Complete genome sequencing and genomic characterization of two Escherichia coli strains co-producing MCR-1 and NDM-1 from bloodstream infection

We previously described the discovery of two Escherichia coli isolates (EC1002 and EC2474) co-harbouring mcr-1 and blaNDM-1 genes, which were recovered from bloodstream infection in China. More importantly, these antibiotic resistance genes were located on different plasmids and signaling the potential spread of pandrug-resistant bacteria. Here, the complete genome sequences of both isolates were determined using Pacbio RS II and Illumina HiSeq2000 systems. The genome of EC1002 consists of a 5,177,501 base pair chromosome and four circular plasmids, while the genome of EC2474 consists of a 5,013,813 base pair chromosome and three plasmids. The plasmid replicon type of pEC1002_NDM and pEC2474_NDM were identified as IncA/C2 and IncF, respectively. The genetic environment of blaNDM-1 in this study was similar to blaNDM-carrying plasmids detected in China, although the overall nucleotide identity and query coverage were variable. The plasmid replicon type of pEC1002_MCR and pEC2474_MCR were identified as IncI2 and IncHI2, respectively. Two different genetic strategies for mcr-1 gene spread were observed in this study and blaNDM-1 genes were also found transferred by two different mobile genetic elements in two plasmids. The findings of this study further support that the diversified transfer mechanisms of blaNDM-1 and mcr-1 present in Enterobacteriaceae.

unclear. In this study, we investigated the genetic features of these two isolates and elaborated on various potential mechanisms by which mcr-1 and bla NDM-1 may be transmitted. In addition, comparative analyses of the genetic contexts of mcr-1 and bla NDM-1 with closely related plasmids were also performed.

Materials and Methods
Bacterial isolation and genome sequencing. E. coli EC1002 and EC2472 carrying both bla NDM-1 and mcr-1 were isolated from BSI patients in the Affiliated Hospital of Jining Medical University and Anhui Provincial Hospital, respectively 13 . Genomic DNA was extracted from overnight cultures using a Gentra Puregene Yeast/ Bact. Kit (Qiaqen, Hilden, Germany) according to the manufacturer's instructions. The harvested DNA was visualized on 1% (w/v) agarose gels, and DNA concentration as well as purity was determined by a NanoDrop 2000 UV-Vis Spectrophotometer (Thermo Scientific, Waltham, MA, USA) and Qubit®2.0 Fluorometer (Thermo Scientific, Waltham, MA, USA). DNA was stored at −20 °C until further processing. The genome of the two isolates was sequenced using the Pacbio RS II (Pacific Biosciences, Menlo Park, CA, USA) and Illumina HiSeq. 2500-PE150 platform (Illumina, San Diego, CA, USA). A 10-kb DNA library was constructed by the PacBio SMRTbell 10 kb Library preparation kit according to the manufacturer's instructions (Pacific Biosciences, Menlo Park, CA, USA). Pair-end index libraries construction followed the NEBNext Ultra DNA Library Prep Kit (Illumina, SanDiego, CA, USA). Library construction and sequencing was performed at Beijing Novogene Bioinformatics Technology Co. Ltd. Genome Assembly. Low quality reads were filtered out and the filtered reads were assembled to generate one contig without gaps by SMRT 2.3.0 using Hierchical Genome Assembly Process (HGAP) V.3.0. Overlaping regions were assessed with Gepard followed by circularization using minimus2 pipeline in the AMOS software package 19 . Subsequently, Illumina HiSeq contigs were mapped over the PacBio-generated contigs to correct the assembled contigs.
Genome annotation, and in silico analyses. Protein-coding genes were initially identified and annotated using RAST 20 and further annotated by BLASTP against UniPort and NR databases, while insertion elements (IS) were identified using IS Finder 21 . Queries were generated using the ResFinder 2.1 database to identify acquired antibiotic resistance genes 22 . Plasmid Finder 1.3 and pMLST were used to identify plasmid incompatibility types 23 . The circular image and circular comparisons between multiple genomes and plasmids was done by BLAST Ring Image Generator (BRIG) 24,25 . Linear comparison figures of multiple plasmids were generated by a Python application Easyfig. 26 .
Nucleotide sequence accession numbers. The complete sequences of E. coli EC1002, EC2474, and other plasmids have been deposited in GenBank under the accession numbers CP021202-CP021210 (Table 1).

Results and Discussion
Basic genomic features. The genomic features and a comparison of EC1002 and EC2474 against other E. coli isolates are summarized in Fig. 1 and Table 1. All plasmids were assembled into a circular ring and the chromosome was assembled into one contig. It was determined that the genome of EC1002 consists of a 5,177,501 base pair chromosome with an average 50.1% GC content and four circular plasmids, while EC2474 consists of a 5,013,813 base pair chromosome with an average 50.6% GC content and three plasmids. Screening for acquired resistance determinants revealed the presence of different kinds of resistance genes ( Table 1). The isolates EC1002 and EC2474 belonged to ST405 and ST131, respectively. E. coli sequence type 131 (ST131) is a worldwide pandemic clone, causing predominantly community-onset antimicrobial-resistant infection and subsequent study has confirmed the worldwide prevalence of ST131 harbouring a broad range of virulence and resistance genes on a transferable plasmid, while ST405, which is an high risk clone found in human, animals and environment usually associated with CTX-M-types 27,28 . At present, several different E. coli ST isolates such as ST167, ST206, ST648, and ST156 have also been reported to carry both bla NDM variants and mcr genes [29][30][31] . Interestingly, a recent study observed significant geographical clustering with regional spread of mcr-1-bearing IncHI2 plasmids in Europe and IncI2 in Asia 32 . The unrelated clonally relationship found in this study suggesting coexist these two plasmids could also be horizontal transferred to other STs. Furthermore, the detection of florfenicol resistance gene, floR, in the genome of isolate EC2474 together with the fact that florfenicol is widely used in veterinary medicine further supports the potential transfer of mcr-1 gene from animals to clinical settings 32 .
Genetic characteristics of plasmids bearing mcr-1. The GC content of mcr-1 bearing plasmids in this study was similar to that of previously reported mcr-bearing plasmids 33 . However, they were also found to be significantly different to other plasmids exist in the same strain (Table 1). pEC1002-MCR is a 63,392 bp circular plasmid encoding the IncI2 replication protein. In contrast, pEC2474-MCR is a 223,982 bp IncHI2 plasmid. The IncI2-type plasmid is considered to be a major genetic event driving the rapid mobilization and acquisition of mcr genes 34 . The IncHI2-type plasmid is characterized by its long as well as conjugation flexible pilus. Furthermore, the thermosensitivity of the conjugative apparatus means that the optimal temperature for conjugation is 22-30 °C rather than 37 °C 35 . Therefore, this strain is more likely to acquire mcr-bearing plasmids in vitro, similar to other reports of mcr-producing strains, which have been mainly isolated from the agriculture industry in China, indicating environmental origins of mcr-1 genes in human pathogens 36,37 . pEC1002-MCR only contained the mcr-1 gene, which is in contrast to other reports where mcr-1 easily co-exists with other resistance genes 38,39 . However, pEC2472-MCR carried several resistance genes, such as bla CTX-M-14 , fosA, and floR. A BLAST search against the nr/nt database indicated that pEC1002_MCR showed an overall nucleotide identity (99-100%) and query coverage (93-97%) similar to several plasmids, such as pMRY16-002_4 (GenBank no. AP017614) 40 , pHeN867 (KU934208), and pEC019 (KY471145) that have been reported in different countries. In addition, the size and backbone structure of these plasmids are quite similar (Fig. 2a). Further analyses revealed three encoding sequence insertions in pEC1002_MCR. The 18,358-19,666 insertion region and 38,741-40,423 region carrying genes encoding for DNA topoisomerase III (topB), integrase (int), and IS1294, respectively (Fig. 2a). The region of 10,223-11,517 encodes for shufflon protein A and two shufflon protein C. These proteins are highly mobile DNA segments that function as a biological switch and generally invert independently or in groups resulting in a complex DNA rearrangement. Furthermore, the shufflon rearrangement is closely related to plasmid transmission in Enterobacteriaceae 40 . Of note, the sequence of nikA-nikB-mcr-1-hp region was identified in pEC1002_MCR, which is in contrast to the 2.6 kb mcr-1-pap2 element usually found in mcr-1-carrying plasmids 41 . The BLASTN comparison of pEC2474_MCR plasmid found 100% nucleotide identity and 100% coverage with pHNSHP45-2 (KU341381), which is the first reported plasmid carrying mcr-1. The main difference between pEC2474_MCR and pHNSHP45-2 is the multidrug resistance region (Fig. 3), suggesting that pHNSHP45-2 may be formed by acquiring an IS region containing several resistance genes.
Genetic context of plasmids bearing bla NDM-1 . The plasmid replicon type of pEC1002_NDM and pEC2474_NDM were identified as IncA/C2 and IncF, respectively. It has been reported that the IncX3-type plasmid is the main type of NDM-producing plasmid spread in China 42 and we first reported the NDM-producing IncA/C2 plasmid in mainland China. While pEC1002_NDM carries a variety of drug resistance genes, pEC2472_ NDM only carries the bla NDM-1 and aph resistance genes (Table 1).
BLASTN comparison of the two NDM-1-producing plasmids revealed that the overall structure of both plasmids showed big differences compared to known NDM-1-producing plasmids. In silico analyses demonstrated that pEC1002_NDM shared 99% nucleotide identity as well as 71%, 87%, and 93% coverage with pNDM-US from K. pneumoniae, pV001-a from E. coli, and pRJ119-NDM from K. pneumoniae, respectively. Although the overall structure of these plasmids was different, the genetic context of bla NDM-1 was relatively similar in all plasmids (Fig. 4a), where bla NDM-1 is located in a mobile region with a structure of rmtc-ISKpn14-bla NDM- Figure 2. Circular representation of the studied plasmids. GC content and GC Skew are represented on the distance scale (in kbp) on the inner map. Each plasmid was compared to its most closely-related plasmid (Genebank accession numbers shown on the right side). The arrows around the map indicate deduced ORFs and their orientation. Certain important genes are also indicated on the ring. The schematics were generated through the 'BLAST Ring Image Generator' (BRIG) program. Yellow arrows indicate conjugal transfer-involved genes. Genes associated with plasmid stability are colored by brown. Antimicrobial resistance genes and mobile elements genes were indicated by red and green arrows, respectively. Grey arrows indicate genes for hypothetical proteins as well as proteins with unknown function. ORFs are labeled above the arrows and colored as described in Fig. 1. for the pEC2474_NDM plasmid, 99% nucleotide identity was found as well as 89%, 89%, and 87% coverage with plasmid pABC143C-NDM, pCC1410-1, and pYHCC, respectively (Fig. 2d). The backbone structure of these plasmids was similar, except for the region containing bla NDM-1 (Fig. 4c). The upstream region of bla NDM-1 encoded a recombinase (recA), while a common gene environment around bla NDM-1 (ISAba125-bla NDM-1 -ble MBL -trpF-dsbC) was identified 36,45 . In addition, the region after this structure also contained four mobile genes (Fig. 4c).

Conclusion
In this study, we report the complete genome sequences of two E. coli strains with coexisting genes, mcr-1 and bla NDM-1 . Two different genetic strategies for mcr-1 transmission were observed in these two strains. Firstly, the transfer of mcr-1 associated with a nikA-nikB-mcr-1-hp structure was observed in pEC1002_MCR, while the ISApl1-mcr-1 mobile element played an important role in pEC2474_MCR. Additionally, a common gene environment around bla NDM-1 (rmtc-ISKpn14-bla NDM-1 -ble MBL -trpF-tat-dsbC) was detected in pEC1002_NDM, while an ISAba125-bla NDM-1 -ble MBL -trpF-dsbC structure was identified in pEC2472_NDM. Taken together, this study further supports that the diversified transfer mechanisms of bla NDM-1 and mcr-1 present in Enterobacteriaceae.