|Home | About | Journals | Submit | Contact Us | Français|
Eighteen carbapenem-resistant, OXA-48-positive enterobacterial isolates recovered from Turkey, Lebanon, Egypt, France, and Belgium were analyzed. In most isolates, similar 70-kb plasmids carrying the carbapenemase gene blaOXA-48 were identified. That gene was located within either transposon Tn1999 or transposon Tn1999.2, which was always inserted within the same gene. This work highlights the current plasmid-mediated dissemination of the OXA-48 carbapenemase worldwide.
Carbapenem-hydrolyzing β-lactamases belonging to Ambler classes A, B, and D have been reported worldwide among Enterobacteriaceae (22). The extensive spread of Ambler class A carbapenemases of the KPC type highlights that carbapenemases may rapidly become threatening (17). Acquired class D ß-lactamases possessing carbapenemase properties have been reported previously, being identified mainly in Acinetobacter sp. (18, 21) and occasionally in Enterobacteriaceae. The chromosome-encoded oxacillinase OXA-23 was previously described for Proteus mirabilis (4), and the oxacillinase OXA-48 was first identified in a Klebsiella pneumoniae isolate from Turkey (20). Since then, several other OXA-48-producing isolates of various enterobacterial species (Citrobacter freundii and Escherichia coli) have been reported, mainly from Turkey (1, 6, 11, 16) but also from Belgium (8), from Lebanon (15), and more recently from the United Kingdom (14, 23a), India (3a), and Argentina (6a). So far, the blaOXA-48 gene has been found to be plasmid borne and located between two identical insertion sequences, IS1999, forming the composite transposon Tn1999 (3). We have analyzed here the genetic backgrounds associated with the blaOXA-48 gene among Enterobacteriaceae isolates collected from different countries.
The study included 18 OXA-48-positive clinical isolates of K. pneumoniae (13 isolates), Enterobacter cloacae (2 isolates), Providencia rettgeri (1 isolate), C. freundii (1 isolate), and E. coli (1 isolate). Isolates were mainly from the Turkish cities Istanbul, Ankara, and Izmir (n = 14) (Table (Table1).1). Among the 13 K. pneumoniae isolates, at least Kp11978 (20) and KpB had been sources of nosocomial outbreaks (6). A single K. pneumoniae isolate (KpBEL) was recovered from Brussels, Belgium (8); another K. pneumoniae isolate (KpL) from Beirut, Lebanon (15); another K. pneumoniae isolate from the Bicêtre Hospital (KpBIC), Paris, France (this study); and another K. pneumoniae isolate from Gizah, Egypt (KpE) (8a). Samples were isolated from blood (KpI1, KpI2, KpB, and Enc1), urine (PR, KpBEL, KpL, Kp11978, and KpBIC), cerebrospinal fluid (Enc2), and catheter (KpE). Isolates from Belgium, France, Egypt, and Lebanon were from patients who did not report recent travel history.
Antibiotic susceptibility of the isolates was determined by the disk diffusion method (7). MICs of β-lactams were determined using Etest strips (AB bioMérieux, Solna, Sweden). All isolates were resistant to penicillins. Fourteen of the 18 isolates were resistant to carbapenems according to the CLSI guidelines (Table (Table1)1) (7). The four remaining isolates (KpBIC, KpE, Enc1, and Enc2) exhibited MICs of carbapenems remaining in the intermediate or in the susceptible range. Resistance to broad-spectrum cephalosporins was observed for most of the isolates. However, isolates Kp3A, Kp7A, KpBEL, KpL, and KpBIC remained susceptible to broad-spectrum cephalosporins (Table (Table1).1). All isolates were resistant to fluoroquinolones, except isolates Kp6A, Enc1, and Enc2. All isolates were resistant to aminoglycosides and sulfamethoxazole, except isolate CF, which remained susceptible to the latter antibiotic.
Carbapenemase- and extended-spectrum-β-lactamase (ESBL)-encoding genes were identified by PCR experiments using previously designed primers (6, 8), followed by sequencing. Additional ESBL production was detected by synergy tests as described previously (12). Positive results for ESBL production were observed for isolates EcA, Enc1, Enc2, Kp4A, Kp5A, Kp6A, KpI-1, KpI-2, and KpE. Several ESBL determinants were identified, including CTX-M-15, SHV-5, SHV-2a, TEM-101, TEM-150, and VEB-1 (Table (Table11).
Isolates belonging to the same species (13 K. pneumoniae isolates or two E. cloacae isolates) were compared by pulsed-field gel electrophoresis (PFGE) as described previously (6). Ten pulsotypes were identified among the 13 K. pneumoniae isolates. The two K. pneumoniae isolates from Izmir were clonally related, and the three K. pneumoniae isolates from Ankara (Kp4A, Kp5A, and Kp6A) shared very similar PFGE patterns. The two E. cloacae isolates recovered from Istanbul were not clonally related (Fig. (Fig.11).
Transferability of the blaOXA-48 gene was studied by conjugation experiments as described previously (6). When conjugation experiments failed, plasmid DNA extract was used for transformation as described previously (20). Transformants were selected on LB agar containing ticarcillin (50 μg/ml). Transconjugants and transformants with decreased susceptibility to carbapenems were obtained for all isolates (Table (Table1),1), and MICs for the transconjugants/transformants remained in the susceptible range. The E. coli transformant obtained from the P. rettgeri isolate exhibited reduced susceptibility to carbapenems associated with resistance to cefotaxime and ceftazidime.
Plasmids were analyzed by using the Kieser technique (13). A 70-kb plasmid was identified in all transconjugants/transformants (data not shown). However, a 150-kb plasmid was identified in the blaOXA-48-positive transformant obtained with the PR isolate. The blaOXA-48 and blaTEM-101 genes were codetected on the same 150-kb plasmid, as confirmed by Southern blot hybridization as described previously (20) (data not shown), explaining the resistance to all β-lactams of the PR isolate and its transformant (Table (Table1).1). Plasmid restriction profiles were compared as described previously (10) (data not shown), and very similar restriction patterns (suggesting highly related structures) were obtained for all of the 70-kb plasmids but not for the 150-kb plasmid pPR.
PCR mapping was used to assess the presence of insertion sequence IS1999 upstream of the blaOXA-48 gene, to confirm the presence of transposon Tn1999, and to identify the transposon insertion site for all of the OXA-48-positive isolates (3, 20). In all isolates, the blaOXA-48 gene was flanked by two copies of IS1999, as described previously (3). The prototype IS1999 located at the left extremity of transposon Tn1999 was identified in isolates Kp3A, Kp4A, Kp5A, Kp6A, Kp7A, CF, PR, Enc1, and Enc2. Insertion of IS1R into IS1999 as described for KpB (6) and giving rise to Tn1999.2 was identified for isolates EcA, KpBIC, KpI, KpL, KpBEL, and KpE (Fig. (Fig.2).2). In isolate Kp11978, transposon Tn1999 had been identified to be inserted into the tir gene, being functionally homologous to the F3 gene encoding the factor F involved in the plasmid replicative machinery (23). By use of a primer located upstream of Tn1999 inserted into the tir gene, insertion of Tn1999 at the same target site was evidenced in all of the blaOXA-48-positive plasmids except for the pPR plasmid (Fig. (Fig.2).2). Inverse PCR performed as described previously (3) was used for identifying the blaOXA-48-surrounding structures in isolate PR. Sequencing of the obtained amplicons indicated that Tn1999 had targeted a gene encoding a phosphoadenosine phosphosulfate reductase (ΔPPR).
Attempts to identify the incompatibility group of the 70-kb OXA-48-positive plasmids failed using a PCR-based replicon typing method as described previously (5). Since rep genes are often located close to the hot spots for resistance gene integration, cloning experiments were performed to study these plasmids further. A gene encoding phage replication protein P (RepP) was identified upstream of the blaOXA-48 gene. Primers specific for the repP gene were designed (RepPA, 5′-AATGGTTAACTTTGACTGTG-3′; RepPB, 5′-GCACGATTTAGAGGTCTAC-3′), and positive results were obtained for all 70-kb plasmids. Association of the repP gene with the blaOXA-48 gene on the 70-kb plasmid was confirmed by hybridization with a specific RepP probe (data not shown). However, the repP gene could not be detected on the 150-kb plasmid identified from isolate PR. Our study showed the spread of a blaOXA-48-carrying plasmid among different enterobacterial species, being identified first in Turkey and later in other European countries and in the Middle East. The present work indicates that dissemination of the blaOXA-48 gene is not driven by the dissemination of a single K. pneumoniae clone. Since the blaOXA-48-carrying plasmid confers by itself a low level of resistance to carbapenems, clinical laboratory detection of OXA-48-producing strains may be difficult. Since the reservoir of blaOXA-48 has been identified in the waterborne Gram-negative organism Shewanella oneidensis (19), it is likely that the process leading to the dissemination of this gene may be the consequence of a wide interspecies exchange. In addition, since plasmids belonging to the RepP group have been described among Pseudomonas sp., phytopathogenic Xanthomonas sp., and samples from soils and sludges (2, 9, 24), it may be hypothesized that the blaOXA-48 gene could also be identified in those species. This work underlines that besides class B (VIM and IMP) and class A (KPC) carbapenemases, the class D carbapenemase OXA-48 type might contribute significantly to carbapenem resistance in Enterobacteriaceae.
This work was funded by a grant from the Ministère de l'Education Nationale et de la Recherche (UPRES-EA3539), Université Paris XI, Paris, France, and mostly by a grant from the European community (LSHM-CT-2005-018705) and by the INSERM, Paris, France.
Published ahead of print on 19 January 2010.