- Email: [email protected]
Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes
Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Diagnostic Microbiology and Infectious Disease journal homepage: www.elsevier.com/locate/diagmicrobio
Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes Luca Agostino Vitali a,⁎, Maria Chiara Di Luca b,1, Manuela Prenna b, Dezemona Petrelli b a b
School of Pharmacy, University of Camerino, Camerino, Italy School of Biosciences and Veterinary Medicine, University of Camerino, Camerino, Italy
a r t i c l e
i n f o
Article history: Received 17 June 2015 Received in revised form 31 July 2015 Accepted 9 August 2015 Available online xxxx Keywords: Group A streptococcus Antimicrobial resistance Macrolide Efflux
a b s t r a c t We investigated the correlation between the genetic variation within mef(A)-msr(D) determinants of efflux-mediated erythromycin resistance in Streptococcus pyogenes and the level of erythromycin resistance. Twenty-eight mef(A)-positive strains were selected according to erythromycin MIC (4–32 μg/mL), and their mef(A)-msr(D) regions were sequenced. Strains were classified according to the bacteriophage carrying mef(A)-msr(D). A new Φm46.1 genetic variant was found in 8 strains out of 28 and named VP_00501.1. Degree of allelic variation was higher in mef(A) than in msr(D). Hotspots for recombination were mapped within the locus that could have shaped the apparent mosaic structure of the region. There was a general correlation between mef(A)-msr(D) sequence and erythromycin resistance level. However, lysogenic conversion of susceptible strains by mef(A)-msr(D)–carrying Φm46.1 indicated that key determinants may not all reside within the mef(A)-msr(D) locus and that horizontal gene transfer could contribute to changes in the level of antibiotic resistance in S. pyogenes. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Macrolide resistance (MR) in Streptococcus pyogenes (group A Streptococcus [GAS]) is a global phenomenon with significant implication for the treatment of streptococcal infections when beta-lactams cannot be used. The high prescription and consumption of macrolides and the dissemination of MR clones have contributed to the high prevalence of resistance in several countries (Huang et al., 2014; Silva-Costa et al., 2006; Syrogiannopoulos et al., 2013; Wajima et al., 2014). Resistance to macrolides in GAS occurs by 2 main mechanisms, namely, the enzymatic modification of the target site onto the ribosome mediated by the product of the erm genes (Weisblum, 1995) and the active efflux of the drug by protein pumps encoded by the mef(A)-msr(D) system (Sutcliffe et al., 1996), which confers resistance to 14- and 15membered macrolides. mef(A) and msr(D) are arranged in tandem and carried by mobile genetic elements. In GAS, 3 of these elements have been discovered: Tn1207.3 (Santagati et al., 2003), Φ10394.4 (Banks et al., 2003), and Φm46.1 (Brenciani et al., 2010). In all the cases, mef(A) and msr(D) are transferred together, either by conjugative transposition or by phage-mediated transfer (Di Luca et al., 2010; Santagati et al., 2003). The phenotype of the mef(A)-msr(D)–mediated macrolide resistance is referred to as M phenotype and is associated with erythromycin ⁎ Corresponding author. Tel.: +39-0737-403282; fax: +39-0737-403281. E-mail address: [email protected] (L.A. Vitali). 1 Present address: Department of Microbiology and Infection Control, University Hospital of North Norway, N-9038 Tromsø, Norway.
MIC90 values of 8–16 μg/mL. Some studies on Streptococcus pneumoniae support the hypothesis that this resistance is able to negatively affect antibacterial therapy outcome (Lonks et al., 2002). Previous works have almost entirely focused on mef(A) only. Actually, both mef(A) and msr(D) play a key role in determining macrolide resistance in streptococci (Ambrose et al., 2005). Given that mef(A)-msr(D)– mediated resistance level is not uniform in GAS, with MIC ranging from 4–8 μg/mL to 32 μg/mL, some questions are still open: i) Which are the structural genetic bases of the observed phenotypic differences? ii) Which is the contribution, if any, of the carrying genetic element in the evolution of resistance? iii) Is the level of MR resistance strain dependent? To try to answer these questions, we investigated erythromycin-resistant isolates expressing the M phenotype due to the efflux pump MefA showing MICs for erythromycin ranging from 4 to 32 μg/mL. We studied the genetics of mef(A)-msr(D) locus aiming at understanding the bases of differences in erythromycin susceptibility and to eventually correlate them to sequence variations, carrying genetic elements, or strain specific genetic background. 2. Material and methods 2.1. Bacterial strains and antimicrobial susceptibility testing S. pyogenes strains showing the M phenotype resistance toward macrolides were selected from our laboratory collection (Seppälä et al., 1993). They were originally isolated from symptomatic patients suffering from pharyngotonsillitis and previously characterized by different molecular methods such as emm typing and Pulsed Field Gel
http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007 0732-8893/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
2
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Electrophoresis (Vitali et al., 2002; Zampaloni et al., 2003). The latter information was used to select a subset of strains so as to avoid the inclusion of clonally related strains. A third-level selection was then done after the determination of the erythromycin MICs, which were assessed by the microdilution method following the guidelines of the CLSI (CLSI, 2011). Finally, 28 strains were considered for the analysis. At a second stage and to study the phenotype of erythromycin resistance upon lysogenic conversion by phage Φm46.1, additional 27 S. pyogenes strains previously generated in our laboratory were used (Di Luca et al., 2010). MIC determinations were repeated 3 times, in triplicate for each test, to have a confident value for each strain. PCRs specific for erm(B) and erm(A) were negative in all strains (Zampaloni et al., 2003). 2.2. Amplification and sequencing of the mef(A)-msr(D) locus Oligonucleotide primers design was done with reference to the sequence of the Φm46.1 phage (acc. nr. FM864213). Their sequences are listed in Table S1 (Supplementary material). Four primers were designed to amplify the mef(A) and msr(D) coding (RAF1-mefA1F, RAF3-mefA3R, RAF4-mefA4F, RAF2-mefA2R) and the intergenic noncoding region. The pair up_CAM1 and up_mefA_R was used to amplify and sequence the mef(A)-msr(D) upstream region up to the insertion site of the Φm46.1-related genetic element into the chromosome. The 10500 series of primers has been designed after chromosome walking into the sequence of the newly recognized variant 5′-region found in a subset of strains (Table 1). 2.3. Phylogenetic reconstruction, tests for recombination, and statistical analysis Phylogenetic, molecular evolutionary analyses, and tree construction were conducted using MEGA version 5 (Tamura et al., 2011). The evolutionary history was inferred using the maximum likelihood
Table 1 List of strains used in this work with their susceptibility to erythromycin expressed as MIC and the PCR analysis of the mef(A)-msr(D) region. According to the expected size of the amplicon, the strain was associated to a known reference phage type (last column). Strain
VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO
m46 9691 00211 12790 1049 20201 41201 82201 31311 04201 70411 20411 3179 34201 84201 11411 40611 11611 40301 11901 6199 0279 10201 53101 58101 00301 01301 61301 00501
MICERY (mg/L)
Approx. size of the amplicon (expected size)
Ref. phage type
16 12 12 12 12 12 12 12 8 12 16 8 8 8 4 16 8 8 8 8 8 8 8 16 32 16 12 12 8
3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) Negative Negative Negative Negative Negative Negative 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp
Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φ10394.4 like Φ10394.4 like Φ10394.4 like Φ10394.4 Φ10394.4 Φ10394.4 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1
bp) bp) bp) bp) bp) bp) bp) bp)
method based on the Tamura 3-parameter model for mef(A) alone and the entire locus and on the Hasegawa–Kishino–Yano model for msr(D) alone (Hasegawa et al., 1985). The maximum chi-square test was used to confirm suspected recombination events between pairs of mef(A)-msr(D) locus from different strains as it compares the distribution of polymorphic sites along sequences with those expected to occur by chance (Smith, 1999). The test was performed using the START package (Jolley et al., 2001), which is available at http:// pubmlst.org/software/analysis/start/ (last accessed on 15 June 2015). Standard statistical analyses were performed using the software Statgraphics Centurion XV. Group to group comparisons were performed by the Kruskal–Wallis test for the medians and by chi-square test for the means. Significance threshold was set at P = 0.05 in all cases. 3. Results and discussion 3.1. mef(A)-mediated erythromycin resistance The results of the susceptibility testing for erythromycin are reported in Table 1. Twelve out of 29 strains (41.4%; 28 test strains plus the reference one) showed an MIC of 8 μg/mL; and 34.5% of the strains, a value of 12 μg/mL. The MIC against 4 strains (13.8%) plus the m46 reference strain was 16 μg/mL. Only strain VP_SPYO58101 showed a value above 16 μg/mL. A level below 8 μg/mL was recorded in 1 strain (VP_SPYO84201). 3.2. Genetic context of the mef(A)-msr(D) locus A PCR screening was used to classify strains in respect to the known genetic elements carrying mef(A) or its homologs (Table 1). Using primers up_CAM1 and up_mefA_R (Table S1, Supplementary material), a PCR product of about 3 kb was expected if the macrolide efflux gene was carried by Φm46.1 phage type. Six strains were positive as well as the control strain m46. In 8 cases, mef(A) subclass was of the MB56Spy045 type (amplicon size 1.7 kb) (Blackman Northwood et al., 2009). An unexpected 4-kb fragment was obtained from 8 strains and was further investigated (see “Sequencing” section). At last, the remaining 6 strains were negative and then found positive to the methylase, and the R28-like genes carried by Φ10394.4 (Banks et al., 2003). They were assigned to the Φ10394.4 phage-type group (Table 1). 3.3. Sequencing The mef(A)-msr(D) locus of each strain was sequenced and was composed by the coding regions of mef(A) and msr(D), the intergenic region between the 2 genes, and the 300 bp 5′ noncoding region upstream to mef(A). The length of the entire locus was in the range of 3112–3127 bp. The 16-bp difference is due to an indel present in the 5′ noncoding region upstream to mef(A) in Φ10394.4. The 4-kb amplicons from the left junction site of the element carrying mef(A) to the 5′ of the mef(A) itself obtained from 8 strains (Table 1 and “Genetic context of the mef(A)-msr(D) locus” section) were also sequenced. Comparative analysis showed a mosaic structure of the locus with contributions from both Φm46.1 and Φ10394.4-like phage-types. The new variant was called VP_00501.1 (Fig. 1). The chromosomal insertion site of the carrying element was the same as in Φm46.1 (GenBank: FM864213), i.e., a gene coding for a 23S rRNA uracil methyltransferase (rumA). Also, the sequence of the Φm46.1 orf1 was maintained, while the downstream sequence corresponding to orf2 was absent. The sequence continued with the orf6 and the 5′-end of orf7 carried by Φ10394.4 (alias SpyM6 mefA phage element, GenBank: AY445042). Downstream of the latter ORF, there was an 800 nt long sequence responsible for the difference in length between the VP_00501.1 and the other phage types (Fig. 1). XBLAST search found a match with a domain of the AlwI family of type II restriction endonucleases (59% identity, 82% similarity to Lactobacillus delbrueckii, GenBank: WP_003617961.1
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
3
Fig. 1. Schematic representation of the 5′ end of the Φm46.1-related phage element VP_00501.1. The box i) delimits the region subjected to recombination test analysis; ii) points the recombination spots with indication of their position relative to the adenine of the mef(A) start codon (position +1); and iii) contains a graphical display of the statistics.
protein). The remaining 300 nt long sequence linking the AlwIcontaining region and the beginning of mef(A) was highly similar to the corresponding region in the other phage types. In addition, the sequence of the whole mef(A)-msr(D) locus (approximate length: 3100 bp) has been determined. The new variant VP_00501.1 has been submitted to GenBank (GenBank: KJ809088). Three strains within the Φ10394.4 group, indicated as “Φ10394.4–like” phage type, did not show the 16-bp duplication at position −140 (relative to the adenine of the mef(A) gene start codon) found in the corresponding region of Φ10394.4. 3.4. Correlation between phage types and the MIC for erythromycin We tempted to correlate the MIC for erythromycin to the sequence features of the mef(A)-msr(D) genetic region (Table 2). We tested the null hypothesis that the medians of the MIC values (4, 8, 12, 16, and 32 μg/mL) within each of the 5 phage-type groups were the same by the Kruskal–Wallis test. The medians of MICs shown by the clades resulting from mef(A) cluster analysis were computed and compared. The same procedure has been independently repeated considering the sequence of msr(D), mef(A)-msr(D) locus, or the carrying genetic element. Even if very close to the 0.05 threshold, the P value was 0.059 meaning that the null hypothesis cannot be rejected at the 95% confidence level. The comparative analysis of paired groups showed a maximum difference between the distribution of the MIC values of the ΦMB56Spy045- in respect to those of Φm46.1- and VP_00501.1carrying strains. Hence, the phage type per se was not associated with
Table 2 Kruskal–Wallis test statistics to compare the medians of MICs. Genetic structure considered
Test statistic
P value
mef(A) msr(D) mef(A)-msr(D) genetic element
13.334 2.08 11.428 9.086
0.01⁎ 0.556 0.022⁎ 0.059
⁎ Statistically significant difference amongst the medians at the 95.0% confidence level.
a peculiar erythromycin resistance level. A correlation was recorded only if either mef(A) alone or mef(A)-msr(D) was considered as the reference genetic structures for the analysis (Table 2). 3.5. Alignment and phylogenetic reconstruction The sequences of the mef(A), msr(D), and mef(A)-msr(D) locus plus the upstream untranslated region were aligned and compared. Cluster analyses have been performed following a stepwise inside-out design considering at first only the contribution of the mef(A)-coding sequence to the observed MIC values. Then the analysis have been extended considering msr(D) alone and at last the mef(A)msr(D) locus plus the noncoding regulatory region upstream to mef(A). Results for mef(A) are shown in Fig. 2. The tree with the highest log likelihood (−1811.0968) is shown. Initial tree(s) for the heuristic search were obtained by applying the neighbor-joining method to a matrix of pairwise distances estimated using the maximum composite likelihood approach. All positions with less than 0% site coverage were eliminated. That is, less than 100% alignment gaps, missing data, and ambiguous bases were allowed at any position. Some clades were composed by strains showing almost identical MIC values, namely, no. 3 and 5, while others were highly heterogeneous (e.g., no. 4) with MICs spanning from 8 μg/mL up to 32 μg/mL. Results for msr(D) revealed that variability of this gene within the set of considered strains is less if compared to that of mef(A). There was a collapse of cluster nos. 4 and 5 identified by mef(A) analysis into a single one when msr(D) alone was considered. Also within cluster nos. 1 and 3, the msr(D) sequences were all identical, while in the mef(A) analysis, strains 31311 and 20411 were slightly different within clade no. 1 and m46 reference strain was different from the other mef(A) sequences clustering within no. 3 (Fig. 2). In the case of msr(D) analysis, the association between sequence of msr(D) and MIC is not appreciable as for mef(A) (P = 0.56). The same phylogenetic and correlation analysis has been performed also for the entire mef(A)-msr(D) sequence locus including the noncoding region upstream to mef(A) (300 nt) and the interspace sequence between the 2 genes (about 90 nt). The result was substantially superimposable on
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
4
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Fig. 2. Molecular phylogenetic analysis of mef(A) by maximum likelihood method and correlation with MIC of erythromycin. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (bootstrap value: 1000). The analysis involved 29 nucleotide sequences (28 isolates plus the reference strain m46). The MIC of erythromycin against each strain is indicated by a colored dot as per the figure legend.
that obtained for mef(A) alone (Fig. 2), indicating that the major contribution to the locus variability and to the measurable degree of association with the MIC values was mostly due to mef(A). 3.6. Single Nucleotide Polymorphisms (SNPs) and protein polymorphisms analysis Nineteen SNPs were mapped in mef(A), 7 of which led to synonymous substitution and 12 to nonsynonymous ones. msr(D) showed an inverted pattern with 14 synonymous and 6 nonsynonymous substitutions. The 2 distributions of synonymous/nonsynonymous substitutions within the 2 genes were significantly different (P = 0.038, chi-square test). In the mef(A) coding region, SNPs were confined to the distal 5′half of the gene. The same general structural feature in mef(A) has been previously observed in a large collection of macrolide-resistant GAS by Blackman Northwood et al. (2009). It is likely that changes in the N-terminal part of the protein are not tolerated. Nonsynonymous replacements were predicted to be neutral with the only exception of K481M in the sequence of MsrD. Differences in the synonymous– nonsynonymous substitution partitioning between mef(A) and msr(D) were significant (chi-square; P b 0.05) and associated with the phylogenetic study that showed a major contribution to variability of the locus deriving from mef(A). Although synonymous mutations have no effect on the protein sequence, the redundancy of the genetic code has been identified as important in shaping gene expression and cellular function (Plotkin and Kudla, 2011). We also computed the dN/dS ratio to investigate the evolutionary significance of this 2-gene system conferring resistance to macrolides in GAS (Table 3). The dN/dS values were determined for each clade to clade comparison for mef(A) alone, msr(D) alone, and the mef(A)-msr(D) locus. A gene is under positive (Darwinian) selection if dN/dS ratio is higher than 1, while a dN/dS ratio b1 indicates a purifying (stabilizing) selection. The average ratio is higher for mef(A) compared to the one computed for either msr(D) alone or the entire locus (mean dN/dS = 0.518; 95% CI [0.518 ± 0.303], P b 0.05). By comparison of the dN/dS ratio computed for sequence of clade 3 with those for clades 1 and 2, we obtained values
of 1.357 and 1.206, respectively. As Mef(A) is a transmembrane protein, we used TOPCONS to predict the cytoplasmic, extracellular, and membrane-spanning regions (Bernsel et al., 2009). We then mapped the SNPs to visualize their position along the protein. The analysis revealed that the N-terminal part of the protein, which was not polymorphic, consisted of 4 transmembrane domains. As expected, no changes mapped in the transmembrane hydrophobic sequences with the exception of position 393, which presented a Phe to Cys substitution in all isolates except those of clade no. 3 (Fig. 2 and Fig. S1). Conversely, all the other nonsynonymous substitutions changed amino acids exposed either to the cytoplasm or the periplasm. No apparent association was found between a particular amino acid substitution or location and MIC of erythromycin. 3.7. Maximum chi-square test for recombination According to phylogenetic reconstruction, different parental mef(A)msr(D) locus sequences belonging to different clades within the tree were considered for analysis. The test result was confirmatory of plausible recombination events for some of the paired variants. The possible Table 3 Interclade analysis of dN/dS ratio. Clades compareda
1-2 1-3 1-4 1-5 2-3 2-4 2-5 3-4 3-5 4-5 a b
dN/dS mef(A)
msr(D)
mef(A)-msr(D)
0.000 1.357 0.346 0.469 1.206 0.308 0.365 0.349 0.436 0.340
0.051 0.282 0.185 n.a.b 0.000 0.110 n.a.b 0.173 n.a.b n.a.b
0.196 0.789 0.244 0.240 0.151 0.169 0.147 0.248 0.245 0.322
Clades numbering corresponds to that of Fig. 2. n.a. = not available. Tree construction for msr(D) resulted in 4 clades only (Section 3.5).
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
recombination sites and the P value are visualized in Fig. 1 (inner box). The 5′ coding region of mef(A) did not show hotspots for recombination due to its invariability (see analysis of SNPs). Two major recombination sites mapped at a distance of 763 and 936 bp downstream the start codon of mef(A) (A of AUG start codon designated as position +1), while a third one was in the intergenic sequence between mef(A) and msr(D) coding sequences (positions 1315 and 1321). The last observation is supporting the mosaic structure of this locus and the hypothesis that mef(A) and msr(D) may follow an independent evolution, although they were associated in all the isolates studied so far. Actually, authors that have cloned mef(A) for the first time were able to express it into Escherichia coli (Clancy et al., 1996). The opposite was found in S. pneumoniae in which efflux-mediated erythromycin resistance is mediated by the mef(A) homolog mef(E). The association of mef(E) with the downstream ATP-dependent pump gene mel (homolog to msr(D)) was essential for the correct expression of erythromycin resistance (Ambrose et al., 2005). A correspondent conclusive demonstration for GAS is lacking. Some authors have presented preliminary results that address msr(D) (therein referred to as matA) as the main driver of erythromycin resistance (Iannelli et al., 2004). Our genetic study supports the mentioned major role of msr(D) since the level of its variation is lower than that of mef(A). Another probable hotspot for recombination mapped at nucleotide −161 bp, within the noncoding sequence upstream to mef(A) (Fig. 1). Interestingly, 3 strains have the same 16bp duplication at position −140 found in the corresponding region of Φ10394.4, which could be the result of recombination. Similar events may also have contributed to the variability observed in the newly described variant VP_00501.1 (Fig. 1 and Table 1). 3.8. Possible role of the strain genetic background on the expression of erythromycin resistance Mef(A) and Msr(D) are likely to be strongly dependent on the metabolism of the bacterial cells, as they rely on the proton motive force throughout the membrane (Mef(A)) and on the intracellular level of ATP (Msr(D)). Given that each and every bacterial strain shows a different relative metabolism, we determined the MIC of erythromycin against susceptible GAS strains with different emm types after lysogenic conversion to an erythromycin-resistant phenotype by phage Φm46.1 (carrying the mef(A)-msr(D) locus). As emm type variability is a good estimate of the genetic variability of a GAS population, it can indicate a different genetic background and possibly a different growth rate or metabolism (Zampaloni et al., 2003). The results are summarized in Table S2. In all lysogens, the insertion site was the same (not shown) and in agreement with published data (Brenciani et al., 2010). After lysogenic conversion, the MIC of erythromycin did not change for the majority of the lysogens (19 out of 27, 70%). Nevertheless, in 8 strains, the MIC rose to 32 μg/mL. Hence, the genetic background of the bacterial cell might affect the expression of erythromycin resistance to some extent. 4. Conclusions In S. pyogenes, mef(A)-msr(D) locus and its carrying genetic elements undergo a significant level of recombination that may shape their genetic structure. In turn, some of the changes could contribute to the heterogeneous expression of the resistance phenotype eventually measured as MIC of erythromycin ranging from 4 μg/mL up to 32 μg/mL. In addition to that, also horizontal genetic transfer of genetic determinants of resistance in a different genomic context could add to the observed phenotypic variability. In light of our findings, it is not improbable that, in the future, we might witness an even increase in the level of mef(A)-msr(D)–dependent macrolide resistance in S. pyogenes. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.diagmicrobio.2015.08.007.
5
Acknowledgments We are grateful to Dr Valerio Iebba, Dr Raffaella Casalena, Dr Stefania D’Ercole, Dr Wael Bahnan, and Dr Loreta Biqiku for their valuable technical assistance and for helpful discussions. This work was supported by MIUR's grants “Futuro in Ricerca” #RBFR10X4YN to D.P. and “PRIN” #200929YFMK_003 to M.P.
References Ambrose KD, Nisbet R, Stephens DS. Macrolide efflux in Streptococcus pneumoniae is mediated by a dual efflux pump (mel and mef) and is erythromycin inducible. Antimicrob Agents Chemother 2005;49:4203–9. http://dx.doi.org/10.1128/AAC.49. 10.4203-4209.2005. Banks DJ, Porcella SF, Barbian KD, Martin JM, Musser JM. Structure and distribution of an unusual chimeric genetic element encoding macrolide resistance in phylogenetically diverse clones of group A Streptococcus. J Infect Dis 2003;188:1898–908. http://dx. doi.org/10.1086/379897. Bernsel A, Viklund H, Hennerdal A, Elofsson A. TOPCONS: consensus prediction of membrane protein topology. Nucleic Acids Res 2009;37:W465–8. http://dx.doi.org/10. 1093/nar/gkp363. Blackman Northwood J, Del Grosso M, Cossins LR, Coley MD, Creti R, Pantosti A, et al. Characterization of macrolide efflux pump mef subclasses detected in clinical isolates of Streptococcus pyogenes isolated between 1999 and 2005. Antimicrob Agents Chemother 2009;53:1921–5. http://dx.doi.org/10.1128/AAC.01065-08. Brenciani A, Bacciaglia A, Vignaroli C, Pugnaloni A, Varaldo PE, Giovanetti E. Phim46.1, the main Streptococcus pyogenes element carrying mef(A) and tet(O) genes. Antimicrob Agents Chemother 2010;54:221–9. http://dx.doi.org/10.1128/AAC.00499-09. Clancy J, Petitpas J, Dib-Hajj F, Yuan W, Cronan M, Kamath AV, et al. Molecular cloning and functional analysis of a novel macrolide-resistance determinant, mefA, from Streptococcus pyogenes. Mol Microbiol 1996;22:867–79. http://dx.doi.org/10.1046/ j.1365-2958.1996.01521.x. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial susceptibility testing; twenty first informational supplement. CLSI; 2011. Di Luca MC, D’Ercole S, Petrelli D, Prenna M, Ripa S, Vitali LA. Lysogenic transfer of mef(A) and tet(O) genes carried by Φm46.1 among group A streptococci. Antimicrob Agents Chemother 2010;54:4464–6. http://dx.doi.org/10.1128/AAC.01318-09. Hasegawa M, Kishino H, Yano T. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J Mol Evol 1985;22:160–74. http://dx.doi.org/10.1007/ BF02101694. Huang CY, Lai JF, Huang IW, Chen PC, Wang HY, Shiau YR, et al. Epidemiology and molecular characterization of macrolide-resistant Streptococcus pyogenes in Taiwan. J Clin Microbiol 2014;52:508–16. http://dx.doi.org/10.1128/JCM. 02383-13. Iannelli F, Santagati M, Doquier JD, Cassone M, Oggioni M, Rossolini G, et al. Type M resistance to macrolides in streptococci is not due to the mef (A) Gene, but to mat (A) encoding an ATP-dependent efflux pump. 44th ICAAC – 2004 Oct 30-Nov 2. Washington, DC; 2004. p. C1‐1188. Jolley KA, Feil EJ, Chan M-S, Maiden MCJ. Sequence type analysis and recombinational tests (START). Bioinformatics 2001;17:1230–1. http://dx.doi.org/10.1093/bioinformatics/17.12.1230. Lonks JR, Garau J, Gomez L, Xercavins M, Ochoa de Echagüen A, Gareen IF, et al. Failure of macrolide antibiotic treatment in patients with bacteremia due to erythromycinresistant Streptococcus pneumoniae. Clin Infect Dis 2002;35:556–64. http://dx.doi. org/10.1086/341978. Plotkin JB, Kudla G. Synonymous but not the same: the causes and consequences of codon bias. Nat Rev Genet 2011;12:32–42. http://dx.doi.org/10.1038/nrg2899. Santagati M, Iannelli F, Cascone C, Campanile F, Oggioni MR, Stefani S, et al. The novel conjugative transposon Tn1207.3 carries the macrolide efflux gene mef(A) in Streptococcus pyogenes. Microb Drug Resist 2003;9:243–7. http://dx.doi.org/10. 1089/107662903322286445. Seppälä H, Nissinen A, Yu Q, Huovinen P. Three different phenotypes of erythromycinresistant Streptococcus pyogenes in Finland. J Antimicrob Chemother 1993;32: 885–91. http://dx.doi.org/10.1093/jac/32.6.885. Silva-Costa C, Ramirez M, Melo-Cristino J. Identification of macrolide-resistant clones of Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2006;12(6):513–8. [cited 2014 Jan 7]. Smith JM. The detection and measurement of recombination from sequence data. Genetics 1999;153:1021–7. Sutcliffe J, Tait-Kamradt A, Wondrack L. Streptococcus pneumoniae and Streptococcus pyogenes resistant to macrolides but sensitive to clindamycin: a common resistance pattern mediated by an efflux system. Antimicrob Agents Chemother 1996;40: 1817–24. Syrogiannopoulos GA, Grivea IN, Al-Lahham A, Panagiotou M, Tsantouli AG, Michoula Ralf René Reinert AN, et al. Seven-year surveillance of emm types of pediatric group A streptococcal pharyngitis isolates in Western Greece. In: de Lencastre H, editor. PLoS One, 8. ; 2013. p. e71558. http://dx.doi.org/10.1371/ journal.pone.0071558. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 2011;28:2731–9. http://dx.doi. org/10.1093/molbev/msr121.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
6
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Vitali LA, Zampaloni C, Prenna M, Ripa S. PCR M typing: a new method for rapid typing of group a streptococci. J Clin Microbiol 2002;40:679–81. http://dx.doi.org/10.1128/ JCM.40.2.679-681.2002. Wajima T, Chiba N, Morozumi M, Shouji M, Sunaoshi K, Sugita K, et al. Prevalence of macrolide resistance among group a streptococci isolated from pharyngotonsillitis. Microb Drug Resist 2014;57:1383–8. http://dx.doi.org/10.1089/mdr.2013.0213.
Weisblum B. Erythromycin resistance by ribosome modification. Antimicrob Agents Chemother 1995;39:577–85. http://dx.doi.org/10.1128/AAC.39.3.577. Zampaloni C, Cappelletti P, Prenna M, Vitali LA, Ripa S. emm Gene distribution among erythromycin-resistant and -susceptible Italian isolates of Streptococcus pyogenes. J Clin Microbiol 2003;41:1307–10. http://dx.doi.org/10.1128/JCM.41. 3.1307-1310.2003.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
Contents lists available at ScienceDirect
Diagnostic Microbiology and Infectious Disease journal homepage: www.elsevier.com/locate/diagmicrobio
Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes Luca Agostino Vitali a,⁎, Maria Chiara Di Luca b,1, Manuela Prenna b, Dezemona Petrelli b a b
School of Pharmacy, University of Camerino, Camerino, Italy School of Biosciences and Veterinary Medicine, University of Camerino, Camerino, Italy
a r t i c l e
i n f o
Article history: Received 17 June 2015 Received in revised form 31 July 2015 Accepted 9 August 2015 Available online xxxx Keywords: Group A streptococcus Antimicrobial resistance Macrolide Efflux
a b s t r a c t We investigated the correlation between the genetic variation within mef(A)-msr(D) determinants of efflux-mediated erythromycin resistance in Streptococcus pyogenes and the level of erythromycin resistance. Twenty-eight mef(A)-positive strains were selected according to erythromycin MIC (4–32 μg/mL), and their mef(A)-msr(D) regions were sequenced. Strains were classified according to the bacteriophage carrying mef(A)-msr(D). A new Φm46.1 genetic variant was found in 8 strains out of 28 and named VP_00501.1. Degree of allelic variation was higher in mef(A) than in msr(D). Hotspots for recombination were mapped within the locus that could have shaped the apparent mosaic structure of the region. There was a general correlation between mef(A)-msr(D) sequence and erythromycin resistance level. However, lysogenic conversion of susceptible strains by mef(A)-msr(D)–carrying Φm46.1 indicated that key determinants may not all reside within the mef(A)-msr(D) locus and that horizontal gene transfer could contribute to changes in the level of antibiotic resistance in S. pyogenes. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Macrolide resistance (MR) in Streptococcus pyogenes (group A Streptococcus [GAS]) is a global phenomenon with significant implication for the treatment of streptococcal infections when beta-lactams cannot be used. The high prescription and consumption of macrolides and the dissemination of MR clones have contributed to the high prevalence of resistance in several countries (Huang et al., 2014; Silva-Costa et al., 2006; Syrogiannopoulos et al., 2013; Wajima et al., 2014). Resistance to macrolides in GAS occurs by 2 main mechanisms, namely, the enzymatic modification of the target site onto the ribosome mediated by the product of the erm genes (Weisblum, 1995) and the active efflux of the drug by protein pumps encoded by the mef(A)-msr(D) system (Sutcliffe et al., 1996), which confers resistance to 14- and 15membered macrolides. mef(A) and msr(D) are arranged in tandem and carried by mobile genetic elements. In GAS, 3 of these elements have been discovered: Tn1207.3 (Santagati et al., 2003), Φ10394.4 (Banks et al., 2003), and Φm46.1 (Brenciani et al., 2010). In all the cases, mef(A) and msr(D) are transferred together, either by conjugative transposition or by phage-mediated transfer (Di Luca et al., 2010; Santagati et al., 2003). The phenotype of the mef(A)-msr(D)–mediated macrolide resistance is referred to as M phenotype and is associated with erythromycin ⁎ Corresponding author. Tel.: +39-0737-403282; fax: +39-0737-403281. E-mail address: [email protected] (L.A. Vitali). 1 Present address: Department of Microbiology and Infection Control, University Hospital of North Norway, N-9038 Tromsø, Norway.
MIC90 values of 8–16 μg/mL. Some studies on Streptococcus pneumoniae support the hypothesis that this resistance is able to negatively affect antibacterial therapy outcome (Lonks et al., 2002). Previous works have almost entirely focused on mef(A) only. Actually, both mef(A) and msr(D) play a key role in determining macrolide resistance in streptococci (Ambrose et al., 2005). Given that mef(A)-msr(D)– mediated resistance level is not uniform in GAS, with MIC ranging from 4–8 μg/mL to 32 μg/mL, some questions are still open: i) Which are the structural genetic bases of the observed phenotypic differences? ii) Which is the contribution, if any, of the carrying genetic element in the evolution of resistance? iii) Is the level of MR resistance strain dependent? To try to answer these questions, we investigated erythromycin-resistant isolates expressing the M phenotype due to the efflux pump MefA showing MICs for erythromycin ranging from 4 to 32 μg/mL. We studied the genetics of mef(A)-msr(D) locus aiming at understanding the bases of differences in erythromycin susceptibility and to eventually correlate them to sequence variations, carrying genetic elements, or strain specific genetic background. 2. Material and methods 2.1. Bacterial strains and antimicrobial susceptibility testing S. pyogenes strains showing the M phenotype resistance toward macrolides were selected from our laboratory collection (Seppälä et al., 1993). They were originally isolated from symptomatic patients suffering from pharyngotonsillitis and previously characterized by different molecular methods such as emm typing and Pulsed Field Gel
http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007 0732-8893/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
2
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Electrophoresis (Vitali et al., 2002; Zampaloni et al., 2003). The latter information was used to select a subset of strains so as to avoid the inclusion of clonally related strains. A third-level selection was then done after the determination of the erythromycin MICs, which were assessed by the microdilution method following the guidelines of the CLSI (CLSI, 2011). Finally, 28 strains were considered for the analysis. At a second stage and to study the phenotype of erythromycin resistance upon lysogenic conversion by phage Φm46.1, additional 27 S. pyogenes strains previously generated in our laboratory were used (Di Luca et al., 2010). MIC determinations were repeated 3 times, in triplicate for each test, to have a confident value for each strain. PCRs specific for erm(B) and erm(A) were negative in all strains (Zampaloni et al., 2003). 2.2. Amplification and sequencing of the mef(A)-msr(D) locus Oligonucleotide primers design was done with reference to the sequence of the Φm46.1 phage (acc. nr. FM864213). Their sequences are listed in Table S1 (Supplementary material). Four primers were designed to amplify the mef(A) and msr(D) coding (RAF1-mefA1F, RAF3-mefA3R, RAF4-mefA4F, RAF2-mefA2R) and the intergenic noncoding region. The pair up_CAM1 and up_mefA_R was used to amplify and sequence the mef(A)-msr(D) upstream region up to the insertion site of the Φm46.1-related genetic element into the chromosome. The 10500 series of primers has been designed after chromosome walking into the sequence of the newly recognized variant 5′-region found in a subset of strains (Table 1). 2.3. Phylogenetic reconstruction, tests for recombination, and statistical analysis Phylogenetic, molecular evolutionary analyses, and tree construction were conducted using MEGA version 5 (Tamura et al., 2011). The evolutionary history was inferred using the maximum likelihood
Table 1 List of strains used in this work with their susceptibility to erythromycin expressed as MIC and the PCR analysis of the mef(A)-msr(D) region. According to the expected size of the amplicon, the strain was associated to a known reference phage type (last column). Strain
VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO VP_SPYO
m46 9691 00211 12790 1049 20201 41201 82201 31311 04201 70411 20411 3179 34201 84201 11411 40611 11611 40301 11901 6199 0279 10201 53101 58101 00301 01301 61301 00501
MICERY (mg/L)
Approx. size of the amplicon (expected size)
Ref. phage type
16 12 12 12 12 12 12 12 8 12 16 8 8 8 4 16 8 8 8 8 8 8 8 16 32 16 12 12 8
3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) 3000 bp (3076 bp) Negative Negative Negative Negative Negative Negative 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 1500–2000 bp (1777 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp 4000 bp
Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φm46.1 Φ10394.4 like Φ10394.4 like Φ10394.4 like Φ10394.4 Φ10394.4 Φ10394.4 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 ΦMB56Spy045 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1 VP_00501.1
bp) bp) bp) bp) bp) bp) bp) bp)
method based on the Tamura 3-parameter model for mef(A) alone and the entire locus and on the Hasegawa–Kishino–Yano model for msr(D) alone (Hasegawa et al., 1985). The maximum chi-square test was used to confirm suspected recombination events between pairs of mef(A)-msr(D) locus from different strains as it compares the distribution of polymorphic sites along sequences with those expected to occur by chance (Smith, 1999). The test was performed using the START package (Jolley et al., 2001), which is available at http:// pubmlst.org/software/analysis/start/ (last accessed on 15 June 2015). Standard statistical analyses were performed using the software Statgraphics Centurion XV. Group to group comparisons were performed by the Kruskal–Wallis test for the medians and by chi-square test for the means. Significance threshold was set at P = 0.05 in all cases. 3. Results and discussion 3.1. mef(A)-mediated erythromycin resistance The results of the susceptibility testing for erythromycin are reported in Table 1. Twelve out of 29 strains (41.4%; 28 test strains plus the reference one) showed an MIC of 8 μg/mL; and 34.5% of the strains, a value of 12 μg/mL. The MIC against 4 strains (13.8%) plus the m46 reference strain was 16 μg/mL. Only strain VP_SPYO58101 showed a value above 16 μg/mL. A level below 8 μg/mL was recorded in 1 strain (VP_SPYO84201). 3.2. Genetic context of the mef(A)-msr(D) locus A PCR screening was used to classify strains in respect to the known genetic elements carrying mef(A) or its homologs (Table 1). Using primers up_CAM1 and up_mefA_R (Table S1, Supplementary material), a PCR product of about 3 kb was expected if the macrolide efflux gene was carried by Φm46.1 phage type. Six strains were positive as well as the control strain m46. In 8 cases, mef(A) subclass was of the MB56Spy045 type (amplicon size 1.7 kb) (Blackman Northwood et al., 2009). An unexpected 4-kb fragment was obtained from 8 strains and was further investigated (see “Sequencing” section). At last, the remaining 6 strains were negative and then found positive to the methylase, and the R28-like genes carried by Φ10394.4 (Banks et al., 2003). They were assigned to the Φ10394.4 phage-type group (Table 1). 3.3. Sequencing The mef(A)-msr(D) locus of each strain was sequenced and was composed by the coding regions of mef(A) and msr(D), the intergenic region between the 2 genes, and the 300 bp 5′ noncoding region upstream to mef(A). The length of the entire locus was in the range of 3112–3127 bp. The 16-bp difference is due to an indel present in the 5′ noncoding region upstream to mef(A) in Φ10394.4. The 4-kb amplicons from the left junction site of the element carrying mef(A) to the 5′ of the mef(A) itself obtained from 8 strains (Table 1 and “Genetic context of the mef(A)-msr(D) locus” section) were also sequenced. Comparative analysis showed a mosaic structure of the locus with contributions from both Φm46.1 and Φ10394.4-like phage-types. The new variant was called VP_00501.1 (Fig. 1). The chromosomal insertion site of the carrying element was the same as in Φm46.1 (GenBank: FM864213), i.e., a gene coding for a 23S rRNA uracil methyltransferase (rumA). Also, the sequence of the Φm46.1 orf1 was maintained, while the downstream sequence corresponding to orf2 was absent. The sequence continued with the orf6 and the 5′-end of orf7 carried by Φ10394.4 (alias SpyM6 mefA phage element, GenBank: AY445042). Downstream of the latter ORF, there was an 800 nt long sequence responsible for the difference in length between the VP_00501.1 and the other phage types (Fig. 1). XBLAST search found a match with a domain of the AlwI family of type II restriction endonucleases (59% identity, 82% similarity to Lactobacillus delbrueckii, GenBank: WP_003617961.1
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
3
Fig. 1. Schematic representation of the 5′ end of the Φm46.1-related phage element VP_00501.1. The box i) delimits the region subjected to recombination test analysis; ii) points the recombination spots with indication of their position relative to the adenine of the mef(A) start codon (position +1); and iii) contains a graphical display of the statistics.
protein). The remaining 300 nt long sequence linking the AlwIcontaining region and the beginning of mef(A) was highly similar to the corresponding region in the other phage types. In addition, the sequence of the whole mef(A)-msr(D) locus (approximate length: 3100 bp) has been determined. The new variant VP_00501.1 has been submitted to GenBank (GenBank: KJ809088). Three strains within the Φ10394.4 group, indicated as “Φ10394.4–like” phage type, did not show the 16-bp duplication at position −140 (relative to the adenine of the mef(A) gene start codon) found in the corresponding region of Φ10394.4. 3.4. Correlation between phage types and the MIC for erythromycin We tempted to correlate the MIC for erythromycin to the sequence features of the mef(A)-msr(D) genetic region (Table 2). We tested the null hypothesis that the medians of the MIC values (4, 8, 12, 16, and 32 μg/mL) within each of the 5 phage-type groups were the same by the Kruskal–Wallis test. The medians of MICs shown by the clades resulting from mef(A) cluster analysis were computed and compared. The same procedure has been independently repeated considering the sequence of msr(D), mef(A)-msr(D) locus, or the carrying genetic element. Even if very close to the 0.05 threshold, the P value was 0.059 meaning that the null hypothesis cannot be rejected at the 95% confidence level. The comparative analysis of paired groups showed a maximum difference between the distribution of the MIC values of the ΦMB56Spy045- in respect to those of Φm46.1- and VP_00501.1carrying strains. Hence, the phage type per se was not associated with
Table 2 Kruskal–Wallis test statistics to compare the medians of MICs. Genetic structure considered
Test statistic
P value
mef(A) msr(D) mef(A)-msr(D) genetic element
13.334 2.08 11.428 9.086
0.01⁎ 0.556 0.022⁎ 0.059
⁎ Statistically significant difference amongst the medians at the 95.0% confidence level.
a peculiar erythromycin resistance level. A correlation was recorded only if either mef(A) alone or mef(A)-msr(D) was considered as the reference genetic structures for the analysis (Table 2). 3.5. Alignment and phylogenetic reconstruction The sequences of the mef(A), msr(D), and mef(A)-msr(D) locus plus the upstream untranslated region were aligned and compared. Cluster analyses have been performed following a stepwise inside-out design considering at first only the contribution of the mef(A)-coding sequence to the observed MIC values. Then the analysis have been extended considering msr(D) alone and at last the mef(A)msr(D) locus plus the noncoding regulatory region upstream to mef(A). Results for mef(A) are shown in Fig. 2. The tree with the highest log likelihood (−1811.0968) is shown. Initial tree(s) for the heuristic search were obtained by applying the neighbor-joining method to a matrix of pairwise distances estimated using the maximum composite likelihood approach. All positions with less than 0% site coverage were eliminated. That is, less than 100% alignment gaps, missing data, and ambiguous bases were allowed at any position. Some clades were composed by strains showing almost identical MIC values, namely, no. 3 and 5, while others were highly heterogeneous (e.g., no. 4) with MICs spanning from 8 μg/mL up to 32 μg/mL. Results for msr(D) revealed that variability of this gene within the set of considered strains is less if compared to that of mef(A). There was a collapse of cluster nos. 4 and 5 identified by mef(A) analysis into a single one when msr(D) alone was considered. Also within cluster nos. 1 and 3, the msr(D) sequences were all identical, while in the mef(A) analysis, strains 31311 and 20411 were slightly different within clade no. 1 and m46 reference strain was different from the other mef(A) sequences clustering within no. 3 (Fig. 2). In the case of msr(D) analysis, the association between sequence of msr(D) and MIC is not appreciable as for mef(A) (P = 0.56). The same phylogenetic and correlation analysis has been performed also for the entire mef(A)-msr(D) sequence locus including the noncoding region upstream to mef(A) (300 nt) and the interspace sequence between the 2 genes (about 90 nt). The result was substantially superimposable on
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
4
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Fig. 2. Molecular phylogenetic analysis of mef(A) by maximum likelihood method and correlation with MIC of erythromycin. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site (bootstrap value: 1000). The analysis involved 29 nucleotide sequences (28 isolates plus the reference strain m46). The MIC of erythromycin against each strain is indicated by a colored dot as per the figure legend.
that obtained for mef(A) alone (Fig. 2), indicating that the major contribution to the locus variability and to the measurable degree of association with the MIC values was mostly due to mef(A). 3.6. Single Nucleotide Polymorphisms (SNPs) and protein polymorphisms analysis Nineteen SNPs were mapped in mef(A), 7 of which led to synonymous substitution and 12 to nonsynonymous ones. msr(D) showed an inverted pattern with 14 synonymous and 6 nonsynonymous substitutions. The 2 distributions of synonymous/nonsynonymous substitutions within the 2 genes were significantly different (P = 0.038, chi-square test). In the mef(A) coding region, SNPs were confined to the distal 5′half of the gene. The same general structural feature in mef(A) has been previously observed in a large collection of macrolide-resistant GAS by Blackman Northwood et al. (2009). It is likely that changes in the N-terminal part of the protein are not tolerated. Nonsynonymous replacements were predicted to be neutral with the only exception of K481M in the sequence of MsrD. Differences in the synonymous– nonsynonymous substitution partitioning between mef(A) and msr(D) were significant (chi-square; P b 0.05) and associated with the phylogenetic study that showed a major contribution to variability of the locus deriving from mef(A). Although synonymous mutations have no effect on the protein sequence, the redundancy of the genetic code has been identified as important in shaping gene expression and cellular function (Plotkin and Kudla, 2011). We also computed the dN/dS ratio to investigate the evolutionary significance of this 2-gene system conferring resistance to macrolides in GAS (Table 3). The dN/dS values were determined for each clade to clade comparison for mef(A) alone, msr(D) alone, and the mef(A)-msr(D) locus. A gene is under positive (Darwinian) selection if dN/dS ratio is higher than 1, while a dN/dS ratio b1 indicates a purifying (stabilizing) selection. The average ratio is higher for mef(A) compared to the one computed for either msr(D) alone or the entire locus (mean dN/dS = 0.518; 95% CI [0.518 ± 0.303], P b 0.05). By comparison of the dN/dS ratio computed for sequence of clade 3 with those for clades 1 and 2, we obtained values
of 1.357 and 1.206, respectively. As Mef(A) is a transmembrane protein, we used TOPCONS to predict the cytoplasmic, extracellular, and membrane-spanning regions (Bernsel et al., 2009). We then mapped the SNPs to visualize their position along the protein. The analysis revealed that the N-terminal part of the protein, which was not polymorphic, consisted of 4 transmembrane domains. As expected, no changes mapped in the transmembrane hydrophobic sequences with the exception of position 393, which presented a Phe to Cys substitution in all isolates except those of clade no. 3 (Fig. 2 and Fig. S1). Conversely, all the other nonsynonymous substitutions changed amino acids exposed either to the cytoplasm or the periplasm. No apparent association was found between a particular amino acid substitution or location and MIC of erythromycin. 3.7. Maximum chi-square test for recombination According to phylogenetic reconstruction, different parental mef(A)msr(D) locus sequences belonging to different clades within the tree were considered for analysis. The test result was confirmatory of plausible recombination events for some of the paired variants. The possible Table 3 Interclade analysis of dN/dS ratio. Clades compareda
1-2 1-3 1-4 1-5 2-3 2-4 2-5 3-4 3-5 4-5 a b
dN/dS mef(A)
msr(D)
mef(A)-msr(D)
0.000 1.357 0.346 0.469 1.206 0.308 0.365 0.349 0.436 0.340
0.051 0.282 0.185 n.a.b 0.000 0.110 n.a.b 0.173 n.a.b n.a.b
0.196 0.789 0.244 0.240 0.151 0.169 0.147 0.248 0.245 0.322
Clades numbering corresponds to that of Fig. 2. n.a. = not available. Tree construction for msr(D) resulted in 4 clades only (Section 3.5).
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
recombination sites and the P value are visualized in Fig. 1 (inner box). The 5′ coding region of mef(A) did not show hotspots for recombination due to its invariability (see analysis of SNPs). Two major recombination sites mapped at a distance of 763 and 936 bp downstream the start codon of mef(A) (A of AUG start codon designated as position +1), while a third one was in the intergenic sequence between mef(A) and msr(D) coding sequences (positions 1315 and 1321). The last observation is supporting the mosaic structure of this locus and the hypothesis that mef(A) and msr(D) may follow an independent evolution, although they were associated in all the isolates studied so far. Actually, authors that have cloned mef(A) for the first time were able to express it into Escherichia coli (Clancy et al., 1996). The opposite was found in S. pneumoniae in which efflux-mediated erythromycin resistance is mediated by the mef(A) homolog mef(E). The association of mef(E) with the downstream ATP-dependent pump gene mel (homolog to msr(D)) was essential for the correct expression of erythromycin resistance (Ambrose et al., 2005). A correspondent conclusive demonstration for GAS is lacking. Some authors have presented preliminary results that address msr(D) (therein referred to as matA) as the main driver of erythromycin resistance (Iannelli et al., 2004). Our genetic study supports the mentioned major role of msr(D) since the level of its variation is lower than that of mef(A). Another probable hotspot for recombination mapped at nucleotide −161 bp, within the noncoding sequence upstream to mef(A) (Fig. 1). Interestingly, 3 strains have the same 16bp duplication at position −140 found in the corresponding region of Φ10394.4, which could be the result of recombination. Similar events may also have contributed to the variability observed in the newly described variant VP_00501.1 (Fig. 1 and Table 1). 3.8. Possible role of the strain genetic background on the expression of erythromycin resistance Mef(A) and Msr(D) are likely to be strongly dependent on the metabolism of the bacterial cells, as they rely on the proton motive force throughout the membrane (Mef(A)) and on the intracellular level of ATP (Msr(D)). Given that each and every bacterial strain shows a different relative metabolism, we determined the MIC of erythromycin against susceptible GAS strains with different emm types after lysogenic conversion to an erythromycin-resistant phenotype by phage Φm46.1 (carrying the mef(A)-msr(D) locus). As emm type variability is a good estimate of the genetic variability of a GAS population, it can indicate a different genetic background and possibly a different growth rate or metabolism (Zampaloni et al., 2003). The results are summarized in Table S2. In all lysogens, the insertion site was the same (not shown) and in agreement with published data (Brenciani et al., 2010). After lysogenic conversion, the MIC of erythromycin did not change for the majority of the lysogens (19 out of 27, 70%). Nevertheless, in 8 strains, the MIC rose to 32 μg/mL. Hence, the genetic background of the bacterial cell might affect the expression of erythromycin resistance to some extent. 4. Conclusions In S. pyogenes, mef(A)-msr(D) locus and its carrying genetic elements undergo a significant level of recombination that may shape their genetic structure. In turn, some of the changes could contribute to the heterogeneous expression of the resistance phenotype eventually measured as MIC of erythromycin ranging from 4 μg/mL up to 32 μg/mL. In addition to that, also horizontal genetic transfer of genetic determinants of resistance in a different genomic context could add to the observed phenotypic variability. In light of our findings, it is not improbable that, in the future, we might witness an even increase in the level of mef(A)-msr(D)–dependent macrolide resistance in S. pyogenes. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.diagmicrobio.2015.08.007.
5
Acknowledgments We are grateful to Dr Valerio Iebba, Dr Raffaella Casalena, Dr Stefania D’Ercole, Dr Wael Bahnan, and Dr Loreta Biqiku for their valuable technical assistance and for helpful discussions. This work was supported by MIUR's grants “Futuro in Ricerca” #RBFR10X4YN to D.P. and “PRIN” #200929YFMK_003 to M.P.
References Ambrose KD, Nisbet R, Stephens DS. Macrolide efflux in Streptococcus pneumoniae is mediated by a dual efflux pump (mel and mef) and is erythromycin inducible. Antimicrob Agents Chemother 2005;49:4203–9. http://dx.doi.org/10.1128/AAC.49. 10.4203-4209.2005. Banks DJ, Porcella SF, Barbian KD, Martin JM, Musser JM. Structure and distribution of an unusual chimeric genetic element encoding macrolide resistance in phylogenetically diverse clones of group A Streptococcus. J Infect Dis 2003;188:1898–908. http://dx. doi.org/10.1086/379897. Bernsel A, Viklund H, Hennerdal A, Elofsson A. TOPCONS: consensus prediction of membrane protein topology. Nucleic Acids Res 2009;37:W465–8. http://dx.doi.org/10. 1093/nar/gkp363. Blackman Northwood J, Del Grosso M, Cossins LR, Coley MD, Creti R, Pantosti A, et al. Characterization of macrolide efflux pump mef subclasses detected in clinical isolates of Streptococcus pyogenes isolated between 1999 and 2005. Antimicrob Agents Chemother 2009;53:1921–5. http://dx.doi.org/10.1128/AAC.01065-08. Brenciani A, Bacciaglia A, Vignaroli C, Pugnaloni A, Varaldo PE, Giovanetti E. Phim46.1, the main Streptococcus pyogenes element carrying mef(A) and tet(O) genes. Antimicrob Agents Chemother 2010;54:221–9. http://dx.doi.org/10.1128/AAC.00499-09. Clancy J, Petitpas J, Dib-Hajj F, Yuan W, Cronan M, Kamath AV, et al. Molecular cloning and functional analysis of a novel macrolide-resistance determinant, mefA, from Streptococcus pyogenes. Mol Microbiol 1996;22:867–79. http://dx.doi.org/10.1046/ j.1365-2958.1996.01521.x. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial susceptibility testing; twenty first informational supplement. CLSI; 2011. Di Luca MC, D’Ercole S, Petrelli D, Prenna M, Ripa S, Vitali LA. Lysogenic transfer of mef(A) and tet(O) genes carried by Φm46.1 among group A streptococci. Antimicrob Agents Chemother 2010;54:4464–6. http://dx.doi.org/10.1128/AAC.01318-09. Hasegawa M, Kishino H, Yano T. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J Mol Evol 1985;22:160–74. http://dx.doi.org/10.1007/ BF02101694. Huang CY, Lai JF, Huang IW, Chen PC, Wang HY, Shiau YR, et al. Epidemiology and molecular characterization of macrolide-resistant Streptococcus pyogenes in Taiwan. J Clin Microbiol 2014;52:508–16. http://dx.doi.org/10.1128/JCM. 02383-13. Iannelli F, Santagati M, Doquier JD, Cassone M, Oggioni M, Rossolini G, et al. Type M resistance to macrolides in streptococci is not due to the mef (A) Gene, but to mat (A) encoding an ATP-dependent efflux pump. 44th ICAAC – 2004 Oct 30-Nov 2. Washington, DC; 2004. p. C1‐1188. Jolley KA, Feil EJ, Chan M-S, Maiden MCJ. Sequence type analysis and recombinational tests (START). Bioinformatics 2001;17:1230–1. http://dx.doi.org/10.1093/bioinformatics/17.12.1230. Lonks JR, Garau J, Gomez L, Xercavins M, Ochoa de Echagüen A, Gareen IF, et al. Failure of macrolide antibiotic treatment in patients with bacteremia due to erythromycinresistant Streptococcus pneumoniae. Clin Infect Dis 2002;35:556–64. http://dx.doi. org/10.1086/341978. Plotkin JB, Kudla G. Synonymous but not the same: the causes and consequences of codon bias. Nat Rev Genet 2011;12:32–42. http://dx.doi.org/10.1038/nrg2899. Santagati M, Iannelli F, Cascone C, Campanile F, Oggioni MR, Stefani S, et al. The novel conjugative transposon Tn1207.3 carries the macrolide efflux gene mef(A) in Streptococcus pyogenes. Microb Drug Resist 2003;9:243–7. http://dx.doi.org/10. 1089/107662903322286445. Seppälä H, Nissinen A, Yu Q, Huovinen P. Three different phenotypes of erythromycinresistant Streptococcus pyogenes in Finland. J Antimicrob Chemother 1993;32: 885–91. http://dx.doi.org/10.1093/jac/32.6.885. Silva-Costa C, Ramirez M, Melo-Cristino J. Identification of macrolide-resistant clones of Streptococcus pyogenes in Portugal. Clin Microbiol Infect 2006;12(6):513–8. [cited 2014 Jan 7]. Smith JM. The detection and measurement of recombination from sequence data. Genetics 1999;153:1021–7. Sutcliffe J, Tait-Kamradt A, Wondrack L. Streptococcus pneumoniae and Streptococcus pyogenes resistant to macrolides but sensitive to clindamycin: a common resistance pattern mediated by an efflux system. Antimicrob Agents Chemother 1996;40: 1817–24. Syrogiannopoulos GA, Grivea IN, Al-Lahham A, Panagiotou M, Tsantouli AG, Michoula Ralf René Reinert AN, et al. Seven-year surveillance of emm types of pediatric group A streptococcal pharyngitis isolates in Western Greece. In: de Lencastre H, editor. PLoS One, 8. ; 2013. p. e71558. http://dx.doi.org/10.1371/ journal.pone.0071558. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 2011;28:2731–9. http://dx.doi. org/10.1093/molbev/msr121.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007
6
L.A. Vitali et al. / Diagnostic Microbiology and Infectious Disease xxx (2015) xxx–xxx
Vitali LA, Zampaloni C, Prenna M, Ripa S. PCR M typing: a new method for rapid typing of group a streptococci. J Clin Microbiol 2002;40:679–81. http://dx.doi.org/10.1128/ JCM.40.2.679-681.2002. Wajima T, Chiba N, Morozumi M, Shouji M, Sunaoshi K, Sugita K, et al. Prevalence of macrolide resistance among group a streptococci isolated from pharyngotonsillitis. Microb Drug Resist 2014;57:1383–8. http://dx.doi.org/10.1089/mdr.2013.0213.
Weisblum B. Erythromycin resistance by ribosome modification. Antimicrob Agents Chemother 1995;39:577–85. http://dx.doi.org/10.1128/AAC.39.3.577. Zampaloni C, Cappelletti P, Prenna M, Vitali LA, Ripa S. emm Gene distribution among erythromycin-resistant and -susceptible Italian isolates of Streptococcus pyogenes. J Clin Microbiol 2003;41:1307–10. http://dx.doi.org/10.1128/JCM.41. 3.1307-1310.2003.
Please cite this article as: Vitali LA, et al, Correlation between genetic features of the mef(A)-msr(D) locus and erythromycin resistance in Streptococcus pyogenes..., Diagn Microbiol Infect Dis (2015), http://dx.doi.org/10.1016/j.diagmicrobio.2015.08.007