MtsR is a dual regulator that controls virulence genes and metabolic functions in addition to metal homeostasis in GAS

SUMMARY

MtsR is a metal-dependent regulator in the Group A Streptococcus (GAS) that directly represses the transcription of genes involved in heme and metal uptake. While MtsR has been implicated in GAS virulence, the DNA recognition and full regulatory scope exerted by the protein are unknown. In this study we identified the shr promoter (P_shr) and mapped MtsR binding to a 69 bp segment in P_shr that overlaps the core promoter elements. A global transcriptional analysis demonstrated that MtsR modulates the expression of 64 genes in GAS, 44 of which were upregulated and 20 were downregulated in the mtsR mutant. MtsR controls genes with diverse functions including metal homeostasis, nucleic acid and amino acid metabolism, and protein fate. Importantly, the MtsR regulon includes mga, emm49, and ska, which are central for GAS pathogenesis. MtsR binding to the promoter region of both negatively and positively regulated genes demonstrates that it functions as a dual regulator. MtsR footprints are large (47-130 bp) and vary between target promoters. A 16 bp motif that consists of an interrupted palindrome is implicated in the DNA recognition by the metalloregulator. In conclusion, we report here that MtsR is a global regulator in GAS that shapes the expression of vital virulence factors and genes involved in metabolic functions and metal transport, and we discuss the implications for the GAS disease process.

INTRODUCTION

The group A streptococcus (GAS, Streptococcus pyogenes) is an obligate human pathogen that produces various infections of the upper respiratory tract, cutaneous, and subcutaneous tissue. Superficial and benign infections, such as pharyngitis and impetigo, are the most frequent outcome of GAS infections. In rare instances GAS disseminates in the human body and produces a number of severe systemic manifestations, such as necrotizing fasciitis, myositis, osteomyelitis, and streptococcal toxic shock syndrome (Cunningham, 2000, Cunningham, 2008). GAS versatility is based on a wide and sometimes redundant repertoire of virulence factors that contribute to the different infection stages, types, and pathology (Tart et al., 2007, Olsen et al., 2009). The antiphagocytic M protein, the hyaluronic acid capsule, C5a peptidase (ScpA), SpeB, and the plasminogen-activator streptokinase (Ska) are expressed by most GAS strains and are considered to be key virulence factors of this human pathogen (Tart et al., 2007, Olsen et al., 2009).

For colonization and survival within different niches in the human host GAS relies on its ability to acquire energy sources and essential nutrients, such as metals, which can be scarce during infection. At least two transport systems for the acquisition of heme, the most abundant iron form in mammals, are employed by GAS. The Sia (Hts) system is a transporter that works in conjunction with the hemoprotein receptor, Shr, and the heme-binding protein, Shp, to obtain and deliver heme (Bates et al., 2003, Nygaard et al., 2006, Zhu et al., 2008). The second machinery involved in hemoprotein utilization is a transporter called Siu (Fts), which was also reported to mediate uptake of ferric-ferrichrome (Montanez et al., 2005, Hanks et al., 2005). Free iron and manganese cations are taken up by GAS using a multi-metal transporter called Mts (Janulczyk et al., 2003). Inactivation of shr, siuG, or the mts transporter leads to virulence attenuation (Janulczyk et al., 2003, Fisher et al., 2008, Montanez et al., 2005), demonstrating that iron acquisition is imperative for disease production by GAS. The importance of zinc (Zn) for GAS growth was only recently established (Weston et al., 2009). Lsp is a ligand binding-protein that mediates Zn uptake in GAS. Mutations in the Zn⁺² binding pocket of Lsp result in decreased virulence, illustrating the significant role Zn homeostasis has on the infection process in GAS (Weston et al., 2009).

GAS adaptation to the human environment relies on complex and cross-interacting regulatory circuits, which interpret diverse environmental signals and produce a dynamic expression profile that is adjusted to the nature of the infection, its stage, and site (Tart et al., 2007). A common regulatory theme used by GAS is the two-component system (TCS). About 13 TCSs are encoded by GAS including the CovR/S (CsrR/CsrS, (Levin & Wessels, 1998)), which is among the better-understood and important systems. CovR/S negatively regulates about 15% of the GAS transcriptome including many virulence factors such as SpeB and Ska (Graham et al., 2002). The CovR/S system is required for GAS growth under stress conditions in vitro (Dalton & Scott, 2004), and is suggested to regulate the transition from mucosal infection to invasive disease in vivo (Sumby et al., 2006). In addition to TCSs, GAS also uses several response regulators (RRs) without known cognate sensor kinases; these encompass Mga, the RofA-like protein family, and Rgg/Rop (McIver, 2009). Mga is a key stand-alone RR in GAS, affecting about 10% of GAS genome during exponential growth. Mga is responsive to CO₂, temperature, iron and metabolized sugars and directly activates a number of important virulence factors. Most of the Mga-regulated genes are anchored to the surface and are involved in adherence, invasion, and immune evasion (including the M and M-like proteins and ScpA). In addition, Mga influences sugar utilization and fatty acid metabolism (Ribardo & McIver, 2006).

In addition to RRs, other types of transcriptional regulators are employed by GAS to coordinate gene expression. PerR belongs to a family of redox-sensing metalloproteins, in which oxidative conditions stimulate the oxidation of selected histidine residues and the subsequent release of the regulator from the DNA (King et al., 2000). GAS uses PerR to control oxidative-stress response, metal homeostasis, and sugar utilization (Gryllos et al., 2008, Brenot et al., 2005). PerR mutants are more sensitive to phagocytic killing and are less virulent in several infection models (Gryllos et al., 2008, Brenot et al., 2005, Ricci et al., 2002). MtsR is a second metal-dependent regulatory protein found in GAS, which belongs to the DtxR-MntR family of metalloregulators. The mtsR gene is located upstream and is divergently transcribed from the mts locus. Previous studies demonstrated that MtsR directly represses the sia and the mts transporters and affects the expression of nrdFEI and ahpCA genes, which are involved in nucleotide metabolism and oxidative stress, respectively (Hanks et al., 2006, Beres et al., 2006, Bates et al., 2005). A null mtsR mutant has elevated level of intracellular iron, is hypersensitive to streptonigrin and hydrogen peroxide, and is attenuated for virulence in a zebrafish infection model (Bates et al., 2005). A comparative genome analysis of GAS clinical isolates revealed that strains containing naturally occurring truncations in mtsR are significantly underrepresented among necrotizing fasciitis cases (Beres et al., 2006). In this study we test the hypothesis that MtsR is a valuable component of GAS regulatory pathways with a significant contribution to pathogenesis by using genomic microarray analysis and carrying out a detailed investigation of the interactions of MtsR with target promoters.

RESULTS

The regulatory elements of P_shr, the sia operon promoter

MtsR directly represses the expression from the sia and the mts operons in iron and manganese dependent manner (Hanks et al., 2006, Bates et al., 2005). The MtsR operator, however, has not been identified to date and none of the promoters it regulates were previously characterized. To obtain a better understanding of gene regulation by MtsR, we examined the interactions of MtsR with the promoter that controls expression from the sia operon. We first used 5′Race to determine the transcription start site(s) in the 360 bp region upstream of shr, the first gene in the sia operon. Since separate initiation sites may be used in the presence or the absence of MtsR, the analysis was performed with RNA from both the wild type and the isogenic mtsR mutant strains. RNA was prepared from GAS cells grown in metal complete medium. The 5′Race produced a single PCR product (320bp) from both the wild type and the mtsR mutant strains (Fig. 1A). Sequence analysis identified an A residue found 52 bp upstream of the shr coding region as the transcription start site to (+1, Fig. 1B). A repeat analysis with different shr-specific primers resulted in an independent confirmation of the same transcription initiation site. Therefore, a single promoter appears to control the expression from the sia operon (P_shr). The region upstream of the P_shr start site contains a putative promoter with a perfect −35 (TTGACA) hexamer, a putative −10 (TAGATT) box, and a 17 base spacer (Fig. 1B).

Fig. 1.

The regulatory region of shr. A. Determination of shr transcription start site by 5′Race. A schematic representation of shr upstream region and an agarose gel with the 5′Race products are shown. The analysis was done with RNA isolated from the wild type (WT) and the mtsR mutant (mtsR⁻) isogenic strains. M stands for DNA size marker. B. An overview of shr regulatory elements. The transcription start site (+1), the −10 and −35 promoter elements, the ribosomal binding site (RBS), and Shr start codon are indicated in bold. The MtsR/DNA region of interaction (including both DNase I protected as well as the hypersensitive residues) is indicated by the solid lines above and below the coding and template strands. The IR1 and IR2 inverted repeats are marked. C. Sequence alignment of MtsR and ScaR footprints (Jakubovics et al., 2000).

To elucidate how MtsR interacts with and regulates P_shr we constructed and investigated several transcriptional fusions to the luc reporter gene. The expression from plasmid pCHT21, in which the entire intergenic region upstream of shr is fused to the luc gene (Fig. 2A), was 2.4 fold higher in the mtsR mutant than it was in the wild type strain. RT-PCR analysis revealed a 6-fold difference in shr expression between these strains (Table S2), suggesting that MtsR repression of P_shr is underrepresented by the luc reporter, possibly due to the plasmid copy number. Transcriptional fusions to two shorter fragments derived from the region upstream of the identified promoter (pCHT22 and pCHT23, Fig. 2A) did not produce any luciferase activity, confirming the absence of additional promoters. We used site directed mutagenesis to study the importance of selected residues in the putative −10 hexamer (TAGATT). Changing the thymines at positions −7 and −8 to guanines (pCHT32, Fig. 2B) resulted in loss of the promoter activity in both strains, providing support for the assertion that the TAGATT sequence is a functional TATA box.

Fig. 2.

Analysis of shr promoter. The schematic representations of the transcriptional fusions to the luc reporter gene and the associated luciferase activity are shown. A. Reporter fusion to fragments from shr upstream sequence. The plasmid names, the fragment from shr upstream sequence that is carried by each plasmid (thin line), the shr transcription start site (+1), and the luc gene (gray arrow) are shown. B. Mutational analysis of shr promoter elements. The sequence of the wild type shr promoter with its −35 and −10 elements (bold) and the transcription start site (+1) found in the transcriptional fusion carried by plasmid pCHT21 are shown. The different mutations introduced in shr promoter and the names of the plasmids that carry them are indicated. Plasmid pCHT31 carries a transcriptional fusion of the entire shr upstream region (1-360 bp) with the G₋₁₄ to A mutation. Luciferase activity expressed by each strain was determined as described in Material and Methods. Each data point represents the average and the standard deviation derived from at least 3 experiments from independent clones done in quadruplicates. The significance of the difference in luciferase activity between the wild type and the mtsR⁻ strain was examined by Student’s t test.

Sequence inspection revealed the presence of a TG motif at the −14 and −15 positions respectively (Fig. 1B), which may function as an extended −10 element (Mitchell et al., 2003, Voskuil & Chambliss, 1998). The role of the TG residues in P_shr was tested by site directed mutagenesis. Replacing the thymine at position −15 with guanine (pCHT33, Fig. 2B) eliminated the promoter activity demonstrating that T₋₁₅ is critical for P_shr function. Changing the guanine at position −14 to adenine resulted in a 4.7-fold increase in the promoter activity in the wild type strain (pCHT31, Fig. 2B). This observation shows that the residue identity in the −14 position is also important for the function of P_shr. The luc expression from pCHT31 was 3.8 fold higher in the mtsR mutant strain in comparison to the wild type strain, indicating that the G₋₁₄ to A mutation increases the strength of P_shr without preventing its repression by MtsR.

Mapping MtsR binding to P_shr

MtsR was previously shown to bind to a DNA fragment containing the entire upstream region of shr (Bates et al., 2005). We used EMSA with 3 fragments encompassing the shr upstream sequence to locate the MtsR operator (Fig. 3A). A DNA shift was generated with a fragment that contained the first 89 bp upstream of shr coding region (F1); this fragment includes the promoter identified by the 5′Race analysis. MtsR bound specifically and with high affinity (Kd of 7.9 × 10⁻¹⁰) as indicated by the competition experiments done in the presence of unlabeled specific DNA (positive control) or P_gyrA fragment (non specific negative control). MtsR did not bind the other fragments (F2 and F3), which contain the DNA region upstream to the first 89 bp, even in the presence of higher MtsR concentrations (Fig. 3B).

Fig. 3.

MtsR binding to shr promoter region. A. Schematic representation of shr promoter region. The thin solid lines represent the 3 DNA fragments used in the EMSAs: F1 (89 bp), F2 (159 bp), and F3 (160 bp). The DNA fragment used in the DNase I footprinting analysis (F4, 193 bp) is indicated by a thick solid line. B. Labeled DNA fragments (50 pmol) from shr promoter were incubated with increasing concentration of rMtsR, and the formation of DNA protein complex was analyzed by EMSA. Some reactions (indicated by the + sign) were done in the presence of 10-fold access of unlabeled DNA fragment (F1) or unlabeled DNA fragment carrying the promoter region of the gyrA gene (P_gyrA).

DNase I footprinting experiments done with a fragment (F4, Fig. 3A) that covers the first 193 bp upstream of shr coding sequence yielded an extensive region of contact between MtsR and the DNA that includes P_shr sequence (Fig. 4). Experiments done with a larger DNA fragment revealed that MtsR did not protect any region upstream of the identified shr promoter (data not shown). MtsR binding to DNA was not symmetrical, resulting in a partial overlap between the footprints on the two strands. On the coding strand, the repressor interacts with a 62 bp region, starting at the −37 position. MtsR footprint on the template strand was larger (69 bp, staring at the −14 position) and the protected residues were not found on a continuous segment (Fig. 1B and Fig. 4). Several residues on both strands were hypersensitive to DNase I; most noticeable was the thymine residue at position +36 on the template strand. The 10 bp immediately upstream of T₊₃₆, were resistant to DNase I digestion even in the absence of MtsR. The DNase I-resistant region is part of an inverted repeat (IR2 in Fig. 1B). A second inverted repeat is also found within the MtsR-protected region upstream of IR2, which overlaps the −10 box of P_shr (IR1 in Fig. 1B). The presence of 2 inverted repeats within the DNA binding region of the shr promoter is reminiscent of the DNA region recognized by ScaR, a closely related metalloregulator from S. gordonii (Jakubovics et al., 2000, Bates et al., 2005). Examination of the footprints of MtsR and ScaR (in P_shr and P_scaC respectively) revealed that significant sequence conservation is limited to the first inverted repeat in both DNA binding regions (IR1, Fig. 1C). Interestingly, while ScaR binds to both repeats, the repressor has significantly higher affinity to the first one (Stoll et al., 2009).

Fig. 4.

DNase I footprinting of MtsR complex with shr promoter. DNA fragments (1 μM) were incubated with DNase I in the presence of increasing concentration of purified rMtsR protein. The 5′end of the nontemplate strand was radiolabeled on the coding strand panel, and the DNA was incubated with 0, 8, 16, 24, 32, 40, 48, 56, and 64 μM MtsR. The 5′end of the template strand was labeled on the template strand panel, and the DNA was incubated with 0, 8, 16, 24, 32, 40, and 48 μM MtsR. The vertical lines specify the 62 bp (on the coding strand) and 69 bp (on the template strand) regions protected by MtsR. Residues hypersensitive to the nuclease digest are indicated by ⋆. The transcription start site (+1) and the −10 and −35 promoter elements are indicated on both strands.

Analysis of the MtsR regulon

Bacterial pathogens often use metallorepressors such as Fur and DtxR to coordinate the expression of genes involved in metal uptake and homeostasis as well as genes that are necessary for the production of infection and pathology. The full spectrum of genes that are regulated by the GAS DtxR-like regulatory protein MtsR has not been previously investigated. To determine the boundaries of MtsR regulation, we compared the transcriptome of a null mtsR mutant (Bates et al., 2005) with that of an isogenic wild type strain. Total RNA was isolated from GAS cells grown in iron complete medium and analyzed using a 70-mer oligonucleotide microarray (representing GAS genomes M1, M3, and M18) (Ribardo & McIver, 2006). This analysis revealed that the expression of 64 GAS genes was changed in response to mtsR deletion, demonstrating that MtsR has an extensive regulatory role in GAS. Out of the 64 MtsR-controlled genes, the transcription of 44 genes was elevated 2-35 fold in the mutant strain in comparison to the wt strain (Table S2). In addition, the expression of 20 genes was down regulated 2-19 fold in the absence of MtsR (Table S3). The array analysis was validated by quantitative RT-PCR (qRT-PCR) on 9 differentially regulated genes (Tables S2 and S3), showing a strong correlation, with an R² value of 0.943 (Fig. S1).

Hypothetical proteins with unknown functions make up of the largest group (43%) among the MtsR-repressed genes and a significant portion (30%) of the genes that were positively controlled by MtsR (Fig. 5). Examination of the rest of the MtsR regulon revealed that it is quite complex and contains genes that participate in a variety of cellular roles including regulation. Genes that are activated by MtsR contribute to nucleotide metabolism, transport, cell envelope and regulation. MtsR-repressed genes cover an overlapping and wider spectrum of functions, which also include protein synthesis and fate, and the biosynthesis of amino acids. Most of the MtsR-controlled genes are found in clusters that are likely to be transcribed into polycystronic RNA molecules (Table S2 and S3).

Fig. 5.

Microarray analysis of MtsR-regulated genes. Repressed (striped bars) and activated (black bars) genes were graphed according to their category role assignment (TIGR). Categories with genes representing less than 5% of the total regulated genes have been omitted for simplification.

MtsR repressed genes

The MtsR–repressed regulon included 4 metal transport systems (Table S2). Consistent with our previous findings (Bates et al., 2005), 2-6 fold induction of the 10-gene sia operon (Spy49_1405c-1395c) was observed in the MtsR⁻ background. Five-fold increase in transcription of the PerR-regulated metal transporter pmtA (Spy49_1146c) was also observed in the mtsR mutant. pmtA encodes a P-type (E1-E2 type) ATPase, which contributes to peroxide resistance and was suggested to mediate metal ion export (Brenot et al., 2007). A moderate increase in the transcription of a second putative heavy metal transporter (copA) from the P-type ATPase family was observed as well (2.5 fold, Spy49_1332c). In contrast to the findings in mtsR mutants in M1 and M3 serotypes (Beres et al., 2006, Hanks et al., 2006) the transcriptome analysis revealed that inactivation of mtsR in NZ131 (M49 background) did not significantly affect mts transcription, although EMSA showed that the repressor binds to the MtsR/A intergenic region in NZ131 (data not shown). RT-PCR analysis confirmed the microarray findings that no change in mtsA transcripts was observed in the mtsR mutant strain. However, about two-fold increase in mtsR transcription was found indicating it is autorepressed (data not shown). Similarly, the MtsR-repression of ahpC transcription that was reported in the M3 strain, MGAS9887 (Beres et al., 2006), was not seen in this study. Therefore, like other global regulators in GAS, MtsR demonstrates some strain-dependent variation of its regulon.

A cluster of 10 genes that starts with the aroE gene (Spy49_0450-0460, Fig. 6A) demonstrated the highest increase in gene expression in the absence of MtsR (35 fold, Table S2). aroE encodes a putative shikimate 5-dehydrogenase, which catalyzes the conversion of 3-dehydroshikimate to shikimate. In addition to the aroE gene, the locus also includes a homologue of metK (encoding S-adenosylmethionine synthatase) that is involved in methionine and selenoamino acid metabolism, a putative efflux protein (mefE) and seven additional orfs encoding hypothetical proteins (Fig. 6A). The close proximity of the 10 genes in the aroE cluster and the similar increase in their expression in the mutant strain suggests that they make up an operon that is repressed by MtsR. A high (7-9 fold) increase in transcript abundance in response to mtsR inactivation was also seen in the nrdFI.2E.2 operon (Spy49_0339-0341, Table S2). MtsR control of this operon in MGAS315 strain was previously reported (Beres et al., 2006). The nrdE.2IF.2 genes encode ribonucleotide reductases (RNR), which play an important role in the metabolism of deoxyribonuclotides, and a flavodoxin (Roca et al., 2008). The dnaK-operon, which encodes the hrcA repressor and the heat shock chaperones grpE and dnaK (Roca et al., 2008), was upregulated 2.5-2.8 fold in the mtsR mutant (Spy49_1376c-1374c, Table S2). This observation suggests that the heat shock response in GAS is linked to metal availability. Among the individual genes that were downregulated by MtsR, the observed 3-fold increase in ska expression in the mtsR mutant is significant. The ska gene codes for the enzyme streptokinase, which converts plasminogen to plasmin, an active serine protease. Multiple experimental and epidemiological data suggest that Ska is an important virulence factor in GAS (Olsen et al., 2009, Shelburne et al., 2008).

Fig. 6.

MtsR-repressed genes. A. Schematic representation of the aroE genomic locus in GAS. The genes and the ORF number in NZ131 strains are indicated. The genomic loci of the streptokinase (ska) and the ribonucleotide reductase genes (nrdFIE) in GAS were previously described (Frank et al., 1995, Beres et al., 2006). B. MtsR binding to DNA fragments carrying the promoter regions of ska (407 bp), nrdF.2 (313 bp), and aroE (246 bp). The formation of DNA-protein complex was analyzed by EMSA as described in figure 3B.

MtsR binding to the promoter region of several negatively regulated genes was investigated by EMSA analysis. DNA fragments derived from the upstream regions of aroE, nrdF.2, and ska were incubated with increasing concentrations of purified MtsR and complex formation was examined following electrophoresis (Fig. 6B). Slow migrating DNA bands, produced in a manner that was dependent on MtsR concentration, were observed with all of the tested promoter fragments. The repressor affinity for the nrdF.2 promoter was the highest; a protein DNA complex was formed with the lowest MtsR concentration, exhibiting a dissociation constant of 1.9 × 10⁻¹⁰. The calculated dissociation constants for MtsR binding to ska and aroE were 7.5 × 10⁻¹⁰ and 1.9 × 10⁻⁹, respectively. In each one of the examined promoters, the specific cold DNA efficiently competed with complex formation while the P_gyrA fragment did not. These observations demonstrate that MtsR directly represses the expression from the aroE, nrdF.2 and ska genes by binding specifically to their regulatory region.

MtsR activated genes

Out of the 20 genes that were positively regulated by MtsR, 7 genes are involved in pyrimidine metabolism (Fig. 7A and Table S3). These genes are organized in two putative operons. The first, which begins with the pyrR gene, encodes 5 genes (Spy49_0650-0654, Fig. 7A) and is down regulated 3-4 fold in the mtsR mutant. The second cluster carries the pyrFE genes (Spy49_0712-0713, Fig. 7A) and demonstrates about 2-fold decrease in expression in the absence of MtsR. The transcription from a 4-gene cluster with unknown function was also reduced 4-5 fold in the absence of MtsR (Spy49_0375-0378). Orfs from this cluster code for enzymes involved in the metabolism of peptidoglycan, purines, and amino groups. In addition to regulating metabolic functions, MtsR positively affected the expression of two major virulence genes: the global transcriptional activator mga and the Mga-regulated emm49 (Spy49_1673c and Spy49_1671c respectively, Table S3). The surface M-protein is a key virulence factor that protects GAS from phagocytosis and mediates adherence and invasion. In addition to activating emm transcription, Mga also controls the expression of multiple surface-associated and secreted products that are important for host colonization and immune evasion (Ribardo & McIver, 2006).

Fig. 7.

MtsR-activated genes. A. Schematic representation of the pyrR (Spy49_0650-0654) and pyrFE loci in GAS. The genes and the ORF number in NZ131 strains are indicated. B MtsR binding to DNA fragments carrying the promoter region of mga (405 bp) and pyrF (346 bp). The formation of DNA-protein complex was analyzed by EMSA as described in figure 3B. C. MtsR binding to the promoter region of emm. MtsR binding to the promoter region of emm genes from NZ131 (emm49), MGAS5005 (emm1), and JRS4 (emm6) was examined by EMSA as described in figure 3B.

To determine if the positive effect of MtsR on transcription is direct, we tested MtsR binding to the promoter region of mga and pyrF (Fig. 7B). MtsR/DNA complexes were observed with DNA fragments that contained the upstream region of both genes. The highest affinity was found with the mga promoter, which showed dissociation constant of 6 × 10⁻¹⁰. The dissociation constant of pyrF was 1.8 × 10⁻¹⁰. MtsR binding to all of the tested promoters was specific, as evidenced by the ability of cold promoter fragment to effectively compete with complex formation and the failure of the gyrA promoter region to prevent complex formation. The reduction in emm49 expression in the mtsR mutant was higher than that observed for mga transcript, suggesting that MtsR may affect emm49 in both mga-dependent and mga-independent manner. We therefore tested the binding of MtsR to emm49 promoter (Fig. 7C); as with the mga and pyrF genes, MtsR bound to emm49 promoter specifically and with high affinity (Kd of 1 × 10⁻⁹). GAS strains are classified into class I and II based on a conserved region in M-protein carboxy-terminus. NZ131 (M type 49), the strain used in this study, belongs to class II. Since significant difference in regulatory circuits are often found between the two classes of GAS we tested MtsR binding to the emm promoter in two strains from class I (MGAS5005, M type 1, and JRS4, M type 6). EMSA analysis demonstrated that MtsR bound to the promoter region of both emm1 and emm6, but requiring higher levels of MtsR (Fig. 7C). These observations are consistent with the reported down regulation of emm6 in JRS4 cells grown in iron depleted medium (McIver et al., 1995) and suggest that MtsR directly controls emm expression in response to iron availability. Together these experiments establish that MtsR is a bifunctional regulator that can be directly involved in both repression and upregulation of gene expression.

MtsR DNA recognition

IR1 in the upstream region of shr is a 16 bp segment containing the 5′ATTAA inverting repeat interrupted by 6 bp (Fig. 1C); sequence analysis demonstrated that shr’s IR1 shares 82% identity with the ScaR binding motif and 94% identity with the complementary strand of the interrupted palindrome hypothesized to serve as MtsR’s binding site at the mtsA promoter (Kitten et al., 2000). As MtsR binds to mtsR-mtsA intergenic sequence (Hanks et al., 2006) and data not shown) we used DNase I protection experiments to map the repressor binding to this region (Fig. 8A). This analysis resulted in a 47 bp protection area, which included the motif mentioned above (Fig. 8A and D). Footprint analysis was also used to define MtsR’s interactions with the sequence preceding the ska and aroE genes (Fig. 8B and C). Two large protection zones of 130 and 112 bp were found upstream of ska coding region (Fig. 8D and data not shown) and 124 bp of MtsR contact area was located in the regulatory region of aroE (Fig. 8D). Therefore, MtsR appears to interact with quite large DNA regions, the size of which varies among different genes. Alignment of the sequence protected by MtsR in the upstream region of shr, mtsR/A, ska and the aroE genes confirmed the 16 bp motif located in the upstream regions of mtsR/A and shr and identified a homologous sequence in each of the MtsR protected areas (Fig. 8 and Fig. 9A). One or two copies of the 16 bp interrupted palindrome was also found in the regions proximal to all other genes that are directly regulated (Table S4). We thus hypothesize that the consensus sequence 5′ATTAAGTTNANTTAAT, generated by weight matrix analysis (Fig. 9B), serves at least in part as an MtsR recognition site in targeted promoters. The limited sequence conservation and the significantly larger protection area, however, suggest that characteristics other than primary sequence, such as DNA topology, also contribute to MtsR detection of its binding sites.

Fig. 8.

DNase I footprint analyses of MtsR complex with the promoter region of mtsR/A (A) ska (B) and aroE (C) genes. MtsR footprint generated as described in figure 4 with DNA fragments labeled with the reverse primers ZE251 (mtsA), ZE361 (ska), and the forward primer ZE375 (aroE) are shown. A second MtsR footprint (containing Ska-I motif) was observed in ska upstream region using a DNA fragment labeled with the forward primer ZE360 (data not shown). The identified MtsR binding motif is indicated by vertical line in each footprint, only Ska-II motif is shown. The sequence 5′ to the coding region of mtsR/A, ska and aroE is provided (D). The nuclease protection areas are shown in grey shading. The suggested MtsR binding motif is indicated in bold and arrows designate the inverted repeats. The MtsR motif identified by Kitten et al (Kitten et al., 2000) is shown in mtsA coding strand, and the motif suggested by this study is shown underneath. The promoter signals provided for the ska genes were adopted form Churchward et al (Churchward et al., 2009).

Fig. 9.

Proposed DNA binding motif for MtsR. A. Alignment of the DNA segments containing the suggested MtsR binding motif found within MtsR footprints at different promoters. B. Weight matrix of the suggested MtsR binding motif as found in the promoter regions of the genes directly regulated by MtsR (Table S4).

DISCUSSION

Bacteria manage metal uptake and resistance mechanisms using transcriptional regulators that sense and modify gene expression in response to the nature and availability of metals in their surroundings. MtsR is known to mediate the repression of two loci involved in metal acquisition in GAS in response to iron and manganese: the sia operon, which allows hemoprotein utilization and heme uptake (shr, shp, and siaABC (Lei et al., 2002)(Bates et al., 2005) and possibly cobalt transport (siaFGH, Table S2), and the mtsABC genes, which encodes an iron and manganese transporter (Hanks et al., 2006). In this study, we analyzed expression from the shr promoter, began investigating MtsR DNA recognition, and used global transcriptional analysis to elucidate the role of MtsR in GAS physiology and virulence.

MtsR function and interactions with DNA

We have determined that a single promoter controls shr expression and identified a putative −35 and an extended −10 promoter elements. The loss of promoter activity in pCHT32 (T₋₇T₋₈ to GG mutation, Fig. 2) supports the suggestion that the TAGATT sequence functions as P_shr −10 element, although it does not exclude possible indirect effect on transcription. A subset of B. subtilis promoters have a TRTG element (R = purine) upstream of the −10 sequence, which enhances transcription by stabilizing Open Complex formation (Voskuil & Chambliss, 2002). The finding that both mutations in the −14 and the −15 positions had a significant effect on P_shr activity demonstrates that the identity of the residues in these positions can be very important for GAS promoters as well. T₋₁₅ to G transversion, which reduces DNA melting, eliminated promoter activity (pCHT33, Fig. 2). G₋₁₄ to A transition, which increases DNA melting, amplified P_shr strength (pCHT31, Fig. 2). Therefore, it is possible that these residues affect promoter melting in GAS as they do in B. subtilis.

MtsR directly represses the shr, ska, nrdF.2, and aroE genes by binding to their upstream region specifically and with high affinity (Fig. 3, 6B, and Table S2). We have also found that MtsR binds to the upstream region of pyrF, mga, and emm genes (Fig. 7B & C), which are downregulated in the mtsR⁻ background (Table S3). Therefore, MtsR can exert either a negative or positive effect on gene expression via direct binding to DNA. This is the first demonstration that MtsR functions as a dual transcriptional regulator. While the mechanism involved remains unexplored, a similar observation was made recently for the related orthologue SloR (O′Rourke et al., 2010) as well as the metalloregulators IdeR (Mycobacterium tuberculosis, (Gold et al., 2001)) and DtxR (Corynebacterium glutamicum, (Brune et al., 2006)).

DNase I protection experiments revealed that MtsR/DNA contact regions are significantly larger then those of DtxR, which protects a 27 bp region (Tao & Murphy, 1994) (Schmitt & Holmes, 1994). Interestingly, a large nuclease protection area (73 bp) was also reported for the related protein MntR of C. diphtheria (Schmitt, 2002). The protection zone generated by MtsR at the intergenic region of mtsR-mtsA (47 bp, Fig. 8D) is similar to the 46 bp footprint produced in P_scaC by ScaR. The MtsR footprint, however, varies between promoters, and significantly larger regions were protected by the metalloregulator upstream of shr (69 bp), ska (130 and 112 bp for sites I & II respectively) and the aroE (124 bp) gene (Fig. 4 and 8). The 46 bp ScaR operator binds two ScaR dimers side by side (Stoll et al., 2009) suggesting that MtsR operators in different GAS genes could accommodate multiple dimers of the repressor (whose size is similar to ScaR). Similarly, the MntR operator at P_mntA was suggested to bind three MntR dimers (Schmitt, 2002). Unfortunately, additional information regarding ScaR binding to other promoters is not available nor have the footprints of the other streptococci orthologues been determined. Future investigations are thus required to determine if the large and variable footprints of MtsR are unique to this GAS metalloregulator or if they are also produced by related regulators.

This study identified a 16 bp sequence motif that consists of 5 bp inverted repeats with 6 bp spacer in the upstream regions of all of the promoters that are regulated by MtsR (Fig. 9B and Table S4). We suggest that this nucleotide sequence serves, at least in part, in MtsR DNA recognition. The MtsR signature is very similar to established and predicted operators of several metal responsive homologues including SloR of S. mutans, PsaR from S. pneumoniae, and S. gordonii ScaR (Kitten et al., 2000, Rolerson et al., 2006, Kloosterman et al., 2008, O′Rourke et al., 2010). The DNA recognition elements of the streptococcal regulators include an AT rich inverted repeat that is interrupted by 6 bp yet some of the binding signatures seem to have more loose sequence conservation. The observation that only a short 16 bp sequence is conserved among the large MtsR binding sites in different promoters suggests that MtsR may recognize other characteristics in its binding region in addition to primary sequence, such as DNA topology. Moreover, it seems possible that MtsR binds initially to the DNA through the 16 bp sequence motif and than multimerizes along the DNA by a more promiscuous step. Support for this suggestion comes from the ScaR operator at P_scaC, which contains two adjacent inverted repeats that are loosely related. While a dimer of ScaR binds to the first repeat with significantly higher affinity than to the second repeat (when it is found along), two ScaR dimers bind to the full operator (one to each inverted repeat) in a cooperative manner and with higher affinity (Stoll et al., 2009).

MtsR regulon and its implication for GAS physiology and virulence

Preceding reports suggesting that MtsR has a significant role in GAS virulence prompted the determination of the complete regulatory scope of MtsR and the mechanisms it employs to exert its functions in GAS. Comparing the transcriptome of the wild type strain to that of an isogenic mtsR mutant yielded 3 major observations: 1) MtsR is a global regulator in GAS which modulates the expression of 64 different transcripts. 2) MtsR transcriptional control expands beyond balancing metal uptake and includes additional metabolic functions as well as central virulence factors. 3) MtsR is an important player in the regulatory circuit that controls metal homeostasis and resistance in GAS.

A significant fraction of the MtsR-affected transcripts encode for hypothetical proteins with unknown function (Fig. 5), an observation that interferes with our ability to fully decipher MtsR contribution to GAS physiology and pathogenesis. The aroE locus, for example, consists of 10 orfs, none of which were previously investigated, and several of which lack homology with known proteins (Table S2, Spy49_0450-0460). The function of the aroE gene cluster is of particular interest because of its large upregulation (35 fold) in the absence of MtsR. MtsR did not regulate the second unlinked aroE gene (aroE.1, Spy49_1225c) in GAS nor did it affect the expression of any other enzymes from the shikimate pathway. This suggests that the aroE cluster (Spy49_0450-0460) is involved in a function separate from the biosynthesis of aromatic amino acids and metabolites.

Regulon analysis demonstrated that MtsR has a significant role in the metabolism of nucleic acids. MtsR was required for the full expression of the pyrR and pyrFE operons (Table S3, Spy49_0650-0654 and Spy49_0712-0713 respectively), which include a uracil permease (pyrP) and genes involved in the pyrimidine biosynthesis (soz00240 pathway in KEGG data base). In addition, MtsR strongly repressed the expression of the nrdE.2IF.2 genes (Spy49_0339-0341, Table 2), which encode a ribonucleotide reductase complex (NrdE.2F.2). The GAS genome includes two types of RNRs: a strictly anaerobic, Class III enzyme (nrdDG) and two aerobic RNRs from Class Ib (nrdHEF and nrdF.2IE.2 operons). Unlike the NrdEF enzyme complex, NrdE.2F.2 proteins require the flavodoxin NrdI to function (Roca et al., 2008). This study demonstrated that MtsR represses the transcription of only the nrdF.2IE.2 cluster, suggesting that transcription of this operon will be induced under low metal conditions. PerR also negatively regulates the nrdF.2IE.2 operon (Gryllos et al., 2008), indicating that oxidative stress will promote nrdF.2IE.2 induction. RNRs have a central role in regulating DNA synthesis and in keeping the ratio of DNA content to cell mass constant during growth. It is possible that under unfavorable growth conditions (such as iron depletion or oxidative stress) NrdF.2E.2 is the RNR of choice in GAS.

In addition to controlling nucleic acid metabolism, MtsR negatively regulates the expression of genes involved in peptide uptake, amino acid metabolism, and protein synthesis (Fig. 5 and Table S2). A moderate increase in the expression of the molecular chaperone dnaK and its co-chaperone grpE was seen in the mtsR mutant (~2.5 fold). HrcA is a negative transcriptional regulator that controls the expression of dnaK and groES operons, though a significant induction of dnaK is still observed following heat shock in a hrcA⁻ background in GAS (Woodbury & Haldenwang, 2003). Our finding that MtsR represses the dnaK operon implicates it as the mediator of the hrcA-independent response of dnaK to heat shock. This suggestion is in agreement with the report that mtsR transcription is significantly downregulated at high temperature (Smoot et al., 2001).

An important observation made in this study is that MtsR is involved in the regulation of key virulence factors in GAS. MtsR directly represses the gene encoding streptokinase (ska), which converts the proenzyme plasminogen to the broad-spectrum protease plasmin (Fig. 6 and 8B). GAS streptokinase contributes to bacterial dissemination and tissue damage, and it is implicated in glomerulonephritis pathology and the coagulopathy seen during GAS invasive infections (Nordstrand et al., 1998, Olsen et al., 2009). The mechanisms that control ska expression during infection are not fully understood. However, it was previously demonstrated that the fasABCX TCS system activates ska transcription while the RR CovR directly represses it (Kreikemeyer et al., 2001, Churchward et al., 2009). Since the CovRS system responds to several stress cues including iron limitation by alleviating CovR- mediated gene repression (Froehlich et al., 2009), iron depletion is expected to ease ska repression by MtsR and CovR in a wild type strain. Interestingly, one of the MtsR nuclease protection areas (with the Ska-II motif) overlaps two of the CovR protected regions and the second footprint of MtsR (with the Ska-I motif) resides downstream of another CovR- protected region (Fig. 8D, (Churchward et al., 2009)). Therefore, it is possible that the two repressors interact at the ska regulatory region when the bacterium is growing under high iron conditions. As CovRS respond to multiple stress conditions and MtsR binds to DNA in the presence of manganese as well as iron, conditions present at different stages of the infection can thus modify ska transcription via one or both repressors.

MtsR positively affects the expression of mga and emm genes (2 and 11 fold decrease respectively), which are central for GAS pathogenesis. It is interesting that MtsR regulates both genes directly (Fig. 7B and C), since Mga is the major activator of emm. MtsR function therefore, may be to help shape emm expression independent of Mga. This regulatory role of MtsR it is not strain-specific, as we found that MtsR binds to emm promoters from several M serotypes (M49, M1, M6). The activation of mga and emm transcription by MtsR is consistent with the observation that emm expression decreases in response to iron depletion in the M type 6 strain JRS4 (McIver et al., 1995), and indicates that MtsR mediates this response. Mga is a central regulator of virulence in GAS, and therefore the finding that MtsR directly regulates Mga is very important and suggests that it can link metal availability input into this critical regulatory circuit. While we did not observe changes in expression of Mga-regulated genes other than emm in the mtsR mutant strain under our experimental conditions, it is possible that under certain in vivo conditions MtsR will help fine tune the expression of virulence factors involved colonization and or invasion (such as M-like proteins, C5a, Sic, and several fibronectin proteins all of which are regulated by Mga (Hondorp & McIver, 2007). The high manganese availability in saliva (36 μM (Chicharro et al., 1999)), for example, is likely to promote MtsR binding to mga and emm promoters, and therefore may contribute to the elevated expression of emm observed during pharyngitis (Virtaneva et al., 2003).

Analysis of the MtsR regulon demonstrated that in addition to repressing the expression from the mts and the sia loci, MtsR negatively controls the P-type cation transporters copA and pmtA (Table S2 and Fig. 10). CopA shares 57% identity with a copper efflux ATPase from S. mutans (Vats & Lee, 2001) and may be involved in heavy metal transport in GAS as well. The ligand and the transport direction mediated by PmtA are unknown as well. However, since PmtA promotes resistance to peroxide stress and zinc toxicity it is suggested to carry out metal efflux (Brenot et al., 2007). The finding that pmtA is also part of the MtsR regulon is intriguing, as it is likely that PmtA links the MtsR regulon to that of AdcR (Fig. 10), at least under some growth conditions. In addition to PmtA, the overlap between MtsR and PerR regulons includes 5 other transporters and enzymes (Fig. 10). PerR and MtsR negatively regulate the expression of 4 out of the 6 common genes, and, therefore, their maximum expression may be seen when GAS is experiencing both oxidative stress and metal starvation. Examination of both MtsR and PerR regulons reveals that many of the GAS genes predicted to contribute to GAS metal homeostasis are regulated by either MtsR or PerR, demonstrating that both regulators are key components in determining the GAS metallome.

Fig. 10.

Comparison of MtsR and PerR regulons. The diagram shows GAS genes that are common to both the MtsR and PerR regulons, as well as genes from each regulon that are involved in metal homeostasis (indicated in red) and resistance to oxidative stress. The putative ligand of each system is provided in parentheses. Asterisks indicate a strain-specific regulation. Gene activation is indicated by arrowheads, and repression is shown by blunt ends. Direct regulation is shown with solid lines and indirect (or unknown) regulation is indicated by dashed lines. The genes displayed in the figure are a compilation of work done in this study and others (Ricci et al., 2002, Brenot et al., 2005, Beres et al., 2006, Brenot et al., 2007, Gryllos et al., 2008).

In summary, this work demonstrates that MtsR is a dual regulator that shapes the expression of a large number of GAS genes. Like the related orthologues SloR and PsaR (Dunning et al., 2008, Rolerson et al., 2006, Hendriksen et al., 2009, Johnston et al., 2006, O′Rourke et al., 2010), the scope of MtsR regulation expands beyond metal homeostasis and includes functions that directly contribute to colonization and disease production.

A study by Olsen et al was published while this manuscript was under review describing the analysis of an mtsR mutant in the M type 3 strain MGAS315. This report established that mtsR inactivation results in a significant attenuation of the pathogen’s ability to produce necrotizing fasciitis, due to deregulation of the chaperon prsA, and therefore underscores the importance of MtsR to GAS pathogenesis (Olsen et al., 2010). A comparison of the MtsR regulons in the NZ131 and MGAS315 reveals a significant overlap as well as genes that appear to be differentially regulated in the two strains, further supporting a strain specific component of the core MtsR regulon. In addition, some of the observed differences may result from different experimental conditions (such as the metal composition in the growth medium).

MATERIALS AND METHODS

Strains, media, and growth conditions

Escherichia coli (E. coli) DH5α and BL21 (DE3) were used for cloning and gene expression. The GAS strains used in this study were NZ131, an M type 49 (Simon & Ferretti, 1991), the mtsR isogenic mutant ZE491 (Bates et al., 2005), the M type 6 JRS4, and MGAS5005, an M type 1 strain. E. coli cells were grown aerobically in Luria Bertani (LB) medium at 37°C. GAS cells were grown statically at 37°C in Chemically Defined Medium (CDM; SAFC Biosciences) as previously described (Montanez et al., 2005) or in Todd Hewitt Broth with 0.2% (w/v) yeast extract (THYB; BD laboratories). When necessary, 100 μg/ml spectinomycin, 70 or 300 μg/ml kanamycin (for E.coli and GAS, respectively) was added to the medium.

DNA manipulations

Chromosomal and plasmid DNA extraction and DNA manipulations, including restriction digest, cloning, and DNA transformation into E. coli or GAS, were done according to the manufacturer’s recommendations and with standard protocols as previously described (Eichenbaum et al., 1998, Sambrook, 1989). PCR for cloning was performed using the High Fidelity AccuTaq LA DNA Polymerase (Sigma). PCR products were purified with the QIAquick PCR Purification Kit (Qiagen). DNA sequencing was performed using the SequiTherm Excel II DNA sequencing kit (Epicentre) or by GSU’s core facility. End labeling with [γ-³²P] ATP was performed using T4 Polynucleotide Kinase (Invitrogen or Epicentre). Site directed mutagenesis was done using the QuikChange II XL Site-Directed Mutagenesis Kit (Stratagene). The oligonucleotide primers used in this study are listed in Table S1. Table 1 lists and describes the plasmids used in this work.

Table 1.

Plasmids used in this study.

Plasmid	Description	Reference or source
pZEDH3.1	E. coli expression vector carrying mtsR under the control of the T7 RNA polymerase	Bates et al. (20051)
pKSM720	E. coli/GAS shuttle vector with the firefly luciferase (luc) gene and ribosomal binding site	Kinkel and Mclver (2008)
pCRS/GW/TOPO	E. coliTOPO PCR cloning vector	Invitrogen
pCHT12	E. coliTOPO vector carrying the DNA fragment produced by the 5′Race protocol using NZ131 RNA as a template	This study
pCHT13	E. coliTOPO vector carrying the DNA fragment produced by the 5′Race protocol using ZE491 RNA as a template	This study
pCHT19	E. coliTOPO vector with a fusion of 3′isp-P shr upstream region to the phoZ reporter gene	This study
pCHT21a	E. coli/GAS shuttle vector with a fusion of shr upstream region (bp 1–360) to the luc gene	This study
pCHT22a	The same as in pCHT21, only with a fusion of a fragment (bp 1–175) from shr upstream region to the luc gene	This study
pCHT23a	The same as in pCHT21, only with a fusion of a fragment (bp 1–126) from shr upstream region to the luc gene	This study
pCHT25	E. coliTOPO vector carrying the DNA fragment produced by the 5′Race protocol using NZ131 RNA as a template	This study
pCHT26	E. coli TOPO vector carrying the DNA fragment produced by the 5′Race protocol using ZE491 RNA as a template	This study
pCHT27a	The same as in pCHT19 only with an E. coli/GAS shuttle vector	This study
pCHT30a,b	Same as pCHT27, only with G−14 to A mutation	This study
pCHT31a,b	Same as pCHT21, only with G−14 to A mutation	This study
pCHT32a,b	Same as pCHT21, only with T−7T−3 to GG mutations	This study
pCHT33a,b	Same as pCHT21, only with T−15 to G mutation	This study

^aA schematic presentation of the luc fusions carried by this plasmid is shown in Fig. 2A.

^bThe mutated residue position is given in relation to the shr transcription start site shown in Fig. 2B.

Expression and purification of recombinant MtsR (His₆-MtsR)

Recombinant MtsR (rMtsR, with a C-terminal fusion to His₆ tag) was purified from E. coli according to Bates et al (Bates et al., 2005) with small modifications. Briefly, BL21 (DE3) cells harboring plasmid pZEDH3.1 were grown at 37°C in LB containing ampicillin (100 μg/ml). Protein expression was induced at OD₆₀₀=0.6 by IPTG (isopropyl-β-D-thiogalactopyranoside; Promega). The cells were harvested, resuspended in phosphate buffer (50 mM NaPO₄, 0.5 M NaCl, pH 8.0), and lysed by sonication. Proteins were precipitated with 1% (w/v) streptomycin sulphate and 40% (w/v) ammonium sulphate. The protein pellet was resuspended in binding buffer (50 mM NaPO₄ pH 7.4, 0.5 NaCl, and 20 mM imidazole) and purified over a nickel affinity column (HisTrap HP column; GE Lifesciences). The purified protein was examined by SDS-PAGE and western blot analysis with anti-MtsR antibodies (Fisher et al., 2008). Protein concentration was determined by Bradford assay (Biorad).

Electrophoretic mobility shift assays (EMSAs)

EMSAs were done as previously described. (Bates et al., 2005) with small modifications. Briefly, DNA fragments were amplified from the bacterial chromosomes, end labeled with [γ-³²P] ATP and purified using the Quick Spin Columns (TE) for radiolabeled DNA (Roche). For the DNA shift assays, increasing concentrations of rMtsR (in 20 mM KPO₄, pH 7.0) were incubated with approximately 50 pmol labeled DNA fragment for 15 minutes at room temperature in a reaction buffer containing 20 mM NaPO₄, 50 mM NaCl, 5 mM MgCl₂, 2 mM dithiothreitol, 0.4 mg/ml bovine serum albumin, 0.2 mg/ml sheared salmon sperm DNA (Ambion), and 9.6% (v/v) glycerol. The reaction mixture was fractionated over a 5% (v/v) polyacryamide gel containing 2.5% (v/v) glycerol, 20 mM NaPO₄, and 2 mM dithiothreitol. Dissociation constants (Kd) were calculated using ImageQuantTL 7.0 (GE Healthcare). The density of the free DNA fragment was determined and plotted as a function of MtsR concentrations. Linear regression was used to determine the Kd value (MtsR concentration at which half of the DNA is shifted).

DNase I protection assay

DNase I footprint assays were performed as previously described (Gusa & Scott, 2005) with small modifications. Briefly, fragments containing the promoter region of shr (ZE227 and ZE228), mtsR/A (ZE251 and ZE252), ska (ZE360 and ZE361), or aroE (ZE363 and ZE375) were amplified using [γ-32P] labeled primers (Table S1). The PCR fragments were purified from an 8% (v/v) polyacrylamide gel. Increasing concentrations of rMtsR (in 20 mM KPO₄) were added to a DNA binding buffer containing 1 μM DNA, 33 mM Tris-acetate pH 8.0, 150 μg/ml BSA, 10 mM MgAc and 0.5 mM dithiothreitol. The reactions were incubated for 15 minutes at room temperature before the addition of 2 ng/μl DNase I and subsequently incubated at 37°C for 50 seconds. DNase was inactivated with pellet paint (Novagen) and DNase I stopbuffer (80% (v/v) EtOH and 22.5 mM NH₄OAc). The DNA was precipitated at −80°C, dried at room temperature, and resuspended in Stop/Loading buffer (Epicentre). The reactions were run for approximately 1.5 hour on a 6% (v/v) sequence gel.

Determination of the transcription start site in P_shr

The transcription initiation site was determined using the 5′Race method (5′Race System for Rapid Amplification of cDNA Ends, Invitrogen). Briefly, RNA was isolated from GAS cells harvested at the mid-logarithmic phase. The first strand cDNA was synthesized using the shr specific primer ZE224. Homopolymeric tails were added to the 3′ ends of the cDNA, which was then used as a template in a PCR reaction with the Abridged Anchor Primer (Invitrogen) and the nested shr primer ZE225. A second PCR reaction was done using the UAP primer (Invitrogen) and the shr-specific primer ZE226. The product was cloned into a TOPO vector (pCR8/GW/TOPO TA Cloning Kit, Invitrogen) producing plasmid pCHT12 and pCHT13 (for NZ131 and ZE491 respectively). The 5′Race procedure was repeated using the shr primer ZE225 to produce the cDNA and the shr primers ZE226 and ZE234 for the nested PCR reactions (producing plasmids pCHT25 & pCHT26 for NZ131 and ZE491 respectively). The sequence of the cloned fragments was analyzed to determine the transcription start site.

Construction of P_shr-luc transcriptional fusions

All of the transcriptional fusions used in this study are described in Table 1. Shr promoter fusions were constructed by amplifying DNA fragments from NZ131 chromosomal DNA and cloning them into pKSM720 (Kinkel & McIver, 2008), which carries a promoterless luc gene. For plasmid pCHT21, the upstream region of shr gene (360bp) was amplified with ZE204 and ZE211 primers, digested with BglII and XhoI, and ligated to the BglII/XhoI fragment of pKSM720. The same method was used to construct pCHT22, which carries a 175 bp fragment from the shr upstream region amplified with ZE205 and ZE211 primers. Plasmid pCHT23 harboring a 126 bp fragment from the shr upstream region was cloned the same way, only using primers ZE206 and ZE211. Plasmid pCHT31, containing shr-luc promoter fusion with a mutation at G₋₁₄ of P_shr was constructed using pCHT30 as a template in a PCR reaction with ZE293 and ZE307 primers. The resulting PCR product, containing the promoter fragment, was digested and cloned into BglII and XhoI sites in pKSM720. Plasmid pCHT32, carrying shr-luc promoter fusion with mutations in T₋₇ and T₋₈ of P_shr was constructed using the same method only with the primers ZE310 and ZE311. Plasmid pCHT33, carrying shr-luc promoter fusion with a mutation at T₋₁₅ in P_shr was generated the same way only with ZE312 and ZE313 primers. Plasmid pCHT19 which encodes an shr-phoZ transcriptional fusion, was generated by amplifying the region from the 3′ end of isp to shr gene using ZE166 and ZE167 primers, and cloned into the HindIII site in pCW1 (Bates et al., 2003).

Luciferase Reporter Assay

Luciferase activity in GAS strains harboring luc transcriptional fusions was done as previously described in Leday T.V. et al (Leday et al., 2008). Briefly, culture samples were harvested at OD₆₀₀= 0.5 and the cell pellets were resuspended in Luciferase Cell Culture Lysis buffer. Britelite plus reagent (Perkin Elmer) was added to the cells in a microtiter plate and luminescence was measured using the 1420 Multilabel Counter Victor ³V (Perkin Elmer).

Microarray and real-time RT-PCR validation

Microarray experiments were performed as previously described (Leday et al., 2008). Briefly, oligonucleotide 70-mer probes (2328) were synthesized (Qiagen Operon) to target unique non-repetitive ORFs in the sequenced genomes of serotypes M1 (SF370), M3 (MGAS315) and M18 (MGAS8232). Probes possess a melting temperature of 76 ± 5^°C, ≤70% cross hybridization identity to another gene within the same strain, ≤20 contiguous bases in common with another gene, and probe location within 3′ end of ORF. Microarrays were printed at Microarrays, Inc (http://microarrays.com), with 10 picoliters of each oligonucleotide probe spotted onto slides (UltraGAPS2; Corning) using a 12 pin contact printer. Total RNA from 3 biological replicates was isolated from NZ131 and the isogenic MtsR⁻ mutant strain ZE491 at mid-logarithmic phase (90 Klett units) using the Ambion RNA purification kit. 18 μg of RNA was treated with DNase I and analyzed for quality on formaldehyde-agarose gel. RNA samples were converted to cDNA with an amino-allyl UTP and were labeled with both Cy3 and Cy5 using the Amino Allyl cDNA Labellng Kit (Ambion) to allow for dye-swap experiments. Yield and incorporation of dye was determined using a Nanodrop ND-1000 (Nanodrop Technologies). Equal volumes (35.42 μl) of labeled Cy5 cDNA and Cy3 cDNA were dried under vacuum, resuspended in 23.8 μl of dH₂O and boiled for 5 min followed by cooling on ice for 5 min. 5XHyb Buffer (GE Healthcare, 17 μl) and formamide (27.2 μl) was added to the cDNA and applied to array slides under raised cover slip (Lifterslip, Inc). Microarray slides were hybridized at 50°C overnight in slide chambers (Array It). Slides were washed twice for 10 min each in the following buffer concentrations and temperatures: 6X SSPE/0.01% Tween-20 at 50°C, 0.8X SSPE/0.001% Tween-20 at 50°C, and 0.8X SSPE at RT. Slides were scanned using a Genepix 4100A personal array scanner and GenePixPro 6.0 software (Axon Instruments).

Data obtained from the wild type and mtsR mutant strains were compared for 2-fold changes in expression, ≥ 2.0 or ≤ 0.50, and were analyzed using Acuity 4.0 software (Axon Instruments). Using a ratio-based normalization, data was normalized by the ratio of the means (635/532) and samples were removed when 4 out of the 6 experiments did not show significance. Array validation was carried out by real-time RT-PCR using 9 differentially regulated genes using primers in Table S1. Correlation coefficients for the arrays were determined by plotting the log value of the array on the X-axis to the log value of the real-time RT-PCR on the Y-axis. An equation determining the line of best fit was determined, and the resulting R² value was calculated to be 0.943 (Figure S1), which represented the fitness of the data.

Real-time RT-PCR

Briefly, total RNA was isolated from each strain using the TritonX-100 isolation protocol (Sung et al., 2003) and 25 ng was DNase I-treated, added to a SYBR Green Master mix (Applied Biosystems) containing 5 μg of each specific real-time primer (Table S1), and combined with 6.25 units of Multiscribe reverse transcriptase (Applied Biosystems) in a 25 μl volume for a one step real-time RT-PCR reaction. The real-time RT-PCR experiments were completed using a Lightcycler 480 (Roche) and transcript levels were detected in the relative quantification mode. Samples were compared to WT gyrA transcript levels, with the levels presented representing ratios of WT/experimental.

In silico analysis

The online directory Kyoto Encyclopedia of Genes and Genomes (KEGG) (Okuda et al., 2008) was used for the analysis of metabolic pathways and orfs from GAS. LALIGN program (Huang & Miller, 1991) and Vector NTi (Invitrogen) were used for sequence analysis. Sequence alignment between the MtsR operator at P_shr, ScaR operator at P_scaC, and the motif suggested by Kitten et al (Kitten et al., 2000) identified a conserved 16 bp segment with an interrupted inverted repeat (5′-ATTAAGTTNAGTTAAT). We therefore searched for homologous motifs in the MtsR protected areas at mtsR/A, ska, and aroE promoters as well as in the upstream regionof all of the genes we found to be directly regulated by MtsR. Sequence with at least 70% homology, containing a palindrome, or sounded by inverted repeats (as found in operators of related metalloregulators (Kitten et al., 2000)) were selected (Table S4) and used to generate a weight position matrix.