Complete Sequence of pJIE186-2, a Plasmid Carrying Multiple Virulence Factors from a Sequence Type 131 Escherichia coli O25 Strain

Abstract

The complete sequence of a 137-kb plasmid, pJIE186-2, from a sequence type 131 (ST131) Escherichia coli strain was determined. pJIE186-2 contained IncF replicons (FIB, FIIA, and FIAΔ), an incomplete conjugative region, and multiple virulence factors (sitABCD, iucABCD-iutA, iroCDEN, etsABC, hlyF, iss, ompT, and vagCD) but no antimicrobial resistance genes. The host strain also had another plasmid, pJIE186-1, carrying multiple resistance genes. The two plasmids conferred selective advantages for the host strain, contributing to the recent emergence of ST131 E. coli.

TEXT

Escherichia coli of sequence type 131 (ST131) has emerged as an important pathogen, mediating the intercontinental spread of bla_CTX-M-type extended-spectrum β-lactamase genes, particularly bla_CTX-M-15 (1, 2). Genome sequencing of ST131 isolates may shed light on mechanisms for the emergence of this clonal group.

An ST131 E. coli clinical isolate, JIE186 (3), isolated in western Sydney, Australia, in 2006 from urine, was subjected to whole-genome sequencing with a 114.18× coverage, using the Solex Sequencer following the manufacturer's protocol. Reads were assembled to contigs and scaffolds using the SOAP program. Sequences of putative plasmid origins were assembled with intervals between scaffolds being filled by PCR with primers designed based on available sequences and Sanger sequencing.

This isolate harbored two plasmids. One plasmid, designated pJIE186-1, was 91 kb in size, had an IncFII-related replicon (RepFIIA), and carried multiple antimicrobial resistance genes, i.e., aac(3)-II, aac(6′)-Ib-cr, bla_CTX-M-15, bla_OXA-30, bla_TEM-1b, and tetA(A). pJIE186-1 was almost identical to pC15-1a (4), an epidemic bla_CTX-M-15-carrying plasmid of an ST131 isolate from Canada, as reported previously (5).

The other plasmid, designated pJIE186-2 (Fig. 1A), was 137 kb in size and had an average 48.8% GC content. A total of 141 open reading frames were identified, and pJIE186-2 had the features summarized below.

Fig 1.

Map of pJIE186-2 and comparison with several other plasmids. (A) Circular map of pJIE186-2. The map was generated using CGView (http://stothard.afns.ualberta.ca/cgview_server/). Some important features are indicated. (B) Comparison of pECOS88, pJIE186-2, pCVM29188_146, pCS0010A, and pAPEC-1. Comparison was performed using the WebACT program (http://www.webact.org/WebACT/home). To facilitate the comparative view, the sequences of pCVM29188_146, pCS0010A, and pAPEC-1 retrieved from GenBank were reversed and set with new beginning positions that correspond to the original positions 94,458, 94,458, and 78,994, respectively. Locations of a few important features of pJIE186-2 are shown with arrows (not to scale). FIB, RepFIB; sit, sitABCD; FIAΔ, truncated RepFIA; traΔ, incomplete conjugative region; FII, RepFIIA; V, colicin V module; iro, iroBCDEN; ets, etsABC.

Multiple replicons.

pJIE186-2 contained multiple replicons, including a FIB replicon (RepFIB), a RepFIIA, and a remnant of a FIA replicon (RepFIAΔ). RepFIA of pJIE186-2 was truncated by the insertion of IS2, and this RepFIAΔ is identical to those of Salmonella plasmids (S. enterica subsp. enterica serovar Heidelberg; GenBank accession numbers in parentheses) pSH163_120 (JN983046) and pSH696_117 (JN983047) except for one less copy of the iteron in the oriS (origin of replication). RepFIB of pJIE186-2 was almost identical (99.8% identity) to those on pECOS88 (CU928146) and pAPEC-02-ColV (AY545598) of E. coli. RepFIIA of pJIE186-2 was closest (98.1% identity) to the first RepFIIA of E. coli plasmids pETN48 (FQ482074) and pIP1206 (AM886293) (6), both of which contained two RepFIIA. Of note, the two plasmids pJIE186-1 and -2 both have a RepFIIA replicon, and their coexistence in a single strain is likely due to the differences between their inc sequence, which encodes an antisense RNA for plasmid incompatibility.

Sophisticated maintenance system.

pJIE186-2 harbored an F-like sopA-sopB-sopC module rather than the R100-like stbA-stbB for plasmid partitioning. In addition, pJIE186-2 carried two toxin-antitoxin postsegregation killing systems, the ccdA-ccdB and srnC-srnB-srnB′ modules. Of note, the sok-mok-hok and pemK-pemI postsegregation killing systems that were widely distributed on IncF plasmids were absent from pJIE186-2.

Incomplete conjugative region.

pJIE186-2 had a 21.4-kb region for conjugation, which was highly similar to that on pIP1206, with 99% nucleotide identity. However, a large (ca. 14-kb) region between traC (encoding the F pilus assembly protein) and traD (encoding a coupling protein connecting the relaxosome and the transferosome) was missing on pJIE186-2 due to unknown mechanisms. In addition, both traC and traD were disrupted into two parts by the direct insertion of ISEc23, evidenced by the presence of 8-bp direct target repeats (DR). Due to the lack of a large part of the conjugation module and the disruption of traD that is essential for conjugation, pJIE186-2 is unlikely to be self-transmissible.

Multiple virulence factors.

Interestingly, pJIE186-2 did not harbor any known antimicrobial resistance determinants except for a mig-14 gene. mig-14 was found to contribute to resistance to antimicrobial peptides, such as polymyxin B, by an unclear mechanism in Salmonella (7).

A few known determinants conferring virulence were found on pJIE186-2 (Table 1). The present study demonstrated that some virulence factors reported previously (2, 8) were actually located on a plasmid. Among the virulence factors on pJIE186-2, iutA and ompT were present in 97% of ST131 E. coli isolates, while hlyF, iss, and iroN were found in 8%, 8%, and 11% of isolates, respectively (2). The significant discrepancy in the carriage of virulence factors suggests that iutA and ompT might have been integrated into the chromosome of ST131 E. coli, while hlyF, iss, and iroN were likely carried by a plasmid that was not widely distributed in the ST131 lineage.

Table 1.

Virulence factors on pJIE186-2

Gene or loci	Products	Position	Closest match(es) (% identity)a
sitABCD	Iron transport protein	5116–8565	pAPEC-1 (100)
iucABCD-iutA	Aerobactin siderophore	11840–19868	pAPEC-1, pSH696_117, pSH163_120 (99.9)
vagC1D1b	Virulence-associated gene	23972–24783	pAPEC-O1-ColBM, pEC_L46, pEC_L8, pEK499, pRSB107, pSal8934b, pSD_88, pSH696_117, pSH163_120, pVM01, U302L, pYT2 (100)d
vagC2D2b	Virulence-associated gene	27302–27945	p746, pAPEC-O1-ColBM, pCoo, pETEC_73, pIP1206, pO145-NM, pO26_1, pO26-CRL, pO26-Vir, pSE11-3 (99.2)
iroBCDEN	Salmochelin	103882–113392	pAPEC-O1-ColBM, pAPEC-O103-ColBM, pO83_CORR (100)
iss	Iss (increasing serum survival) protein	116538–116846	pAPEC-1, pAPEC-O103-ColBM, pAPEC-O1-ColBM, pAPEC-O2-ColV, pCS0010A, pCVM29188_146, pECOS88, pO83_CORR, pSSAP03302A, pVM01 (100)c
etsABC	Type 1 secretion system	124931–129429	pAPEC-1, pAPEC-O1-ColBM, pECOS88, pVM01 (99.9)
ompT	Outer membrane protease	133711–134664	pAPEC-1, pAPEC-O2-ColV, pECOS88, pO83_CORR, pSMS35_130, pVM01 (100)
hlyF	Hemolysin	135097–136206	p1658/97, pAPEC-O2-ColV, pCS0010A, pCVM29188_146, pECOS88, pETN48, pO83_CORR, pSMS35_130, pSSAP03302A, pVM01 (100)

^aGenBank accession numbers for plasmids listed in the Table: p746 (FN822748), pAPEC-1 (CP000836), pAPEC-O103-ColBM (CP001232), pAPEC-O1-ColBM (DQ381420), pCoo (CR942285), pCS0010A (CP002090), pCVM29188_146 (CP001122), pEC_L46 (GU371929), pEC_L8 (GU371928), pECOS88 (CU928146), pEK499 (EU935739), pETEC_73 (CP000797), pIP1206 (AM886293), pO145-NM (HM138194), pO26_1 (AP010954), pO26-CRL (GQ259888), pO26-Vir (FJ386569), pO83_CORR (CP001856), pSE11-3 (AP009243), pSH696_117 (JN983047), pSH163_120 (JN983046), pSSAP03302A (CP002089), pU302L (AY333434), pVM01 (EU330199). Plasmids from Salmonella spp. are underlined.

^bpJIE186-2 harbored two copies of vagCD with only 86% nucleotide identity, designated vagC1D1 and vagC2D2 here.

^ciss is also present on the chromosome of E. coli strain NRG 857C (CP001855).

^dvagC1D1 is also present on the chromosome of Salmonella strain T000240 (AP011957).

pJIE186-2 also contained multiple bacteriocin modules, including mcjABCD encoding microcin J25, cbi-cba for colicin B, cma-cmi encoding colicin M, and cvaABC for colicin V (ColV). The host strain JIE186 indeed produced bacteriocin, tested as described previously (9). The presence of mcjABCD appears to be uncommon, as none of the completely sequenced plasmids from the Enterobacteriaceae deposited in GenBank carried this module. Modules encoding colicin M and B (ColBM) are usually clustered together. However, clustering of all of the modules encoding colicin M, B, and V remains uncommon. Among plasmids with sequences deposited in GenBank (GenBank accession numbers in parentheses), only pAPEC-O103-ColBM (CP001232), pO83_CORR (CP001856), and pETN48 contained colicin B and M genes plus a remnant of the colicin V module in various sizes. pJIE186-2 is the only completely sequenced plasmid containing all of the complete modules of colicin B, M, and V thus far.

Insertion sequences.

pJIE186-2 carried multiple insertion sequences, including 6 copies of IS1, 3 copies of IS2, 2 copies of ISEc23, and a single copy each of IS629, IS26, ISEc32, and a 590-bp IS21-like element. No DR were identified flanking any of 6 copies of IS1 and 2 copies of IS2, suggesting that these elements were likely to have resulted from homologous recombination rather than direct insertion. In contrast, DR were present abutting IS629, IS26, 1 copy of IS2, and both copies of ISEc23, suggesting that these elements were due to direct insertion. ISEc32, an element of the IS110 family, generates no DR on insertion.

Mosaic overall structure.

On pJIE186-2, a ca. 66-kb large region that contains the colicin V module, multiple virulence factors, and RepFIB and was bracketed by IS1 has 99% identity with the counterparts on (GenBank accession numbers in parentheses) pAPEC-1 (CP000836), pCS0010A (CP002090), and pCVM29188_146 (Fig. 1B). This large region was also highly similar (99% identity) to that on pECOS88 (Fig. 1B), except for an inversion of a ca. 16-kb region containing sitABCD and iucABCD-iutA that was likely due to homologous recombination between two copies of IS1. As this large, 66-kb region is present on all completely sequenced ColV plasmids, it has been proposed as the core part of ColV pathogenomics (10). The conjugation module and RepFIIA of pJIE186-2 had 99% identity with the counterpart on pIP1206, while RepFIAΔ and the ccdAB postsegregation killing system were 99% identical to those on pSH163_120 and pSH696_117. All of these findings revealed that pJIE186-2 is a large plasmid of a highly mosaic structure with genetic components originating from various sources.

In general, pJIE186-2 does not possess many novel genetic components but contains various new combinations of known structures seen in different plasmids. The mosaicism of pJIE186-2 was largely generated by insertion sequences. In particular, as genetic components between different copies of IS1 did not have the same origins, IS1 apparently serves as an “adapter” in pJIE186-2, allowing genetic components to be linked and clustered via homologous recombination between different copies of IS1.

In conclusion, the ST131 isolate JIE186 had two large IncFII-related plasmids, one carrying multiple resistance genes and the other harboring multiple virulence determinants. The coexistence of these two plasmids could confer significant selective advantages for the ST131 host strain.

This work was presented in part at the 49th Annual Meeting of the Infectious Diseases Society of America (11).

Nucleotide sequences accession number.

The complete sequence of pJIE186-2 is deposited in GenBank under accession number JX077110.