Nonmolecular Test for Detection of Low-Level Resistance to Fluoroquinolones in Streptococcus pneumoniae

Abstract

With respect to pneumococci, there is a need to detect first-step mutants with reduced fluoroquinolone (FQ) susceptibility from which second-step, resistant mutants are likely to be selected in the presence of antipneumococcal FQs. Here, we describe an interpretative disk diffusion test, of which three options are presented, that allows the distinction between first- and second-step mutants. Using five FQ disks (pefloxacin, norfloxacin, levofloxacin, ciprofloxacin, and sparfloxacin, option 1), all known mechanisms of altered FQ susceptibility found in first-step mutants (ParC, ParE, GyrA, or efflux) and in second-step mutants (ParC and GyrA or ParE and GyrA) can be accurately detected, making this option a useful epidemiological tool. Using three FQ disks (pefloxacin, norfloxacin, and levofloxacin, option 2), the most prevalent FQ-resistant mutants, but not the first-step GyrA mutants, can be detected. With only two FQ disks (norfloxacin and levofloxacin) in the third and simplest option, first-step mutants can be distinguished from second-step mutants, however, without differentiation of ParC, ParE, or efflux alterations.

Since the 1990s, fluoroquinolone (FQ) resistance among strains of Streptococcus pneumoniae has been emerging worldwide. The increasing use of FQs for the treatment of a variety of community-acquired infections has led to an increased prevalence of the FQ-resistant strains, which, however, remains low, ranging from 1 to 7% (2, 3, 6, 8, 16, 20, 23, 27, 32, 33). The prevalence of FQ-resistant strains of S. pneumoniae in France is currently about 2%, the highest rate (4%) being found in strains isolated from the respiratory tract in adults (Centre National de Référence des Pneumocoques [CNRP], rapport d'activité 2003, http://www.invs.sante.fr/surveillance). Decreased susceptibility to FQs results mainly from amino acid substitutions in the quinolone resistance-determining region (QRDR), either in the topoisomerase IV, preferentially in the ParC subunit, or in the DNA gyrase, preferentially in the GyrA subunit, or in both (5, 16, 17, 18, 26, 31).

FQ resistance associated with target mutations is acquired through a stepwise process (15). First-step mutants (also designated in this work as low-level-resistant mutants) generally result from mutations in the preferential target for a given FQ, ParC for pefloxacin, ciprofloxacin, and levofloxacin or GyrA for sparfloxacin, moxifloxacin, gatifloxacin, gemifloxacin, and garenoxacin (7, 10, 11, 12, 13, 21, 22). Another mechanism underlying first-step resistance is an increase in active efflux which affects quinolones such as norfloxacin and ciprofloxacin (34) more than the antipneumococcal FQs (11, 13). In the second-step mutants (designated in this work as high-level-resistant mutants), amino acid substitutions are present in both topoisomerase IV and gyrase, most frequently affecting ParC and GyrA, and less so ParE and GyrA (25). They result in resistance to all the currently available antipneumococcal FQs (levofloxacin, moxifloxacin, and gatifloxacin).

Outbreaks of respiratory tract infections due to FQ- and multidrug-resistant pneumococci have been reported, reflecting the propensity of these strains to spread (32). Furthermore, first-step mutants with topoisomerase IV (4, 29) or gyrase (24) mutations have been shown to be associated with treatment failures in some cases of pneumonia. As emphasized by these authors, since FQs can be used as first-line drugs for the treatment of community-acquired pneumonia, clinicians need to be informed about the FQ susceptibilities of the pneumococcal strains isolated from their patients. This implies that microbiologists must be able to detect first-step, low-level-resistant mutants (13, 30) since they have the potential to generate second-step, high-level-resistant mutants.

According to the recommendations of the Clinical and Laboratory Standards Institute (CLSI) (19) and the Comité de l'Antibiogramme of the Société Française de Microbiologie (CA-SFM, http://www.sfm.asso.fr), the breakpoints for antipneumococcal FQs allow the detection only of high-level-resistant mutants affected in at least two targets. Consequently, the prevalence of pneumococci nonsusceptible to FQs may not be accurately evaluated if surveillance systems rely simply on the criterion of susceptibility to antipneumococcal fluoroquinolones. Among strains with a levofloxacin MIC of 2 μg/ml (which are considered susceptible according to both CLSI and CA-SFM rules), 59 to 71% were shown to have a parC mutation (5, 18). Attempts to distinguish the low-level-resistant mutants from the wild-type strains by testing for ciprofloxacin or levofloxacin susceptibility failed since neither test had adequate sensitivity or specificity (28). Thus, today, no phenotypic test is available for the accurate detection of the first-step mutants.

We previously constructed isogenic, R6-derived strains harboring various mutations in ParC or GyrA to evaluate their impact on FQ activity and showed that, depending upon the particular FQ considered, topoisomerase IV or gyrase appeared to be the preferential target (30). Subsequent analysis of the results of that study revealed that the disk diffusion method with four “older” FQs, i.e., pefloxacin, norfloxacin, ciprofloxacin, and sparfloxacin, allowed discrimination between low-level-resistant mutants and the fully FQ-susceptible strain R6 (unpublished data). Based on this observation and exploiting a large collection of S. pneumoniae strains with defined relevant mechanisms, the aim of the present study was to design a phenotypic test for the detection of the low-level FQ- resistant mutants and therefore to allow their distinction from the FQ-susceptible strains.

MATERIALS AND METHODS

Susceptibility testing.

The antimicrobial agents pefloxacin, norfloxacin, ciprofloxacin, sparfloxacin, levofloxacin, and moxifloxacin were provided by their manufacturers. MICs were determined on Mueller-Hinton agar supplemented with 5% horse blood using a Steers replicator device and an inoculum of 10⁴ to 10⁵ CFU per spot and read after 18 h at 37°C. In order to detect resistance due to increased active efflux (1, 9), the MICs of ciprofloxacin and norfloxacin were also determined in the presence of reserpine (10 μg/ml) (Sigma, St. Quentin-Fallavier, France). Disk diffusion testing was performed according to the recommendations of the CA-SFM using disks with 5 μg of pefloxacin, norfloxacin, ciprofloxacin, sparfloxacin, levofloxacin, and moxifloxacin (Bio-Rad Diagnostics, Marnes-La-Coquette, France). CA-SFM and CLSI interpretative standards were used when available.

Strains.

As part of our ongoing surveillance program at the CNRP, representative strains of the French epidemiological situation were collected between 2001 and 2003 from the 22 regional surveillance teams (Observatoires Régionaux du Pneumocoque) which participate in the French CNRP surveillance network. In addition, many strains which were assumed to be of reduced quinolone susceptibility were sent to the CNRP. A total of 97 clinical isolates of reduced susceptibility or resistant to FQs were analyzed further and the QRDRs of parC, gyrA, parE, and gyrB were sequenced. No low-level-resistant mutants harboring a GyrA mutation were found. To compensate for the lack of in vivo-selected gyrase mutants, 12 gyrase (gyrA) mutants were selected in vitro on moxifloxacin from different susceptible clinical isolates (13; this study).

Strains with well-defined mechanisms of resistance (20 isogenic, R6-derived transformants and 15 in vitro-selected mutants) were also included in this study (13, 30). The following isogenic strains were used as internal controls for susceptibility testing and detection of the different phenotypes: S. pneumoniae R6, a derivative of the encapsulated Rockefeller University strain R36A, for the wild-type phenotype; Tr₅₉₂₉ for the efflux phenotype; Tr1, which harbors an S79Y ParC substitution, for the ParC phenotype; Tr6, which harbors an S81F GyrA substitution, for the GyrA phenotype; and Tr10, which harbors two substitutions, S79Y in ParC and S81F in GyrA, for the ParC and GyrA phenotype (30).

Chromosomal DNA extraction and PCR experiments.

The regions encompassing the QRDRs of ParC, ParE, GyrA and GyrB were amplified as previously described (13, 14, 15). Direct sequencing was performed using the dRhodamine BigDye Terminator sequencing kit (Perkin-Elmer, Applied Biosystems Division) with the oligonucleotides used for amplification.

RESULTS AND DISCUSSION

The MICs of six different FQs as well as the genotypes of 144 pneumococcal strains expressing either low- or high-level resistance are listed in Table 1. For each FQ, the correlation between MICs and inhibition zone diameters for a given mechanism of resistance is shown in Fig. 1, with the breakpoints or the cutoff values when they exist. Table 2, derived from the data in Table 1 and Fig. 1, gives the mean inhibition zone diameters for the six FQs used and the pneumococcal strains according to their genotype. From these results, using a diffusion test with five FQ disks containing pefloxacin, norfloxacin, ciprofloxacin, sparfloxacin, and levofloxacin but not moxifloxacin, we considered possible the detection of four phenotypes, each corresponding to a defined mechanism of resistance, three for the low- and one for the high-level-resistant mutants.

TABLE 1.

Phenotypes and genotypes of, and MICs of different fluoroquinolones for, strains of S. pneumoniae with decreased susceptibility to FQs

Strains	Genotype (no. of strains)	Amino acid substitution(s)				No. of strains	MIC rangea (μg/ml)
		Topoisomerase IV		Gyrase			NOR	PEF	CIP	LVX	SPX	MXF
		ParC	ParE	GyrA	GyrB
Wild-type strains
R6	Wild type	—d					8	8	1	0.5	0.25	0.12
Wild-type clinical isolates	NAb					1,150	2-16 (16)	2-16 (8)	0.5-4 (2)	0.25-2 (1)	0.06-0.5 (0.5)	0.06-0.25 (0.25)
First-step low-level-resistant mutants (n = 77)
Efflux mutantsc	Efflux (14)	—d	—	—	—	14	32-64	8-16	4-8	1-2	0.25-0.5	0.12-0.25
ParC mutants	parC (44)	S79F	—	—	—	19	32-≥128	32-64	4-16	1-2	0.5-1	0.12-0.5
		S79Y	—	—	—	15	32-64	32-64	4-16	1-4	0.5-1	0.12-0.5
		D83N	—	—	—	5	32-64	32	4-8	2-4	0.5-1	0.12-0.5
		D83Y	—	—	—	3	64	32	4	2	0.5	0.25
		D83G	—	—	—	1	≥128	64	4	1	0.5	0.25
		S79Y, D83Y	—	—	—	1	≥128	64	4	1	0.5	0.25
ParE mutants	parE (2)	—	D435N	—	—	2	32	32	4	2	0.25-0.5	0.25
GyrA mutants	gyrA (16)	—	—	S81F	—	8	8	4-8	1-2	1-2	0.5-2	0.5-1
		—	—	S81Y	—	3	4-8	8-16	1-2	0.5-2	1-2	0.5-1
		—	—	E85K	—	2	4-8	8	1	0.5-1	1-2	0.5
		—	—	S81Y, E85K	—	1	8	8	1	0.5	2	0.5
		—	—	E85G	—	1	8	8	1	2	1	0.5
		—	—	S81A	—	1	8	8	1	1	0.5	0.5
ParC + efflux mutants	parC + efflux (3)	S79Y	—	—	—	2	64	64	16	4	1	0.5-0.25
		S79A	—	—	—	1	≥128	32	16	4	1	0.5
Second-step mutants (n = 65)
ParC + GyrA mutants	parC + gyrA (57)	S79F	—	S81F	—	22	64-≥128	64-≥128	16-64	8-16	8-16	2-8
		S79F	—	E85K	—	2	64-≥128	64-≥128	16-64	8-32	8-32	2-4
		S79F	—	S81C	—	1	64	64	16	8	4	2
		S79Y	—	S81F	—	11	64-≥128	64-≥128	16-≥128	4-32	8-16	2-8
		S79Y	—	S81Y	—	2	≥128	≥128	32	16	8-16	4
		S79Y	—	E85K	—	2	64-≥128	≥128	32-64	16-32	32-64	4-8
		S79Y	—	S81Y, E85K	—	1	≥128	≥128	32	64	64	8
		D83G	—	S81F	—	2	64	64-≥128	32	8	8	2
		D83G	—	S81Y, E85K	—	1	≥128	≥128	32	8	16	4
		D83G	—	S81A	—	1	64	≥128	16	8	4	1
		D83G	—	S81Y	—	1	≥128	≥128	32	8	8	4
		S79Y, D83Y	—	S81Y, E85K	—	1	≥128	≥128	64	64	≥128	≥128
		S79Y, D83Y	—	S81F	—	1	≥128	≥128	32	16	16	8
		S79Y, D83Y	—	E85K	—	1	≥128	≥128	32	64	64	8
		S79Y, D83Y	—	S81Y	—	1	≥128	≥128	64	16	16	2
		D83N	—	S81F	—	3	32-≥128	64-≥128	16	8-16	4-16	2-4
		D83N	—	E85G	—	1	32	32	8	8	16	4
		D83N	—	S81Y	—	1	32	64	16	16	16	4
		S79I	—	S81F, G114	—	1	≥128	≥128	64	16	16	4
		S79A	—	S81F	—	1	32	64	16	8	8	2
ParE + GyrA mutants	parE + gyrA (7)	—	D435N	S81F	—	5	64-≥128	32-64	8-32	8-16	2-4	2-4
		—	D435N	S81Y	—	1	32	32	8	8	1	1
		—	E474K	S81F	—	1	≥128	≥128	64	16	4	2
ParC + GyrB mutants	parC + gyrB (1)	S79Y	—	—	D435N	1	≥128	64	64	32	2	4

^aMIC range (MIC₉₀) for 1,150 susceptible clinical isolates of S. pneumoniae. NOR, norfloxacin; PEF, pefloxacin; CIP, ciprofloxacin; LVX, levofloxacin; SPX, sparfloxacin; MXF, moxifloxacin.

^bNA, not applicable. The QRDR was examined only for a minority of the 1,150 susceptible strains.

^cIncreased efflux was deduced from a fourfold decrease in the MIC of norfloxacin or ciprofloxacin in the presence of reserpine (10 μg/ml). The corresponding genotype designation is used only for practical purposes.

^dAbsence of mutations in the QRDR.

FIG. 1.

Distribution of 144 low-level and high-level fluoroquinolone-resistant strains of S. pneumoniae according to their genotypes, and MICs and inhibition zone diameters of six different fluoroquinolones (panels A to F). The strains are represented by colored disks, with the disk area corresponding to the number of strains and the color corresponding to a given FQ resistance mechanism. The rectangle above the abscissa represents the zone diameter range for 1,150 susceptible strains; the mean diameter is represented by the vertical bar inside the rectangle. CLSI breakpoints when available are represented by vertical lines. The vertical dotted line represents the cutoff value proposed for the detection of low-level-resistant mutants using pefloxacin (this work) and norfloxacin (recommended by the CA-SFM).

TABLE 2.

Correlation between phenotype, genotype, and inhibition zone diameter around FQ disks (5 μg) for S. pneumoniae strains with decreased susceptibility to FQs

Strains	Genotype	No. of strains	Mean zone diama (mm) ± SD
			PEF	NORb	CIP	SPX	LVX	MXF
Wild-type strains	Wild type	1,151	18 ± 3	15 ± 3	24 ± 2	27 ± 3	24 ± 2	32 ± 3
Low-level-resistant mutants	parC/parE	46	6 ± 1	6 ± 0	17 ± 3	24 ± 2	20 ± 2	29 ± 3
	parC + efflux	3	6 ± 0	6 ± 0	8 ± 3	25 ± 0	17 ± 0	30 ± 1
	Efflux	14	16 ± 2	6 ± 0	18 ± 2	24 ± 2	22 ± 1	29 ± 3
	gyrA	16	17 ± 2	15 ± 3	21 ± 3	20 ± 2	21 ± 2	25 ± 3
High-level-resistant mutants	parC + gyrA	57	6 ± 0	6 ± 0	6 ± 1	8 ± 3	7 ± 2	17 ± 3
	parE + gyrA	7	7 ± 1	6 ± 0	11 ± 4	18 ± 2	8 ± 2	20 ± 2

^aPEF, pefloxacin; NOR, norfloxacin; CIP, ciprofloxacin; SPX, sparfloxacin; LVX, levofloxacin; MXF, moxifloxacin. The CLSI and CA-SFM breakpoints (susceptible to resistant) for sparfloxacin, levofloxacin, and moxifloxacin are ≥19 to ≤15 mm and ≥20 to <16 mm, ≥17 to ≤13 mm and ≥17 to <17 mm, and ≥18 to ≤14 mm and ≥24 to <24 mm, respectively. The CA-SFM breakpoint for norfloxacin is <10 mm. Bold characters indicate the most useful FQs for the detection of the corresponding mechanism of resistance. A diameter of 6 mm is indicative of no inhibition zone around the disk tested.

^bAccording to the CA-SFM, zone diameters of <10 mm for norfloxacin (5-μg disk) suggest the presence of a mechanism of FQ resistance. It is not considered a breakpoint but a cutoff value for the detection of FQ resistance mechanisms.

Low-level-resistant mutant phenotypes.

Neither levofloxacin nor moxifloxacin was useful for the detection of low-level-resistant strains with mutations in either ParC, ParE, or GyrA or with increased active efflux (Table 2, Fig. 1A and B). Thus, the detection of these single-step mutants on the antibiogram is based on the four remaining quinolones, pefloxacin, norfloxacin, ciprofloxacin, and sparfloxacin.

Strains with increased efflux as the only mechanism of FQ resistance showed a significant reduction in zone diameter around the norfloxacin (Table 2, Fig. 1C) but not, or to a much lesser degree, around the remaining disks. This resulted in an easily recognizable phenotype on the antibiogram (Fig. 2C).

FIG. 2.

Disk diffusion test of fluoroquinolone susceptibility for five strains with defined resistance genotypes. The corresponding phenotypes are indicated in panels A to E. The position of the fluoroquinolone disks (5 μg) is indicated in panel F; the colored rectangles include the FQ disks to be tested in each of the three options (Table 3). PEF, pefloxacin; NOR, norfloxacin; LVX, levofloxacin; CIP, ciprofloxacin; SPX, sparfloxacin.

For the strains harboring only topoisomerase IV mutations, which predominantly affect ParC (Table 1), a specific phenotype could also be recognized. All these strains showed either no zone diameter or, more rarely, a significantly reduced zone around both the norfloxacin and pefloxacin disks (Table 2, Fig. 1C and D). In contrast to the strains harboring both gyrase and topoisomerase IV mutations, a significant zone diameter was observed with all the other FQs tested. Again, this resulted in a phenotype easy to detect, as illustrated in Fig. 2B. This applied to ParE mutants too (data not shown).

The strains harboring only a GyrA mutation were more difficult to detect. Considering the norfloxacin, pefloxacin, and ciprofloxacin zone diameters, strains with a GyrA mutation were not distinguishable from the wild-type strains (Table 2, Fig. 1C, D, and E). This was not surprising since the corresponding MICs differ only slightly from those of the wild-type strains (Table 1). In contrast, the sparfloxacin zone diameters were smaller than those recorded for the wild-type strains (Table 2, Fig. 1F). Therefore the GyrA phenotype would include an apparent susceptibility to all these FQs except that a reduced zone diameter should be seen with sparfloxacin in comparison to that in wild-type strains. Another approach to detect such mutants, the easiest in our experience, was comparison of the diameters around sparfloxacin and ciprofloxacin. While for the wild-type strains the zone diameters around sparfloxacin were always larger than those around ciprofloxacin, the zone diameters around sparfloxacin were smaller than or at most equal to those around ciprofloxacin in strains harboring a single GyrA mutation (Table 2, Fig. 2D).

High-level-resistant mutant phenotypes.

The highest MICs were determined for strains with mutations in ParC and GyrA, ParE and GyrA, or ParC and GyrB, known or considered to be second-step mutants (Table 1). The corresponding absence of an inhibition zone (or a small diameter) around all FQ disks tested resulted in an unambiguous resistance phenotype (Fig. 2E). However, levofloxacin, which displayed smaller zone diameters than moxifloxacin (Table 2, Fig. 1A and B), was more efficient for the detection of this phenotype, including ParE and GyrA mutants which were categorized as moxifloxacin susceptible according to CLSI interpretative breakpoints (Table 2).

If the aim is to identify one FQ mechanism of resistance in low-level-resistant mutants, this simple phenotypic method is helpful, in particular for those with alterations in ParC, ParE or active efflux: ParC or ParE mutants will display either no or only a very small inhibition zone diameter (<10 mm) around disks of both pefloxacin and norfloxacin (Table 2); in the case of the increased active efflux, a similar change concerns only norfloxacin; and although somewhat less easily, a single GyrA mutation may be detected by comparing the ciprofloxacin and sparfloxacin zone diameters, the latter typically being smaller.

Conclusion.

We tried to translate these results in three practical test options (Table 3, Fig. 2). Each option will allow distinction between susceptible strains and the important-to-detect low-level-resistant mutants in addition to detection of the high-level-resistant mutants. However, recognition of the resistance mechanism present in these low-level-resistant mutants will depend upon the option chosen. In Table 3, the categories R (for resistance) and S (for susceptibility) are suggested in a way that does not always correspond to the accepted clinical categories, in particular for norfloxacin and pefloxacin, which are not antipneumococcal FQs. Similar to the use of the oxacillin disk, the purpose of which is the detection of penicillin resistance, the use of these two FQs is only for the detection of low-level FQ resistance mechanisms, which is not possible with the new antipneumococcal FQs.

TABLE 3.

Three test options for detection of the mechanisms of FQ resistance in S. pneumoniae

Mechanism of resistance	FQ testinga (interpretive values)
	Option 1				Option 2			Option 3
	NOR (R <10 mm)	PEF (R <10 mm)	SPX/CIP	LVX	NOR (R <10 mm)	PEF (R <10 mm)	LVX	NOR (R <10 mm)	LVX
Topoisomerase IVc	R	R	SPX>CIP	S	R	R	S	Rb	S
Effluxc	R	S	SPX>CIP	S	R	S	S	R	S
Gyrase (GyrA)c	S	S	SPX<CIP	S	—d	—	—	—	—
Topoisomerase IV + gyrase	R	R	—	I or R	R	R	I or R	R	I or R

^aFor each option the indicated FQs have to be tested and this should allow detection of the indicated mechanisms of resistance. Results should be interpreted by comparing the zone diameters around all FQ disks tested in the option. Bold characters indicate the affected FQ which is the most helpful to detect the corresponding mechanism of resistance. High-level-resistant mutants are detected whatever is the option. Interpretive values correspond to the breakpoints according to CLSI or CA-SFM rules for levofloxacin (LVX). For pefloxacin and norfloxacin (5-μg disks) the cutoff value (<10 mm) is only used to detect the presence of a mechanism of resistance which would not be detected using the antipneumococcal FQs. This corresponds to results from this work for pefloxacin (PEF) and to the value defined by the CA-SFM for norfloxacin (NOR) (Table 2). The SPX/CIP test, only helpful for the detection of a single GyrA mutation(s), is based on the comparison of the inhibition zone diameter around sparfloxacin (SPX) and ciprofloxacin (CIP). I, intermediate.

^bIn this option topoisomerase IV and efflux mutants are detected but cannot be distinguished unless the effect of reserpine is determined.

^cMechanism detected in low-level-resistant mutants.

^d—, in this option GyrA mutations cannot be detected.

Option 1 (Table 3), using disks containing norfloxacin, pefloxacin, ciprofloxacin, sparfloxacin, and, among the available antipneumococcal FQs, levofloxacin, should be chosen in order to recognize each of the different resistance mechanisms present in the low-level-resistant mutants (topoisomerase IV or DNA gyrase mutation or increased efflux). This option, the only one adequate to screen for single GyrA mutations, would be particularly helpful for comprehensive epidemiologic studies for which the different mechanisms must be accurately identified.

Option 2 (Table 3) is proposed with the use of norfloxacin, pefloxacin, and levofloxacin. This allows the detection of low-level-resistant mutants harboring either mutations in topoisomerase IV (ParC or ParE) or an increased efflux, but not the recognition of GyrA mutants, which are exceptional to date. Options 1 and 2 would require that pefloxacin disks (5 μg) be made available to microbiologists.

Option 3 (Table 3), using only norfloxacin and levofloxacin, is the simplest alternative. With this option detection of the currently most frequently encountered low-level-resistant mutants is possible, though without discrimination between increased efflux and altered topoisomerase IV (ParC or ParE), which requires testing of the effect of reserpine. In our experience, two out of three low-level mutants identified using option 3 were ParC mutants. A similar frequency was recently reported in the United States (M. W. R. Pletz, A. P. Shergill, L. McGee, B. Beall, C. G. Whitney, K. P. Klugman, and the Active Bacterial Core Surveillance Team, Program Abstr. 44th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1724, 2004). However, the prevalence of these mutants will depend upon the epidemiological situation in a given country.

Although suboptimal, options 2 and 3 should be considered worthwhile alternatives to option 1 since they allow the detection of almost all low-level-resistant mutants with the potential to acquire high-level FQ resistance.