Expression of the mef(E) Gene Encoding the Macrolide Efflux Pump Protein Increases in Streptococcus pneumoniae with Increasing Resistance to Macrolides

Abstract

Active macrolide efflux is a major mechanism of macrolide resistance in Streptococcus pneumoniae in many parts of the world, especially North America. In Canada, this active macrolide efflux in S. pneumoniae is predominantly due to acquisition of the mef(E) gene. In the present study, we assessed the mef(E) gene sequence as well as mef(E) expression in variety of low- and high-level macrolide-resistant, clindamycin-susceptible (M-phenotype) S. pneumoniae isolates (erythromycin MICs, 1 to 32 μg/ml; clindamycin MICs, ≤0.25 μg/ml). Southern blot hybridization with mef(E) probe and EcoRI digestion and relative real-time reverse transcription-PCR were performed to study the mef(E) gene copy number and expression. Induction of mef(E) expression was analyzed by Etest susceptibility testing pre- and postincubation with subinhibitory concentrations of erythromycin, clarithromycin, azithromycin, telithromycin, and clindamycin. The macrolide efflux gene, mef(E), was shown to be a single-copy gene in all 23 clinical S. pneumoniae isolates tested, and expression post-macrolide induction increased 4-, 6-, 20-, and 200-fold in isolates with increasing macrolide resistance (erythromycin MICs 2, 4, 8, and 32 μg/ml, respectively). Sequencing analysis of the macrolide efflux genetic assembly (mega) revealed that mef(E) had a 16-bp deletion 153 bp upstream of the putative start codon in all 23 isolates. A 119-bp intergenic region between mef(E) and mel was sequenced, and a 99-bp deletion was found in 11 of the 23 M-phenotype S. pneumoniae isolates compared to the published mega sequence. However, the mef(E) gene was fully conserved among both high- and low-level macrolide-resistant isolates. In conclusion, increased expression of mef(E) is associated with higher levels of macrolide resistance in macrolide-resistant S. pneumoniae.

Active macrolide efflux is a major mechanism of macrolide resistance in Streptococcus pneumoniae in many parts of the world, especially North America (23, 24, 27). It confers low-level resistance (MIC, 1 to 16 μg/ml) to 14- and 15-member macrolides but not to 16-member macrolides, lincosamides, and streptogramin B and is phenotypically referred to as M-type resistance, in contrast to the macrolide-lincosamide-streptogramin B phenotype, which confers constitutive high-level resistance (MIC, ≥256 μg/ml) to 14-, 15-, and 16-member macrolides, licosamides, and streptogramin B (23, 24). This active macrolide efflux in S. pneumoniae is due to acquisition of the mef(A) gene, originally described in Streptococcus pyogenes and then identified as the mef(E) gene in S. pneumoniae (3, 24). The two genes, originally grouped into one mef(A) class based on their high (90%) sequence homology (21), are considered as separate entities because it has been shown that a number of marked differences exist between them (11, 12). For instance, the genetic elements carrying mef(A) or mef(E) have been studied by Santagati et al., Gay and Stephens, and Del Grosso et al. and were shown not only to be quite different but also to behave quite differently (6, 8, 19, 22). The antibiotic susceptibility profiles of the mef(A)-carrying and mef(E)-carrying S. pneumoniae isolates have been shown to be significantly different, suggesting that although the genes are highly conserved, the elements containing them confer significantly different characteristics on the strains carrying them (6). In addition, the two genes have disseminated differently and are being recognized in an ever-growing number of microbial species (11). Presently, both the mef(A) and the mef(E) genes have unambiguously been identified in five streptococcal species, whereas mef(E) has been identified in five more streptococcal species and in nine additional nonstreptococcal species (11).

Recently, we studied the prevalence of mef(A) and mef(E) genes in macrolide-resistant S. pneumoniae isolates obtained from all regions of Canada and we identified mef(E) carried by the macrolide efflux genetic assembly (mega) element as the predominant gene responsible for macrolide efflux (24). Although uncommon, mef(A) was also present in Canadian macrolide-resistant S. pneumoniae (25). In the same study, in contrast to the mef(A)-carrying isolates, which were shown to be genetically related by both pulsed-field gel electrophoresis and serotype, the mef(E)-carrying isolates were shown to be genetically unrelated and belonged to many different serotypes, consistent with the findings of others (25). Interestingly, the macrolide resistance level of the mef(E)-carrying S. pneumoniae isolates, although low for the majority of isolates (MIC₉₀, 4 μg/ml), varied from 1 μg/ml to 32 μg/ml and this broad range of macrolide MIC distributions has been reported but never studied before by other investigators (10).

In the present study, we investigated further the properties of the mef(E) gene in S. pneumoniae. Specifically, we assessed the mef(E) gene sequence and mef(E) expression and induction in isolates with various degrees of resistance to macrolide antibiotics. We also studied the mega element upstream sequence, the potential promoter region, and sequence downstream of the efflux pump gene, assessing the association between mef(E) and mel—the msrA-like homolog believed to be a part of the macrolide efflux system.

MATERIALS AND METHODS

Streptococcus pneumoniae isolates.

Twenty-three macrolide-resistant (erythromycin MIC, 1 μg/ml to 32 μg/ml) and clindamycin-susceptible (MIC, ≤0.25 μg/ml) (M-phenotype) S. pneumoniae clinical isolates were selected from among more than 3,000 isolates collected between 1997 and 1999 as part of an ongoing annual national surveillance study (Canadian Respiratory Organism Susceptibility Study) (9). The isolates were selected to represent a variety of low- and high-level macrolide-resistant mef(E) strains, representing all regions of Canada. Isolates were also chosen to contain a variety of serotypes and different pulsed-field gel electrophoresis profiles.

Antibiotic susceptibility.

Erythromycin, clarithromycin, azithromycin, and clindamycin susceptibilities were determined using the Clinical and Laboratory Standards Institute (CLSI; formerly NCCLS) M7-A6 broth microdilution method (4). MIC interpretive standards for macrolides and clindamycin were defined according to the CLSI breakpoints (4).

DNA isolation.

Genomic DNA was prepared by sequential incubations of the cell suspensions grown overnight in Todd-Hewitt broth supplemented with 0.5% yeast extract in lysozyme (15 min at 37°C), 10% sodium dodecyl sulfate (5 min at 65°C), phenol-chloroform-isoamyl alcohol purification (25:24:1), ethanol, or isopropanol precipitation.

PCR-based detection of mef(E).

The presence of the mef gene was determined by a previously described PCR assay that did not distinguish between the two variants (6). Discrimination between mef(A) and mef(E) was performed by PCR-restriction fragment length polymorphism analysis according to a previously described protocol, and selected isolates confirmed to carry the mef(E) gene were included in further study (6). Nucleotide primers designed based on mega sequence (GenBank accession no. AF274302) that were used in PCR and DNA sequencing are shown in Table 1. Primer set mef(E)-UP and orf1-UP were used to define the 600-bp upstream (promoter) region of the mef(E) gene. mef(E)-UP2 was paired with mef(E)-DN2 to amplify the mef(E) gene, resulting in a 1,480-bp product. PCR amplification consisted of 35 cycles at 95°C, 54°C, and 72°C for 1 min each using a Perkin-Elmer GeneAmp PCR system; Applied Biosystems). Each reaction was carried out in 50 μl of mix containing 25 μl of 2× master mix (2× GeneAmp PCR buffer II [Applied Biosystems], 6 mM MgCl₂ [Applied Biosystems], 0.4 mM deoxynucleoside triphosphates [dNTPs; Invitrogen]), 0.5 μM primers, 2.5 U of AmpliTaq Gold DNA polymerase, and 100 ng of DNA template.

TABLE 1.

Oligonucleotide primers used for PCR, DNA sequencing, and RT-PCR

Primer	Primer sequence (5′→3′)	Position in mega 5′→3′	Reference
mef-1	AGTATCATTAATCACTAGTGC	1181-1201	5, 6
mef-2	TTCTTCTGGTACTAAAAGTGG	1526-1506	5, 6
mef(E)-UP	ATGGCACTAGTGATTAATG	1205-1186	mega (7, 8)
orf1-UP	TGAGGTTGAGTTAGAAAATCC	562-582	mega (7, 8)
mef(E)-UP2	GCCTATAATGCTATTCAAAAT	1066-1086	mega (7, 8)
mef(E)-DN	TACTAAACCAATACGGTCATA	2569-2548	mega (7, 8)
mef(E)-RT1	AGCTACCTGTCTGGATGATT	1420-1439	mega (7, 8)
mef(E)-MID	ATAGAAATATGCACAGGCGTT	1900-1880	mega (7, 8)
mef(E)-RT2	TCGTTAGCTGTTCTTCTGGT	1536-1517	mega (7, 8)
mef(E)-RT-PCR probe	HEX-ACCCCAGCACTCAATGCGGTTACAC	1482-1506	mega (7, 8)
Spglck-RT1	CATCGATAATGATGCCAACGT		GenBank
Spglck-RT2	AGTACCGAGTGTCATAAAGAC		GenBank
Spglck-RT-PCR probe	FAM-GCACCCATCCAGCGCTCACCAA		GenBank

Phenotypic induction of the mef(E) gene.

Erythromycin, clarithromycin, azithromycin, clindamycin, or telithromycin (each at 1/8× MIC) was added to an overnight diluted growth culture (1.5 × 10⁵ to 1.5 × 10⁶ CFU/ml), and cells were grown in a shaker bath at 37°C for 3 to 4 h. Cultures were sampled pre- and postinduction, and erythromycin, clarithromycin, azithromycin, clindamycin, and telithromycin MICs were determined by Etest (AB Biodisk, Solna, Sweden) according to the manufacturer's instructions. All Etest susceptibility testing was performed in duplicate, and the mean MIC was determined. Fold increase in the MIC was calculated by dividing the MIC after induction by the MIC before induction. A fourfold increase in the MIC following the induction was considered as significant.

Southern blot hybridization.

Extracted DNA was digested with the restriction enzyme EcoRI, electrophoresed, denatured, and transferred to a Hybond N+ membrane (Amersham Pharmacia) using a Turboblotter (Schleicher & Schuell) according to the manufacturer's instructions. Prehybridization, hybridization, probe labeling, and detection were performed using the ECL (enhanced chemiluminescence) direct labeling and detection system (Amersham Life Sciences). The 346-bp mef PCR product of mef-1 and mef-2 was used as a probe in the Southern blot hybridization experiments (Table 1).

Real-time RT-PCR.

RNA isolation was performed using the RNeasy Mini kit (QIAGEN) according to the manufacturer's instructions. Complete removal of DNA was verified by direct PCR with the RNA as a template. Real-time reverse transcription-PCR (RT-PCR) quantification based on the relative expression of a target gene, mef(E), versus a reference gene, gki (coding for glucose kinase), was utilized to study the expression of the mef(E) gene. RT-PCR primers mefE-RT1, mefE-RT2, and Spglck-RT1, Spglck-RT2 (DNA CORE Facility, Health Canada, Winnipeg, Manitoba) and hybridization probes mef(E) and gki (Synthegen) were designed according to the rules in the TaqMan One-Step RT-PCR master mix reagents kit (Applied Biosystems) and are shown in Table 1 together with PCR and sequencing primers. RT-PCR was carried out in 50 μl of mix containing TaqMan One-Step RT-PCR master mix without UNG, Multiscribe and RNase inhibitor, dNTPs, MgCl₂, and AmpliTaq Gold (Applied Biosystems) according to the manufacturer's specifications. The probe/primer ratio of 1:2 (0.25 μM to 0.5 μM) with 100 ng of RNA template was optimal for coamplification of mef(E) and gki. The coamplification was performed using a Mx4000 Multiplex quantitative PCR system (Stratagene) and consisted of an initial 30 min of reverse transcription at 48°C, followed by AmpliTaq Gold activation at 95°C for 10 min, and 40 cycles of PCR at 95°C for 30s (denaturation) and at 58°C for 60 s (annealing). Data analysis was performed in accordance with the Mx4000 Multiplex quantitative PCR system (Stratagene) instructions. All statistical tests were performed using the 2002 NCSS statistical program. Pearson correlation was utilized to study the degree of linear relationship between the erythromycin MIC and the relative mef(E) gene expression. The Pearson correlation coefficient has an absolute value between 0 and 1, with 1 indicating a perfect linear relationship and 0 meaning no linear relationship exists. Analysis of variance was used to determine the degree of statistical significance between isolates with different MICs.

DNA sequence analysis.

PCR products were purified using the Microcon microconcentrators according to the manufacturer's instructions (Millipore, Bedford, MA). Automated sequencing was done at the DNA CORE Facility (Health Canada, Winnipeg, Manitoba, Canada) using the same four primers as those used for PCR, plus two primers, mef(E)-MID and mef(E)-RT1, were designed to create more overlap and are depicted in Table 1. Sequence analysis was conducted using Lasergene (DNA Star Inc., Madison, WI) Seqman II module.

RESULTS

Phenotypic induction of mef(E)-mediated resistance in S. pneumoniae.

Following the induction with subinhibitory concentrations of macrolides, clindamycin, and telithromycin, mef(E)-mediated resistance in S. pneumoniae appeared to be inducible by 14- and 15-member macrolide antibiotics and was expressed at higher levels. Of all the macrolides, clarithromycin appeared to be the strongest inducer of mef(E) gene expression, where MICs of 78% (18/23) of S. pneumoniae isolates showed a significant (≥4-fold) increase following exposure to subinhibitory concentrations of clarithromycin. Subinhibitory levels of erythromycin and azithromycin showed the ability to induce mef(E) gene expression to the same extent, and both significantly increased the MICs of 57% (13/23) of S. pneumoniae isolates in comparison to those of susceptible controls, which showed a nonsignificant fold increase. Similarly to the susceptible controls, the MICs of all 23 isolates of mef(E)-carrying S. pneumoniae isolates showed a nonsignificant (<4-fold) increase following exposure to telithromycin and clindamycin.

Analysis of the clarithromycin MIC distribution before and after induction for the 23 M-phenotype mef(E)-carrying S. pneumonaie isolates showed a significant change, while no change in the MIC distribution was noted for the two susceptible mef(E)-negative S. pneumonaie isolates. Table 2 depicts the Etest MIC₉₀ values for erythromycin, clarithromycin, azithromycin, clindamycin, and telithromycin before and after induction of 23 M-phenotype and two susceptible S. pneumoniae isolates. The average erythromycin MIC₉₀ value of 64 μg/ml before induction increased fourfold to 256 μg/ml following induction in all the M-phenotype S. pneumoniae isolates. The azithromycin MIC₉₀ value of 96 μg/ml before induction increased threefold to 256μg/ml following induction. A sixfold increase from 32 μg/ml to 196 μg/ml in the MIC₉₀ value was observed following the induction with clarithromycin. As expected, no significant changes in the MIC₉₀ values before and after induction with clindamycin, as well as with telithromycin, were observed. Similarly, no significant changes in the MIC₉₀ values were observed for the susceptible controls (Table 1).

TABLE 2.

Phenotypic induction of macrolide resistance in 23 mef(E)-positive and 2 macrolide-susceptible S. pneumoniae isolates

S. pneumoniae isolate type	n (%)	Exptl condition	Etest MIC (μg/ml)a
			ERY		CLR		AZI		CLI		TEL
			MIC90	Range	MIC90	Range	MIC90	Range	MIC90	Range	MIC90	Range
M-phenotype	23 (92)	Before induction	64	3-128	32	2-64	96	16-128	0.19	0.023-0.25	0.5	0.125-0.75
		After induction	256	8-256	196	12-256	256	32-256	0.25	0.047-0.25	1	0.25-1.5
Susceptible	2 (8)	Before induction			0.25	0.25			0.047	0.032-0.047	0.008	0.008
		After induction			0.25	0.25			0.047	0.047	0.008	0.008

^aERY, erythromycin; CLR, clarithromycin; AZI, azithromycin; CLI, clindamycin; TEL, telithromycin.

mef(E) gene copy number.

Southern blot hybridizations using a mef(E) probe and EcoRI digestion of all 23 M-phenotype S. pneumoniae isolates tested indicated that mef(E) was present as a single band. Southern blot hybridization confirmed that the mef(E) gene inserts in more than four distinct sites within the pneumococcal genome as indicated by five different band sizes of 2,800 for 1 isolate, 3,100 for 3 isolates, 4,000 for 2 isolates, 8,000 for 3 isolates, and ≥10,000 bp for the majority (14 isolates).

Expression of mef(E) in clinical S. pneumoniae isolates.

Relative real-time RT-PCR was used to assess the expression of the macrolide efflux gene, mef(E), in 23 clinical isolates and in two susceptible controls. Each relative expression value is the mean of three replicas. The relative expression of mef(E) versus gki (a housekeeping gene coding for glucose kinase) increased linearly in isolates with decreased susceptibility to macrolides (Fig. 1). The most significant difference of 200-fold in the relative expression of mef(E) was observed between isolates with MICs of 1 μg/ml (0.0042) and 32 μg/ml (0.84). Comparing the mean relative expression of mef(E) in isolates with the erythromycin MIC of 2 μg/ml (0.034), 4 μg/ml (0.15), and 8 (0.24) μg/ml to the relative expression of 0.84, that for isolates with an erythromycin MIC of 32 μg/ml revealed increases in relative mef(E) expression by approximately 20-fold, 6-fold, and 4-fold, respectively (Fig. 1). The Pearson correlation coefficient of 0.9476 for the analysis of the degrees of linear relationship between the expression of the mef(E) gene and the erythromycin MICs indicated a nearly perfect relationship between these two variables. Therefore, a positive correlation (direct relationship) exits between the expression of the mef(E) gene and the erythromycin MIC. To assess whether the increases are actually statistically significant, an analysis of variance was utilized to study the degree of statistical significance of mef(E) gene expression between isolates with different MICs of erythromycin. Results from this study revealed that the differences in the mef(E) expression in isolates with different MICs of erythromycin are statistically significant (P < 0.05).

FIG. 1.

Mean relative expression of the efflux pump gene, mef(E) in 23 M-phenotype S. pneumoniae isolates with increasing resistance to macrolides.

Sequencing of the mef(E) gene, promoter, and the mef(E)-mel integenic region.

A 1,945-bp nucleotide sequence encompassing the upstream and downstream intergenic region of mef(E) and mel genes was amplified and sequenced using overlapping primers from all 23 M-phenotype S. pneumoniae isolates containing mef(E) and compared with each other and with the published sequence of mega. A 542-bp upstream region of mef(E) in the 23 M-phenotype S. pneumoniae isolates contained 24 nucleotide changes, including a 16-bp deletion 153 bp upstream and a single T→C substitution 30 bp upstream of the putative mef(E) start site, where all 23 isolates differed from the published mega sequence. A 16-bp deletion was noted previously in Italian mef(E) isolates (6). In addition to these changes, eight isolates showed a single-base-pair substitution in this region. Analysis of putative −10 and −35 regions of all isolates did not reveal any changes.

A 1,217-bp nucleotide sequence of the mef(E) gene differed by a single A→T substitution 165 bp downstream of the putative mef(E) start site in three isolates from the published mega sequence. No other changes were observed. A 119-bp intergenic region between mef(E) and mel contained a 99-bp deletion in 11 of the 23 M-phenotype S. pneumoniae isolates as compared to the published mega sequence. The intergenic region in these isolates was only 20 bp long. The remaining 12 isolates did not contain the 99-bp deletion in this region and were 100% in concordance with the published mega sequence. All the changes are depicted in Table 3.

TABLE 3.

Nucleotide sequence differences identified in 23 mef(E)-negative S. pneumoniae isolates and the published mega sequence

Position relative to mef(E)a	Change	No. of isolates with change
−30	T→C	23
−52	T→G	23
−60	Deletion of T	23
−75	A→T	23
−78	T→G	23
−79	A→G	23
−153	16-bp deletion	23
−360	G→T	5
−361	T→A	1
−372	A→C	2
+165	A→G	3
+1234	99-bp deletion	11

^aPosition relative to the start site of the mef(E) gene.

DISCUSSION

Efflux-mediated resistance in S. pneumoniae is one of the major mechanisms of macrolide resistance, and it is the predominating mechanism of macrolide resistance in North America (9). In fact, the increasing rates of macrolide resistance in the United States in recent years have been directly linked to the emergence of efflux-mediated resistance (8). Although once not very common, efflux-mediated macrolide resistance is not only now present at significant rates in Germany and Italy (6, 16, 20) but also predominates in European countries, such as Greece, Norway, and Scotland (1, 15). Recent work has shown that the efflux mechanism of resistance in S. pneumoniae responsible for acquired macrolide resistance can be distinguished as either mef(E) or mef(A) (6, 8, 25). Since it has been reported that both mef(A) and mef(E) genes are located on different genetic elements in S. pneumoniae that may influence the dissemination of M-phenotype macrolide resistance among streptococcal and other species, it has been suggested that the two genes remain separate (11). Generally, mef(A) predominates in Europe (1, 2, 6, 15) and mef(E) predominates in North America and Asia (5, 25). However, countries such as The Netherlands and Italy have reported that as many as 30% to 40% of macrolide-resistant S. pneumoniae isolates contain mef(E) (17). Also in a recent study from Hungary, no mef(A) genes were found (7).

In the present study, we investigated the properties of the mef(E) gene and the element carrying it in 23 genetically different strains of S. pneumoniae exhibiting increasing levels of macrolide resistance. It has been hypothesized that efflux-mediated resistance in S. pneumoniae is inducible by erythromycin and its derivatives and is expressed at moderately high levels (13, 14, 26). We showed here that efflux-mediated resistance in S. pneumoniae is inducible by all macrolides (erythromycin, clarithromycin, and azithromycin) tested and results in higher MICs. As expected mef(E)-mediated resistance was not induced by clindamycin and telithromycin. The induction of the mef(E) gene indicates some level of mef(E) gene expression regulation. Southern blot hybridization was employed to study the presence of multiple copies of the mef(E) gene in isolates with decreased susceptibility to macrolides. The Southern blot hybridization experiments using a mef(E) probe and EcoRI digestion revealed that the mef(E) gene in the 23 macrolide-resistant isolates of S. pneumoniae was present as a single copy. Consistent with previously published studies, we also found that the mef(E) gene inserts into the genome at more than four different locations (8). Insertion in different locations may have an effect on transcription of the efflux pump gene. We found some correlation between insertion sites (as identified by a band pattern) and erythromycin MIC for several isolates. Three isolates with a band size of 3,100 bp and two isolates with a band size of 4,000 bp had MICs of 2 and 4 μg/ml, respectively. However, 14 isolates with a band size of ≥10,000 bp had MICs ranging from 1 to 32 μg/ml; therefore, it is not possible to conclude from these data that the mega insertion site has an effect on erythromycin resistance.

It has been speculated before that the macrolide efflux system in S. pneumoniae might be a dual-efflux system as mef(E) and mel, the gene coding for the homolog to the Msr(A) protein in Staphylococcus, are not only present on the same genetic element but are cotranscribed (6, 8). However, there might exist some regulation of mel gene translation as the intergenic region of 119 bp contains a Shine-Dalgarno consensus sequence upstream of mel, but this sequence is eliminated in a class II insert which has a deletion of 99 bp in the intergenic region between mef(E) and mel (8). In addition, recently, it has been shown that msr(A)-like genes alone can confer the efflux phenotype in S. pneumoniae (7). This 99-bp deletion between mef(E) and mel of mega appeared to be common among macrolide-resistant S. pneumoniae isolates as it was identified in 62% of mef(E) isolates in comparison to 15.7% of isolates having no deletion in the study by Gay and Stevens (8). In contrast to this earlier study, we found equal numbers of isolates with and without the 99-bp deletion between mef(E) and mel in our isolates. Some of the isolates with the 119-bp intergenic region had higher erythromycin MICs, and some with the deletion had lower erythromycin MICs.

Utilization of relative real-time RT-PCR allowed the characterization of mef(E) gene expression in macrolide-resistant S. pneumoniae isolates. We showed that higher levels of expression of the mef(E) gene were associated with higher MICs of erythromycin. This to the best of our knowledge is the first study of macrolide efflux gene expression of macrolide-resistant S. pneumoniae clinical isolates with various macrolide MICs that were induced with macrolides. Expression studies of the efflux pump gene pmrA in fluoroquinolone-resistant isolates of S. pneumoniae have been reported and showed a similar direct correlation between the level of expression and the MIC to the antibiotic (18). Real-time RT-PCR quantifies steady-state mRNA levels, and therefore quantification of mRNA levels under these conditions does not explain transcription levels or mRNA stability. Further, the mRNA levels detected may not reflect the levels of protein produced by the cell, as regulation may occur at the posttranscriptional stage. The only conclusion that can be drawn from the RT-PCR experiments is that the increased levels of mef(E) mRNA are present in isolates with increasing macrolide resistance, but whether this increase in mef(E) mRNA level translates to an increased amount of the Mef(E) efflux pump protein remains to be determined.

In conclusion, this study presents new findings of the association between the level of expression of the mef(E) gene and resistance to macrolides in S. pneumoniae. Further studies assessing Mef(E) protein levels and function alone and also as part of the proposed dual-efflux system [Mef(E)-MsrA] are needed to fully understand the function of the mef(E) efflux gene in macrolide resistance. It would also be interesting to assess if a similar association between expression and resistance to macrolides exists in strains carrying the mef(A) macrolide efflux gene.