Fecal Carriage and Genetic Characterization of CTX-M-1/9/1-Producing Escherichia coli From Healthy Humans in Hangzhou, China

CTX-M-199, a novel chimeric β-lactamase which mediated resistance to sulbactam and tazobactam, was recently identified in Hangzhou, China. This study investigated the prevalence of fecal carriage of bacteria producing CTX-M-199 and other CTX-M-1/9/1-type enzymes among healthy individuals and characterized the genetic features of blaCTX–M–1/9/1-bearing mobile elements. A total of 74 Enterobacterales strains carrying various blaCTX–M–1/9/1 genes, including blaCTX–M–64 (n = 40, carriage rate of 0.74%), blaCTX–M–199 (n = 23, 0.40%), blaCTX–M–123 (n = 5, 0.10%), novel blaCTX–M–153 (n = 5, 0.10%), and blaCTX–M–132 (n = 2, 0.04%), were isolated from 68 out of 5,000 (1.36%) fecal samples of healthy adults in Hangzhou City. Phylogenetic analysis based on whole-genome sequencing data showed that 72 blaCTX–M–1/9/1-bearing Escherichia coli isolates were clustered into four major clades, three of which included CTX-M-199 producers. Sixty out of 75 blaCTX–M–1/9/1 genes were located on plasmids belonging to four Inc types: IncI2, IncI1, IncFIB, and IncHI2. The blaCTX–M–199 genes were harbored by three of the four types of plasmids except for IncHI2. All these blaCTX–M–1/9/1 genes were carried on an ISEcp1-mediated transposition unit. In conclusion, human fecal carriage of blaCTX–M–1/9/1 was low in healthy populations of China. The ISEcp1 was commonly associated with blaCTX–M–1/9/1 and may mediate its transmission on various mobile elements. Our findings provide insights into the dissemination and the development of further measures for the control of pathogens producing CTX-M-1/9/1-type enzymes.


INTRODUCTION
β-Lactam antibiotics are currently the most widely used antibiotics in the treatment of infectious diseases (Bush and Bradford, 2016). The emergence and spread of extended-spectrum β-lactamases (ESBLs), which hydrolyze expanded-spectrum cephalosporins and monobactams, severely limit therapeutic options and pose a serious public health threat (Pitout and Laupland, 2008). β-Lactamase inhibitors, including clavulanate, sulbactam, tazobactam, and the new avibactam, with high affinity for β-lactamases, can greatly enhance the activity of their partner β-lactams against ESBL producers. Therefore, β-lactam/β-lactamase inhibitor (BLBLI) combinations can be considered for the treatment of non-severe infections (Gutierrez-Gutierrez and Rodriguez-Bano, 2019). Unfortunately, several TEM and SHV variants have been reported to confer resistance to BLBLIs since the 1990s (Drawz and Bonomo, 2010). These inhibitor-resistant variants have been found more frequently in Europe than in the USA and have never been reported in China (Canton et al., 2008;Manageiro et al., 2012). It is worrying that the inhibitor-resistant TEM enzymes were still prevalent in Escherichia coli isolates from both outpatients and hospitalized patients of a Spanish hospital (Rios et al., 2015). The first natural inhibitor-resistant CTX-M-type β-lactamase, CTX-M-190, has been found in E. coli from Shanghai, China, in 2017 . Soon after, we identified a chimeric CTX-M-64 variant carrying S130T substitution, CTX-M-199, which mediated resistance to sulbactam and tazobactam, in three E. coli strains from Hangzhou, China (Cai et al., 2017). Mass spectrometry and crystallography analysis revealed that the binding of the sulbactam to the active site rendered the formation of the inhibitor enzyme complex inefficient and resulted in inhibitor resistance in CTX-M-199 (Cheng et al., 2020). CTX-M-64, the first described chimeric CTX-M (a hybrid of CTX-M-14-like and CTX-M-15-like β-lactamases), was identified in a Shigella sonnei strain recovered from a Japanese tourist who had returned from China in 2009 (Nagano et al., 2009). Subsequently, several chimeric CTX-Ms, including CTX-M-64, CTX-M-123, CTX-M-132, and CTX-M-37, have been detected in E. coli from pets, food animals, and humans in China (Sun et al., 2010;He et al., 2013;Tian et al., 2014;Gu et al., 2015;Liu et al., 2015). CTX-M-64 or/and CTX-M-123 producers have also been reported in a domestic pet dog and wastewater in Japan (Kawamura et al., 2017;Tanaka et al., 2019), fecal samples from healthy children in Laos (Stoesser et al., 2015), and a patient in Germany (Pfeifer et al., 2018). Our previous study indicated that bla CTX−M−199 gene was located on the mcr-1-bearing IncI2 plasmid and harbored in E. coli isolated from stool samples (Cai et al., 2017). However, the prevalence of CTX-M-199 and other CTX-M-1/9/1 enzymes in fecal E. coli isolates from healthy populations and the range of plasmids carrying this gene are unclear. The study aims to determine the fecal carriage rates of CTX-M-1/9/1-producing Enterobacterales in the intestinal tract of healthy individuals and to depict potential mechanisms underlying the dissemination of the bla CTX−M−1/9/1 gene by investigating the genetic characterization of CTX-M-1/9/1 producers, bla CTX−M−1/9/1 -bearing plasmids, and the genetic context of bla CTX−M−1/9/1 gene.

Isolation and Species Identification of Enterobacterales
A convenient sample of 5,000 stools collected from asymptomatic healthy adults (3,218 male and 1,782 female) receiving annual physical examinations in a hospital of Hangzhou, China, from July to September 2019 were investigated. In order to eliminate the effect of antibiotics or intestinal diseases on intestinal flora, participants showing any symptoms related to gastroenteritis or had been exposed to antimicrobial agents or to a hospital environment in the 3 months prior to sample collection were excluded from the study. Within 4 h of collection, 1 g of stool sample was inoculated into 5 ml of Luria-Bertani (Oxoid Ltd., Basingstoke, England) broth without shaking at 37 • C for 24 h. Twenty microliters of the mixture was streaked onto a plate of China Blue agar (Hangzhou Binhe Microorganism Reagent Co., Ltd., Hangzhou, China) supplemented with 10 mg/L of cefotaxime and incubated at 37 • C for 24 h for the selection of potential CTX-M producers. To avoid the omission of isolates carrying bla CTX−M−1/9/1 gene, suspected Enterobacterales with different colonial morphology were selected from each sample and sub-cultured on fresh selective medium. Thus, more than one Enterobacterales isolates were obtained from the same sample in some cases. Species identification of bla CTX−M−1/9/1positive isolates was performed by MALDI-TOF MS (Bruker Daltonik GmbH, Bremen, Germany). This study was approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine, and consent was given by the participants.

Whole-Genome Sequencing and Genome Analysis
Genomic DNA extracted from bla CTX−M−1/9/1 -positive isolates was sequenced using the MGISEQ-2000 platform (BGI Complete Genomics), and the reads were de novo assembled by SPAdes (v3.11.1) (Bankevich et al., 2012). The assembly scaffolds were subjected to a screening of bla CTX−M−1/9/1 genes. The contigs containing the target gene were searched in the NCBI database by the BLASTN program, and the putative gaps were filled by PCRs based on the available reference sequences in an attempt to obtain the complete plasmid sequences.
The Transferability of bla CTX−M−1/9/1 Genes The transferability of bla CTX−M−1/9/1 genes was also assessed by conjugation experiments which were carried out using filter mating method. Rifampin-resistant E. coli EC600 was used as the recipient strain. Transconjugants were selected on MacConkey agar plates containing 500 mg/L rifampin and 30 mg/L cefoperazone/sulbactam (2:1) or 20 mg/L aztreonam as appropriate.

Experimental Evolution Assays
To assess the possibility of emergence of S130T substitution in CTX-M-64, E. coli transconjugants harboring bla CTX−M−64carrying plasmids with different Inc types (including pM-64-1028, pM-64-826, pM-64-799, and pM-64-3814) were subjected to parallel experimental evolution assays in the presence of increasing concentrations of cefoperazone/sulbactam as previously described (Ripoll et al., 2011).

Nucleotide Sequence Accession Numbers
Only one of the bla CTX−M−1/9/1 -containing plasmids or chromosomal fragments sharing identical sequence or the same genetic structure surrounding bla CTX−M−1/9/1 has been deposited in GenBank. The accession numbers of 19 representative sequences are listed in Supplementary Table S3.

Antimicrobial Susceptibility Results
All 74 Enterobacterales were resistant to cefotaxime but were susceptible to tigecycline. Remarkably, resistance against carbapenems and colistin was observed in six (one CTX-M-199 producer, three CTX-M-64, and two CTX-M-153) and 22 (12 CTX-M-199, eight CTX-M-64, and two CTX-M-132) E. coli isolates, respectively. Regardless of their status of carbapenemase production, all bla CTX−M −199 -positive E. coli isolates were resistant or intermediate to piperacillin/tazobactam and cefoperazone/sulbactam but were susceptible to carbapenems, cefotaxime/clavulanate, ceftazidime/clavulanate, ceftazidime/avibactam, and cefmetazole. These E. coli isolates were also susceptible or intermediate to aztreonam and ceftazidime except five isolates co-producing CTX-M-55. On the contrary, isolates carrying other bla CTX−M −1/9/1 genes were susceptible to piperacillin/tazobactam and cefoperazone/sulbactam but were resistant to aztreonam FIGURE 2 | Continued and ceftazidime. All isolates except for carbapenemase producers were susceptible to the compound of clavulanate or avibactam combined with their partner cephalosporins.
Genomic and Phylogenetic Analysis of bla CTX−M−1/9/1 -Positive Enterobacterales Seventy-four bla CTX−M −1/9/1 -positive isolates and two E. coli transconjugants of E. coli 1028 which carried bla CTX−M −199 and bla CTX−M −64 genes, respectively, were subjected to whole-genome sequencing (WGS). The genome sizes of the tested isolates ranged from approximately 4.6 to 5.9 Mbp. The MLST analysis showed that 72 CTX-M-1/9/1producing E. coli isolates belonged to a diverse range of sequence types, including 39 known STs and nine novel STs. Among 23 CTX-M-199 producers, a total of 21 STs were identified. ST48 was the most common type among these (n = 8), followed by ST457 (n = 4), ST746 (n = 4), and ST2973 (n = 3) (Supplementary Table S1). The phylogenetic analysis based on SNPs showed that 72 E. coli isolates were clustered into four major clades. The CTX-M-199 producers were grouped into three clades, and CTX-M-64 and CTX-M-123 producers can be found in each clade. Isolates belonging to the same ST were generally divided into one cluster (Figure 1).

Genetic Characterization of bla CTX−M−1/9/1 -Bearing Plasmids
Seventy-five DNA fragments harboring bla CTX−M −1/9/1 genes were obtained, among which 60 plasmids belonging to four Inc types, including 22 IncI2, 14 IncI1, 11 IncFIB, and 13 IncHI2, were identified in 58 isolates (strain 1028 carried two IncFIB plasmids encoding CTX-M-199 and CTX-M-64, respectively). The bla CTX−M−199 genes in this study were harbored by the first three types of plasmids. Fifteen bla CTX−M −64 -containing DNA fragments, which ranged from 6 to 277 kb, were found to closely match known chromosomes, but not plasmid sequences, in the NCBI database by BLASTN, with coverage of >98% and identity of >99% (Supplementary Table S1).

FIGURE 3 | Continued
The sequence alignments showed that plasmids belonging to the same Inc type shared a high degree of sequence homology (Figure 2). IncI2 plasmid, in which the bla CTX−M−199 gene was firstly identified, was dominant among bla CTX−M−1/9/1bearing plasmids, accounting for 12 bla CTX−M−199 , eight bla CTX−M−64 , and two bla CTX−M−132 . Almost all these IncI2 plasmids carried the mcr-1 gene simultaneously except three bla CTX−M−64 -bearing plasmids from strains 400, 1966, and 3354 (Supplementary Table S1). The IncI1-type replicon can be detected in four out of 23 bla CTX−M−199 -carrying plasmids and one out of 40 bla CTX−M−64 -carrying plasmids. Remarkably, it can also be detected in all four bla CTX−M−123 -carrying plasmids and all five bla CTX−M−153 -carrying plasmids. These plasmids were closely related to a bla CTX−M−123 -bearing IncI1 plasmid pHNAH4-1 (GenBank accession no. NC_024955.2) which was obtained from an E. coli isolated from the stool sample of a chicken in Anhui province of China (Liu et al., 2015). Seven bla CTX−M−199 -bearing and four bla CTX−M−64 -bearing plasmids with ca. 110 kb in size belonged to IncFIB type (Supplementary Table S1 and Figure 2C). Interestingly, the sequences of pM-199-1028 and pM-64-1028, both harbored in the same strain 1028, were almost identical, differing only by two nucleotide substitutions that distinguish bla CTX−M−199 from bla CTX−M−64 gene ( Figure 2C). Similar plasmids can also be found in E. coli isolated from anal swab and feed samples from a breeding goose farm in China, such as pL100-3 (GenBank accession no. NZ_CP034747.1) and pL41-1-3 (GenBank accession no. NZ_CP034729.1), two IncFIB plasmids carrying bla CTX−M−199 , which were identified in Jiangsu province recently (Liu et al., 2019). The IncHI2-type replicon was only detected in bla CTX−M−64 -bearing plasmids in this study (Supplementary Figure S2).
Genetic Context of bla CTX−M−1/9/1 Genes Based on a comparison of the genetic environment surrounding bla CTX−M −1/9/1 genes, transposition units flanked by different direct repeats were determined in all samples (Figure 3). ISEcp1 was present in all transposition units; however, their sizes varied. Apart from the 3,080-and 5,800-bp ISEcp1 transposition units located on IncI2 and IncI1 plasmids, respectively, a 2,968 bp FIGURE 3 | Continued transposition unit was identified in the plasmids of IncFIB and IncHI2 and chromosome. For the IncHI2 plasmid and chromosome, more than one type of transposition units can be found. These results suggested that bla CTX−M −1/9/1 genes can be carried by an ISEcp1 mobile genetic element in various sizes of transposition units and disseminated among different types of plasmids and chromosomes.

DISCUSSION
The intestinal tract is an important reservoir and source for transmission of many antibiotic resistance genes, including bla genes, and intestinal colonization is associated with infections caused by ESBL-producing organisms (Donskey, 2004;Carlet, 2012;Karanika et al., 2016). This study revealed a low carriage rate of bla CTX−M−199 , which was located on three types of plasmids-IncI2, IncFIB, and IncI1-in healthy populations of a Chinese city. However, the bla CTX−M−199 gene can be located on IncHI2-type plasmid by searching the NCBI database. One example was the pCTXM199_025970 (GenBank accession no. CP036178.1) (Figure 2D), which was harbored in an E. coli isolated from Sichuan province. A large number of CTX-M-199 producers recovered from clinical specimens have been found in different hospitals, including two sporadic E. coli isolates from Zhejiang and Sichuan provinces (Li et al., 2018) (GenBank accession no. CP036178.1) and several E. coli, K. aerogenes, and K. quasivariicola isolates from multiple samples of the same patient in another hospital of Zhejiang province (GenBank accession nos. VLPD01000017.1, VLOS01000013.1, FIGURE 3 | Alignment of the genetic environment surrounding bla CTX−M−1/9/1 genes located on different Inc types of plasmids (A,B) and chromosomes (C). The genes of different function or different origins were labeled with different colors. The position and sizes of the transposition units were indicated. Only the representative sequences of plasmids and chromosomes were compared. The sequences of the remaining isolates shared the same or similar genetic structure of the bla CTX−M−1 , 9 , 1 genes with that of the representative sequences and were listed in Supplementary Table S3. VLOY01000014.1, and VLPG01000031.1). CTX-M-199 has also been found in E. coli from goose, Gallus, and chicken in three different provinces (Jiangsu, Jilin, and Zhejiang) (Liu et al., 2019) (GenBank accession nos. CP034747.1, CP034729.1, RRZS01000003.1, and RDCW01000100.1) and K. pneumoniae from hospital sewage in Sichuan province (GenBank accession nos. NQVM01000026 and NQVL01000026). The majority of these isolates harbored the IncI2 plasmid, rare harbored the IncFIB and IncHI2 plasmids, and none harbored the IncI1 plasmid. Among the 22 IncI2-type plasmids identified in this study, 19 carried the mcr-1 gene (except three plasmids from strains 400, 1966, and 3354). Twenty-two E. coli isolates (including 12 CTX-M-199 producers, eight CTX-M-64 producers, and two CTX-M-132 producers) were positive for mcr-1 gene, 19 of which were located on IncI2 plasmids (except three mcr-1 from strains 197, 1079, and 3814). The coexistence of two resistance genes-mcr-1 and bla CTX−M −1/9/1 -in the same IncI2 plasmid may increase the risk of co-selection of this plasmid type and may contribute to the predominance of IncI2 plasmid among isolates of various origins.
The phylogenetic analysis showed that no significant association was found between a specific clone and the production of CTX-M-1/9/1. However, two remarkable dominant STs of E. coli, ST48, which was frequently identified in NDM and ESBL producers (Smet et al., 2010;Liu et al., 2019), and ST746, which was the most prevalent type of CTX-M-123 producers from chicken farms of China (Liu et al., 2015), were observed. ST48 and its derivative can be detected in multiple CTX-M-1/9/1 producers (five CTX-M-64, three CTX-M-199, one CTX-M-123, and one CTX-M-153), two of which coproduced NDM. Unlike isolates from chicken, three of four ST746 E. coli produced CTX-M-64 (Supplementary Table S1). The presence of each subtype of bla CTX−M−1/9/1 gene in E. coli from different clades in the phylogenetic tree based on SNPs (Figure 1) indicated that E. coli of diverse genetic backgrounds contributed to the dissemination of bla CTX−M−1/9/1 genes, although few clonal spreads seemed to have occurred.
The bla CTX−M−199 gene may be generated via two possible routes. One possibility was that CTX-M-199 was derived directly from CTX-M-64 by two substitutions of A 109 T and S 130 T. The E. coli 1028 in this study appeared to support this hypothesis. In this isolate, we found the coexistence of two 112,177-bp plasmids (pM-199-1028 and pM-64-1028 carrying bla CTX−M−199 and bla CTX−M−64 genes, respectively), which differed by only two bases. However, no S 130 T mutation was found in CTX-M variants which were resistant to βlactamase inhibitors by in vitro selection (Ripoll et al., 2011). Consistent with this, the serial passage of E. coli strains, each carrying pM-64-1028 and three other bla CTX−M−64 -carrying plasmids with different Inc types at growing concentrations of cefoperazone/sulbactam, also failed to generate S 130 T mutation (data not shown). Thus, the bla CTX−M−199 gene was more likely to be formed by the recombination between bla CTX−M−1−like and bla CTX−M−14−like carrying natural S 130 T mutation. The generated inhibitor-resistant bla CTX−M−1/9/1 hybrid gene further disseminated among different types of plasmids via ISEcp1-mediated transposition. As one of the most important elements associated with the mobilization and dissemination of bla CTX−M genes, transposition units mediated by ISEcp1 varied (Zhao and Hu, 2013). Two ISEcp1-mediated transposition units, with 3,080 and 5,800 bp, respectively, identified in this study can also be found in the previous study (Liu et al., 2015). Our current data were unable to depict the evolution of bla CTX−M−1/9/1 hybrid gene. However, from the analysis of the genetic backgrounds of these types of genes carrying different gene cassettes and plasmids, we confirmed that ISEcp1 was the most common element that mediated the transmission of these types of genes. Allele change between different ISEcp1-mediated bla CTX−Ms mobile elements might facilitate the formation of hybrid bla CTX−M genes.
One limitation of this study is that the risk factors for CTX-M-199 carriage are not assessed, as information on occupation, diet, behavioral characteristics, or travel from the participants were not available.

CONCLUSION
This study reported the intestinal colonization of healthy populations of China by CTX-M-1/9/1-producing E. coli with a fecal carriage rate of 1.36%. The bla CTX−M−1/9/1 gene, carried by the mobile element ISEcp1, was located on four major types of plasmids and disseminated among E. coli with diverse genetic backgrounds. Our findings provide insights into the dissemination and further development of appropriate control measures for this inhibitor-resistant β -lactamase.

DATA AVAILABILITY STATEMENT
The data presented in the study are deposited in GenBank and accession numbers can be found in the Supplementary Material.

ETHICS STATEMENT
This study was approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine, and written informed consent was given by participants.

AUTHOR CONTRIBUTIONS
JCa and SC conceived and designed the work. JCh and YJ provided samples and collected enterococcal isolates. JCh performed the whole-genome sequencing and sequence assembly. JCa, JCh, and SC analyzed the data and interpreted the results. JCh and JCa drafted the manuscript. SC and RZ improved the English. All authors revised the manuscript and approved the final version.