DNA Gyrase and Topoisomerase IV Mutations in Clinical Isolates of Ureaplasma spp. and Mycoplasma hominis Resistant to Fluoroquinolones

Ureaplasma urealyticum and Mycoplasma hominis are commonly isolated from the lower urogenital tract of healthy adults but could be responsible for genital and extragenital infections (7). The species U. urealyticum has been recently divided into two new species, namely, Ureaplasma parvum (previously U. urealyticum biovar 1) and U. urealyticum (previously U. urealyticum biovar 2) (10, 12). In this study, they will be considered together as Ureaplasma spp.

Fluoroquinolones interact in bacteria with the type II topoisomerases DNA gyrase and topoisomerase IV, both of which are composed of two A and two B subunits; these subunits are encoded by the gyrA and gyrB genes for DNA gyrase and by the parC and parE genes for topoisomerase IV (8). We recently reported in vitro and in vivo fluoroquinolone-resistant mutants of M. hominis associated with alterations in GyrA, ParC, and ParE quinolone resistance-determining regions (QRDRs) (4, 6). The Ureaplasma spp. quinolone-resistant clinical isolates identified during this study have been characterized for their species identification and subtyping and for their QRDR status in regard to their fluoroquinolone susceptibility. Furthermore, we describe also a new clinical isolate of M. hominis that is highly resistant to fluoroquinolones.

Growth conditions and antibiotic susceptibility testing of the Ureaplasma and M. hominis strains have been previously described (14). The Ureaplasma sp. reference strains, the serovar 3 standard (ATCC 27815) designated as the type strain of U. parvum, and the serovar 8 standard (ATCC 27618) designated as the type strain of U. urealyticum (12), as well as the clinical isolates, were grown in Shepard medium. Thirteen clinical isolates comprising UUb to UUg5 (Table 1), which were obtained at the Pellegrin Hospital in Bordeaux, and UUa, which was isolated at the Hospital of Montpellier, France, all of which were grown in our laboratory between 1989 and 1999, were studied. The M. hominis reference strain PG21 (ATCC 23114) and the clinical isolate MHe obtained at the Foch Hospital in Suresnes, France, were grown in Hayflick modified medium.

PCR was carried out with 1 μM each primer and 5 μl of template DNA for the Ureaplasma species identification and subtyping (10) and for the QRDR amplification of the four topoisomerase genes for Ureaplasma spp. and M. hominis, as described elsewhere (4, 5). PCR products were directly sequenced by using an ABI PRISM dRhodamine terminator cycle sequencing ready reaction kit (Applied Biosystems).

Thirteen clinical isolates of Ureaplasma spp. from seven patients were examined; their origin is described in Table 1. The first one, UUa, was previously reported (2, 5). Among the 1,224 Ureaplasma spp. strains isolated in our laboratory between 1989 and 1999, 37 isolates were classified as resistant to ofloxacin by the susceptibility testing kit Mycoplasma SIR (Bio-Rad). Determining the ofloxacin MICs for these 37 isolates led us to study 11 strains for which the ofloxacin MIC is ≥4 μg/ml, namely, UUb to UUd and UUe2 to UUg5 (Table 1). We report here also a new clinical isolate of M. hominis, named MHe, isolated from a knee synovial fluid of a male patient with hypogammaglobulinemia and a lung transplant. He received at least one regimen of ofloxacin.

We investigated our Ureaplasma isolates for species identification and subtyping using the algorithm recently described by Kong et al. (10). Thus, isolates UUe1 and UUe2, isolated from the same patient, were identified as U. parvum (biovar 1). The other isolates belonged to the species U. urealyticum (biovar 2), subtype 1 or subtype 2. It is tempting to speculate that the paired isolates (UUe1 and UUe2, from the same patient) and the seven isolates from the sexual partners (UUf1-2 from the male, and UUg1-5 from the female) are from the same clonal origin. Furthermore, among the strains resistant to fluoroquinolones, we isolated more U. urealyticum (from six patients) than U. parvum (from one patient), which is not the case commonly. Indeed, U. parvum clinical isolates usually outnumber U. urealyticum clinical isolates (1, 10). There is a probable bias in our study, with nine isolates being possibly from the same clonal origin and with the number of patients studied being statistically too small to be conclusive.

The 13 ureaplasmal clinical strains were characterized for their susceptibilities to eight fluoroquinolones and for the QRDR status of their gyrA, gyrB, parC, and parE genes (Table 2). Except the fluoroquinolone-susceptible isolate UUe1, all other isolates had a high-level resistance to the older fluoroquinolones, such as norfloxacin, pefloxacin, ofloxacin, and ciprofloxacin, with MICs ranging from 4 to >128 μg/ml. Susceptibility to the newer fluoroquinolones, like sparfloxacin, levofloxacin, moxifloxacin, and gemifloxacin, depends on the resistant isolate, but increases in MICs were much less pronounced except for mutant UUa. Thus, the MICs of these fluoroquinolones varied between 0.25 and 16 μg/ml.

The 12 quinolone-resistant isolates were found to carry target mutations in either parC (one isolate) or gyrA and parC (11 isolates). No mutation was detected in the gyrB or parE QRDR (Table 2). Except isolate UUe2, all other resistant isolates presented two alterations in ParC, namely, the Ala125(123)-to-Thr and Ala136(134)-to-Thr changes (Escherichia coli numbering). Our data did not allow an assessment of the contribution of these two mutations to the observed resistance phenotype. However, these positions are located close to the Tyr122(120) active site of the protein, and several mutations in this area have been frequently described for other quinolone-resistant bacteria (3, 9), but most of these mutations could not be clearly related to resistance to fluoroquinolones. All other alterations found in this study (Table 2) have been already described for other bacteria concerning either the altered position or the amino acid change. The hot spots ParC Ser83(80) and Glu87(84) and GyrA Gln100(83) have been found previously to be mutated in numerous other bacteria and in mycoplasmas (4, 6, 11). The GyrA Asp112-to-Glu alteration, which is rare and which is homologous to position GyrA 95 in E. coli, has been found to be mutated in Streptococcus pneumoniae (13) without a clear direct link to quinolone resistance. Furthermore, in vitro studies with laboratory mutants would be helpful to determine the preferential target of fluoroquinolones in ureaplasmas. Some isolates, like UUb and UUf1 or UUd and UUf1, which had the same amino acid changes in the QRDR of GyrA and ParC, harbored fourfold differences in MICs of ciprofloxacin or pefloxacin, respectively. Other undefined resistance mutations seem likely to be present in some of the strains studied.

The drug susceptibility profile and the QRDR status of the M. hominis isolate are summarized in Table 2. Compared to the reference strain PG21, isolate MHe harbored a high level cross-resistance to the eight fluoroquinolones tested, with the lowest MIC being that of gemifloxacin at 2 μg/ml. This resistance in an immunodeficiency context, which was found also for the previously described clinical strains of quinolone-resistant M. hominis (6), was associated with four target alterations (Table 2): GyrA 83, ParC 80 and 134, and ParE 462.

In summary, this is the first extensive description of fluoroquinolone resistance in clinical isolates of Ureaplasma spp. In our laboratory during a 10-year period, we found 12 Ureaplasma spp. isolates resistant to fluoroquinolones. The clinical histories of patients showed one or several quinolone regimens before the appearance of resistant bacteria and a correlation with immunosuppression. These findings confirm the necessity to monitor the antibiotic susceptibility of urogenital mycoplasmas isolated in humans and to control it routinely in immunocompromised patients.

• American Society for Microbiology