First description of a TEM-30 (IRT-2)-producing Klebsiella oxytoca isolate.

Abstract

Both Klebsiella oxytoca and Klebsiella pneumoniae synthesize a chromosomal class A β-lactamase (KOXY and LEN-1, respectively), but K. oxytoca harbors acquired plasmid-mediated β-lactamases much less often (10%) than K. pneumoniae (90%) (8). Increased and extended resistance to β-lactams in K. oxytoca (10 to 20% of the clinical isolates) is generally imputed to the overproduction of KOXY β-lactamase due to point mutations in the promoter region (3, 8). However, some K. oxytoca isolates producing plasmid-mediated extended-spectrum β-lactamases have been described (5) and, more recently, two isolates which displayed a β-lactamase inhibitor-resistant (IR) phenotype were described (1, 9). The first isolate produced an OXY-2 β-lactamase with the amino acid substitution Ser130Gly responsible for the IR phenotype (9). The second isolate harbored a plasmid-mediated TEM-2 derivative (TEM-59, IRT-17) which also showed the amino acid substitution Ser130Gly (1). We report here on a K. oxytoca isolate (SG 81) which also displayed an IR phenotype on the basis of the disk diffusion method: highly resistant to amoxicillin, even in the presence of clavulanate; relatively resistant to ticarcillin; and fully susceptible to cephalosporins. As no associated resistance was observed, the KOXY-encoding gene was studied first, with gene amplification and sequencing using the primers OXY-A (5′-TCGGTAACTGTGACGGGA-3′) and OXY-B (5′-CCGAATTTCGGGAAGCCA-3′). This gene was affiliated to blaOXY-2 and differed from the blaOXY-2 gene of K. oxytoca strain SL911 by two mutations, one of them leading to the amino acid substitution Val260Ile (4). As this substitution had never been involved in any IR phenotype, the plasmidic content of this strain was then studied. Conjugation between K. oxytoca SG 81 and Escherichia coli K-12 C600S1000 (lac negative; streptomycinr) and thereafter between E. coli transconjugants and K. oxytoca reference strain SL911 (OXY-2 β-lactamase) were carried out by a broth mating procedure. E. coli transconjugants were selected at a frequency of 4 × 10−6. MIC determinations for donors, recipients, and transconjugants were performed by the agar dilution method using fixed concentrations of β-lactamase inhibitors and variable concentrations of antibiotics, and vice versa. The two types of transconjugants displayed the same IR phenotype as K. oxytoca SG 81, suggesting that the genetic support of the IR phenotype was a conjugative plasmid (Table ​(Table1).1). TABLE 1. MICs of β-lactam antibiotics and β-lactamase inhibitors against K. oxytoca strain SG 81, recipient strains, and transconjugants This plasmid was extracted and its size was estimated to be 40 kb. In order to further characterize this plasmid which displayed no other markers of antibiotic resistance, mercury resistance was tested as recommended by Liebert et al. (7), but the isolate was found susceptible to this compound. The incompatibility group of this plasmid was also studied as previously described (2), and we found that our plasmid was compatible with the T, 6-C, 7-M, FII, N, I1, X, M, 10-B-O, I2, FIV, J, and W groups. As IR phenotype is essentially due to TEM-type enzymes, a blaTEM-specific PCR was performed on a crude extract of DNA from K. oxytoca SG 81 as previously described (10). The sequence analysis of the amplified product showed that the blaTEM gene displayed two mutations in comparison toblaTEM-1A, namely, the silent mutation T → C at position 436 and the mutation A → C at position 929, leading to the amino acid substitution Arg244Ser according to Ambler numbering. It was thus demonstrated that the blaTEM gene derived from blaTEM-1C had a weak promoter P3 and encoded the TEM-30 (IRT-2) β-lactamase. The previously published TEM-30 in E. coli isolates, derived from blaTEM-1A, -1B, -1D, and -1F but not from blaTEM-1C, had, in the majority of cases, a strong promoter, Pa/Pb or P4 (6). To our knowledge, this is the first description of TEM-30 in K. oxytoca.