Complete Sequence of the IncT-Type Plasmid pT-OXA-181 Carrying the bla_OXA-181 Carbapenemase Gene from Citrobacter freundii

Abstract

The gene encoding the carbapenemase OXA-181 (an OXA-48 variant) was identified from a Citrobacter freundii isolate coproducing NDM-1. The whole sequence of plasmid pT-OXA-181 bearing the bla_OXA-181 gene was determined and revealed a 84-kb mobilizable but non-self-conjugative IncT-type plasmid. It totally differs from the 7.6-kb ColE-type and bla_OXA-181-bearing plasmid recently identified in a Klebsiella pneumoniae isolate. However, in both plasmids, insertion sequence ISEcp1 might have played a role in acquisition of the bla_OXA-181 gene.

TEXT

The recent emergence of carbapenemase-encoding genes among Enterobacteriaceae is increasing worldwide. Recently, an extensively drug-resistant Citrobacter freundii isolate (strain STE) was collected in France from a patient who had been transferred from India (1). This isolate was highly resistant to all ß-lactams, including carbapenems, aminoglycosides, sulfonamides, tetracycline, nitrofurantoin, fluoroquinolones, and tigecycline, and remained susceptible to fosfomycin, according to the CLSI guidelines, and EUCAST breakpoints (2, 3). This isolate harbored nine ß-lactamase genes, including the extended-spectrum-ß-lactamase (ESBL) gene bla_CTX-M-15, the narrow-spectrum-ß-lactamase genes bla_OXA-1, bla_OXA-9, bla_OXA-10, and bla_TEM-1, the intrinsic bla_CMY-2-like gene, and three carbapenemase genes, namely, the bla_NDM-1, bla_VIM-4, and bla_OXA-181 genes (1). Preliminary mating-out assays allowed the identification of numerous plasmids in the C. freundii clinical isolate. Multiple plasmids were cotransferred in transconjugants, in several cases impairing the correct estimation of the plasmid size, and assigning of the plasmid family; however, no transconjugants were obtained carrying the bla_OXA-181 gene (1). OXA-181 differs from OXA-48 by four amino acid substitutions, but the two enzymes share exactly the same hydrolysis spectrum (4). OXA-181 hydrolyzes penicillins and (at a lower level) carbapenems and spares broad-spectrum cephalosporins (4). Interestingly, the origin of the bla_OXA-181 gene has been identified, being the waterborne Gram-negative species Shewanella xiamenensis (5). Recently, the bla_OXA-181 gene has been identified in a Klebsiella pneumoniae KP3 isolate from Oman on a small non-self-conjugative but mobilizable ColE2-type plasmid backbone on which no additional resistance gene was present (4, 6).

The aim of our study was to identify the plasmid and genetic environment of the bla_OXA-181 gene in the STE C. freundii strain. Preliminary experiments using primers designed to target sequences specific for ColE-type plasmids revealed that the bla_OXA-181 gene in the STE strain was not located on such a plasmid type, in contrast to what had been reported from K. pneumoniae KP3 (4).

Repeated conjugation experiments have been conducted as described previously (7) using Escherichia coli J53 as the recipient and selecting transconjugants on Trypticase soy agar plates supplemented with imipenem at 0.5 μg/ml and sodium azide at 100 μg/ml. We successfully obtained a single E. coli transconjugant harboring the bla_OXA-181 gene, which, however, contained three plasmids, with two of them being nontypeable by PCR-based replicon typing (PBRT), whereas one was identified as an IncT type (8). It is noteworthy that this E. coli transconjugant was susceptible to broad-spectrum cephalosporins and showed reduced susceptibility to carbapenems, in accordance with the OXA-181 hydrolysis profile, suggesting that there was no associated ß-lactamase resistance marker and, in particular, no ESBL production. In addition, it was resistant to sulfonamides and trimethoprim (data not shown). Sequencing of the plasmid content of the bla_OXA-181-positive E. coli transconjugant was performed using the 454-Genome Sequencer FLX procedure, using libraries obtained from total plasmid DNA, purified by the Invitrogen PureLink HiPure plasmid filter midiprep kit (Invitrogen, Milan, Italy) according to the manufacturer's protocol. Contigs with at least 50-fold coverage obtained by the use of GS-FLX gsAssembler software were assembled into continuous plasmid sequences. Contig assembly and the remaining gaps were confirmed and filled by PCR-based gap closure, by Sanger DNA sequencing of the amplicons (Applied Biosystems, Foster City, CA).

Plasmid pT-OXA-181 (GenBank accession no. JQ996150) was 83,698 bp in size and possessed an IncT-type replication initiation module. Sequence analysis showed that it was a derivative of the IncT-type plasmid Rts1 (217,182 bp; GenBank accession no. NC_003905) that had been recovered from Proteus vulgaris (9), but a large portion of the IncT plasmid backbone was missing here, including several mobile elements, the PvuII restriction-antirestriction system, the SOS repair system (umuC, umuD), and a large part of the transfer system. Among the conserved features (corresponding to 98% nucleotide identity and 60% sequence coverage), only the replicon, the origin of transfer, and the genes encoding the partitioning proteins were identified.

The insertion of the Tn1696 transposon occurred between the repA gene and the mucA and mucB genes of the IncT backbone, as suggested by the presence of direct repeats (DR) and inverted repeats (IR) located at boundaries of the transposon (Fig. 1). The Tn1696 transposon contained a class 1 In4-like integron carrying the dfrA1 gene cassette encoding trimethoprim resistance, the orfC cassette, and a partial mer locus interrupted by an IS6100 element (Fig. 1). The region flanking the Tn1696 transposon and the mucA, mucB, and DNA terminase genes were identical to those identified on the IncT-type plasmid Rts1 (Fig. 1).

Fig 1.

Major structural features of plasmids pT-OXA-181, pN-Cit, and pMobC identified in the E. coli transconjugant. Plasmid pT-OXA-181 was compared with the most closely related IncT-type plasmid, pRts1 (GenBank accession no. NC_003905) (9). Arrows indicate the deduced open reading frames (ORFs) and their orientations. White arrows were used for ORFs encoding plasmid basic functions, with the exception of the tra gene clusters, which are indicated with a unique white arrow that includes capital letters, each referring to the respective tra gene (i.e., J, traJ; D, traD; I, traI, etc.). Resistance genes are indicated by orange arrows, except for a blue arrow that indicates the bla_OXA-181 gene. Transposon-related genes (tnpA, tnpR, and tnpM), class 1 integrase, resolvases, and insertion sequences are indicated by red arrows. Other genes are indicated by colored arrows as follows: violet, replicase genes; green, restriction enzymes and DNA methylase genes. Locations of transposons Tn1696, Tn2013, Tn6901, Tn9199, and Tn2680 are indicated above the map. The gray regions between the pT-OXA-181 and Rts1 plasmids indicate nucleotide identity ≥ 98% by BLASTN.

In accordance with our experimental observations, the conjugation capability of pT-OXA-181 was impaired by a large deletion that had occurred inside the transfer system genes, with only the TraN- and partial TraG-encoding genes remaining (Fig. 1). Basically, the traG gene was truncated by the integration of a Tn3-like transposon. At the 3′-end of the tnpR of the Tn3-like transposon (from nucleotide [nt] 12797 to nt 15771), a deleted version of the Tn2013 transposon (2,974 bp in length) was identified, showing 99% nucleotide identity to the Tn2013 transposon sequenced from plasmid pKP3-A and harboring ISEcp1 and the bla_OXA-181 and ΔlysR genes but lacking the Δere gene (4). In this position, 115 bp of insertion sequence ISKpn14 were detected. In plasmid pKP3-A, transposon Tn2013 was made of ISEcp1, which had mobilized the bla_OXA-181 gene by a one-ended transposition. During such transposition, the left boundary corresponded to the left inverted repeat of ISEcp1 (IR_L) and the right boundary corresponded to an imperfect right inverted repeat sequence, which has been misrecognized by ISEcp1 transposase (IRR₂). Interestingly, the IRR₂ was not identified in pT-OXA-181. At this site, the relict of the ISKpn14 element was identified instead. However, a 5-bp target site duplication (ATTTA) was identified at the extremities of that acquired fragment, suggesting that it resulted from a transposition process in the pT-OXA-181 plasmid (Fig. 1). A Tn3-like transposase and resolvase genes similar to those found in the AbaR1 resistance island from Acinetobacter baumannii were also identified in the region flanking Tn2013 (10) (Fig. 1).

The hypothesis of mobilization in trans of the pT-OXA-181 plasmid by a coresident plasmid was reinforced by the presence of the second plasmid in the bla_OXA-181-positive E. coli transconjugant. That plasmid, termed pN-Cit, was 34,826 bp in size (GenBank accession no. JQ996149), and its backbone was that of an IncN-type plasmid, possessing genes encoding a complete transfer locus (Fig. 1). This plasmid might therefore have efficiently promoted the conjugation and mobilized plasmid pT-OXA-181 in trans. However, it is still possible that plasmids coresident in the C. freundii STE strain, other than those identified in the OXA-181 transconjugant, may have promoted the conjugative transfer of both the pT-OXA-181 and pN-Cit plasmids. Detailed analysis of the sequence of pN-Cit revealed that it belonged to a new IncN subgroup, here designated IncN3, which is significantly different from plasmids previously classified as IncN1 or IncN2 (11, 12). In fact, there was not significant nucleotide homology but the order and type of genes were similar to those of IncN2-type plasmid p271-A, carrying the bla_NDM-1 gene recently identified in E. coli (12). Furthermore, BLASTP analysis demonstrated that the RepA replicase showed 80% amino acid identity to that of p271-A but showed less than 20% amino acid identity to that of the IncN1 reference plasmid R46. Nevertheless, the proteins from the transfer locus of pN-Cit showed variable levels (70% to 80%) of identity to those of the IncN1 or IncN2 plasmids, thus suggesting that this plasmid might represent a derivative of the heterogeneous IncN plasmid family rather than a novel plasmid family.

Plasmid pMobC, with a size of 8,969 bp (GenBank accession no. JQ996148), was also identified in the bla_OXA-181-positive E. coli transconjugant. This plasmid did not show any significant nucleotide identity to previously described plasmids available in GenBank. However, it encoded MobA- and MobB-like proteins likely contributing to its mobilization, sharing 46% amino acid identities with MobA and MobB sequences of plasmid pPNS3 (accession no. NZ_AKKN01000017) identified in a whole-genome shotgun sequencing of Providencia sneebia DSM 19967 (Fig. 1). pMobC did not carry any gene that could apparently provide any advantage to the host strain.

This study identified a novel genetic vehicle of the emerging bla_OXA-181 carbapenemase gene which recently has been identified in different enterobacterial species, including K. pneumoniae and Providencia rettgeri, and in different countries (13). Interestingly, the OXA-181-producing isolate was recovered from a patient who had a link with India, and coproduced the NDM-1 carbapenemase encoded by a different plasmid, which was recently fully sequenced (14). Thus, as recently evidenced by the identification of a large variety of plasmid backbones at the origin of the acquisition and spread of the bla_NDM and bla_KPC genes (7, 15–19), this report highlights the finding that the genetic structures at the origin of the mobilization and spread of the bla_OXA-181 gene are diverse and do not correspond to one single epidemic plasmid. This is in contrast to what has been shown with bla_OXA-48, whose spread is mainly linked to the spread of an epidemic IncL/M-type plasmid. The location of the bla_OXA-181 gene on two distinct plasmid scaffolds can be explained by the fact that this gene is associated with ISEcp1, which can frequently promote its mobilization by a one-ended transposition process, as previously demonstrated for the bla_CTX-M genes (20).

Nucleotide sequence accession numbers.

The nucleotide sequences reported in this work have been deposited in the GenBank nucleotide database under accession no. JQ996148 (pMobC), JQ996149 (pN-Cit), and JQ996150 (pT-OXA-181).