Milk samples collection
A total of 535 milk samples of mastitic cows were collected from
different commercial dairy farms located in three provinces of China from May 2018 to June 2019 Table 1. Milk samples were aseptically collected by farm veterinarians from each cow's affected quarter. All the samples were shipped in iceboxes to the laboratory within 24 hours for further microbial diagnosis.
E. coli isolates were obtained as outlined according to the previous study (Suojala et al. 2011). Briefly, the bovine mastitis milk specimens were plated on MacConkey agar the plates were raised 24 h at 37°C. Presumptive
E. coli colonies were further confirmed microscopically and verified by species-specific PCR as defined earlier (Shafiq et al. 2019).
Briefly, after the preliminary growth of E. coli on MacConkey agar plates enriched with cefotaxime (2mg/L) and colistin (2mg/L) to screen out the putative ESBL and colistin-resistant isolates. The ESBL-producing E. coli isolates were further confirmed by a double-disc synergy test (DDST) (CLSI 2016).
Antimicrobial susceptibility testing
The antimicrobials used for susceptibility testing were ampicillin (AMP), cefotaxime (CTX), cefoxitin (CFX), ceftiofur (CEF), chloramphenicol (CHL), ciprofloxacin (CIP), colistin (COL), gentamycin (GEN), kanamycin (KAN), meropenem (MEM), polymyxin-B (POL), tetracycline (TET), and sulfamethoxazole-trimethoprim (SXT) in this study. The antimicrobial resistance (AMR) level was measured by the minimum inhibitory concentrations (MIC). The results were interpreted according to the principles of the Clinical and Laboratory Standards Institute (CLSI) (CLSI 2016).
A single colony of fresh bacterial culture from MacConkey agar was picked and inoculated into 3 mL of sterile Luria–Bertani (LB) medium at 37°C for 24 h in an incubator. Total DNA was extracted by a conventional boiling method (Queipo-Ortuño et al. 2008). E. coli isolates were used to find out ESBL genes (blaCTX-M, blaTEM, and blaSHV) as well as COL-resistant genes mcr-1 to mcr-9. All ESBL-positive isolates were further screened for the detection of major blaCTX-M groups (i.e., blaCTX-M-1, blaCTX-M-2, blaCTX-M-8, blaCTX-M-9, and blaCTX-M-25), using a multiplex PCR approach with specific primers, as previously defined Table S1. PCR amplicons were sequenced in TSKINGNE Corporation (Nanjing, China). Specific blaCTX-M alleles were identified using the online database for Genomic Epidemiology (https://cge.cbs.dtu.dk/services/ResFinder/), which is using to detect acquired AMR and chromosomal mutations in sequenced isolates of bacteria.
Phylogenetic typing
The
E. coli isolates were phylogenetically classified into seven phylogroups: based on four genetic markers, namely
arpA,
yjaA,
chuA, and
TspE4.C2. The phylogenetic grouping of the
E. coli isolates was carried out by using a quadruplex PCR reaction
(Clermont et al. 2013)
. All the primer sequences and PCR conditions are given in Table S1.
Detection of virulence genes
Virulence genes for this study were chosen based on those previously reported in E. coli isolates from clinical bovine mastitis. PCR tested the virulence linked genes: ompC, fimH, Ecs3703, fyuA, eaeA, ler, iucD, ompF, colV, and irp2 as previously defined (Zhang et al. 2018). All the primer sequences and PCR conditions for the virulence encoding genes are listed in Table S2.
Multi-locus sequence typing (MLST)
Evolutionary affiliation of the analyzed
E. coli isolates was determined through MLST for which all seven housekeeping genes (
adk, fumC, gyrB, icd, mdh, purA, and
recA) were amplified by PCR listed Table S3 (Wirth et al. 2006), and sequenced following the guidelines of the MLST databases (http://mlst.warwick.ac.uk/mlst/dbs/Ecoli/). Sequences yielded by the PCR amplification of each locus were utilized to obtain allelic numbers using an online database
https://pubmlst.org/bigsdb?=pubmlst_escherichia_seqdef) followed by sequence type (ST) mapping according to
E. coli MLST scheme (
https://enterobase.warwick.ac.uk/species/index/ecoli). A minimum spanning tree was built with PHYLOViZ 2.0 software (Instituto de Microbiologia, Portugal) according to the associations among MLST alleles (Ribeiro-Goncalves et al. 2016) and to visualize the genetic relatedness and diversity of different isolates. As proposed earlier, the clonal complexes of STs were determined through the goeBURST algorithm (Feil et al. 2004). Similarly, the association among the STs made in this study with the available STs in the MLST database (
https://pubmlst.org/bigsdb) was also evaluated through goeBURST (Francisco et al. 2009). Sequence alignment of all the examined strains was performed using MUSCLE implemented in MEGA 7 and the maximum probability tree constructed using General Time Reversible model, Gamma distributed and invariant sites (G+1) with 1000 bootable replicas using MEGA version 7.0 (Tamura et al. 2013).
Xbal pulsed-field gel electrophoresis (Xbal-PFGE)
To know the genetic relatedness, DNA fingerprinting profiles of the 22 mcr-1-positive E. coli strains were determined by Xbal-PFGE typing, as previously described (PulseNet 2017). Salmonella enterica serotype Braenderup H9812 standard was used as a size marker to adjust each fragment of XbaI-digested genomic DNA of donor strains, especially those strains from different gels. Fragment patterns were analyzed using BioNumerics software 7.6 (Applied-Maths, Kortrijk, Belgium), and a dendrogram was generated based on the Dice coefficient with 1.0% of tolerance in the band position and 1.5% optimization. A similarity level of 80% was considered the criterion for identifying genetically related strains (Tenover et al. 1995).