Emerging Extended-Spectrum β-Lactamases in Proteus mirabilis

β-Lactam susceptibility and β-lactamase production in P. mirabilis isolates.

During a 1-year period (February 1997 to January 1998), 282 consecutive nonduplicate P. mirabilis isolates were collected from inpatients at the S. Matteo I.R.C.C.S. Hospital of Pavia (northern Italy). Most isolates (65%) were from urine or other urogenital specimens, while a minority were from respiratory specimens (11%), discharge of infected wounds/ulcers (10%), pus (6.5%), blood (3%), or other sources (4.5%). All isolates were identified using the GNI card of the Vitek system (BioMérieux, Rome, Italy).

Screening by Vitek GNS cards (BioMérieux) revealed that 135 (48%) of the 282 isolates were susceptible to ampicillin and other β-lactams, while the remaining 147 were intermediate or resistant to ampicillin and produced β-lactamase activity detectable by the nitrocefin hydrolysis test (Oxoid-Unipath, Milan, Italy). Susceptibility testing by disk diffusion (15) showed that (i) all the β-lactamase producers were susceptible to cefoxitin, suggesting that production of AmpC-like enzymes was not involved; and (ii) none of them were resistant to amoxicillin-clavulanate but susceptible to cephalothin, suggesting that production of inhibitor-resistant TEMs or oxacillin-type enzymes was also not involved (5). Production of ESBL activity was screened by a double-disk synergy test as described previously (8), placing the disks at a distance of 23 mm (center to center) from each other. A potentiation of the inhibitory zones of cefotaxime, ceftazidime, cefepime, and/or aztreonam by clavulanate, suggesting the production of ESBL activity, was observed with 70 (48%) of the β-lactamase producers.

Susceptibility testing by a broth macrodilution procedure (14) revealed that the putative ESBL producers exhibited variable susceptibility patterns to cefotaxime (MIC range, 4 to >64 μg/ml), ceftazidime (MIC range, 2 to 32 μg/ml), cefepime (MIC range, 2 to 64 μg/ml), and aztreonam (MIC range, 2 to 32 μg/ml), while the ESBL-negative β-lactamase-producing isolates were always susceptible to the above drugs; for those isolates, MICs were ≤1 μg/ml.

Characterization of ESBLs.

Analytical isoelectric focusing (IEF) of crude extracts of the 70 putative ESBL producers and detection of β-lactamase bands by nitrocefin were carried out as described previously (18). The activity against cefotaxime, ceftazidime, cefepime, and aztreonam of the β-lactamase bands separated by IEF was assayed by a substrate overlaying procedure as reported previously (18), using a final substrate concentration of 1 μg/ml in the medium overlay and Escherichia coli ATCC 25922 as an indicator strain. Substrate hydrolysis was indicated by the occurrence of bacterial growth above the enzyme bands. IEF detected two β-lactamases of pIs 5.2 and 5.6, respectively, in 52 isolates (in 7 of them the pI 5.2 enzyme appeared to be consistently less abundant than in the other 45) and a single β-lactamase of pI 5.9 in the remaining 18 isolates. In the bioassay, the pI 5.2 and 5.9 enzymes were active on cefotaxime, ceftazidime, cefepime, and aztreonam, while the pI 5.6 enzyme did not show activity with any of the above compounds.

In the 70 ESBL producers the presence of β-lactamase genes was investigated by colony blot hybridization (26). The probes were PCR-generated amplicons containing the entire bla_TEM-1 (27) or bla_PER-1 (16) coding sequence, labeled with ³²P using a commercial kit (Rediprime II; Amersham Pharmacia). Results of these experiments showed that all 70 isolates were recognized by the bla_TEM probe, while 52 (corresponding to those producing a pI 5.2 enzyme) were also recognized by the bla_PER probe (Table 1).

PCR amplification of bla_TEM alleles was carried out with primers TEM/f (5′-GGAAGAGTATGAGTATTCAACAT) and TEM/r (5′-ATATGAGTAAGCTTGGTCTGACAG) and the following parameters: initial denaturation at 94°C for 7 min, denaturation at 94°C for 45 s, annealing at 52°C for 45 s, and elongation at 72°C for 60 s, repeated for 35 cycles; and a final extension at 72°C for 7 min. PCR amplification of bla_PER alleles was carried out with primers BLAPER-f (5′-GGGACA(A/G)TC(G/C)(G/T)ATGAATGTCA) and BLAPER-r (5′-GGG(C/T)(G/C)GCTTAGATAGTGCTGAT) as described previously (20). PCR was performed using the Expand PCR system (Roche Biochemicals, Mannheim, Germany) under the conditions recommended by the manufacturer, with crude bacterial lysates (obtained by boiling cells in distilled water) as templates. Sequencing of the β-lactamase determinants was performed directly on PCR-generated amplicons on both strands as described previously (20). For each gene, the sequenced regions included those encoding the mature protein and part of the signal peptide. Direct sequencing of the bla_PER and bla_TEM amplicons obtained from three isolates producing the pI 5.2 and 5.6 enzymes, randomly selected (two among those producing a higher amount of the pI 5.2 enzyme and one among those producing a lower amount of the enzyme, all of which were clonally related; see below), always revealed the presence of alleles encoding mature enzymes identical to PER-1 (16) and TEM-2 (1). Direct sequencing of the PCR-amplified bla_TEM genes obtained from three isolates producing the pI 5.9 enzyme (selected as representatives of the three circulating clones, see below) revealed, in all cases, the presence of an allele encoding a mature enzyme identical to TEM-52 (24) (Table 1).

Unlike in other reports, where TEM-type enzymes were found to represent the most common ESBLs in P. mirabilis (5, 22), in this case an “unconventional” class A enzyme was found to be the most prevalent ESBL. PER-1 has recently been reported in Pseudomonas aeruginosa and other gram-negative nonfermenters from northern Italy (12, 20). Present findings indicate that, in the same area, PER-1 is also emerging in P. mirabilis, a species where this enzyme was not previously reported. The presence of TEM-52 was not surprising since it had already been detected in Klebsiella pneumoniae isolates from the same hospital (17) and is the most widespread TEM-type ESBL circulating in Enterobacteriaceae in Italy (22).

Concerning in vitro susceptibility, the PER-1 producers were always resistant or intermediate to cefotaxime, ceftazidime, and cefepime and, in some cases, also to aztreonam. On the contrary, the TEM-52-producers were intermediate or susceptible to cefotaxime and cefepime and were always susceptible to ceftazidime and aztreonam (Table 1), a finding that underscores the need for definition of breakpoints for screening ESBL production also in this species. The success of PER-1 could be due, at least in part, to the higher levels of resistance exhibited by the PER-1-producing strain, which also produced TEM-2, and emphasizes the potential clinical relevance that ESBLs other than of the TEM or SHV type could acquire in Enterobacteriaceae.

Clonal relationships and nosocomial distribution of ESBL producers.

The clonal relationships among ESBL producers were investigated by comparing the pulsed-field gel electrophoresis (PFGE) profiles of genomic DNA digested with SfiI. PFGE profiles were analyzed by the Bio-Rad Gene Path Procedure (Bio-Rad Laboratories, Richmond, Calif.) using the no. 5 pathogen group reagent kit. DNA fragments were electrophoresed in 1% agarose gels in 0.5× Tris-borate-EDTA buffer (26) with the Gene Path system (Bio-Rad) at 14°C and 6 V/cm for 20 h, using pulse times ranging from 5 to 50 s. Bacteriophage lambda concatemers (Bio-Rad) were used as DNA size markers. Clonal relationships based on PFGE patterns were interpreted according to the criteria proposed by Tenover et al. (28). Profiles identical or different by no more than four bands, indicating clonal relatedness, were observed with all the PER-1 producers (Fig. 1 and data not shown), suggesting that spreading of this resistance determinant had been most likely contributed by clonal expansion of a strain that had acquired it. On the other hand, profiles different by more than four bands, indicating the presence of three different clonal lineages, were detected among the isolates producing TEM-52 (Fig. 1 and data not shown). None of the TEM-52-producing clones was apparently related to that producing PER-1 (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

PFGE patterns of the genomic DNAs of P. mirabilis isolates producing ESBLs, after digestion with SfiI. Lane 1, DNA size standards (in kilobases, on the left); lanes 2 to 4, isolates producing the TEM-52 enzyme representative of the three identified clonal lineages (A, B, and C) (the PFGE patterns of the other 15 TEM-52-producing isolates were identical or related to those shown in the figure); and lanes 5 to 7, isolates producing the PER-1 enzyme; the PFGE profiles of the other 49 PER-1-producing isolates were identical or related to those shown in the figure.

Both the PER-1 and the TEM-52 producers were present since the beginning of the survey and were isolated throughout the survey period. The PER-1-producing strain was initially isolated from patients of a general intensive care unit and, subsequently, also from patients of 15 additional wards (Table 2), suggesting a remarkable spreading ability. The TEM-52 producers exhibited a more restricted distribution (Table 2), and a clear relationship was observed between clonality and ward (Table 2), suggesting that horizontal transfer of the resistance determinant could have played a relevant role in its dissemination.

Concluding remarks.

The prevalence of ESBL-producing P. mirabilis isolated from this hospital was notably higher than that observed in other settings (5). The remarkable epidemiological variability of this phenomenon, which can acquire a notable impact, suggests that screening for ESBL production should be routinely considered also for this species.