Inhibitory Activities of Quinolones against DNA Gyrase and Topoisomerase IV of Enterococcus faecalis

Abstract

We have cloned the DNA gyrase and topoisomerase IV genes of Enterococcus faecalis to examine the actions of quinolones against E. faecalis genetically and enzymatically. We first generated levofloxacin-resistant mutants of E. faecalis by stepwise selection with increasing drug concentrations and analyzed the quinolone resistance-determining regions of gyrA and parC from the resistant mutants. Isogenic mutants with low-level resistance contained a mutation in gyrA, whereas those with higher levels of resistance had mutations in both gyrA and parC. These results suggested that gyrA is the primary target for levofloxacin in E. faecalis. We then purified the recombinant DNA gyrase and topoisomerase IV enzymes of E. faecalis and measured the in vitro inhibitory activities of quinolones against these enzymes. The 50% inhibitory concentrations (IC₅₀s) of levofloxacin, ciprofloxacin, sparfloxacin, tosufloxacin, and gatifloxacin for DNA gyrase were found to be higher than those for topoisomerase IV. In conflict with the genetic data, these results indicated that topoisomerase IV would be the primary target for quinolones in E. faecalis. Among the quinolones tested, the IC₅₀ of sitafloxacin (DU-6859a), which shows the greatest potency against enterococci, for DNA gyrase was almost equal to that for topoisomerase IV; its IC₅₀s were the lowest among those of all the quinolones tested. These results indicated that other factors can modulate the effect of target affinity to determine the bacterial killing pathway, but the highest inhibitory actions against both enzymes correlated with good antienterococcal activities.

The main mechanism of antibacterial action of quinolones is the inhibition of the target enzymes DNA gyrase and topoisomerase IV. DNA gyrase and topoisomerase IV are both composed of two subunits, subunits GyrA and GyrB and subunits ParC and ParE, respectively, which are encoded by the gyrA and gyrB genes and the parC and parE genes, respectively. Studies with many bacteria have shown that mutations in the gyrA and parC quinolone resistance-determining region (QRDR) were related to quinolone resistance. In gram-negative bacteria, such as Escherichia coli, Neisseria gonorrhoeae, and Pseudomonas aeruginosa, strains with low-level resistance have been found to contain only gyrA mutations, whereas those with higher levels of resistance have been found to have mutations in both the gyrA and the parC genes (1, 6, 11, 14). This evidence indicated that GyrA is the first target of quinolones in these bacteria. On the other hand, in gram-positive bacteria, such as Staphylococcus aureus, mutations in parC (grlA) have been found to confer low-level resistance and preceded those in gyrA (3, 20). In addition, the primary targets in Streptococcus pneumoniae were reported to be different among quinolones. The primary target for the majority of quinolones (ciprofloxacin, levofloxacin, trovafloxacin) was found to be topoisomerase IV (7, 16, 17), whereas that of sparfloxacin and gatifloxacin was DNA gyrase (3, 18). In cases in which the inhibitory actions of quinolones against DNA gyrase and topoisomerase IV are different, the possibility of the occurrence of stepwise mutations that confer high levels of resistance would be higher.

Recent studies of Enterococcus faecalis suggested that ParC is the primary target for quinolones and that gyrA mutations are associated with high-level resistance (10). The evidence, however, is based on data for one clinical isolate, and there has been little enzymatic analysis of DNA gyrase and topoisomerase IV. In the study described in this report, we cloned E. faecalis DNA gyrase and topoisomerase IV genes, characterized the mutants selected in a stepwise manner with levofloxacin, and examined the inhibitory activities of quinolones against the purified enzymes to analyze the antibacterial effects of quinolones against E. faecalis.

MATERIALS AND METHODS

Antibacterial agents and bacterial strains.

All quinolones used in this study were synthesized at New Product Research Laboratories I, Daiichi Pharmaceutical Co., Ltd., Tokyo, Japan. The bacterial strain used in this study was quinolone-susceptible strain E. faecalis ATCC 19433.

Determination of MICs.

The MICs were determined by a standard agar dilution method (15) with Mueller-Hinton agar (Difco Laboratories, Detroit, Mich.). Drug-containing agar plates were inoculated with 1 loopful (5 μl) of an inoculum corresponding to about 10⁴ CFU per spot and were incubated for 18 h at 37°C. The MIC was defined as the lowest drug concentration that prevented the visible growth of bacteria.

Cloning of DNA gyrase and topoisomerase IV genes.

Nucleotide sequence data from The Institute for Genomic Research (Rockville, Md.) E. faecalis Genome Project were screened against the EMBL prokaryote library by using the BLAST software program.

PCR and protocols.

The sequences of the primers used in the PCRs are shown in Table 1. PCR was performed with the Expand High-Fidelity PCR system (Boehringer Mannheim, Indianapolis, Ind.), according to the manufacturer's recommendations.

TABLE 1.

Primers used for PCR and sequencing

Primer	Sequencea	Location
Pr-EFGA1	5′-ATGAGTGAAGAAATTAAAGAAAACATTCA-3′	bp 1-29 of gyrA
Pr-EFGA2	5′-GCGTTTTTCCCAATCTCTAGATTCG-3′	321 bp downstream of stop codon
Pr-EFGA4	5′-ACTCATACGTGCTTCGGTATAACGC-3′	bp 363-387 of gyrA
Pr-EFGB1	5′-ATGACAGAAGAAGAAAAAAACATGAGAG-3′	bp 1-28 of gyrB
Pr-EFGB2	5′-ATATATGTTCTAGAGTCTCTTACAACG-3′	37 bp downstream of stop codon
Pr-EFPC1	5′-GTGACAATTTTGGAAAAACGCCAAG-3′	bp 1-25 of parC
Pr-EFPC2	5′-AACACATAACCTTCTAGATGCAAGCG-3′	224 bp downstream of stop codon
Pr-EFPC4	5′-CACCACTTAACTGTGATAAACGAGC-3′	bp 379-403 of parC
Pr-EFPE1	5′-TTGGCTAAAAAAATTAACAATGAGTAC-3′	bp 1-27 of parE including the stop codon
Pr-EFPE2	5′-TTTTTCCAAGCTTGTCACTCCCTACTC-3′

^aThe initiation codon and the stop codon (underlined) and the introduced XbaI or HindIII site (underlined italic type) are indicated.

Selection and characterization of levofloxacin-resistant E. faecalis isolates.

Approximately 10⁸ CFU of E. faecalis ATCC 19433 was plated onto brain heart infusion agar plates containing increasing concentrations of levofloxacin in multiples of the MIC. Second-step, third-step, and fourth-step mutants were obtained similarly by using the first-step, second-step, and third-step mutants, respectively. The gyrA and parC QRDRs were amplified by PCR and sequenced by the dye termination method. For the gyrA QRDR, the forward primer was Pr-EFGA1 and the reverse primer was Pr-EFGA4. The primers for the parC QRDR were Pr-EFPC1 and Pr-EFPC4. The PCR conditions were 25 cycles of 94°C for 0.5 min, 60°C for 0.5 min, and 72°C for 0.5 min.

Construction of expression vectors.

Four sets of oligonucleotide primers were designed for amplification of the gyrA, gyrB, parC, and parE genes and their subsequent insertion into the pMAL-c2 fusion protein expression vector (New England Biolabs, Beverly, Mass.). In each case, the sequence of the forward primer was chosen at the initiation codon. For reverse primers, either the XbaI or the HindIII site was introduced for cloning purposes. For gyrA, the forward primer was Pr-EFGA1 and the reverse primer was Pr-EFGA2. The primers used for gyrB gene amplification were Pr-EFGB1 and Pr-EFGB2, the primers used for parC gene amplification were Pr-EFPC1 and Pr-EFPC2, and the primers used for parE gene amplification were Pr-EFPE1 and Pr-EFPE2. PCR was carried out with genomic DNA from E. faecalis strain ATCC 19433. Each gene was amplified for 20 cycles, in which the PCR conditions were 0.5 min at 94°C for denaturation, 0.5 min at 63°C for annealing, and 2 min at 72°C for polymerization. The amplified gyrA, gyrB, and parC products were digested with XbaI, and the amplified parE product was digested with HindIII. The digested DNA fragments were ligated into the XmnI and XbaI sites (gyrA, gyrB, and parC) or the XmaI and HindIII sites (parE) of the pMAL-c2 expression vector and transformed into E. coli MC1061 (12).

Purification of enzymes.

The GyrA and GyrB proteins of DNA gyrase and the ParC and ParE proteins of topoisomerase IV were purified separately as maltose-binding protein (MBP) fusion products from an overproducing strain of E. coli. E. coli MC1061/pMAL-c2 cells containing one of the genes mentioned above were incubated in Luria-Bertani broth until the log phase of growth (optical density at 600 nm, 0.5), and then isopropyl-β-d-thiogalactopyranoside was added to the culture at a final concentration of 0.3 mM to induce protein synthesis. After a 2-h incubation, the cells were harvested and resuspended in TGED buffer (50 mM Tris-HCl [pH 8.0], 10% glycerol, 1 mM EDTA, 1 mM dithiothreitol [DTT]) supplemented with 0.5 mg of lysozyme per ml and were then incubated for 30 min at 4°C. The suspension was centrifuged at 100,000 × g for 40 min, and the supernatant was loaded onto an amylose resin column previously equilibrated with TGED buffer. The column was then washed with 10 volumes of TGED buffer, and the fusion proteins were eluted with 10 mM maltose. The eluted fusions were dialyzed twice against TGED buffer at 4°C for 6 h and concentrated by dialysis against 50 mM Tris-HCl (pH 8.0)-50% glycerol-1 mM EDTA-1 mM DTT. The purified fusion proteins were digested with factor Xa to remove the MBP.

Determination of inhibitory activities of drugs.

The supercoiling activity of DNA gyrase and the decatenation activity of topoisomerase IV were measured by previously described methods (20). One unit of supercoiling activity was defined as the amount of GyrA and GyrB proteins required to supercoil 50% of 0.2 μg of relaxed pBR322 plasmid DNA. One unit of decatenation activity was defined as the amount of ParC and ParE proteins required to fully decatenate 0.4 μg of kinetoplast DNA. The 50% inhibitory concentration (IC₅₀) was defined as the drug concentration that reduced the enzymatic activity observed with drug-free controls by 50%.

For supercoiling assays, the standard reaction mixture (20 μl) contained 20 mM Tris-HCl (pH 7.5), 2 mM MgCl₂, 50 mM KCl, 1 mM DTT, 1 mM spermidine, 1 mM ATP, 20 μg of tRNA per ml, 20 μg of bovine serum albumin per ml, 0.2 μg of relaxed pBR322 DNA, and GyrA and GyrB proteins. The reaction mixtures were incubated at 37°C for 1 h, the reactions were terminated by the addition of a dye mix, and then the products were analyzed by electrophoresis in 0.8% agarose.

Decatenation assay mixtures (20 μl) containing 39 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 50 mM KCl, 1 mM DTT, 0.5 mM ATP, 50 μg of bovine serum albumin per ml, 0.4 μg of kinetoplast DNA, and topoisomerase IV subunits were incubated at 37°C for 1 h. The reactions were terminated, and the DNA products were examined by electrophoresis on 0.8% agarose gels.

Nucleotide sequence accession numbers.

The DNA sequences corresponding to the DNA gyrase and topoisomerase IV genes have been assigned GenBank accession nos. AB059405 and AB059406, respectively.

RESULTS

Screening of DNA gyrase and topoisomerase IV subunit genes.

The gyrA and gyrB genes of DNA gyrase and the parC and parE genes of topisomerase IV were identified from The Institute for Genome Research genome database. The gyrA, gyrB, parC, and parE genes code for proteins of 833, 649, 823, and 685 amino acids, respectively. The sequence of each protein showed 56.6 to 78.0% identity to the sequences of the corresponding proteins in S. aureus and S. pneumoniae, whereas the sequences of the proteins showed 31.6 to 54.4% identity to those of the corresponding proteins of E. coli. The initiation codons of parC and parE were GTG and TTG, respectively; although infrequent, these codons have previously been described in S. aureus and S. pneumoniae as well (3, 16).

Sequence analysis of levofloxacin-resistant mutants of E. faecalis selected in a stepwise manner.

To analyze the target of the quinolones in E. faecalis, we developed mutants of susceptible strain E. faecalis ATCC 19433 by stepwise exposure to levofloxacin. Approximately 10⁸ CFU of E. faecalis ATCC 19433 was plated onto brain heart infusion agar plates containing levofloxacin at the MIC (1.56 μg/ml). Mutants appeared after 48 h of growth at 37°C, and two clones (clones Lr1-1 and Lr1-2) were chosen for further analysis. Second-step selection was then carried out by plating strain Lr1-1 on plates containing 3.13 μg of levofloxacin per ml, from which two clones (clones Lr2-1 and Lr2-2) were chosen. Third- and fourth-step mutants, which grew in the presence of 25 and 50 μg of levofloxacin per ml, respectively, were generated similarly. The MICs of the quinolones for the mutant strains were measured simultaneously, and the presence of mutations in the QRDRs of their gyrA and parC genes was investigated by DNA sequence analysis of selected clones (Table 2). The nucleotide sequences of the QRDRs of Lr1-1 and Lr1-2 were identical to the nucleotide sequence of the QRDR of strain ATCC 19433. This result indicated that resistance in the first-step mutants could ensue from a non-topoisomerase-related mechanism, such as an efflux pump (2, 8). Second-step mutants had acquired a single mutation in the gyrA QRDR: Lr2-1 and Lr2-2 had a substitution of Ser-84 (AGT) to Arg (AGA) and had no mutation in parC. Third-step mutants, for which the levofloxacin MICs were higher than those for the second-step mutants, also carried no parC mutations. However, parC mutations did appear in a fourth-step mutant: Lr4-1 carried the parC mutation of Ser-85 (AGT) to Ile (ATT). Thus, mutations in parC occurred after those in gyrA and were associated with high-level resistance to levofloxacin. This result suggested that gyrA is the primary target in E. faecalis. This result is in conflict with those of a former study (10), which reported that topoisomerase IV is the primary target of quinolones in clinical isolates of E. faecalis. A recent study showed, however, that either DNA gyrase or topoisomerase IV could be the primary target, depending on the structure of the quinolone (4, 18, 19).

TABLE 2.

Properties of mutants of E. faecalis ATCC 19433 selected for resistance by stepwise exposure in vitro to levofloxacin

Strain	MIC (μg/ml)a						Mutation
	LVX	STX	CIP	SPX	TFX	GAT	GyrA	ParC
ATCC 19433	1.56	0.1	0.78	0.39	0.39	0.39
Lr1-1	3.13	0.2	1.56	0.78	0.78	0.78
Lr2-1	6.25	0.39	3.13	12.5	3.13	3.13	Ser-84→Arg
Lr3-1	50	1.56	25	12.5	3.13	12.5	Ser-84→Arg
Lr4-1	100	3.13	100	50	50	25	Ser-84→Arg	Ser-85→Ile

^aLVX, levofloxacin; STX, sitafloxacin; CIP, ciprofloxacin; SPX, sparfloxacin; TFX, tosufloxacin; GAT, gatifloxacin.

Purification and characterization of DNA gyrase and topoisomerase IV subunit proteins.

The GyrA and GyrB proteins of DNA gyrase and the ParC and ParE proteins of topoisomerase IV of E. faecalis ATCC 19433 were purified separately as MBP fusion proteins. After digestion with factor Xa, the bands for each protein on a sodium dodecyl sulfate-polyacrylamide gel stained with Coomassie brilliant blue were about 95, 70, 95, and 70 kDa for GyrA, GyrB, ParC, and ParE, respectively (data not shown). Although none of the single-protein subunits had enzymatic activity (Fig. 1, lanes 1, 2, 6, and 7), GyrA and GyrB together showed supercoiling activity and ParC and ParE together showed decatenation activity (Fig. 1, lanes 3 and 8, respectively). These enzymes were also ATP and Mg²⁺ dependent because enzymatic activities were not detected in the absence of ATP and Mg²⁺ (Fig. 1, lanes 4, 5, 9, and 10). These results indicate that E. faecalis DNA gyrase and topoisomerase IV were successfully reproduced in vitro.

FIG. 1.

Enzymatic activities of purified DNA gyrase (A) and topoisomerase IV (B) of E. faecalis ATCC 19433. Lane 1, GyrA; lane 2, GyrB; lane 3, GyrA and GyrB; lane 4, GyrA and GyrB without ATP; lane 5, GyrA and GyrB without Mg²⁺; lane 6, ParC; lane 7, ParE; lane 8, ParC and ParE; lane 9, ParC and ParE without ATP; lane 10, ParC and ParE without Mg²⁺.

Comparison of inhibitory activities of antibacterial agents against DNA gyrase and topoisomerase IV.

The quinolones inhibited enzyme activities in a concentration-dependent manner (Fig. 2). In contrast, benzylpenicillin, which did not inhibit either enzyme, had no effect on their activities (data not shown). The IC₅₀s of the quinolones were calculated by quantifying the bands corresponding to supercoiled DNA or fully decatenated substrate. The IC₅₀s of sitafloxacin, levofloxacin, ciprofloxacin, sparfloxacin, tosufloxacin, and gatifloxacin for DNA gyrase were 1.38, 28.1, 27.8, 25.7, 11.6, and 5.60 μg/ml, respectively (Table 3). Thus, sitafloxacin was the most potent among the quinolones tested. The IC₅₀s of sitafloxacin, levofloxacin, ciprofloxacin, sparfloxacin, tosufloxacin, and gatifloxacin for topoisomerase IV were 1.42, 8.49, 9.30, 19.1, 3.89, and 4.24 μg/ml, respectively (Table 3). The IC₅₀s of all quinolones except sitafloxacin for topoisomerase IV were lower than those for DNA gyrase. These IC₅₀s suggested that topoisomerase IV is more sensitive than DNA gyrase to the quinolones, but these data are not consist with the prediction indicated by the genetic approach. The IC₅₀s of all quinolones except sparfloxacin for the type II topoisomerases correlated with their MICs. Among the quinolones tested, sitafloxacin showed the highest inhibitory activities against both enzymes.

FIG. 2.

Inhibitory activities of sitafloxacin against supercoiling activity of DNA gyrase (A) and decatenation activity of topoisomerase IV (B) from E. faecalis ATCC 19433. Lanes 1 to 6, 0, 0.39, 0.78, 1.56, 3.13, and 6.25 μg of sitafloxacin per ml, respectively; lanes 7 to 13, 0, 0.2, 0.39, 0.78, 1.56, 3.13, and 6.25 μg of sitafloxacin per ml, respectively.

TABLE 3.

Inhibition of E. faecalis ATCC 19433 topoisomerase IV and DNA gyrase by quinolones

Quinolone	MIC (μg/ml)	IC50 (μg/ml)
		Topoisomerase IV	DNA gyrase
Sitafloxacin	0.10	1.42	1.38
Levofloxacin	1.56	8.49	28.1
Ciprofloxacin	0.78	9.30	27.8
Sparfloxacin	0.39	19.1	25.7
Tosufloxacin	0.39	3.89	11.6
Gatifloxacin	0.39	4.24	5.60

DISCUSSION

We have cloned E. faecalis DNA gyrase and topoisomerase IV and analyzed the actions of quinolones against E. faecalis genetically and enzymatically. First, we have examined isogenic levofloxacin-resistant mutants and have shown that GyrA is the primary target in E. faecalis. This result is in conflict with those of a former study (10) showing that ParC is the primary target of quinolones in E. faecalis, as the investigators found a clinical isolate with a parC mutation but not one with a gyrA mutation. That study, however, was based on data for just one clinical isolate, and clinical isolates with a gyrA mutation but not a parC mutation would be expected to exist. Another hypothesis would be to assume that the primary target changes, depending on the structure of the quinolones. Recent studies with S. pneumoniae showed that either DNA gyrase or topoisomerase IV could be the primary target: ciprofloxacin and sparfloxacin act preferentially through topoisomerase IV and DNA gyrase, respectively (4, 18). As the background of the clinical isolate is not clear, it might not have been exposed to levofloxacin. More data will be needed to determine the primary target for quinolones in E. faecalis.

For isogenic mutants, some features of the resistance to quinolones do not involve alteration of the enzymes. No mutation was detected in mutants with low-level resistance (clones Lr1-1 and Lr1-2). As the level of resistance was modest in the first-step mutants, it is conceivable that an efflux pump (8) is related to their quinolone resistance. The levofloxacin, sitafloxacin, ciprofloxacin, and gatifloxacin MICs for the third-step mutants (clones L3-1 and L3-2) were higher than those for the second-step mutants (clones L2-1 and L2-2), but no other mutation was detected in either gyrA or parC. These results suggested that mutations in other regions, such as gyrB and parE, or a combination of mutations and some efflux pump activity may occur to confer quinolone resistance. On the other hand, the sparfloxacin and tosufloxacin MICs for the third-step mutants were the same as those for the second-step mutants. Some efflux pumps are known to confer resistance to hydrophilic quinolones, such as ciprofloxacin, but not to hydrophobic quinolones, such as sparfloxacin (5), suggesting that efflux pumps rather than mutations in gyrB or parE are responsible for the resistance observed in third-step mutants. Some multidrug resistance efflux pumps were reported in E. faecalis (2), and a combination of these pumps with a mutation in topoisomerases would contribute to quinolone resistance.

Enzymatic analysis showed that topoisomerase IV was more sensitive than DNA gyrase to inhibition by levofloxacin, ciprofloxacin, sparfloxacin, tosufloxacin, and gatifloxacin. This result indicated that topoisomerase IV would be the primary target for these quinolones. The fact, however, that this suggestion conflicts with the results obtained by mutant analysis suggests that DNA gyrase is the primary target (Table 2). To reconcile the enzymatic and genetic results, it is necessary to consider the killing pathway after inhibition of the enzymes. Previous work with E. coli has indicated that a ternary complex of quinolone, enzyme, and DNA (the cleavable complex) triggers bacterial cell death in vivo. It is conceivable that in vivo these quinolone-gyrase-DNA complexes could be more lethal than those formed with topoisomerase IV. At present, little is known about the formation and breakdown of cleavable complexes in E. faecalis. More work will need to be done to resolve the various mechanistic possibilities.

In a comparison of the IC₅₀s of the quinolones tested, sitafloxacin showed the highest levels of inhibitory activity against both enzymes. As sitafloxacin showed good in vitro antienterococcal activity (9, 13), the IC₅₀ is one parameter of antibacterial activity. Moreover, the IC₅₀s of sitafloxacin for DNA gyrase and topoisomerase IV were almost equal, whereas the other quinolones were shown to inhibit the two enzymes at different levels. Although the primary target was not obvious, mutations in the DNA gyrase or topoisomerase IV gene could be the first step toward drug resistance, and then these mutants could gradually acquire higher levels of resistance to quinolones in a stepwise manner. As the frequency of any single mutation is low, multiple mutations are unlikely to occur at the same time. Therefore, a quinolone which has multiple targets would be desirable. Furthermore, such a quinolone would be effective against parC mutants with wild-type gyrA and against gyrA mutants with wild-type parC because it has the ability to inhibit the other wild-type enzyme.

The IC₅₀s of sparfloxacin for both enzymes (25.7 μg/ml for DNA gyrase and 19.1 μg/ml for topoisomerase IV) were higher than those of tosufloxacin and gatifloxacin, although the MIC of sparfloxacin (0.39 μg/ml) was as low as those of tosufloxacin (0.39 μg/ml) and gatifloxacin (0.39 μg/ml). These results might indicate that other factors affect the antienterococcal activity of sparfloxacin, such as good permeability or less efficient efflux. It may also be suggested that the cleavable complex of sparfloxacin is more lethal than those of the other quinolones.

In addition to mutations in gyrA and parC, mutations in gyrB and parE and mutations that prevent drug accumulation as a result of activation of drug efflux pumps cause resistance to quinolones. To overcome these problems, some approaches for the development of antibiotics have been advanced. The efficient expression system for E. faecalis type II topoisomerases described in this report should facilitate studies on the screening of antienterococcal drugs and open the way to crystallographic approaches for structure-based drug design.