Food Protection Trends - July/August 2023 - 334

or associated animal feces, whether that be through roof
runoff or dust kicked up by foot traffic. This is supported by
Lévesque et al. (41), who found that soil and dust carried
by wind can be washed into freshwater containers after
analyzing rainwater collection tanks in Bermuda, with 66%
of the 102 tanks sampled containing E. coli. Flood and
heavy rains have also been known to wash debris and fecal
pollutants into rain catchment systems and cisterns (18). It
is important to note that maintaining the water quality of a
cistern can be difficult, with numerous routes of microbial
contamination, and should be regularly monitored (18).
The well was the only groundwater source and had a covered
opening with a hand pump to collect water. The well was
next to a pond (pond E), and even though the well was not
exposed to the environment, large quantities of E. coli could
be found. Based on our observations, cross-contamination
from the pond could have been possible. Andrade et al. (4)
addresses three ways E. coli could contaminate well water:
(i) directly through well opening, (ii) water recharge/deep
drainage (geological pathways including shifts in soil), and
(iii) direct migration (contaminated groundwater mixing
with noncontaminated groundwater). Any of these factors
could explain the transfer of E. coli contaminates into the
well sampled during this study. Sasakova et al. (59) found
contaminated surface water could eventually lead to the
contamination of groundwater sources and that an aquifer
environment could facilitate pathogen survival. Many factors
can facilitate the microbial quality of groundwater sources, so
it is important to note that these sources are still susceptible
to contaminants.
Isolate characterization
In total, 99 different serotypes of E. coli could be identified
(Table 3). Reports of E. coli in water have exposed high
discrepancies of serotypes with both pathogenic and
nonpathogenic isolates. Maloo et al. (45) and Ramteke
and Tewari (54) also noted the remarkable diversity found
among the serotypes collected from both recreational water
and drinking water. The most prevalent serotype identified
in our study was E. coli O65:H38, with 12 identified
isolates, all of which were collected during spring. Of the
total E. coli isolates, only two were found to be a part of
the " Big 6, " including O26 (SAMN28816604) and O45
(SAMN23828792) (10). Each isolate was however sourced
from two different ponds during different seasons: O26 was
isolated during spring and O45 during summer.
In Table 3 and Figure 3, isolates are grouped based on the
same flagellar H-antigens. The genetic relations inferred from
phylogenetic relatedness (Fig. 3) can be used in source attribution
and might indicate the same contamination source.
Therefore, to quantify the relative importance of animal hosts
as pathogen reservoirs, we performed a literature review to
understand which serotypes could be historically linked to
a potential host and therefore contamination source. The
334 Food Protection Trends July/August
link is only theoretical, nevertheless it is useful to initially
understand the potential risks for these water sources. Papers
published earlier than 2010 were excluded from our search,
and only publications linked to water source illness and E. coli
contamination were considered. Literature supporting serotype-based
identification shows isolates deriving from more
than one potential host. Bovine sources were found to be the
most common source of likely contamination, encompassing
53% of the E. coli serotypes. Other isolates could likely be
traced back to human (38%), ovine (17%), poultry (11%),
porcine (8%), and deer (2%) sources.
Observations collected during the sampling periods
support the likely causes of contamination linked through
serotype identification. As mentioned, all water samples
were collected on farms with domesticated animals. Even
if ponds were protected by a fence, the presence of E. coli
in our samples indicates that the pathogen could be spread
regardless of the confinement. Osman et al. (52) found
domestic calves, sheep, and goats to harbor many of the same
E. coli serovars as confirmed in this research. This finding
supports the hypothesis that there could be a direct or
indirect animal-to-human or animal-to-animal transmission
of E. coli, creating overlap in serotype profiles (33). Half of
the isolates characterized in this study could be traced back
to bovine origination. E. coli is a natural part of the ruminant
microbiota; nevertheless, it has been suggested that even
wildlife could be an asymptomatic reservoir (22, 61). A
deer running through a cow field could carry E. coli from
that field on its hooves and eventually contaminate a water
source. E. coli can be carried on animals' fur, hooves, paws,
skin, feathers, and feces, thereby spreading E. coli to the areas
where they roam (16). E. coli contamination in agricultural
water as observed is difficult to track, with numerous direct
and indirect routes of transmission.
At least two antimicrobial resistance genes were expressed
in all the isolates collected (Table 3). More than 15 different
AMR genes were reported through in silico analysis, covering
many of the antibiotic classes. Each antibiotic class
was grouped based on chemical structure (77). The most
prevalent AMR gene was blaEC, expressing resistance to
β-lactam, spanning across 98% of the samples. The next most
prevalent Amr gene was mdtM (94%), a gene expressing a
multidrug-resistant protein including resistance to nucleoside,
phenicol, lincosamide, and fluoroquinolone antibiotics
through an efflux pump complex (3). Fifty-seven percent
of the isolates had the AMR gene acrF (57%), encoding an
efflux pump resistance mechanism to fluoroquinolone (3).
Finally, the other most frequent genes expressed resistance
to tetracyclines (tetA [14%] and tetB [10%]) and aminoglycosides
(aph(3′′)-Ib [15%] and aph(6)-Id [15%]). Coinciding
with the results of our study, Liao et al. (42) found
all collected E. coli isolates from an aquafarm to have at least
two drug resistance genes, indicating a multidrug-resistant
sampling pool. One study suggests bodies of water are ideal
Food Protection Trends - July/August 2023

Table of Contents for the Digital Edition of Food Protection Trends - July/August 2023
