Three Paralogous LysR-Type Transcriptional Regulators Control Sulfur Amino Acid Supply in Streptococcus mutans

Abstract

The genome of Streptococcus mutans encodes 4 LysR-type transcriptional regulators (LTTRs), three of which, MetR, CysR (cysteine synthesis regulator), and HomR (homocysteine synthesis regulator), are phylogenetically related. MetR was previously shown to control methionine metabolic gene expression. Functional analysis of CysR and HomR was carried out by phenotypical studies and transcriptional analysis. CysR is required to activate the transcription of cysK encoding the cysteine biosynthesis enzyme, tcyABC and gshT genes encoding cysteine and glutathione transporter systems, and homR. HomR activates the transcription of metBC encoding methionine biosynthesis enzymes, tcyDEFGH involved in cysteine transport, and still uncharacterized thiosulfate assimilation genes. Control of HomR by CysR provides evidence of a cascade regulation for sulfur amino acid metabolism in S. mutans. Two conserved motifs were found in the promoter regions of CysR and HomR target genes, suggesting their role in the regulator binding recognition site. Both CysR and HomR require O-acetylserine to activate transcription. A global sulfur amino acid supply gene regulatory pathway is proposed for S. mutans, including the cascade regulation consequent to transcriptional activation of HomR by CysR. Phylogenetic study of MetR, CysR, and HomR homologues and comparison of their potential regulatory patterns among the Streptococcaceae suggest their rapid evolution.

Adaptation of bacteria to new environments often requires acquisition of new functions. Mutation of genes is not considered the primary mechanism for adaptation since it often leads to abnormal products and, after further changes, could lead to loss of the initial function. Alternatively, lateral gene transfer is proposed to play a most important role in the acquisition of new functions, but it does not explain how functions initially arose in the donor. In this respect, gene duplication might be a valuable source for function evolution, since a duplicated gene may evolve without loss of a possibly essential function. Therefore, duplication events lead to the appraisal of a family of genes, which have related sequences but may exert different functions. Genome studies show that such cases occur frequently, and genes that retain their initial functions in different bacteria are often referred to as orthologs, whereas those that underwent changes after duplication are called paralogs (37). However, recent computational studies have suggested that bacterial transcriptional factors that may be orthologous by criteria such as alignment or synteny may have different functions and may regulate different genes due to a very rapid evolution (50). Therefore, study of orthologous and paralogous genes is an important focus of research to ensure correct assignment of their functions in the ever-increasing number of sequenced genomes.

The availability of many genomes of related species offers a good opportunity to detect duplication events, as shown with yeast (34, 35, 48), enterobacteria (50), bacilli (15, 32, 60), and streptococci (9, 23, 41). Streptococcus is a genus that contains a wide variety of species, mostly commensal or pathogenic for humans and animals but also used as food starters. Within the host, streptococci occupy a broad range of ecological niches, such as mucosal surfaces and skin or muscle tissues, and the factors responsible for their niche adaptations remain poorly understood. The particular importance of these bacteria, both in severe pathologies and in food productions, prompted studies aiming to understand their adaptation versatility, in particular, at the genome level. Gene acquisition was shown to be an important feature in the evolution of their genome (3, 17, 36, 40, 45, 62), but also were other processes, such as recombination and positive selection within their core genomes (10, 11, 39).

In this paper, we study the function of two new members of the LysR-type transcriptional regulator (LTTR) family involved in sulfur amino acid metabolism in streptococci. Cysteine and methionine have crucial roles in different metabolisms in the cell, due in particular to the reactivity of their sulfur group, which plays an essential role in the catalytic sites of many enzymes and participates in ion and redox metabolisms (6). Interestingly, regulation of these pathways involves a large variety of mechanisms in different groups of bacteria so that it appears to evolve more rapidly than that of many other regulatory pathways. For example, Bacillus subtilis and Lactococcus lactis, which are used as models of the Firmicutes, do not share orthologous transcriptional regulatory factors for sulfur amino acid pathways. Furthermore, in L. lactis, a member of the Streptococcaceae, most of the cys and met genes are regulated by a single LTTR, CmbR (cysteine/methionine biosynthesis regulator, also named FhuR) (19, 58), which recognizes a 13-bp box located in their promoter regions (19, 26, 58). The absence of CmbR binding sites in the other streptococcal genomes (26, 58) and the presence of a different motif upstream of several potential cys genes in their genomes (38) suggest that the control of sulfur amino acid metabolism may be different in the other members of the closely related Streptococcaceae family. Lastly, study of the streptococcal MetR regulator suggested that, although MetR and its regulatory binding site are conserved among streptococci, the regulatory scheme mediated by MetR is significantly different among several bacteria of this genus (38, 52, 57). Here, we report the role of two additional LTTRs involved in sulfur amino acid transport and synthesis in streptococci. Their respective coinducer, putative binding sites, and regulatory functions were determined. The first, CysR (for cysteine synthesis regulator), appears to be orthologous to CmbR but controls cysteine synthesis genes only by the means of a specific regulatory motif. The second, HomR (for homocysteine synthesis regulator) is a CysR/MetR paralog that controls homocysteine synthesis genes. Finally, likely evolutionary schemes for reporting the diversity of cysteine and methionine regulatory pathways in streptococci are discussed in the light of their specific environmental niches.

MATERIALS AND METHODS

Bacterial strains, plasmids, culture media, and growth conditions.

The bacterial strains used in this study are listed in Table 1. Escherichia coli TG1 was used for plasmid propagation. E. coli was grown in Luria-Bertani medium at 37°C (44). Streptococcus mutans strains were grown at 37°C in M17 glucose medium or in chemically defined medium (CDM) (61) containing glucose (20 g/liter) as a carbon source and all amino acids except methionine and cysteine. The CDM was supplemented with 1 mM sulfide (Na₂S; Sigma), 0.5 mM thiosulfate (Na₂S₂O₃; Merck), 1 mM l-glutathione oxidized (Sigma), or 1 mM l-cystine (Sigma), when indicated. CDM+M designates CDM with 0.67 mM l-methionine (Merck), CDM+C designates CDM with 1.65 mM l-cysteine (Sigma), and CDM+CM designates CDM with l-cysteine and l-methionine. The effect of O-acetylserine (OAS; Sigma), thiosulfate (Merck), N-acetylserine (NAS; Sigma), sulfide (Merck), or dl-homocysteine (Sigma) on transcription was tested by the addition of 4 mM concentrations of these compounds to exponentially growing cells (optical density at 600 nm [OD₆₀₀] of 0.2) in CDM+CM. When required, 5-bromo-4-chloro-3-indolyl-β-d-galactoside (0.04 g/liter), isopropyl 1-thio-β-d-galactopyranoside (IPTG; 0.04 g/liter), ampicillin (100 μg/ml for E. coli), erythromycin (8 μg/ml for S. mutans and 100 μg/ml for E. coli), and tetracycline (3 μg/ml for S. mutans) were added to the culture medium.

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant characteristics	Reference or source
E. coli strain TG1	supE Δthi(lac-proAB) hsdD5 (F′+traD36 proAB lacIqZΔM15)	25
S. mutans strains
UA159	S. mutans wild-type strain (ATCC 700610)	ATCC
JIM8877	UA159 ΔcysR	This work
JIM8878	UA159 ΔhomR, Emr	This work
JIM8879	UA159 ΔgshT, Emr	This work
JIM8880	UA159 ΔcysK, Emr	This work
JIM8882	UA159 ΔtcyA, Emr	This work
JIM8884	UA159 ΔtcytE, Emr	This work
JIM8886	UA159 ΔtcyA ΔtcytE, Emr Tetr	This work
JIM8890	UA159 containing pJIM5788 integrated at the PhomR locus, Emr	This work
JIM8892	UA159 containing pJIM5790 integrated at the PtcyD locus, Emr	This work
JIM8887	UA159 containing pJIM5786 integrated at the PcysK locus, Emr	This work
JIM8888	UA159 containing pJIM5787 integrated at the PcysK locus, Emr	This work
JIM8889	JIM8877 containing pJIM5786 integrated at the PcysK locus, Emr	This work
Plasmids
pGEM-T easy	Apr, M13ori pBR322ori, linear T-overhang vector	Promega
pGhost 9	Emr, ori+ ΔrepA, integrative promoter probe vector	43
pJIM4900	Emr, ori+ ΔrepA, integrative promoter probe vector containing luxAB genes	57
pJIM5785	SalI fusion of pGhost 9 and pGEM-T Easy containing 818-bp fragment upstream and 850-bp fragment downstream cysR gene ligated in BamHI	This work
pJIM5788	PHomR-lux-SpeI site fusion of pJIM4900 and pGEM-T Easy containing PhomR on a 551-bp fragment	This work
pJIM5786	PcysK-lux-BclI/BamHI site fusion of pJIM4900 and pGEM-T Easy containing Pcysk on a 591-bp fragment	This work
pJIM5787	PcysK-lux-BclI/BamHI site fusion of pJIM4900 and pGEM-T Easy containing PcysK carrying nucleotide substitution on the CysR box on a 591-bp fragment	This work
pJIM5790	PtcyD-lux-SpeI site fusion of pJIM4900 and pGEM-T Easy containing PtcyD on a 591-bp fragment	This work

DNA manipulation procedures.

DNA manipulations and E. coli transformation were carried out as described previously (58). All enzymes for DNA technology were used according to the manufacturer specifications. Transformation of S. mutans by natural competence was performed as described by Murchison et al. (47). Primers used in this work were synthesized by Sigma Genosis (see Table S1 in the supplemental material). DNA sequencing was performed on both strands using a fluorescent sequencing procedure.

Gene deletion in S. mutans.

The chromosomal homR (SMU.930c; GeneID 1028282), gshT (SMU.1942c; GeneID 1029139), tcyA (SMU.459; GeneID 1027964), tcyE (SMU.933; GeneID 1028287), and cysK (SMU.496; GeneID 1027990) genes were deleted in S. mutans strain UA159 as described previously (57). Briefly, two DNA fragments of about 750 bp corresponding to the upstream and the downstream regions of each gene were amplified by PCR using specific primers that contained 5′ sequences of 19 nucleotides complementary to an erythromycin cassette (see Table S1 in the supplemental material) devoid of a transcriptional regulatory region (e.g., promoter or terminator). A joining PCR was carried out using these PCR fragments and the erythromycin cassette by the use of the created overlap at their extremities. It yielded a template allowing the amplification of a fragment carrying the antibiotic cassette flanked by the upstream and downstream chromosomal regions of the respective genes. The PCR fragments were used directly to transform S. mutans strain UA159. The antibiotic resistance cassette was integrated at the corresponding locus by a double-crossover event leading to the deletion of the different genes. Its transcription was driven by the initiation and termination of regulatory signals from the deleted genes in order to avoid interference with surrounding gene expression. The chromosomal cysR gene (SMU.852; GeneID 1028225) was deleted as follows. DNA fragments of 818 and 850 bp carrying, respectively, the upstream and the downstream regions of the cysR gene were generated by PCR. These fragments contain a BamHI restriction site at their 3′ and 5′ ends, respectively, introduced in cysR up and down primers (Table S1). The PCR-amplified products were digested with BamHI, ligated, and cloned into pGEM-T Easy cloning vector. This plasmid was fused to thermosensitive pGhost 9 in SalI to yield pJIM5785 (Table 1). In S. mutans UA159, the cysR gene was deleted using pJIM5785 by double-crossing over as described by Biswas et al. (8). The mutants obtained are listed in Table 1. Regions flanking the deletions were verified by sequencing to ensure the absence of PCR-induced mutations.

Construction of lux transcriptional fusions and determination of Lux activity in S. mutans.

The luxAB reporter genes were placed under the control of the cysK, tcyD, and homR promoter regions and integrated at their respective loci as described previously (57). Briefly, chromosomal DNA from S. mutans UA159 was used as a template to amplify DNA fragments containing wild-type tcyD, homR, and cysK or the mutant cysK promoter variant. PCR fragments were cloned into the pGEM-T Easy vector, sequenced, and introduced into the streptococcal integration vector pJIM4900 upstream of the luxAB genes (Table 1). The resulting plasmids were integrated by a single-crossover event at the tcyD, homR and cysK loci in the S. mutans UA159 chromosome. Strains carrying a single copy of the integrated plasmids were identified by PCR with specific primers. The resulting strains (Table 1) contained the corresponding transcriptional fusion as well as an intact copy of the gene. Luciferase assays were carried out on a Lumat LB9501 apparatus (Berthold Technologies, Bad Wildbad, Germany) (28).

RNA isolation, transcriptional start mapping, and real-time qRT-PCR.

Total RNA was isolated as described by Sperandio et al. (57). Residual chromosomal DNA was removed by treating total RNA preparations with RNase-free DNase I (Roche) according to the manufacturer's protocol. The RNA concentration was determined by absorbance at 260 nm, and the quality of RNA preparations was checked as described previously (30). Transcriptional start sites of the cysK, gshT, homR, tcyA, and tcyD genes were determined using the 5′/3′ rapid amplification of cDNA ends (RACE) kit (Roche). Real-time quantitative RT-PCR (qRT-PCR) was carried out using first-strand cDNA as the template. The primers used are listed in Table S1 in the supplemental material. Synthesis of cDNAs and quantitative PCR were carried out as described previously (57). For each gene, qRT-PCR experiments were performed in triplicate. Results were normalized by using the translational elongation factor Tu coding gene. Statistical significance between the mean ratios of genes was evaluated by Student's t test by using R (R: A language and environment for statistical computing: reference index; R Foundation for Statistical Computing, Vienna, Austria; ISBN 3-900051-07-0; http://www.R-project.org). A P value of <0.05 was considered significant.

Bioinformatics analysis.

Research of orthologous proteins for LTRRs and sulfur amino acid supply proteins were obtained from iMoMi databases (49). Phylogenic analyses were performed by using PHYLOGENY pipeline facilities (www.phylogeny.fr) (16). Further studies, including analysis of additional streptococcal genome sequences, were performed using the Scissors and iMoMi database (49), MUSCLE for multiple alignment (18), and PhyML for phylogeny (maximum likelihood). A search for potential regulatory motifs was performed in regions −250 to +50 relative to translational gene starts by using the MEME algorithm implemented in the iMoMi database (http://locus.jouy.inra.fr/imomi). WebLogo (http://weblogo.berkeley.edu) was used to generate CysR and HomR motifs from sequence alignments.

RESULTS

Identification of two LTRRs involved in sulfur amino acid supply in streptococci.

In previous studies, CmbR (FhuR) and MetR were shown to control sulfur amino acid metabolism genes in L. lactis and methionine genes in S. mutans and Streptococcus agalactiae, respectively (12, 19, 56-58). A phylogenic analysis of these proteins and their homologues from 14 Streptococcus genomes, presented in Fig. S1 in the supplemental material, shows that they can be clustered in three distinct groups whose representatives in S. mutans were named MetR, CysR, and HomR (see below). Each Streptococcus species encodes one protein in the CysR and MetR clusters, which contain the previously studied L. lactis CmbR and S. mutans MetR regulators, respectively. Members of the third group, the HomR cluster, are found only in S. mutans, Streptococcus salivarius, Streptococcus infantarius, and Streptococcus thermophilus. Proteins of the MetR, CysR, and HomR clusters share 57 to 88% identity with members of the same clusters, 27 to 36% with members of the other clusters, and only about 15% with the other streptococcal LTRRs. This relatedness suggests that these regulators may have diverged recently and prompts the question of their exact role in the Streptococcaceae.

To address this question, a functional study of mutants for genes coding for a representative protein of each cluster was carried out with S. mutans. With this bacterium, we have previously shown that MetR is specifically involved in the control of methionine synthesis and uptake (57). In contrast, the function of CysR and HomR (formerly SMU.852 and SMU.930c, respectively) is not known yet, although their homology with CmbR (57% and 36% of identity, respectively) suggests a role in sulfur metabolism. To test this hypothesis, ΔcysR and ΔhomR mutants were constructed and their growth was tested in the presence of various sulfur compounds (Table 2). Compared to the UA159 parental strain, the ΔcysR and ΔhomR mutants were unable to grow on CDM containing thiosulfate as the sole sulfur source, indicating their involvement in the regulation of the assimilation of this compound. In addition, the ΔcysR strain growth was completely abolished in the presence of sulfide or glutathione. These data indicate that both CysR and HomR play a role in the control of sulfur amino acid supply in S. mutans.

TABLE 2.

Growth of S. mutans UA159 and its derivatives on different sulfur sources^a

Supplement(s) in CDM	Morphology of colonies of S. mutans UA159	Morphology of colonies of strain expressing:
		ΔcysR	ΔhomR	ΔcysK	ΔgshT	ΔtcyA	ΔtcyE	ΔtcyA ΔtcyE
Thiosulfate	+	−	−	−	+	+	+	+
Sulfide	+	−	+	−	+	+	+	+
Glutathione	+	−	+	+	−	+	+	+
l-cystine	+	+	NT	+	+	+/−	+/−	+/−
CM	+	+	+	+	+	+	+	+
C	+	+	+	+	+	+	+	+
M	+/−	−	+/−	−	+/−	+/−	+/−	+/−
M + thiosulfate	+	−	+	−	+	+	+	+
M + sulfide	+	−	+	−	+	+	+	+
M + glutathione	+	−	+	+	−	+	+	+
M + l-cystine	+	+	NT	+	+	+	+	+

^aWashed precultures of S. mutans UA159 and its derivatives were streaked onto CDM plates supplemented with different sulfur sources, and the growth was evaluated as a function of colony morphology on agar plates. C, l-cysteine; M, l-methionine; CM, l-cysteine and l-methionine; +, normal colonies; +/−, pinpoint translucent colonies; −, no colonies; NT, not tested. Pinpoint colonies were approximately 10 times smaller than normal colonies, and growth on CDM+CM or M17 medium restored a normal colony morphology.

Role of OAS as a regulatory signal for sulfur amino acid supply genes in S. mutans.

Transcriptional activation by LTTRs requires coeffectors, which are typically intermediates of the pathways they regulate (42, 55, 57). Previously, we have shown that O-acetyl serine (OAS) was a coeffector of CmbR (58), making this metabolite a good candidate to be an effector of LTTRs involved in cysteine supply control, such as CysR and HomR. Therefore, we tested the effect of OAS on the expression of genes coding for enzymes involved in the synthesis of cysteine (CysE and CysK) or methionine (CysD, MetA, MetBC, and MetEF), for potential l-cystine uptake systems (TcyABC [formerly SMU.459-460-461], TcyDEFGH [SMU.932-933-934-935-936], and GshT [SMU.1942c]), and CysR and HomR regulators.

The transcriptional levels of these genes in the wild-type UA159 S. mutans strain growing in CDM+CM with/without OAS were compared by qRT-PCR (Table 3). The expression levels of cysK, gshT, tcyA, homR, tcyD, and metB increased 2- to 30-fold after OAS addition in the medium, whereas those of cysR, cysE, cysD, metA, and metE remained unchanged. The effect of other sulfur amino acid metabolic intermediates, such as NAS, sulfide, or thiosulfate, was tested on homR and tcyD, but no change was detected (Fig. 1). These results show that OAS is a key metabolite for regulation of sulfur amino acid supply in S. mutans and a good candidate as an effector for CysR or HomR.

TABLE 3.

Effect of OAS on the expression of several sulfur amino acid supply genes in S. mutans UA159 wild-type strain and ΔcysR and ΔhomR mutants^a

Gene	Organizationb	CysR box	HomR box	Mean ratio ± SD of gene expression levels in:c
				UA159	ΔcysR mutant	ΔhomR mutant
cysK	cysK	+	−	13.9 ± 0.3	1.4 ± 0.1*	25.7 ± 0.7*
gshT	gshT	+	−	6.3 ± 0.4	1.1 ± 0.05*	8.7 ± 0.4*
tcyA	tcyABC	+	−	2.4 ± 0.03	0.9 ± 0.02*	7.0 ± 0.3*
homR	homR	+	−	16.6 ± 1.7	2.4 ± 0.2*	NA
tcyD	tcyDEFGH	−	+	31.8 ± 3.3	8.5 ± 0.1*	1.2 ± 0.4*
metB	metBC	−	+	21.0 ± 0.1	3.3 ± 0.02*	1.1 ± 0.1*
cysR	cysR-lspA-SMU.854	−	−	1.0 ± 0.1	NA	1.0 ± 0.1
cysE	pnpA-SMU.156-cysE	−	−	1.01 ± 0.02	1.06 ± 0.03	1.0 ± 0.01
cysD	cysD-SMU.1172c	−	−	0.9 ± 0.02	0.9 ± 0.02	1.0 ± 0.1
metE	metEF	−	−	0.8 ± 0.03	1.2 ± 0.3	1.0 ± 0.1

^aGene expression in S. mutans UA159 and the ΔcysR and ΔhomR mutants was evaluated by qRT-PCR. The effect of OAS on gene expression was evaluated 45 min after the addition of 4 mM OAS to exponentially growing cells (OD₆₀₀ of 0.2) in CDM+CM.

^bGene clusters were deduced from sequence analysis and transcriptional start site mapping.

^cRatios correspond to expression levels in CDM+CM+OAS (OAS induced) relative to CDM+CM (uninduced). NA, not applicable; *, ratios in ΔcysR and ΔhomR mutants significantly different (P < 0.05 by Student's t test) from those in the UA159 wild-type strain.

FIG. 1.

Effect of OAS, thiosulfate, NAS, or sulfide on homR and tcyD expression. Luciferase activities (10³ Lux/OD₆₀₀) were measured for JIM8890 and JIM8892 strains carrying P_homR-lux and P_tcyD-lux fusions, respectively, grown in CDM+CM (□) supplemented with OAS (▪), thiosulfate (Δ), NAS (○), or sulfide (×). The x axis shows time of incubation relative to the addition of the different sulfur compounds (time zero). One curve representative of the results from at least three experiments is presented.

Respective effect of CysR, HomR, and OAS in the control of sulfur amino acid genes.

To test the possible role of OAS as an effector of CysR or HomR regulators, induction levels mediated by OAS on the previously studied set of genes were measured in the ΔcysR and ΔhomR mutant strains (Table 3). No change was detected in the transcription level of the genes whose expression is not affected by OAS addition, such as cysE, cysD, metA, and metE. In contrast, the activation observed for cysK, gshT, and tcyA transcription was abolished in the ΔcysR mutant, whereas it remained effective in the ΔhomR background, indicating the requirement of CysR in their OAS-dependent activation. Moreover, the induction of homR, tcyD, and metB transcription by OAS decreased in the ΔcysR mutant (2.5 to 8.5 versus 18 to 32), suggesting a role for CysR in their regulation. Finally, the activation of tcyD and metB transcription was abolished in the ΔhomR mutant, indicating that HomR is involved in their activation. To conclude, CysR and HomR are both involved in OAS-dependent transcription activation but target different sets of genes. CysR would directly regulate cysK, gshT, and tcyABC transcription, while HomR would control that of metBC and tcyDEFGH. The positive effect of CysR on the HomR-dependent gene could be indirect, through a control of homR transcription.

Regulatory CysR sites.

In order to identify a putative regulatory site of CysR, a search for conserved sequence elements in upstream regions of potentially CysR-regulated genes in Streptococcus genomes was carried out using the MEME algorithm (7, 49). As an initial step, this search was performed by phylogenetic footprint using, as a set of genes, the Streptococcus orthologs of cysK, tcyABC, homR, and gshT, which are controlled by CysR in S. mutans. A 13-bp motif, named the CysR box, organized as an interrupted dyad symmetry element, with TATCACNGTGATA as a consensus, is present in the upstream region of all cysK, gshT, tcyA, and homR orthologs and most tcyB orthologs (9 out of 13) (Fig. 2). In the vicinity of the CysR box, two other well-conserved boxes were identified: box1, which may be a direct repeat of the right arm of the CysR box (GTGATA), and box2 (TAT), which corresponds to the inverted sequence of the 3′ half of box1. The relative position of these boxes is perfectly conserved: box1 and box2 are placed 9 nucleotides upstream and downstream of the CysR box, respectively. Finally, the CysR box is centered in a larger inverted repeat sequence covering a 37-nucleotide DNA region from the 3′ half of box1 (ATA) to the 5′ half of box2 (TAT) (Fig. 2, dotted inverted arrows).

FIG. 2.

Alignment of CysR-regulated gene promoter regions of S. mutans and of their orthologs in other streptococci. Conserved regions (box1, CysR box, box2) are shaded, and inverted regions (arrows) are indicated below. Transcription start sites (+1) of cysK, gshT, tcyA, and homR of S. mutans were identified by 5′ RACE. The deduced −35, −10, and extended −10 regions are boxed. Logo of the region containing the CysR box was generated from the sequence alignments of promoter regions of cysK, gshT, tcyA, and tcyB orthologs with WebLogo (http://weblogo.berkeley.edu).

As a second step, additional regulatory sites were searched over the entire Streptococcus genome with the consensus sequence previously computed (ATA-N₉-TATCACNGTGATA-N₉-TAT) with PATSCAN software, allowing for two changes (see Table S2 in the supplemental material). The number of genes found by this procedure varied from a single gene (cysK in Streptococcus equi and Streptococcus mitis) to 12 genes in S. thermophilus LMG18311. Among the newly identified genes, cysD of S. thermophilus and S. salivarius is the only gene directly related to sulfur amino acid supply (Fig. 2; see also Table S2). No target in addition to those already pointed out was identified in S. mutans.

To further asses regulatory features of CysR-regulated genes, we determined the transcriptional start sites of cysK, gshT, homR, and tcyA genes from S. mutans by 5′ RACE (Fig. 2, +1). Transcription is initiated 21, 22, 24, and 25 bp upstream of the translational start point of gshT, cysK, tcyA, and homR genes, respectively. CysR boxes are thus located 48 to 56 bp upstream of the transcriptional start sites of regulated genes, fitting the general rule for LTTR operators (55). The CysR box is therefore a good candidate to be part of the CysR regulatory site and to participate in the control of cysteine synthesis in streptococci. Finally, in order to demonstrate the regulatory role of the CysR box, the effect of modification in the CysR box at positions corresponding to the conserved bases (underlined in the following sequence) for LTTRs (T-N₁₁-A) was examined on the most strongly CysR-dependent S. mutans gene, cysK, using cysK-lux transcriptional fusion (Table 4). The expression of cysK was approximately 30-fold lower when cysR was deleted or when the CysR box was substituted, demonstrating the role of CysR and of the CysR box.

TABLE 4.

Effect of CysR and CysR-box mutations on cysK expression in S. mutans

Strain	CysR boxa	Luciferase activityb
UA159 PcysK::lux	TATCACGGTGATA	9,733 ± 180
UA159 PcysK::lux	GATCACGGTGATG	423 ± 15
ΔcysR PcysK::lux mutant	TATCACGGTGATA	315 ± 11

^aNucleotide substitutions on the CysR box of the cysK gene are underlined.

^bLuciferase activity (10³ Lux/OD₆₀₀) was measured for strains JIM8887 (UA159 P_cysK::lux), JIM8888 (UA159 P_cysK::lux carrying nucleotide substitutions in CysR box), and JIM8889 (ΔcysR P_cysK::lux) cultivated in CDM+CM. The luciferase activities shown are the averages ± standard deviations of three independent cultures at an OD₆₀₀ of 0.4.

Regulatory HomR sites.

Our transcriptional analysis showed that the expression of S. mutans metBC and tcyDEFGH clusters is specifically affected in a ΔhomR mutant (Table 3). To identify a putative HomR DNA-binding site, conserved sequence elements were searched in the upstream regions of the metBC and tcyDEFGH orthologs from Streptococcus species that contain HomR. This analysis revealed the presence of a potential LTRR motif (T-N₁₁-A) in the upstream region of metB orthologs and the S. mutans tcyD gene, with TADCYAACYATMA as a consensus (Fig. 3). The motif, named HomR box, is located 43 bp upstream of the transcriptional start site of tcyD, a correct location for an LTRR recognition site. Two conserved elements, similar to box1 and box2 previously defined in the CysR-regulated genes, were found 9 and 8 nucleotides upstream and downstream of the HomR box, respectively. Finally, no additional HomR motifs upstream of genes related to sulfur amino acid metabolism were identified in the genome of S. infantarius, S. mutans, S. salivarius, and S. thermophilus with the computed motif ATA-N₉-TATADCYAACYATMA-N₈-WAT, in allowing for one change.

FIG. 3.

Alignment of HomR-regulated gene promoter regions of S. mutans and of their orthologs in other streptococci. Conserved regions (box1′, HomR box, box2′) are shaded. Transcription start site (+1) of tcyD of S. mutans was identified by 5′ RACE, and the deduced extended −10 regions are boxed. Logo of the region containing box1′ to box2′ (including the HomR box) was generated from the sequence alignments with WebLogo (http://weblogo.berkeley.edu).

Functional analysis of substrate-binding proteins regulated by CysR and HomR.

The results presented above indicated that CysR and HomR activate the expression of three transcriptional units (tcyABC, tcyDEFGH, and gshT) encoding components of ABC transporters sharing similarities with l-cystine uptake systems. TcyA, TcyE, and GshT exhibit 29 to 32% identity to YtmJ and YckK l-cystine binding proteins of B. subtilis (13). Transcription of tcyABC and gshT is activated by CysR, while that of tcyDEFGH by HomR. In order to clarify the function of these genes and better define the metabolic role of CysR and HomR, we have performed a functional analysis of mutants for the tcyA, tcyE, and gshT genes encoding substrate-binding proteins of ABC transporters in S. mutans (Table 2).

The growth of the ΔtcyA and ΔtcyE mutant strains was significantly reduced in the presence of l-cystine as a sole sulfur source, suggesting that these genes encode transport systems involved in the uptake of this molecule. The ΔgshT mutant strain was unable to grow in CDM containing glutathione (GSH) as a sole sulfur source, indicating that GshT is required for GSH assimilation. gshT is located upstream of the atmBCDE cluster involved in methionine uptake and thiosulfate assimilation (57). Since the ΔatmB and ΔatmBCDE mutants and strain UA159 display similar growth levels in the presence of GSH (data not shown), GshT is likely a substrate-binding protein necessary for GSH uptake.

DISCUSSION

In this work, CysR and HomR, two new LTRRs involved in sulfur amino acid metabolism in S. mutans were characterized. These regulators respond to OAS as a coeffector but control different sets of genes. CysR activates the transcription of cysK, tcyABC, and gshT, which are required for de novo cysteine synthesis and transport of l-cystine and glutathione, a cysteine-containing tripeptide, respectively. HomR positively regulates the transcription of metBC and tcyDEFGH, involved in homocysteine synthesis and l-cystine uptake, respectively. Furthermore, the absence of growth of a homR mutant in a medium containing thiosulfate as the sole sulfur source indicates that HomR controls additional genes remaining to be determined. Interestingly, homR expression is activated by CysR, leading to a regulatory cascade. The fact that the two regulators are responding to the same coeffector suggests that this cascade is a fine-tuning system to control HomR targets.

On the basis of this and previous work we propose a global regulatory pathway for sulfur amino acid supply in S. mutans (Fig. 4). Its upper part, corresponding to cysteine synthesis, is activated by CysR and OAS, and its intermediate part, allowing homocysteine supply, is under the control of HomR and OAS and indirectly of CysR. Finally, the lower part, leading to methionine synthesis, is activated by MetR, with homocysteine as a coeffector (57), and by an S-adenosylmethionine (SAM)-responsive riboswitch mechanism (22). While the three regulators participate in the control of cysteine or of its precursor's uptake, the transport of methionine and of its precursor homocysteine is regulated only by MetR. This regulatory network, composed essentially of 3 paralogous LTTRs, differs significantly from those described for other Gram-positive bacteria, such as B. subtilis, Corynebacterium glutamicum, and Clostridium acetobutylicum, in which a large diversity of molecular mechanisms controls sulfur metabolism (2, 4, 5, 14, 27, 31, 51-54, 59).

FIG. 4.

Overall regulatory scheme for control of sulfur amino acid supply in S. mutans. Colored arrows represent the steps controlled by CysR (black), HomR (gray), MetR (white), and riboswitch (hatched). Discontinuous arrows indicate steps for which no regulation has been yet evidenced. Genes of S. mutans proposed to encode proteins involved in the corresponding reaction are indicated. OAS, O-acetylserine; OAH, O-acetylhomoserine; CYS, cysteine; HC, homocysteine; MET, methionine; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; SRH, S-ribosylhomocysteine; TS, thiosulfate; S, sulfide; GSH, glutathione; RBS, riboswitch; NR, not regulated.

Orthologs of the S. mutans CysR and MetR regulators are encoded by all Streptococcus and Lactococcus genomes, while HomR is found only in S. mutans, S. infantarius, S. salivarius, and S. thermophilus (Fig. S1 in the supplemental material). As deduced from specific motifs found for these regulators, common lines for sulfur amino acid gene regulation are conserved among the Streptococcaceae, even if differences do exist. MetR displays the most conserved regulatory pattern and recognizes a motif present in the upstream region of Streptococcus and Lactococcus metE, encoding the last step for methionine synthesis. In few species, MetR may display a limited number of additional targets (52, 57). In contrast, L. lactis CmbR and S. mutans CysR, which both belong to the CysR cluster, display significantly different regulatory patterns (58; this study). CmbR positively regulates 18 genes (7 transcriptional units), including most cysteine and methionine metabolic and transporter genes, while S. mutans CysR controls 6 genes (4 transcriptional units) involved only in cysteine supply. The phylogenic analysis presented here and several aspects of their regulatory pattern suggest that these regulators are orthologs. For example, CmbR and CysR respond to OAS as a coeffector and control several orthologous genes involved in cysteine metabolism, such as cysK, the most strongly regulated genes in both species (19, 58; this work). However, the LTTR sites recognized by these regulators exhibit a different consensus motif (TATCACNGTGATA for CysR and TAAAWWTTTYTWA for FhuR/CmbR). The CysR motif is partly similar to a motif found by a comparative genomic approach (TGATA-N₉-TATCA-N_2-4-TGATA) and that was proposed to be the CmbR motif, although absent in L. lactis (38). The absence of potential CysR and CmbR boxes in lactococci and streptococci, respectively, confirms the divergence between the recognition sites of the 2 regulators (E. Guédon, unpublished data). Since functional and phylogenic analysis clearly links CmbR to the CysR cluster, we propose that CmbR separated from the CysR cluster and now recognizes a new and less conserved motif, allowing an enlarged regulatory spectrum. However, it is also possible that the CysR ancestor in streptococci evolved to recognize a strict consensus and a smaller number of targets. Both hypotheses are in agreement with the fact that regulatory functions are often evolving faster than other functions in microorganisms (50).

Besides this major divergence between lactococci and streptococci, further data suggest that CysR regulation is rapidly evolving within streptococci. CysR orthologs and CysR boxes are present in all Streptococcus genomes, suggesting a conserved control by CysR of cysK, tcyABC, and gshT orthologs when present. Nevertheless, in S. agalactiae, upstream of the cysK, gshT, and tcyA genes, a base is missing from the CysR boxes, suggesting a loss of functionality. A lack of activation of cysK transcription would be in agreement with the requirement of cysteine for S. agalactiae growth (46). Conversely, the CysR box is overrepresented in the S. thermophilus genome in which, in addition to the above-mentioned sulfur metabolic genes, CysR boxes are found upstream of genes coding for two substrate-binding proteins of oligopeptide ABC transporters (24), two amino acid transporters (including a gshT paralogous gene with 96% identity), a protein associated with bacteriocin production (21), and a potential redox-sensitive transcriptional regulator. Since cysteine or cystine is not detectable as a free amino acid in milk, cysteine-containing peptides may represent the main source of cysteine for S. thermophilus in this environment (33). These facts suggest modification of the CysR target set in relation to S. thermophilus adaptation to its ecological niche.

In addition to MetR and CysR, several streptococci, such as S. mutans, S. infantarius, S. salivarius, and S. thermophilus, contain HomR, a third LTTR involved in sulfur amino acid metabolism. Phylogenic analysis of HomR indicates that it is related to CysR, although it forms a cluster separate from CysR orthologs. The involvement of OAS as an effector in both groups indicates a functional conservation. Beside these conserved features, CysR and HomR differ by their recognition sites (TATCACNGTGATA and TADCYAACYATMA, respectively) and target genes. In this work, we have shown that S. mutans HomR is required for the expression of two homocysteine synthesis enzymes (MetB and MetC) and an l-cystine transport system (TcyDEFG) and is likely required for the expression of thiosulfate assimilation genes. These data indicate that HomR is a CysR paralog that drifted to control specific genes, likely to increase environmental fitness of these bacteria.

Conclusion.

In the Streptococcaceae, sulfur amino acid supply is controlled by 2 or 3 related LTTRs. Their main regulatory features appear to be conserved among this family of bacteria, but several differences could be evidenced. First, HomR, which is present in a limited number of species, may be the result of duplication from cysR followed by a drift and specialization to few targets. Second, CmbR, the lactococcal CysR ortholog, may have diverged from a CysR ancestor to recognize a new motif for DNA binding and regulate an enlarged set of genes. Lastly, regulation functions of CysR in different streptococci display subtle differences that could reflect adaptation of regulatory function to cope with their different lifestyles. In this genus, several species display versatile metabolic ability; these include S. mutans (1, 20) and S. salivarius (P. Renault, unpublished data) living in an oral environment or those of more limited ability, such as S. agalactiae, in which CysR may have lost its regulatory role. Previous studies indicated that streptococci evolved by gene acquisition, recombination, and positive selection in their core genomes. This study underlines the rapid drift of regulatory function in the Streptococcaceae as an additional paradigm of evolution, allowing their adaptation to different niches within hosts and shaping their behaviors toward being commensal or pathogenic.