Rapid detection and characterization of tet(X4)-positive Escherichia coli strains with nanopore sequencing

Sir,

Novel antimicrobial resistance genes harboured by pathogens pose a great public health concern, as exemplified by the emergence of bla_NDM-1 and mcr-1 over the past decade.^1,2 Recently, two novel plasmid-mediated tigecycline resistance genes, tet(X3) and tet(X4), were identified in Enterobacteriaceae and Acinetobacter in two studies,^3,4 which has inevitably caused worldwide concern.⁵ Tigecycline was introduced into clinical use in 2005 and regarded as the last-resort drug for severe infections caused by pathogens resistant to multiple antimicrobials including carbapenems and colistin.⁶ Traditionally, tigecycline resistance has been mediated by chromosomal mutations, overexpression of efflux pumps or a plasmid-mediated tet(A) mutant,^6,7 which were confined to specific bacterial clones via vertical transmission. Thus, the detection and genetic characterization of plasmid-encoded tet(X) genes among pathogens warrant serious attention.

Rapid characterization of MDR pathogens harbouring tet(X) genes from the perspective of genomics is crucial for informing antimicrobial choices for infections, monitoring their prevalence and investigating the transmission routes of the resistance genes. Nanopore sequencing, with its features of handheld, USB-sized, long-read, real-time sequencing and easy accessibility, has been used to characterize bacterial antibiotic resistance islands and resistance plasmids in an efficient way.⁸ Following the recent studies reporting plasmid-mediated tet(X) genes,^3–5 we utilized nanopore sequencing to characterize three tet(X4)-positive Escherichia coli strains of pork origin and highlighted the feasibility of nanopore sequencing in investigating and tracking emerging novel resistance genes.

In June 2019, 16 pork samples were collected randomly from 16 markets in Yangzhou, China in order to detect tigecycline-resistant bacterial strains. Briefly, the purchased pork was rinsed with LB broth (10 mL) containing tigecycline (2 mg/L), incubated at 37°C overnight and subcultured on MacConkey agar plates with tigecycline (2 mg/L) to screen for tigecycline-resistant isolates, which were further subjected to PCR, targeting tet(X) genes, and species identification by 16S rRNA gene sequencing with previously reported primers.³ Three tet(X4)-positive E. coli strains were subjected to antimicrobial susceptibility testing (AST) against different antimicrobials (Table S1, available as Supplementary data at JAC Online). Conjugation assays were conducted to investigate the transferability of tet(X4) and S1-PFGE was utilized to characterize the plasmid profiles. The genomes were extracted with the TIANamp Genomic DNA Kit then libraries were constructed with the rapid barcoding kit RBK004 and subjected to nanopore sequencing in a portable MinION with a flow cell R9.4 (Figure 1a). As nanopore sequencing data alone is insufficient to obtain a reliable bacterial genome sequence, short-read Illumina sequencing data were also generated to obtain accurate sequences.⁸ Different bioinformatics tools including Unicycler and ResFinder were utilized to assemble and characterize the tet(X4)-bearing genomes.

Figure 1

(a) Nanopore sequencing and data analysis workflow of tigecycline-resistant strains used in this study. The tet(X4)-positive strains were isolated from pork samples and subjected to nanopore sequencing via the MinION platform in a rapid way, combining the Rapid Barcoding Kit and the fast basecalling mode, followed by bioinformatics analysis and in-depth analysis to resolve the genomic characterization of tet(X4)-bearing genomes. The workflow could be finished within a few days, depending on the sampling scale. Other techniques such as AST, conjugation, S1-PFGE and Illumina sequencing were utilized to evaluate the phenotypes and obtain accurate sequences without indels. (b) Alignment of tet(X4)-bearing core structures derived from three plasmids in this study and two plasmids reported recently. Pink arrows denote the tet(X4) gene and red ones the other resistance genes. Yellow arrows represent ISs and orange ones the other genes. ISCR2, the alias of ISVsa3 in pG3X16-2-3, was annotated as ISVsa3 here, as in the reported paper.³ The grey areas show the homologous regions between plasmids. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Three tet(X4)-positive isolates, YPE3, YPE10 and YPE12, were identified in 3 different samples out of 16 pork samples (18.75%). They were resistant to multiple antimicrobials including tigecycline (MIC 8–16 mg/L), doxycycline (MIC 32–64 mg/L) and tetracycline (MIC >128 mg/L) (Table S1). The complete genome sequences of these three E. coli strains were resolved (Table S2). Two different STs, ST761 and ST10, were confirmed and the two ST761 strains (YPE3 and YPE12) were phylogenetically related, with 84 SNPs detected in their core genomes by Roary tool analysis. In addition to the three reports about tet(X4),^3–5 we found that tet(X4) had existed in different E. coli clones in China. All three strains harboured multiple plasmids, ranging from 1 kb to 190 kb, some of which encoded multiple resistance genes and replicon genes (Table S2).

To investigate the genetic contexts of tet(X4), detailed comparison of the tet(X4)-bearing plasmids was performed (Figures S1 and S2). pYPE3-92k-tetX4 was 92 973 bp long and encoded 112 predicted ORFs including resistance genes mef(B), sul3, dfrA5, tet(X4) and floR. pYPE12-101k-tetX4, sharing a similar structure to pYPE3-92k-tetX4 at 85% coverage, was 101 987 bp in length and coded for 125 predicted ORFs including floR, tet(X4), dfrA5, sul3, mef(B), qnrS1, bla_TEM-1B and tet(A). Online BLASTn of pYPE3-92k-tetX4 against the nr database retrieved a similar plasmid pG3X16-2-3 (NZ_CP038140, 97% identity at 86% coverage), which was also positive for tet(X4).⁴ Linear alignment of these three plasmids indicated that they were MDR plasmids characterized by a conserved region and an MDR region harbouring various ISs and resistance genes (Figure S1). ISVsa3-abh-tet(X4)-ISVsa3-yheS-cat-zitR-ISVsa3 was found in pYPE12-101k-tetX4; however, duplication and reversion of this structure happened in pYPE3-92k-tetX4 with the help of ISVsa3 and IS26 (Figure 1b). Tandem repeats of ISVsa3/ISCR2-abh/catD-tet(X4)-ISVsa3/ISCR2 were observed in pG3X16-2-3, which may result from the tet(X4)-positive circular intermediates,⁴ but lacked the ISVsa3-yheS-cat-zitR-ISVsa3 structure (Figure 1b). Three different replicon genes (IncFIA, IncFIB and IncX1) were identified in these three plasmids, which indicates they were mosaic MDR plasmids and evolved in different ways to generate the promiscuous tet(X4)-bearing segments mediated by various ISs. In contrast, tet(X4) in YPE10 was located on a conjugative IncHI1 and IncFIA hybrid plasmid, pYPE10-190k-tetX4, being 190 128 bp in length and encoding 227 predicted ORFs. pYPE10-190k-tetX4 was found to be similar to plasmid pYSP8-1 (NZ_CP037911) from ST542 E. coli isolated from pig faeces,⁴ which harboured another 43 kb region encoding emr(B) and an IncX4 replicon surrounded by ISVsa3/ISCR2 (Figure S2). The segment IS1-abh-tet(X4)-ISVsa3-yheS-cat-zitR-ISVsa3 existed in both pYPE10-190k-tetX4 and pYSP8-1. We proposed that pYPE10-190k-tetX4 was the progenitor plasmid and ISVsa3-mediated recombination of pYPE10-190k-tetX4 and another IncX4 plasmid generated the hybrid plasmid pYSP8-1. Based on available sequence data, ISVsa3/ISCR2 was a conserved mobile element accelerating transfer of tet(X4) among different plasmids (Figure 1b).^3–5

In addition to the tet(X4)-bearing MDR plasmids, other MDR plasmids encoding various resistance genes were found in YPE10 and YPE12 (Table S2). The tigecycline resistance phenotype conferred by tet(X4) was successfully transferred into E. coli C600 (Rif^r) with YPE10 and YPE12 as donor strains but this failed with YPE3 after three repeats. Based on the S1-PFGE result, pYPE10-190k-tetX4 was transferred into C600, but the plasmid in CYPE12 (the transconjugant of YPE12) was ∼210 kb in size, larger than the plasmids in donor strain YPE12. Complete plasmid sequencing with nanopore long-read data implied that plasmid reorganization occurred during this process (Figure S3). Although the two tet(X4)-bearing plasmids in YPE3 and YPE12 were similar in structure (Figure S1), the transferability of them differed and the underlying mechanism warrants further study.

The emergence of mobile tet(X) genes conferring resistance to tigecycline among pathogens constitutes a serious public concern and genomic epidemiology surveillance of tigecycline-resistant bacteria from different sources including humans, food, animals and environments should be urgently implemented. Using portable MinION nanopore sequencing, the genomic characterization of three tet(X4)-positive strains was completed within a few days in a rapid fashion (Figure 1a). Together with traditional methods such as AST, conjugation and S1-PFGE, the resistance phenotypes and transferability of tet(X4)-bearing plasmids were evaluated.

This efficient investigation paradigm paves the way for investigating novel resistance genes harboured by pathogens causing outbreaks in real time, similar to the application of nanopore sequencing in Ebola molecular surveillance and rapid clinical diagnosis of pathogens.^9,10 This study exemplifies an efficient approach to performing genomic studies of MDR pathogens in laboratories with limited available resources, which could revolutionize the practice of antimicrobial resistance research. The prevalence of tet(X4)-bearing E. coli strains of different clones highlights that the tet(X4) gene has dispersed widely.

Acknowledgements

We thank Yan Li, Kai Peng and Kangkang Li in Professor Wang’s laboratory for sample collection.

Funding

This work was supported by National Natural Science Foundation of China (31872523 and 31872526), the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), and the Natural Science Foundation of Jiangsu Province (BK20180900).

Transparency declarations

None to declare.