Genetic characterization of commensal Escherichia coli isolated from laboratory rodents

Isolation of E. coli

A total of 67 randomly selected healthy rodents (mice; n = 38 and rats; n = 29) aged
between 6 and 12 weeks old, housed in the Animal Experimental Unit, Faculty of Medicine,
University of Malaya during March 2014–April 2015, were assessed as part of the unit’s
rodents health monitoring program (Table 1). The sex of the rodents were distributed quite proportionately between males (n = 32)
and females (n = 35) (Table 1). Healthy mice were euthanized and serologically assayed for selected pathogens according
to previously published protocols (Loong et al. 2016]). The rats however, were anaesthetized with Ketamine (50 mg/kg) and Xylazine (5 mg/kg)
prior to blood collection through cardiac puncture. The collected rat blood was tested
by serology (SMART-SPOT Sentinel Panel Test, Rat Sera, Biotech Trading Partners, USA)
for the following agents; Kilham rat virus, Mycoplasma pulmonis, pneumonia virus of mice, rat coronavirus, respiratory enteric orphan virus, rat
parvovirus, Sendai virus and Theiler’s murine encephalomyelitis virus. Following that,
the rats were euthanized by overdosing with the same drugs at three times the anaesthetic
dosage. Rodents were housed in open cages and other detailed procedures were described
in a previous study (Loong et al. 2016]). Selected rodent organs (gastrointestinal tract, liver and, the trachea and lung)
were harvested, homogenized and cultured onto MacConkey agar. Organs from every rodent
were cultured onto individual agar plate. Cultured bacterial strains suspected to
be E. coli were identified using Gram staining, microscopy and 16S rDNA sequencing (Misbah et
al. 2005]). One strain was selected from each suspected culture plates after 24 h incubation
at 37 °C for this study.

Table 1. Breed of rodents assessed between March 2014 and April 2015

Multi locus sequence typing (MLST) and phylogroup classification

Commensal E. coli strains were subjected to MLST (referred to as the Pasteur scheme), where the internal
portions of eight housekeeping genes (dinB, icdA, pabB, polB, putP, trpA, trpB and uidA) were sequenced and compared with the database at http://bigsdb.web.pasteur.fr/ecoli/ecoli.html (Jaureguy et al. 2008]). A specific allele number was assigned to each distinct sequence within a locus
and a specific sequence type (ST) was assigned to each distinct combination of alleles.
A novel allele number was assigned to sequences that could not be found in the database,
subsequently establishing a novel ST. For phylogroup classification, E. coli strains were designated into the phylogroups; A, B1, B2, C, D, E, or F, according
to a previous study (Clermont et al. 2013]). This quadruplex PCR method differentiates the E. coli phylogroups by nucleic acid amplification of the arpA, chuA and yjaA genes, and the DNA fragment, TspE4.C2.

Identification of Shiga toxin-producing E. coli serotypes

E. coli serogroups O5, O15, O26, O45, O55, O76, O91, O103, O104, O111, O113, O118, O121,
O123, O128, O145, O146, O157, O165, O172 and O177 are the O-antigen types of the most
clinically significant Shiga toxin-producing E. coli serotypes. These 21 serotypes were identified using three serotype-specific multiplex
PCR assays, and it can also detect the respective O-antigens even when they are not
expressed by the E. coli strains (Sánchez et al. 2015]).

Extended spectrum beta lactamases (ESBL) phenotypic confirmatory method

All commensal E. coli strains were cultured onto Mueller–Hinton agar after adjustment to 0.5 McFarland
standards. Antimicrobial discs containing cefotaxime (30 ?g), cefotaxime/clavulanic
acid (30/10 ?g), ceftazidime (30 ?g) and ceftazidime/clavulanic acid (30/10 ?g) were
placed onto the agar plate so that the distances between them are approximately twice
the radius of the inhibition zones, and the plate was incubated overnight at 35 °C.
ESBL activity was indicated when a ?5 mm increase in the inhibition zone diameter
around either cephalosporin disc in combination with clavulanic acid as compared to
the inhibition zone diameter of the cephalosporin when tested alone (CLSI 2012]). The E. coli strain ATCC 25922 and Klebsiella pneumoniae strain ATCC 700603 were used as controls as per CLSI recommendation.

Detection of genes encoding for ESBL

Five multiplex and one singleplex PCRs were employed in this study to detect genes
encoding important beta-lactamases; one CTX-M multiplex PCR including groups 1, 2,
8, 9 and 25 (Woodford et al. 2006]); one TEM/SHV/OXA-1-like multiplex PCR (Dallenne et al. 2010]); one plasmid-mediated AmpC gene multiplex PCR including ACC, FOX, MOX, DHA, CIT
and EBC (Dallenne et al. 2010]); one VEB/GES/PER multiplex PCR (Dallenne et al. 2010]); one VIM/IMP/KPC multiplex PCR (Dallenne et al. 2010]); and one OXA-48-like singleplex PCR (Dallenne et al. 2010]). In addition, the New Delhi metallo-beta-lactamase, NDM-1 was detected using PCR
as previously described (Peirano et al. 2011]).

Detection of E. coli from the animal housing environment, rodent feed and drinking water

Testing for the presence of E. coli in the animal housing environment was performed by taking swabs from the tabletop,
tap, trolley and wash basin in the room housing the rodents. Sterile, irradiated rodent
feed (#1320-Maintenance diet for rats and mice, Altromin International, Germany) was
vortexed in tubes containing nuclease-free water. Following that, the swabs, the vortexed
feed and the autoclaved drinking water were cultured onto MacConkey agar and incubated
at 37 °C under aerobic condition for 24 h. One colony resembling the morphological
appearance of E. coli was selected from each positive plate and the E. coli phylogroup was determined. Additionally, detection of genes encoding for ESBL was
also performed.