Prevalence, Characteristics, and Molecular Epidemiology of Macrolide and Fluoroquinolone Resistance in Clinical Isolates of Streptococcus pneumoniae at Five Tertiary-Care Hospitals in Korea

Abstract

The genes erm(B), mef(A), and both erm(B) and mef(A) were identified in 42.6, 10.1, and 47.3%, respectively, of the erythromycin-resistant Streptococcus pneumoniae isolates. Of the strains, 3.8% were nonsusceptible to levofloxacin and had 1 to 6 amino acid changes in the quinolone resistance-determining region, including a new mutation, Asn94Ser, in the product of parC. Levofloxacin with reserpine was highly specific for efflux screening.

The increase in the resistance of Streptococcus pneumoniae to primary antibiotic agents has become a serious problem (1). Previous reports documented high rates of penicillin and erythromycin resistance among clinical isolates from Asia with the spread of the Taiwan19F and the Spain23F clones (19). Fluoroquinolones are now recommended for the empirical treatment of S. pneumoniae infection (9). Although increasing rates of resistance have been described in a few previous reports, the rate of fluoroquinolone resistance among S. pneumoniae isolates remains low worldwide (4, 7, 14, 15, 17). We describe the antimicrobial susceptibility of S. pneumoniae to commonly used antimicrobial agents and focus on the prevalence of genotypes and the molecular epidemiology of macrolide and fluoroquinolone resistance among isolates of S. pneumoniae from five tertiary-care hospitals in Korea.

A total of 235 isolates of S. pneumoniae were collected from five tertiary-care hospitals between 2002 and 2006. The organisms were isolated from the lower respiratory tract (68.5%), blood (14.9%), the upper respiratory tract (6.8%), the ear (3.4%), pus (2.6%), the eye (1.7%), and other sites (2.1%). Antimicrobial susceptibility testing was performed using the broth microdilution method, and results were interpreted according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (5, 6). The S. pneumoniae ATCC 49619 strain was used as a control. The bacterial genomic DNA was prepared with a genomic DNA preparation kit (Solgent Co., Daejeon, Korea). The PCRs for macrolide resistance genes and mutations in the quinolone resistance-determining region (QRDR) were performed with the primers listed in Table 1. The purified DNA was sequenced with the ABI PRISM 3130xl genetic analyzer (Applied Biosystems, Foster City, CA). The efflux phenotype was inferred from a fourfold or greater reduction in the MICs of any of the three fluoroquinolones (norfloxacin, ciprofloxacin, and levofloxacin) tested in the presence of reserpine (10 μg/ml; Sigma, St. Louis, MO). A reverse transcriptase PCR was performed to detect pmrA expression (12). Multilocus sequence typing and serotyping by Pneumotest-Latex (Statens Serum Institut, Copenhagen, Denmark) were conducted with nine levofloxacin-nonsusceptible isolates to assess clonality.

TABLE 1.

Primers of macrolide and fluoroquinolone resistance genes used in this study

Gene	Primer type	Sequence (5′ to 3′)	Reference
erm(B)	Forward	GAA AAG GTA CTC AAC CAA ATA	20
	Reverse	AGT AAC GGT ACT TAA ATT GTT TAC
mef(A)	Forward	AGT ATC ATT AAT CAC TAG TGC	20
	Reverse	TTC TTC TGG TAC TAA AAG TGG
erm(TR)	Forward	AGA AGG TTA TAA TGA AAC AGA A	16
	Reverse	GGC ATG ACA TAA ACC TTC AT
gyrA	Forward	TTC TCT ACG GAA TGA ATG	8
	Reverse	GAT ATC ACG AAG CAT TTC CAG
gyrB	Forward	TTC TCC GAT TTC CTC ATG	10
	Reverse	AGA AGG GTA CGA ATG TGG
parC	Forward	TGG GTT GAA GCC GGT TCA	8
	Reverse	CAA GAC CGT TGG TTC TTT C
parE	Forward	CCA ATC TAA GAA TCC TG	11
	Reverse	GCA ATA TAG ACA TGA CC
pmrA	Forward	TGT TCC TAA TGC AAC GGC AC	12
	Reverse	GCA TTG GCA CAG AGG AGA TA

Antimicrobial susceptibility results.

The rate of resistance to penicillin was 67.2%. The highest rates of resistance were those to erythromycin (80%) and tetracycline (81.3%) (Table 2). Nine strains (3.8%) were nonsusceptible to levofloxacin, with seven resistant and two intermediate, and 185 isolates (78.7%) were multidrug resistant.

TABLE 2.

Susceptibility of 235 isolates of S. pneumoniae collected from five tertiary-care hospitals in Korea to 12 agents

Antimicrobial agent	MIC (μg/ml)			% of isolates found to be:
	50%	90%	Range	Susceptible	Intermediate	Resistant
Penicillin	4	16	≤0.03-≥128	26.0	6.8	67.2
Amoxicillin	4	16	≤0.03-≥128	40.9	26.0	33.2
Cefotaxime	1	2	≤0.03-≥128	78.7	17.9	3.4
Ceftriaxone	1	2	≤0.03-≥128	76.2	20.4	3.4
Erythromycin	4	≥128	≤0.03-≥128	20.0	0.0	80.0
Clarithromycin	64	≥128	≤0.03-≥128	23.4	3.8	72.8
Levofloxacin	1	2	0.06-64	96.2	0.9	3.0
Moxifloxacin	0.12	0.25	≤0.03-8	97.9	0.4	1.7
Tetracycline	32	64	0.06-≥128	18.3	0.4	81.3
Trimethoprim-sulfamethoxazole	4/76	16/304	≤0.03/0.57-32/608	33.6	7.2	59.1
Linezolid	0.5	1	≤0.03-2	100.0	0.0	0.0
Vancomycin	0.25	0.25	0.06-0.5	100.0	0.0	0.0

Distribution of macrolide resistance genes and QRDR mutations.

The genes erm(B), mef(A), and both erm(B) and mef(A) were identified in 42.6, 10.1, and 47.3% of the erythromycin-resistant isolates. No erm(TR)-positive isolates were detected. Levofloxacin-nonsusceptible strains had 1 to 6 amino acid changes in the QRDRs, including a new mutation, Asn94Ser, in the product of parC (Table 3). All of the levofloxacin-nonsusceptible strains contained an Ile460Val alteration in the product of parE. Three isolates possessed only 1 amino acid substitution within the product of parE. The clonal profile in general was heterogeneous, although a few isolates may have a clonal relation.

TABLE 3.

QRDR mutations, results of multilocus sequence typing, serotypes, and efflux phenotypes characterizing nine levofloxacin-nonsusceptible S. pneumoniae isolates

Strain no.	Mutation(s) in QRDR of product of:				Levofloxacin MIC (μg/liter)	Drug(s) corresponding to efflux phenotypea	Serotype	Sequence type
	gyrA	gyrB	parC	parE
1	Glu85Lys	None	Ser79Phe	Ile460Val	32	C	19	271
2	Ser81Phe, Ser114Gly	None	Ser52Gly, Ser79Phe, Asn91Asp	Ile460Val	8	None	6	Unknown
3	None	None	Ser79Tyr	Ile460Val	64	L	19	320
4	Ser81Phe	None	Lys137Asn	Ile460Val, Asp435Asn	8	None	11	166
5	None	None	Ser79Tyr, Lys137Asn	Ile460Val	4	N, C	19	166
6	None	None	None	Ile460Val	8	C, L	19	320
7	None	None	None	Ile460Val	64	C, L	3	1374
8	None	None	None	Ile460Val	64	N, C, L	3	180
9	None	None	Asn94Ser	Ile460Val	4	N, C, L	3	180

^aN, norfloxacin; C, ciprofloxacin; L, levofloxacin. The combination of N, C, and L indicates that the strain showed an efflux phenotype with each drug in the presence of reserpine.

Reserpine efflux screening and pmrA expression.

Among the 235 isolates, 177 showed an efflux phenotype (norfloxacin, n = 134 [57.0%]; ciprofloxacin, n = 109 [46.4%]; and levofloxacin, n = 7 [3.0%]). Among nine levofloxacin-nonsusceptible isolates, an efflux phenotype corresponding to norfloxacin, ciprofloxacin, and levofloxacin was detected in three, six, and five strains, respectively. Only 2 of 226 levofloxacin-susceptible isolates exhibited the efflux phenotype when levofloxacin was used with reserpine (MIC, 2 μg/ml). All three isolates showing only a parE mutation had an efflux phenotype. Among the isolates showing an efflux phenotype, 85.4% expressed the pmrA gene.

These study results indicate high rates of resistance of S. pneumoniae to many antimicrobial agents and an uncommon increase in levofloxacin resistance in Korea. We ascertained that the presence of both erm(B) and mef(A) was highly related to multidrug resistance, as Brown et al. reported previously (3). All levofloxacin-nonsusceptible isolates had an Ile460Val alteration in the product of parE, with three isolates showing this change as the only substitution. This finding differs from results in previous reports. The high prevalence of a single Ile460Val mutation in the parE product was also described in a previous report (18), but the authors of that study reported that all except 1 of 78 isolates with a single parE mutation were susceptible to levofloxacin. We presume that the high-level levofloxacin resistance of isolates with parE-only mutations was associated with efflux pump mechanisms (see below). The numerous QRDR mutations may not have any major effect on the fluoroquinolone MICs, as noted in a previous report (3).

It is considered that for S. pneumoniae, ciprofloxacin is the drug most susceptible to efflux mechanisms (2, 7, 13, 21). In this study, efflux mechanisms were also found more commonly when efflux screening was performed with norfloxacin or ciprofloxacin than when levofloxacin was used. However, for the fluoroquinolone-nonsusceptible isolates, levofloxacin was superior to norfloxacin and similar to ciprofloxacin in detecting an efflux phenotype. Moreover, the majority of strains that were positive for an efflux phenotype active with ciprofloxacin or norfloxacin were characterized as susceptible by levofloxacin susceptibility testing. In almost all previous reports, norfloxacin- or ciprofloxacin-resistant strains were used to study the fluoroquinolone resistance mechanism of S. pneumoniae and norfloxacin or ciprofloxacin was used to study the reserpine-sensitive efflux mechanism. If the low rate of levofloxacin resistance among S. pneumoniae strains in the world is considered, there is a significant likelihood that no or very few isolates of levofloxacin-resistant S. pneumoniae were included in previous studies. Neither norfloxacin nor ciprofloxacin is recommended for respiratory infections caused by S. pneumoniae, and those drugs are not included in the CLSI guidelines for S. pneumoniae. From this result, we can infer that levofloxacin is the most appropriate and highly specific drug for efflux screening for high-level levofloxacin resistance. Brown et al. (3) thought that parE might play a greater role in high-level fluoroquinolone resistance. In the present study, we obtained similar results but found that all three isolates showing only the parE mutation had an efflux phenotype. This efflux mechanism may play a greater role in high-level fluoroquinolone resistance along with the QRDR mutations.