Dissemination of CTX-M-Type Extended-Spectrum β-Lactamase Genes to Unusual Hosts

CTX-M-type extended-spectrum β-lactamases (ESBLs) have recently acquired a major role as emerging resistance determinants to expanded-spectrum cephalosporins in Enterobacteriaceae (3). A worldwide distribution of these enzymes has now been reported, with several variants (3). In some epidemiological settings the prevalence of CTX-M-type enzymes can be even higher than that of the TEM- or SHV-type ESBL variants (3, 22).

Escherichia coli was the first species in which CTX-M-type enzymes were identified as acquired ESBLs (2), and it remains the principal host of these enzymes (3, 5). Dissemination of CTX-M-type enzymes in Klebsiella pneumoniae and Salmonella enterica has increasingly been reported (5, 9, 10, 12, 16), whereas the prevalence of these enzymes in other enterobacterial species remains apparently low (1, 3, 8).

In the present study we report the detection of CTX-M-1-producing clinical isolates of Citrobacter amalonaticus and Morganella morganii from an area of northern Italy where CTX-M producers were found to be widespread in E. coli (6), and we show that, in both cases, the ESBL determinant was possibly acquired by these unusual hosts in vivo, after a coinfection with E. coli strains carrying conjugative plasmids bearing the bla_CTX-M-1 gene.

Screening for ESBL producers among clinical isolates was routinely carried out by the Phoenix System (Becton Dickinson, Milan, Italy). In selected clinical isolates, E-test ESBL strips (AB Biodisk, Solna, Sweden) were used to confirm ESBL production. The E-test (AB Biodisk) was used to determine MICs. Identification of the C. amalonaticus, M. morganii, and E. coli isolates was confirmed by using the API 20E identification system (bioMérieux, Marcy l'Etoile, France). The presence of CTX-M determinants was investigated by a PCR assay using primers CTX-MU1 (5′-ATGTGCAGYACCAGTAARGT) and CTX-MU2 (5′-TGGGTRAARTARGTSACCAGA), designed on conserved regions of bla_CTX-M genes, as described previously (13). PCR amplification of bla_CTX-M-1 group alleles was carried out with primers CTX-M3G-F (5′-GTTACAATGTGTGAGAAGCAG) and CTX-M3EG-R (5′-AACGGAATGAGTTTCCCCATT) as described previously (13). The presence of TEM- and SHV-type genes was investigated by PCR as described previously (17). Nucleotide sequences were determined for both strands of PCR products, as described previously (19). Plasmid DNA was extracted by using the alkaline lysis method (20). Conjugation experiments were performed on solid medium with the E. coli strain MKD-135 (argH rpoB18 rpoB19 recA rpsL) as recipient, as described previously (15). Transconjugants were selected on Mueller-Hinton agar (Difco Laboratories, Detroit, Mich.) containing rifampin (250 μg/ml) plus cefotaxime (2 μg/ml). Southern hybridization was performed as described previously (13) with, as a probe, a bla_CTX-M-1-containing a PCR amplicon generated with primers CTX-M3G-F and CTX-M3EG-R (see above) and labeled with ³²P by the random-priming technique (20). Analytical isoelectric focusing for detection of β-lactamases was carried out as described previously (14).

C. amalonaticus (isolate VA-1340/03) was isolated in October 2003 from the urine of an outpatient who had received a renal allograft 10 months earlier. The patient had been subjected to antimicrobial prophylaxis with amoxicillin-clavulanate at the time of transplantation but did not report any subsequent antimicrobial treatment. M. morganii (isolate VA-1342/03) was isolated in October 2003 from the urine of an inpatient at a long-term care facility, where the patient had been admitted 2 years previously. The patient was suffering from catatonic schizophrenia with other comorbidities (chronic obstructive bronchopneumopathy, hypertension, and decubitus ulcers) and had a history of recurrent urinary tract infections. In July 2003 the patient had received a course of ciprofloxacin (500 mg twice daily for 7 days) for a urinary tract infection and, in September 2003, a course of ceftriaxone (1 g daily for 7 days) for pneumonia. Both isolates were identified as ESBL producers by the BD Phoenix Expert System. Susceptibility testing showed cefotaxime and ceftriaxone MICs of ≥16 μg/ml, while ceftazidime, cefepime, and aztreonam MICs were consistently lower (Table 1). E-test ESBL results confirmed ESBL production since cefotaxime MICs were significantly decreased (>8-fold) in the presence of clavulanate (Table 1). A >8-fold MIC decrease in the presence of clavulanate was also observed with cefepime and, for C. amalonaticus, also with ceftazidime (Table 1). Interestingly, E. coli showing an overall similar resistance phenotype (isolates VA-1339/03 and VA-1341/03, respectively) (Table 1) were isolated from the same urine samples, yielding C. amalonaticus VA-1340/03 and M. morganii VA-1342/03. The infection caused by M. morganii and E. coli was successfully treated with gentamicin. No data were available on the treatment of the other infection.

PCR analysis detected, in all four isolates, the presence of bla_CTX-M genes, which sequencing identified as bla_CTX-M-1. No bla_SHV genes were detected in any of the isolates, but a bla_TEM gene was detected in E. coli VA-1339/103. Analytical isoelectric focusing detected a pI 8.4 β-lactamase (likely CTX-M-1) in all four isolates, plus a pI 5.4 enzyme in VA-1339/03 (likely TEM-1) and a pI 7.3 enzyme in VA-1340/03 (likely the chromosomal β-lactamase of C. amalonaticus) (21).

Cefotaxime resistance could be transferred by conjugation to E. coli MKD-135 from C. amalonaticus VA-1340/03, E. coli VA-1339/03, M. morganii VA-1342/03, or E. coli VA-1341/03 (with transfer frequencies of 7.3 × 10⁻², 1.3 × 10⁻¹, 9.6 × 10⁻³, and 4.6 × 10⁻² transconjugants/recipient, respectively). Analysis of the transconjugants obtained from C. amalonaticus VA-1340/03 and E. coli VA-1339/03 revealed that they carried an apparently identical plasmid, and the same was true for transconjugants obtained from M. morganii VA-1342/03 and E. coli VA-1341/03 (Fig. 1A). However, the plasmids present in the two couples of transconjugants were notably different from each other (Fig. 1A). The β-lactam resistance phenotype of transconjugants was consistent with CTX-M-1 production (Table 1). Their plasmids were confirmed to carry a bla_CTX-M gene by Southern hybridization (Fig. 1B), which was identified as bla_CTX-M-1 by PCR and sequencing.

Although other CTX-M-type enzymes have been previously reported in C. amalonaticus and in M. morganii (1, 4, 18), to our best knowledge this is the first report of CTX-M-1 in isolates of those species. This finding underscores the ability of CTX-M-type ESBL genes to spread among different species of Enterobacteriaceae.

CTX-M-type enzymes, mostly CTX-M-1, were found to be increasingly prevalent during the past few years in E. coli isolated at the Clinical Microbiology Laboratory of the Varese University Hospital (a 800-bed teaching hospital located in northern Italy and serving a population of ca. 250,000) from either inpatients or outpatients (6). C. amalonaticus VA-1340/03, and M. morganii VA-1342/03 were the first isolates of these species producing a CTX-M-type ESBL detected in this laboratory. In both cases, coinfection with E. coli producing CTX-M-1, encoded by conjugative plasmids apparently identical to those found in C. amalonaticus and M. morganii, respectively, would suggest the possibility that transmission of the ESBL determinant could have occurred in vivo. However, transfer could have also occurred prior to coinfection, possibly in the environment. Anyway, this finding underlines the potential role of E. coli strains producing CTX-M-type enzymes, once they have spread in the clinical setting, as a source of similar ESBL determinants for other pathogenic bacteria.

Present results also underscore the need for clinical microbiologists to be aware that these types of organisms can harbor these enzymes. In our case, C. amalonaticus VA-1340/03 and M. morganii VA-1342/03 were correctly detected as ESBL producers by the BD Phoenix expert system, although they would have not been classified as suspect for ESBL production according to Clinical and Laboratory Standards Institute (formerly NCCLS) guidelines (7). This may result in underestimation of the prevalence of ESBL producers in species other than E. coli, klebsiellae, and Proteus mirabilis in surveillance studies, as well as in an impaired predictivity of in vitro susceptibility testing for similar infections. The use of modified cephalosporin breakpoints for all Enterobacteriaceae, as recommended by other agencies (11), might be helpful in reducing this bias at least in areas where a consistent prevalence of ESBL production among Enterobacteriaceae is currently recognized or suspected.

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Plasmid profiles of the MKD-135 transconjugants obtained from C. amalonaticus VA-1340/03, E. coli VA-1339, M. morganii VA-1342, and E. coli VA-1341, either undigested or after digestion with the restriction endonuclease PstI. DNA size standards (in kilobases) are shown in lanes M1 and M2. (B) Results of Southern hybridization of the gel shown in panel A with the bla_CTX-M-1 probe.

• American Society for Microbiology