Molecular Basis of Resistance to Macrolides and Other Antibiotics in Commensal Viridans Group Streptococci and Gemella spp. and Transfer of Resistance Genes to Streptococcus pneumoniae

Abstract

We assessed the mechanisms of resistance to macrolide-lincosamide-streptogramin B (MLS_B) antibiotics and related antibiotics in erythromycin-resistant viridans group streptococci (n = 164) and Gemella spp. (n = 28). The macrolide resistance phenotype was predominant (59.38%); all isolates with this phenotype carried the mef(A) or mef(E) gene, with mef(E) being predominant (95.36%). The erm(B) gene was always detected in strains with constitutive and inducible MLS_B resistance and was combined with the mef(A/E) gene in 47.44% of isolates. None of the isolates carried the erm(A) subclass erm(TR), erm(A), or erm(C) genes. The mel gene was detected in all but four strains carrying the mef(A/E) gene. The tet(M) gene was found in 86.90% of tetracycline-resistant isolates and was strongly associated with the presence of the erm(B) gene. The cat_pC194 gene was detected in seven chloramphenicol-resistant Streptococcus mitis isolates, and the aph(3′)-III gene was detected in four viridans group streptococcal isolates with high-level kanamycin resistance. The intTn gene was found in all isolates with the erm(B), tet(M), aph(3′)-III, and cat_pC194 gene. The mef(E) and mel genes were successfully transferred from both groups of bacteria to Streptococcus pneumoniae R6 by transformation. Viridans group streptococci and Gemella spp. seem to be important reservoirs of resistance genes.

Viridans group streptococci (VGS) and Gemella spp. are commensal bacteria of the human upper respiratory tract, although they also cause systemic diseases, including bacterial endocarditis, bacteremia (especially in neutropenic patients), and pneumonia (12, 15, 23, 24). These bacteria can exchange genetic material with other bacteria sharing their habitat, making their resistance profiles good markers for the risk of the emergence of resistance to some antibiotics in Streptococcus pneumoniae and Streptococcus pyogenes (4).

Resistance to macrolides and related antibiotics has spread among VGS (10, 17, 27, 36, 48, 54). Two major mechanisms account for resistance to macrolide, lincosamide, and streptogramin B (MLS_B) antibiotics in streptococci: the first is mediated by methylation of the ribosomal target of these antibiotics (MLS_B resistance). The methylase responsible for this activity is encoded by the erm genes (erythromycin ribosome methylase) (26). Expression of MLS_B resistance can be constitutive (cMLS_B) or inducible (iMLS_B) (25). The streptococcal erm(B) gene is associated with conjugative transposons of the Tn916-Tn1545 family that also confer resistance to tetracycline [by the tet(M) gene] and/or kanamycin [by the aph(3′)-III gene]. These elements also contribute to the dissemination of multidrug resistance by integration into larger conjugative transposons, like Tn5253 and Tn3872, that encode additional antimicrobial resistance determinants (i.e., the cat_pC194 gene). A core element in the biology of these transposons is the integrase, encoded by the intTn gene, which is absolutely required for their transposition movements (11, 30, 39, 40).

The second major mechanism of resistance to MLS_B antibiotics is mediated by an active efflux pump, encoded by the mef(A/E) gene. The expression of the mef(A/E) gene produces resistance to 14- and 15-membered macrolide compounds, and the resulting phenotype is designated M. There are two subclasses of the mef(A/E) gene: mef(A), originally found in S. pyogenes, and mef(E), originally found in S. pneumoniae (5, 51, 53). mef(A) and mef(E) are 90% identical at the nucleotide level but are characterized by major genetic differences (9). Genetic elements carrying the mef(A) and mef(E) genes were recently characterized in S. pneumoniae. The mef(A)-carrying element is a 7.2-kb defective transposon (Tn1207.1) that contains eight open reading frames (ORFs) (46) and that appears to be part of a longer conjugative transposon, named Tn1207.3 (45). The element that contains the mef(E) gene, designated MEGA (macrolide efflux genetic assembly), is approximately 5.5 kb and contains five ORFs (16). Both elements contain an ORF adjacent to mef, designated ORF5 in Tn1207.1 and mel in MEGA, which shares approximately 35% identity with the msr(A) gene of Staphylococcus aureus.

The purpose of this study was to study the prevalence and genetic basis of resistance to MLS_B, tetracycline, chloramphenicol, and kanamycin among VGS and Gemella spp. from the human microbiota. We also investigated the elements that carry erythromycin resistance genes and their possible transfer mechanisms.

MATERIALS AND METHODS

Bacterial strains.

Between October 2001 and March 2003, 164 isolates of VGS and 28 Gemella sp. isolates resistant to erythromycin were isolated from 178 patients in the Microbiology Department of the “Lozano Blesa” Clinical University Hospital (Zaragoza, Spain). These isolates originated from pharyngeal exudates (n = 88), sputa (n = 54), bronchial aspirates (n = 17), nasal swab samples (n = 13), oral swab samples (n = 5), and other samples (n = 15). Strains were identified on the basis of colony morphology, α-hemolysis, optochin susceptibility, and Gram staining. Isolates were identified to the species level by using the API 20 Strep system (bioMerieux, Marcy-l'Etoile, France). The isolates were assigned to a group or species according to the criteria described by Facklam (13). Different isolates from the same patient were differentiated by colony morphology, species identification, antimicrobial resistance phenotype and genotype, and/or pulsed-field gel electrophoresis pattern (data not shown).

Detection of MLS_B resistance phenotypes.

Macrolide resistance phenotypes were classified as described by Seppälä et al. (47). The M phenotype was confirmed by the induction test described by Malke (29).

Antimicrobial agents.

The following antibiotics were tested: erythromycin, clindamycin, tetracycline, minocycline, and chloramphenicol (all from Sigma, St. Louis, Mo.); azithromycin, (Pfizer, Madrid, Spain); miocamycin (Menarini, S.A., Badalona, Spain); and kanamycin (Amersham Life Science, [manufactured for USB in China]).

Antimicrobial susceptibility testing.

Antimicrobial susceptibility testing was performed by a standard agar diffusion test with commercially available disks (Bio-Rad, La Coquette, France) and a standard agar dilution method according to the guidelines established by the National Committee for Clinical Laboratory Standards (NCCLS) (32, 33). S. pneumoniae ATCC 49619 was used as a control strain. The cutoff points for resistance to each antibiotic were those recommended by the NCCLS (34). The MIC breakpoint for miocamycin resistance was as defined by the Comité de l′Antibiogramme de la Societé Française de Microbiologie (Antibiogram Committee of the French Microbiology Society) (7). As there are no defined MIC breakpoints for Gemella spp., we used those for VGS, given the similarities between these two bacterial genera.

Detection of antibiotic resistance genes.

Antibiotic resistance genes were detected by PCR with oligonucleotide primers specific for each gene. DNA samples were prepared as described by Ausubel and Frederick (3). The erm(A), erm(B), erm(C), and mef(A/E) genes were amplified as reported by Sutcliffe et al. (52). The erm(A) subclass erm(TR) gene was detected as described by Seppälä et al. (49). The mel gene was detected with primers designed on the basis of the sequence of MEGA and mel in S. pneumoniae (GenBank accession no. AF274302). The forward and reverse primers were 5′-CAT GAG CGG TGG TGA AGA-3′ and 5′-TAG GGA TTT AGC GGC ATT AT-3′, respectively. Cycling conditions were 1 cycle at 94°C for 4 min; 35 cycles of 94°C for 1 min, 56°C for 1 min, and 72°C for 1 min; and 1 cycle at 72°C for 10 min. The PCR mixture contained 4 mM MgCl₂, 0.2 mM deoxynucleoside triphosphates, 0.25 mM each primer, and 0.5 U of Taq polymerase in a final volume of 20 μl. The tetracycline resistance gene tet(M) and the intTn gene were examined by established protocols (11). The cat_pC194 gene was detected with oligonucleotides catD (5′-GAA ACA TAA AAC AAG AAG GA-3′) and catR (5′-ATA GAA AGA GAA AAA GCA TT-3′), designed from the sequence of staphylococcal plasmid pC194 (GenBank accession no. NC_002013). The PCR conditions were as follows: 35 cycles of 94°C for 30 s, 46°C for 45 s, and 72°C for 2 min with 2.5 mM MgCl₂, 0.5 mM deoxynucleoside triphosphates, 1 mM each primer, and 1.75 U of Taq polymerase. The aph(3′)-III gene, which encodes the aminoglycoside-modifying enzyme, was amplified as described by van de Kludert and Wiegenthart (56).

Amplifications were performed in a DNA thermal cycler (Perkin-Elmer Cetus, Norwalk, Conn.). The PCR products were resolved by electrophoresis on 1.5% agarose gels. To discriminate between mef(A) and mef(E), the mef(A/E) amplicon was digested with BamHI. The mef(A) amplicon contains one BamHI site, so restriction generates two fragments of 282 and 64 bp, respectively, whereas the mef(E) amplicon contains no BamHI restriction sites. PCR reagents and BamHI enzyme were purchased from Promega (Madison, Wis.).

Transformation assays.

Precompetent S. pneumoniae R6 cells were prepared as described previously (18). Total genomic DNA (1 μg/ml) from strains carrying either the mef(A/E) or the erm(B) gene and competence-stimulating peptide (1 μg/ml) were added to precompetent S. pneumoniae R6 cells. Erythromycin-resistant transformants were isolated on Mueller-Hinton agar (Bio-Rad), supplemented with 5% blood (MHB) containing 2 or 8 μg of erythromycin per ml. The transformation frequency is expressed as the number of CFU of the transformants divided by the number of CFU of the recipients. The stabilities of the transformant strains were assessed by successive plating on MHB. Erythromycin resistance was determined by the agar diffusion test.

Statistical analysis.

The χ² test was used to determine whether differences in resistance rates among the isolates with different macrolide resistance phenotypes and between groups of isolates were significant.

RESULTS

Of the 164 isolates of VGS examined in this study, 125 were identified as Streptococcus mitis, 28 were identified as Streptococcus oralis, 7 were identified as Streptococcus sanguinis, 3 were identified as Streptococcus salivarius, and 1 was identified as Streptococcus anginosus. Among the Gemella spp. identified, 12 were identified as Gemella haemolysans and 16 were identified as Gemella morbillorum.

MLS_B resistance phenotypes and macrolide resistance genes.

Table 1 shows the distributions of the MLS_B phenotypes and the macrolide resistance genes among the species tested. The M phenotype was the most prevalent among the VGS (59.15%) and Gemella spp. (60.71%); all the isolates with this phenotype gave the expected 346-bp PCR product corresponding to the mef(A/E) gene. This was the only phenotype found among the S. sanguinis isolates. All M-phenotype strains were negative by the induction test. The cMLS_B phenotype was the second most prevalent. S. oralis displayed the highest percentage of strains with this phenotype (46.43%). Finally, the iMLS_B phenotype was the rarest phenotype in both VGS and Gemella spp. The erm(B) gene was detected in all strains with the cMLS_B or the iMLS_B phenotype. This gene was found in combination with the mef(A/E) gene in 45.76% of the cMLS_B isolates and 52.63% of the iMLS_B isolates. None of the isolates contained the erm(A) subclass erm(TR), erm(A), or erm(C) gene. The amplicons of mef(A/E) were subjected to restriction analysis to differentiate between mef(A) and mef(E). The mef(E) gene was predominant in VGS (95.42%) and gemellae (95.00%). Only six VGS isolates (three S. mitis isolates, two S. oralis isolates, and one S. sanguinis isolate) and one G. haemolysans isolate carried the mef(A) subclass. The mel gene, ORF2 of MEGA, was found in all mef(A/E)-containing Gemella sp. isolates and 96.95% of the VGS strains containing mef(A/E). The four mef(A/E)-containing VGS that did not carry this ORF included one S. mitis isolate and three S. oralis isolates.

TABLE 1.

Distributions of VGS and Gemella sp. strains according to macrolide resistance phenotypes and genotypes

Bacterium	Total no. of isolates	No. (%) of M-phenotype isolatesa	cMLSB-phenotype isolatesb		iMLSB-phenotype isolatesb
			No. (%) of isolates	No. with mef(A/E)	No. (%) of isolates	No. with mef(A/E)
VGS	164	97 (59.15)	50 (30.49)	24	17 (10.37)	10
S. mitis	125	76 (60.80)	35 (28.00)	15	14 (11.20)	8
S. oralis	28	28 (46.43)	13 (46.43)	8	2 (7.14)	1
S. sanguinis	7	7 (100.00)
S. salivarius	3	1 (33.33)	1 (33.33)	1	1 (33.33)	1
S. anginosus	1		1 (100.00)
Gemella spp.	28	17 (60.71)	9 (32.14)	3	2 (7.14)
G. haemolysans	12	9 (75.00)	3 (25.00)	2
G. morbillorum	16	8 (50.00)	6 (37.50)	1	2 (12.50)

^aAll strains with the M phenotype harbored the mef(A/E) gene.

^bAll strains with the cMLS_B or iMLS_B phenotype harbored the erm(B) gene.

Non-MLS_B resistance genes tet(M), cat_pC194, and aph(3′)-ITIII.

Tetracycline resistance (including intermediate resistance) was the most common non-MLS_B resistance found in VGS (41.46%) as well as in Gemella spp. (57.14%); it was mainly found in isolates with the cMLS_B and iMLS_B phenotypes (P < 0.0001) (Table 2). In all cases, resistance to tetracycline was associated with resistance to minocycline. Among the species tested, S. sanguinis, the isolates of which displayed only the M phenotype, showed the lowest rates of resistance to these antibiotics. The tet(M) gene was detected in a large proportion (86.90%) of tetracycline-resistant strains, particularly in those with the erm(B) gene (cMLS_B and iMLS_B phenotypes) (P < 0.005) (Table 2).

TABLE 2.

Resistance to tetracycline correlated with the presence of the tet(M) gene in VGS and Gemella spp. with different macrolide resistance phenotypes

Bacterium	M-phenotype isolates		cMLSB-phenotype isolates		iMLSB-phenotype isolates
	No. (%) Tetra	No. with tet(M)	No. (%) Tetr	No. with tet(M)	No. (%) Tetr	No. with tet(M)
VGS	13 (13.40)	7	40 (80.00)	35	15 (88.24)	15
S. mitis	10 (13.16)	7	30 (85.71)	26	12 (85.75)	12
S. oralis	1 (7.69)		9 (69.23)	8	2 (100.00)	2
S. sanguinis	1 (14.29)
S. salivarius	1 (100.00)				1 (100.00)	1
S. anginosus			1 (100.00)	1
Gemella spp.	4 (14.29)	4	8 (88.89)	8	2 (100.00)	2
G. haemolysans	1 (8.33)	1	2 (66.67)	2
G. morbillorum	3 (18.75)	3	8 (75.00)	8	2 (100.00)	2

^aTet^r, tetracycline resistant.

Only 4.27% of VGS isolates were resistant (including intermediately resistant) to chloramphenicol. There were no significant differences between the three macrolide resistance phenotypes (P > 0.5). The seven chloramphenicol-resistant strains were all S. mitis, and all of them carried the cat_pC194 gene. None of the Gemella sp. isolates were resistant to chloramphenicol.

High-level kanamycin resistance was found in two S. mitis isolates and two S. oralis isolates. The aph(3′)-III gene was detected only in the isolates that also harbored the erm(B) gene (one S. mitis isolate and two S. oralis isolates).

In one strain with the mef(A/E) gene (M phenotype), the tet(M) gene was associated with the cat_pC194 gene, whereas three isolates harboring the erm(B) gene showed this combination. The erm(B), tet(M), and aph(3′)-III genes were found together in one S. mitis isolate with the iMLS_B phenotype.

intTn gene.

The intTn integrase gene, from conjugative transposons of the Tn916-Tn1545 type and from Tn5253 and Tn3872, was detected in all the isolates with one or more of the following genes: erm(B), tet(M), aph(3′)-III, and cat_pC194.

Transformation of macrolide resistance genes.

Transformation experiments were performed with S. pneumoniae R6 as the recipient. The mef(E) gene could be transferred from different species of VGS and from Gemella spp., always in association with the mel gene. Although 38 erm(B)-carrying strains were used as DNA donors, only strain 208 could transfer it to S. pneumoniae R6, and this occurred at a frequency of 3.3 × 10⁻⁹ (Table 3). All the transformants were stable, as erythromycin resistance remained after 10 consecutive platings and after 1 year of storage at −80°C.

TABLE 3.

Transfer by transformation of erythromycin resistance determinants from VGS to S. pneumoniae R6

Donor	Donor phenotype	Donor gene(s)	Frequency of transformation	Transformant phenotype	Transformant gene(s)
S. mitis 130	M	mef(E), mel	1.5 × 10−7	M	mef(E), mel
S. mitis 143	M	mef(E), mel	3.0 × 10−7	M	mef(E), mel
S. mitis 147	M	mef(E), mel	7.8 × 10−7	M	mef(E), mel
S. mitis 183	M	mef(E), mel	5.3 × 10−7	M	mef(E), mel
S. mitis 220	M	mef(E), mel	1.1 × 10−7	M	mef(E), mel
S. mitis 208	cMLSB	erm(B)	3.3 × 10−9	cMLSB	erm(B)
S. oralis 47	M	mef(E), mel	6.7 × 10−7	M	mef(E), mel
S. oralis 288	M	mef(E), mel	1.7 × 10−6	M	mef(E), mel
S. oralis 296	M	mef(E), mel	2.1 × 10−7	M	mef(E), mel
S. salivarius 137	iMLSB	erm(B), mef(E), mel	1.1 × 10−7	M	mef(E), mel
S. salivarius 172	cMLSB	erm(B), mef(E), mel	2.0 × 10−6	M	mef(E), mel
G. haemolysans 70	M	mef(E), mel	1.7 × 10−7	M	mef(E), mel
G. haemolysans 78	M	mef(E), mel	1.9 × 10−6	M	mef(E), mel
G. haemolysans 79	M	mef(E), mel	6.3 × 10−6	M	mef(E), mel
G. morbillorum 95	M	mef(E), mel	1.7 × 10−7	M	mef(E), mel
G. morbillorum 256	M	mef(E), mel	8.9 × 10−8	M	mef(E), mel

DISCUSSION

The M phenotype was predominant (59.38%) among the erythromycin-resistant VGS and Gemella spp. analyzed here, whereas the iMLS_B phenotype was the rarest (9.90%). Other investigators have also shown that the M phenotype is predominant in VGS from the oropharynx (1, 21, 48). However, the cMLS_B phenotype is the one the most commonly encountered among VGS isolated from the bloodstream (2, 8, 43, 58). In this study, only S. sanguinis displayed the M phenotype, in agreement with the results of Rodriguez-Avial et al. (42), who found a statistically significant (P < 0.01) prevalence of the M phenotype among blood isolates of this bacterial group. PCR amplification of macrolide resistance genes was performed to clarify the mechanisms of resistance. Among our strains, the erm(B) gene was always detected in isolates with the cMLS_B or the iMLS_B phenotype, as reported by other investigators (20, 38, 43, 48). However, the erm(A) subclass erm(TR), erm(A), and erm(C) genes were not detected. The only description of the erm(A) subclass erm(TR) gene in VGS was in clinical isolates of S. anginosus by Jacobs et al. (22). As far as we know, the erm(A) and erm(C) genes have never been detected in VGS or Gemella spp. The mef(A/E) gene was found in all the strains with the M phenotype and 45.76 and 52.63% of strains with the cMLS_B and iMLS_B phenotypes, respectively, with mef(E) being the predominant subclass (95.36%). Although the mef(A/E) gene has been described in erythromycin-resistant VGS with the three macrolide resistance phenotypes (20, 22, 27, 38, 43, 48), most investigators did not differentiate between the two subclasses of the mef(A/E) gene. Only Arpin et al. (2) reported on the prevalence of mef(E) in clinical isolates of VGS, and they obtained results similar to ours. It is important to differentiate between the two subclasses of mef(A/E) because of the genetic differences between them (9) and the information that it gives about the possible origins of these genes and gene transfer with other bacteria. Recently, Gay and Stephens (16) and Santagati et al. (46) described two genetic elements, MEGA and Tn1207.1, carrying mef(E) and mef(A), respectively. Both elements possess an ORF (mel in MEGA) adjacent to the mef gene that shares approximately 35% identity with the msr(A) gene from staphylococci. The mel gene was amplified by PCR to localize the mef(A/E) gene in our strains. The mel gene was detected in all but four isolates, suggesting that these strains have elements similar to MEGA or Tn1207.1. The absence of this ORF from the four VGS could be due to the lack of mel or the complete elements or to the presence of another mel-like gene with major nucleotide sequence differences. We are performing experiments to determine which of these hypotheses is true. Only a few studies have looked at the carriage of mef(A) and mef(E) elements. The first description of such elements was in S. pneumoniae, which remains the most studied bacterium (16, 46). However, Luna et al. (28) performed a broad study with gram-positive and gram-negative bacteria and detected the eight ORFs of Tn1207.1 in seven Streptococcus spp. and two Enterococcus spp. That group detected only ORF5 to ORF8 in four Staphylococcus intermedius isolates.

Conjugative transposons from the Tn916 family and large composite structures like Tn5253 and Tn3872 are found in different species of the genus Streptococcus (11, 40, 44). Tn916 encodes resistance to tetracycline [tet(M)] alone, but some of the other elements mentioned carry other resistance determinants, such as the cat_pC194, aph(3′)-III, and erm(B) genes (40). To detect these conjugative transposons, we amplified the cat_pC194, aph(3′)-III, and erm(B) resistance genes and the intTn gene, which encodes a protein responsible for transposition of Tn916-like elements and composite transposons. Tetracycline was the drug to which resistance was the most frequently encountered among our isolates, and tetracycline resistance had a strong association (P < 0.0001) with the erm(B) gene. The same was true for the tet(M) gene (P < 0.005), which was found in 79.49% of the erm(B) isolates. Rodriguez-Avial et al. (43) also described this association. They found that 75.80% of VGS isolated from blood carried both genes. Clermont and Horaud (6) detected the tet(M) gene in clinical S. anginosus isolates that also carried chromosomal elements similar to Tn916. Olsvik et al. (35) reported on two tetracycline-resistant G. morbillorum strains carrying the tet(M) gene. The tet(M) gene appears to be widespread in erythromycin-resistant VGS and Gemella spp., as is the case in pneumococci (39, 50). The cat_pC194 gene is believed to be the main gene responsible for chloramphenicol resistance in pneumococci, whereas other classes of cat genes are more prevalent in streptococci of groups A, B, and G (55, 57). We detected the cat_pC194 gene in all chloramphenicol-resistant VGS, but we found no significant association with the erm(B) gene, as reported in pneumococci by Seral et al. (50). To our knowledge, this is the first description of the cat_pC194 gene in S. mitis. The aph(3′)-III gene was detected in two high-level kanamycin-resistant isolates of VGS (two S. oralis isolates and one S. mitis isolate) harboring the erm(B) gene. Although some studies have investigated high-level aminoglycoside resistance in VGS (6, 14, 19), the aph(3′)-III gene has been described in only six kanamycin-resistant S. anginosus isolates (6).

The association between the tet(M), erm(B), aph(3′)-III, and/or cat_pC194 genes with the intTn gene suggests the presence of elements similar to the Tn916 family of conjugative transposons and composite elements like Tn3872 and Tn5253. It also demonstrates that resistance to tetracycline, erythromycin, kanamycin, and chloramphenicol in VGS and Gemella spp. could be due to the acquisition of these highly mobile elements. In this sense, filter mating experiments have demonstrated the conjugation of tet(M) and erm(B) determinants from oral streptococci to Enterococcus faecalis JH2-2 and also the transfer of a Tn916 native conjugative transposon from the same bacteria to Streptococcus parasanguinis in an oral biofilm model (6, 41).

The transfer of the mef(E) determinant by transformation clearly indicates that gene transfer can occur between VGS or Gemella spp. and S. pneumoniae, as is the case for genes encoding penicillin-binding proteins (37). In pneumococci, the mef(E) gene is part of a genetic insertion element that lacks the genes necessary for transposition and is inserted at various chromosomal sites (16). As mentioned above, this element could be found in our mef(E)-positive isolates; therefore, transformation probably plays an important role in the transfer of this resistance determinant. The fact that mef(E) and mel were always transferred together suggests that mel could be involved in the efflux mechanism of erythromycin resistance, as suggested by Gay and Stephens (16). However, our results suggest that transformation does not play an important role in the transfer of the erm(B) gene, because only 1 of the 38 erm(B)-positive isolates used as donors could transfer this gene to S. pneumoniae R6, and transfer occurred at a very low frequency. Furthermore, both S. salivarius isolates that possessed the erm(B) and mef(E) genes were able to transfer only the latter gene. The erm(B) gene is located in conjugative transposons from the Tn916 family, meaning that conjugation is the major means by which it is transferred (6, 40, 44). However, the main reason why mef(E) gene transfer was more efficient than erm(B) gene transfer could be differences in the percent identities with the host DNA. The critical step of transformation is the integration of the transforming DNA into the bacterial chromosome. Homologous recombination can occur only if the introduced DNA and the host DNA share a minimum of 70% nucleotide sequence identity. Larger allelic differences can be incorporated if substantial amounts of flanking identity are present on both sides of a heterologous marker (31). The efficiency of transformation varies with the length of identity available for recombination; therefore, it seems probable that the mef(E)-mef(E) element shares a higher degree of identity with DNA host than the erm(B)-erm(B) transposon.

The nasopharynx is one of the environments most colonized in humans. While in the nasopharynx, VGS and Gemellae spp. come into contact with other bacteria, such as S. pneumoniae and S. pyogenes. This work demonstrates that VGS and Gemella spp. are important reservoirs of genes conferring resistance to macrolides and related antibiotics. The genetic pattern found in VGS and Gemella spp. [i.e., the presence of the erm(B) gene; the absence of the erm(A) subclass erm(TR), erm(A), and erm(C) determinants; and the prevalence of the mef(E) subclass], together with evidence of the presence of Tn916-related transposons and the transfer of the mef(E) by transformation, suggests that these groups of bacteria can exchange genetic information, especially with S. pneumoniae. These results emphasize the need to monitor the epidemiology and genetic basis of antibiotic resistance in VGS and Gemella spp. from the normal flora, not only because they are reservoirs of antibiotic resistance genes but also because of the serious invasive diseases that they can cause.