INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Quinolones are synthetic antimicrobial agents which target topoisomerases II and IV in bacteria (4, 7-9, 11). Most mutations to quinolone resistance involve changes in gyrA and/or parC, although mutations are known in gyrB and parE. Mycoplasma hominis is an opportunistic pathogen commonly found in the male and female genital tracts. The organism is unusual: it has no cell wall, a small genome (700 kbp), a low guanine-plus-cytosine ratio (30%) in its DNA, and a high mutation rate (5). Its susceptibility to quinolones closely parallels that of Staphylococcus aureus (13). The purpose of the current study was to determine the rates of mutation to quinolone resistance in first- and second-step mutants of M. hominis and to determine the roles of GyrA and ParC in this resistance. Some of these data have been reported in a preliminary communication (14).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mycoplasma hominis GX55, a strain from Seattle, Wash., was grown in H broth (12) supplemented with 20% horse serum, 5 mM arginine, 10% fresh yeast extract, 0.001% phenol red, and 200 U of penicillin per ml. The H-agar medium (12) was similar: arginine was omitted and agarose was added at 0.7%. For selection of mutants, H agar was supplemented with various concentrations of a quinolone. Strain PG-21 was obtained from the American Type Culture Collection (Bethesda, Md.) as strain 23114 (6). Actively growing cultures were disaggregated by filtration through a 0.6-µm-pore-size polycarbonate filter (Nuclepore; Corning, Acton, Mass.) and were inoculated (0.1 ml) onto H-agar plates (10 ml in 60-mm-diameter plates) containing a quinolone at 0, 1.0, 2.0, 4.0, and 8 times the MIC determined for unselected organisms. The number of CFU was determined by plating serial 10-fold dilutions on control agar. The frequency of occurrence of mutants (mutant frequency) was the number of colonies found on selective media divided by the number of CFU found on control plates. Mutant colonies were excised with a plastic pipette, and the agar plugs were inoculated into quinolone-free broth medium. After growth for 24 h, broth cultures were stored at 70°C. Mutants were cloned on quinolone-free agar. The susceptibilities of the mutants were determined by the agar dilution method (15).

Prototypical sequences of the proposed gyrA and parC regions of M. hominis were initially obtained by PCR with DNA from strain PG-21 (6) and degenerate primers in the N-terminal regions of the genes as described previously (10). The assignment of the M. hominis gyrA and parC gene fragments was based on their amino acid sequence homology with other well-characterized gyrA and parC sequences. M. hominis-specific primers were subsequently derived from these prototypical sequences. The primers selected for gyrA were 5'-GCACCGTAGAATTTTATATGG-3' and 5'-CATACCGACCGCTATTCCACT-3', which yielded a product of 361 bp, excluding the primers. The primers for parC were 5'-CGTCGGATTTTATATTCAATG-3' and 5'-GGTGATTCCTTTAGCACCGTT-3', which yielded a 348-bp fragment, excluding the primers. Mycoplasmal cultures (1.5 ml) were centrifuged, the pellet was solubilized with 10 µl of 0.25% deoxycholate in 10 mM Tris-HCl-1 mM EDTA at (pH 8.3), and 1 µl of proteinase K (200 µg/ml) was added. Two microliters of GyrA and 3 µl of ParC primers (at 5 pmol/ml) were combined with 3 µl of template in the reaction mixture (18). Amplification was carried out by heating the sample to 93°C for 4 min, followed by 35 cycles of amplification (30 s each at 93, 56, and 72°C) with a final synthesis of 5 min at 72°C. The PCR products were evaluated by electrophoresis in 2% SeaKem LE agarose (FMC Bioproducts, Rockland, Maine), followed by ethidium bromide staining (19). The PCR products were sequenced with the Big Dye Terminator Ready Reaction Kit on an ABI Prism 377XL DNA Sequencer (Perkin-Elmer, Foster City, Calif.) at the Biochemistry Department, University of Washington. The sequences in both directions were determined for most but not all mutants. Sequences were aligned by using the program Sequencher 3.1 (Gene Codes Corp., Ann Arbor, Mich.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Small mutant colonies of various sizes were observed on plates containing ofloxacin at 1×, 2×, and 4× the MIC beginning on day 2. Colonies increased in size for several days, and additional colonies appeared day by day, whereas the sizes and numbers of colonies on control plates with approximately 100 colonies were stable after day 2. Mutant colonies selected with sparfloxacin were smaller on day 3 and took longer to reach the same diameter as those selected with ofloxacin. However, subcultures of mutants grew as rapidly as the parent strain in quinolone-free medium. The inoculum size averaged 2.8 × 10⁷ CFU (range, 1 × 10⁶ to 1.8 × 10⁸ CFU) in the 0.1-ml sample applied to the selection plates. The average mutant frequency was 1.5 × 10⁶ (range, 1.4 × 10⁵ to 1.6 × 10⁷). The mutant frequency among first-step mutants was the same by selection with either sparfloxacin or ofloxacin. The average mutant frequency for second-step mutants was closely similar at 1.8 × 10⁶ (range, 1.3 × 10⁵ to 6.6 × 10⁷). Independent cultures were prepared by inoculating separate broth tubes with 10¹ to 10⁴ mycoplasmas. Selection of five cultures (examples of two cultures [cultures A and B] are shown in Table 1) independently derived with sparfloxacin at 0.03 µg/ml (1× the MIC) or 0.125 µg/ml gave rise to strains with 8- to 16-fold increased levels of resistance to sparfloxacin (MICs, 0.25 to 0.5 µg/ml) but no increase in the level of resistance to ofloxacin. A similar picture was seen when the same independent cultures were selected with ofloxacin (cultures C and D, Table 1). The MIC of ofloxacin increased four- to eightfold, whereas the susceptibility to sparfloxacin was unchanged. Mutant colonies were seen only on plates containing 1× to 4× the MIC of either sparfloxacin or ofloxacin. The number of colonies diminished with an increase in the quinolone concentration. No mutant colonies were observed on plates containing 8× the MIC, even though the MICs for the first-step mutants were two- to eightfold higher than that of the selecting quinolone.

The nucleotide sequence of the amplified portion of gyrA of strain GX-55 was compared with that of gyrA of M. hominis PG-21. Strain GX-55 showed two nucleotide changes at amino acid positions 158 (GTCGTT; Escherichia coli numbering) and 159 (TTGTTA). Two nucleotide changes were observed in the parC fragment: one at the hot spot for resistance mutations at position 84 (GAAGAG) and the other at position 133 (AAAAGA). None of these nucleotide changes affected the amino acid sequence. Our results for PG-21 were the same as those of Bébéar et al. (1-3).

First-step, sparfloxacin-selected mutants showed changes in GyrA (Table 1) at amino acid position 83 (SerLeu) or at position 87 (GluLys). First-step mutants selected with ofloxacin showed mutations in ParC: at position 80 (SerIle) or at position 84 (GluLys). Selection of a first-step mutant with a mutation at ParC80 (strain D), again with ofloxacin, yielded two different step-two mutants: GyrA83 (SerTrp) and GyrA83 (SerLeu) (mutants D1 and D2, Table 1). The MIC of sparfloxacin increased 64- to 128-fold, from 0.015 to 4 or 8 µg/ml, and that of ofloxacin increased 16- to 32-fold. Selection of the single-step ParC84 mutant with sparfloxacin gave two-step mutants with a mutation at GyrA87 (GluLys), with similar large increases in resistance to both quinolones (mutant C1). Selection of a first-step mutant with a mutation at GyrA87 (mutant A) with sparfloxacin or ofloxacin resulted in step-two mutants that had changes at ParC80 (SerIle) or ParC84 (GluLys). A second selection of the mutant with a mutation at GyrA83 (mutant B) with sparfloxacin or ofloxacin gave second-step mutants with substitutions at ParC80 (SerIle). Overall, susceptibility to ofloxacin decreased 128-fold and that to sparfloxacin decreased 4- to 8-fold compared to the susceptibilities of strains with single mutations.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A principal finding of the study was that the initial topoisomerase targets in M. hominis selected with sparfloxacin or ofloxacin differ: sparfloxacin selected for GyrA mutants and altered the gyrase target, whereas ofloxacin selected for ParC mutants and affected the topoisomerase IV target. Because mutations in parC did not increase the level of resistance to sparfloxacin, first-step parC mutants could be selected with ofloxacin but not with sparfloxacin. Similarly, first-step gyrA mutants were selectable only with sparfloxacin because they showed no increase in resistance to ofloxacin.

A comparison of our results for first- and second-step mutants with those from the recent studies by Bébéar et al. (2, 3) shows the variety of mutations which can occur on selection of M. hominis with ofloxacin and sparfloxacin. They recovered three single-step mutants: those with mutations at ParE426 (AspAsn), GyrA83 (SerLeu), and GyrA84 (SerTrp). Only one of these overlapped with the four that we derived. Their two-step mutants, those with mutations at GyrA83 (SerLeu), ParC84 (GluLys), and GyrA83 (SerLeu)-GryA84 (SerTrp), differed from the four two-step mutants that we found. Thus, six different single-step mutants and six distinct two-step combinations are possible. Closely similar results have been reported by Pan and Fisher (16) for selection of mutants of Streptococcus pneumoniae with ciprofloxacin and sparfloxacin. Sparfloxacin selected for first-step GyrA mutants and ciprofloxacin selected for a first-step ParC mutant. First-step GyrA mutants had increased levels of resistance to sparfloxacin but not to ciprofloxacin. The ParC mutant had increased levels of resistance to ciprofloxacin but not to sparfloxacin. Second-step mutants had mutations in the other topoisomerase and greatly increased levels of resistance to both quinolones.

One mutant (mutant D1) showed a GyrA83 (SerTrp) substitution as a result of a TCATGA change. This mutation would be lethal to conventional bacteria, but mycoplasmas decode TGA as tryptophan rather than as a stop (5). Bébéar et al. (2) found this mutation for both serines at GyrA83 and GyrA84. The two tryptophans in the ParC fragment from the parent culture were encoded by TGA at amino acid positions 50 and 93. The GyrA sequence amplified from the parent culture had no tryptophans.

Single mutations in either gyrase or topoisomerase IV gave modest increases in the levels of resistance of M. hominis to the selecting quinolone, but high-level resistance was seen for strains that have mutations in both topoisomerases, as is true for bacteria in general (4, 7, 8, 10, 20). Even though ofloxacin did not select for first-step GyrA mutations, it readily selected for GyrA mutations in strains with mutations already present in ParC (Table 1). For one group of double mutants (GyrA83 [SerLeu] and ParC80 [SerIle]; strains B1, B2, and D2, Table 1), sparfloxacin MICs were 8 µg/ml, which is fourfold greater than those for another group of mutants with double mutations (GyrA87 [GluLys] and ParC84 [GluLys]; strains A2, A3, and C1).

The recognition of selective targeting of GyrA or ParC provides some theoretical concepts for the development of strategies for the treatment of infections caused by organisms for which selective targeting similar to that observed with M. hominis is shown. The use of ofloxacin would be disadvantageous for organisms for which MICs are close to the levels known to be achievable in blood because a single-step mutation could result in clinical resistance. In contrast, two mutations would be required to achieve clinical resistance to sparfloxacin, given its relatively greater activity. However, the use of sparfloxacin for the treatment of infections caused by organisms with ParC mutations (as a result of previous treatment with ofloxacin) could likely fail because a single mutation could lead to clinical resistance. Several strategies have been proposed: one is to find quinolones which equally target gyrase and topoisomerase IV (21). Recently, Pan and Fisher (17) have shown dual targeting of S. pneumoniae gyrase and topoisomerase IV by clinafloxacin. A related proposal would be to use both ofloxacin and sparfloxacin simultaneously to control infections with M. hominis or organisms with similar susceptibilities. In both of these cases, the frequency of development of resistance is expected to be much less because the organism needs to acquire mutations in both the gyrase and topoisomerase IV targets simultaneously to develop resistance. In the case of M. hominis, this could give a mutant frequency of 10¹², whereas that for strains with single-step mutations is 10⁶. The result that first-step mutations may well appear to be "silent" when tested with one quinolone suggests that quinolone susceptibility testing needs to be done with quinolones which prefer different targets, as has been suggested by Bébéar et al. (3) for mycoplasmas. Since the patterns of resistance in M. hominis closely parallel those in gram-positive bacteria (13), the high M. hominis mutation rates make this organism a good model for studies of quinolone resistance.