Erythromycin-Resistant Pharyngeal Isolates of Streptococcus pyogenes Recovered in Italy

Abstract

Three classes of macrolide resistance phenotypes and three different erythromycin resistance determinants were found among 127 erythromycin-resistant group A streptococcal (GAS) isolates recovered from 355 (35.8%) pediatric pharyngitis patients in Rome, Italy. According to emm and sof sequence typing results, erythromycin-resistant isolates comprised 11 different clonal types. Remarkably, 126 of the 127 macrolide-resistant isolates were serum opacity factor (sof) gene positive. These data suggest a strong association between macrolide resistance and the presence of sof among GAS isolates recovered from Italian pediatric pharyngitis patients.

Group A streptococci (GAS) have remained susceptible to benzylpenicillin; however, macrolides are preferred by some physicians (2, 39). Recently, a dramatic increase in macrolide-resistant GAS has been documented in several countries, including Italy (1, 2, 9, 10).

In GAS, macrolide resistance is conferred by the 23S rRNA methylase genes erm(B) (7) and erm(TR) (34) [now considered a variant of erm(A) (31)], as well as by the efflux determinant mef(A) (8, 36). These erm genes often confer inducible resistance to macrolides, lincosamides, and streptogramin B (iMLS_B resistance), but also may encode constitutive MLS_B (cMLS_B) resistance due to upstream attenuator sequence alterations. mef(A) encodes resistance to erythromycin (and other 14- or 15-member ring macrolides), while the organism remains susceptible to clindamycin (and other lincosamides) and streptogramin B (8).

Previous pulsed-field gel electrophoresis (PFGE) work demonstrated that macrolide-resistant GAS isolates in Italy are polyclonal, with a few clones of undefined strain types predominating (17, 30). This study identifies the macrolide resistance genes and strain types, as defined by variable gene sequence typing (VGST), of erythromycin-resistant GAS strains in Rome, Italy.

The isolates consisted of two groups. One hundred nine pediatric pharyngitis GAS isolates were collected in spring 2000 at Ospedale Bambino Gesù in Rome and genotyped by PFGE and VGST (of emm and sof), as described previously (13). In the present study, these isolates were tested for macrolide resistance and were found to consist of 41 macrolide-resistant isolates and 68 macrolide-sensitive isolates (Table 1). A second group of 86 erythromycin-resistant GAS isolates were screened from 246 consecutive pharyngitis isolates at the same hospital during October to December 2000. Thus, a total of 127 macrolide-resistant isolates (Table 2) were found among 355 isolates. This resistance frequency (35.7%) is consistent with previous results in Italy (1, 2, 9, 10, 32).

TABLE 1.

Distribution of erythromycin resistance and sensitivity among 109 pharyngeal GAS isolates representing 21 different genotypes^a

Sequence type	No. of isolates with erythromycin:		Total
	Resistance	Sensitivity
emm12/sof12	18	10	28
emm89/sof89	8	4	12
emm4/sof4	5	6	11
emm28/sof28	3	3	6
emm77/sof77	3	0	3
emm2/sof2	2	0	2
emm22/sof22	1	0	1
emm5.3/sof−	1	3	4
emm1/sof−	0	11	11
emm87/sof87	0	7	7
emm29.1/sof−	0	5	5
emm6/sof−	0	5	5
emm11/sof11	0	4	4
emm3.1/sof−	0	3	3
emm44-61/sof44	0	1	1
emm75/sof75	0	1	1
emm59/sof59	0	1	1
emm78/sof78	0	1	1
emm118/sof118	0	1	1
emm80.1/sof−	0	1	1
emm14.2/sof−	0	1	1
Total	41	68	109

^aThe 41 macrolide-resistant isolates included here are also represented in the results shown in Table 2.

TABLE 2.

Distribution of sequence types (emm and sof), macrolide resistance genes, and macrolide resistance phenotypes among 127 erythromycin-resistant GAS isolates isolated in Italy

Sequence type	No. of isolates with resistancea
	erm(B)/MLSB	erm(B)/iMLSB	mef(A)/M	erm(A)/iMLSB	erm(B)/iMLSB + mef(A)/M	Total
emm12/sof12	21		13			34
emm89/sof89		19			1	20
emm77/sof77				16		16
emm4/sof4			14			14
emm28/sof28	3	10	1			14
emm2/sof2			13			13
emm22/sof22	6	2				8
emm44-61/sof44	2					2
emm75/sof75	1		3			2
emm9/sof9	1					1
emm5.3/sof−		1				1
Total	34	32	44	16	1	127

^aMLS_B and iMLS_B represent erythromycin resistance and inducible resistance, respectively, to macrolide, lincosamide, and streptogramin B antibiotics. M represents macrolide resistance with sensitivity to lincosamide and streptogramin B antibiotics.

Sensitivity testing was done with the Etest (AB BIODISK, Solna, Sweden) according to the manufacturer's instructions and was verified by disk diffusion according to the National Committee for Clinical Laboratory Standards (NCCLS) guidelines (25). The erythromycin-clindamycin double-disk test was applied as described previously (17) by a modification of a previous method (33) employing disks of erythromycin (30 μg) and clindamycin (10 μg) (Oxoid, Ltd., Basingstoke, United Kingdom). If there was no inhibition zone around both disks, the result was simply interpreted as resistance (MLS), since this method may not detect inducibility by clindamycin. Blunting of the inhibition zone around the clindamycin disk proximal to the erythromycin disk was considered inducible resistance (iMLS), and susceptibility to clindamycin with no blunting in the inhibition zone proximal to the erythromycin disk indicated the M phenotype.

A multiplex PCR employing AmpliTaq Gold (Perkin-Elmer, Roche Molecular Systems, Branchburg, N.J.) was used to amplify regions of erm(B), erm(A)/erm(TR), and mef(A) as described previously (17, 34, 35). emm and sof sequence types were obtained as described previously (3, 4) for the 109 isolates depicted in Table 1 (13), as well as an additional 86 erythromycin-resistant isolates (in Table 2, including 41 resistant isolates from Table 1).

The macrolide resistance phenotypes, resistance determinants, and VGST results are shown in Table 2. VGST (13) and erythromycin susceptibility (this work) of the 109-isolate set consisting of macrolide-resistant and -sensitive isolates are shown in Table 1. These 109 isolates consisted of 79 sof (serum opacity factor gene)-positive isolates (14 M protein gene [emm] types) and 30 sof-negative isolates (7 emm types). The 41 macrolide-resistant isolates represented 8 different emm sequence types, 7 of which are typically sof gene positive (4, 11, 20, 22, 29). While 40 of the 79 sof-positive isolates (50.6%) were macrolide-resistant, only 1 of the 30 (3.3%) sof-negative isolates was resistant.

Eleven emm types of the 22 found accounted for all macrolide-resistant isolates (Table 2). The majority of these isolates were within types emm12 [21 with MLS_B resistance and 13 containing mef(A)], emm89 [20 with iMLS_B resistance, one of which also contained mef(A)], emm28 (10 with iMLS_B and 3 with MLS_B resistance), and emm22 (6 with MLS_B and 2 with iMLS_B resistance).

The erm(B) determinant was found in 66 of the 127 erythromycin-resistant isolates, and about half were clearly iMLS_B (Table 2). Forty of the 44 erythromycin-resistant, mef(A)-positive isolates that were clindamycin sensitive occurred within types emm4, emm12, and emm2. Inducible MLS_B resistance conferred by erm(A)/erm(TR) was found only within type emm77 (16 isolates).

Five types (emm12, emm89, emm4, emm28, and emm5.3) consisted of both macrolide-resistant and macrolide-sensitive isolates among the set of 109 isolates comprised of both resistant and sensitive isolates (Table 1). All 68 erythromycin-sensitive isolates, representing 18 emm types, were PCR negative for mef(A), erm(B), and erm(A)/erm(TR). Thirteen of these 18 emm types consisted entirely of macrolide-sensitive isolates (42 isolates).

Type emm89, emm77, emm4, emm44/61, emm11, and emm118, isolates infrequently contain sof genes that diverge extensively from the emm type reference strains and encode serologically distinct opacity factor proteins (4). The sof sequences from the isolates within the types investigated here were 98 to 100% identical to the corresponding sof sequences from the corresponding CDC emm type reference strains. The classical serotype M44 and M61 reference strains share the same emm sequence and M serotypes, yet contain dissimilar sof genes (4). The single type emm44/61 isolate described here is of the strain M44 lineage, since it contains the sof44 sequence.

These data are consistent with previous results indicating that macrolide resistance occurs at a high frequency (about 35%) in GAS pharyngitis isolates recovered in Italy (1, 2, 9, 10, 32). In this study, 64 to 100% of isolates within the common types emm12, emm77, and emm89 were macrolide resistant (Table 1). In contrast, population-based analysis of 496 pharyngitis GAS isolates during 1998 in Quebec, Canada, revealed an incidence of erythromycin resistance of 4.6%, with no resistance detected within types M12, M77, and M89 isolates (40). The majority of the resistant isolates were type M28, all of which harbored erm(A)/erm(TR). Again, this result differs from our results, in which all 14 type emm28 macrolide-resistant isolates carried either erm(B) or mef(A) (Table 2). There has been a 100% correlation between the M28 serotype and the sequence type emm28 from isolates recovered from various countries (2, 4), and based on PFGE (G. Gherardi and B. Beall, unpublished data) and multilocus sequencing data (13, 14), these isolates appear to constitute a clone. From the accumulated data, it is clear that there are distinct geographic differences in both the frequency and the genetic basis of macrolide resistance within individual clonal types.

We found that macrolide resistance within our sample set of 127 resistant isolates, with the exception of a single erm(B)-positive type emm5.3 isolate, was restricted to sof-positive strains (Table 2) which have been estimated to comprise about 65% of invasive and noninvasive GAS isolates recovered in the United States (B. Beall, unpublished data). Analysis of the published literature concerning macrolide resistance in GAS revealed similar results in Chile, Canada, Finland, Germany, Spain, and Sweden in that each study primarily described macrolide resistance in sof-positive types (including types M2, M4, M11, M12, M22, M28, M58, M73, M75, and M77) (6, 12, 19, 21, 27, 28, 40). (For a compilation of sof PCR-positive types and opacity factor-positive types, see references 4 and 20.) These studies along with our work indicate that macrolide resistance is widespread among sof-positive GAS strains of many different clonal types and that resistant GAS within types emm2, emm4, emm12, emm28, emm75, and emm77 are geographically widely distributed. (Although emm12 isolates are opacity factor negative [20], they are sof PCR positive [4]). However, in at least one instance, a localized outbreak of pharyngitis within the United States has been due to a sof-negative, erythromcin-resistant type emm6 strain (23). To our knowledge, our work is the first documentation of macrolide resistance within types emm89, emm44/61, emm9, and emm5 (emm5.3).

Sof is a fibronectin-binding protein that appears to contribute to virulence (11). Differences in binding specificities could effect physical separation of the two broad divisions (sof negative and sof positive) of GAS strains, preventing horizontal transfer events (such as transfers of macrolide resistance determinants) between them. There is circumstantial evidence of natural barriers between sof-negative and sof-positive GAS strains preventing recombination at the mga locus. emm sequences are generally phylogenetically divisible into sof-positive and -negative types (16, 38, 41), even though studies of mga loci have revealed mosaic structures of emm and emm-like genes indicative of intraspecies horizontal recombination events (5, 42).

One study found that the proportion of prtF1-containing strains was much higher among erythromycin-resistant isolates than within susceptible isolates (15). The prtF1 gene encodes another fibronectin-binding protein (37) that appears to be required for efficient entry of GAS into epithelial cells (24). Although limited data indicate that prtF1 is most often present among sof-positive strains (serotypes M28, M62, M12, M49, M9, and M2 and sof sequence type sof75) (22, 26) and may be absent in type emm1 isolates (26), prtF1 has also been found among the sof-negative serotypes M3 and M6 (18, 26). It will be interesting to determine the association between prtF, sof, and other genes encoding extracellular matrix binding proteins with erythromycin resistance in GAS.