Newly identified colistin resistance genes, mcr-4 and mcr-5, from upper and lower alimentary tract of pigs and poultry in China

Antimicrobial resistance against colistin has emerged worldwide threatening the efficacy of one of the last-resort antimicrobials used for the treatment of Enterobacteriaceae. To investigate the presence of the recently identified colistin resistance genes (mcr-4, mcr-5) in China, we established PCRs to detect mcr-4 and mcr-5 on 213 anal and 1,339 nasal swabs from apparently healthy pigs (n = 1,454) in nine provinces, and 1,696 cloacal and 1,647 oropharyngeal samples from poultry (n = 1,836) at live-bird markets in 24 provinces of China. The prevalence of the mcr-4 in swine swabs (41.4%; 642/1,552) was significantly higher than in swabs from poultry (11.5%; 384/3,343). The mcr-4 gene was found in geese (49.5%, 54/109), chickens (17.2%, 257/1,498), pigeons (17.2%, 17/99) and ducks (15.4%, 20/130). In a similar trend, the prevalence of the mcr-5 in swine swabs (33.1%; 514/1552) was significantly higher than in swabs from poultry (5.6%; 187/3,343). The mcr-5 was identified in geese (17.4%, 19/109), chickens (9.9%, 148/1,498), ducks (7.7%, 10/130) and pigeons (3%, 3/99). The mcr-4 prevalence in the nasal swabs from pigs (59.2%, 58/98) was significantly higher than that in anal swabs (29.6%, 29/98) (P<0.001). Similarly, the mcr-5 prevalence in the nasal swabs from pigs (61.2%, 60/98) was significantly higher than in anal swabs (44.9%, 44/98) (P = 0.02), and significantly higher in oropharyngeal swabs (7.2%, 109/1,507) than in the cloacal swabs (3.7%, 56/1,507) (P<0.001). This study further confirms the presence of the mcr-4 and mcr-5 in animals and indicates these genes are prevalent and widespread in food producing animals (pig and poultry) in China. Future studies are needed to characterize the bacteria carrying the mcr-4 and mcr-5 and their locations on plasmids and/or the bacterial chromosomes, and determine co-resistances in the mcr-4 and mcr-5 positive strains.


Swab samples from swine and poultry
This study was reviewed and approved by the Institutional Animal Care and Use Committee of Yangzhou University College of Veterinary Medicine (YZU-CVM IACUC 2013#87, YZU-CVM IACUC 2015#57).
Nasal (n = 1,339) and anal (n = 213) swabs collected from apparently healthy pigs (n = 1,454) from nine provinces of China in 2016 (Table 1, S1 Table) [27] were used in this study to investigate the prevalence of mcr-4 and mcr-5. In addition, oropharyngeal and cloacal samples were obtained from poultry (n = 1,836) at 38 live-bird markets in 24 provinces in China between 2014 and 2015 (Table 1) [28], and 1,647 oropharyngeal and 1,696 cloacal samples from 1,836 birds were used in this investigation. The swabs from pigs and poultry were collected into tubes containing 400μl DNA/RNA Stabilization Buffer (Roche Molecular Biochemicals, IN, USA), and frozen at -80˚C until DNA extraction. Swabs were centrifuged in the DNA/RNA Stabilization Buffer (3,000×g, 4˚C for 5 min), and DNAs were extracted from the supernatants using the High Pure PCR Template Preparation Kit (Roche Diagnostic, USA) as described before [22,28]. All samples from previous studies that had sufficient residual DNA extract were included in this investigation.
The specificity of the primers for the mcr-4 and mcr-5 PCRs was verified by BLASTN and DNA sequencing of the amplicons obtained using synthesized plasmids containing portions of the target mcr-4 and mcr-5 that were cloned into the SacI site (Takara Biothechnology, Dalian, China). The sensitivity of the mcr-4 PCRs and mcr-5 PCRs was determined by amplifying dilutions of the synthesized plasmids. The PCRs were quantified using the PicoGreen DNA fluorescence assay (Molecular Probes, Eugene, OR, USA) with standards prepared with the synthesized plasmids (10 4 , 10 3 , 10 2 , 10 1 , and 10 0 copies/reaction) (Genscript, Nanjing, China).
The PCRs were performed on a Roche LightCycler 480 II PCR instrument. The PCRs with short amplicons were SYBR based, and were used to determine the presence of mcr-4 and mcr-5 in swabs in this study. The positive samples determined by PCRs were further amplified by PCRs with long amplicons. The PCR products of both short and long amplicons were sequenced using forward and reverse primers (BGI, Shanghai, China).

Phylogenetic analysis
The mcr sequences obtained from this study and those from GenBank for the mcr-4 and mcr-5 were aligned used the MEGA 6.0 software to compare their similarities.

Statistical analysis
Multiple Pearson Chi-square test was used to compare differences between animal species as well as between oropharyngeal/anal and oral/nasal swabs with Bonferroni adjusted p-values. P value below 0.05 was considered significantly different.

Establishment of PCRs for mcr-4 and mcr-5
The SYBR-based real-time PCRs detected the positive control plasmids containing the target mcr-4 and mcr-5 sequences with a detection limit of one gene copy per reaction. The detection limit was 10 copies per reaction for the mcr-4 and mcr-5 PCRs with long amplicons. The specificity of the PCRs was verified by gel electrophoresis and DNA sequencing.

Prevalence of mcr-4
Overall, the prevalence of the

Comparison of the presence of mcr in swabs from the upper and lower alimentary tract
Swabs from both the upper (represented by nasal and oropharyngeal swabs) and lower alimentary tract (represented by anal and cloacal swabs) were available for each of 98 pigs and of 1,507 poultry we studied. Both of the swabs were positive for mcr-4 in 21 (21.4%, 21/98) of the pigs but only the anal swab was mcr-4 positive in eight pigs (8.2%, 8/98), and only the nasal swab was mcr-4 positive in 37 pigs (37.8%, 37/98) (Fig 1, S3 Table). The prevalence of mcr-4 in the nasal swabs from pigs (59.2%, 58/98) was significantly higher (p<0.001) than that in anal swabs (29.6%, 29/98) (Fig 1).
All nucleotide sequences were submitted to GenBank with accession numbers MG586909 to MG586912 for mcr-4 and MG586913 to MG586915 for mcr-5.

Discussion
The usefulness of colistin, the last-resort antibiotic used to treat multidrug resistant Gram-negative bacterial infections, is being compromised with the recent identification of the mobile colistin resistance gene, mcr-1 [3], and the subsequent finding of mcr-2, mcr-3, mcr-4 and mcr-5 [4][5][23][24][25][26]. The mcr-1, mcr-2 and mcr-3 have been detected in bacteria or swabs from a variety of hosts in China and elsewhere in the world [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] However, there are yet only few reports on mcr-4 and mcr-5 [23][24][25][26]. Our study shows the mcr-4 and mcr-5 occur widely in pigs and poultry in China (Table 1, S2-S6 Tables). The prevalences of the genes we detected using PCR of swabs from animals were considerably higher than those obtained with studies that relied on bacterial isolates [23,24]. The sensitive and specific PCR we used to detect the mcr-4 and mcr-5 directly in swabs avoided the limitations introduced by bacterial isolation and the associated underestimation of the prevalence of the mcr's, the so-called 'phantom resistome' [29]. Although bacterial isolation for resistance testing is expensive, laborious, time consuming, and limits the resistant strains detected in a sample, it is an important adjunct to detection by molecular methods and enables a more complete understanding of colistin resistance and its epidemiology. Our molecular study might have provided more accurate data on the true prevalence of the mcr. However, the data did not enable us to determine the bacterial species that carried the resistance genes, or the location of the mcr in plasmids or in the bacterial chromosomes.
The mcr-1 gene has spread to most continents, and has been detected in various bacterial isolates from animals, human and the environment, including Escherichia coli, Klebsiella pneumoniae, Enterobacter cloacae and Enterobacter aerogenes [30]. The mcr-2 gene was found on rarely occasion, in Escherichia. coli isolates from porcine of Belgium and in flies of China [4,22]. After its first characterization on a IncHI2-type plasmid, pWJ1, from Escherichia coli isolated from a Chinese pig [5], the mcr-3 gene was shown to be present in bacteria isolated from humans in Denmark, chickens and flies in China, pigs in Japan and cattle in Spain [20-22, 26, 31].
As far as we know, the mcr-4 gene was first detected in two Salmonella enterica serovar Typhimurium strains isolated from human fecal samples and in Salmonella and E. coli isolated from pigs in Italy, Spain and Belgium [25]. After the initially discovery of the mcr-5 gene in Salmonella enterica subsp. enterica serovar Paratyphi B, the gene was detected in Escherichia coli from diseased pigs and healthy pigs in Japan [26].
Our findings of very high prevalences of the mcr-4 and mcr-5 in pigs and poultry from large areas in China are most likely associated with the prolonged and widespread use of colistin as a growth promoter in food animals in China. However, it should be noted that these two genes might also be more prevalent in other countries as few studies looked for them.
It is noticeable that the prevalences of the mcr-4 and mcr-5 were generally significantly higher in the nasal/oropharyngeal swabs than in the anal/cloacal swabs in both pigs and poultry. This suggests that bacteria in saliva and respiratory secretions might play important roles in the maintenance and transmission of colistin resistance genes in pigs and poultry. Further comparative studies are needed to determine the bacterial species carrying the mcr-4 and mcr-5 in the upper and lower alimentary tract and how there might be transmission of the resistance genes between these populations.
In conclusions, our study further confirms the presence of the mcr-4 and mcr-5 in bacteria from animals and indicates that these genes are widespread in food producing animals (pigs and poultry) in China. Future studies are needed to characterize the bacteria carrying the mcr-4 and mcr-5 and their locations on plasmids and/or the bacterial chromosomes.
Supporting information S1 Table. List of farms from where swine swabs were collected for mcr detection in this study.