|Home | About | Journals | Submit | Contact Us | Français|
The emergence of blaVIM-1 within four different genetic platforms from distinct Enterobacteriaceae and Pseudomonas aeruginosa isolates in an area with a low prevalence of metallo-β-lactamase producers is reported. Forty-three VIM-1-producing isolates (including 19 Enterobacter cloacae, 2 Escherichia coli, and 2 P. aeruginosa isolates, 18 Klebsiella pneumoniae isolate, and 2 Klebsiella oxytoca isolate) recovered from 2005 to 2007 and corresponding to 15 pulsed-field gel electrophoresis types were studied. The Enterobacteriaceae isolates corresponded to a hospital outbreak, and the P. aeruginosa isolates were sporadically recovered. The genetic context of the integrons carrying blaVIM-1 (arbitrarily designated types A, B, C, and D) was characterized by PCR mapping based on known Tn402 and mercury transposons and further sequencing. Among Enterobacteriaceae isolates, blaVIM-1 was part of integrons located either in an In2-Tn402 element linked to Tn21 (type A; In110-blaVIM-1-aacA4-aadA1) or in a Tn402 transposon lacking the whole tni module [type B; In113-blaVIM-1-aacA4-dhfrII (also called dfrB1)-aadA1-catB2] and the transposon was associated with an IncHI2 or IncI1 plasmid, respectively. Among P. aeruginosa isolates, blaVIM-1 was part of a new gene cassette array located in a defective Tn402 transposon carrying either tniBΔ3 and tniA (type C; blaVIM-1-aadA1) or tniC and ΔtniQ (type D; blaVIM-1-aadB), and both Tn402 variants were associated with conjugative plasmids of 30 kb. The dissemination of blaVIM-1 was associated with different genetic structures and bacterial hosts, depicting a complex emergence and evolutionary network scenario in our facility, Ramón y Cajal University Hospital, Madrid, Spain. Knowledge of the complex epidemiology of blaVIM-1 is necessary to control this emerging threat.
The emergence of acquired metallo-β-lactamases (MBLs) is of clinical concern since they confer resistance to all available β-lactams except aztreonam and they are not inhibited by class A β-lactamase inhibitors (42). MBLs have been categorized into different groups; IMP and VIM are the most commonly identified types, and they are predominant in Asia and Europe, respectively. VIM enzymes have been grouped into three main clusters designated VIM-1, VIM-2, and VIM-7 (42). Although they are more prevalent among nonfermenting gram-negative bacteria, they are increasingly recognized among members of the family Enterobacteriaceae (42). To date, VIM-2 is more widely spread among Pseudomonas aeruginosa isolates, whereas VIM-1 is normally confined to Enterobacteriaceae isolates (41, 42). MBL-producing isolates have been identified rarely in Spain, in contrast to other European countries such as Greece and Italy (28, 34, 41). Nevertheless, the prevalence of MBL producers in Spain may have been underestimated since Enterobacteriaceae and P. aeruginosa isolates with VIM-1 and VIM-2 enzymes have been identified in different Spanish cities (11, 25, 26, 32, 37).
The blaVIM allelic variants are part of class 1 integrons which are derivatives of Tn402 (also called Tn5090), a transposon characterized by the presence of a transposition module comprising a suite of four genes (tniR/tniC, tniQ, tniB, and tniA) (12, 21). Although diverse gene cassette arrays have been identified among class 1 integrons containing blaVIM alleles, a detailed characterization of the whole Tn402-harboring structure has hardly been achieved. Recently, MBL genes have been identified as part of sul1 integrons linked to a Tn402 element lacking tniR/tniC and tniQ, as reported for VIM-1- and VIM-2-producing P. aeruginosa isolates from Italy, Poland, and Australia (5, 29, 34, 35). Integrons lacking the 3′ conserved sequence (3′-CS) and associated with either a complete Tn402 transposon, as described previously for an IMP-4-producing Citrobacter youngae isolate (GenBank accession no. AF288045), or a Tn402 element containing tniC, like the integrons in P. aeruginosa VIM-2 producers from Italy, Norway, Russia, Ghana, India, Taiwan, and the United States, have also been reported previously (13, 16, 31, 36, 44). Some of these Tn402 elements were identified within different Tn3-like transposons (5, 35, 40).
The objective of this work was to characterize the genetic platforms linked to the blaVIM-1 gene from clinical isolates of different bacterial species recovered at Ramón y Cajal University Hospital over a period of 3 years (2005 to 2007). These isolates included the VIM-1-producing Enterobacteriaceae isolates involved in a recent outbreak (32) and two sporadic VIM-1-producing P. aeruginosa isolates contemporarily identified in the same hospital. In addition, these genetic platforms were compared with those already described.
Forty-three VIM-1-producing isolates (18 Klebsiella pneumoniae, 19 Enterobacter cloacae, 2 Klebsiella oxytoca, 2 Escherichia coli, and 2 P. aeruginosa isolates) recovered from 2005 to 2007 in our hospital were studied. Some of these isolates were detected at the time of the first report of VIM-1-producing Enterobacteriaceae in our hospital (32). Isolates were obtained from clinical and surveillance specimens from 38 patients (15 in intensive care units, 13 located in medical wards, 8 in surgical wards, and 2 outpatients) (Table (Table11).
Clonal relatedness was established by comparison of XbaI- and SpeI-digested genomic DNA patterns from Enterobacteriaceae and P. aeruginosa isolates, respectively, as described previously (32). The electrophoresis conditions used in the CHEF-DR III system (Bio-Rad, Hemel Hempstead, United Kingdom) were 6 V/cm2 and 14°C for 24 h with initial and final pulse times of 10 and 40 s for XbaI and 20 h with initial and final pulse times of 1 and 36 s for SpeI.
The genetic environment of blaVIM-1 was evaluated by amplification of the integrase genes and the variable regions of class 1 integrons and sequences of the Tn402 backbone (20) (Fig. (Fig.1).1). The association with mercury transposons Tn21 and Tn5051 was preliminarily inferred by screening for the presence of tnpA from Tn21 (tnpA21), tnpR21, IS1326, IS1353, merA21, tnpA from Tn5051 (tnpA5051), and tnpR5051 by PCR. The entire genetic environment of blaVIM-1 was fully identified by PCR mapping using primers based on the known Tn402 and Tn21 sequences (GenBank accession no. U67194 and AF071413.3) and further sequencing (20) (Fig. (Fig.1;1; Table Table22).
Transfer of blaVIM-1 was screened for by filter mating at a 1:10 donor-recipient ratio using E. coli strain BM21 (a plasmid-free lactose fermenter resistant to rifampin [rifampicin] and fusidic acid) and P. aeruginosa strain PAO as recipients (32). Transconjugants were selected on Luria-Bertani agar plates containing ceftazidime (4 μg/ml) plus rifampin (300 μg/ml), the plates were incubated at 37°C for 24 h, and the presence of blaVIM in the transconjugants was confirmed by PCR. Plasmid sizes in E. coli transconjugants (or wild-type strains in the absence of transfer) were determined (20). Enterobacterial plasmids were classified according to their incompatibility group by using the PCR replicon-typing scheme described by Carattoli et al. (3). Confirmation of the incompatibility group was achieved by hybridizing the S1-digested plasmid DNA with probes specific for blaVIM-1 and the plasmid replicons amplified by PCR.
The sequences corresponding to the genetic elements described in this work were assigned to the following NCBI accession numbers: GQ422826 (type A), GQ422827 (type B), GQ422828 (type C), and GQ422829 (type D).
The increase in VIM-producing isolates in our institution was due to the spread of the blaVIM-1 cassette among multidrug-resistant strains of Enterobacteriaceae, mainly strains of K. pneumoniae and P. aeruginosa (Table (Table1).1). The 18 VIM-producing K. pneumoniae isolates showed indistinguishable pulsed-field gel electrophoresis (PFGE) patterns designated KPMBL-A, and both K. oxytoca strains showed the same KOMBL-1 type. Clonal diversity was detected among the 19 E. cloacae isolates (with nine PFGE types designated ECLMBL-1 to ECLMBL-9), the 2 E. coli isolates (with PFGE types ECOMBL-1 and ECOMBL-2), and the 2 P. aeruginosa isolates (with PFGE types PAMBL-1 and PAMBL-2). Clonal outbreaks caused by K. pneumoniae strains producing VIM-1 have recently been reported in different southern European countries such as Greece, France, Italy, and Spain, although the lack of information about the relatedness among strains precludes conclusions about international clonal expansion (2, 13, 27, 30, 32, 41). Outbreaks caused by VIM producers of other enteric species such as E. coli, Proteus mirabilis, or P. aeruginosa have been documented less frequently (7, 13, 38, 39, 41). Horizontal transfer of particular genetic elements among strains seems to have greatly contributed to the recent increase in the prevalence of VIM-1 producers among European hospitals, giving rise to multiclonal epidemics. However, the full characterization of the involved mobile elements is not addressed in most studies (14, 19, 26, 27, 32, 41).
The blaVIM-1 gene was detected in four class 1 integrons arbitrarily designated types A, B, C, and D (Fig. (Fig.1)1) (32). Our analysis not only enlarges the list of MBL integrons (to include types C and D) but also confirms the wide distribution of particular variants such as In110 (type A), previously found among Pseudomonas isolates from Italy and Spain and now identified in Enterobacteriaceae, and In113 (type B), previously detected among Enterobacteriaceae isolates from another Spanish hospital (11, 17, 28, 37) (see Fig. Fig.3).3). Detailed molecular characterization of the integron genetic environment revealed a level of heterogeneity not made obvious by the analysis of gene cassette arrays as highlighted previously (18).
The blaVIM-1 gene was identified in integrons located either in an In2-Tn402 element within Tn21 (type A) or in a Tn402 transposon lacking the whole tni module (type B).
Type A (blaVIM-1-aacA4-aadA1), recovered from K. oxytoca isolates and 11 E. cloacae isolates, was identical to In110 integrons from P. aeruginosa and Pseudomonas putida isolates from Italy and Spain (11, 28) (GenBank accession no. AJ867812 and EF577408.1). The type A integron comprises a defective Tn402 tni (tni402) module (tniBΔ1 tniA), truncated by the insertion of IS1326 and IS1353, which is usually associated with the In2 family of class 1 integrons (21). The sequences corresponding to this tni402 module and the region beyond inverted repeat IRi of this In2-Tn402 element display 100% identity to the sequence of In2 found in Tn21 (GenBank accession no. AF071413). Moreover, the inverted repeat of tnp21was interrupted by a copy of IS5075 identical to that in Tn21 of pRMH760 from K. pneumoniae (GenBank accession no. AY123253.3). Sequences beyond the tni402 module were not characterized, although merA21 was identified in isolates carrying type A integrons. This study constitutes the first complete characterization of a class 1 integron containing MBL genes in Enterobacteriaceae. Until now, the location of MBL integrons within Tn3/mercury transposons had been observed only in P. aeruginosa, as is the case for integrons containing blaVIM-1, blaVIM-2, blaVIM-3, and blaIMP-13, which are linked to Tn5051, a member of the Tn501 family, or Tn1403 (5, 28, 34, 35) (Fig. (Fig.2).2). Despite the host preferences (Enterobacteriaceae versus Pseudomonadaceae) exhibited by the distribution of Tn21 and Tn501 or Tn1403 subfamily elements, it is of interest to highlight that Tn21 variants with the inverted repeat of tnp interrupted by copies of IS5075/IS4321 may be located in broad-host-range conjugative plasmids able to spread among different genera or to be integrated into chromosomes (6, 9, 19a, 22, 23). In our case, plasmids of the incompatibility group HI2 with variable sizes (300 to 435 kb) and transfer properties were identified among isolates carrying a type A integron (Table (Table22).
Type B [blaVIM-1-aacA4-dhfrII (also called dfrB1)-aadA1-catB2], recovered from one E. cloacae clone, a K. pneumoniae clone, and both E. coli clones, showed 100% identity to In113 found in the first VIM-1-producing K. pneumoniae and E. coli isolates reported in Spain (37). An intI1 gene with a hybrid P1 promoter (comprising the TGGACA −35 hexamer of the weak version of the P1/Pant promoter and the TAAACT −10 hexamer of the strong version of the P1/Pant promoter) similar to that of In1 was identified. The lack of association of type B with tni402 and Tn3-like elements was not surprising since MBL integrons in Enterobacteriaceae may be flanked by particular insertion sequences such as IS26 or IS1, creating a composite element able to recombine among plasmids of different incompatibility groups as demonstrated in Greece (19, 27). In our case, a common IncI1 plasmid of approximately 60 kb was detected in all transconjugants carrying the type B integron.
We detected two new integrons linked to defective Tn402 variants containing either tniBΔ3 and tniA (type C) or tniC and ΔtniQ (type D). Both variants had a typical 5′-CS containing an intI1 gene with a weak P1 promoter followed by an inactive P2 promoter (lacking the GGG insertion prior to the −10 hexamer) and an attI1 site. They were both located on 30-kb plasmids which were not amplified with any of the primers included in the scheme used.
Type C (blaVIM-1-aadA1) was part of a Tn402 transposon with a deleted tni402 module (tniBΔ3 tniA) which is identical to the modules of In70 and In70.2 (5, 28, 29). We could not detect the presence of ISpA7or Tn5051-like transposons, often associated with MBL integrons in Pseudomonas (28, 29, 35). However, the location of the type C integron in other elements such as Tn1403 transposons (GenBank accession no. AF313470.2) currently spread in Italy, Tunisia (GenBank accession no. FM897214), Taiwan, and Australia cannot be discarded (5, 28, 40). Interestingly, most of these transposons are chromosomally located. Clonal expansion may explain chromosomal predominance in some instances (10, 11). Nevertheless, a plasmid origin for these Tn402 derivatives, the elements of which could have been transposed into the res sites of different transposon backbones in the P. aeruginosa chromosome, should be considered.
The type D integron (blaVIM-1-aadB) lacks the 3′-CS region (including qacEΔ1, sul1, and orf5), tniB, and tniA. The incomplete tni module comprises a region displaying 100% identity to the tniR/tniC gene and a sequence of 161 bp showing 86% similarity to tniQ of Tn402 from IncP-1β plasmid R751 (GenBank accession no. U67194). Although integrons lacking the 3′-CS and containing tniC are spread among P. aeruginosa VIM-2 producers from the United States, Italy, Norway, Russia, India, Taiwan, and Ghana, this study constitutes the first documentation involving blaVIM-1 (13, 16, 31, 36, 44) (Fig. (Fig.22).
The aadA1 gene cassettes present in type A, B, and C integrons showed differences from previously described aadA1 gene cassettes. The variant present in types A and C had a shorter attC site than previously described cassettes, containing the heptanucleotide boxes called R′ and R" but lacking most of the central region, including parts of the recombination core sites known as L′ and L" (24). This variant constitutes a new alternative form of the aadA1 gene cassette which was previously observed in the widespread In70/In70.2 integrons (22, 29). The identity and synteny of the gene cassettes present in In70 and In110, as well as the unusual hybrid configuration of their intI1 promoters (with the TTGACA −35 hexamer of the strong version of the P1/Pant promoter and the TAAGCT −10 hexamer of the weak P1/Pant promoter), suggest a common origin for gene cassette arrays of In70 and In110. However, differences in the Tn402 variants linked to In70 (tniBΔ3 tniA) and In110 (tniBΔ1 tniA) may reflect recombination of these integrons with different platforms (Fig. (Fig.2).2). Interestingly, the attC site of blaVIM-1 located in integron type C showed a deletion identical to that observed in the attC sites of the aadA1 cassettes in types A and C and may represent a particular recombination crossover event in this area.
We describe the emergence and dispersal of blaVIM-1 in an allodemic scenario (1) in which both clonal expansion and the transfer of conjugative plasmids carrying different integrons/transposons among distinct clones of Enterobacteriaceae and P. aeruginosa have contributed to increasing the presence of this MBL gene. At first glance, the complex epidemiology of blaVIM-1 in our institution would suggest independent events of selection in the hospital environment. However, recombination between different blaVIM-1-carrying platforms with unusual gene cassettes might also have occurred. The capture of blaVIM or blaIMP mostly by integrons containing aacA4 and aadA1 cassettes, which are overrepresented in environmental plasmids of Enterobacteriaceae or Pseudomonas, may have influenced the recent and fast dissemination of blaVIM-1 (6, 15, 19, 19a, 33, 43) (Fig. (Fig.3).3). Although blaVIM-2 and blaVIM-1 have been predominant among P. aeruginosa and Enterobacteriaceae strains, respectively, our results show that both genes have the potential to be located in different platforms suitable for successful dissemination and eventually favored by recombination events. Moreover, the diversity of platforms associated with blaVIM-1 may facilitate the rapid spread of this allele among P. aeruginosa isolates (7, 8, 10, 13, 28, 39).
To date, the epidemiological studies of MBL producers have been focused on the description of gene cassette arrays containing MBL genes. Further studies addressing the full characterization of MBL genetic elements are needed for the better understanding of their spread, as this and other studies reveal (5, 18). The roles of selective pressures for clonal expansion of certain strains, genetic platform rearrangement, and/or possible plasmid dissemination among different clones need to be clarified in order to establish infection control measures for this emerging threat.
We thank Ângela Novais for helping with the sequence annotation and for critical reviewing of the manuscript.
M.T. is supported by the Fondo de Investigaciones Sanitarias, belonging to the Instituto de Salud Carlos III, Spanish Ministry of Science and Innovation (reference no. CM06/00044, Rio Hortega program). This study was funded by research grants from the European Commission (LSHM-CT-2009-223031 and KBBE-2008-2B-227258) and the CIBERESP Network for Biomedical Research in Epidemiology and Public Health (Instituto Carlos III, Spanish Ministry of Science and Innovation; reference no. CB06/02/0053).
Published ahead of print on 9 November 2009.