A novel double-tryptophan peptide pheromone is conserved in mutans and pyogenic Streptococci and Controls Competence in Streptococcus mutans via an Rgg regulator

Summary

All streptococcal genomes encode the alternative sigma factor SigX and 21 SigX-dependent proteins required for genetic transformation, yet no pyogenic streptococci are known to develop competence. Resolving this paradox may depend on understanding the regulation of sigX. We report the identification of a regulatory circuit linked to the sigX genes of both mutans and pyogenic streptococci that uses a novel small, double-tryptophan-containing competence-inducing peptide (CIP) pheromone. In both groups, the CIP gene, which we designate comS, and sigX have identical, noncanonical promoters consisting of 9-bp inverted repeats separated from a −10 hexamer by 19 bp. comS is adjacent to a gene encoding a putative transcription factor of the Rgg family and is regulated by its product, which we designate ComR. Deletion of comR or comS in S. mutans abolished transformability, as did deletion of the oligopeptide permease subunit oppD, suggesting that CIP is imported. Providing S. mutans with synthetic fragments of CIP revealed that seven C-terminal residues, including the WW motif, cause robust induction of both sigX and the competent state. We propose that this circuit is the proximal regulator of sigX in S. mutans, and we infer that it controls competence in a parallel way in all pyogenic streptococci.

Introduction

Horizontal gene transfer is thought to be an important facilitator of the evolution of bacterial genomes. The single mechanism of such transfer that does not depend on fortuitous extra-chromosomal genetic elements such as phages, plasmids, or transposons is natural genetic transformation. Genetic transformation has been directly observed in scores of bacterial species, and likely is important for the survival and evolution of many more (Johnsborg et al., 2007; Lorenz and Wackernagel, 1994). In those species where it has been reported, the incidence of competence in the laboratory ranges from constitutive and ubiquitous to rare, sometimes being limited to a single strain and a peculiar growth state (Gromkova et al., 1998; Sexton and Vogel, 2004; Sikorski et al., 2002). Natural transformation depends on expression of a large, dispersed set of genes encoding proteins responsible for DNA uptake and recombination, as well as other associated functions (Chen and Dubnau, 2004; Dagkessamanskaia et al, 2004; Maughan and Redfield, 2009; Peterson et al., 2004; Redfield et al., 2005; Tortosa and Dubnau, 1999). In some species, these genes are apparently expressed constitutively, and the bacteria are always capable of acquisition of DNA. In many species, however, including those in the genera Hemophilus, Vibrio, Bacillus, and Streptococcus, expression of competence genes is limited to specific growth circumstances, where coordinated expression of the entire set of genes creates a temporary physiological state of DNA receptivity, termed competence.

While many effector genes of genetic transformation are widely conserved, regulatory circuits that control their expression are more diverse (Claverys and Martin, 2003). In several species, the coordination of the competence regulon(s) has been traced to a master regulator, which controls transcription of the effector genes. In Hemophilus influenzae, expression of the competence regulon depends on both the cAMP regulatory protein CRP and a specific co-regulator named Sxy, which mediate a response to nutritional deprivation (Redfield et al., 2005). The same regulon is well conserved among all Pasteurellaceae (Redfield et al., 2006). In Bacillus subtilis, the master regulator is a protein, ComK, which stimulates transcription of target operons by sigmaA-polymerase (E•SigA), is positively auto-regulated, and responds both to nutritional signals and to at least two cell-to-cell communication systems mediated by peptide pheromones (Hamoen et al., 2003; Solomon et al., 1996).

In the streptococci, the master regulator is an alternative sigma factor, SigX (also known as ComX), of the Sigma-70 family (Lee and Morrison, 1999). Although first identified in S. pneumoniae, genome sequencing has revealed that homologs of SigX are ubiquitous in the streptococci and common in certain close relatives within the phylum Firmicutes including enterococci, lactococci, and lactobacilli (Claverys and Martin, 2003). While natural transformation has been reported in only a minority of these species, it is an open question whether this alternative sigma factor (the sole known alternative sigma factor in many of these species) universally controls competence or whether it may have diverged to control transcription of various other gene sets and traits. In fact, ~20 late competence effector genes identified directly in S. pneumoniae (Oggioni and Morrison, 2008; Peterson et al., 2004; Dagkessamanskaia et al., 2004) are widely co-distributed with SigX. Interestingly, these conserved effector genes are linked to conserved non-canonical, SigX-driven promoters identifiable by shared sequence motifs, indicating they form a SigX-dependent regulon (Martin et al., 2006). Three such SigX-dependent promoter sequence patterns have been identified among the Firmicutes, as illustrated in Table S1. In each case, there is at least one experimental demonstration that artificial over-expression of SigX causes induction of expression of competence effector genes, suggesting that the SigX homologues are widely conserved as active sigma factors (Morikawa et al., 2003; Woodbury et al.; 2006; Wydau et al., 2006).

The core competence regulatory components (SigX and SigX-dependent non-canonical promoters) and effector genes of the DNA uptake and recombination machinery are found in all streptococci sequenced to date, but the regulators of sigX differ between groups of species within this genus (Martin et al., 2006; Johnsborg et al., 2007). Three different regulatory schemas controlling sigX, have been described; each involves a peptide signal mediating cell-cell communication and illustrates the phenomena of microbial cooperation that are often termed quorum sensing. In S. pneumoniae and other species of the mitis and anginosis groups of streptococci, the membrane-bound receptor (ComD) of a two-component signal transduction system (TCSTS) participates in cell-to-cell signaling through a secreted peptide of the double-glycine class (ComC) and activates transcription of sigX through a response regulator (ComE) and a direct-repeat site at −40 in its promoter (Ween et al., 1999). In S. mutans, the single well-studied member of the mutans group, a completely different regulatory circuit controls sigX expression (Martin et al., 2006). In this case, a paralogous TCSTS (known as ComDE or BlpHR) acts through a different double-glycine peptide signal (CSP or BIP) as a transcriptional regulator to control production of the non-lantibiotic mutacin, NlmC (Mutacin V) (as well as other mutacin loci (van der Ploeg, 2005; Kreth et al., 2007)). ComDE/BlpHR are orthologous to widespread streptococcal bacteriocin regulators, which are paralogous in the mitis and anginosus groups to the ComDE systems controlling sigX (Martin et al., 2006). Importantly, it also strongly stimulates expression of sigX and competence effector genes by an unknown mechanism (Kreth et al., 2007; Perry et al., 2009). As has been discussed elsewhere (Martin et al., 2006; Kreth et al., 2007; Perry et al., 2009; Ahn et al., 2006), the proximal regulator of sigX in S. mutans appears to have eluded detection, as sigX expression is temporally separated from the CSP-induced bacteriocin induction that depends on ComE, competence does not depend absolutely on the comDE TCSTS, and the sigX gene is not known to share any cis-acting site with ComE-dependent genes. Finally, a third peptide-signal-dependent regulator of sigX, ComR, was recently described in the salivarius group species, S. salivarius and S. thermophilus (Fontaine et al., 2010; Gardan et al., 2009). ComR is a ‘stand-alone regulator’ of the Rgg family (McIver, 2009; Qi et al., 1999; Sanders et al., 1998; Sulavik et al., 1992), that is distinct from the classical response regulators of TCSTS quorum-sensing circuits. Proteins of the Rgg family are homologous to the transcription factor discovered in S. gordonii and called Rgg, for regulator gene of glucosyltransferase (Sulavik et al., 1992). The family includes GadR of Lactococcus lactis, MutR of S. mutans, Rgg (RopB) of S. pyogenes, and other proteins, mostly with unknown function, that are widespread in Gram-positive genomes. In the case of ComR of S. thermophilus, response to the inter-cellular peptide signal requires the oligopeptide permease transporter, Opp, rather than a surface receptor. The salivarius group competence pheromone is derived from a propeptide that is in a class termed SHP (small hydrophobic peptide (Ibrahim, et al., 2007b)), distinct from the double-glycine propeptide signals associated with streptococcal bacteriocin regulation and with competence regulation in the mutans, anginosus, and mitis groups. sigX and the SHP gene comS share common promoter sequences, proposed as targets of the pheromone-associated form of ComR so that the ComR/ComS pair form an autocatalytic loop linked directly to expression of sigX (Fontaine et al., 2010).

Here we report results of genomic analysis and direct genetic experiments that identify a new small hydrophobic peptide pheromone class and a conserved upstream regulator of sigX that are shared by the mutans and pyogenic groups of streptococci. The results suggest that most streptococcal species may be naturally competent and use regulation of sigX to control the expression of competence genes.

Results

Identification of a conserved sigX promoter in the pyogenic group of streptococci

Genomic sequences are now available for over 70 streptococcal strains in 17 species and four principal phylogenetic groups. Inspection of these reveals, in all species except S. thermophilus and S. mutans (which have one copy of sigX each), duplicate copies of genes encoding sigX orthologs, located in group-specific gene contexts. Within the naturally transformable mitis group, the duplicate sigX copies are positioned at two conserved sites between an upstream unique gene (fstH or nusG, respectively) and duplicate tRNA-Glu downstream neighbors (Fig. S1). In each case, identical upstream sequences include a promoter with a direct-repeat −40 element to which the TCSTS response regulator, ComE, binds (Ween et al., 1999). In contrast, the sigX genes in S. pyogenes and the other species of the pyogenic group (except S. agalactiae) are found at a different pair of conserved sites (Fig. S1), downstream from duplicate tRNA-Arg genes, and upstream of different unique genes. Again, within each species the two sigX copies and upstream sequences are identical to each other.

In Fig. 1A sequences upstream of the sigX gene in seven pyogenic genomes are aligned at the centers of symmetry of the tRNA-Arg terminator stem-loops. This alignment reveals an apparent overlap between the tRNA transcriptional terminator and the sigX transcriptional promoter. The terminator consists of a conserved 9-bp stem, a 4-nt loop, and a T-rich region at the 3’ side of the stem-loop (Fig. 1A). This choice of alignment immediately revealed an additional conserved motif which is identical to the canonical SigA −10 promoter hexamer (TATA[A/C]T) and is positioned 25–48 nt upstream of the sigX translation initiation signal, but there is no apparent corresponding canonical −35 element. In fact, this hexamer is the only conserved sequence between the tRNA terminator and the ribosome-binding site of sigX. Because the distance between the tRNA-Arg terminator stem and the conserved putative −10 element of sigX was uniformly short (19 nt), we inspected the tRNA terminators themselves for conserved sequences that might participate in initiation of sigX transcription. Remarkably, most positions in the ‘stem’ of the terminator signal are conserved not only in secondary structure but also in sequence, providing a conserved 9-bp inverted repeat element centered at −45. The same conserved elements are also found at the sigX gene in S. agalactiae, which is not adjacent to a tRNA gene. Taken together, these features are suggestive of a conserved Class II activator-dependent promoter (Browning and Busby, 2004) associated with sigX genes of the pyogenic group. If this interpretation is correct, it provides an example of parsimonious use of genomic resources: most of the bases encoding the RNA transcriptional terminator signal for tRNA-Arg also serve as part of the DNA transcriptional initiation signal for sigX.

Figure 1. Genomic analysis of sigX promoters in pyogenic and mutans groups of streptococci.

A. Comparison of ~100 bp upstream of sigX genes in seven genomes of pyogenic species, aligned (except for Saga) at dyad centers of the terminators of upstream tRNA-Arg genes. Highlighted bases: , conserved −10 hexamer 19 bp from the terminator stem; , , other bases conserved across species. Above, terminator stem-loop pattern designated with > and < symbols; below, consensus pattern P1; distance to sigX shown at right. B. Alignment of sequences at sites upstream of comS identified by scanning pyogenic genomes for P1. Highlights as in panel A. Stop codon triplet of the flanking comR gene is underlined. C. Alignment of sequences upstream of sigX and downstream of SMU.61 in S. mutans. Highlights as in panel A. Genome sources: Saga1, S. agalactiae A909 (NC_007432); Saga2 S. agalactiae NEM316 (NC_004368); Sdys, S. dysgalactiae subsp. equisimilis GGS_124 (NC_012891); Sequ, S. equi subsp. zooepidemicus (NC_012470); Sgal, S. gallolyticus UCN034 (NC_013798); Sinf, S. infantarius subsp. infantarius ATCC BAA-102 (NZ_ABJK00000000); Smut, S. mutans UA159 (NC_004350); Spyo1, S. pyogenes M1 (NC_002737); Spyo2, S. pyogenes Manfredo (NC_009332); Sube, S. uberis 0140 (NC_012004).

Identification of a single additional copy of the consensus sigX promoter reveals a conserved Rgg-SHP locus in pyogenic genomes

To investigate the phylogenetic distribution and specific genetic contexts of the conserved sigX promoter structure more broadly, we used the pattern of its conserved features (designated P1) to scan complete streptococcal genomes for similar sequences. The P1 pattern was found at several sites in multiple streptococcal genomes, primarily in species of the pyogenic group (Table 1), but never in any of the four salivarius group genomes available. The scan with pattern P1 found, as expected, matches at the promoters of sigX genes (Fig. 1A) in all pyogenic group genomes. In each pyogenic genome, one additional site matching pattern P1 was found (Table 1), falling between an open reading frame (ORF) encoding a putative transcription factor of the Rgg/GadR/MutR family that is homologous to the salivarius comR gene, which we designate comR, and a DNA repair gene, ruvB, in a region whose organization is well conserved among the pyogenic streptococci (Fig. S2). In all cases, the comR stop triplet is located near or within the P1 inverted repeat (Fig. 1B). This organization resembles the ComRS locus of the salivarius group, where a sequence matching the promoter of the salivarius sigX is also positioned at the 3’ end of the Rgg gene, comR (STER_0316). In addition, the sequence of the P1 pattern is itself reminiscent of the sigX promoter of salivarius species (designated P2 in Table 1), where a conserved inverted repeat (termed the Ecombox) occurs 21 bp upstream of a canonical −10 hexamer (Fontaine et al., 2010). Despite this similarity, P1 and P2 differ in the first and last four bases of their inverted repeats.

Table 1.

Genome scans with sigX promoter consensus patterns P1 or P2

		No. sites matching sigX promoter consensus pattern From a
Species	Strain	S. pyogenes (P1)	S. thermophilus (P2)	Nearest gene b
Mitis group
S. gordonii	CH1	2	0	SGO_2080 (plcR), SGO_0149
S. mitis	B6	0	0	-
S. pneumoniae	R6	0	0	-
S. sanguinis	SK36	0	0	-
Pyogenic group
S. agalactiae	NEM316	3	0	sigX, sigX, rgg
S. dysgalactiae	GGS_124	3	0	sigX, sigX, rgg
S. equi subsp zooepidemicus	H70	4	0	sigX, sigX, rgg, rggc
S. gallolyticus	UCN34	3	0	sigX, sigX, rgg
S. pyogenes	M1	3	0	sigX, sigX, rgg
S. uberis	0140J	3	0	sigX, sigX, rgg
Salivarius group
S. thermophilus	LMD9	0	7	sigX, comR, 5 x bcnd
Mutans group
S. mutans	UA159	2	2	P1: sigX, rgg P2 SMU.28, SMU.370

^aSearch patterns: P1 AACA [TG]GACA N(4) TGTC [AC] TGTT N(19) TATAAT, Mismatches = 3

P2 T [AG]GT GACAT N(2) ATGTCACTA N(21) TATAAT, Mismatches = 3

^brgg, promoter at 3’ end of comR homolog and upstream of a WW SHP.

^c4^th site, 3’ of Rgg homolog and upstream of a GG-propeptide

^dbcn, genes implicated in bacteriocin biosynthesis.

In S. thermophilus, sequences matching the P2 pattern of the sigX promoter are found at other genomic sites immediately upstream of co-regulated ORFs, one encoding ComS, the SHP precursor to the intercellular signal, Shp0316, and three others thought to encode bacteriocin production proteins (Fontaine et al., 2010). The SHP is processed by an unknown secretion pathway for release of a product that acts as an intercellular signal and appears to correspond to seven to ten C-terminal residues of Shp0316. Response to the pheromone requires an oligopeptide transporter, Opp, and the ComR protein. Together, ComR and ComS work to induce transcription of target genes from the Ecombox promoters. Collectively designated the Early Competence Genes, these are proposed to contribute to a coordinated system for genetic exchange in which bacteriocins allow competent cells to acquire DNA from susceptible neighbors.

As the positioning of the third sigX-like P1 promoter pattern at the 3'-end of the pyogenic comR homologues is reminiscent of the organization of the ComR locus of the salivarius group, and as the comR – ruvB interval within the pyogenic genomes had not been characterized, we inspected the intervening sequences (~100–200 bp) for additional shared features, especially small ORFs. Remarkably, each of these regions contains a small open reading frame not annotated in GenBank (Table 2). Between the −10 element of the P1 sigX-like promoter and the ATG of each of these ORFs is an interval of 23 to 41 bp that includes an apparent Shine-Dalgarno ribosome binding site. The predicted translation products of these small ORFs share several structural properties. All are SHPs, have a net positive charge (except in the case of S. gallolyticus, where the peptide is neutral), and contain a double-tryptophan (WW) motif near the C-terminus (Table 2). The conserved structural features of these predicted peptides, taken together with their conserved genomic context, strongly suggest a common function.

Table 2.

Putative Competence Pheromone Propeptides in Streptococci

Species	Strain	Upstream Gene	Peptide Sequence
Pyogenic group, putative ComS
S. agalactiae	A909	SAK_0081	MFKVFFTVMTGVFWWG
S. agalactiae	NEM316	gbs0048	MTLVIKLVGTLLTMGWWGL
S. dysgalactiae       subsp. equisimilis	GGS_124	SDEG_0056	MFKRYHYYFILTAMLAFKAAQMISQVDWWRL
S. equi subsp zooepidemicus	H70	SZO_00390	MFKKYQYYLFLAALFLLHSAQLLSDIDWWRVG
S. gallolyticus	UCN34	GALLO_0038	MLNIFSIVITGWWGL
S. infantarius	BAA	STRINF_00393	MLKGFTVLLTAWWGL
S. pyogenes	M1	Spy0037	MLKKYKYYFIFAALLSFKVVQELSAVDWWRL
S. pyogenes	Manfredo	SpyM50034	MLKKVKPFLLLAAVVAFKVARVMHEFDWWNLG
S. uberis	0140J	SUB0057	MFKKIHFYVTTFSFLAVALITFLSEKDWWHIG
Mutans group, ComS
S. mutans	UA159	SMU.61	MFSILTSILMGLDWWSLa
Salivarius group, ComS
S. thermophilus	LMD9	STER_0316	MKTLKIFVLFSLLIAILPYFAGCLb
S. salivarius	SK126		MKKLKLFTLFSLLITILPYFTGCLb
Mitis group, ComC
S. gordonii	Challis	SGO_2148	MKKKNKQNLLPKELQQFEILTERKLEQVTGG / DVRSNKIRLWWENIFFNKKc
S. mitis	B6	smi_2086	MKNTVKLEQFVALKEKELQKIKGG / EMRKPDGALFNLFRRRc
S. pneumoniae	R6	Sprt5	MKNTVKLEQFVALKEKDLQKIKGG / EMRLSKFFRDFILQRKKd
Anginosus group, ComC
S. anginosus	10713	tRNA	MKKLFAKKEVVKKEVVEFKELNDEQLDKIIGG / DSRIRMGFDFSKLFGKc
S. milleri	10708	tRNA	MKKIFSKKEITKVEVESFKELTDEQLNKIVGG / DRRDPRGIIGIGKKLFGc

^aLi et al.,2001.;

^bFontaine et al.,2010;

^cHavarstein et al., 1997;

^dHavarstein et al., 1995.

As ComR is proposed to control expression of comS and sigX directly via the P2 promoter in S. thermophilus (Fontaine et al., 2010), the parallel association of the P1 promoters with both an SHP gene and sigX in the pyogenics suggests the possibility of their co-regulation in these species as well. We propose that species of the pyogenic group possess a ComR/SHP/sigX regulatory circuit homologous to the S. thermophilus circuit, and that it acts similarly to mediate cell-to-cell signaling in control of expression of sigX and, by extension, of competence. While the outline of this regulatory pathway appears the same as that of the salivarius system, many of its details are not yet established, and may differ. For example, the novel composition of this group of SHPs suggests that processing steps in the pyogenic streptococci could be distinct from those in the salivarius group. Unfortunately, conditions for development of competence or expression of comR or sigX in pyogenic species are unknown, precluding an immediate direct test of the overall model in any species of the group.

Identification of a ComR/SHP/sigX locus in S. mutans orthologous to the pyogenic ComR/SHP/sigX circuits

Although the P1 promoter pattern was primarily limited to species of the pyogenic group (Table 1), an exception was the mutans group, in which both sequenced genomes have the P1 promoter pattern at two genomic sites. Because the P1 pattern is so strongly conserved and so restricted in its distribution, it was of interest to learn whether it might be linked to sigX expression in S. mutans, which, in contrast to the pyogenic species, is readily transformable. Interestingly, one mutans P1 site precedes sigX (SMU.1997) and the other follows an Rgg family gene (SMU.61) (Fig. 1C). Remarkably, inspection of the SMU.61 locus further revealed an organization similar in detail to that of comR/SHP loci described above for the pyogenic species, with a P1 match at the 3’ end of SMU.61 and with a downstream ORF encoding a 17-residue SHP with a WW motif near its C-terminus (Table 2). We designate this ORF associated with SMU.61 as shp61. This pattern strongly suggests that SMU.61 may encode the proximal regulator of sigX in S. mutans as does comR in S. thermophilus and as we propose above for pyogenic species’ comR genes. It also suggests that SMU.61 may mediate cell-to-cell signaling via the translation product of shp61. Since conditions for S. mutans competence are known, we tested several predictions of this hypothesis directly.

The ComR/SHP/sigX circuit of S. mutans is required for competence development

To assess the involvement of SMU.61 or Shp61 in competence development in S. mutans, we constructed deletion mutants of the wild type strain UA159 and examined their competence under standard conditions for endogenous competence development in Todd-Hewitt Broth (THB) with horse serum (Perry et al., 1983; Lindler and Macrina, 1986). Both ΔSMU.61 and Δshp61 mutations reduced the endogenous rate of transformation below the limits of detection, at least 400-fold below wild type (Fig. 2). Indeed, in multiple transformation trials under these conditions, we have never recovered a transformant from the Δshp61 mutant MW05 or from two different ΔSMU.61 mutants, MW01 and MW02 (data not shown). We also constructed a strain with a plasmid-borne copy of SMU.61. The SMU.61 plasmid pWAR300 restored transformation in the ΔSMU.61 background nearly to wild type levels, showing that the loss of competence in the deletion mutants was largely or entirely due to loss of SMU.61, rather than to any polar effects on flanking loci. Introduction of plasmid pWAR300 into the wild type strain by itself enhanced competence development, as evidenced by a yield of transformants increased 10-fold over that observed in the wild type (Fig. 2). This dramatic phenotype is reminiscent of the strong effect achieved by supplementation with CSP (Li et al., 2001), or by over-expression of the high density response regulator HdrR (Okinaga et al., 2010), and suggests that the level of SMU.61 products is ordinarily limiting for competence development.

Figure 2. Requirement of SMU.61 locus for competence development in S. mutans.

Genetic maps indicate locations of deletion/replacement mutations and presence of complementing plasmid carrying comR. Competence of indicated strains was determined in 1-ml cultures under standard conditions (Lindler and Macrina, 1986) for the wild type, UA159, and for derivatives with indicated gene disruptions. Competence is presented as Cm^R transformants per ml culture per OD₅₅₀ after exposure to 50 ng of Cm^R genomic DNA from strain MW01.

We conclude that SMU.61 and Shp61 are critical for endogenous development of competence in S. mutans during culture in rich media and propose the names comR and comS, respectively. Furthermore, it appears that expression of the former gene may itself be a limiting parameter in operation of this system. Overall, the parallel between the organization of the comR/comS loci of mutans and pyogenic species and the organization of comR/comS in the salivarius group suggests that the three systems function in a similar manner, through small peptide signals, but with divergent peptide sequences and distinct DNA target sequences.

A second ComR homologue in S. mutans is apparently linked to bacteriocin transporters

S. mutans has two paralogous comR loci: SMU.61 is an ortholog of the pyogenic group's comR, while SMU.381c has greater homology to the salivarius comR (Table S2). Concomitantly, two copies of the salivarius P2 promoter pattern exist in S. mutans. One lies 107 bp upstream of SMU.28, a homolog of bacteriocin ABC exporters. The second is upstream of SMU.370, also an ABC exporter, adjacent to a cluster of bacteriocin-related genes and 7 kb away from SMU.381c (Fig. 3). These bacteriocin-related loci of S. mutans are distinct from those controlled by the comCDE TCSTS (van der Ploeg, 2005). Together with the preceding results, this pattern suggests that in S. mutans two Rgg homologs have diverged to control two phenotypes, one for competence, and another for bacteriocins. In contrast, both phenotypes are controlled by a single Rgg in salivarius species.

Figure 3. Paralogous Rgg regulators in S. mutans vs S. pyogenes and S. thermophilus.

Chevron, ORF encoding short hydrophobic peptide. Pentagons, ORFs flanking genomic sites matching promoter patterns P1 and P2. Promoter pattern elements: inverted repeats and −10 hexamer. Raised arrow, putative transcription start. Bcn1, etc., ORFs associated with known or putative bacteriocins, as described in the text. Orthologous pairs indicated by matching colors. Sequences of P1 and P2 from Fig. 1 are compared in lower panel. Inner box: conserved core; outer box: divergent inverted repeat sequences.

Extracellular ComS induces competence development in S. mutans

To investigate the mechanism by which the mutans comR/S genes affect competence development, and to probe the proposed analogy with the salivarius circuit, we asked if the full-length ComS product could elicit competence development when supplied extra-cellularly. We also examined the response to truncated C-terminal derivatives of ComS since some peptide pheromones of other Gram-positive quorum sensing pathways (e.g. PapR of Bacillus thuringensis, the enterococcal sex phermone inhibitors, the Phr peptides of B. subtilis, and the SHP competence pheromone of S. thermophilus) undergo one or more processing steps that release the C-terminal five to eight amino acids of the peptide as the active form (Slamti and Lereclus, 2002; Bouillaut et al., 2008; Nakayama et al., 1994; Solomon et al., 1996), and since the conserved WW motif indicated the possible importance of the terminal portion of ComS. The peptides were tested in a chemically defined medium (CDM), containing individual amino acids but no polypeptides (van de Rijn and Kessler, 1968). In this medium, S. mutans grows to high densities and at the same rate as in THB, but does not develop competence in the early exponential growth phase, providing a suitable background for evaluating peptide signals.

To assess the ability of ComS peptides to induce competence in S. mutans, we added synthetic peptides to low-density logarithmic-phase cultures and measured the frequency of transformation 60 minutes thereafter. As shown in Fig. 4, competence was induced by ComS_11–17 at concentrations as low as 300 nM, and approached a maximum rate at 10 µM. Neither full-length ComS nor the ComS_13–17 derivative gave rise to transformants during the one-hour assay period, suggesting post-translational processing of ComS and suggesting that ComS_11–17 is, or closely resembles, the active form of this pheromone. We conclude that ComS encodes an intercellular signal similar or identical to ComS_11–17 that is important for competence development in S. mutans.

Figure 4. Induction of Competence by exogenous peptide.

Transformation in chemically defined medium with addition of exogenous pheromone candidates. Frozen CDM stock cultures of S. mutans wild type UA159 (A) or isogenic mutants (B) were diluted to OD=0.01 and grown until OD=0.1, at which time synthetic peptides and donor DNA were added to 1-ml aliquots at the indicated final concentrations. Transformants were determined after 60 min at 37 C. Panel A. MW02 donor DNA,: ○, ComS_11–17; □, ComS_1–17; △, ComS_13–17; ◊, CSP. Panel B. pXPL16 donor DNA and indicated amounts of ComS_11–17 were added to recipient mutant strains: □, wild type UA159; Δ, MW04 (ΔoppD); ○, MW05 (ΔcomS); ◇, MW01 (ΔcomR). Error bars indicate standard deviation of triplicate assays.

To test the proposed role of ComS further, we asked if ComS_11–17 induced competence in comS and comR mutants. As shown in Fig. 4, addition of ComS_11–17 to the comS mutant induced high levels of competence, levels that were only slightly lower than that seen with the wild type. We conclude that the comS deletion mutant remains responsive to ComS_11–17 and infer that the only role of ComS is as a pheromone precursor. Finally, but importantly, ComS_11–17 was unable to rescue the transformation-deficient phenotype of the comR deletion, supporting the hypothesis that the competence-inducing properties of ComS are propagated through ComR.

Gram-positive bacteria detect peptide pheromones via two major pathways: a membrane-spanning sensor kinase that initiates a phospho-relay signaling pathway (e.g. ComAP of B. subtilis, ComDE of S. pneumoniae, and AgrCA of Staphylococcus aureus), or a permease that enables the active pheromone to reach an intracellular target (e.g. enterococcal sex pheromones, B. thuringensies PapR, the B. subtilis Phr peptides, and the S. thermophilus ComS pheromone). Since ComR, the putative receptor for ComS_11–17, lacks secretory signals and apparent trans-membrane domains, and since the ComS_11–17 product is active when supplied exogenously, it seemed reasonable that it would be imported to the cytoplasm to interact with ComR. Streptococci import oligo-peptides by means of ATP-dependent multi-subunit transporters known as Opp or Ami (Alloing et al., 1990). To ask whether the Opp transporter is needed for competence induction by ComS_11–17 in S. mutans, we constructed a strain in which one of the ATP-binding subunits of this transporter, OppD (SMU.258), was disrupted. Addition of ComS_11–17 peptide to this ΔoppD strain resulted in no tranformants, even at concentrations as high as 10 µM (Fig. 4), indicating that active ComS pheromone is imported to the cytoplasm rather than acting at the cell surface through a trans-membrane signal-transduction pathway, and that Opp is the principal or only permease for this transport.

ComR/ComS induce transcription of sigX in S. mutans

To test the hypothesis that ComS and ComR act by controlling the transcription of sigX, we constructed a transcriptional reporter plasmid that fuses DNA upstream of sigX, containing P1, to the bacterial luciferase genes, luxAB. The resulting P_sigX-luxAB plasmid was introduced into the S. mutans wild type strain UA159 and luciferase activity was monitored during growth in CDM. Without addition of exogenous pheromone, the level of P_sigX-luxAB expression was low and increased modestly (approximately ten-fold) as the culture reached high cell densities (Fig. 5A), consistent with the idea that sigX responds to endogenously accumulating pheromone in wild-type cultures. However, addition of ComS_11–17 pheromone to low-density cultures, even at concentrations as low as 50 nM, led to an immediate and much larger increase in LuxAB activity (Fig. 5A). Interestingly, full-length ComS also led to reporter induction, but with a long delay and at lower intensity. Finally, consistent with the transformation results above, the ComS_13–17 peptide caused no increase in LuxAB activity. Thus, we conclude that ComS_11–17 induces sigX transcription in CDM and infer that this effect explains the activity of this peptide in inducing competence.

Figure 5. Dependence of sigX transcription on exogenous peptides, comR and oppD, but not comS or comE.

Expression of the P_sigX-luxAB reporter in various genetic backgrounds was determined in chemically defined medium with exogenous addition of pheromones. Overnight cultures of S. mutans strains were diluted to OD=0.01 and grown until OD=0.1, at which time synthetic peptides were added (arrows) to a final concentration of 50 nM. Data shown are representative of three similar results. A) Wild-type UA159, B) MW18 (ΔcomR), C) MW17 (ΔcomS), D) MW22 (ΔoppD), E) MW21(ΔcomE). O, DMSO vehicle; □, ComS_1–17, full length peptide; ▼, ComS_11–17; △, ComS_13–17; ◇, CSP.

The P_sigX-luxAB reporter expression pattern in comR, comS, and oppD mutants was also consistent with the mechanism proposed above. Thus, the comR mutation abrogated both the immediate response of P_sigX-luxAB to ComS_11–17 (Fig. 5B) and the delayed response to full-length ComS peptide. In contrast, ComS_11–17 elicited a robust increase in reporter expression in the ΔcomS strain (Fig. 5C). Furthermore, the sigX transcriptional response to ComS_11–17 was fully dependent on the oligopeptide transporter (Fig. 5D). Taken together, these results demonstrate that comR is required for sigX induction in CDM and suggest strongly that the P1 promoter element is the link between the ComR/S regulatory circuit and development of competence for transformation.

ComR/ComS act downstream of ComE/CSP in S. mutans, to regulate expression of sigX

Among four different regulatory systems known to influence competence development in S. mutans, the comCDE pathway has been most thoroughly characterized. To compare the contributions of ComR and ComE to competence development directly, we determined transformation rates in parallel cultures of isogenic comR and comE strains. As shown in Table 3, the comE mutant was transformed in rich medium at a greatly reduced, but detectable, rate (~0.2% of the ComE⁺ strain), whereas the transformation frequency of a comR mutant was below the limits of detection (< 0.001% of the ComR⁺ strain). This pattern itself suggests a more central role for comR than comE in competence regulation. To ask more directly whether ComR acts upstream or downstream of ComE in the pathway regulating expression of sigX, we measured the response to competence stimulating peptide (CSP) in ΔcomR cells, and the response to ComS_11–17 in ΔcomE cells. Addition of CSP to cultures in THBHS increased transformation in the wild type approximately 100-fold, but not at all in either comE or comR mutants (Table 3). Also, CSP failed to induce expression of sigX in the comS or comR deletion strains (Fig. 5B&C). Thus, CSP stimulation does not bypass comR control. Conversely, as shown in Fig. 5E, deletion of comE did not block ComS_11–17 induction of sigX expression, showing that ComS_11–17 does bypass comE control. Both results imply that ComR acts downstream of ComE, but neither reveals the mechanism linking the two regulators. The model in which ComR is the proximal regulator of sigX is further supported by comparison of the kinetics of the responses to the two pheromones, as ComS_11–17 induced expression of sigX much more quickly than did CSP (Fig. 5A).

Table 3.

Effects of deletion of comR or comE on transformation in S. mutans

			Transformants per ml (std. dev.)
Strain	Genotype	DNA CSP	+ +		+ −		− +
UA159	WT		572,000	(45,000)	5,040	(670)	0	(4)
MW01	ΔComR		2	(1)	5	(4)	6	(4)
MW19	ΔComE		1,000	(150)	900	(47)	0	(4)

^aCompetence of indicated strains was determined after exposure at OD 0.1 in triplicate to 50 ng/ml of Spc^R genomic DNA from strain MW02 for 60 min under standard conditions in THBHS (Lindler and Macrina, 1986).

The foregoing experiments may be summarized by the model presented in Fig. 6. In this model, we propose that the previously documented regulators of competence, (ComCDE, HtrA, CiaHR, and HdrMR) lie upstream of ComR and ComS, and that their signaling information is integrated through the control of ComR/S activity, though their effects may not be direct and could be relayed through one or more undiscovered components. The translated product of comS is secreted from the cytoplasm by a transporter that has not been identified. Processing of ComS may occur outside the cell, as we have found that full-length synthetic ComS exhibits activity, but with a markedly delayed response, perhaps because it is the substrate for a peptidase whose product resembles ComS_11–17. We designate the mature form of ComS as CIP, for competence inducing peptide. The CIP pheromone is imported to the cytoplasm by the oligopeptide permease and interacts directly with ComR. Although there is no evidence for a direct interaction between ComR and ComS, this hypothesis is based on analogies to other structurally related peptide-binding transcription factors, PlcR and PrgX (see below). ComR is represented as a dimer to provide a simple model for direct binding to the inverted repeat of the P1 promoter element. The ComR/CIP complex then functions as a transcriptional activator of sigX, enabling development of competence.

Figure 6. Model for regulation of initiation of development of competence in S. mutans.

SigX, master regulator of late competence effector genes, and ComR, proximal regulator of sigX, are distinguished from upstream regulatory pathways mentioned in the text, whose effects may be integrated at the level of a ComRS quorum-sensing circuit.. The pheromone CSP, encoded by ComC(BIP), is secreted and processed at a GG site by the NlmTE ABC transporter, and sensed by ComD (BlpH), an autophosphorylating histidine kinase. The cognate response regulator ComE (BlpR) activates genes including cipB (Mutacin V), which establishes an undefined link to sigX expression. Components involved in secretion and processing of ComS are unknown, whereas uptake of the signal depends on the Opp/Ami transporter. ComR is drawn as a dimer binding to the P1 inverted repeat element, in analogy to PlcR and PrgX proteins. Upstream pathways are drawn for simplicity as acting to control transcription of comR, but they may act at any part of the ComR/ComS circuit. Upstream regulators include HtrA, CiaHR, and HdrRM, implicated as sensors of unknown extracellular or cell-surface-located signals.

Discussion

While S. pyogenes, S. agalactiae, and the other pyogenic species have long been considered ‘non-transformable’ members of the streptococci, accumulating evidence on their population structures shows that their gene allele distribution is non-clonal and highly mixed, as has been measured both for individual loci and at the genomic level (Bessen and Hollingshead, 1994; McShan et al., 2008). The source of the implied high rate of gene exchange has not been identified, but is usually attributed to transfers occurring via phages, plasmids, or transposons, with genetic transformation discounted. With the finding that S. pyogenes and all pyogenic streptococcal species possess both a conserved competence effector gene set and complete sigX regulatory structure (SigX and SigX-targeted late gene promoters) as well as upstream regulators that parallel the ComR/S regulators of the salivarius and mutans groups, we propose that transformation makes continuing important contributions to the evolution and population plasticity of the pyogenic streptococci. Although the natural conditions for competence are poorly understood in any streptococcal species, they may well be frequent enough that much of the population mixing known to occur could reflect exchange via transformation. It now appears likely that most or all streptococcal species regulate competence strictly, in a manner that uses one of three classes of quorum-sensing signals to coordinate the proximal regulator of sigX expression.

Development of competence for genetic transformation in S. mutans is absolutely dependent on expression of the alternative sigma factor, SigX, which allows transcription of effector genes of DNA uptake and recombination by permitting recognition of the non-canonical cin-box promoter motif TACGAATA (Okinaga et al., 2010). A variety of upstream regulators of sigX expression, including CiaRH, HtrA, ComCDE, CipB, and HrdRM, have a significant effect on sigX expression under one condition of growth or another, but none is individually absolutely required for sigX induction or development of competence (Ahn et al., 2005; Ahn et al., 2006; Okinaga et al., 2010; Qi et al., 2004; Perry et al., 2010). While multiple parallel pathways for stimulation of sigX expression have therefore been proposed, it remains unknown how the effects of such pathways are combined, or how they affect sigX expression. Standard competence-inducing conditions (THB+HS) have provided adequate conditions to test transformability, but how this artificial condition stimulates expression of sigX remains unknown and may work through more than one upstream regulator. While it is possible that the ComR/S system described here represents yet another sigX regulatory pathway parallel to those already identified, we propose instead that ComR is the proximal regulator of sigX, and may serve as a focal point for integration of information from the other pathways, for four reasons. First, complete loss of competence is unique to comR, comS, and oppD mutants, as loss of each of the other pathways mentioned causes only a partial competence deficiency. For example, deletion of comD in strain NG8 caused a 100-fold reduction in endogenous competence (Li et al., 2001), while in strain UA159, the reduction was only 3-fold and deletion of comC caused no reduction at all (Ahn et al., 2006). Second, a direct molecular link is apparent between ComR and sigX transcription, in the form of the P1 promoter that is shared by sigX and comS genes, whereas no other pathway has been linked so directly to the sigX promoter. Third, induction of sigX expression and of competence remains robust in response to ComS_11–17 in a comE mutant. Finally, both exposure to CSP and loss of hdrM or over-expression of hdrR have been reported to cause 3–5-fold increases in transcript levels for comR (SMU.61) and/or downstream flanking genes (supplemental tables in (Perry et al., 2009; Okinaga et al., 2010)), suggesting a possible link between upstream regulatory pathways and the ComR/S circuit.

The ComR proteins of mutans, pyogenic, and salivarius groups have multiple Rgg paralogues within each species. For example S. thermophilus contains at least six Rgg members and S. pyogenes, four. Rgg proteins have received attention as being 'stand-alone' regulators, known to respond to changes in growth conditions, but through unknown signaling pathways. They are characterized by an N-terminal 60-residue HTH domain and an ~220-residue C-terminus, rich in predicted alpha-helical structure. These regulators are conserved throughout the phylum Firmicutes, but very little is yet understood about their role in gene regulation beyond a few cases in a few species. Perhaps the most studied prototype is RopB of S. pyogenes, which is required for transcription of a secreted cysteine protease virulence factor as cultures reach high cell densities (Lyon et al., 1998; Chausee et al., 1999; Kreikemeyer and. McIver, 2003). Recent studies of two Rgg proteins in S. thermophilus have revealed a new mechanism that controls Rgg activity: they respond to peptide pheromones (Fontaine et al., 2010; Ibrahim et al., 2007a)

Curiously, each Rgg member has some degree of homology to the PlcR proteins of the cereus group of Bacillales, and among Rggs, ComR proteins are often the most similar to PlcR. Indeed, the ComR/ComS group of regulators could reasonably be regarded as either an Rgg subfamily or a PlcR subfamily on the basis of sequence similarity. PlcR is a transcriptional regulator that directly binds a small peptide ligand, PapR, forming an active transcriptional activator that controls at least 45 genes in the B. cereus genome (Slamti and Lereclus, 2002; Gohar et al., 2008). The genetic organization of comR-comS also resembles that of plcR and papR genes, where the peptide-encoding ORFs are located downstream of the regulator gene, distinguishing them from the 'rgg-associated SHP' gene organization that was recognized by Ibrahim and co-workers (2007a). This study found that non-annotated open reading frames encoding short hydrophobic peptides (SHPs) are commonly found adjacent to rgg genes, but their biological significance has only been demonstrated twice, both in S. thermophilus gene regulation (Fontaine et al., 2010; Ibrahim et al., 2007). Not all rgg genes are coupled to ORFs that encode a recognizable signaling peptide; such exceptions may instead respond to other types of signals; or to peptides from genetically distant origins.

Analyses of the four S. pyogenes Rgg protein sequences (ComR, Spy0496, Spy0533, and RopB) using algorithms that predict secondary and tertiary structures of proteins (PHYRE (Kelley and Sternberg, 2009) and I-Tasser (Roy et al., 2010)) found high degrees of similarity between the predicted structure of each Rgg and the crystal structure of PlcR. Predicted similarities include both functional domains of PlcR: an N-terminal xenobiotic regulatory element (XRE) containing a helix-turn-helix (HTH) motif and a large C-terminal alpha-helical bundle domain containing a tetratricopeptide repeat (TPR). The PapR peptide interacts directly with the TPR domain to induce conformational changes that modulate interactions of PlcR with its DNA targets, altering expression of target genes (Declerck et al., 2007; Gohar et al., 2008; Slamti and Lereclus, 2002). Although TPR prediction algorithms (Karpenahalli et al., 2007) have not identified a TPR domain in Rgg proteins, the proposed interaction of CIP with ComR suggests that the Rgg C-terminus directly binds to peptide pheromones, and may offer a new structural fold for binding peptides. Furthermore, the present evidence that ComR proteins of S. mutans and S. thermophilus rely on short peptides for activation of transcription of target genes, together with predicted structural similarities to PlcR, provides further support to the recently introduced notion (Ibrahim et al., 2007; Fontaine et al., 2010) that Rgg proteins fit well with PlcR in a class of transcription factors that directly bind small peptides to modify their activities. A recent study further suggests a large family of regulatory proteins exists that bind cytoplasmic peptide signals. Members of this family include, in addition to PlcR, the transcriptional regulators PrgX of Enterococcus faecalis and NprR of B. thuringiensis, and the Rap phosphatases of the bacilli, generating a superfamily designated RNPP (Rap/NprR/PlcR/PrgX) (Declerck et al., 2007). With the present finding that ComR responds to a peptide stimulus and with its predicted structural similarity to PlcR, it now appears that the RNPP super-family should be considered to include the Rgg family as well, suggesting designation as the RRNPP group. The domain by which peptide ligands bind to (R)RNPP proteins unifies the group, whereas functional effector modules, encoding DNA-binding or phosphatase domains, distinguishes subgroups (Declerck et al. 2007).

Rgg proteins have been designated 'stand-alone regulators', meaning it is not understood what signals and/or interacting cognate sensory elements control Rgg activity (Kreikemeyer et al., 2003). The demonstration that ComR of S. mutans is the third Rgg homolog to rely on a signaling peptide raises the question of whether any other Rgg proteins also mediate peptide signaling, and may offer the key to understanding one category of 'stand-alone' regulator. With Rgg homologs found throughout most Gram-positive genomes, their newly proposed function as quorum-sensing regulators may require reassessment of regulatory circuits that include Rgg proteins.

Interaction of (R)RNPP regulators with DNA have been studied for PlcR and PrgX. Both proteins form dimers (and PrgX a tetramer) (Shi, et al. 2005; Declerck et al. 2007) and bind to inverted repeats consisting of six or eight base pair half-sites (Gohar et al. 2008; Bae et al., 2002). Consistent with these family members’ DNA binding properties, we predict that ComR proteins recognize and directly interact with the P1 and P2 promoter elements, which are comprised of direct inverted repeats of nine bp half-sites. If this is true, the ComR proteins may also form dimers or higher-ordered multimers whose interactions could be regulated by the ComS pheromone ligand.

Mature peptide pheromones of Gram-positive bacteria are products of precursor polypeptides that are processed and sometimes modified. Each pheromone contains signatures that serve as secretion substrates, sites for processing and modification, and sites conferring ligand specificity for interaction with their target receptors. The ComS peptides described here for pyogenic and mutans streptococci appear to contain characteristics unseen in other quorum-sensing peptides. The most obvious property is the consistent presence of a double tryptophan near the C-terminus of the peptide. This motif may encode a specialized processing or modification site required for maturation, as has been seen in other extracellular signaling peptides (for example, Lanigan-Gerdes et al., 2008, Ansaldi, et al., 2002; Okada et al., 2005), but the high activity of ComS11–17 shows that any such modifications are not critical for activity, unless the synthetic peptide is efficiently modified upon addition to cells or is modified during transport. The WW motif may instead be a unique identifier for the target receptor, presumably ComR in this case. The shared motif may offer opportunities for cross-species recognition or antagonism.

The overall structure of four CIP pheromones (S. pyogenes, S. uberis, S. dysgalactiae, and S. equi) somewhat resemble secretion signal sequences (Table 2), containing a basic N-terminus, hydrophobic central core, and polar C-terminus (von Heijne, 1985), but lack an attached C-terminal peptide destined for secretion, and in this way resemble the composition of enterococcal conjugation inhibitors (Nakayama et al.,1994). More interesting is that the other ComS orthologs (S. mutans, S. agalactiae, S. gallolyticus, and S. infantarius) lack resemblance to a signal sequence in that they are very short (15–17 residues), and do not contain appropriate N-and C-terminal signatures, although they do contain an overall hydrophobic quality. Therefore, if these peptides are secreted by homologous pathways, how these dissimilar peptides are recognized by the cell could reveal undescribed secretion signatures. An alternative explanation may be that the two groups of peptides are secreted by separate systems, yet undiscovered.

Application of the new class of CIPs for inducing the competent state and making genetic transformation possible in streptococci of the pyogenic group will depend, at a minimum, on discovery of the biologically active forms of the CIP signal products for each species. Since the WW motif is conserved, characteristics that define activity may be universal and allow for quick identification of functional pheromones for all members of the group. Roadblocks likely remain in determining environmental factors required to induce ComR expression. Even in S. mutans, where many upstream regulators have been identified, we were able to enhance transformability by increasing the gene copy number of comR on a plasmid. Another barrier to transformation will be the high amounts of nucleases many of these species produce and secrete, indicating the need to study kinetics of the optimal competent state during growth.

If our prediction that Rgg proteins bind small peptides for so as to regulate activities that reflect population density and peptide production by the bacterial community, it will be interesting to identify other behaviors regulated by quorum sensing among streptococci containing Rgg regulators. For example, RopB, an Rgg family member of S. pyogenes, induces speB transcription at high cell densities. An unknown density-dependent factor has been proposed, but has yet to be identified (Neely et al., 2003). Considering the emerging possibility that Rgg proteins are peptide receptors, the factor might be a peptide pheromone that accumulates in cultures as cells reach high density. Important questions to address for each new Rgg-dependent pathway include identifying the active structures of signaling peptides, processing and transport components, and integration of upstream and parallel signaling pathways.

Experimental Procedures

Bacterial strains, media, and oligonucleotides

Strain UA159 was a kind gift from Lin Tao (UIC Dental College). Other strains and plasmids are described in Table 4. Oligonucleotide primers are listed in Table S3. Cultures of S. mutans were grown in closed tubes at 37 C in Todd-Hewitt Broth (Difco) (THB) or chemically defined medium (CDM) and stored at −80 C in the same medium supplemented with 10% glycerol. For selection, samples were embedded in THB containing 0.75% agar, and covered with selective drug agar. Plates were incubated in 5% CO₂ or in a candle jar at 37 C for 40 hrs before counting the resulting colonies. Selective levels of antibiotics were 1.5 µg erythromycin/ml, 200 µg spectinomycin/ml, or 7.5 µg chloramphenicol/ml. The CDM described previously (van de Rijn and Kessler, 1980) was used after supplementation with L-asparagine to 100 µg/ml.

Table 4.

Strains and plasmids used in this study

S. mutans strains	Description	Source
UA159	Transformable S. mutans isolate	Tao et al., 1993
MW01	UA159 but ΔcomR∷cat; CmR	This study
MW02	UA159 but ΔcomR∷spc; SpcR	This study
MW03	UA159 with pWAR300(ComR); EmR	This study
MW04	UA159 but ΔoppD∷spc; SpcR	This study
MW05	UA159 but ΔcomS∷spc; SpcR	This study
MW08	MW03 but ΔcomR∷spc (from strain MW02); SpcR EmR	This study
MW14	UA159 with pWAR304(PxLux); EmR	This study
MW17	MW14 but ΔcomS∷spc (from strain MW05); SpcR EmR	This study
MW18	MW14 but ΔcomR∷spc (from strain MW02);; SpcR EmR	This study
MW19	UA159 but ΔcomE∷erm (from strain ΔSMU.1917); EmR	This study
MW21	MW19 with pWAR307(PxLux); EmR CmR	This study
MW22	MW04 with pWAR304(PxLux); SpcR EmR	This study
Plasmid
pCN59	pAMβ-1 shuttle plasmid contaiing promoterless luxAB of V fischeri (accession #X06758)	Charpentier, et al, 2004
pGh9-ISS1	pG+host shuttle plasmid containing ISS1 transposable element; ts, EmR	Maguin et al., 1996
pFED760	pGh9-ISS1 derivative deleted for ISS1	This study
pFED761	Heat-resistant pFED760 derivative containing wild-type repA allele (repA 972S, 977D, 980V, 987R; see Maguin et al., 1992)	This study
pJC156	pFED761 derivative containing cat in place of ermB; 3,790 bp; CmR	This study
pWAR300	pFED761 derivative carrying a 1414-bp fragment with comR and 454 bp upstream, inserted between NotI and SalI sites; 5,098 bp; EmR	This study
pWAR303	pFED761 derivative carrying a 2222-bp fragment with luxAB from pCN59 inserted between PstI and NotI sites; 5,946 bp; EmR	This study
pWAR304	pWAR303 derivative carrying a 561-bp fragment with PsigX between the SalI and PstI sites; 6,466 bp; EmR	This study
pWAR307	pJC156 derivative carrying PsigX-luXAB between the SalI and NotI sites; 6,504 bp; EmR	This study
pXPL16	pMSP3535 derivative carrying comW-His6 between PstI and XbaI sites; 8,589 bp; EmR	Luo and Morrison, 2004

Abbreviations: Cm, chloramphenicol; Em, erythromycin; Spc, spectinomycin.

Genome analysis

Whole-genome scans were performed using the FUZZNUC program of Emboss (Rice et al., 2000) provided by the MOBYLE@Pasteur server. Protein structure prediction was performed on PHYRE and I-Tasser serveres (Kelley and Sternberg, 2009; Roy et al., 2010) located at sbg.bio.ic.ac.uk/~phyre, and zhanglab.ccmb.med.umich.edu/I-TASSER.

Transformation assay

Transformation was carried out as described previously (Lindler and Macrina, 1986). Briefly, an overnight culture in THB supplemented with 5% v/v sterile heat-treated horse serum (Sigma Chemical, St Louis) (THBHS) was aliquoted for storage at −80 C. For transformation, 25 µl of the frozen stock was added to 1 ml THBHS and incubated at 37 C in closed 1.5-ml Eppendorf tubes. After 3 hrs, 50 ng donor DNA was added, and incubation continued for 60 min more, before plating to select transformants. For assay in CDM, cultures growing from 1:100 dilution in CDM were incubated 60 min at 37 C after addition of small volumes of peptide and/or DNA stocksr before selection in THB as above.

Construction of mutants and complementation plasmids

New mutants were constructed according to Lau et al. (2002), using primers listed in Table S3. The Spc^R comR, comS, and oppD mutants (MW02, MW05, and MW04) were constructed by insertion of a spectinomycin resistance cassette (aad9) within the specified ORFs. One-kb fragments upstream and downstream of comR (designated as comR-US and comR-DS) were amplified by PCR using the primer pairs LMW1/LMW2 and LMW3/LMW4 and UA159 DNA as template. These primers created SalI and PstI sites at the 3’ end of comR-US and the 5’ end of comR-DS respectively. The aad9 gene was amplified from pLZ12Spec (Husmann et al., 1995) using primers LMW5/LMW6 generating a SalI site at the 5’ and a PstI site at the 3’ ends. Two resulting PCR fragments were fused at SalI sites by in vitro ligation. This ligation reaction was amplified using primers LMW1/LMW6, creating the comR-US-aad9 fusion. The comR-US-aad9 product and comR-DS were ligated in vitro at PstI sites, resulting in the linear comR knockout construct, which was used to transform UA159 as described above, creating strain MW02. The knock-out constructs for comS and oppD were assembled in the same manner, using primers LMW7-LMW12 and LMW17–21/LMW12 respectively. The comR knockout construct used to create strain MW01 (ΔcomR; Cm^R) contained a chloramphenicol resistance cassette (cat) flanked by comR-US and comR-DS and was constructed similarly, but with a cat gene amplified from pEVP3 (Claverys et al., 1995) using primers LMW15/16. Confirmation of mutations was completed by analytical PCR using genomic DNA from the mutant strains to verify the presence of expected new junctions and the absence of intact target genes. To create strain MW19, genomic DNA from a UA159 ΔSMU.1917∷erm mutant, obtained as a gift from Celine Levesque (Perry et al., 2009), was transformed into UA159.

The plasmid pFED760 was obtained by removing the ISS1 element from pGh9:ISS1 (Maguin et al., 1996) by inverse PCR using primers MF1/MF2 and introducing an EcoRI site. A variant of pFED760, capable of replication at temperatures above 30 C, designated pFED761, was isolated by growing an Escherichia coli transformant at 37 C to select for the wild-type allele of repA (Maguin et al, 1992). The comR complementation plasmid was assembled by amplifying comR with 454 bp upstream to include the comR promoter region using primers LMW13/LMW14. This DNA fragment was cloned into pFED761 using NotI and SalI restriction sites, resulting in the comR complementation plasmid designated pWAR300. pWAR300 was transformed into E. coli BH10c (Howell-Adams and Seifert, 2000) by electroporation and selection on Luria-Bertani agar containing 500 µg/ml erythromycin. To create the comR complementation strain (designated MW03), pWAR300 was transformed into S. mutans UA159 as described above. To place the comR deletion within the complementation strain, chromosomal DNA from strain MW02 was transformed into MW03 to create strain MW08.

Construction of reporter plasmids and strains

The P_sigX-luxAB reporter plasmid was assembled by amplification of luxAB from pCN59 (Charpentier et al., 2004) using primers LMW22/23, followed by insertion into pFED761 using PstI and NotI sites, resulting in the plasmid pWAR303. The promoter region (550 bp) upstream of sigX was amplified using primers LMW24/25 and cloned into pWAR303 using SalI and PstI sites to create the P_sigX-luxAB reporter plasmid pWAR304. Both constructs were transformed into E. coli BH10c as described above. To generate the P_sigX-luxAB reporter strain (designated MW14), pWAR304 was transformed into S. mutans UA159 as described above. To create the comR, comS, and oppD deletions within the P_sigX-luxAB reporter strain, chromosomal DNA from mutant strains MW02, MW05, and MW04 were transformed into MW14 to create strains MW18, MW17, and MW22 respectively. A P_sigX-luxAB reporter plasmid containing cat was obtained by amplifying the P_sigX-luxAB fusion from pWAR304 using primers LMW24/23. This DNA fragment was cloned into pJC156 using NotI and SalI restriction sites to create pWAR307. pJC156 is derivative of pFED761 in which the Em^R gene has been replaced with cat (Cm^R) using primers JC98-JC101 and methods described previously (Datta et al.). pWAR307 was transformed into the comE mutant MW19 to produce strain MW21 as described above.

Preparation of Synthetic Peptides

Peptides were purchased from NEO-Peptide (Cambridge, MA) at a 48–52% purity grade. Stock solutions were dissolved in DMSO at a concentration of 1 mM, taking into consideration their specific purity. Peptide sequences are as follows: Smu-UA159-ComS _1–17 (MFSILTSILMGLDWWSL); Smu-UA159-ComS_11–17 (GLDWWSL); Smu-UA159-ComS_13–17 (DWWSL); and Smu-sCSP (SGSLSTFF RLFNRSFTQALGK).

Reporter Assays

Throughout growth, optical densities were measured using a Milton Roy Spectronic 20D spectrophotometer (Fisher Scientific, Pittsburgh, PA) and luminescence was assayed with a Wallac 1450 Microbeta Plus Liquid Scintillation Counter (Perkin Elmer, Walthum, MA). Reporter strains were grown at 37 C in CDM after dilution from mid-logarithmic stocks to OD_600nm 0.01. Cultures were grown to OD_600nm 0.1 and supplemented with exogenous peptides. 100-µL aliquots were placed in a Falcon white flat-bottom 96-well plate (Becton Dickinson Labware, Franklin Lakes, NJ) and exposed to 50 µL of decyl aldehyde spread on the plate lid for ~30 seconds prior to light detection.