Pathogenic bacteria and their antibiotic resistance profile in irrigation water in farms from Mexico

Annually, the consumption of contaminated food causes illness in approximately 600 million individuals and around 420,000 deaths worldwide (World Health Organization 2022). Pathogens commonly associated with foodborne diseases include norovirus, non-typhoidal Salmonella spp., Shigella spp., Listeria monocytogenes, Clostridium perfringens, Staphylococcus aureus, Vibrio spp., Campylobacter jejuni/coli, and the enteropathogenic groups of Escherichia coli (E. coli) (Alegbeleye et al. 2018).

Contamination of agricultural products can occur before, during, and after harvesting. Pre-harvest contamination factors are related to soil, manure use, livestock/wildlife, and irrigation water. Agricultural water is the water used to grow fresh produce and feed livestock; it is obtained from a variety of sources, including surface water (rivers, streams, irrigation ditches, open canals, impounded water such as ponds, reservoirs, and lakes), groundwater (wells) or rainwater (locally collected water in cisterns and rain barrels). Municipal, urban, and rural water systems can also be used for agriculture (Canizalez-Roman et al. 2019).

The presence of foodborne pathogens in irrigation water is common (Heredia et al. 2016). These organisms could remain viable and survive for a very long time, as well as develop resistance to antibiotics, thereby posing a great threat to public health. Pathotypes of E. coli and Enterococcus are microorganisms that survive under these conditions (Heredia et al. 2016). Enterococcus spp. are part of the microbiota present in humans and animals and are found in high concentrations in human feces (i.e., 10⁴–10⁶ cells/g wet weight). However, when Enterococcus faecalis (E. faecalis) contaminates water, it can cause serious health problems, particularly nosocomial diseases (Rice & Baird 2017). E. coli is part of the intestinal microbiota, and several pathogenic groups have been associated with diarrhea, peritonitis, colitis, bacteremia, and urinary tract infections. These pathogenic groups have been classified into the following diarrheagenic pathotypes based on their virulence factors and pathogenicity mechanisms: enteropathogenic (EPEC), enteroaggregative (EAEC), enterotoxigenic (ETEC), diffuse adherent, enteroinvasive, and enterohemorrhagic (EHEC) (Corzo-Ariyama et al. 2019). Salmonella can also be found in the agricultural environment, and is one of the most common causes of severe foodborne diseases, according to the Centers for Disease Control & Prevention (2019). In the United States of America, it is responsible for approximately 1.35 million cases of illness, 26,500 hospitalizations, and up to 420 deaths per year.

As a major producer and exporter of fresh fruits and vegetables, Mexico must prevent the use of contaminated irrigation water for the cultivation of export products, as it would affect public health internationally. These products are often consumed raw and, if previously contaminated, can transmit infectious agents to humans. Microbial contamination of vegetables is always latent due to the processes that take place during production, harvesting, and handling (Zoellner et al. 2016). In Mexican agriculture, the soil, produce, farmworker's hands, and water have been regarded as reservoirs of pathogens and antibiotic-resistant bacteria (Heredia et al. 2016).

Antibiotic resistance is an evolving problem worldwide that impacts bacteria in the agro-environment. Currently, there is limited information available regarding this phenomenon in microbial contamination of irrigation water in Mexico. In the present study, the presence of Salmonella spp., Enterococcus spp., and diarrheagenic E. coli, as well as their antibiotic resistance, were determined in agricultural water samples obtained from northern and central Mexico.

The strains used as positive controls were E. coli O157:H7 ATCC 43895 (EHEC, kindly provided by Dr Lynne McLandsborough, University of Massachusetts, Amherst, MA, USA); E. coli ATCC 25922 (non-pathogenic, donated by Becton Dickinson Co., Mexico); E. coli O111:NM ATCC 43887 (EPEC), E. coli O78:H11 ATCC 35401 (ETEC) (both commercially acquired); E. coli 042 (EAEC) (donated by Dr Fernando Navarro-García, CINVESTAV, Mexico), Salmonella enterica ser. Typhimurium ATCC 14028 (donated by Becton Dickinson Co.), and E. faecalis ATCC 19433 (commercially acquired).

In Mexico, 61% of consumptive water use is based on rivers, streams, lakes, or ponds, while the remaining 39% is derived from underground aquifers. The agricultural sector consumes 76% of this water, whereas 63% of the water used by humans is from rivers and lakes. Notably, 87% of the country's surface water is obtained from 51 rivers and is stored in ponds and lakes (INEGI 2018).

Water samples were analyzed for the presence of Salmonella spp., Enterococcus spp., and E. coli pathotypes (EHEC, ETEC, EPEC, EAEC) according to the Food and Drug Administration Bacteriological Analytical Manual (BAM) Chapter 5, Salmonella method (FDA 2018), Rice & Baird (2017), and Corzo-Ariyama et al. (2019), respectively. An aliquot (0.5 mL) of each sample was inoculated in 5 mL of specific enrichment broth, universal pre-enrichment broth (UPEB, BD), and Rappaport–Vassiliadis broth (BD) for Salmonella, dextrose azide broth for Enterococcus spp., and tryptone phosphate broth (Bioxon, Mexico) for E. coli pathotypes. The UPEB and dextrose azide broth were incubated for 24 h at 37 °C, while the tryptone phosphate broth was incubated at 44 °C for 20 h. For Salmonella spp., a secondary enrichment broth was used; UPEB (50 μl) was inoculated into tubes with 5 mL of Rapapport–Vassiliadis broth, and the mixture was incubated at 42 °C for 20 h.

After incubation, an aliquot of each enrichment broth was streaked onto a specific agar: xylose-lysine deoxycholate agar (Bioxon) for Salmonella, KF Streptococcus agar (Acumedia, USA) supplemented with triphenyl tetrazolium chloride (1%; Sigma) for Enterococcus spp., and MacConkey (Bioxon), MacConkey sorbitol (BD), and eosin methylene blue (Bioxon) agars for E. coli pathotypes. Representative and presumptive colonies were confirmed by polymerase chain reaction.

DNA extraction and polymerase chain reaction assay for Salmonella and Enterococcus was performed according to Corzo-Ariyama et al. (2019) with minor modifications. The final reaction volume was 25 μl, consisting of 0.2 U Taq (Bioline, USA), 5 μl of 5× MyTaq reaction buffer (Bioline), 0.24 μM of each oligonucleotide (invA for Salmonella and efaA, ccf, eda, and gelE for Enterococcus), and 2 μl of template DNA.

In the case of E. coli pathotypes (EHEC, EPEC, ETEC, EAEC), the method described by Corzo-Ariyama et al. (2019) was utilized. The mixture (final reaction volume: 25 μl) consisted of 1 U Taq and 5 μl of 5× MyTaq reaction buffer, 0.1 μM of oligonucleotides stx1, stx2, eae, and bfp, or 0.5 μM for aafll and lt, and 3 μl of template DNA. All amplicons (10 μl) were separated using a 1.5% agarose gel, stained with GelRed™ Nucleic Acid Gel Stain (1×), and visualized under an ultraviolet light photo-documentation device (Gel Logic 200 Imaging System; Kodak).

Antibiotic resistance or susceptibility was evaluated using previously confirmed isolates according to the Clinical and Laboratory Standards Institute (CLSI 2022). An aliquot (10 μl) of a culture previously grown to 0.5 McFarland standard (1.5 × 10⁸ CFU/mL) was inoculated onto Mueller–Hinton agar (Bioxon) spiked with various antibiotics: ampicillin (10 μg/mL), vancomycin (6 μg/mL), erythromycin (30 μg/mL), polymyxin B (30 μg/mL), colistin (4 μg/mL), ciprofloxacin (5 μg/mL), trimethoprim (25 μg/mL), and tetracycline (30 μg/mL). In addition, the Kirby–Bauer method was performed according to the CLSI (2022); 1 mL of culture adjusted to 1 × 10⁶ CFU/mL was inoculated by extension onto Mueller–Hinton agar without antibiotics. Antimicrobial Susceptibility Test Discs (ampicillin [25 μg/mL], erythromycin [15 μg/mL], cefotaxime [30 μg/mL], and sulfamethoxazole/trimethoprim [25 μg/mL]) were placed over the inoculated surface. Plates were incubated at 37 °C for 24 h.

Microbial resistance was determined by observing microbial growth, inhibition halos were measured in the Sensi-Disc assay, and the results were interpreted according to the CLSI guidelines as follows: ampicillin (susceptible ≥17 mm, intermediate 14–17 mm, resistant ≤13 mm); erythromycin (susceptible ≥23 mm, intermediate 14–22 mm, resistant ≤13 mm); cefotaxime (≥26 mm, intermediate 23–25 mm, resistant ≤22 mm); sulfamethoxazole/trimethoprim (susceptible ≥16 mm, intermediate 11–15 mm, resistant ≤10 mm).

For strains that showed resistance to colistin, confirmation of resistance was performed according to the CLSI guidelines. For this purpose, bacterial growth was determined in agar supplemented with 1, 2, and 4 μg/mL colistin.

Data were analyzed using the general linear model (GLM) procedure, and the differences between antibiotic resistance and susceptibleness were determined using Fisher's least significant difference (LSD) multiple-comparison test. p ≤ 0.05 was considered significant.

E. faecalis, ETEC, and EPEC were detected in the river water samples obtained from Veracruz. Two isolates of E. faecalis were eda1+ , ccf+ , and gel E− ; however, one isolate was efaA+ whereas another was efaA− ; in addition, two strains of EPEC, namely one typical (eae+ and bfp+) and one atypical (eae+ , bfp−), were detected. All these isolates were derived from a single sample. E. coli, Salmonella, or E. faecalis were not detected in the well and dam water samples obtained from Chihuahua. In contrast, isolates of pathogenic E. coli were found in three different pond samples, two atypical EPEC (eae+ and bfp− ) and one ETEC (lt+), collected from Puebla.

Enterococci are commonly found in human and animal feces and can persist in the environment. The sources of enterococci in waters include wastewater, agricultural and urban runoff, stormwater, direct input from animals through defecation, bathhouses, boats, plant debris, contaminated groundwater, soils, sediments, and sands. However, enterococcal infections in humans are predominantly caused by E. faecalis and Enterococcus faecium; these microorganisms are increasingly resistant to antibiotics worldwide. Genes encoding different virulence factors were identified in isolated enterococci. It has been reported that these factors increase the pathogenicity of bacteria, thus enabling tissue colonization, invasion, translocation, and evasion of the host immune response (Davis et al. 2022).

In general, it was observed that the samples collected from rivers and ponds harbor pathogenic bacteria, whereas there were no pathogens detected in well water samples. Megchún-García et al. (2015) detected E. coli in groundwater (27% of samples) and surface water (92%) in the sugarcane agroecosystem in Veracruz, México. These results indicated that contaminated irrigation water in that location poses a risk to human health. Gonzales-Siles & Sjöling (2016) analyzed the environments in which ETEC develops. They revealed that plants and aquatic environments are common reservoirs for this pathotype, partly due to their ability to adapt to pH 7.2–7.8 and their growth temperature. The investigators also showed that the detection rate of ETEC increased between April and May (Qadri et al. 2005). This is consistent with the results of our study since the samples were collected in April.

Corzo-Ariyama et al. (2019) showed that only four isolates (1.2%) of 341 E. coli strains isolated from the agricultural environment in north Mexico were positive for E. coli pathotypes, while three were positive for atypical EPEC and one was positive for ETEC. These data are in agreement with the findings of our study, as both typical and atypical EPEC were found in Veracruz water samples; both have been associated with diarrheal disease in developed countries (Haymaker et al. 2019). Detection was based on the presence of specific genes previously reported; for typical EPEC, the eae (intimin) and bfp (‘bundle forming pili’ encoded genes must be positive by the E. coli adherence factor plasmid), and in the atypical EPEC eae and bfp were positive and negative, respectively. Haymaker et al. (2019) reported a high incidence of atypical EPEC (9.0%) in untreated surface and reclaimed water in the mid-Atlantic USA. This also corresponds to our results, as only three of the 30 samples (10%) analyzed were positive for EPEC.

The seven confirmed isolates were analyzed for their susceptibility or resistance to antibiotics. All isolates were resistant to at least three of the 10 antibiotics tested (Table 1); most isolates (5/7, 71.5%) were resistant to vancomycin. All E. coli isolates were resistant to vancomycin (5/5, 100%), while four isolates (4/5, 80%) were resistant to ampicillin and erythromycin and susceptible to ciprofloxacin, trimethoprim, and sulfamethoxazole/trimethoprim. Resistance to polymyxin B, colistin, and cefotaxime was detected in E. faecalis isolates (2/2, 100%). All E. coli and E. faecalis isolates were sensitive to trimethoprim, ciprofloxacin, and sulfamethoxazole/trimethoprim (Table 1).

Odonkor & Addo (2018) investigated the prevalence of antibiotic resistance in E. coli strains isolated from water sources. They found that 32.99% of isolates showed resistance to penicillin, cefuroxime (28.87%), erythromycin (23.71%), tetracycline (21.45%), ampicillin (11.32%), and ciprofloxacin (8.25%). In our study, the observed rate of resistance to ampicillin in E. coli isolates was 20%. For erythromycin, ampicillin, and ciprofloxacin, the resistance rate was 80, 80, and 0%, respectively.

Regarding Enterococcus, isolates with resistance to antimicrobials have been reported in environmental water used as irrigation water (Carey et al. 2016), suggesting that contaminated irrigation water may lead to contamination of food crops (Ben Said et al. 2016). In our study, resistance to polymyxin B, cefotaxime, and colistin was observed in the two E. faecalis isolates. Colistin is considered a last resort for the treatment of bacterial infections due to its toxicity to humans.

Importantly, in the agricultural environment, pathogens can be disseminated through contaminated irrigation water, agricultural soil, raw or poorly composted manure, and feces from nearby domestic or wild animals. Identifying sources of contamination in the pre-harvest period is crucial, since reducing or eliminating contaminations after harvest could be challenging (Zoellner et al. 2016).

In this study, pathogens were not detected in the well water samples, and groundwater is generally considered microbiologically safe. Nevertheless, pathogens can enter water sources through overflowing sewage, polluted stormwater, and agricultural runoff. Wells located excessively close to floodplains, livestock areas, manure storage areas, cesspools, septic tanks, and trickle fields are more susceptible to contamination by pathogens.

Human activities and natural processes occurring during crop production must be monitored to reduce or eliminate this risk to public health. This study provides information that may drive a broader study of additional sites, taking into account the variability of the ecosystems and crops grown in Mexico.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.