blaNDM and mcr-1 to mcr-5 Gene Distribution Characteristics in Gut Specimens from Different Regions of China

Antibiotic resistance has become a global public health concern. To determine the distribution characteristics of mcr and blaNDM in China, gene screening was conducted directly from gut specimens sourced from livestock and poultry, poultry environments, human diarrhea patients, and wild animals from 10 regions, between 2010–2020. The positive rate was 5.09% (356/6991) for mcr and 0.41% (29/6991) for blaNDM, as detected in gut specimens from seven regions, throughout 2010 to 2019, but not detected in 2020. The detection rate of mcr showed significant differences among various sources: livestock and poultry (14.81%) > diarrhea patients (1.43%) > wild animals (0.36%). The detection rate of blaNDM was also higher in livestock and poultry (0.88%) than in diarrhea patients (0.17%), and this was undetected in wildlife. This is consistent with the relatively high detection rate of multiple mcr genotypes in livestock and poultry. All instances of coexistence of the mcr-1 and blaNDM genes, as well as coexistence of mcr genotypes within single specimens, and most new mcr subtypes came from livestock, and poultry environments. Our study indicates that the emergence of mcr and blaNDM genes in China is closely related to the selective pressure of carbapenem and polymyxin. The gene-based strategy is proposed to identify more resistance genes of concern, possibly providing guidance for the prevention and control of antimicrobial resistance dissemination.


Introduction
Antibiotic resistance has become a major global public health concern in the 21st century. Carbapenems and polymyxins are among the last-resort antibiotics for defending against Gram-negative bacterial infections [1]. Among the various mechanisms, the bla NDM (New Delhi metallo-β-lactamase) gene and the mcr (mobile colistin resistance) gene, conferring resistance to carbapenem and polymyxin respectively, exhibit cross-species and cross-region transmission [2]. The bla NDM-1 gene was first discovered in patients in 2009 [3], and 29 genotypes have since been identified [4][5][6][7][8][9]. Carbapenems are mainly used to treat human respiratory infections, and resistant bacterial strains often exhibit multidrug and broad-spectrum drug resistance [10]. Although carbapenem usage has not been approved for use in the breeding industry in China, new bla NDM subtypes and an epidemic of resistant bacterial strains have appeared in livestock and poultry [11,12]. The mcr-1 gene was first discovered in pigs in 2016 [13] and 10 genotypes and multiple subtypes have since been found [14][15][16][17][18]. Polymyxins were once widely used as feed additives and for disease prevention in livestock and poultry in China, but have been banned as animal growth 14 * The positive rate shows significant differences between different sources (p < 0.05). a , b , c : each subscript letter denotes a subset of source categories whose column proportions do not differ significantly from each other.
(257/1179) and 2.02% (13/644) for the mcr gene. In wild animals, the mcr gene was detected in various species ranging from marmots and rats to bats ( Figure 1).

Sequence Analysis of blaNDM and mcr
Among the blaNDM-positive gut specimens, no new mutations were found. Among these gut specimens, 89.66% (26/29)  All 20 mcr-2 positive gut specimens in this study were new subtypes. Compared to MCR-2.1 (Accession No.: WP_065419574.1), there were numerous sense and nonsense mutations ( Figure 3). Compared with MCR-2.1, the aa identity of MCR-2.4 was 97.0%, and of the 3% mutations, 81.81% were nonsense and 18.18% were sense; the aa identity of MCR-2.5 was 98.5%, and of the 1.5% mutations, 84.91% were nonsense and 15.09% were sense; the aa identity of MCR-2.6 was 98.7%, and of the 1.3% mutations, 83.39% were nonsense and 10.61% were sense; the aa identity of MCR-2.7 was 98.5%, and of the 1.5% mutations, 85.96% were nonsense and 14.04% were sense.

Distribution of Coexisting Genes/Genotypes and New Subtypes
The gut specimens with coexisting genes/genotypes or new subtypes were mostly from livestock and poultry and poultry environment, with only the new subtype mcr-3.31 being derived from wild animals ( Table 2). Ten gut specimens harbored both bla NDM and mcr-1, and 18 gut specimens harbored two mcr genotypes. The coexistence of bla NDM and mcr-1 within a single gut specimen was only observed in Anhui. The coexistence of mcr genotypes a within single gut specimen was mostly found in Guangxi. The new mcr subtypes were from Guangxi, Anhui and Yunnan.

Discussion
Antibiotic resistance may be a survival strategy for bacteria, with antibiotics triggering specific bacterial responses [30,31]. This study shows that antibiotic selective pressure might be reflected by resistance gene pools of various sources. In combination with the findings of our previous study, this shows that the emergence of polymyxin and carbapenem resistance strains in China is closely related to the selective pressure of antibiotics. The mcr or bla NDM strains originating from livestock and poultry, patients, and wildlife, are mainly nonpathogenic organisms [20], which is consistent with findings from studied conducted in 47 countries across six continents with mcr-positive strains [32], which showed a tendency to be increased under antibiotic selection pressures. Due to the limited sensitivity of isolation, some normal flora with resistance may not be detected. To further determine the distribution of mcr and bla NDM from different sources-carried by the bacteria selected by antibiotic selective pressures and normal flora with resistance-this study was conducted based on gut specimen detection strategy and a One Health approach. Overall, the positive rate of the mcr gene was much higher than that of the bla NDM gene for each specimen source. This is in accordance with positive-strain isolation [20]. The positive rates of the mcr gene showed significant differences among sources: livestock and poultry (14.81%) > diarrhea patients (1.43%) > wild animals (0.36%) ( Table 1), consistent with the relative isolation rates of polymyxin-resistant strains among these sources [20]. Though polymyxinresistant strains had not been isolated in wildlife, the mcr gene was detected. Livestock and poultry (0.88%) were found to contain the bla NDM gene more frequently than diarrhea patients (0.17%), but this gene was not detected in wildlife (Table 1). Carbapenem-resistant strains were also not isolated from wildlife in a previous study [20]. Compared with other sources, no polymyxin-or carbapenem-resistant strains [20], lower rates of mcr and bla NDM genes (Table 1) and less mcr genotypes (Table 2) were found in wildlife samples, which supports the hypothesis that wild animals are a net sink rather than a source of clinically relevant drug resistance [33]. The phenotypic diversity of drug resistant strains in wildlife is also low [33]. Since wild animals have less chance of being exposed to antibiotics, the emergence of resistance genes possibly reflects the resistance genes carried by normal flora. Similarly, Salmonella enterica-isolated from diarrhea patients and asymptomatic individuals-showed equal carriage of mcr carriers, suggesting the mcr gene is carried by normal flora [34]. In this study, the detection rates of mcr and bla NDM in diarrhea patients were far lower than in livestock and poultry, and higher than in wild animals ( Table 2). This is in accordance with the relatively low use of polymyxin and carbapenem in this population.
The gene pools of mcr or bla NDM reflect resistance genes carried by normal flora when antibiotic pressure is low, and genes carried by the bacteria selected by antibiotic pressure. When the pressure is relatively high, such as in livestock, poultry and humans, the relative levels of the mcr and bla NDM genes-to a certain extent-possibly reflects the antibiotic selective pressure. In particular, in livestock and poultry, there higher rates of the bla NDM and mcr gene (Table 1) and more mcr genotypes were found (Table 2 and Figure 3). Polymyxins are often used as therapeutic drugs and feed additives for animals, and they are used more frequently for farmed animals in China [29], where the highest number of mcr-positive strains was reported [32]. During the intensive feeding period, antibiotics are required for animal treatment and disease prevention, which involves large doses and long-term use [35]. We found that the positive rate of the mcr gene was much higher in intensively reared animals (21.80%, 257/1179, swine, chickens, etc.) than in non-intensively reared breeding animals (2.02%, 13/644, yak, goats, canine, etc.) (Figure 1). The ban of polymyxin use as an animal growth promoter in 2017 seems to have reduced CREC and MCRPEC [20,36]. However, the observation that the mcr detection rates peaked in 2019 in this study (10.60%, 167/1575) (Figure 2), is consistent with the notion that mcr-1 isolates successively recovered from 2017 to 2019, which indicates the possibility that polymyxin resistance still exists in livestock and poultry. Carbapenem drug-resistant strains have appeared and are prevalent in poultry and livestock. New genotypes of the bla NDM gene have been found in livestock and poultry-derived strains around the world [11,28]. Firstly, carbapenems might be applied when treating animal diseases. Secondly, their use in humans pollutes the environment and results in indirect exposure of animals to the drug. Last but not least, bacteria with the bla NDM gene may exist in normal gut flora [37]. In summary, the emergence of drug resistance genes is due to the selective pressure caused by the overuse of antibiotics. The strategy of gene detection can be used for resistance gene profiles and supervision.
In this study, livestock and poultry were not only the main source of the mcr and bla NDM gene pool (Table 1), but they were also sources of mcr-1 and bla NDM co-harbored genes. Additionally, livestock and poultry were the source of multiple mcr genotypes within single gut specimens ( Table 2). Similar findings were not shown in diarrhea patients or wild animals. In general, the coexistence of the mcr-1 and bla NDM genes was only found in Anhui, and the coexistence of mcr genotypes mostly came from Guangxi, indicating that livestock and poultry in some regions may be exposed to higher or more complex antibiotic selective pressures. Considering that no strain carrying both the mcr and bla NDM genes had been isolated in the previous study [20], a past and present coexistence of the mcr-1 and bla NDM genes within one gut specimen is more likely to come from different clones (e.g., one clone harboring mcr-1, other clone harboring bla NDM ). It is also possible that a single clone carried both genes. In either case, the drug resistance conferred by mcr and bla NDM genes may be transmitted from livestock and poultry to humans, possibly even resulting in the emergence of polymyxin and carbapenem resistant strains. Recently, mcr-1 and bla NDM coexistence was also reported in the United States, Venezuela, and Japan [38][39][40], which reduces treatment options for multidrug-resistant bacterial infections and increases the incidence and mortality of the infections, leading to stricter antibiotic controls. It is necessary to strengthen antimicrobial resistance surveillance in livestock and poultry.
This study revealed the gene distribution of mcr and bla NDM in livestock and poultry, diarrhea patients and wild animals, demonstrating that relative level of the resistance genes may reflect the selective pressure of antibiotic exposure of various hosts, which is expected to become a strategy of antibiotic usage oversight. Potential antimicrobial usage of colistin, and others, plays a role in the enrichment of antimicrobial resistance genes in gut specimens, which are needed to further support culture-based data. Compared with the culture-based strategy, the gene-based strategy is more sensitive. The positive rates of gene detection among various gut specimens were about two to three times those of isolation rates [20]. On the other hand, bacterial culture and genetic background information is not available through the gene-based strategy. The fact that more positive specimens found by gene detection than culture detection, may come from the normal flora with resistance that cannot always be isolated, and gene positive results do not always equate to phenotype positive results [25]. Additionally, searching for new variants is limited by the current PCR method. Although this method is improving over time [34,41], it is based on known genotype data which often cannot be used to discover an unknown variant. The gene detection method could be developed into a strategy based on metagenomic sequencing [42], identifying more concerned drug resistance genes and genetic information coming from various sources, and providing guidance for the prevention and control of drug-resistant bacteria and for supervision of antibiotic usage.

Gut Specimen Sources
Nucleic acid samples were obtained from 6991 gut specimens from livestock and poultry (26.08%, 1823/6991) including swine, chickens, canine, yak, goats, etc., poultry environments (5.01%, 350/6991), including breeding or slaughter environment, human diarrhea patients (16.96%, 1186/6991), and wild animals (51.95%, 3632/6991), including bats, marmots, rats, etc. (Table 1). The gut specimens were obtained in 2010-2020 from 10 regions of China (Beijing, Anhui, Gansu, Yunnan, Guangxi, Guizhou, Ningxia, Inner Mongolia, Qinghai, and Zhejiang) ( Figure S1), and they were retrospectively screened for Antibiotics 2021, 10, 233 9 of 12 the target genes. Gut specimen types of this study included human feces, animal anal swab, feces, intestinal content/swab or oral-pharyngeal swab, and poultry environment specimens related to gut environment, including drinking water, cage swab, depilator swab, cleaning sewage, chopping board swab, and soil. Unified protocols for specimen collection, transportation, and process were applied by professionals from local CDC (Center for Disease Control and Prevention) facilities, Institutes for Endemic Disease Prevention and Control, and hospitals. Specimens were collected and transported in Cary-Blair Transport Medium, processed and nucleic acids extracted using a genomic extraction kit (TIANamp Bacteria DNA Kit, Beijing, China). The nucleic acid samples were frozen for storage.

bla NDM and mcr-1 to mcr-5 Screening of Gut Specimens and Sequence Analysis
The target genes bla NDM and mcr were screened for, sequenced, and aligned with reference sequences from the National Center for Biotechnology Information (NCBI) database. The screening primers (Table S2) for bla NDM and mcr (mcr-1 to mcr-5) were previously described [20,41]. The original amplification of mcr-1 to mcr-5 involved multiplex PCR, but single PCR was conducted in this study. The CDS (coding sequences) of gut specimens with mutations in the screening sequences were further amplified, cloned (Transgene, Beijing, China), and sequenced. The number of PCR cycles for gene screening is 25 to 30, for CDS amplification it is 30. The PCR was performed using a 20 µL volume containing 10 µL Premix Taq version 2.0 (Takara, Beijing, China), 8 µL ultrapure distilled water, 0.5 µL (10 µM) of each forward and reverse primer and 1 µL of DNA template. The amplified products were detected using gel electrophoresis and sequenced in both directions using an Applied Biosystems 3730xl DNA Analyzer (Tsingke Biological Technology, Beijing, China). Phylogenic tree was constructed based on CDS sequences of mcr gene including sequences of this study and reference sequences (mcr-1 to mcr-4) and sequence analysis of mcr and MCR were conducted (Figure 3).

Statistical Analysis
Pearson's chi-square test (theoretical frequency T ≥ 5) was used to compare positive rates among different sources. As one theoretical frequency is 1 < T < 5, the Fisher exact test was also applied when comparing rates among different sources. Bonferroni correction was used to compare the positive rates between two sources. p < 0.05 was considered statistically significant. The statistical analysis was conducted by SPSS Version 19.0.

Ethics Statement
The study was approved by the ethics committee of the National Institute for Communicable Disease Control and Prevention of the Chinese Center for Disease Control and Prevention (IACUC Issue No. 2020-008). Verbal consent was obtained from the included diarrhea patients.

Conclusions
This study is first to determine the distribution characteristics of bla NDM and mcr genes from various sources of China. The positive rate of the mcr gene was much higher than that of the bla NDM gene for all sources, from highest to lowest was: livestock and poultry, diarrhea patients, and wild animals. The mcr or bla NDM gene pool of certain source reflect the resistance gene carried by normal flora when antibiotic pressure is low, and genes carried by the bacteria selected by antibiotic pressure. Livestock and poultry were not only the main source of the mcr and bla NDM gene pool, but also the source of co-harbored mcr-1 and bla NDM genes. The antimicrobial resistance surveillance in livestock and poultry needs to be strengthened. In conclusion, the study demonstrated that the selective pressure of antibiotic exposure of various hosts maybe reflected by relative level of the resistance genes, which is expected to become a strategy of antibiotic usage oversight.