarded and wells

202 were rinsed five times with 200µl of distilled water, and then air-dried. To extract

203 the CV from the biofilm, 200μl of 96% ethanol (Sigma Aldrich, Mexico) was added,

204 and then incubated 5 min at room temperature. The ODs of the solutions were

205 measured at 570 nm. Biofilm formation was quantitated by determining the biofilm

206 formation index (BFI) obtained from the formula

207 BFI = (AB – CW)/G,

9
208 where AB was the optical density of the stained attached microorganism, CW was

209 the optical density of the stained control wells containing microorganism-free

210 medium only, and G corresponded to the optical density of the cell growth in

211 suspended culture (Teh et al. 2010). According to BFI values, biofilm formation

212 was categorized as strong (>1.10), moderate (0.70–1.09), weak (0.35–0.69), or no

213 biofilm formed (<0.35) (García-Heredia et al. 2016).

214

215 2.5 Determination of antibiotic resistance profile of E. coli

216 The Kirby-Bauer disk diffusion method reported in the Clinical and Laboratory

217 Standards Institute (CLSI 2012) was used for these studies. Antibiotics tested were

218 chosen on the basis of their use to treat infections of Gram-negative bacteria

219 (Amabile-Cuevas 2010) and included: nalidixic acid (NA, 30μg),

220 sulfamethoxazole/trimethoprim (SXT, 1.25/ 23.75μg), tetracycline (TE, 30μg),

221 gentamicin (CN, 10μg), ciprofloxacin (CIP, 5μg), ampicillin (AMP, 10μg),

222 ceftazidime (CAZ, 30μg), cefotaxime (CTX, 30μg) and chloramphenicol (C, 30μg)

223 (Oxoid Company, Cambridge, UK).

224 Bacterial strains were streaked onto MH agar plates and incubated for 18-20

225 h at 37°C. Isolated colonies were homogenized with sterile saline solution (0.85%

226 w/v), adjusted to No. 0.5 McFarland standard at 600nm absorbance (approximately

227 1.5 x108 CFU/ml), and decimal serial dilutions were made with sterile saline

228 solution. One hundred µl of the 106 CFU/ml dilution was spread onto MH agar

229 plates using a Drigalski spatula. After 15 min, the antibiotic discs were placed on

230 the plates using sterile forceps. Plates were incubated at 37°C for 16-18h, and the

231 inhibition halos were measured using a caliper. The zones of inhibition were

10
232 interpreted according Clinical and Laboratory Standards Institute (CLSI 2014) and

233 the bacteria categorized as sensitive (S: ≥15mm of diameter for CN and TE,

234 ≥16mm of diameter for SXT; ≥17mm of diameter for AMP; ≥18 of diameter for C;

235 ≥19 mm of diameter for NA; ≥21 mm of diameter for CAZ and CIP and ≥26mm of

236 diameter for CTX); intermediate (I: 11-15 mm of diameter for SXT; 12-14 mm of

237 diameter for TE; 13-14 mm of diameter for CN; 13-17 mm of diameter for C; 14-16

238 mm of diameter for AMP; 14-18 mm of diameter for NA, 16-20 mm of diameter for

239 CIP, 18-20 mm of diameter for CAZ and 23-25 mm of diameter for CTX); or

240 resistant (R: ≤10 mm of diameter for SXT; ≤11 mm of diameter for TE; ≤12 mm of

241 diameter for CN and C; ≤13 mm of diameter for AMP and NA; ≤15 mm of diameter

242 for CIP; ≤17 mm of diameter for CAZ and ≤22 mm of diameter for CTX).

243

244 2.6 Statistical Analyses

245 At least two separate experiments, each with triplicates, were conducted for each

246 assay type. Statistical analyses were performed using SPSS software (version

247 22.0.0.0, SPSS Inc., Chicago, IL). Results were analyzed by the Chi-square (X2)

248 test to determine associations between phylogenetic groups, type of sample,

249 biofilm formation and antibiotic resistance. Differences were considered significant

250 if the P value was < 0.05.

251

252 3. Results

253 3.1 Phylogroups of E. coli isolates

254 Phylogroup analyses of the 341 isolates indicated that 293 (86%) isolates

255 belonged to phylogroup A; 25 (7.3%) to B1; 3 (0.9 %) to B2; and 20 (5.9%) to D

11
256 (Table 1). These groups were sub-classified into 7 subgroups as follow: 199

257 (58.4%) isolates belonged to phylogroup A0; 94 (27.6 %) to A1; 25 (7.3%) to B1; 1

258 (0.3%) to B22; 2 (0.6%) to B23; 20 (5.9%) to D1; whereas none of the strains

259 analyzed belonged to genotype D2 (Table 1). A0 was the phylogroup with highest

260 prevalence (58.4%) whereas B22 showed the lowest prevalence (0.3%).

261 Phylogroups A0, A1, B1, and D1 were represented by isolates from all the three

262 produce chains analyzed, while the subgroup B22 was detected in only one sample

263 (0.3%, from water source of a tomato farm) and the subgroup B23 was detected in

264 two samples (0.6%; from cantaloupe at distribution and soil from jalapeño farm).

265 The most prevalent phylogroup was A0 with 60.7%, 57.1% and 57.1% of all

266 isolates from jalapeño pepper, tomato and cantaloupe chains, respectively (Table

267 1). However, phylogroup A1 was predominant among isolates from produce

268 (before harvest) and soil of jalapeño pepper, and from water (source), farmworkers

269 hands (distribution), and produce at packaging point of cantaloupe. The phylogroup

270 B1 was predominant in isolates of the cantaloupe at their distribution point (Table

271 1). In general, no statistically significant association (P>0.05) was observed

272 between phylogroup and origin of the isolates, or between phylogroup and type of

273 produce. No differences among the collection points of the water source, produce

274 at distribution, produce pre-harvest and soil were detected.

275

276 3.2 Pathotypes of E. coli isolates

277 From the 341 strains of E. coli analyzed, only 1.2% (4 isolates) were positive for at

278 least one of the tested genes. Three isolates (0.9%, two from the source water

279 from a tomato farm, and one from the soil from a jalapeño pepper farm) were

12
280 positive for the eae gene and belonged to the EPEC pathotype. However, these

281 strains were classified as atypical EPEC (aEPEC) since only the eae (intimin) gene

282 was amplified, whereas the bfpA gene was not detected. Another isolate (0.3%,

283 from the soil from a jalapeño pepper farm) was positive for the lt gene, and was

284 classified as pathotype ETEC. Other pathotypes were not detected. The EPEC

285 strains detected belonged to the A0, B22, and B23 phylogroups, whereas the ETEC

286 strain was grouped as phylogroup A1.

287

288 3.3 Biofilm formation by strains

289 Biofilm formation was analyzed on 340 isolates. 117 (34.4%) formed strong, 80

290 (23.5%) moderate, and 61 (17.9%) weak biofilms; 82 (24.1%) did not form biofilms

291 (Table 2). Most strains classified as strong biofilm formers (40.3 %) were isolated

292 from tomato farms, followed by those isolated from jalapeño and cantaloupe farms

293 (31.9 and 25.7%, respectively). Strains that formed moderate and weak biofilms

294 were mainly isolated from cantaloupe farms (24.3 and 20%, respectively) followed

295 by tomato (24 and 16.9%, respectively) and jalapeño pepper farms (22.4 and

296 18.1%, respectively, Tables 2 and 3). Strains that did not form biofilms were mainly

297 from cantaloupe farms (30%), followed by jalapeño pepper and tomato farms (27.6

298 and 18.8%, respectively) (Table 2). Strong biofilm forming isolates were present in

299 water from tomato and jalapeño farms, whereas the water from cantaloupe farms

300 contained mainly weak biofilm forming organisms. In the case of hands, produce,

301 and soil isolates, the ability of isolates to form biofilms varied (Table 2). A similar

302 pattern was observed between the degree of biofilm formation (strong, moderate,

303 weak and non-forming) and phylogroup (Table 3): A0 was the most common for

13
304 each biofilm-forming category, followed by A1. The relationships between

305 pathotypes and biofilm formation were inconsistent. The EPEC isolates (one from

306 source water of tomato farm, and one from soil of jalapeño farm) produced weak

307 biofilms, whereas another isolate from the source water of a tomato farm produced

308 moderate biofilms. The ETEC isolate from the soil of a jalapeño pepper farm

309 formed a strong biofilm.

310

311 3.4 Antibiotic resistance profiles of E. coli isolates

312 Of the 341 strains analyzed, the greatest proportion of strains were resistant to

313 tetracycline (79/341, 23.2%), followed by ampicillin (68/341,19.9%), ceftazidime

314 (39/341,11.4%), chloramphenicol (31/341, 9.1%), sulfamethoxazole/trimethoprim

315 (29/341, 8.5%), cefotaxime (27/341, 7.9%), nalidixic acid (24/341, 7%), gentamicin

316 (23/341, 6.7%) and ciprofloxacin (4/341, 1.2%; Table 4). It is noteworthy that only 4

317 isolates (all from tomato at the packing point) were resistant to ciprofloxacin.

318 Isolates from tomato and jalapeño were more resistant to antibiotics than

319 those from cantaloupe. Isolates from the tomato chain showed greater resistance

320 to antibiotics, especially to ampicillin and tetracycline (29.9 and 28.6% of isolates,

321 respectively). Almost 24% of isolates from jalapeño samples showed resistance

322 against tetracycline, whereas strains isolated from the cantaloupe chain generally

323 showed lower resistance against all antibiotics analyzed (Table 4).

324 Multidrug resistance, defined as resistance to two or more classes of

325 antimicrobial agents, was observed mainly in strains isolated from jalapeño and

326 tomato chains, whereas most isolates from cantaloupe chains showed resistance

327 to only one or two antibiotics (Table 4). Bacterial isolates with multidrug resistance

14
328 came from all points of the production chain; eight isolates from the tomato chain

329 and four from jalapeño pepper chains showed resistance to six antibiotics. An

330 interesting finding was that six strains isolated from tomato (3 at the distribution

331 point and 3 at packaging) showed resistance against at least 5 of the antibiotics

332 analyzed (Table 4).

333 Of the multidrug resistant isolates (6-7 antibiotics), seven corresponded to

334 phylogroup A1, and five isolates to A0; these were isolated mainly from tomato and

335 jalapeño pepper farms. The ETEC isolate from the soil of a jalapeño pepper farm

336 (phylogroup A1) was resistant to four antibiotics, whereas the EPEC isolates were

337 sensitive to all antibiotics tested (data not shown).

338

339 4. Discussion

340 Escherichia coli is an important inhabitant of the gastrointestinal tract of humans

341 and warm-blooded animals; however, it also can be present, in a transient stage,

342 outside the host and is capable of contaminating water and soil (Tenaillon et al.

343 2010).

344 Comparative genomic studies of E. coli strains shows that only around 2,000 of the

345 4721 genes are conserved (called core genoma, Hendrickson 2009). The rest of

346 these genes show a high degree of genomic plasticity, readily gaining and losing

347 genes (Rasko et al. 2008; Touchon et al. 2009). Because of this, a genomic

348 classification of E. coli strains has been used as a simple and inexpensive method

349 for assigning E. coli isolates to a specific phylogroup. Evidence suggests that

350 organisms of the same phylogroup share phenotypic and genotypic characteristics,

351 ecological niches, life history traits and disease causing ability (Tenaillon et al.

15
352 2010). Several phylogroups have been found in specific hosts and appear to

353 display the same level of environmental adaptability (Bergholz et al. 2011).

354 In this study, of the 341 isolates analyzed, phylogroup A was the most

355 prevalent (86%), whereas phylogroup B2 (5.9%) was the least commonly found.

356 When the subgroups were analyzed, the most common were A0 and A1, while the

357 least common subgroup was B22. These results differ from a previous report from

358 the Mid-Atlantic region of the USA, where 445 E. coli strains isolated from water,

359 animal bodies, and human and animal (pets and zoo) feces were analyzed, and

360 31% belonged to phylogroup B1, 26% to A, 25% to D and 17% to B2 (Higgins et al.

361 2007).

362 Most extra-intestinal E. coli strains belongs to subgroup B2 and, to a lesser

363 extent, to group D (Johnson and Russo 2002), and these appear to have a greater

364 propensity to carry more virulence factors than members of phylogroups A and B1,

365 although some members of the latter have been classified as diarrheogenic strains

366 (Clermont et al. 2000; Le Gall et al. 2007). Several reports indicate that strains

367 from phylogroup B1 persist in the environment (soils and waters) (Julian et al.

368 2015; Walk et al. 2007).

369 The abundance of strains belonging to phylogroup A found in this study, is in

370 agreement with Walk et al. 2007 which reported that phylogroups A and B are

371 found in environmental samples. The phylogroups A and B1 are considered sister

372 groups, whereas subgroup B2 is the most primitive taxon in terms of branching

373 pattern (Lecointre et al. 1998). Humans can host all the phylogroups (majors and

374 subgroups) with exception of the subgroup A0, with phylogroups A and B2 being

375 those of higher prevalence. In fact, the presence of phylogroup B23 has been

16
376 suggested as an indicator of contamination by human feces (Carlos et al. 2010). In

377 our study, only two isolates belonged to the B23 phylogroup, one isolate originating

378 from the distribution point of cantaloupe and the other from soil of jalapeño pepper

379 (which corresponded to pathotype EPEC).

380 Detection of potentially pathogenic E. coli strains in the fresh produce

381 production chain is very important. Contamination of water and soil could come

382 from the feces of humans, wild and domestic animals. In our study, only 1.2% of

383 the isolates were positive for an E. coli pathotype; three of them (two isolated from

384 water of tomato farm and one from soil of a jalapeño pepper farm) belonged to the

385 EPEC pathotype, and one (isolated from soil of a jalapeño farm) to the ETEC

386 pathotype. The phylogroups of the EPEC isolates were A0, B22, and B23, whereas

387 the ETEC strain belonged to phylogroup A1. Although in this study no pathogens

388 were detected in final produce samples, the risk of contamination exists, since in-

389 field vegetables can come into contact with soil and water (Franz and van Bruggen

390 2008; García-Heredia et al. 2013). ETEC strains have been consistently found in

391 surface waters, leafy vegetables, serrano and jalapeño pepper (Begum et al. 2005;

392 Cerna-Cortes et al. 2012; Lothigius et al. 2008; Singh et al. 2010). Infection with

393 ETEC is the leading cause of travelers' diarrhea and an important cause of

394 diarrheal disease in lower-income countries, especially among children (CDC

395 2014).

396 Biofilms are communities of microorganisms attached to surfaces by means

397 of polysaccharides, proteins and nucleic acids (Sauer et al. 2007). Thirty-four

398 percent of the strains analyzed in this study showed the ability to form strong

399 biofilms, particularly isolates from jalapeño and tomato farms; however, 24% of the

17
400 all isolates did not form biofilms at all. Most isolates from produce during harvest

401 and from soil formed strong biofilm (56.3 and 47.4%, respectively). Since biofilm

402 formation has been associated with tolerance to stress conditions and to increased

403 pathogenicity (Naves et al. 2008), the ability of E. coli strains to form biofilms in the

404 environment may represent a survivorship strategy, allowing them to persist longer

405 and as a consequence, have greater likelihood of contaminating food (Méric et al.

406 2013). However, the in-vitro test used for biofilm formation does not directly predict

407 environmental persistence of bacteria, since biofilm formation is not the only

408 determinant of persistence. In produce, biofilms can protect against different

409 stresses such as desiccation and bactericidal agents (Morris and Monier 2003). It

410 is recognized that soil represents a special reservoir for microorganisms, including

411 those potentially pathogenic for humans (Opelt et al. 2007).

412 Many members of phylogroup B1 isolated from plants can form strong

413 biofilms, produce high quantity of extracellular matrix and utilize sucrose, which

414 can aid in colonization of surfaces (Méric et al. 2013). These usually form

415 significantly more biofilm and extracellular matrix than mammalian-associated

416 strains (Méric et al. 2013). In our study, many strains belonging to phylogroups A0,

417 A1, B1 and D1 formed strong biofilms (50.4, 31.6, 11.1 and 6.8%, respectively).

418 Agarwal et al. (2013) studied the phylogenetic background, virulence genotypes

419 and biofilm formation of 172 E. coli strains, and found no significant differences in

420 intensity and biofilm formation ability when comparing phylogroups or virulence

421 scores. In agreement with these results, the EPEC isolates of this study showed

422 weak and moderate biofilm production, whereas the ETEC strain formed a strong

423 biofilm.

18
424 The evolution of antibiotic resistance among bacteria results from complex

425 interactions with the environment where strains grow exposed to antibiotics and

426 genomic constraints (Davies and Davies, 2010). Bacteria frequently are exposed to

427 antibiotics due to their extensive use in human and veterinary medicine, providing

428 a source of antibiotic resistance genes even to environmental bacteria (Walsh and

429 Duffy 2013). E. coli resistant to tetracycline, ampicillin and sulfonamide have been

430 detected from water isolates of aquatic ecosystems in the USA (Hamelin et al.

431 2007). E. coli isolated from water and soil for agricultural use in Sinaloa, Mexico

432 showed resistance to tetracycline, streptomycin and gentamicin (Lopez-Cuevas et

433 al. 2009). Our study suggested that isolates from the fresh produce production

434 chain were most often resistant to tetracycline (23.2%) and ampicillin (19.9%).

435 Multidrug resistance was observed mainly in the isolates from jalapeño and tomato

436 chains, whereas most isolates from the cantaloupe chain showed resistance to one

437 or two antibiotics only. Differences observed in farms included less employee

438 turnover rate in cantaloupe farms and a dryer region compared to tomato and

439 pepper farms (data not shown).

440 Twelve strains isolated from tomato and jalapeño farms (six from produce

441 packaging, four from farmworkers hands at the distribution point, one from source

442 water, and one from irrigation water) belonged primarily to phylogroups A1 and A0

443 and showed resistance to six or seven antibiotics (mainly nalidixic acid,

444 tetracycline, gentamicin, ampicillin and chloramphenicol). Our results agree with

445 previous reports that indicated that phylogroup A is detected as the most prevalent

446 in water samples, and could exhibit multidrug resistance against sulfamethoxazole,

447 tetracycline, streptomycin and β-lactam antibiotics (Hamelin et al. 2007). E.

19
448 coli isolates from ready-to-eat salads in Portugal, belonging to phylogroups A, B1

449 and D, showed resistance to tetracycline, streptomycin, sulfamethoxazole,

450 trimethoprim, ampicillin, nalidixic acid, and less frequently to ciprofloxacin and

451 chloramphenicol (Campos et al. 2013). Multi-drug resistant E. coli isolated from

452 lettuce, irrigation water and soil from farms in Belgium also exhibited higher

453 resistance to ampicillin and cephalothin and other antibiotics (Holvoet et al. 2013).

454 It is important to recognize that antibiotic resistant bacteria in multispecies biofilms

455 could protect other non-resistant strains in this environment, and that biofilms can

456 serve as reservoirs or sources of antibiotic resistance genes to be transferred to

457 other non-resistant bacteria (Balcazar et al. 2015).

458 In general, our study showed no real association between E. coli

459 pathotype or phylogroup with respect to the origin of the strains, their patterns of

460 resistance to antibiotics, and their ability to form biofilms. However, multidrug

461 resistance was observed, with more than 3.5% of the strains resistant to six or

462 seven antibiotics. Although in low proportion, potential pathogenic E. coli strains

463 were detected, and a relatively high percentage of the isolates were able to form

464 strong biofilms. The presence in the fresh produce production chain of E. coli

465 strains having multi-drug resistance, their ability to form biofilms, and the potential

466 pathogenicity of various isolates, could present significant opportunities for

467 contamination that results in health risks for consumers. Efforts to keep E. coli out

468 of the human food chain continue to be necessary throughout the farm-to-fork

469 continuum.

470

471 Acknowledgments

20
472 This research was supported by the Consejo Nacional de Ciencia y Tecnología de

473 México (CONACYT) grant number 183620 and the National Institute of Food and

474 Agriculture, U.S. Department of Agriculture, under award numbers 2010-85212-

475 20608, 2011-67012-30762, and 2015-67017-23080. We are thankful to CONACYT

476 for the scholarship granted to Hesperia Andrea Corzo Ariyama.

477

478 Conflict of interest

479 The authors declare no conflicts of interest.

480

481 References

482 Agarwal, J., Mishra, B., Srivastava, S., Srivastava, R. 2013. Genotypic

483 characteristics and biofilm formation among Escherichia coli isolates from

484 Indian women with acute cystitis. T. Roy. Soc. Trop. Med. H. 107, 183-187.

485 Amabile-Cuevas, C. 2010. Antibiotic resistance in Mexico: a brief overview of the

486 current status and its causes. J. Infect. Dev. Ctries. 4, 126-131

487 Balcazar, J. L., Subirats, J., Borrego, C. M. 2015. The role of biofilms as

488 environmental reservoirs of antibiotic resistance. Front. Microbiol. 6, 1-9.

489 Begum, Y. A., Talukder, K. A., Nair, G. B., Qadri, F., Sack, R. B., Svennerholm, A.

490 M. 2005. Enterotoxigenic Escherichia coli isolated from surface water in

491 urban and rural areas of Bangladesh. J. Clin. Microbiol. 43, 3582-3583.

492 Bergholz, P.W., Noar, J.D., Buckley, D.H. 2011. Environmental patterns are

493 imposed on the population structure of Escherichia coli after fecal

494 deposition. Appl. Environ. Microbiol. 77, 211-219.

21
495 Boehme, S., Werner, G., Klare, I., Reissbrodt, R., Witte, W. 2004. Occurrence of

496 antibiotic‐resistant enterobacteria in agricultural foodstuffs. Mol. Nutr. Food.

497 Res. 48, 522-531.

498 Campos, J., Mourão, J., Pestana, N., Peixe, L., Novais, C., Antunes, P. 2013.

499 Microbiological quality of ready-to-eat salads: an underestimated vehicle of

500 bacteria and clinically relevant antibiotic resistance genes. Int. J. Food

501 Microbiol., 166, 464-470.

502 Canizalez-Roman, A., Gonzalez-Nuñez, E., Vidal, J. E., Flores-Villaseñor, H.,

503 León-Sicairos, N. 2013. Prevalence and antibiotic resistance profiles of

504 diarrheagenic Escherichia coli strains isolated from food items in

505 northwestern Mexico. Int. J. Food. Microbiol. 164, 36-45.

506 Carlos, C., Pires, M. M., Stoppe, N. C., Hachich, E. M., Sato, M. I., Gomes, T.

507 Amaral L. Ottoboni, L. M. 2010. Escherichia coli phylogenetic group

508 determination and its application in the identification of the major animal

509 source of fecal contamination. BMC Microbiol. 10, 161.

510 CDC 2014. Center for Disease Control and Prevention. Enterotoxigenic E. coli

511 (ETEC). CDC [www Document] URL: https://www.cdc.gov/ecoli/etec.html

512 (access 06.16.2016)

513 Cerna-Cortes, J. F., Gómez-Aldapa, C. A., Rangel-Vargas, E., del Refugio Torres-

514 Vitela, M., Villarruel-López, A., Castro-Rosas, J. 2012. Presence of some

515 indicator bacteria and diarrheagenic E. coli pathotypes on jalapeno and

516 serrano peppers from popular markets in Pachuca City, Mexico. Food

517 Microbiol. 32, 444-447.

22
518 Clermont, O., Bonacorsi, S., Bingen, E. 2000. Rapid and simple determination of

519 the Escherichia coli phylogenetic group. Appl. Environ. Microb. 66, 4555-

520 4558.

521 Clermont, O., Christenson, J. K., Denamur, E., Gordon, D. M. 2013. The Clermont

522 Escherichia coli phylo‐typing method revisited: improvement of specificity

523 and detection of new phylo‐groups. Env. Microbiol. Rep. 5, 58-65.

524 Clinical and Laboratory Standards Institute (CLSI). 2012. Performance Standards

525 for Antimicrobial Disk Susceptibility Tests; Approved Standard—Eleventh

526 Edition. CLSI document M02-A11. Wayne, PA.

527 Clinical and Laboratory Standards Institute (CLSI). 2014. Performance Standards

528 for Antimicrobial Susceptibility Testing; Twenty-Fourth Informational

529 Supplement. CLSI document M100-S24. Wayne, PA.

530 Costerton, J. W., Lewandowski, Z., Caldwell, D. E., Korber, D. R., Lappin-Scott, H.

531 M. 1995. Microbial biofilms. Annu. Rev. Microbiol. 49, 711-745.

532 Davies, J., and Davies, D. 2010. Origins and evolution of antibiotic resistance.

533 Microbiol. Mol. Biol. R. 74, 417-433.

534 Dombek, P. E., Johnson, L. K., Zimmerley, S. T., Sadowsky, M. J. 2000. Use of

535 repetitive DNA sequences and the PCR to differentiate Escherichia coli

536 isolates from human and animal sources. Appl. Environ. Microb. 66, 2572-

537 2577.

538 FDA 2009. Salmonella Saintpaul outbreak. [WWW Document] URL:

539 https://www.fda.gov/newsevents/publichealthfocus/ucm179116.htm

540 (accessed 10.01.2017)

23
541 FDA 2012. Food and Drug Administration. Bad bug book: Foodborne pathogenic

542 microorganisms and natural toxins. Chapter: Pathogenic E. coli group. 2012.

543 FDA [WWW Document] URL:

544 https://www.fda.gov/downloads/food/foodsafety/foodborneillness/foodborneil

545 lnessfoodbornepathogensnaturaltoxins/badbugbook/ucm297627.pdf

546 (accessed 06.16.2016)

547 FDA 2017. FDA Investigates multiple Salmonella outbreak strains linked to

548 papayas. [WWW Document] URL:

549 https://www.fda.gov/Food/RecallsOutbreaksEmergencies/Outbreaks/ucm56

550 8097.htm (accessed 10.01.2017)

551 Franz, E., and van Bruggen, A. H. 2008. Ecology of E. coli O157: H7 and

552 Salmonella enterica in the primary vegetable production chain. CRC Cr.

553 Rev. Microbiol. 34, 143-161.

554 García-Heredia, A., Orue, N., Heredia, N., and García, S. 2013. Efficacy of citrus-

555 based disinfectants to inhibit growth, swarming, and biofilm formation of

556 Salmonella and decontaminate parsley. J. Food. Agric. Environ. 11, 1295-

557 1299.

558 Garcia-Heredia, A., S. Garcia, J.A. Merino-Mascorro, P. Feng, N. Heredia. 2016.

559 Natural plant products inhibits growth and alters the swarming motility,

560 biofilm formation, and expression of virulence genes in enteroaggregative

561 and enterohemorrhagic Escherichia coli. Food Microbiol. 59, 124-132.

562 Gordon, D. M., and Cowling, A. 2003. The distribution and genetic structure of

563 Escherichia coli in Australian vertebrates: host and geographic effects.

564 Microbiology. 149, 3575-3586.

24
565 Gordon, D. M., Clermont, O., Tolley, H., Denamur, E. 2008. Assigning Escherichia

566 coli strains to phylogenetic groups: multi‐locus sequence typing versus the

567 PCR triplex method. Environ. Microbiol. 10, 2484-2496.

568 Gould, L., Kline, J., Monahan, C., and Vierk, K. 2017. Outbreaks of Disease

569 Associated with Food Imported into the United States, 1996–2014. Emerg

570 Infec Dis. 23: 525-528.

571 Hamelin, K., Bruant, G., El-Shaarawi, A., Hill, S., Edge, T. A., Fairbrother, J.,

572 Harel., C., Maynard. L., Masson, Brousseau, R. 2007. Occurrence of

573 virulence and antimicrobial resistance genes in Escherichia coli isolates

574 from different aquatic ecosystems within the St. Clair River and Detroit River

575 areas. Appl. Environ. Microbiol. 73, 477-484.

576 Hendrickson, H. 2009. Order and disorder during Escherichia coli divergence.

577 PLoS Genet. 5, 1. e1000335

578 Heredia, N., Caballero, C., Cardenas, C., Molina, K., Garcia, R., Solis, L.,

579 Burrowes, V., Bartz, F., Fabiszewski de Aceituno, A., Jaykus, L. A, Garcia,

580 S. León. J. 2016. Microbial indicator profiling of fresh produce and

581 environmental samples from farms and packing facilities in northern Mexico.

582 J. Food Protect. 79, 1197-1209.

583 Higgins, J., Hohn, C., Hornor, S., Frana, M., Denver, M., Joerger, R. 2007.

584 Genotyping of Escherichia coli from environmental and animal samples. J.

585 Microbiol. Meth. 70, 227-235.

25
586 Holvoet, K., Sampers, I., Callens, B., Dewulf, J., Uyttendaele, M. 2013. Moderate

587 prevalence of antimicrobial resistance in Escherichia coli isolates from

588 lettuce, irrigation water, and soil. Appl. Environ. Microbiol. 79, 6677-6683.

589 Johnson, J. R., and Russo, T. A. 2002. Extraintestinal pathogenic Escherichia coli:

590 “the other bad E coli”. J. Lab. Clin. Med. 139, 155-162.

591 Julian, T. R., Islam, M. A., Pickering, A. J., Roy, S., Fuhrmeister, E. R., Ercumen,

592 A., Harris, A., Bisahi, J. Schwab, K. J. 2015. Genotypic and phenotypic

593 characterization of Escherichia coli isolates from feces, hands, and soils in

594 rural Bangladesh via the Colilert Quanti-Tray system. Appl. Environ.

595 Microbiol. 81, 1735-1743.

596 Kaper, J. B., Nataro, J. P., and Mobley, H. L. 2004. Pathogenic Escherichia

597 coli. Nat. Rev. Microbiol. 2,123-140.

598 Le Gall, T., Clermont, O., Gouriou, S., Picard, B., Nassif, X., Denamur, E.,

599 Tenaillon, O. 2007. Extraintestinal virulence is a coincidental by-product of

600 commensalism in B2 phylogenetic group Escherichia coli strains. Mol. Biol.

601 Evol. 24, 2373-2384.

602 Lecointre, G., Rachdi, L., Darlu, P., and Denamur, E. 1998. Escherichia coli

603 molecular phylogeny using the incongruence length difference test. Mol.

604 Biol. Evol. 15, 1685-1695

605 Lee, C. 2011. Genotyping Escherichia coli isolates from duck, goose, and gull fecal

606 samples with phylogenetic markers using multiplex polymerase chain

607 reaction for application in microbial source tracking. JEMI. 15, 130-135.

26
608 León, B. D., Gómez‐Aldapa, C. A., Rangel‐Vargas, E., Vázquez‐Barrios, E.,

609 Castro‐Rosas, J. 2013. Frequency of indicator bacteria, Salmonella and

610 diarrhoeagenic Escherichia coli pathotypes on ready‐to‐eat cooked

611 vegetable salads from Mexican restaurants. Lett. Appl. Microbiol. 56,414-

612 420.

613 Liu, C., Hofstra, N., Franz, E. 2013. Impacts of climate change on the microbial

614 safety of pre-harvest leafy green vegetables as indicated by Escherichia coli

615 O157 and Salmonella spp. Int. J. Food. Microbiol. 163, 119-128.

616 Lopez-Cuevas, O., León Félix, J., Jimenez Edeza, M., and Chaidez Quiroz, C.

617 2009. Detection and antibiotic resistance of Escherichia coli and Salmonella

618 in water and agricultural soil. Rev. Fitotec. Mex. 32, 119-126.

619 Lothigius, Å., Janzon, A., Begum, Y., Sjöling, Å., Qadri, F., Svennerholm, A. M.,

620 Bölin, I. 2008. Enterotoxigenic Escherichia coli is detectable in water

621 samples from an endemic area by real‐time PCR. J. Appl. Microbiol. 104,

622 1128-1136.

623 Méric, G., Kemsley, E. K., Falush, D., Saggers, E. J., Lucchini, S. 2013.

624 Phylogenetic distribution of traits associated with plant colonization in

625 Escherichia coli. Environ. Microbiol. 15, 487-501.

626 Morris, C. E., and Monier, J. M. 2003. The ecological significance of biofilm

627 formation by plant-associated bacteria. Annu. Rev. Phytopathol. 41, 429-

628 453.

629 Naves, P., del Prado, G., Huelves, L., Gracia, M., Ruiz, V., Blanco, J., Dahbi, G.,

630 Blanco, M., Ponte, M. Soriano, F. 2008. Correlation between virulence

27
631 factors and in vitro biofilm formation by Escherichia coli strains. Microb.

632 Pathogenesis. 45, 86-91.

633 Nowrouzian, F. L., Adlerberth, I., Wold, A. E. 2006. Enhanced persistence in the

634 colonic microbiota of Escherichia coli strains belonging to phylogenetic

635 group B2: role of virulence factors and adherence to colonic cells. Microbes

636 Infec. 8, 834-840.

637 Opelt, K., Berg, C., and Berg, G. 2007. The bryophyte genus Sphagnum is a

638 reservoir for powerful and extraordinary antagonists and potentially

639 facultative human pathogens. FEMS Microbiol. Ecol. 61, 38-53.

640 Radosavljevic, V., Finke, E. J., Belojevic, G. 2016. Analysis of Escherichia coli

641 O104: H4 outbreak in Germany in 2011 using differentiation method for

642 unusual epidemiological events. Cent. Eur. J. Publ. Heal. 24, 9-15.

643 Rasko, D. A., Rosovitz, M. J., Myers, G. S., Mongodin, E. F., Fricke, W. F., Gajer,

644 P., Crabtree, J., Sebaihia, M., Thomson, N. R., Chaudhuri, R., Henderson, I.

645 R., Sperandio, V. Ravel, J. 2008. The pangenome structure of Escherichia

646 coli: comparative genomic analysis of E. coli commensal and pathogenic

647 isolates. J. Bacteriol. 190, 6881-6893.

648 Russo, T. A., and Johnson, J. R. 2000. Proposal for a new inclusive designation for

649 extraintestinal pathogenic isolates of Escherichia coli: ExPEC. J. Infect.

650 Dis. 181, 1753-1754.

651 Sauer, K., Rickard, A.H., Davies, D.G. 2007. Biofilms and biocomplexity. Microbe

652 2, 347–353.

28
653 Schwaiger, K., Helmke, K., Hölzel, C. S., Bauer, J. 2011. Antibiotic resistance in

654 bacteria isolated from vegetables with regards to the marketing stage (farm

655 vs. supermarket). Int. J. Food. Microbiol. 148, 191-196.

656 Singh, G., Vajpayee, P., Ram, S., Shanker, R. 2010. Environmental reservoirs for

657 enterotoxigenic Escherichia coli in south Asian Gangetic riverine

658 system. Environ. Sci. Technol. 44, 6475-6480.

659 Teh, K. H., Flint, S., French, N. 2010. Biofilm formation by Campylobacter jejuni in

660 controlled mixed-microbial populations. Int. J. Food Microbiol. 143, 118-124.

661 Tenaillon, O., Skurnik, D., Picard, B., Denamur, E. 2010. The population genetics

662 of commensal Escherichia coli. Nat. Rev. Microbiol. 8, 207-217.

663 Touchon, M., Hoede, C., Tenaillon, O., Barbe, V., Baeriswyl, S., Bidet, P., Bingen,

664 E., Bonacorsi, S., Bouchier, C., Bouvet, O., Calteau, A., Chiapello, H.,

665 Clermont, O., Cruveiller. S., Danchin, A., Diard, M., Dossat, C., Karoui, M.,

666 Frapy, E., Garry, L., Marc Ghigo. J., Gilles, A. M., Johnson, J., Le

667 Bouguénec, C., Lescat, M., Mangenot, S., Martinez-Jéhanne, V., Matic. I.,

668 Nassif, X., Oztas, S., Agnès Petit, M., Pichon, Z. R., Saint Ruf, C.,

669 Schneider, D., Tourret, J., Vacherie, B., Vallenet, D., Médigue, C., Rocha, E.

670 P. C., Denamur, E. 2009. Organized genome dynamics in the Escherichia

671 coli species results in highly diverse adaptive paths. PLoS Genet. 5, 1.

672 e1000344.

673 Vidal, M., Kruger, E., Durán, C., Lagos, R., Levine, M., Prado, V., Toro, C., Vidal,

674 R. 2005. Single multiplex PCR assay to identify simultaneously the six

675 categories of diarrheagenic Escherichia coli associated with enteric

676 infections. J. Clin. Microbiol. 43, 5362-5365.

29
677 Walk, S. T., Alm, E. W., Calhoun, L. M., Mladonicky, J. M., and Whittam, T. S.

678 2007. Genetic diversity and population structure of Escherichia coli isolated

679 from freshwater beaches. Environ. Microbiol. 9, 2274-2288.

680 Walsh, F., and Duffy, B. 2013. The culturable soil antibiotic resistome: a community

681 of multi-drug resistant bacteria. PLoS One. 8, e65567.

682 Wang, R. F., Cao, W. W., Cerniglia, C. E. 1997. A universal protocol for PCR

683 detection of 13 species of foodborne pathogens in foods. J. Appl.

684 Microbiol. 83, 727-736.

30
 Table 1. Phylogroups of E coli strains isolated along the in-field production chain of jalapeño pepper, tomato and cantaloupe.
Sample Number of isolates (jalapeño, Jalapeño pepper Tomato Cantaloupe
type tomato, cantaloupe) % Phylogroup (%)
A0 (58.6), A1 (34.5), B1 (5.2),
Source (22, 58, 5) A0 (54.5)*, A1 (22.7), D1 (22.7) A0 (20), A1 (80)
24.9 B22 (2)
Water
A0 (54.5), A1 (27.3), B1 (15.9),
Irrigation hose (9, 44, 15) A0 (66.7), A1 (22.2), B1 (11.1) A0 (73.3), B1 (6.7), D1 (20)
19.9 D1 (2.3)
Harvest (29, 8, 15) 15.2 A0 (69), A1 (17.2), D1 (13.8) A0 (62.5), A1 (25), B1 (12.5) A0 (66.7), A1 (20), B1 (13.3)
A0 (22.2), A1 (55.6), B1
Hands A0 (42.9), A1 (14.3), B1 (28.6),
Distribution (14, 7, 9) 8.8 A0 (71.4), A1 (28.6) (11.1),
D1 (14.3)
D1 (11.1)
Packaging (5, 0, 4) 2.6 A0 (100) ND A0 (75), B1 (25)

Before harvest (19, 7, 6) 9.4 A0 (47.4), A1 (52.6) A0 (71.4), B1 (28.6) A0 (83.3), B1 (16.7)

During harvest (3, 10, 3) 4.7 A0 (66.7), A1 (33.3) A0 (60), A1 (10), D1 (30) A0 (66.7), A1 (33.3)
Produce
A0 (20), A1 (20), B1(40),
Distribution (6, 3, 5) 4.1 A0 (83.3), B1 (16.7) A1 (100)
B23 (20)
A0 (70), A1 (30) A0 (40), A1 (60)
Packaging (1, 10, 5) 4.7 A0 (100)

Around produce-plant sampled

Soil 5.6 A0 (11.1), A1 (77.8), B23 (11.1) A0 (57.1), A1 (14.3), D1 (28.6) A0 (100)
(9, 7, 3)
100
A0 (58.4),
A1 (27.6),
A0 (60.7), A1 (29.1), B1(1.7), A0 (57.1), A1 (27.9), B1(9.7), A0 (57.1), A1 (24.3), B1(11.4),
Total (117, 140, 70 = 341) B1 (7.3),
B23 (0.9), D1 (7.7) B22 (0.6), D1 (4.5) B23 (1.4), D1 (5.7)
B22 (0.3),
B23 (0.6),
D1 (5.9)
* bold letter: phylogroup with the highest percentage of isolates; ND: not detected

31
Table 2. Biofilm formation by E. coli strains isolated along the in-field production chain of jalapeño pepper, tomato and cantaloupe.

Origin of isolates Jalapeño pepper Tomato Cantaloupe Total per sample type
Sample (number of isolates
type from jalapeño, tomato % strains forming biofilm (S, M, W, NF)1
and cantaloupe)
S (34/85, 40)3, M (19/85,
S (45.5)2, M (22.7), W S (39.7), M (24.1), W S (20), M (0), W (60),
Source (22, 58, 5) (18.2), NF (13.6) (6.9), NF (29.3) NF (20)
22.4), W (11/85, 12.9),
NF (21/85, 24.7)
Water S (24/68, 35.3), M
S (44.4), M (11.1), W S (36.4), M (29.5), W S (26.7), M (26.7), W
Irrigation (9, 44, 15) (22.2), NF (22.2) (22.7), NF (11.4) (26.7), NF (20)
(18/68, 26.5), W (16/68,
23.5), NF (10/68, 14.7)
S (14/51, 27.5), M
S (25), M (28.6), W S (37.5), M (25), W (25), S (26.7), M (46.7), W
Harvest (28, 8, 15) (10.7), NF (35.7) NF (12.5) (6.7), NF (20)
(13/51, 33.3), W (6/51,
11.8), NF (14/51, 27.5)
S (10/30, 33.3), M (7/30,
S (28.6), M (28.6), W S (42.9), M (42.9), W S (33.3), M (0), W
Hands Distribution (14, 7, 9) (14.3), NF (28.6) (14.3), NF (0) (11.1), NF (55.6)
23.3), W (4/30, 13.3), NF
(9/30, 30)
S (2/9, 22.2), M (2/9,
S (20), M (40), W (20), S (25), M (0), W (25),
Packaging (5, 0, 4) NF (20)
ND
NF (50)
22.2), W (2/9, 22.2), NF
(3/9, 33.3)
S (10/32, 31.3), M (4/32,
Before harvest (19, 7, S (26.3), M (5.3), W S (71.4), M (14.3), W S (0), M (33.3), W
12.5), W (9/32, 28.1), NF
6) (21.1), NF (47.4) (14.3), NF (0) (66.7), NF (0)
(9/32, 28.1)
S (9/16, 56.3), M (4/16,
During harvest (3, 10, S (0), M (33.3), W (66.7) S (80), M (20), W (0), NF S (33.3), M (33.3), W
25), W (2/16, 12.5), NF
3) , NF (0) (0) (0), NF (33.3)
(1/16, 6.3)
Produce S (5/14, 35.7), M (2/14,
S (33), M (16.7), W (33), S (0), M (0), W (66.7), S (60), M (20), W (0),
Distribution (6, 3, 5) NF (16.7) NF (33.3) NF (20)
14.3), W (4/14, 28.6), NF
(3/14, 21.4)
S (0), M (4/16, 25), W
S (0), M (100), W (0), NF S (0), M (20), W (50), NF S (0), M (20), W (0),
Packaging (1, 10, 5) (0) (30) NF (80)
(5/16, 31.3), NF (7/16,
43.8)
S (9/19, 47.4), M (3/19,
Around plant sampled S (44.4), M (22.2), W S (57.1), M (0), W (14.3), S (33.3), M (33.3), W
Soil (0), NF (33.3)
15.8), W (2/19, 10.5), NF
(9, 7, 3) (11.1), NF (22.2) NF (28.6)
(5/19, 26.3)
S (37/116, 31.9)3, S (62/154, 40.3), S (18/70, 25.7), S (117/340, 34.4),
Total by M (26/116, 22.4), M (37/154, 24), M (17/70, 24.3), M (80/340, 23.5),
(116, 154, 70= 340)
produce W (21/116, 18.1), W (26/154, 16.9), W (14/70, 20), W (61/340, 17.9),
NF (32/116, 27.6) NF (29/154, 18.8) NF (21/70, 30) N (82/340, 24.1)
1
S: strong, M: moderate, W: weak, NF: no forming, 2Type of biofilm(s) formed by the highest number of isolates, 3: number of positive
samples/total of samples, percentage

32
Table 3. Biofilm formation and phylogroup from E. coli isolates from the chain product of jalapeño pepper, tomato and cantaloupe and their
genotypes.

Type of Jalapeño pepper Tomato Cantaloupe Total

biofilm n (%) Phylogroup n (%) Phylogroup n (%) Phylogroup # (%)
A0 (59/117, 50.4)
A0= 18 A0= 29
37/116 A0= 12 A1 (37/117, 31.6)
A1= 17 62/154 A1= 20 18/70
Strong (31.6) B1= 5 B1 (13/117, 11.1)
B1= 1 (40.3) B1= 7 (25.7)
D1= 1 D1(8/117, 6.8)
D1= 1 D1= 6
117 (34.4)
A0 (52/80, 65)
A1 (16/80, 20)
A0= 20 A0= 20 A0= 12
B1 (6/80, 7.5)
A1= 1 37/154 A1= 12 17/70 A1= 3
Moderate 26/116 B22 (1/80, 1.3)
B1= 1 (24) B1= 4 (24.3) B1= 1
(22.4) B23 (1/80, 1.3)
D1= 4 B22= 1 B23= 1
D1 (4/80, 5)
80 (23.5)
A0 (36/61, 59)
21/116 A0= 16 A0= 14 A0= 6 A1 (11/61, 18)
(18.1) A1= 1 26/154 A1= 7 14/70 A1= 3 B1 (6/61, 9.8)
Weak
B23= 1 (16.9) B1= 4 (20) B1= 2 B23 (1/61, 1.6)
D1= 3 D1= 1 D1= 3 D1 (7/61, 11.5)
61 (17.9)
A0 (51/82, 62.2)
A0= 16
32/116 29/154 A0= 25 21/70 A0= 10 A1 (30/82, 36.6)
None A1= 15
(27.6) (18.8) A1= 4 (30) A1= 11 D1 (1/82, 1.2)
D1= 1
82 (24.1)
n: Strains forming biofilm / number total of isolates by produce

#: Strains that formed the specific biofilm

33
Table 4. Antibiotic resistance of E. coli isolates from the chain product of jalapeño pepper, tomato and cantaloupe.

No isolates (% resistance)
Product Type of sample n NA TE CN CIP AMP CTX CAZ C SXT
Source water 22 4 (18.2) 4 (18.2) 3 (13.6) 3 (13.6) 2 (9.1)
Irrigation water 9 1 (11.1) 2 (22.2)
Preharvest product 19 8 (42.1) 3 (15.8) 2 (10.5) 3 (15.8) 2 (10.5)
Hands in harvest 29 2(6.9) 6 (20.7) 3 (10.3) 3 (10.3) 1 (3.4) 3 (10.3) 3 (10.3)
Jalapeño pepper During harvest product 3 1 (33.3) 1 (33.3) 1 (33.3) 1 (33.3) 1 (33.3)
Hands in distribution 14 3 (21.4) 4 (28.6) 3 (21.4) 6 (42.9) 6 (42.9) 5 (35.7) 3 (21.4)
Product in distribution 6 2 (33.3) 3 (50) 1 (16.7) 2 (33.3) 2 (33.3)
Hands in packaging 5 1 (20) 2 (40)
Product in packaging 1
Soil 9 1 (11.1) 1 (11.1) 1 (11.1) 1 (11.1)
Total 117 7 (5.9) 28 (23.9) 10 (8.5) 0 17 (14.5) 15 (12.8) 18 (15.4) 9 (7.7) 6 (5.1)
Source water 58 1 (1.7) 9 (15.5) 1 (1.7) 12 (20.7) 6 (10.3) 9 (15.5) 1 (.7) 3 (5.2)
Irrigation water 44 2 (4.5) 11 (25) 1 (2.3) 9 (20.5) 2 (4.5) 6 (13.6) 7 (15.9) 11 (25)
Preharvest product 7 5 (71.4) 1 (14.3)
Hands in harvest 8 2 (25)
During harvest product 10 5 (50) 6 (60) 1 (10) 4 (40)
Tomato
Hands in distribution 7 1 (14.3) 1 (14.3) 1 (14.3) 2 (28.6) 1 (14.3) 3 (42.9) 1 (14.3)
Product in distribution 3 3 (100) 3 (100) 3 (100) 3 (100) 3 (100)
Hands in packaging 0
Product in packaging 10 8 (80) 10 (100) 7 (70) 4 (40) 10 (100) 3 (30) 3 (30) 9 (90) 5 (50)
Soil 7 1 (14.3)
Total 154 15(9.7) 44 (28.6) 13 (8.4) 4 (2.6) 46 (29.9) 12 (7.8) 21 (13.6) 22 (14.3) 23 (14.9)
Source water 5 3 (60) 3 (60)
Irrigation water 15 1 (6.7) 1 (6.7)
Preharvest product 6
Hands in harvest 15 1 (6.7)
During harvest product 3 1 (33.3)
Cantaloupe
Hands in distribution 9
Product in distribution 5 1 (20) 1 (20)
Hands in packaging 4 2 (50)
Product in packaging 5
Soil 3
Total 70 2 (2.9) 7 (10) 0 0 5 (7.1) 0 0 0 0
TOTAL 341 24 (7) 79 (23.2) 23 (6.7) 4 (1.2) 68 (19.9) 27 (7.9) 39 (11.4) 31 (9.1) 29 (8.5)
* % resistance of total samples by produce

34
NA: nadilixic acid, TE: tetracycline, CN: gentamicin, AMP: ampicillin, CTX: cefotaxime, CAZ: ceftazidime, C: chloramphenicol, STX:
sulfamethoxazole/trimethoprim.

35