Mobile colistin resistance gene mcr-5 in porcine Aeromonas hydrophila.

Objectives: To characterize the mobile colistin resistance gene mcr-5 in Aeromonas hydrophila from backyard pigs in rural areas of China.
Methods: Pig faecal samples from 194 households were directly tested for the presence of mcr-5 by PCR assay and the phenotypic antimicrobial susceptibility profiles of the mcr-5-positive isolates were determined using the broth dilution method. The genomic location and transferability of mcr-5 were analysed by S1-PFGE with Southern blotting and DNA hybridization, and natural transformation, respectively. One strain isolated from an mcr-5-positive sample was subjected to WGS and the stability of the mcr-5-harbouring plasmid over successive generations was examined by subculturing.
Results: One mcr-5-positive A. hydrophila isolate showing resistance, with a colistin MIC of 4 mg/L, was isolated from a backyard pig faecal sample. mcr-5 was located on a 7915 bp plasmid designated pI064-2, which could naturally transform into a colistin-susceptible A. hydrophila strain of porcine origin and mediated colistin resistance in both the original isolate and its transformants. The plasmid backbone (3790 bp) of pI064-2 showed 81% nucleotide sequence identity to the corresponding region of the ColE2-type plasmid pAsa1 from Aeromonas salmonicida, while similar replication primases are widely distributed among aeromonads, Enterobacteriaceae and Pseudomonas species. Conclusions: To the best of our knowledge, this is the first identification of the novel colistin resistance gene mcr-5 in an A. hydrophila isolate from the faeces of a backyard pig. mcr-5 is expected to be able to disseminate among different bacterial species and genera.

Introduction

Colistin is one of the ‘last-resort’ drugs for treatment of serious clinical infections caused by MDR Gram-negative bacteria, especially carbapenem-resistant Enterobacteriaceae. However, increasing use of this antibiotic in clinical and veterinary practice has led to the emergence of mobile colistin resistance genes, including mcr-1, which was first reported in Escherichia coli in 2015, and mcr-2, identified in E. coli in 2016., Recently, three further plasmid-mediated colistin resistance genes, mcr-3, mcr-4 and mcr-5, were identified, with Enterobacteriaceae, particularly E. coli and Salmonella spp., being the predominant hosts. However, MCR-5 is distinct from MCR-1, MCR-2, MCR-3 and MCR-4, showing only 34%–36% amino acid sequence identity to the other four proteins. Despite the sequence discrepancy, all five proteins are phosphoethanolamine transferases, adding phosphoethanolamine to the lipid A moiety of LPS, leading to a more cationic LPS structure and consequently resistance to colistin. To date, only mcr-1 and mcr-3 have been widely reported among Enterobacteriaceae, while the mcr-5 gene has only been found in Salmonella spp. from poultry and animal-derived food products in Germany.

Bacteria belonging to the genus Aeromonas, especially Aeromonas hydrophila, Aeromonas caviae and Aeromonas veronii, frequently cause diarrhoeal diseases and wound infections in both humans and animals, and it appears to be ubiquitous in the aquatic environment. The mcr-3 variants have been identified in 10 Aeromonas species of various origins across four continents,, with mcr-3.3 and mcr-3.7 identified in A. veronii and A. caviae from chicken meat and a domestic duck, respectively, in China., Thus, Aeromonas species may play an important role in the acquisition and dissemination of the mcr genes. Here, we report the identification of an mcr-5-positive A. hydrophila isolate from the faeces of backyard pigs from rural areas of Shandong Province, China.

Materials and methods

Sample collection, bacterial identification and susceptibility testing

A total of 336 faecal samples were collected from backyard pigs from 194 households across 12 villages in rural areas of Shandong Province, China, in August 2017, using the ESwab Collection and Transport System (Copan, Brescia, Italy). The village and household selection methods have been described previously. All samples were enriched in 1 mL of LB broth containing 10 mg/L vancomycin at 37°C overnight. Total DNA was extracted from 500 μL of each of the enriched cultures using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) and the DNA used for screening of mcr-5 by PCR, as previously reported. The enriched samples positive for mcr-5 were inoculated onto Salmonella Shigella agar (Binhe Microorganism Reagent Co., Hangzhou, China) containing 2 mg/L colistin. One random colony of each morphotype was selected for mcr-5 gene detection. Species identification of mcr-5-positive isolates was conducted by 16S rRNA gene sequencing and MALDI-TOF MS analysis (Bruker Daltonik GmbH, Bremen, Germany).mcr-5-positive isolates were also subjected to antimicrobial susceptibility testing for colistin and eight other commonly used antibiotics, using the broth dilution method and interpreted according to CLSI M100-S25, VET01-S2 and EUCAST (version 6.0, http://www.eucast.org/clinical_breakpoints/). E. coli strain ATCC 25922 served as the quality control strain.

Location and transferability of mcr-5

S1-PFGE followed by Southern blotting and DNA hybridization with a digoxigenin-labelled mcr-5 probe were performed to determine the genomic location of mcr-5, as previously described. The transferability of mcr-5 was then examined using natural transformation, with plasmids extracted from an mcr-5-positive strain used as donor DNA and a colistin-susceptible porcine A. hydrophila strain (2ZF0081a) used as the recipient, as previously reported. Competent cells of the control strain 2ZF0081a were induced in 20% nutrient broth (NB) in late stationary phase after 24 h of incubation at 30°C. A 40 μL aliquot of competent cells was added to 100 μL of transformation buffer consisting of 53.5 mM Tris, pH 7.9, 20 mM MgSO4 and 33 mM NaCl. Plasmids extracted from an mcr-5-positive strain using a Plasmid Midi Kit (Omega, Norcross, GA, USA) were added in a volume of 10 μL (final concentration of 3.3 ng/μL). The mixture was incubated at 30°C for 120 min, then diluted in 0.85% NaCl and inoculated onto LB agar with 2 mg/L colistin at 30°C for 48 h. Transformants were confirmed by mcr-5 PCR assays.

WGS analysis of the mcr-5-carrying isolate

DNA was extracted from the mcr-5-positive isolate using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA). DNA libraries were prepared using a KAPA Hyper Prep Kit (Kapa Biosystems, Wilmington, MA, USA) and 150 bp paired-end sequencing was conducted using an Illumina HiSeq 2500 platform (Annoroad Genomics Co., Beijing, China). The draft genomes were assembled using SPAdes with ×50 minimum assembled coverage. Identification of antibiotic resistance genes and plasmid replicon typing were carried out via the Center for Genomic Epidemiology website (http://www.genomicepidemiology.org/). The mcr-5-carrying contigs were annotated using the online automated PATRIC server (version 3.5.1, https://www.patricbrc.org).

Stability of the mcr-5-carrying plasmid

The stability of the mcr-5-carrying plasmid was confirmed by subculturing successive generations in triplicate without colistin, as previously reported. Briefly, three colonies were enriched in 5 mL of LB broth containing 2 mg/L colistin for 24 h at 30°C. Colistin was removed by centrifuging 1 mL of culture (4000 g, 5 min) and resuspending the pellet in 1 mL of saline. A total of 4.88 μL of the suspension was then inoculated into 5 mL of fresh LB broth minus antibiotic (timepoint zero) and incubated at 30°C with shaking at 200 rpm for 24 h to obtain approximately 10 generations. Thereafter, 4.88 μL of each culture was transferred to 5 mL of fresh LB broth every 24 h. Cultures were diluted and plated onto LB plates every second day. A total of 50 resulting colonies were randomly selected for mcr-5-specific PCR assays to determine the proportion of mcr-5-positive bacteria in each population.

Results and discussion

PCR assays showed that eight pig faecal samples from seven households in four villages were positive for mcr-5, while all the samples were negative for the other three novel mcr-2, mcr-3 and mcr-4 genes. However, we were only able to detect one mcr-5-positive isolate (I064-2), identified as A. hydrophila, from one sample, while E. coli, A. veronii and A. caviae strains isolated from the other seven samples were negative for mcr genes. The considerably higher rate of positive results obtained from direct sample testing compared with bacterial isolation is indicative of a substantial segment of unseen resistome, which is in line with the previous observation of the ‘phantom resistome’ of blaNDM and mcr-1 resistance genes in samples from commercial chicken farms. The identified A. hydrophila isolate, I064-2, was resistant to colistin (MIC = 4 mg/L) and intermediate to tetracycline (MIC = 8 mg/L), but was susceptible to amikacin (MIC = 4 mg/L), gentamicin (MIC = 2 mg/L), florfenicol (MIC = 1 mg/L), ceftazidime (MIC = 0.5 mg/L), ciprofloxacin (MIC = 0.5 mg/L), tigecycline (MIC = 0.5 mg/L) and meropenem (MIC = 0.03 mg/L). S1-PFGE and DNA hybridization analyses revealed that the mcr-5 gene was located on an ∼8 kb plasmid (data not shown).

The draft genome of A. hydrophila I064-2, assembled from WGS reads, contained 135 contigs (N50=112971 bp). Among these, contig 56 (8064 bp, ×476.625 assembled coverage) harboured a 3670 bp segment consisting of chrB-mcr-5-Δmsf-Δmsf that showed 100% nucleotide sequence identity to the mcr-5-carrying operon of plasmid pSE13-SA01718 (GenBank accession number KY807921.1) from a Salmonella Paratyphi B isolate from Germany (Figure 1). We also observed 149 bp repeats at each end of contig 56. Reverse PCR using primers pI0642-F (5′-CGGCTCGTATTATGGCTGTCG-3′) and pI0642-R (5′-CGCTCGGGTGCGAAATCA-3′) (Figure 1) yielded a 1940 bp amplicon that successfully closed the contig, fully covering the 149-repeated base, confirming that mcr-5 was located on a 7915 bp plasmid, designated pI064-2 (GenBank accession number MG800820). The 100% nucleotide sequence identity between the mcr-5-carrying fragments of pSE13-SA01718 from Salmonella Paratyphi B and pI064-2 implied high stability of the mcr-5-carrying operon, as well as the dissemination of mcr-5 between different bacterial species and genera.

Figure 1.

Comparison of the genetic environment of mcr-5 and the plasmid backbone of pI064-2 with pSE13-SA01718 (GenBank accession no. KY807921.1) and pAsa1 (GenBank accession no. AY301063.1). Arrows indicate the positions and directions of the genes. Δ indicates a truncated gene. Regions with 100% homology are indicated by dark grey shading and those with 81% homology are indicated by light grey shading. F and R represent the primers pI0642-F and pI0642-R used in the reverse PCR assay. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

In addition to the mcr-5-carrying operon, pI064-2 contains a 3790 bp segment consisting of genes involved in plasmid replication (rep) and mobilization (mobC and mobA) and coding for proteins of unknown function. This region showed 81% nucleotide sequence identity to the corresponding region of the ColE2-type plasmid pAsa1 (GenBank accession number AY301063.1) from a piscine Aeromonas salmonicida isolate (Figure 1). The 1967 bp segment containing mobC-mobA showed 91% nucleotide sequence identity and 82% query coverage to the corresponding region of pAsa1, while the replicase-encoding gene rep showed 76% nucleotide sequence identity and 95% coverage to the repA region of pAsa1. BLASTp analysis of the pI064-2-encoded replicase (311 amino acids) revealed that a large section of this protein shows high amino acid identity to the replicase of a Leisingera sp. isolate obtained from an Euprymna scolopes sample from the USA (GenBank accession number WP_082020692.1; 91% identity and 75% query coverage) and to the deduced amino acid sequence of repA from pAsa1 (75% identity and 99% query coverage) (Figure 2). In addition, the A. hydrophila pI064-2 replicase showed 65%–76% amino acid sequence identity to the replication primases from Serratia marcescens, Klebsiella pneumoniae, Pseudomonas putida, A. caviae, A. salmonicida and A. hydrophila isolates derived from humans, chicken, fish, sewage and sand samples collected from eight countries in Asia, Europe and North America [Figure 2 and Table S1 (available as Supplementary data at JAC Online)]. Similarities between ColE-like replication primases in Enterobacteriaceae, Pseudomonas species and aeromonads suggests the likelihood of potential wide dissemination of mcr-5 among these bacteria, especially Aeromonas species containing closely related plasmids. BLAST analysis revealed the presence of other resistance genes in A. hydrophila I064-2, including the AmpC β-lactamase gene blaMOX-6 and the tetracycline resistance gene tet(E).

Figure 2.

Phylogenetic tree generated from the deduced amino acid sequences of 18 replication primases from different bacterial species and from pI064-2. The tree was generated using CLC Genomics Workbench 9 (CLC Bio-Qiagen, Aarhus, Denmark).

Conjugation experiments did not result in the transfer of pI064-2 into a recipient E. coli J53AzR strain using filter mating, but purified plasmid DNA obtained from I064-2 was successfully transformed into a colistin-susceptible A. hydrophila strain in transformation buffer. The mcr-5-positive transformant had an MIC of 2 mg/L for colistin, 8-fold higher than the control recipient (MIC = 0.25 mg/L). Stability testing showed that all I064-2 colonies were positive for mcr-5 for ∼200 generations (20 days) in the absence of colistin selection, confirming that pI064-2 is stable in the WT parent. However, pI064-2 was only maintained in the mcr-5-positive transformant for 20 generations without the pressure of colistin. As previously described, natural transformation is a general property of Aeromonas environmental isolates. Therefore, mcr-5 might have the potential to be transferred among the aquatic Aeromonas species in the environment.

In conclusion, to the best of our knowledge, this is the first report of mcr-5 in A. hydrophila isolated from backyard pigs in rural areas of China. The ColE-like replication primase in mcr-5-harbouring plasmid, which is widely distributed among Enterobacteriaceae, Pseudomonas species and aeromonads, indicates that pI064-2 has the potential to disseminate among different bacterial genera. In addition, the observed 100% nucleotide sequence identity between the mcr-5-carrying operons in Salmonella and Aeromonas spp. also suggests the possibility of horizontal gene transfer of the mcr-5-carrying segment among these bacterial species and genera. Further studies should focus on the surveillance of mcr-5 in colistin-resistant Gram-negative pathogens derived from animals, humans and the environment.

Click here for additional data file.