IS1R-Mediated Plasticity of IncL/M Plasmids Leads to the Insertion of bla_OXA-48 into the Escherichia coli Chromosome

Abstract

The OXA-48 carbapenemase is mainly encoded by ∼62-kb IncL/M plasmids. However, chromosome-mediated genes have been observed in Escherichia coli isolates. In this work, we investigated the genetic environment of OXA-48 in members of the family Enterobacteriaceae (n = 22) to understand how the OXA-48-encoding gene is transferred into the E. coli chromosome. The OXA-48-encoding gene was located within intact Tn1999.2 transposons in the ∼62-kb plasmids or within a truncated variant of Tn1999.2 for the OXA-48-encoding genes located in the chromosomes of E. coli bacteria. The analysis of the Tn1999.2 genetic environment revealed an inverted orientation of the transposon in five ∼62-kb plasmids (5/14 [35%]) and in all chromosome inserts (n = 8). The sequencing of pRA35 plasmid showed that this orientation of Tn1999.2 and the acquisition of an IS1R insertion sequence generated a 21.9-kb IS1R-based composite transposon encoding OXA-48 and designated Tn6237. The sequencing of a chromosomal insert encoding OXA-48 also revealed this new transposon in the E. coli chromosome. PCR mapping showed the presence of this element in all strains harboring an OXA-48-encoding chromosomal insert. However, different insertion sites of this transposon were observed in the E. coli chromosome. Overall, these findings indicate a plasticity of the OXA-48 genetic environment mediated by IS1R insertion sequences. The insertion sequences can induce the transfer of the OXA-encoding gene into E. coli chromosomes and thereby promote its persistence and expression at low levels.

INTRODUCTION

Carbapenems are broad-spectrum antibiotics that constitute the last-line therapeutic option available to treat infection caused by multidrug-resistant members of the family Enterobacteriaceae (1). However, owing to the emerging resistance to carbapenems worldwide, the antimicrobial activity of these drugs is no longer guaranteed (2). An underlying mechanism is the acquisition of carbapenem-hydrolyzing β-lactamases encoded by genes located on mobile genetic supports, which facilitate their diffusion among bacteria (3).

The OXA-48 carbapenemase was initially identified from a Klebsiella pneumoniae isolate in Turkey (4). It then spread to various Enterobacteriaceae species especially throughout the southern Mediterranean area and in Europe (4, 5). The OXA-48-encoding genes (bla_OXA-48) are mostly found in K. pneumoniae and, to a lesser extent, in Escherichia coli and Enterobacter spp. (5). Their emergence is mediated by the rapid spread of strains containing broad-host-range conjugative IncL/M plasmids (6–8) harboring bla_OXA-48 located within the Tn1999-type composite transposon (6, 9–11). The plasmid implicated in the diffusion of bla_OXA-48 was a widespread 62-kb IncL/M plasmid (6). Derivatives have also been reported (7, 8). However, bla_OXA-48 was observed in E. coli chromosomes, including in a strong killer strain (12).

The aim of this work was to investigate the genetic environment of bla_OXA-48 in order to understand how this gene is transferred into the E. coli chromosome.

MATERIALS AND METHODS

Bacterial isolates.

Twenty-two Enterobacteriaceae strains with reduced susceptibility or resistance to ertapenem and/or imipenem were used in the study (Table 1). Nineteen strains were recovered from clinical samples of patients hospitalized in Nini Hospital, Tripoli, Lebanon, between January 2008 and December 2012. The other three strains were collected in December 2012 during a study investigating intestinal carriage in healthy children. The strains were identified using the matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (MS) Vitek microbial identification system (bioMérieux, La Balme, France). Rifampin-resistant E. coli C600 was used for mating-out assays.

TABLE 1.

Phenotypic and genotypic characteristics of OXA-48-producing isolates

Isolate	STa (E. coli phylogroup)	Clinical sampleb (yr)	OXA-48 genetic environment
			Transposon	PCR resultc			Genetic supportd
				parA	repA	traU
E. coli isolates
EC37	88 (A)	Stoole (2012)	Inverted ΔTn1999.2	+	−	−	Chrom
EC267	88 (A)	Sputum (2012)	Inverted ΔTn1999.2	+	−	−	Chrom
EC264	227 (A)	Urine (2012)	Inverted ΔTn1999.2	+	−	−	Chrom
EC15	127 (B2)	Sputum (2011)	Inverted ΔTn1999.2	+	−	−	Chrom
EC8	131 (B2)	Urine (2012)	Inverted ΔTn1999.2	+	+	+	Chrom
EC254	38 (D)	Urine (2012)	Inverted ΔTn1999.2	+	−	−	Chrom
EC265	38 (D)	Pus (2012)	Inverted ΔTn1999.2	+	−	−	Chrom
EC49	38 (D)	Stoole (2012)	NTf	+	−	−	Chrom
EC253	617 (A)	Urine (2012)	Tn1999.2	+	+	+	62kb
EC269	617 (A)	Urine (2012)	Tn1999.2	+	+	+	62kb
EC119	3877 (A)	Stoole (2012)	Tn1999.2	+	+	+	62kb
EC260	1711 (B1)	Urine (2012)	Tn1999.2	+	+	+	62kb
K. pneumoniae isolates
KP78	1157	Blood (2012)	Tn1999.2	+	+	+	62kb
KP87	1156	Pus (2012)	Tn1999	+	+	+	62kb
KP34	45	Urine (2011)	Tn1999.2	+	+	+	62kb
KP9	866	Urine (2009)	Inverted Tn1999.2	+	+	+	62kb
KP26	866	Urine (2008)	Tn1999.2	+	+	+	62kb
E. aerogenes EA2		Pus (2012)	Inverted Tn1999.2	+	+	+	62kb
E. cloacae isolates
ECL6		Urine (2010)	Inverted Tn1999.2	+	+	+	62kb
ECL8		Urine (2009)	Tn1999.2	+	+	+	62kb
C. koseri CK8		Urine (2011)	Inverted Tn1999.2	+	+	+	62kb
R. planticola RA35		Pus (2011)	Inverted Tn1999.2	+	+	+	62kb

^aSequence types for E. coli and K. pneumoniae isolates.

^bClinical sample that the strain was isolated from.

^cPCR result for “pOXA-48a genes.”

^dChrom, chromosome-mediated OXA-48; 62kb, ∼62-kb IncL/M plasmid encoding OXA-48.

^eFecal carriage in children having no known contact with a hospital.

^fNT, nontypeable Tn1999-type environment.

Determination of genetic relatedness and E. coli phylogroup.

The genetic relatedness of the strains was determined by multilocus sequence typing (MLST). Allelic profiles and sequence types (STs) were assigned using the MLST schema of Diancourt et al. (13) and Wirth et al. (14) for K. pneumoniae and E. coli isolates, respectively.

Molecular characterization of bla_OXA-48 and the associated transposon.

The OXA-48-encoding genes were detected by PCR amplification as previously reported (12). The genetic environment of the bla_OXA-48 gene was further investigated by PCR and sequencing using specific primers to map the transposon composites Tn1999 to Tn1999.4 (6, 9–11) (Table 2). The orientation of these transposons in their genetic environments was investigated by PCR and sequencing using combinations of primers targeting bla_OXA-48 (5′-AATACACGCATAACGTCCCC-3′ and 5′-GCCATCACAAAAGAAGTGCTC-3′), lysR (5′-GCTGCAAATAGCATCATACC-3′), orf67 (5′-CAGCCAGAATAAGAGCAATC-3′), and pemI (5′-TGACCATGCCAACTTCATT-3′). The sequences of PCR products were established on both strands by dideoxy chain terminator using an Applied Biosystems sequencer.

TABLE 2.

Primers used in this study

Target	Primer sequence (5′ to 3′ end)	Positioning of primers in pOXA-48a sequence	Reference
IS1999	CAGCAATTCTTTCTCCGTG	2.780–2.798/7.520–7502a	6
	CAAGCACAACATCAAGCGC	3.764–3.746/6.536–6.554a
blaOXA-48	AATACACGCATAACGTCCCC	6.232–6.251	32
	GCCATCACAAAAGAAGTGCTC	5.489–5.469
lysR	GCTAGTGCCAATCTTACAGG	4.983–5.002	10
	GCTGCAAATAGCATCATACC	4.262–4.243
pemI	TGACCATGCCAACTTCATT	8.022–8.004	This study
	AAACATACCAGCCTTTAACG	7.869–7.850
	CCCTCCTGACCAGCCGCC	8.163–8.149
trbB	CACCGCTTCACCAGAAAA	61.698–61.715	This study
trbA	GAAGCGTTCCGAGAATTTG	519–501	This study
trbN	CAGAGAATTTTGGTAAACC	1.314–1.332	This study
	CCGCCCTCAATCCCGCTGC	1.688–1.670
mucA	CGTTCACTGTCATGAGACAC	8.995–9.014	This study
	GGATCAATACCTTCTCCG	10.203–10.187
parA	GCAGTGAAAACGTTGATCAG	18.746–18.765	6
	GATCGCAATGCGTCTTGGTG	19.277–19.258
nuc	GATAAGAATCTGGAAGAAAC	20.047–20.066	This study
	GAGCCTTGATAATACGACG	20.745–20.727
orf22	CCTGCAGGGCTATTTATTC	22.482–22.500	This study
korC	TTTTGCGCTGCTATTTCC	23.859–23.842	This study
orf25	GGAAATAGCAGCGCAAAA	23.842–23.859	This study
	CCATAACATCGCCATCAT	24.048–24.031
	ATGATGGCGATGTTATGG	24.031–24.048
	GGTTCGATGCCCGTATTG	25.237–25.220
orf28	GACTGGCGCAGGAGTAAA	28.046–28. 063	This study
orf30	GTGGTTCAGGGTGAACAG	29.678–29.661	This study
mobB	TATCTGATGAGCCGTAGC	33.189–33206	This study
mobA	CGCTCTTCCATCGACTTC	34.487–34.470	This study
traJ	GCACCAGGAAGGTTTGAT	37.849–37.866	This study
	CGGGTTCTTTCAGTTGCA	38.251–38.234
traU	ATCTCACGCAATCTTACGTC	48.129–48.148	6
	TCGCGTCATGCGTGATCTTC	48.705–48.686
repA	GACATTGAGTCAGTAGAAGG	56.427–56.446	6
	CGTGCAGTTCGTCTTTCGGC	57.351–57.332

^aPositioning of primers corresponding to the two copies of IS1999.

Plasmid analysis.

The transferability of the bla_OXA-48 gene was studied by a mating-out assay. Selection was performed on agar plates supplemented with ticarcillin (32 μg/ml) and rifampin (300 μg/ml). The plasmid content of bacteria and the size of plasmids were determined using plasmid DNA extracted by the method of Kado and Liu (15) with the plasmids Rsa (39 kb), TP114 (61 kb), pCFF04 (85 kb), and pCFF14 (180 kb) as standards. Plasmid restriction analysis was performed from DNA extracted by alkaline lysis and digested with EcoRI and HindIII restriction endonucleases (Boehringer Mannheim, Meylan, France). PCR-based replicon typing (PBRT) was used to identify plasmid incompatibility groups in transconjugants (16). The repA, traU, and parA genes were detected by PCR to relate the OXA-48-encoding plasmids to the pOXA-48a IncL/M plasmid as previously described (6).

Chromosome analysis.

The chromosomal location of the bla_OXA-48 gene was investigated by pulsed-field gel electrophoresis (PFGE) using the I-CeuI endonuclease and hybridization with probes specific for the 16S rRNA genes, bla_OXA-48 gene and IS1999, as previously described (17). The chromosomal insertion of bla_OXA-48 was further mapped by PCR using bla_OXA-48-containing DNA fragments purified from pulsed-field gels and primers specific for pOXA-48a. Seventeen primer pairs (Table 2) covering pOXA-48a plasmid backbone were used for PCR mapping; the targets were trbN, pemI, mucA, nuc, orf22, orf23, orf24, korC, orf28, orf29, orf30, mobB, mobA, traJ, traU, repA, and parA.

Amplification by nested TAIL-PCR.

The sites flanking the pOXA-48 insertion site in the E. coli chromosome were amplified by a nested thermal asymmetric interlaced PCR (TAIL-PCR) approach. The primer targeting the pOXA-48 insert was ORF-25-R1 (R stands for reverse) (5′-GCCAGCGAGAAGCGAACAAAACG-3′), and the arbitrary degenerate primers were TAIL-pcr-F1 (F stands for forward) (5′-GTAATACGACTCACTATAGGGCACGCGTGGTNTCGASTWTSGWGTT-3′), TAIL-pcr-F2 (5′-GTAATACGACTCACTATAGGGCACGCGTGGTNGTCGASWGANAWGAA-3′), and TAIL-pcr-F3 (5′-GTAATACGACTCACTATAGGGCACGCGTGGTWGTGNAGWANCANAGA-3′). The amplifications comprised 15 standard PCR cycles, including 10 high-stringency cycles (annealing temperature, 62°C) and then 5 low-stringency cycles (annealing temperature, 35°C). The PCR product (1 μl) was then amplified under high-stringency conditions by nested PCR (annealing temperature, 62°C) using the primer ORF25-R2 (5′-CGCAACATCAAAACGAGCTCC-3′) and a primer targeting the adaptor part of the degenerated primers (5′-GTAATACGACTCACTATAGGGC-3′).

Plasmid and chromosome sequencing.

Plasmid DNA of Raoultella planticola strain RA35 was purified with the NucleoBond Xtra Midi kits (Macherey-Nagel) following the manufacturer's instructions. After fragmentation, DNA was sequenced with Roche 454 GS-FLX system. The genomic DNA of E. coli EC15 was extracted and purified using the Gentra Puregene Yeast/Bact kit (Qiagen, Valencia, CA) and sequenced using Illumina sequencing technology with 500-bp paired-end libraries.

Bioinformatic analyses.

The de novo assemblies were performed with MIRA package (mimicking intelligent read assembly implemented in Linux) (18). The contiguous sequences (contigs) were then aligned with the pOXA-48a plasmid sequence (JN626286) using the BLAST algorithm (19). The gaps and insertions between contigs were closed or checked by PCR and Sanger sequencing. Open reading frames (ORFs) were predicted and annotated using RATT (20). The resulting annotation was manually checked and curated using Artemis (21). The resulting DNA sequences were compared with the pOXA-48a sequence using Easyfig (22).

RESULTS

Genetic background of bacteria.

The 22 isolates included in this study produced the carbapenemase OXA-48. The isolates belonged to the E. coli (clinical strains [n = 9]; colonizing strains [n = 3]), K. pneumoniae (n = 5), Enterobacter cloacae (n = 2), Enterobacter aerogenes (n = 1), Raoultella planticola (n = 1), and Citrobacter koseri (n = 1) species. The genetic background of K. pneumoniae and E. coli producing OXA-48 was investigated by MLST (Table 1). Three K. pneumoniae isolates (KP9, KP26, and KP112) belonged to sequence type 866 (ST866). The three other K. pneumoniae isolates were assigned to ST45 (KP34), ST1156 (KP78), and ST1157 (KP87). Seven STs were represented among E. coli isolates (ST227 [n = 2], ST617 [n = 2], ST38 [n = 3], ST88 [n = 2], ST131 [n = 1], ST1711 [n = 1], and ST127 [n = 1]), and they covered all E. coli phylogroups (Table 1).

Location of bla_OXA-48 in plasmids or chromosome.

OXA-48-encoding transconjugants were obtained from 14 strains with OXA-48-encoding genes. Plasmids from natural isolates and their transconjugants were extracted and hybridized with a bla_OXA-48 probe. The results showed that the bla_OXA-48 gene was located on ∼62-kb plasmids (64% [14/22] [Table 1]). Further analysis of these plasmids from transconjugants revealed similar restriction profiles (data not shown). None of the plasmids could be assigned to an incompatibility group by the PBRT method. However, the repA, traU, and parA genes of pOXA-48a were detected in all plasmids, suggesting that the plasmids had an IncL/M pOXA-48a-like backbone.

Despite many attempts, no transconjugants or transformants were obtained from eight unrelated E. coli isolates belonging to the A, D, and B2 E. coli phylogroups. The insertion of bla_OXA-48 into the bacterial chromosome was therefore investigated by PFGE after I-CeuI nuclease treatment and hybridization. Distinct DNA fragments, with a range of 600 to 880 kb, hybridized with the bla_OXA-48 probe (data not shown), suggesting that the insertion of bla_OXA-48 in E. coli chromosomes is not rare (8/22 [36%]) and may occur at different sites.

Genetic environment of bla_OXA-48.

The bla_OXA-48 gene was identified as part of the transposons designated Tn1999 to Tn1999.4 (6, 9–11). To characterize the genetic environment of bla_OXA-48 in our strains, we mapped by PCR and sequencing the Tn1999-type transposons harboring bla_OXA-48. The bla_OXA-48 gene was predominantly associated with Tn1999.2 (68% [15/22]). However, the eight bla_OXA-48 genes on E. coli chromosomes (32% [7/22]) were located in incomplete Tn1999.2 elements, in which the IS1999 element flanking bla_OXA-48 was truncated. Finally, bla_OXA-48 was carried by Tn1999 in K. pneumoniae KP87, and we were unable to identify a Tn1999-type element in E. coli EC49.

PCR mapping covering the whole sequence of pOXA-48a was then performed by targeting 17 genes and the sequences overlapping Tn1999.2 edges. The complete backbone of pOXA-48a was observed in all bacteria, except for E. coli harboring a chromosome-mediated OXA-48 gene, in which we observed PCR products only for sequences located between bla_OXA-48 and orf25 of pOXA-48a. In addition, no amplification was obtained with the primers targeting the sequences overlapping Tn1999.2 edges for five plasmids and in all E. coli harboring a chromosome-mediated OXA-48 gene, suggesting a significant rearrangement in the vicinity of the transposon encoding OXA-48.

Sequence of pRA35 plasmid.

To investigate the rearrangement of Tn1999.2, we sequenced the ∼62-kb plasmid of R. planticola RA35(pRA35) using a next-generation sequencing approach. The sequencing reactions of pRA35 plasmid yielded 9,607 reads with an average length of 407 bases. A total of 9,303 reads with an average length of 355 bases were conserved (97%) after trimming, amounting to a total number of 3,302,032 bases of DNA. After contig assembling, gap closing, and repeat sequence mapping by Sanger sequencing, a total of 8,576 reads (92% of reads with an average length of 354 bases) totaling 3,037,308 bases were assembled to obtain the complete sequence of pRA35, resulting in an average of 47× sequence coverage (range, 38× to 121×). Overall, the complete sequencing of pRA35 showed an IncL/M circular plasmid of 63.434 kb with an average G+C content of 51.2%. The pRA35 and pOXA-48a nucleotide sequences shared overall 99% identity, and most of the synteny was conserved. Differences in the DNA sequence were observed in several genes or open reading frames; the most divergent genes were the nuclease-encoding gene and the DNA primase gene. However, the pRA35 plasmid differed from the reference plasmid pOXA-48a in three major respects (Fig. 1): (i) the insertion of an additional copy of the IS1R insertion sequence in orf25, (ii) the presence of the Tn1999.2 transposon, which differs from Tn1999 by the insertion of the IS1R insertion sequence in the IS1999 insertion sequence located upstream of bla_OXA-48, and (iii) an inverted orientation of Tn1999.2. The orientation of Tn1999.2 was investigated in all OXA-48-encoding plasmids by PCR mapping and sequencing. The inverted orientation was detected in five plasmids (5/14 [35%]).

FIG 1.

Comparison of pRA35 plasmid and bla_OXA-48 chromosome-mediated insert (Tn6237) with pOXA-48a plasmid. (A) Schematic diagram showing the rearrangements in Tn1999-type transposons encoding OXA-48. (B) Global comparison performed using the bioinformatics tools Easyfig. The transposases of insertion sequences are represented by red arrows for IS1R and black arrows for IS1999, bla_OXA-48 by gray arrows, genes implicated in transfer and mobility (mob and tra) by navy blue arrows, genes specific to pOXA-48a-like plasmids (parA, traU, and parA) by light blue arrows, and E. coli chromosome genes by green arrows.

Sequence of the chromosome insert containing the bla_OXA-48 gene.

To understand the insertion of bla_OXA-48 into the E. coli chromosome, the bla_OXA-48 environment was sequenced in E. coli strain EC15 using a next-generation sequencing approach. The sequencing reactions yielded 18,710,176 reads with an average length of 101 bases. A total of 18,441,693 reads of 93-base average lengths were conserved (98.6%) after trimming. Three contigs containing sequences corresponding to pOXA-48a were obtained by de novo assembling. After gap closing and repeat sequence mapping by Sanger sequencing, we obtained a contiguous sequence of 27.6 kb, which contained a 21.8-kb central fragment related to the pOXA-48a sequence flanked by 2.7-kb and 3.1-kb E. coli chromosome sequences. A total of 112,059 reads (0.6%) with an average length of 93 bases totaling 10,415,879 bases were aligned to 100% of this sequence, resulting in an average of 377× sequence coverage (range, 62× to 795×).

The sequence of the bla_OXA-48 chromosomal insert shared 100% identity with the pRA35 sequence and identical synteny (Fig. 1). The insert was flanked by two copies of IS1R, the first provided by the truncated and inverted Tn1999.2 transposon, and the second inserted into orf25, as observed in the pRA35 plasmid. The deletion in Tn1999.2 was due to the loss of residual IS1999 sequences upstream of Tn1999.2-located IS1R. Between the IS1R insertion sequences, after the truncated Tn1999.2 transposon, we observed 9 genes (pemI, pemK, mucA, mucB, resD, parA, parB, nuc, and korC) and 18 hypothetical ORFs, as previously observed in pRA35 and pOXA-48a. These genes and ORFs, associated with the flanking IS1R insertion sequences, formed an IS1R-based composite transposon designated Tn6237. The flanking regions of this element corresponded to the hypothetical gene orf33 belonging to the II₅₃₆ pathogenicity island, which has been observed in the reference uropathogenic strain E. coli 536 (23). Twelve-base imperfect direct repeats (GATCACTTAGAT and GATCTTTTAGAT) were observed at the site of inversion, suggesting that the insertion of Tn6237 into the E. coli chromosome is the result of a transposition event. PCR mapping of Tn6237 in E. coli showed that Tn6237 is conserved in all strains harboring a chromosome-mediated OXA-48-encoding gene. However, sequencing of nested TAIL-PCR products revealed distinct insertion sites. In the EC264 strain (ST227, phylogroup A), the bla_OXA-48 insert was detected in the yqjF gene, which encodes a putative quinol oxidase subunit. In E. coli EC8 (ST131, phylogroup B2), the bla_OXA-48 insert was detected in an intergenic region ∼375 bp at the 5′ site of the pgi gene (glucose-6-phosphate isomerase). Overall, these data showed that the chromosome-mediated bla_OXA-48 inserts exhibited similar organization in all strains and corresponded to a novel composite transposon, which can be inserted in different sites of the E. coli chromosome.

DISCUSSION

Mobile genetic elements, including large broad-host-range plasmids, transposons, and insertion sequences (IS) have played a part in the evolution and generation of diversity of bacterial populations. They performed a critical role in the horizontal and vertical transfer and rapid spread of virulence factors and drug resistance genes in Enterobacteriaceae. IncL/M-type plasmids are currently detected worldwide in Enterobacteriaceae isolates of different origin and are considered to be epidemic resistance plasmids (24). They were implicated in the large dissemination of specific genes encoding extended-spectrum β-lactamases (ESBL), in particular the bla_CTX-M-3 gene. IncL/M-type plasmids have also been reported to contribute to the diffusion of the carbapenemase-encoding genes bla_NDM-1 and bla_OXA-48.

In this study, most bla_OXA-48 genes were also located in ∼62-kb IncL/M plasmids related to pOXA-48a, as previously observed in Europe and the Mediterranean region (25). However, most plasmids harbored bla_OXA-48 inserted into the Tn1999.2 transposon and not into the Tn1999 transposon, as previously observed in pOXA-48a. In addition, 67% (8/12) of OXA-48-producing E. coli strains investigated in this study harbored a chromosome-mediated gene encoding OXA-48. The chromosome location of bla_OXA-48 is therefore not rare and is observed in unrelated E. coli isolates. This chromosomal location of bla_OXA-48 was associated with the presence of a truncated form of Tn1999.2 and significant rearrangement in the vicinity of this element.

To investigate the mechanism underlying the chromosomal location of the OXA-48-encoding gene, we sequenced plasmid pRA35, the chromosomal insert encoding bla_OXA-48 in E. coli EC15 and flanking regions of two additional chromosomal inserts. We observed a 21.9-kb bla_OXA-48-encoding and composite transposon designated Tn6237 in the E. coli chromosome and pRA35. This element was detected in all strains harboring an OXA-48-encoding chromosomal insert. It forms an IS1R-based composite transposon harboring bla_OXA-48, which is responsible for its insertion into the E. coli chromosome.

IS1 is one of the smallest bacterial insertion sequences isolated so far. It is 768 bp long, includes two ∼23-bp imperfect inverted repeats (IRL [inverted repeat left] and IRR [inverted repeat right]) located at its ends, and two partly overlapping open reading frames (insA and insB′) located in the 0 and −1 relative translational phases, respectively (26). IS1 transposition may appear as a simple insertion or as a replicative fusion event (27, 28). The origin of the second copy of IS1R observed in pRA35 and Tn6237 may therefore be the result of the replicative transposition of the IS1R insertion sequence located in Tn1999.2.

The IS1 transposition can occur in both orientations, and the direction of IS1 is not an important factor in transposition. Reports have mentioned that the gene being mobilized should be flanked by short terminal inverted repeats within the IS1 sequence regardless of the orientation of the two IS1 elements (29). The orientation of the IS1R elements in Tn6237 and pRA35 is therefore compatible with the formation of a functional composite transposon. IS1-based composite transposons (elements Tn9 and Tn1681) have already been reported (30). In Tn1681, the mobile element is flanked by two inverted repeats of IS1, which bracket the E. coli heat-stable toxin. In Tn9 and derivatives, a gene encoding chloramphenicol acetyltransferase is bracketed by two direct repeats of the IS1 insertion sequence, as observed in our bla_OXA-48-encoding transposon. In the E. coli EC15 chromosome, this IS1R-based transposable element is flanked by direct repeats of the insertion site, which may be marks of a transposition event.

Investigation of the target specificity of the IS1 element indicated that acquisition of IS1 is generally located within or directly adjacent to the A+T-rich content segments (31). The Tn6237 insertion site in the EC15 strain is the orf33 gene of the pathogenicity island (PAI) II₅₃₆, which has been previously reported in the uropathogenic Escherichia coli 536 strain (23). The orf33 gene (1.668 kb) and surrounding sequence are characterized by a very low-GC percentage (∼33%), as at the insertion sites in E. coli EC264 (∼44%) and EC8 (∼37%). Likewise, the IS1R insertion sites in the orf25 gene in pRA35 and Tn6237 are identical and characterized by a low GC percentage (40%), which makes a putative hot spot of IS1R insertion in the backbone of pOXA-48a-like IncL/M plasmids and therefore a region likely to be variable.

In this work, we provide new insights into the bla_OXA-48 genetic environment and describe its high plasticity, which is promoted by Tn1999.2 inversion and IS1R elements responsible for the transposition of the OXA-48-encoding fragment derived from pOXA-48 in E. coli chromosomes.