Novel Plasmid and Its Variant Harboring both a bla_NDM-1 Gene and Type IV Secretion System in Clinical Isolates of Acinetobacter lwoffii

Whole-genome sequencing of A. lwoffii 10621.The genomic draft sequence of A. lwoffii 10621 includes a ∼3.42-Mb chromosome and a ∼45-kb plasmid. Two gaps in the plasmid were closed by PCR and sequencing, and a final complete plasmid (named pNDM-BJ01) with length of 47,274 bp was obtained. The structure of pNDM-BJ01 was further confirmed by PCR using 12 pairs of primers with overlap ends of the product (see Table S1 in the supplemental material).

General features of plasmid pNDM-BJ01.The plasmid pNDM-BJ01 contains 46 open reading frames (ORFs) with an average GC content of 40.8% (Fig. 1; see Table S2 in the supplemental material). Three different functional regions were predicted (Fig. 1): a putative transfer and replication region containing plasmid transfer genes traA and traC and plasmid-partitioning gene parA, a type IV secretion system (T4SS) gene cluster region containing the quintessential T4SS conjugative transfer genes traB and traC, and a bla_NDM-1-carrying a composite transposon region flanked by two copies of ISAba125. A BLAST search showed that approximately 20 kb of pNDM-BJ01 displayed more than 70% identity with the draft genomic sequence of Acinetobacter radioresistens SK82 contig 00019 (GenBank Genome Project ID 55855). However, pNDM-BJ01 cannot be classified by the PCR-based replicon typing method (2), and no plasmid backbone sequence homologous with it can be found, indicating that it may be a novel plasmid with an unknown maintenance mechanism. The plasmid showed a relatively high transfer frequency, ranging from 9.1 × 10⁻³ to 1.3 × 10⁻² per donor cell, to E. coli J53 Azi^r, suggesting a high horizontal transfer ability.

• Open in new tab
• Download powerpoint

Fig 1

Circular map of plasmid pNDM-BJ01. The circles display (from the outside) (i) coordinates in kilobase pairs, (ii) predicted coding sequences, (iii) GC skew ([G+C]/[G-C]) in a 500-bp window, and (iv) GC content plotted against 50% GC content, with light gray indicating <50% and dark gray indicating >50%. The bla_NDM-1 gene and insertion element ISAba125 are colored with red and purple, respectively. ORFs in white indicate hypothetical proteins. Three major parts corresponding to the bla_NDM-1-carrying composite transposon, the type IV secretion system gene cluster, and the putative transfer and replication region are shaded with light red, light blue, and light green, respectively. Homologous regions with Acinetobacter radioresistens SK82 contig 00019 are shaded with gray.

Ten genes in the predicted T4SS region show high homology with genes for putative type IV secretion system proteins in A. radioresistens SK82. Previous studies showed that genes traB/virB10 and traC/virB4, encoding quintessential T4SS proteins, are ideal for comparative amino acid alignment and appear to be sufficient to define membership in a T4SS type (9, 10). A phylogenetic tree was therefore constructed by using concatenated amino acid sequences for virB10 and virB4. The results indicated that the T4SS in pNDM-BJ01 was closely related to that in A. radioresistens SK82, and it was classified as a known P-type T4SS, members of which include those in RP4 (Pseudomonas aeruginosa), Ti (Agrobacterium tumefaciens C58), and pRi1724 (Agrobacterium rhizogenes) (Fig. 2).

• Open in new tab
• Download powerpoint

Fig 2

Phylogenetic tree of representative members of the T4SS family. The tree was constructed by the neighbor-joining method using concatenated amino acid sequences of virB10 and virB4 of the pNDM-Bj01 T4SS (●) and the previously described T4SSs. Bootstrap values of more than 50% are shown at the respective nodes (1,000 replications). The lengths of the branches indicate the divergence among the amino acid sequences. The scale bar corresponds to 20% estimated sequence divergence.

The bla_NDM-1 gene was organized in a composite transposon structure within a ∼7-kb abnormal GC region bracketed between two copies of insert sequence ISAba125 which differed by 6 nucleotides, one of which resulted in the amino acid substitution Arg-60 → Gln. With both ISAba125 insertion elements, this composite transposon was 10,099 bp in length, and the transposition event of this composite transposon was confirmed by (i) a different G+C content compared with the rest of the plasmid and (ii) a 3-bp (GTT) target site duplication at the point of insertion between the aphA6 gene and its 3′ flanking region, which were continuous in Alcaligenes faecalis transposon Tn5393d (Fig. 3A). It was interesting that a nearly identical composite transposon structure (with a difference of 2 bp in a 10,099-bp length) was found to be inserted in the gene encoding a major facilitator superfamily (MFS) metabolite/H⁺ symporter in the chromosome of A. baumannii 161/07, which was isolated from a patient who had been repatriated to Germany from Serbia (19). Moreover, the whole component of this composite transposon (excluding ISAba125) was also captured by the IS903 mobile element in pNDM_Dok01 (Fig. 3A).

• Open in new tab
• Download powerpoint

Fig 3

Comparative analysis of bla_NDM-1-carrying composite transposon sequences (A), promoter sequences of bla_NDM-1 (B), and 3′-end sequences of trpF (C) in different genetic environments. (A) Different colors of ORFs denote different sequence origins. Light gray and pink shadings indicate more than 97% and less than 75% nucleotide identity, respectively. ISAba125 is colored with red. The dashed line box indicates the complete composite transposon. (B) Gray shading indicates the right inverted repeat (IR-R) of ISAba125.

Comparative analysis of the genetic environment of bla_NDM-1.The bla_NDM-1 gene-containing region between two copies of ISAba125 in pNDM-BJ01 comprised three major parts (Fig. 3A). In the 5′ region, the bla_NDM-1 gene was adjacent to a bleomycin resistance protein followed by a truncated trpF gene. This three-gene combination was frequently found in other genetic backgrounds: E. coli plasmids pNDM-HK, p271A, and pNDM_Dok01; the NDM-1 plasmid in E. coli DVR22; and K. pneumoniae plasmid pKpANDM-1. In the 3′ region, genes encoding the chaperonin subunits GroS (truncated) and GroEL and the transposase InsE constitute a 4-kb element showing high homology to that described in E. coli plasmids peH4H and pAR060302 and Salmonella enterica plasmid pSN254. The third part consisted of two ORFs between the truncated trpF and groS genes displaying ∼70% nucleotide homology with the genome of Stenotrophomonas maltophilia K279a. It seemed that this part was related to insertion events. However, no feature sequences, such as direct repeats or inverted repeats, can be found flanking this region therefore, exactly how the insertions happened needs to be further investigated. These results indicated that the bla_NDM-1-containing region in pNDM-BJ01 was probably formed via multiple recombination events by genetic elements of different origins.

Detailed analysis of the sequence spanning ISAba125 and bla_NDM-1 revealed the existence of a promoter of bla_NDM-1 gene in which the −35 region was located inside the right inverted repeat (IR-R). Sequence alignment showed that all the reported NDM-1-containing sequences have the same promoter region and harbor residual nucleotides of the ISAba125 3′ sequence (including the right inverted repeat of ISAba125) in different lengths (pNDM-HK, 256 bp; pKpANDM-1, 255 bp, p271A, 194 bp; pNDM_Dok01, 273 bp; E. coli DVR22, the same copy of ISAba125 as in pNDM-BJ01) (Fig. 3B). These truncated ISAba125 ORFs had been disrupted by other insertion sequences, such as ISEC33 in p271A, IS26 in pNDM-HK and pKpANDM-1, and IS903 in pNDM_Dok01. Interestingly, available sequence information showed that the 3′ end of the bla_NDM-1 gene was always adjacent to a bleomycin resistance protein followed by a trpF gene. Sequence alignment indicated that in E. coli DVR22, pNDM-HK, and pKpANDM-1 there is a complete ORF of trpF (660 bp) but that in pNDM-BJ01 and pNDM_Dok01 79 bp of its 3′ end is missing and that in p271A 271 bp of its 3′ end is missing (Fig. 3C). These results suggested that the bla_NDM-1 gene was probably originally linked to ISAba125 and then disseminated in the form of truncated ISAba125 and that the sequence in E. coli DVR22 may be closely related to the ancestral form, as both ISAba125 and trpF flanking the bla_NDM-1 gene were intact.

Upstream of the ISAba125 gene was a complete ORF carrying the aminoglycoside resistance gene aphA6 (12); however, the results of antibiotic susceptibility testing showed that A. lwoffii 10621 harboring pNDM-BJ01 was susceptible to amikacin, an aminoglycoside antibiotic. Analysis of the 5′ flanking region of the aphA6 gene indicated that no obvious promoter sequence could be detected by using different promoter searching programs, and upstream of the aphA6 gene were two transpose genes which were very similar to those in the genomes of A. lwoffii and A. baumannii (two copies) (Fig. 3A). Therefore, we suggest that the promoter sequence of the aphA6 gene was probably disrupted by the transposition event, thus resulting in the failure to resistant amikacin.

Characterization of pNDM-BJ02.In the second case described above, we had the opportunity to isolate another bla_NDM-1 gene-positive A. lwoffii strain harboring a plasmid (designated pNDM-BJ02) very similar in size to pNDM-BJ01. To investigate whether the two plasmids were identical, the whole pNDM-BJ02 was sequenced by primer walking of 12 PCR products; a final assembly with a length of 46,165 bp was obtained. Sequence comparison of pNDM-BJ02 with pNDM-BJ01 revealed that an 1,109-bp fragment adjacent to transposase gene insE was missing in pNDM-BJ02 (Fig. 3A). This deletion includes 94 bp of the 5′ flanking region of the ISAba125 ORF and 1,015 bp of the ORF itself (total, 1,026 bp), thus leaving 11 nucleotides of the 3′ end of ISAba125 gene and its downstream 3-bp transposition target site duplication on pNDM-BJ02. The rest of the sequences of the two plasmids are identical to each other. We therefore suggest that plasmid pNDM-BJ02 is a recent variant of pNDM-BJ01.