Molecular characterization of fluoroquinolone resistant Streptococcus pneumoniae clinical isolates obtained from across Canada

1. Introduction 

Streptococcus pneumoniae is a key pathogen in many respiratory tract infections. In a Canada-wide surveillance study published in 1999, it was reported that 21.2% of pneumococcal isolates tested were intermediately susceptible or resistant to penicillin (Zhanel et al., 1999b). As the newer respiratory fluoroquinolones, including levofloxacin, gatifloxacin, gemifloxacin, and moxifloxacin, are active against both penicillin-susceptible and penicillin-resistant isolates of S. pneumoniae, they will be increasingly important in the future treatment of infections caused by this pathogen Zhanel et al 1999a, The Canadian Respiratory Infection Study Group Hoban D. J et alThe Canadian Respiratory Infection Study Group Hoban D. J et al 1999b, Zhanel et al 2002. Of concern, however, is the increase in prevalence of fluoroquinolone-resistant pneumococci in Canada from 0% in 1993 to 1.7% in 1997 and 1998 (Chen et al., 1999). The purpose of this study was to characterize the resistance mechanisms em-ployed by fluoroquinolone-resistant S. pneumoniae clinical isolates.

Fluoroquinolones exert their bactericidal effect on S. pneumoniae by inhibiting the type II topoisomerases DNA gyrase and topoisomerase IV (Zhanel et al., 1999a). Resistance to the fluoroquinolones is mediated by mutations in the quinolone resistance determining regions (QRDRs) of the genes coding for the target enzymes (DNA gyrase: gyrA and gyrB, Topoisomerase IV: parC and parE) and by active efflux Bast et al 2000, Broskey et al 2000. Efflux of fluoroquinolones is carried out, at least in part, by the reserpine-sensitive efflux pump PmrA (Bast et al., 2000).

In this study, the QRDRs of gyrA, parC, and parE from 34 clinical isolates were sequenced to determine any mutations associated with resistance in the clinical setting. The contribution of efflux was assessed for ciprofloxacin, gatifloxacin, gemifloxacin, levofloxacin and moxifloxacin. Finally, the isolates were typed using pulsed field gel electrophoresis (PFGE) to investigate their genetic relatedness.

2. Materials and methods 

2.1. Bacterial strains and growth conditions 

The 34 fluoroquinolone-resistant (ciprofloxacin MIC ≥4 μg/mL) S. pneumoniae clinical isolates investigated in this study were obtained from medical centers in 8 of the 10 Canadian provinces between 1997 and 2000 (Table 1) (Zhanel et al., 1999b). Isolates were subcultured twice on 5% sheep blood agar and incubated overnight at 35°C in 5 to 10% CO2 prior to bacterial lysate preparation and efflux experiments.

Table 1. Mechanisms of fluoroquinolone resistance investigated for 34 fluoroquinolone resistant Streptococcus pneumoniae clinical isolates

									Fluoroquinolone Resistance Mechanisms Investigated
	Isolate	MIC (μg/mL)							Target Site Alterations*			Reserpine Inhibited Efflux**
		Location	Pen	Cipro	Gati	Levo	Moxi	Gemi	GyrA- Amino Acid Changes	ParC-Amino Acid Changes	ParE-Amino Acid Changes	Fold Reduction in MIC with Reserpine
												Cipro	Gati	Levo	Moxi	Gemi
	3104	Winnipeg, MB	0.03	4	0.5	2	0.25	0.03	None observed	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2	2	0	0	0
	4610	Montreal, QB	0.03	4	0.5	2	0.25	0.06	None observed	Ser79 to Phe	Ile460 to Val	0	0	0	2	2
	9286	Vancouver, BC	0.12	4	0.5	2	0.25	0.03	Ala17 to Thr, Ser114 to Gly	Ser79 to Arg, Asn91 to Asp, Glu125 to Gln, Glu135 to Asp	None observed	8	2	2	2	2
	18705	Hamilton, ON	0.03	4	0.5	2	0.25	0.06	None observed	Ser79 to Tyr	None observed	4	2	0	0	2
	10277	Montreal, QB	0.03	4	1	2	0.5	0.12	None observed	Ser79 to Phe	Ile460 to Val	4	2	0	0	2
	11434	Montreal, QB	1	4	0.5	2	0.06	0.03	None observed	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2	0	0	0	0
	12070	Winnipeg, MB	0.03	4	1	2	0.25	0.06	None observed	Ser79 to Tyr	Ile460 to Val	2	0	2	0	0
	12547	Montreal, QB	2	4	0.5	2	0.12	0.03	None observed	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2	0	0	0	2
	12883	Moncton, NB	0.03	4	0.5	2	0.25	0.03	None observed	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2	0	2	0	2
	13817	Montreal, QB	0.03	4	0.5	2	0.25	0.06	None observed	Ser52 to Gly, Ser79 to Tyr, Asp83 to Ala, Asn91 to Asp	Ile460 to Val	2	0	0	0	0
	14744	Montreal, QB	0.03	4	0.5	2	0.25	0.06	None observed	Ser79 to Phe	None observed	2	2	0	0	0
	12291	Montreal, QB	0.03	4	1	2	0.25	0.06	None observed	Ser79 to Tyr	None observed	2	0	0	0	0
	12292	Montreal, QB	0.03	4	0.5	2	0.12	0.03	None observed	Ser79 to Phe	Ile460 to Val	2	0	0	0	2
	15017	Hamilton, ON	0.03	4	0.25	2	0.12	0.06	None observed	None observed	Ile460 to Val	4	0	2	0	2
	1282	Calgary, AB	1	4	0.5	2	0.25	0.06	None observed	Lys137 to Asn	Ile460 to Val, Glu474 to Lys	4	0	0	2	2
	12873	Moncton, NB	0.03	4	1	2	0.25	0.06	None observed	Ser79 to Phe	Ile460 to Val	4	2	0	0	0
	16072	Winnipeg, MB	0.03	4	0.5	2	0.12	0.06	None observed	None observed	Ile460 to Val	4	0	0	0	2
	14769	Montreal, QB	0.5	8	0.25	2	0.25	0.03	None observed	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2		0	0	2
	17913	Hamilton, ON	0.03	8	0.5	2	0.25	0.06	None observed	Asp83 to Gly	None observed	4	2	0	0	2
	10733	Victoria, BC	0.03	8	4	8	2	0.5	None observed	Ser79 to Phe	None observed	2	0	2	0	2
	9989	Sherbrook, QB	0.03	8	4	8	2	0.5	Ser81 to Phe	Ser79 to Phe	Ile460 to Val	2	0	0	0	2
	10280	Montreal, QB	0.03	8	4	8	2	0.25	Ser81 to Phe	Ser79 to Phe	None observed	2	0	0	0	2
	14904	Regina, SK	2	8	4	8	4	0.5	Glu85 to Lys	Ser79 to Phe, Lys137 to Asn	Ile460 to Val	2	0	0	0	2
	12818	Calgary, AB	0.03	8	4	8	2	0.25	Ser81 to Phe	Asp83 to Tyr	Ile460 to Val	4	0	0	0	2
	19120	Vancouver, BC	0.03	16	0.5	2	0.12	0.25	None observed	Asp83 to Asn	Ile460 to Val	4	2	2	0	2
	17012	Victoria, BC	0.03	16	4	8	2	0.12	Ser81 to Phe	Ser79 to Phe	None observed	0	0	0	0	0
	18410	Toronto, ON	0.03	16	4	8	2	0.12	Ser81 to Phe	Ser79 to Phe	None observed	0	0	0	0	0
	16071	Winnipeg, MB	0.03	16	4	8	2	0.12	Ser81 to Phe	Ser79 to Tyr	None observed	2	2	0	0	2
	4030	Calgary, AB	0.03	32	4	4	2	0.25	Ser81 to Phe	Asp83 to Asn	Ile460 to Val	4	2	2	0	2
	11361	Hamilton, ON	0.03	32	8	8	4	0.5	Ser81 to Phe	Ser79 to Phe	Ile460 to Val	2	0	0	0	2
	14033	Hamilton, ON	0.03	32	8	8	4	0.5	Ser81 to Phe	Ser79 to Tyr	None observed	2	0	2	0	2
	18397	Toronto, ON	0.25	32	4	16	2	0.25	Ser81 to Phe	Ser79 to Phe	Ile460 to Val	0	0	0	0	2
	18955	Halifax, NS	0.03	32	16	32	8	1	Glu85 to Lys	Ser79 to Phe	Ile460 to Val	2	0	0	0	2
	16078	Winnipeg, MB	0.03	32	8	32	2	0.5	Ser81 to Phe	Ser79 to Phe	None observed	4	2	0	2	2

Cipro, ciprofloxacin; Gati, gatifloxacin; Gemi, gemifloxacin; Levo, levofloxacin; Moxi, moxifloxacin;

* Changes observed are relative to reference strains with known fluoroquinolone susceptibility ATCC 49619, 2587, and 2670 for GyrA and ParC and relative to 2587 and 2670 for ParE; ** a 4 fold or greater reduction in MIC in the presence of reserpine was interpreted as efflux positive, S. pneumoniae ATCC 49619, E. faecalis ATCC 29212, P. aeruginosa ATCC 27853, S. aureus ATCC 29213, and CA813, an efflux positive laboratory S. pneumoniae strain were used as controls; MB, Manitoba; QB, Quebec; BC, British Columbia; ON, Ontario; AB, Alberta; NB, New Brunswick; SK, Saskatchewan; NS, Nova Scotia.

3. Results 

The MIC values of penicillin and 5 fluoroquinolones for the 34 clinical isolates included in this study are presented in Table 1. For the purpose of data analysis, the isolates were subdivided into 2 groups: those demonstrating lower level fluoroquinolone resistance (ciprofloxacin MIC = 4 to 8 μg/mL, n = 24) and those demonstrating higher level fluoroquinolone resistance (ciprofloxacin MIC ≥16 μg/ml, n = 10) (Bast et al., 2000).

The results of the PFGE typing are summarized in a dendrogram (Figure 1). Five small clusters of isolates having greater than 95% genetic similarity were observed. Both the location of the isolates and the MICs observed among the five clusters vary greatly. Additionally, the isolates within each cluster originate from diverse geographical locations, with the exception of one cluster in which both isolates were obtained from Hamilton, Ontario. The ciprofloxacin MICs of the isolates within each cluster also vary. Only two clusters contain isolates with identical MICs, 16 μg/mL in one cluster and 4 μg/mL in the second cluster.

View Large Image Download to PowerPoint

Fig. 1. Genetic Relatedness Among 34 Fluoroquinolone Resistant Streptococcus pneumoniae Clinical Isolates

The QRDRs of gyrA, parC and parE were sequenced for each of the 34 S. pneumoniae clinical isolates. Amino acid changes resulting from base pair substitutions (determined by comparison with sequences from reference strains) are presented in Table 1. 21%, 92%, and 71% of the lower level fluoroquinolone-resistant isolates (n = 24) had mutations in gyrA, parC, and parE, respectively. 90%, 100%, and 50% of the higher level fluoroquinolone-resistant isolates (n = 10) had mutations in gyrA, parC, and parE, respectively.

The MICs of ciprofloxacin, gatifloxacin, gemifloxacin, levofloxacin, and moxifloxacin in the presence and absence of reserpine were determined by agar dilution for all 34 clinical isolates. The magnitude of the reduction in MIC observed for the isolates with the addition of reserpine is reported in Table 1. Twelve of 34 (35%) isolates were considered to be positive for reserpine-inhibited ciprofloxacin efflux based on the criteria set out previously. None of the isolates demonstrated efflux of gatifloxacin, gemifloxacin, levofloxacin or moxifloxacin.

4. Discussion 

PFGE typing demonstrated that the majority of isolates investigated were genetically unrelated. This suggests that, for the most part, the fluoroquinolone resistant isolates included in this study arose independently in different regions of the country, rather than through the dissemination of a few highly resistant isolates.

Of the 34 fluoroquinolone resistant S. pneumoniae clinical isolates investigated, those demonstrating lower level fluoroquinolone resistance were associated with mutations in parC, while those demonstrating higher level resistance were associated with mutations in both parC and gyrA. Mutations in parC were observed most frequently at positions encoding Ser79 (Ser to Phe, Tyr) and Asp83 (Asp to Asn, Ala, Gly, Tyr). GyrA mutations were observed most often at positions encoding Ser81 (Ser to Phe) and Glu85 (Glu to Lys). These data are consistent with those published by other investigators Bast et al 2000, Broskey et al 2000.

No association between mutations in parE and fluoroquinolone resistance was demonstrated. The Ile460 to Val substitution observed in 22 isolates has been previously reported by other investigators who have also been unable to assign a role for it in fluoroquinolone resistance Bast et al 2000, Broskey et al 2000. GyrB was not sequenced in this study, but gyrB mutations have not been demonstrated to be associated with fluoroquinolone resistance in S. pneumoniae Bast et al 2000, Broskey et al 2000.

Only the quinolone resistance determining region (QRDR) was sequenced in our study. Some investigators have hypothesized that mutations outside of the QRDR may play a role in fluoroquinolone resistance (Morrissey & George, 1999). The exact role of mutations outside the QRDR in the development of fluoroquinolone resistance remains to be determined. However, it is quite possible that strains 10733 and 18955 (Table 1) have mutations in regions outside the QRDR as their MIC’s to all fluoroquinolones were more significantly elevated than would be expected relative to the QRDR resistance mutations identified.

Reserpine-sensitive ciprofloxacin efflux was demonstrated in 30% (3/10) of higher level and 38% (9/24) of lower level ciprofloxacin-resistant isolates. The efflux data are in agreement with those published by Broskey et al., who demonstrated reserpine-sensitive efflux at all levels of ciprofloxacin resistance (Broskey et al., 2000).

Efflux alone was able to account for the development of clinical resistance in 3 isolates, 15017, 16072, and 1282, each having a ciprofloxacin MIC of 4 μg/mL (Table 1). It was also observed that efflux in combination with a parC mutation resulted in higher level ciprofloxacin resistance (isolate 19120). The MICs of gatifloxacin, gemifloxacin, levofloxacin, and moxifloxacin for all isolates were reduced by twofold or less in the presence of reserpine suggesting that they are poor substrates for efflux in S. pneumoniae. Beyer et al. have similarly observed reduced reserpine-sensitive efflux of new fluoroquinolones relative to ciprofloxacin in S. pneumoniae (Beyer et al., 2000). This finding may be explained on the basis of structural differences between the various fluoroquinolones Beyer et al 2000, Broskey et al 2000. In S. pneumoniae, bulkier and/or more hydrophobic fluoroquinolones are reported to be less affected by reserpine-sensitive efflux (Beyer et al., 2000). The C-7 and C-8 substituents of gatifloxacin, gemifloxacin, levofloxacin and moxifloxacin are more bulky than those of ciprofloxacin as well as being more hydrophobic molecules (Zhanel et al., 2002). Of clinical importance, those fluoroquinolones less affected by efflux may be less likely to select out resistant S. pneumoniae, by virtue of their ability to achieve higher intracellular drug concentrations (Beyer et al., 2000).

One group of investigators has hypothesized that there may exist a fluoroquinolone efflux mechanism in S. pneumoniae which is not inhibited by reserpine (Nagai et al., 2000). Such an efflux system could have a different substrate specificity than PmrA. Fluoroquinolone accumulation studies with clinical isolates may be useful in clarifying whether this theoretical mechanism is important in the development of clinical resistance.