Association of the Extended-Spectrum β-Lactamase Gene bla_TLA-1 with a Novel ISCR Element, ISCR20

Abstract

The bla_TLA-1 gene encoding an extended-spectrum β-lactamase was identified in 11 enterobacterial isolates from Mexico City, Mexico. This gene was located on different plasmids and plasmid types with different sizes and incompatibility groups. It was associated with a novel insertion sequence, ISCR20, encoding a putative transposase that shared only 20% amino acid identity with the most closely related transposase of ISCR1. The ISCR20 element provided specific promoter sequences for expression of the bla_TLA-1 gene.

Although the majority of Ambler class A extended-spectrum β-lactamases (ESBLs) identified in the family Enterobacteriaceae worldwide are mostly of the TEM, SHV, and CTX-M types, some “minor” ESBLs have been reported, such as VEB-like, GES-like, BES-1, SFO-1, and TLA-1 enzymes (5). The bla_TLA-1 gene was first reported in 2000 from an Escherichia coli clinical isolate from Mexico (8). From April 2000 until February 2002, the bla_TLA-1 gene was detected from an epidemic Klebsiella pneumoniae clone in Mexico, and both the bla_TLA-1 and bla_SHV-5 ESBL genes were located on the same plasmid (1). The bla_TLA-1 gene was identified on a 150-kb conjugative plasmid named pRZA92, but its genetic context had not been further characterized (8). Considering that many ESBL genes may be transposon or integron associated, the aim of this study was to characterize the genetic elements at the possible origin of acquisition of the bla_TLA-1 gene from isolates of members of the family Enterobacteriaceae recovered from several hospitals in Mexico City, Mexico.

The 150-kb plasmid pRZA92 of E. coli R170 (8) was extracted by using the Qiafilter plasmid purification maxikit (Qiagen, Courtaboeuf, France) and partially sequenced using the primers indicated in Table 1. A sequence of 5,923 bp was identified through a primer-walking approach and contained several open reading frames (ORFs) (Fig. 1). Upstream of the bla_TLA-1 gene, another gene (named orf432) was identified; this gene encoded a 432-amino-acid protein sharing 86% sequence identity with an uncharacterized orf430 gene (83% amino acid identity) located in a class 1 integron from a Proteus mirabilis isolate recovered from Mediterranean herring gulls in Italy (GenBank accession number DQ520941) (3). The orf432 gene encoded a putative transposase belonging to the IS91 family of transposases that includes the newly described ISCR elements. The Orf432 transposase shared 20%, 21%, and 18% amino acid identity with the transposases encoded by the ISCR1, ISCR2, and ISCR3 elements, respectively. Detailed analysis allowed us to identify the orf432 gene as part of a novel ISCR element, termed ISCR20 according to the official nomenclature (http://medicine.cf.ac.uk/en/research/research-groups/i3/research/antibacterial-agents/iscr-elements/). The ISCR20 element is 1,705 bp long and possesses a G+C content of 47.1% (Fig. 1). A putative oriIS sequence was identified 255 bp downstream of the stop codon of the transposase gene and shared 10 identical bp (5′ACTGATAGGAACTGTCATTTC3′ [the identical bases are shown in boldface type and underlined]) with the 19-bp consensus sequence reported for the other ISCRs (5′XXGTATAGGAAGTTCAAACGC3′) (10) (Fig. 1). A putative terIS sequence forming a hairpin structure (5′GGACCCGCACGCAGGGTGTT3′ [the boldfaced and underlined nucleotides indicate the complementary nucleo-tides forming the hairpin structure]) was also evidenced 136 bp upstream of the transposase start codon. The transposase of ISCR20 shared an overall significant degree of identity with members of the IS91 family transposases especially in the five functional domains previously described (Fig. 2). A tyrosine residue at position 218, which has been shown to be critical for the transposase activity of IS91 and speculated to be also critical for those of the ISCR elements, was present in motif IV of the ISCR20 transposase.

TABLE 1.

Sequences of primers used for detection of the bla_TLA-1 gene and its circular form with ISCR20 and for mapping the genetic environment

Primer name	Primer sequence	Positiona
TLA-1F	TGTGTGCTTTTTGCTTCTGC	1
TLA-1R	GCTTCCGGTTTTATGAGCAA	2
TLA-1-5′-TR1	TACTGCTTTTAAGCGAATCCG	3
TLA-1-5′-TR2	TTAGCCGCAAAAGCAGAAGC	4
TLA-1-5′-TR3	AAGCAAAAAGCACACAAAATGC	5
IS91-5′-TR1	ACAACCGACATTGACCGTGG	6
IS91-5′-TR4	CTAAAACATGCCAGTGCGGG	7
IS91-5′-TR5	AATGCTCTAGGCGGTCAAGG	8
IS91-5′-TR6	GACATGGCTTTTTGCTGGCG	9
IS91-5′-TR7	GATATTGCTTCGATGCAGCAC	10
TLA1-3′-TR1	CATTGCTGTTTATGTGTCGG	11
TLA1-3′-TR4	CATCGCGTCATTAATCCTGC	12
TLA1-3′-TR5	GATCGAACAGATTTCCGCCT	13
TLA1-3′-TR6	ACGAACAACAGAAAAGCGCC	14
IS91-3′extR	CTTTTGTAACAACCTTGCCC	15
TLA1-3′extF	GTATTATTACTCTGCCGAACGG	16
IS91-5′extR	TCTGAAGTACGAATGGCCAC	17
IS91-3′extF	ACCAAGATTATTGACTGGGC	18
ICR20F	CAAAGCTGATCGTGAAGCCC	19
Maturase F	AACATCATGCTCGACCCAC	20
Methyl R	CATCGAATTGAGCGAGTCG	21

^aPositions of the primers indicated in Fig. 1. All the primers were designed in this study.

FIG. 1.

Schematic representation of the sequences surrounding the bla_TLA-1 gene in E. coli X170. The positions of primers shown in Table 1 are indicated by arrows; primers 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, and 21 are primers TLA-1F, TLA-1R, TLA-1-5′-TR1, TLA-1-5′-TR2, TLA-1-5′-TR3, IS91-5′-TR1, IS91-5′-TR4, IS91-5′-TR5, IS91-5′-TR6, IS91-5′-TR7, TLA-1-3′-TR1, TLA-1-3′-TR4, TLA-1-3′-TR5, TLA-1-3′-TR6, IS91-3′-extR, TLA-1-3′extF, IS91-5′-extR, IS91-3′extF, ISCR20F, maturase F, and methyl R, respectively. Genes are shown by large arrows, and the transcription orientations of the genes are indicated. The putative origin of replication, oriIS, and termination of transposition, terIS, of the ISCR20 element are indicated by a solid black circle and a black hairpin, respectively.

FIG. 2.

Comparison of the sequence motifs of ISCR20 and those of the IS91 family transposases. The sequences (with their GenBank accession numbers shown in parentheses) shown are as follows: ISCR1 (FJ187822), ISCR2 (AY055428), ISCR3 group, including ISCR3, ISCR4, ISCR6, ISCR14, and ISCR16 (FJ183463), ISCR7 (AJ250371), ISCR8 (AF028594), ISCR15 (AM998375), and ISCR19 (EU503121). The five motifs found within IS91 group elements are indicated. The amino acid residues conserved in all IS91 group elements are shown boxed. Gaps introduced to maximize alignment are indicated by dashes.

The bla_TLA-1 gene was followed by part of a gene encoding a methyl-accepting chemotaxis protein that shares 58% amino acid identity with that found in Desulfovibrio magneticus RS-1 (GenBank accession number YP_002952376). A group II intron, containing an orf458 gene, was identified at the left extremity of the ISCR20 element (Fig. 1). The orf458 gene encoded a putative reverse transcriptase and maturase sharing 69% amino acid identity with those of Vibrio cholerae (GenBank accession number ABV21790). Analysis of the nucleotide sequence surrounding the bla_TLA-1 gene did not yield other insertion sequences that could be part of a transposon or any class 1 integron.

A collection of 11 TLA-1-producing Enterobacteriaceae clinical strains (3 E. coli strains [including E. coli R170], 1 K. pneumoniae strain, and 7 Enterobacter cloacae strains) isolated from different samples (urine, blood, and tracheal fluid samples) from 11 patients hospitalized over a 10-year period since 1991 from several hospitals in Mexico City, Mexico, were retrospectively analyzed. These isolates were selected among clinical ESBL-producing enterobacterial isolates on the basis of two criteria: their ESBL profile and their production of a β-lactamase with an isoelectric point of 9.0. By DNA restriction with XbaI endonuclease, followed by pulsed-field gel electrophoresis (PFGE) analysis, we showed that most of the isolates were not clonally related, with four distinct E. cloacae PFGE types and three distinct E. coli PFGE types, suggesting a diffusion of bla_TLA-1-positive plasmids. The bla_TLA-1 gene was transferred from seven of those distinct PFGE types to E. coli J53 by conjugation performed as described previously (6). Together with the expression of an ESBL phenotype, the E. coli transconjugants were resistant to trimethoprim, sulfamethoxazole, and tetracycline, and some of these transconjugants were resistant to amikacin (n = 5) or gentamicin (n = 1). Plasmid analysis performed by the Kieser technique (4) allowed us to visualize several plasmids in all the clinical isolates but only a single plasmid from each E. coli transconjugant. Southern blot hybridization with a bla_TLA-1-specific probe indicated that this ESBL gene was located on a single plasmid, being of either ca. 110 or 150 kb. PCR-based replicon typing (PBRT) analysis performed on the E. coli transconjugants (2) showed that the plasmids carrying the bla_TLA-1 gene belonged to the IncA/C, IncL/M, or IncN incompatibility group, corresponding to the 150-, 150-, and 110-kb plasmids, respectively, highlighting that three different plasmid types were at the origin of the diffusion of the bla_TLA-1 gene.

Except for E. coli isolate R2915 and its transconjugant, PCR amplification identified in all cases the same ISCR20 element upstream of the bla_TLA-1 gene. It has been suggested that the promoter of the bla_TLA-1 gene was located 71 bp from the translational start site by researchers using a computer-based promoter analysis (8). However, by using a 5′ rapid amplification of cDNA end PCR system (Invitrogen, Cergy-Pontoise, France) and primers TLA-1-5′-TR1 to TLA-1-5′-TR3 (Table 1), the initiation site of transcription of the bla_TLA-1 gene was mapped from E. coli R170. Promoter sequences were located 144 bp upstream of the translational start codon, corresponding to the −35 promoter sequence TTGACA separated by the 17-bp distance from the −10 promoter sequence TTAAAG. This result showed that the bla_TLA-1 gene expression was driven by the ISCR20 element. One pair of outward primers chosen at each end of the ISCR20 insertion sequence was used to detect the circular form of this novel ISCR associated with the bla_TLA-1 gene. The amplifications performed using strain R170 and its transconjugant as templates failed.

Nineteen ISCR elements with weak structural relationship (18 to 96%) have been found to be associated with different families of antibiotic resistance genes. As demonstrated with the transposition model of the IS1294 elements (9), it is proposed that a transposase encoded by the ISCR has the ability to cotranspose DNA adjacent to its terminal terIS sequence through a rolling-circle (RC) transposition mechanism (10). It has been suggested that the mobilization by ISCR elements of adjacent DNA sequences likely involving sequences located only at their left extremities and the frequent observation of resistance genes bracketed by two copies of ISCR may be the consequence of secondary recombination events (10). This has been recently shown with the identification of the bla_VEB-1a ESBL gene flanked by two copies of ISCR2 in Acinetobacter baumannii (7). Here, no ISCR20 element was identified downstream of the bla_TLA-1 gene, casting doubt on the involvement of that element in the mobilization process of the bla_TLA-1 gene. Further work is needed to understand the process that gave rise to the mobilization of the bla_TLA-1 gene.

Besides describing a novel ISCR element, our study indicates its involvement in the expression of the bla_TLA-1 gene and its likely role in its acquisition. In addition, we showed here that the dissemination of the bla_TLA-1 gene may have already occurred in Mexico and that it was related to different plasmid backbones. Further screening of TLA-1-positive isolates should be performed at least in neighboring countries, such as countries in Central America and in the southern United States. We showed a heterogeneity of plasmid backgrounds, supporting the hypothesis that the bla_TLA-1 gene was very likely harbored by a mobile genetic structure.

Nucleotide sequence accession number.

The nucleotide sequence reported in this study has been deposited in the GenBank database under accession number GU441460.