Four carbapenem-resistant Enterobacter strains causing nosocomial infections were collected from seven tertiary health care hospitals in different provinces or cities in China between January 2017 and March 2021. Detailed bacterial strain collection was described in our previous study (Zhu et al., 2022). Clinical data were obtained from medical chart using the same standardized questionnaire in each hospital.

Genomic DNA of carbapenem resistant Enterobacter strains were prepared using the Omega Bio-Tek Bacterial DNA Kit (Doraville, GA, USA). The extracted DNA was detected by the agarose gel electrophoresis and quantified by Qubit (Thermo Scientific). Sequencing libraries were generated using NEBNext Ultra DNA Library Prep Kit (NEB, USA) following manufacturer’s recommendations and index codes were added to attribute sequences to each sample. At last, PCR products were purified (AMPure XP system) and libraries were analyzed for size distribution by Agilent2100 Bioanalyzer and quantified using real-time PCR. Whole-genome sequencing was performed using the Illumina NovaSeq 6000 platform (Illumina Inc., San Diego, CA, USA) with the 350 bp paired-end protocols. Reads were filtered using fastP 0.17.1, and evaluated through FastQC 0.9.1. The above methods can ensure the quantity and quality of DNA data. The qualified data were then de novo assembled to contigs using SPAdes 3.11. Further plasmid assembly was obtained by mapping contigs on reference plasmid sequence (> 99% identical), checking overlapping paired ends and gap closure by PCR and Sanger sequencing.

Precise species identification was performed by calculating the pairwise isDDH and ANI value between the genome sequence of the query strain and the type strains of Enterobacter spp. described recently (Zong et al., 2021). The pairwise ANI with a ≥ 96% cutoff and isDDH with a ≥ 70.0% cutoff were used for precise species identification as previously suggested (Meier-Kolthoff et al., 2013; Rosselló-Móra and Amann, 2015). The isDDH analysis was performed by calculating identities/HSP length value through GGDC 3.0 (Meier-Kolthoff et al., 2022), while ANI analysis was performed by using OAT 0.93.1 (Lee et al., 2016).

Open-reading frames (ORFs) and pseudogenes were predicted using RAST 2.0 (Brettin et al., 2015) combined with BLASTp/BLASTn searches. Multilocus sequence typing (MLST) as well as annotation of resistance genes, mobile elements and other features were carried out using online databases including PubMLST (Jolley et al., 2018), CARD (Alcock et al., 2020), ResFinder (Bortolaia et al., 2020), PlasmidFinder (Carattoli et al., 2014), ISfinder (Siguier et al., 2006), and INTEGRALL (Moura et al., 2009). Multiple and pairwise sequence comparisons were performed using BLASTn. Gene organization diagrams were drawn through scripts from Danmel (Wang et al., 2022), and displayed using Inkscape 1.0.¹

Zong et al. collected 4899 high-quality Enterobacter genome sequences within GenBank and carried out the precise species identification (Zong et al., 2021). Using the method described in this document, we further determined publicly available 15 E. chengduensis genome sequences within GenBank (excluding low-quality and duplicated sequences; accessed by January 19, 2022).

A phylogenetic analysis was performed on 19 identified E. chengduensis strains, including 15 strains from NCBI (Supplementary Table 1), and the four strains from this study. The genome sequences of E. chengduensis were aligned to the complete chromosome sequence (GenBank accession number CP004345) of E. chengduensis strain WCHECh050004, and the sequence of E. cloacae strain ATCC 13047 (GenBank accession number NC_014121) was used as an out group. The core single nucleotide polymorphisms (SNPs) were identified by Mummer 3.25 (Delcher et al., 2003). A maximum-likelihood phylogenetic tree was constructed using MEGAX 10.1.8 (Kumar et al., 2018) based on the core SNPs with a bootstrap iteration of 1000, and displayed using iTOL.²

We conducted conjugal transfer experiments between four E. chengduensis strains (HD5030, HD7411, HD7423 and HD7427) and the recipient E. coli J53 strain (NaN3-resistant) by combining mid-log-phase cultures of the donor and recipient (in a 1:10 donor to recipient ratio). Transconjugants were selected on plates with 2 μg/ml meropenem and 200 μg/ml NaN₃. The conjugants were then confirmed by PCR for bla_NDM as previously described (Nordmann et al., 2011).

Strains were grown overnight in 3 ml of LB with shaking (200 rpm) at 37°C, and then diluted to an OD600 of 0.25. 2 μl of the solution was added to 200 μl of LB in a 96-well plate in triplicate. Culture densities were determined every 10 min by measuring the OD600 for 16 h with shaking (200 rpm) at 37°C via FLUOstar Omega (BMG Labtech, Germany). Growth curves were estimated via GraphPad Prism 5.0 (GraphPad Software, Inc.) using two-way analysis of variance (ANOVA), and p-values < 0.05 were reported as statistically significant.

The draft genome sequences of four E. chengduensis strains HD5030, HD7411, HD7423 and HD7427 were submitted to GenBank under BioProject PRJNA824645. The sequences of four mapped plasmids were submitted to GenBank under accession numbers ON209151-ON209154.

We identified four non-duplicate E. chengduensis strains (Supplementary Table 2) from the 92 Enterobacter strains described in our previous study (Zhu et al., 2022). The four strains (HD5030, HD7411, HD7423 and HD7427) were collected from the emergency intensive care unit (ICU) or gastroenterology department in a tertiary care hospital in Southwest China (Supplementary Table 2). All strains exhibited resistance to cephalosporins (ceftazidime, cefepime, cephalothin, cefazolin, ceftriaxone etc.), carbapenems (meropenem, imipenem, ertapenem), piperacillin/tazobactam, and fluoroquinolones (ciprofloxacin, levofloxacin, moxifloxacin), while they were susceptible to amikacin, sulfamethoxazole. Antibiotic resistance gene screening through ResFinder revealed that all four E. chengduensis strains carried the bla_NDM–1 gene (Supplementary Table 3).

The results of MLST analysis showed that all 19 E. chengduensis strains analyzed in this study (including 15 ones from NCBI and four ones reported in this study) belonged to ST414 (Supplementary Tables 1, 2). A total of 2066 core genome SNPs were identified from these 19 strains and were used to construct a maximum-likelihood phylogenetic tree (Figure 1). The 19 strains were divided into three clades (clade I-III). A total of seven strains carried the carbapenemase-encoding gene, of which two strains from Colombia carried the bla_KPC–2 gene, while five strains from China carried the bla_OXA–48 gene (n = 1) and bla_NDM–1 gene (n = 4). All four bla_NDM–1-carrying strains belonged to clade III, while the other three belonged to clade II.

The core SNPs of 19 ST414 E. chengduensis strains were further pairwise compared to determine their clonal relatedness (Supplementary Table 4). In the nine strains from China, there were 5 to 592 SNP differences, suggesting that they did not belong to a single clone. However, the three strains in this study, HD7411, HD7423 and HD7427, only had only 5–10 SNP differences. Furthermore, these three strains were collected from different patients in two different wards in a hospital between April and June 2020. The hospitalization times of the three patients infected with these three strains overlapped, and the collection interval of these three strains was within one month (Supplementary Figure 1). These results suggest the potential intrahospital clonal transmission of ST414 E. chengduensis strains. Additionally, SNP comparison also showed that two bla_KPC–2-carrying E. chengduensis strains did not belong to a single clone (having 302 SNPs difference).

The bla_NDM–1 genes of four strains were all carried by IncC plasmids (Table 1). Each IncC plasmid could be divided into the backbone and two or three accessory modules, which resulted from exogenous DNA region insertion at different sites of the backbone (Figure 2). The backbones of the four plasmids were almost identical (Table 1) to the IncC reference plasmid p5_SCLZS62 (GenBank accession number CP082173), while these five plasmids were mainly different in accessory modules (Figures 3, 4). First, compared with p5_SCLZS62 and the other three plasmids, the plasmid pHD5030-NDM was interrupted by a 27.6-kb inserted region_orf2454 containing several inserted sequences (IS) (Figure 4A). Second, the four plasmids and reference plasmid p5_SCLZS62 all carried unit transposon Tn6358 variants (Figure 4B). Among them, p5_SCLZS62 harbored Tn6358b, whereas the four plasmids in this study harbored Tn6358c. Tn6358b and Tn6358c were inserted within the orf240 gene from the backbone of IncC plasmids and were flanked by 5-bp direct repeats (DRs). Moreover, Tn6358b and Tn6358c differed from Tn6358a (GenBank accession number JX141473) by exchanging integron In834 for truncated integron In7. However, truncated integron In7 from Tn6358b harbored three resistance loci: gene cassette array (GCA) aadB (VR1: variable region 1), ISCR1–truncated Tn125 (carrying bla_NDM–1) region (VR2), and ISCR1–qnrA1 unit (VR3), while truncated integron In7 from Tn6358c further lost a 4.3-kb region including the VR3, a truncated qacED1 gene and a sul1 gene. In addition, the four plasmids all harbored the 15.1-kb inserted region_uvrD (Figure 4C), which was inserted within uvrD from backbone of IncC plasmid and was flanked by 6-bp DRs. The inserted region_uvrD carried a predicted efflux RND transporter (acrAB-tolC-like) and its function was not clear.

Conjugal transfer experiment showed that all four bla_NDM–1-carrying IncC-type plasmids could be transferred from host strains into the recipient E. coli J53. To assess the impact of the IncC-type plasmid on the growth of the strain, the growth of E. coli J53 was compared with the growth of its corresponding transformant carrying the IncC-type plasmid pHD5030-NDM. The growth curves of E. coli J53 and J53/pHD5030-NDM showed no significant difference (p > 0.05; Supplementary Figure 2). These results indicated that the IncC-type plasmid was self-transmissible, and had little impact on the growth of the E. coli J53 strain in LB.