Contribution of Topoisomerase IV and DNA Gyrase Mutations in Streptococcus pneumoniae to Resistance to Novel Fluoroquinolones

Abstract

In this study, we assessed the activity of ciprofloxacin, levofloxacin, sparfloxacin, and trovafloxacin against clinical isolates of Streptococcus pneumoniae that were resistant to the less-recently developed fluoroquinolones by using defined amino acid substitutions in DNA gyrase and topoisomerase IV. The molecular basis for resistance was assessed by using mutants selected with trovafloxacin, ciprofloxacin, and levofloxacin in vitro. This demonstrated that the primary target of trovafloxacin in S. pneumoniae is the ParC subunit of DNA topoisomerase IV, similar to most other fluoroquinolones. However, first-step mutants bearing the Ser79→Phe/Tyr substitution in topoisomerase IV subunit ParC were susceptible to trovafloxacin with a minimum inhibitory concentration of 0.25 μg/ml, and mutations in the structural genes for both topoisomerase IV subunit ParC (parC) and the DNA gyrase subunit (gyrA) were required to achieve levels of resistance above the breakpoint. The data also suggest that enhanced activity of trovafloxacin against pneumococci is due to a combination of factors that may include reduced efflux of this agent and an enhanced activity against both DNA gyrase and topoisomerase IV.

The gram-positive bacterium Streptococcus pneumoniae is one of the most important pathogens responsible for upper and lower respiratory tract infections, acute otitis, and meningitis (22). The rapid spread of pneumococcal clones resistant to β-lactam and macrolide antibiotics has led some to suggest that the use of selected fluoroquinolones may be appropriate for the treatment of pneumococcal infections (24). Of concern is the increase in S. pneumoniae resistance to less-recently developed fluoroquinolones that has recently been reported (4). A continuous search for newer, more potent agents has led to the clinical development of hydrophobic quinolones such as sparfloxacin and trovafloxacin, both of which are reported to be two- to eightfold more active than ciprofloxacin against penicillin-sensitive and -resistant pneumococci (1, 6, 23). Understanding the targets of the fluoroquinolones and improving how they are used in clinical practice are key to avoiding the rapid emergence of resistance to these agents in pathogenic microbes.

We have recovered several S. pneumoniae strains resistant to less-recently developed fluoroquinolones from patients infected with these bacteria. We analyzed the activity of trovafloxacin and sparfloxacin against clinical isolates of S. pneumoniae that demonstrate different levels of resistance to ciprofloxacin and levofloxacin [S-(−)-ofloxacin]. We also measured the in vitro activity of these agents against ciprofloxacin-, levofloxacin-, and trovafloxacin-resistant in vitro-selected mutants in order to determine the relative contributions of gyrase and topoisomerase IV mutations in these isolates. The quinolone resistance-determining regions (QRDRs) of gyrA, gyrB, parC, and parE also were sequenced to determine the association of specific DNA changes in these areas with phenotypic expression of resistance.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The clinical isolates of S. pneumoniae tested were recovered at Northwestern Memorial Hospital (Chicago, Ill.) from January 1996 through March 1997. These isolates were designated SP30, 6406, 6711, 6513, and 6678. Strain SP30 is susceptible to all tested fluoroquinolones. Strain RT1 was provided by Evanston Northwestern Healthcare in whose facilities it was isolated from a patient with pneumonia that failed treatment with levofloxacin. A highly susceptible laboratory strain, designated CP1000, has been described previously (20). It is an isolate that was recovered before the introduction of fluoroquinolones and was used in our studies as a susceptible control as well as the strain for selection of resistant mutants. For these experiments, organisms were grown in the laboratory at 35°C in Todd-Hewitt broth (Difco Laboratories, Detroit, Mich.) supplemented with 0.5% yeast extract (THBY) or on tryptic soy agar plates (Difco) supplemented with 5% sheep’s blood. A casein hydrolysate-yeast extract-tryptone medium (CAT) was used for mutant selection (11). S. pneumoniae transformation was performed in CAT supplemented with 0.2% bovine serum albumin and 1 mM CaCl₂, as described previously (18).

Susceptibility testing.

The susceptibility of the isolates to antimicrobial agents was determined by the microdilution method using Mueller-Hinton broth (Difco) supplemented with 5% lysed horse blood or, alternatively, by twofold agar dilution with corresponding antimicrobial agents prepared in Mueller-Hinton agar supplemented with 5% sheep’s blood (13). The following agents were provided by their manufacturers: levofloxacin (Ortho-McNeil Pharmaceuticals, Raritan, N.J.), ciprofloxacin (Bayer Corporation, West Haven, Conn.), sparfloxacin (Rhône-Poulenc Rorer R-D, Vitry-sur-Seine, France), and trovafloxacin (Pfizer Pharmaceuticals Group, New York, N.Y.). Prior to testing, individual strains were grown overnight in CO₂ at 35°C on tryptic soy agar plates (Difco) supplemented with 5% sheep’s blood.

Selection of mutants.

First-step mutants were obtained by exposing S. pneumoniae CP1000 to twice the minimum inhibitory concentration (MIC) of each agent: 2 μg of levofloxacin per ml, 1 μg of ciprofloxacin per ml, and 0.25 μg of trovafloxacin per ml. Exposure to each drug was achieved through plating 1 ml of an S. pneumoniae CP1000 culture, grown in THBY to an optical density at 550 nm of 0.4 (2 × 10⁹ cells/ml), onto a second layer of CAT agar on a two-layer plate with the concentration of the antimicrobial agent in the bottom layer doubled. In each experiment, approximately 10¹⁰ cells were used for mutant selection at each drug concentration. Individual clones were subcultured from selection plates into THBY broth with the same concentration of drug as that on which they were selected.

Stability testing of selected mutants.

The stability of the acquired resistance for all selected mutants was tested by subculture of the organisms to drug-free sheep’s blood agar plates (DiMed, St. Paul, Minn.) with a subsequent passage onto a second plate for a total of 48 h of growth. Organisms were then subcultured onto blood agar plates containing one-half the MICs, determined after selection, of the respective antimicrobial agents in order to document that resistance was stable in a drug-free environment.

PCR amplification and DNA sequence analysis.

To assess whether clinical isolates 6406, 6513, 6678, and RT1 carried amino acid substitutions in ParC, ParE, GyrA, or GyrB, the nucleotide sequences of parC, parE, gyrA, and gyrB gene fragments that included regions corresponding to QRDRs of the respective proteins for each of these strains were determined and compared to the corresponding sequences from sensitive isolates and the reference strain CP1000. The gene sequences of gyrA, gyrB, parC, and parE were retrieved from GenBank (accession no. AF053121, Z67740, AF065151, and AF065153, respectively). A 253-bp fragment of gyrA (bp 342 to 595; amino acids [aa] 115 to 198), a 453-bp fragment of gyrB (bp 1080 to 1533; aa 361 to 511), a 337-bp fragment of parC (bp 164 to 501; aa 55 to 167), and a 413-bp fragment of parE (bp 1175 to 1587; aa 392 to 529) were amplified by using the following pairs of primers: GyrA1 (5′-CGTCGCATTCTCTACGGA-3′) and GyrA2 (5′-CGTCGCATTCTCTACGGA-3′), GyrB1 (5′-CTCTTCAGTGAAGCCTTCTCC-3′) and GyrB2 (5′-CTCCATCGACATCGGCATC-3′), ParC1 (5′-TGACAAGAGCTACCGTAAGTCG-3′) and ParC2 (5′-TCGAACCATTGACCAAGAGG-3′), and ParE1 (5′-ACGTAAGGCGCGTGATGAG-3′) and ParE2 (5′-CTAGCGGACGCATGTAACG-3′). Amplification was performed with AmpliTaq DNA polymerase (Perkin-Elmer Cetus) on an MJ Research Peltier Thermal Cycle PTC-100. Either 0.1 μg of chromosomal DNA or 1 μl of bacterial culture at an optical density at 550 nm of 0.2 was used as a template in standard 50-μl PCRs. Sequencing was carried out on the amplified PCR products by using the ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction kit according to the protocol of the manufacturer (Perkin Elmer). An ABI PRISM 310 genetic analyzer was used for sequencing. All testing was performed in duplicate.

Transformation of detected mutants.

To confirm which of the detected amino acid substitutions in GyrA, ParC, and ParE are responsible for the ciprofloxacin resistance, we transferred the corresponding gyrA, parC, and parE mutations into the laboratory strain CP1000 by genetic transformation. PCR products corresponding to regions bp 342 to 595 in gyrA, bp 164 to 501 in parC, and bp 1175 to 1587 in parE were used as donor DNA. Transformants were selected on 0.75 μg of ciprofloxacin per ml and screened after a 48-h incubation at 35°C. Individual colonies of transformants were transferred to THBY broth and incubated at 35°C overnight, and the resulting cultures were preserved for nucleotide sequence analysis. In each transformation experiment, chromosomal DNA from strain CP1500 (hex nov-r1 bry-r str-r1 ery-r2 ery-r6) was used as a source of the novobiocin resistance (Nov-R) marker for the monogenic transformation control.

RESULTS

Antimicrobial susceptibility of S. pneumoniae clinical isolates.

The seven S. pneumoniae strains with various levels of fluoroquinolone susceptibility used for this study were confirmed to be of nonclonal origin by restriction endonuclease analysis with HaeIII (data not shown). Susceptibility of these strains to ciprofloxacin, levofloxacin, sparfloxacin, and trovafloxacin is presented in Table 1. According to their susceptibility to ciprofloxacin, these isolates could be categorized as susceptible (SP30 and CP1000; MIC < 1 μg/ml) or resistant (6513, 6711, 6678, 6406, and RT1; MIC ≥ 2 μg/ml) to this widely used agent. The MICs of the less-recently developed hydrophilic quinolone agents (levofloxacin and ciprofloxacin) were higher than the MICs of sparfloxacin and trovafloxacin. Interestingly, clinical isolates 6406, 6711, 6513, 6678, and RT1 (resistant to ciprofloxacin) were still susceptible to trovafloxacin at concentrations less than or equal to 2 μg/ml (Table 1).

TABLE 1.

Susceptibilities of S. pneumoniae clinical isolates to ciprofloxacin, levofloxacin, and trovafloxacin and detected amino acid substitutions in DNA gyrase and topoisomerase IV

Strain	MIC (μg/ml) ofa:				Amino acid substitution inb:
	CIPRO	LEVO	SPAR	TROVA	GyrA	GyrB	ParC	ParE
CP1000	0.5	1.0	0.25	0.12	—	—	—	—
SP30	0.5	1.0	0.25	0.12	—	—	—	—
6711	2.0	2.0	0.5	0.12	Ser114→Gly	—	Asn91→Asp	—
6406	2.0	4.0	0.5	0.5	—	—	Ser79→Phe	Ile460→Val
6513	4.0	4.0	1.0	0.5	Ser114→Gly	—	Ser79→Tyr	Ile493→Leu
							Asn91→Asp
6678	4.0	4.0	0.5	0.5	—	—	Asp83→Tyr	Ile460→Val
RT1	>4.0	>8.0	4.0	2.0	Ser81→Tyr	—	—	Ile460→Val
								Asp435→Asn

^aCIPRO, ciprofloxacin; LEVO, levofloxacin; SPAR, sparfloxacin; TROVA, trovafloxacin. 

^b—, no change in the amino acid sequence. 

Nucleotide sequence analysis of the QRDRs of gyrA, gyrB, parC, and parE.

The ciprofloxacin-susceptible strains of S. pneumoniae did not bear DNA mutations in their respective QRDRs, and therefore, they contained no amino acid substitutions in GyrA, GyrB, ParC, or ParE. However, four of the resistant isolates, 6711, 6406, 6513, and 6678, had substitutions in the ParC subunit of topoisomerase IV. Isolates 6406, 6513, 6678, and RT1 bore changes in the ParE subunit of DNA topoisomerase IV. Three isolates, 6711, 6513, and RT1, also contained GyrA alterations. Interestingly, strain RT1, which demonstrated the highest resistance of the tested strains, did not have amino acid substitutions in ParC, but rather in the ParE subunit of topoisomerase IV. This QRDR had not been assessed in most prior reports (9, 17), and therefore, its role in some phenotypic resistance evaluations may have been overlooked (8, 16).

Transfer of fluoroquinolone resistance by genetic transformation.

To demonstrate that the detected amino acid substitutions in GyrA, ParC, and ParE are in fact responsible for ciprofloxacin resistance, we transferred the corresponding gyrA, parC, and parE mutations into the laboratory strain CP1000 (Table 2). The number of clones transformed to ciprofloxacin resistance was normalized to the number of transformants resulting from a monogenic transformation with a Nov-R chromosomal marker. The levels of monogenic transformation with Nov-R ranged from 6.5 × 10³ to 1.6 × 10⁴ transformants/μl, corresponding to transformation frequencies of 4.5 × 10⁻³ to 1.1 × 10⁻² in five independent transformation experiments.

TABLE 2.

Transfer of gyrA, parC, and parE mutations into S. pneumoniae CP1000 by genetic transformation

Donor strain	Donor gene	Resulting amino acid substitution(s) transferreda	No. of ciprofloxacin-resistant transformants/μlb	No. of putative transformants analyzed	MIC (μg/ml) for transformants ofc:				Amino acid substitution(s) in transformant(s)
					C	L	S	T
RT1	gyrA	Ser81→Tyr	2.80 × 10−5						NDd
	parE	Asp435→Asn	1.78	4	2.0	2.0	0.5–1.0	0.5	Asp435→Asn
		Ile460→Val							Ile460→Val
6406	parC	Ser79→Phe	1.43	2	1.0–2.0	2.0	0.5	0.5	Ser79→Phe
	parE	Ile460→Val	3.12 × 10−5						ND
6711	gyrA	Ser114→Gly	1.2 × 10−5						ND
	parC	Asn91→Asp	2.1 × 10−5						ND
6678	parC	Asp83→Tyr	1.77	2	2.0	2.0	0.5	0.5	Asp83→Tyr
	parE	Ile460→Val	1.72 × 10−4	2					No change
6513	gyrA	Ser114→Gly	2.53 × 10−5						ND
	parE	Ile493→Leu	1.04 × 10−5						ND

^aPCR-amplified regions bearing mutations corresponding to the indicated amino acid substitutions were used as donor DNA. 

^bFor normalization of transformation values, each recipient was transformed in parallel with a segment of chromosomal DNA from strain CP1550 carrying a chromosomal marker for Nov-R (nov-r1), and the ratio of ciprofloxacin-resistant and novobiocin-resistant transformants listed here was determined. 

^cC, ciprofloxacin; L, levofloxacin; S, sparfloxacin; T, trovafloxacin. 

^dND, no data available. 

As is evident in Table 2, levels of transformation to ciprofloxacin resistance upon the introduction of parC fragments from strains 6406 and 6678 were consistent with a monogenic transformation event. Nucleotide sequence analysis of selected transformants confirmed the transfer of mutations encoding the Ser79→Phe and Asp83→Tyr substitutions, respectively. In addition, the MICs of ciprofloxacin, levofloxacin, sparfloxacin, and trovafloxacin for the analyzed transformants increased to 1.0 to 2.0, 2.0, 0.5, and 0.5 μg/ml, respectively. Therefore, amino acid substitutions at aa 79 and 83 in ParC, considered essential for low-level resistance to fluoroquinolones (12, 14, 21), conferred ciprofloxacin resistance on the recipient strain.

Several individual mutations did not appear to lead to the development of ciprofloxacin resistance. The transformation of CP1000 with parC fragments from strain 6711 did not yield resistant transformants, suggesting that the Asn91→Asp substitution in ParC does not contribute to ciprofloxacin resistance in this strain. The transfer of mutations encoding amino acid substitutions Ser81→Tyr and Ser114→Tyr in GyrA also did not lead to an increase in resistance, based on transformation with gyrA fragments from RT1, 6711, and 6513. Similarly, neither Ile460→Val nor Ile493→Leu in ParE conferred ciprofloxacin resistance in transformation experiments.

Transformation of S. pneumoniae CP1000 with a parE fragment from RT1, however, produced ciprofloxacin-resistant transformants at the levels of transformation with a single chromosomal marker. Four transformants were analyzed; each was shown to bear Asp435→Asn and Ile460→Val substitutions in ParE that correlated with a two- to fourfold increase in the MICs of ciprofloxacin, levofloxacin, and sparfloxacin, thus strongly suggesting that the amino acid substitution Asp435→Asn is the primary alteration responsible for fluoroquinolone resistance in RT1.

Genetic analysis of the in vitro resistance mutants.

Ciprofloxacin-, levofloxacin-, and trovafloxacin-resistant mutants were selected in vitro to further examine the role of topoisomerase IV and DNA gyrase in the development of resistance. Mutant selection was attempted at two, four, and eight times the MIC of each fluoroquinolone. In three replicate experiments, the observed frequencies of resistance selection were similar at concentrations twice the MICs of ciprofloxacin, levofloxacin, and trovafloxacin (2.2 × 10⁻⁷ to 4.2 × 10⁻⁷ transformants/μl). Most interestingly, mutant clones resistant to both four and eight times the MICs of ciprofloxacin and levofloxacin were obtained at frequencies of 5.0 × 10⁻¹⁰ to 4.5 × 10⁻⁹ transformants/μl. However, in all three independent selection experiments, we failed to recover any mutants resistant to more than twice the MIC of trovafloxacin.

The amino acid substitutions in GyrA, GyrB, ParC, and ParE of the first-step mutants selected at twice the MICs of the studied agents are listed in Table 3. All analyzed mutants bore homogeneous amino acid substitutions at aa 79 in ParC. The Ser79 substitution in the ParC subunit of topoisomerase IV resulted in a fourfold increase in resistance to ciprofloxacin and twofold increases in resistances to levofloxacin, trovafloxacin, and sparfloxacin.

TABLE 3.

Selection and properties of the fluoroquinolone-resistant mutants^a

Mutant strain	Parent strain	Selection agent (concentration in μg/ml)	MIC (μg/ml) of:				Amino acid substitution in:
			CIPRO	LEVO	SPAR	TROVA	GyrA	GyrB	ParC	ParE
First-step selection
1C1, 1C3, 1C4	CP1000	CIPRO (1.0)	2.0	2.0	0.5	0.25	—b	—	Ser79→Tyr	—
1C2	CP1000	CIPRO (1.0)	2.0	2.0	0.5	0.25	—	—	Ser79→Phe	—
1L1, 1L2, 1L3, 1L4	CP1000	LEVO (2.0)	2.0	2.0	0.5	0.25	—	—	Ser79→Phe	—
1T1	CP1000	TROVA (0.25)	2.0	2.0	0.5	0.25	—	—	Ser79→Phe	—
1T2	CP1000	TROVA (0.25)	2.0	2.0	0.5	0.25	—	—	Ser79→Tyr	—
Second-step selection
2C1, 2C2	1C2	CIPRO (4.0)	8.0	8.0	8.0	4.0	Ser81→Phe	—	Ser79→Phe	—
2L1	1L3	LEVO (4.0)	8.0	16.0	8.0	4.0	Ser81→Tyr	—	Ser79→Phe	—
2L2	1L3	LEVO (4.0)	8.0	16.0	8.0	4.0	Ser81→Phe	—	Ser79→Phe	—
2T1	1T1	TROVA (0.5)	8.0	8.0	8.0	4.0	Ser81→Phe	—	Ser79→Tyr	—
2T2	1T1	TROVA (0.5)	8.0	16.0	8.0	4.0	Ser81→Phe	—	Ser79→Phe	—

^aCIPRO, ciprofloxacin; LEVO, levofloxacin; SPAR, sparfloxacin; TROVA, trovafloxacin. 

^b—, no change in the amino acid sequence. 

Using first-step mutants 1L3 (levofloxacin), 1C1 (ciprofloxacin), and 1T1 (trovafloxacin) as parental strains, we selected second-step mutants resistant to ciprofloxacin, levofloxacin, and trovafloxacin on plates containing the corresponding fluoroquinolone agent at twice its MIC for each first-step mutant (Table 3). The observed frequency of second-step resistance selection in this experiment was similar to frequencies of first-step selection. All of the analyzed second-step mutants carried amino acid substitutions both in the ParC subunit of topoisomerase IV and in the GyrA subunit of DNA gyrase, indicating a similar adaptive resistance mechanism for the pneumococci against all three agents. Sparfloxacin was not analyzed in this portion of our study since it has already been shown to first give rise to mutations in GyrA, but it should be noted that an increase in resistance to sparfloxacin was seen even with an isolated ParC mutation and that trovafloxacin was at least twice as active as any of the other fluoroquinolones tested, even in the presence of mutations (Table 3).

DISCUSSION

Fluoroquinolone agents exert their antibacterial actions via the inhibition of homologous type II topoisomerases, DNA gyrase and DNA topoisomerase IV. The gyrase is a tetramer composed of two subunits encoded by the gyrA and gyrB genes, respectively. It catalyzes ATP-dependent negative supercoiling of DNA and is implicated in the replication, recombination, and transcription of the bacterial chromosome. Topoisomerase IV is believed to participate in the partitioning of replicated chromosomes before cell separation. The two subunits of this enzyme are encoded by the parC and parE genes. Since these enzymes interact with DNA in a similar manner, fluoroquinolone action on either gyrase or topoisomerase IV can be lethal to the bacterial cell. This suggests that relatively good potency against both enzymes may be a drug strategy superior to high potency against one with inherent resistance in the other.

Bacterial resistance to quinolones is thought to arise primarily through point mutations at the highly conserved amino acid residues in the QRDR of the GyrA subunit of DNA gyrase and the ParC subunit of DNA topoisomerase IV or, less frequently, in the QRDRs of GyrB and ParE (17). ParC is considered the primary target for ciprofloxacin in gram-positive organisms (7, 12, 14, 21). The Ser79→Tyr and Asp83→Tyr substitutions likely alter important target affinity, since they are known to be responsible for initial, low-level resistance of gram-positive bacteria to ciprofloxacin (7, 12, 15, 21). Mutations at the equivalent positions of the GyrA subunit of DNA gyrase A are secondary events and lead to very high levels of resistance, presumably by making both topoisomerase IV and gyrase relatively resistant to fluoroquinolone action. While most quinolone antimicrobials appear to have an affinity for topoisomerase IV, sparfloxacin was recently reported to first target GyrA, an observation based on the fact that mutations in strains selected on this fluoroquinolone accumulated primarily in gyrA (16). However, our data show that isolated amino acid substitutions in ParC affect the activity of sparfloxacin, which strongly suggests drug action against both enzymes influences the overall action of most, if not all, fluoroquinolones.

The QRDRs of GyrA, GyrB, ParC, and ParE in clinical isolates with elevated levels of resistance to ciprofloxacin, ofloxacin, levofloxacin, and sparfloxacin showed that these isolates carried combinations of amino acid substitutions in GyrA, ParC, and/or ParE. These combinations of amino acid substitutions in DNA gyrase and DNA topoisomerase IV also led to a decreased susceptibility to trovafloxacin in most strains. Thus, in agreement with earlier observations (9, 12, 17), these data indicate that mutations in both the DNA gyrase and topoisomerase IV genes are required for high-level fluoroquinolone resistance and suggest the same conclusion for trovafloxacin, although it appeared inherently more active than the other agents studied, even when a dual mutation was present. This suggests novel structural modifications in this agent have provided a relatively broad activity against both gyrase and topoisomerase IV.

It is also interesting to note the marked heterogeneity of the DNA gyrase and topoisomerase mutations observed in the clinical strains recovered from patients with pneumococcal infections, as opposed to the homogeneity of mutations observed in the laboratory-derived mutants. We noted even more diversity than was reported in the recent work by Jorgensen et al. (9). Such an observation indicates that continued study of clinical isolates will be crucial to our further understanding of how bacteria adapt to the pressures of escalating antimicrobial agent use.

Nongyrase targets were previously implicated in the resistance of S. pneumoniae to certain fluoroquinolones (2, 3, 5). It is highly likely that other factors, such as a lesser efficiency of trovafloxacin efflux by the pneumococci, contributes significantly to the high activity of this fluoroquinolone (3). The observation we made on clinical isolate 6711 provides additional support for this hypothesis. This strain demonstrated elevated resistance to ciprofloxacin and levofloxacin but the same sensitivity to trovafloxacin as susceptible isolates SP30 and CP1000 (trovafloxacin MIC, 0.12 μg/ml). Although 6711 bore amino acid substitutions Ser114→Gly in GyrA and Asn91→Asp in ParC, these substitutions failed to confer ciprofloxacin resistance in transformation experiments. Therefore, ciprofloxacin and levofloxacin resistances in this strain are likely due to factors other than mutations in the genes encoding DNA gyrase or DNA topoisomerase IV. Ciprofloxacin is a good substrate for active efflux in gram-positive bacteria (2, 3, 25), possibly due to an overexpression of a putative efflux (NorA) mechanism (10). This mechanism, however, does not appear to contribute to the resistance of S. pneumoniae to trovafloxacin (3). The reduced efflux of trovafloxacin is also suggested by our experiments which demonstrated that first-step resistance selection does not occur at concentrations exceeding twice the MIC of trovafloxacin. Any agent that avoided active efflux would provide increased intracellular drug concentrations, leading to early death of the entire bacterial population with markedly reduced survival of any mutants. Two reports have suggested that the prevalence of expression of quinolone efflux in S. pneumoniae is nearly 50%, highlighting the importance of this mechanism (5, 19). Additional work in this area is clearly warranted (10).

The results from our experiments suggest that the enhanced activity of certain new fluoroquinolones against pneumococci is due to a combination of factors, ranging from the resistance to selection of drug-resistant mutants at drug levels only slightly above the MIC to an enhanced activity against both DNA gyrase and topoisomerase IV. It is crucial to evaluate the mechanisms that contribute to the enhanced action of novel agents such as trovafloxacin, as these mechanisms provide insights into optimal therapeutic use as well as potential directions for ongoing discovery of drugs needed for our future.