A unique open reading frame within the comX gene of Streptococcus mutans regulates genetic competence and oxidative stress tolerance

Abstract

Streptococcus mutans displays complex regulation of genetic competence, with ComX controlling late competence gene transcription. The rcrRPQ operon has been shown to link oxidative stress tolerance, (p)ppGpp metabolism and competence in S. mutans. Importantly, an rcrR polar (ΔrcrR-P) mutant is hyper-transformable, but an rcrR non-polar (ΔrcrR-NP) mutant cannot be transformed. Transcriptome comparisons of the rcrR mutants using RNA-Seq and quantitative real-time polymerase chain reaction revealed little expression in the 5′ region of comX in ΔrcrR-NP, but high level expression in the 3′ region. Northern blotting with comX probes revealed two distinct transcripts in the ΔrcrR-P and ΔrcrR-NP strains, and 5′ Rapid Amplification of cDNA Ends mapped the 5′ terminus of the shorter transcript to nt +140 of the comX structural gene, where a unique 69-aa open reading frame, termed XrpA, was encoded in a different reading frame than ComX. Two single-nucleotide substitution mutants (comX::T162C; comX::T210A) were introduced to disrupt XrpA without affecting the sequence of ComX. When the mutations were in the ΔrcrR-NP genetic background, ComX production and transformation were restored. Overexpression of xrpA led to impaired growth in aerobic conditions and decreased transformability. These results reveal an unprecedented mechanism for competence regulation and stress tolerance by a gene product encoded within the comX gene that appears unique to S. mutans.

Introduction

Horizontal gene transfer in prokaryotes offers a method for adaptation to changing environmental conditions through acquisition of new genetic traits. Natural bacterial transformation in Streptococcus mutans, a member of the oral microbiome and the primary causative agent of dental caries, was first described in 1981 by Perry and Kuramitsu (Perry and Kuramitsu, 1981). Development of genetic competence during natural transformation is defined as the transient physiological state during which proteins necessary for DNA uptake and processing are expressed, leading to internalization of exogenous DNA (Cvitkovitch, 2001). The development of competence benefits bacteria in a variety of ways, including genome repair (Gagne et al., 2013), genome diversification (Johnston et al., 2013), use of DNA as a nutritional source (Sinha et al., 2013) and release of extracellular DNA for utilization as a structural lattice in biofilm formation (Bakkali, 2013). In many bacteria, but in S. mutans in particular, competence development is tightly integrated with pathways that are required for establishment, persistence and virulence of the organism (Lemos et al., 2013).

In S. mutans, competence development can be stimulated by two small signaling peptides: competence-stimulating peptide (CSP) and comX-inducing peptide (XIP) (Li et al., 2002; Mashburn-Warren et al., 2010). The comC gene encodes the precursor of CSP, a ribosomally translated 46-amino acid (aa) peptide (Senadheera and Cvitkovitch, 2008). This peptide requires the ComAB ATP-binding cassette (ABC) transporters for export and maturation into the 21-residue CSP (Hui et al., 1995). Further processing by the surface-localized SepM protease yields an 18-residue CSP molecule that is the more biologically active form of the peptide (Hossain and Biswas, 2012). When CSP accumulates in sufficient concentrations outside of the cell (~ 30 nM), it stimulates the ComDE two-component signal transduction system (TCS). CSP of S. mutans can cause increased expression of the comX gene and greatly enhance transformation efficiency, but this effect does not appear to occur as a direct result of ComE binding to the comX promoter (Hung et al., 2011), as is the case for the ComCDE pathway of Streptococcus pneumoniae. Rather, ComDE of S. mutans activates a set of class II bacteriocins (Kreth et al., 2007) and appears to have the same or similar ancestral origins as the bacteriocin regulatory TCS BlpRH of S. pneumoniae (Håvarstein, 2010; Johnston et al., 2014). It is now established that direct activation of the comX gene in S. mutans occurs via the ComRS system. The comS gene encodes the 17-aa precursor of the 7-aa XIP, although the gene products required for secretion and processing of ComS to XIP have not yet been identified. Exported XIP can be found in concentrations as high as 0.9 μM in cellular supernates (Wenderska et al., 2012), and XIP is internalized through the oligopeptide permease OppA (Desai et al., 2012; Khan et al., 2012). Once inside the cell, XIP interacts with ComR, an Rgg-like transcriptional regulator. The ComR-XIP complex activates transcription from the promoters for comX and comS, the latter creating a positive feedback loop (Fontaine et al., 2013). The ability of CSP or XIP to activate competence is highly dependent on growth conditions and growth phase, and in the case of CSP, a bimodal (subpopulation) response by the bacteria is commonly observed (Son et al., 2012; Wenderska et al., 2012; Guo et al., 2014).

A ComX ortholog is ubiquitous in streptococci and is also present in many lactobacilli (Martin et al., 2006). Of note, S. mutans carries only one copy of the comX gene, whereas other streptococci, including S. pneumoniae, harbor two copies of comX (Lee and Morrison, 1999). Recently, the anginosus group of streptococci was shown to contain three comX loci in most strains (Olson et al., 2013). When ComX levels in the cell increase, genes that encode DNAuptake machinery, which are orthologs of type IV pilus systems (Laurenceau et al., 2013), and enzymes that facilitate homologous recombination are activated (Mair et al., 2012). However, the effects of enhanced expression of comX are not restricted to competence-related genes, as ComX was found to affect the expression of about 240 genes in S. mutans (Perry et al., 2009), although many of these genes are not directly activated by ComX. Direct activation of genes by ComX requires an 8 bp sequence (TACGAATA) that is recognized by the RNA polymerase-ComX holoenzyme (Luo et al., 2003). In both Bacillus subtilis and S. pneumoniae, induction of the ComX regulon is more broadly referred to as the K- or X-state, respectively, as close to three-quarters of the genes that are activated are not required for transformation (Berka et al., 2002; Claverys et al., 2006). The K- or X-states are considered to be a global stress response to repair DNA damage and generate genetic diversity (Prudhomme et al., 2006). Thus, understanding the factors that govern the commitment to competence should provide insights into the ability of organisms to adapt to environmental fluctuations, a trait that is considered essential for persistence and virulence of S. mutans in the oral cavity.

Recent studies into multiple stress response pathways, such as VicRK (Senadheera et al., 2012), HtrA (Ahn et al., 2005), CiaRH (Ahn et al., 2006), HdrMR (Okinaga et al., 2010) and others (Merritt et al., 2005a; Abranches et al., 2006; Banu et al., 2010; Dmitriev et al., 2011; Kajfasz et al., 2011; Kim et al., 2013) suggest there is strong overlap between the competence circuit and tolerance of environmental stressors. An additional operon, rcrRPQ, was recently characterized and appears to serve as a molecular link between (p)ppGpp metabolism, acid and oxidative stress tolerance, and genetic competence (Seaton et al., 2011). The operon consists of genes for a MarR-like protein (rcrR) that negatively regulates the expression of the operon and a pair of ABC transporters that are predicted to function as exporters (rcrPQ). Two different deletion-replacement mutations of the rcrR gene confer drastically different phenotypes: a polar ΔrcrR mutant (ΔrcrR-P) is constitutively hyper-transformable, whereas a non-polar mutant (ΔrcrR-NP) is non-transformable under all conditions tested. Of relevance here, the polar rcrR mutant expresses very high levels of the late competence gene comYA, whereas comYA mRNA levels in the non-polar strain are lower than baseline levels in a wild-type (WT) strain that has not been treated with exogenous CSP or XIP (Seaton et al., 2011). The differences in genetic competence are associated with major changes in the expression levels of theABC exporters, with the non-polar mutant expressing over 100-fold higher rcrPQ mRNA than the polar and WT strains. Additionally, two peptides that are encoded as part of the rcrRPQ operon, at the end of rcrQ, and that are overexpressed in the non-polar mutant have been shown to function as effector molecules that adversely affect transformation efficiency in S. mutans (Ahn et al., 2014). RcrR may also directly regulate com genes by binding to the promoter regions of comX and comYA, albeit at lower efficiency than to the rcrRPQ promoter (Seaton et al., 2014). The main objective of this study was to examine in more detail the ΔrcrR-P and ΔrcrR-NP strains to determine the molecular basis for the difference in competence behaviors. The data reveal a previously unrecognized effector molecule and a novel regulatory mechanism for the control of genetic competence that appears to be unique to S. mutans.

Results

Transcriptome profiling of ΔrcrR-P and ΔrcrR-NP strains

A variety of single or combination mutations introduced into genes in the rcrRPQ operon have significant effects on (p)ppGpp metabolism, acid and oxidative stress tolerance, and genetic competence (Seaton et al., 2011). Of note, two different mutations of rcrR display drastically different phenotypes in relation to competence development. A polar mutation of the rcrR gene (ΔrcrR-P) is hyper-transformable, displaying a 10⁴-fold increase in transformation efficiency in the absence of added sCSP. In contrast, a non-polar mutation (ΔrcrR-NP) renders the strain non-transformable under all conditions tested. Seaton and colleagues (2011) used quantitative real-time polymerase chain reaction (qRT-PCR) to show that the expression of comYA, a late competence gene that is required for natural competence in S. mutans (Merritt et al., 2005b), is upregulated more than 100-fold in the hyper-transformable ΔrcrR-P strain, but downregulated nearly 15-fold in the non-transformable ΔrcrR-NP, compared with the WT UA159 strain. To begin to understand the molecular genetic basis for these observations, whole transcriptome shotgun sequencing of enriched mRNA (RNA-Seq) was performed as previously described (Zeng et al., 2013) to compare the gene expression profiles of these two mutant strains. In the ΔrcrR-P strain, we found the most highly upregulated transcripts to be late competence genes in the comY, comG and comE operons, as well as several genes for bacteriocins [bsmB (SMU.1906c) and nlmB (SMU.151)] (Table 1) compared with the WT strain. In contrast, these transcripts were not upregulated in the ΔrcrR-NP strain (Table 2). The genes showing the greatest degree of downregulation in both strains compared with the WT were those for pyruvate-formate lyase (pfl), the pyruvate-formate lyase-activating enzyme (pflA) and superoxide dismutase (sod). Given the effects of rcrRPQ mutations on oxidative stress tolerance, it is noteworthy that Pfl of S. mutans is inhibited by oxygen and that Sod is a major defense against oxidative stress. We were able to identify a sequence with some homology to the consensus RcrR-binding sites (Seaton et al., 2014) in the pfl promoter region and a weaker potential RcrR-binding site in the sod promoter region (data not shown), so it is possible these genes may be direct targets of RcrR. Several predicted amino acid ABC transporters (SMU.460, SMU.933–936), glycerol dehydrogenase (gldA) and the mannitol PTS EII (mltA) were also down-regulated in the ΔrcrR-NP strain compared with the parental UA159. The promoter regions of these genes were also examined for the consensus RcrR-binding site, but no homology was found. Taken together, these data show that the RcrRPQ may have a fairly broad impact on cellular physiology.

Table 1.

Genes differentially regulated in the ΔrcrR-P strain versus WT strain UA159 via RNA sequencing analysis.

Gene ID	Gene description	Fold change ΔrcrR-P/WT	P value
SMU.1001	DNA processing protein	279	0
SMU.1987c	Late competence protein; type II secretion system protein E	247	1.58E-32
SMU.1980c	Conserved hypothetical protein	227	5.36E-86
SMU.498	Late competence protein F	210	0
SMU.1983c	Competence protein ComYD	209	1.99E-145
SMU.1984c	Competence protein ComYC	203	4.35E-165
SMU.1981c	Competence protein G	203	1.62E-34
SMU.1982c	Conserved hypothetical protein	200	2.26E-134
SMU.1985c	Competence protein; general (type II) secretory pathway protein	191	3.47E-26
SMU.499	Late competence protein required for DNA uptake	170	2.81E-164
SMU.626	Competence protein; possible integral membrane protein	143	8.63E-299
SMU.625	Competence protein	112	2.48E-172
SMU.1967c	Single-stranded DNA-binding protein	84.5	7.18E-50
SMU.644	Competence protein CoiA	53	3.21E-220
SMU.1912c	Hypothetical protein	25.7	7.58E-16
SMU.1910c	Hypothetical protein	24.7	3.37E-17
SMU.1997c	Competence-specific sigma factor	22.8	2.18E-219
SMU.539c	Prepilin peptidase type IV	21	4.04E-07
SMU.1906c	Bacteriocin-related protein	16	4.95E-06
SMU.151	Non-lantibiotic mutacin IV B	10	3.89E-05
SMU.1576c	Hypothetical protein	8.66	1.46E-23
SMU.1055c	DNA repair protein RadC	8.65	3.37E-101
SMU.10	Cell-division protein DivIC	7.3	8.20E-07
SMU.46	Hypothetical protein	7.27	2.39E-09
SMU.2086c	Competence damage-inducible protein A	6.59	2.66E-58
SMU.196c	Immunogenic secreted protein (transfer protein)	4.1	1.99E-07
SMU.2085c	Recombinase A	3.23	7.55E-62
SMU.1915	Streptococcus mutans specific competence stimulating peptide	3.23	5.48E-05
SMU.1345c	Peptide synthetase similar to mycA	3.25	7.34E-07
SMU.1828	Universal stress protein family	0.652	1.46E-08
SMU.132	Amino acid amidohydrolase (hippurate amidohydrolase)	0.39	1.46E-20
SMU.1692c	Pyruvate-formate lyase activating enzyme	0.385	8.13E-32
SMU.540	Peroxide resistance protein/iron-binding protein	0.385	2.55E-45
SMU.923	ABC-type multidrug/protein/lipid transport system	0.368	1.11E-51
SMU.932	Conserved hypothetical protein	0.335	6.55E-11
SMU.629	Superoxide dismutase	0.301	3.04E-71
SMU.924	Thiol peroxidase	0.218	1.35E-68
SMU.921	Transcriptional regulator	0.086	4.55E-38

Table 2.

Genes differentially regulated in the ΔrcrR-NP strain versus WT strain UA159 via RNA sequencing analysis.

Gene ID	Gene description	Fold change (ΔrcrR-NP/WT)	P value
SMU.1997c	Competence-specific sigma factor	37.1	1.80E-273
SMU.922	ABC-type multidrug/protein/lipid transport system	16.6	9.31E-280
SMU.923	ABC-type multidrug/protein/lipid transport system	13.6	5.04E-246
SMU.09	Small RNA binding protein	7.91	9.83E-56
SMU.43	Conserved hypothetical protein	7.76	3.71E-31
SMU.1576c	Hypothetical protein	7.62	3.26E-21
SMU.44	Conserved hypothetical protein	7.39	1.07E-25
SMU.64	Holliday junction DNA helicase	7.3	2.58E-139
SMU.10	Cell-division protein DivIC	4.55	0.0001
SMU.66	Conserved hypothetical protein	3.71	2.65E-36
SMU.533	Anthranilate synthase	3.62	3.81E-33
SMU.1340c	Bacitracin synthetase 1/tyrocidin synthetase III	3.22	2.71E-07
SMU.537	Tryptophan synthase	3.22	2.24E-43
SMU.532	Anthranilate synthase	3.09	3.93E-39
SMU.928	Sensor histidine kinase	0.476	1.09E-06
SMU.463	Thioredoxin reductase (NADPH)	0.438	6.04E-26
SMU.1692c	Pyruvate-formate lyase-activating enzyme	0.426	4.65E-25
SMU.540	Peroxide resistance protein/iron-binding protein	0.422	1.16E-36
SMU.924	Thiol peroxidase	0.414	2.42E-24
SMU.179	Conserved hypothetical protein (possible oxidoreductase)	0.365	4.21E-10
SMU.494	Transaldolase family protein	0.361	5.46E-16
SMU.491	Transcriptional regulator	0.356	2.23E-11
SMU.1955c	Co-chaperonin 10 kDa	0.322	1.03–27
SMU.495	Glycerol dehydrogenase	0.319	2.13E-27
SMU.460	Amino acid ABC transporter	0.319	5.13E-22
SMU.629	Superoxide dismutase	0.316	7.12E-64
SMU.996	ABC transporter	0.304	7.23E-06
SMU.961	Macrophage infectivity potentiator-related protein	0.297	2.30E-19
SMU.1425c	ATP-dependent Clp protease	0.295	4.72E-12
SMU.929c	Conserved hypothetical protein	0.292	3.35E-40
SMU.82	Chaperone protein	0.285	4.56E-14
SMU.493	Formate acetyltransferase (pyruvate-formate lyase)	0.282	7.74E-43
SMU.962	Acyl-CoA dehydrogenase	0.239	6.88E-45
SMU.496c	Cysteine synthetase A	0.237	7.36E-17
SMU.1701c	Conserved hypothetical protein	0.236	0.0001
SMU.80	Heat-inducible transcription repressor	0.229	1.43E-18
SMU.1185c	Mannitol PTS EII	0.11	5.79E-66
SMU.935	Amino acid ABC transporter	0.095	4.32E-30
SMU.933	Amino acid ABC transporter	0.094	4.63E-39
SMU.936	Amino acid ABC transporter	0.094	2.31E-27
SMU.921	Transcriptional regulator	0.08	3.40E-39
SMU.932	Conserved hypothetical protein	0.08	1.40E-44
SMU.1395c	Hypothetical protein	0.038	1.52E-10

The 5′ region of comX is differentially expressed in the ΔrcrR-P and ΔrcrR-NP strains

The expression level of the comX gene was previously measured by qRT-PCR and shown to be upregulated nearly 100-fold in both the ΔrcrR-P and ΔrcrR-NP strains (Seaton et al., 2011). Thus, it was hypothesized that rcrRPQ affects competence by interfering with ComX activation of late competence genes, including comYA. However, while analyzing the number of reads that mapped to the comX gene (SMU.1997), we discovered that a very low number of reads (29 ± 4) could be mapped to the 5′ region of comX in the ΔrcrR-NP mutant (chromosomal position 1 872 351–1 872 503) (Table 3, Supporting Information Fig. S1A), which is in stark contrast to the ΔrcrR-P mutant, where 1067 ± 57 reads mapped to the same region (Supporting Information Fig. S1B). Importantly, more reads were observed in the central and 3′ portions of the comX coding sequence, between positions 1 872 020 and 1 872 351, in the ΔrcrR-NP strain. In fact, the average read map for the non-polar mutant in this region exceeded that for the ΔrcrR-P by about twofold. To further explore the RNA-Seq results, expression of comX was measured in both strains by qRT-PCR using 5′ and 3′ comX primer sets, such that the 5′ set amplified a region between nucleotides 28 and 109 of comX and the 3′ set amplified nucleotides 223–311 of comX (Fig. 1A, Supporting Information Table S1). Although the levels of 5′ and 3′ comX mRNA were about 100-fold higher in the ΔrcrR-P mutant compared with the WT strain, there were no statistically significant differences observed between the expression levels of the 5′ and 3′ regions in either of these strains (Fig. 1B). However, in the ΔrcrR-NP strain, comX expression was dramatically lower in the 5′ region (8.2 × 10¹ copies × μg RNA⁻¹) compared with the 3′ region (1.3 × 10⁶ copies × μg RNA⁻¹). These data confirm the RNA-Seq results and indicate that a sense transcript is highly expressed from the 3′ portion of the comX gene in the ΔrcrR-NP mutant. Note that the discrepancy between the RNA-Seq and qRT-PCR data presented here and the qRT-PCR data reported by Seaton and coworkers (2011) arises from the fact that the earlier study utilized PCR primers that amplified the 3′ portion of comX. In fact, the two results are consistent when the Seaton et al.’s study is compared with results obtained by RNA-Seq and qRT-PCR with the 3′ primers used here.

Table 3.

Counted number of reads mapped to the comX coding region between the ΔrcrR-P and ΔrcrR-NP strains via RNA Sequencing analysis.

Sample	5′ region of comX chromosomal position 1 872 352–503 RPKM normalized	3′ region of comX chromosomal position 1 872 020–351 RPKM normalized	Fold change (5′ counts/3′ counts)	P valuea	Total comX chromosomal position 1 872 020–503 RPKM normalized
UA159 Replicate #1	71	55			52
UA159 Replicate #2	69	51			50
Average	70	53	1.321	3.72E-02	51
ΔrcrR-P Replicate #1	1035	1585			1267
ΔrcrR-P Replicate #2	1033	1275			1014
ΔrcrR-P Replicate #3	1133	1346			1078
Average	1067	1402	0.761	5.72E-02	1120
ΔrcrR-NP Replicate #1	26	2663			1831
ΔrcrR-NP Replicate #2	28	2866			1971
ΔrcrR-NP Replicate #3	33	2897			1992
Average	29	2809	0.010	6.90E-04	1931

Counts are normalized by RPKM (reads per kilobase per million mapped reads).

^aP value determined by Student’s t-test between 5′ and 3′ RPKM normalized counts.

Fig. 1. qRT-PCR of the comX coding region using 5′ and 3′ primer sets.

A. Diagram of 5′ and 3′ primer sets by location within the comX coding region (SMU.1997) and relation to RNA-Seq reads mapped in the ΔrcrR-NP and ΔrcrR-P strains.

B. qRT-PCR results of the UA159, ΔrcrR-P and ΔrcrR-NP strains using 5′ (white bars) and 3′ (black bars) comX primer sets. Strains were grown in BHI medium to OD₆₀₀ = 0.5 before cells were harvested and RNA extracted. Data represent averages of three biological replicates measured as triplicate samples. Differences in copies per microgram of RNA in all strains were evaluated for statistical significance between the 5′ and 3′ primer sets by Student’s t-test. **P value 0.009.

ComX production is reduced in the ΔrcrR-NP strain

It would be predicted that because of the very low levels of expression in the 5′ region of the comX gene, production of a functional ComX sigma factor might be reduced in the ΔrcrR-NP mutant. To test this, an antiserum against a full-length, recombinant S. mutans ComX protein purified from Escherichia coli was generated. For detection of ComX by immunoblotting, strains UA159, ΔrcrR-P and ΔrcrR-NP were grown in brain heart infusion (BHI) medium to an OD₆₀₀ of 0.2, at which point synthetic competence-stimulating peptide (sCSP), corresponding to the mature 21-mer, was added to a subset of the cultures to a final concentration of 1 μM. The cells were then collected by centrifugation after further incubation for 1 h. Clarified whole cell lysates were prepared and 10 μg of protein from each lysate was separated by Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). ComX could only be detected in the WT strain when cells were treated with sCSP, as previously noted (Tian et al., 2013). No ComX could be detected in a ΔcomX mutant, regardless of whether cells were treated with sCSP (Fig. 2). Consistent with the RNA-Seq and qRT-PCR data, high levels of ComX protein could be detected in the ΔrcrR-P strain, both in the presence and absence of exogenous sCSP. In contrast, no ComX was visualized in the ΔrcrR-NP strain with or without sCSP, similar to the ΔcomX mutant. Thus, ComX expression is markedly reduced in the ΔrcrR-NP strain, likely because of the absence of expression of the 5′ region of comX. These findings are also sufficient to explain the poor expression of comYA in the non-polar mutant and the high levels of comYA mRNA measured in the polar rcrR strain (Seaton et al., 2011).

Fig. 2.

Immunoblotting of ComX in the ΔrcrR-P and ΔrcrR-NP strains. Strains UA159, ΔcomX, ΔrcrR-P and ΔrcrR-NP were grown in BHI medium to an OD₆₀₀ = 0.5 before cells were harvested and whole cell lysates collected. One micromolar sCSP was added at OD₆₀₀ = 0.2 to a subset of the growing cultures. Ten micrograms of protein was loaded and separated on a 12.5% SDS-PAGE gel before transfer onto a polyvinylidene difluoride (PDVF) membrane. ComX was detected by a 1:1000 dilution of primary antisera against full-length recombinant ComX from Streptococcus mutans. Molecular mass standards (in kDa) are shown to the left. The calculated molecular mass of ComX is 19 kDa.

Identification of two distinct comX transcripts

The data detailed above provide evidence that two transcripts may be produced from the comX coding region: a full-length comX in the ΔrcrR-P and WT strains, and a shorter transcript in the ΔrcrR-NP strain that originates internal to, and consists of only the 3′ region of, comX. To further validate the existence of these two transcripts, a northern blot was performed using RNA isolated from the WT (± 1 μM sCSP), ΔrcrR-P and ΔrcrR-NP strains. To detect the comX transcript, a 352 bp probe that covers both the 5′ and 3′ regions of the comX coding region was designed and randomly labeled with biotin. After separation of 5 μg of RNA and hybridization of the probe, two distinct comX transcripts could be identified in the ΔrcrR-P and ΔrcrR-NP strains (Fig. 3). A transcript of 0.5 kb, consistent with the 483 bp comX, was detectable in sCSP-treated cells of the WT strain. In the hyper-transformable ΔrcrR-P mutant, the 0.5 kb transcript was highly expressed, even in the absence of sCSP. This result parallels the RNA-Seq and western blot data, supporting that ComX production in this strain is robust. Interestingly, in the ΔrcrR-NP strain, a strong signal at 0.35 kb was evident. The size of this transcript correlates with what would be expected for the non-polar rcrR mutant based on the RNA-Seq and qRT-PCR data. It is important to note that the shorter 0.35 kb transcript could not be detected in the WT or the rcrR polar mutant, nor could the full-length comX transcript be detected in the non-polar rcrR mutant. It is also evident that transcripts of a larger size than that predicted for comX can be observed in the ΔrcrR-P and ΔrcrR-NP strains, both of which highly express mRNA that hybridizes to comX. We attribute these transcripts to transcriptional read-through of the normal comX termination sequence because of strong activation of the 0.35 and 0.5 kb transcripts in the mutant strains, and to a lesser extent in the WT strain treated with sCSP. It is important to note that a 146 bp intergenic region separates comX from ipk (SMU.1996), which is 849 bp. The 0.8 and 0.9 kbp transcripts observed would not correspond to a full-length comX-ipk co-transcript. Thus, it is our belief these transcripts most likely terminate within the ipk gene. Collectively, the results clearly demonstrate that two transcripts of different lengths are expressed from the comX coding region in the ΔrcrR-P and ΔrcrR-NP strains.

Fig. 3.

Northern blot of comX transcripts. Strains UA159, ΔrcrR-P (ΔP) and ΔrcrR-NP (ΔNP) were grown in BHI medium to an OD₆₀₀ = 0.2, at which point 1 μM sCSP was added, if desired. Cells were grown for an additional hour after sCSP addition before harvest by centrifugation. Five microgram of purified RNA was separated on a 1% agarose-formaldehyde gel and transferred to a BrightStar-Plus membrane. A biotin-labeled, 352 bp comX-specific probe was utilized for hybridization and detection. Transcript sizes (right) were estimated by comparison with an RNA Millennium Marker. An image of 23S and 16S rRNA (bottom) from each sample serves as the loading control.

Both comX transcripts are expressed from the comX promoter

We hypothesized that one mechanism for loss of the 5′ region in the comX transcript was due to silencing of the PcomX promoter in the ΔrcrR-NP strain and expression of a second, intragenic comX promoter. We monitored PcomX activity with a PcomX::gfp reporter plasmid (Son et al., 2012) in either WT, ΔrcrR-P and ΔrcrR-NP backgrounds. While observing the strains grown in the chemically defined medium FMC, PcomX activity was detected in the ΔrcrR-P genetic background but not in ΔrcrR-NP (Supporting Information Fig. S2). Therefore, PcomX could, in fact, be silenced in the ΔrcrR-NP mutant. However, when sXIP was added, PcomX activity was robust in all three strains. Notably, the WT and ΔrcrR-NP strains showed no differences in the concentration of sXIP needed to activate the promoter; as little as 100 nM sXIP was sufficient (data not shown). A northern blot with cells grown in FMC with and without 2 μM sXIP was performed to determine if addition of sXIP could restore expression of the full-length comX transcript. We found that the 0.35 kb comX transcript in the ΔrcrR-NP strain was still present, even with addition of a high concentration of sXIP (Fig. 4A). Additionally, the non-polar mutant still could not be transformed, even when treated with 2 μM sXIP (Ahn et al., 2014). We also probed for comX mRNA in the ΔrcrR-NP strain carrying a deletion of the comS gene. In this strain, expression of the 0.35 kb comX transcript was lost and was not restored by addition of sXIP, and a very weak expression of the full-length 0.5 kb comX transcript was instead detected (Fig. 4A).

Fig. 4. comX transcripts evaluated by northern blot.

A. Strains UA159, ΔrcrR-P (ΔP), ΔrcrR-NP (ΔNP) and ΔcomSΔrcrR-NP (ΔcomS/ΔNP) were grown in FMC medium to an OD₆₀₀ = 0.2, at which point 1% dimethyl sulfoxide (vehicle) or 2 μM sXIP was added. Cells were grown for an additional hour after addition of sXIP, then cells were harvested by centrifugation and RNA prepared. Asterisk (*) in lane ΔcomSΔNP + sXIP denotes the predicted point to which a full-length comX mRNA would migrate.

B. Strains UA159 (WT), ΔrcrR-P (ΔP), ΔrcrR-NP (ΔNP), ΔcomS/ΔrcrR-NP (ΔS), ΔcomRΔrcrR-NP (ΔR) and Δ PcomXΔrcrR-NP (ΔPx) were grown in FMC medium to an OD₆₀₀ = 0.5, at which point cells were harvested. Five microgram of purified RNA was separated on a 1% agarose-formaldehyde gel and transferred to a BrightStar-Plus membrane. A biotin-labeled, 352 bp comX-specific probe was utilized for hybridization and detection. Transcript sizes (right) were estimated by comparison with an RNA Millennium Marker. An image of 23S and 16S rRNA (bottom) from each sample serves as the loading control.

To further determine whether expression of the internal 0.35 kb transcript required PcomX activation, we conducted northern blots using total RNA from a comS mutant, a comR mutant and from a mutant in which the -10 element of the comX promoter was mutated from TATAAT to TCCCCT, all in the non-polar genetic background. We found that neither the 0.5 kb nor the 0.35 kb comX transcripts were present in these three strains (Fig. 4B). We also cloned various fragments upstream of the 0.35 kb transcript and internal to the comX structural gene, i.e. lacking PcomX, behind a LacZ reporter gene, but could not find any evidence that the 0.35 kb transcript was driven from a promoter internal to comX (data not shown). Collectively, then, these results show that expression of the 0.35 kb comX transcript in the ΔrcrR-NP strain is dependent on a functional ComR, ComS and PcomX promoter, and is not expressed from a second promoter internal to comX.

Mapping of the 5′ end of the internal transcript

To further characterize the smaller transcript that is expressed from the 3′ coding region of comX in the ΔrcrR-NP strain, the 5′ Rapid Amplification of cDNA Ends (5′-RACE) technique was used to locate the 5′ termini of the mRNAs expressed in the ΔrcrR-P and ΔrcrR-NP strains. cDNA was generated from three biological replicates, grown in BHI medium and harvested at an OD₆₀₀ = 0.5, using a comX gene-specific primer located at the 3′ end of the comX coding region, such that the primer would recognize regions present in both transcripts (Supporting Information Table S1). Following homopolymeric tailing, several additional rounds of PCR were performed to obtain products suitable for DNA sequence analysis (Supporting Information Fig. S3A). A potential transcriptional start site (TSS) was mapped 23 bp upstream of the comX start codon in the ΔrcrR-P mutant, coinciding with a previously published start site for the comX transcript (Mashburn-Warren et al., 2010). In contrast, we were only able to map a 5′ terminus that began at position +140 of the comX structural gene in the non-polar strain (Fig. 5A, Supporting Information Fig. S3B). Thus, the transcript expressed in ΔrcrR-NP is 163 bp shorter than the full-length comX mRNA produced in the ΔrcrR-P strain, with its 5′ terminus positioned internal to the comX coding region. This finding agrees with the size of the 0.35 kb transcript observed by northern blotting, and correlates wholly with the RNA-Seq data.

Fig. 5. Mapping of 5′ termini within the comX coding region.

A. 5′ termini of the comX coding region transcripts were mapped using 5′ Rapid Amplification of cDNA Ends (5′-RACE) and are bolded, underlined and labeled (5′ Terminus) within the DNA sequence. Elements regulating comX transcription are labeled by underlining (ComRS box) and bolding (-35, -10). Shine-Dalgarno sequences are underlined (RBS) and start (ATG) and stop (TGA; TAG) codons are bolded for the ComX and XrpA protein coding sequences. Sequence shown is chromosomal position 1 872 020–1 872 623 of the Streptococcus mutans UA159 genome.

B. Illustration of the organization of xrpA within the comX coding region. comX is 483 bp in length with the 210 bp xrpA ORF beginning at the 161st bp of comX.

Identification of an intragenic open reading frame within the comX coding region

When the transcript expressed in the ΔrcrR-NP strain, designated henceforth as comX regulatory peptide A (xrpA), was examined in more detail, a start codon for a 69-aa open reading frame (ORF) could be identified beginning at base 161 of the comX coding region, in a reading frame that was different (+1 nt) than comX (Fig. 5, Supporting Information Fig. S4A). A potential Shine-Dalgarno sequence (AGAAAGA) was correctly positioned upstream of the xrpA ORF. BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) results with the predicted XrpA protein sequence (Supporting Information Fig. S4B) showed no hits to known proteins (blastp suite, taxid: 1239) and an InterPro (http://www.ebi.ac.uk/interpro/) search did not find any known protein domains. XrpA was predicted to contain a transmembrane domain starting at amino acid 45 and continuing until aa 67, with residues 1 to 44 predicted to reside in the cytoplasm (Supporting Information Fig. S4C). The coding sequence of comX orthologs in multiple other streptococci, including S. pneumoniae, Streptococcus thermophilus and Streptococcus pyogenes, was examined, but an intragenic ORF similar to XrpA could not be identified (Supporting Information Fig. S5). We also examined the comX coding sequence of 57 isolates of S. mutans for sequence conservation and the presence of the xrpA ORF. Only three isolates carried single-base changes that would result in a different ComX sequence than for strain UA159, with all three mutations located in the 3′ region of the gene. However, there were no sequence polymorphisms in the xrpA coding sequence, so the XrpA protein was 100% identical in all 57 isolates. Therefore, ComX and XrpA are very highly conserved in S. mutans.

Mutation of xrpA restores transformation in the ΔrcrR-NP strain

To validate that xrpA encodes a protein product, we created two different mutants: mutation of the XrpA ATG start codon at position 162 of comX by changing the T to a C (comX::T162C) and insertion of a premature stop codon at the 17th aa position of XrpA by changing a T to an A (comX::T210A) (Fig. 6A). Importantly, neither of these mutations changed the ComX protein sequence; an aspartate residue in comX::T162C and an isoleucine residue in comX::T210A remained unchanged. Introduction of both of these mutations in the UA159 chromosome did not elicit any apparent changes in the growth of the strains in rich medium (Supporting Information Fig. S6).

Fig. 6. Mutation of xrpA in the ΔrcrR-NP background.

A. Sequence of the comX::T162C and comX::T210A mutations. A thymine is mutated to cytosine at the 162nd bp of comX (bolded), changing the xrpA start codon from ATG to ACG (underlined) to prevent translation initiation. In comX::T210A, a thymine at the 210th bp of comX is mutated to adenosine to create a premature stop codon. In both mutations, the ComX protein coding sequence is unchanged.

B. Transformation efficiency between strains UA159, ΔrcrR-NP, comX::T162C/ΔrcrR-NP and comX::T210A/ΔrcrR-NP in BHI without sCSP (white bars) or BHI with addition of 1 μM sCSP (black bars). Transformation efficiency is determined by dividing the number of transformants by the total number of viable bacteria and is expressed as a percentage. Differences in transformation efficiency between strains were evaluated for statistical significance by Student’s t-test. *P value < 0.05.

C. Growth curves of strains ΔrcrR-NP (circles), comX::T162C/ΔrcrR-NP (triangles) and comX::T210A/ΔrcrR-NP (squares) in BHI with addition of 1 μM sCSP. Growth was monitored with a Bioscreen C lab system every 30 min for 24 h. Growth is displayed as an average of three biological replicates monitored in three individual wells.

D. Immunoblot against ComX in the UA159, ΔrcrR-NP, comX::T162C/ΔrcrR-NP and comX::T210A/ΔrcrR-NP strains. One micromolar sCSP was added at OD₆₀₀ = 0.2 to a subset of the cultures, and cells were harvested at OD₆₀₀ = 0.5. Ten microgram of protein from extracted whole cell lysates was loaded and separated on a 12.5% SDS-PAGE gel before transfer onto a polyvinylidene difluoride (PDVF) membrane. ComX was detected by a 1:1000 dilution of primary antiserum against full-length recombinant ComX from Streptococcus mutans. Molecular mass standards (in kDa) are shown to the left. The calculated molecular mass of ComX is 19 kDa.

While the ΔrcrR-NP strain is not transformable under any of a variety of conditions that have been previously tested, the mutations introduced into xrpA in the ΔrcrR-NP genetic background resulted in strains that displayed transformation efficiencies, in the absence of exogenous sCSP, that were more than 100-fold higher than that of WT (Fig. 6B), i.e. more similar to the rcrR polar mutant. We have also observed that the non-transformable ΔrcrR-NP mutant is much more resistant to apparent growth inhibition by sCSP than the WT strain (Ahn et al., 2014). Interestingly, when we compared the sensitivity with sCSP of the ΔxrpA mutants in the ΔrcrR-NP background, significant increases in the apparent doubling times and decreases in final yields were noted in the strains lacking XrpA, compared with the non-polar rcrR mutant (Fig. 6C), indicative of enhanced sensitivity to sCSP. It was also determined that ComX expression (Fig. 6D) and the production of full-length comX mRNA (Supporting Information Fig. S7) were restored in both the comX::T162C/ΔrcrR-NP and comX::T210A/ΔrcrR-NP strains, consistent with the transformation phenotypes. Likewise, sensitivity to sXIP and transformability in the chemically defined medium FMC were also restored in the comX::T162C/ΔrcrR-NP strain (Supporting Information Fig. S8). Collectively, these data support that xrpA encodes a functional protein that is able to interfere with the development of competence, perhaps by altering comX mRNA processing or the production and/or stability of ComX.

Regulatory elements of xrpA can drive translation in vivo

To observe xrpA expression in vivo, a translational fusion of XrpA was constructed on a GFP reporter plasmid (Son et al., 2012), such that the first three codons of xrpA were in frame with the GFP protein and translation would be driven by the cognate ribosome-binding site (RBS) of xrpA. Two different constructs were created: XrpA::gfp containing the 5′ region of comX adjacent to xrpA (comX bp 1–160) but not the comX promoter, and PcomXXrpA::gfp that contained both the 5′ region of comX and the comX promoter (Fig. 7A). Consistent with the results that showed that comX promoter activity was required for xrpA expression (Fig. 4), no activity could be observed in the XrpA::gfp construct in either the WT or non-polar rcrR mutant genetic backgrounds. However, GFP expression could be detected in the PcomXXrpA::gfp construct in either the WT background with addition of sXIP, or in the ΔrcrR-NP background (Fig. 7B). In the WT background, xrpA-dependent GFP expression was marginal in FMC alone, but was robust with addition of sXIP. In contrast, GFP expression in the non-polar mutant in FMC medium alone, or with exogenous sXIP, could be detected, as would be expected based on RNA-Seq and northern blot data. We could also monitor GFP in these strains through immunoblotting, as the GFP contained a C-terminal 6X His tag, and the results were consistent with the fluorescence measurements (Fig. 7B). It is important to note that expression observed in the PcomXXrpA::gfp strain showed different characteristics from what was observed in the PcomX::gfp/ΔrcrR-NP reporter mutant (Supporting Information Fig. S2). Specifi-cally, the PcomX::gfp fusion in the/ΔrcrR-NP genetic background did not show fluorescence in chemically defined medium alone, whereas fusion to xrpA resulted is readily detectable GFP fluorescence in the same genetic background. This finding is consistent with the observed lack of comX mRNA and strong xrpA expression in the non-polar rcrR mutant. Because PcomX is necessary for xrpA expression, we posit that post-transcriptional regulation of comX, probably involving cleavage within, and degradation of the 5′ portion of the comX mRNA must be an important control point for ComX and XrpA production.

Fig. 7. xrpA mRNA and start codon drive GFP production in vivo.

A. Diagram of translational fusion constructs used to drive GFP production in vivo. XrpA::gfp contains the 5′ comX coding sequence (bases 1–169) fused to GFP. PcomXXrpA::gfp is the same construct as XrpA::gfp, but the 5′ end is extended to include the comX promoter.

B. Relative expression of the XrpA::gfp and PcomXXrpA::gfp strains in either the UA159 or ΔrcrR-NP genetic backgrounds, grown in FMC with addition of either 1% dimethyl sulfoxide (vehicle) or 2 μM sXIP. The strains indicated in the figure, from left to right, are UA159, XrpA::gfp/UA159, PcomXXrpA::gfp/UA159, XrpA::gfp/ΔrcrR-NP, and PcomXXrpA::gfp/ΔrcrR-NP. Fluorescence of GFP was measured using a Synergy 2 multimode microplate reader (filters: excitation 485/20 nm, emission 528/20 nm). Differences in relative expression between strains were evaluated for statistical significance by Student’s t-test. **P value < 0.01. Relative expression (RFU/OD₆₀₀) is calculated as fluorescent units minus background, divided by OD₆₀₀ (~OD₆₀₀ 0.5). An immunoblot of detecting the expressed GFP, which contains a 6-His tag, using an anti 6-His antibody (0.5 ng μl⁻¹ Anti-6X His tag^®) against extracted whole cell lysates (5 μg of protein loaded on 12.5% SDS-PAGE gel) is shown below for each referenced strain. The calculated molecular mass of XrpA-GFP fusion is 28.5 kDa.

XrpA is an antagonist of ComX

To begin to understand how XrpAmay exert its influence on competence development in S. mutans, we compared the transformation phenotypes of an xrpA mutant (comX::T162C) with the WT strain. In BHI medium, transformation efficiency in the strain lacking XrpA increased almost threefold in the absence of sCSP (2.0 × 10⁻⁴ to 5.5 × 10⁻⁴) and did not change when the transformation assays were performed with sCSP (6.2 × 10⁻₂ to 4.8 × 10⁻²) (Fig. 8A), compared with the WT strain. However, a more significant increase in transformation efficiency was seen in FMC medium, with the XrpA-deficient stain showing a 6.7-fold increase in transformation efficiency over the WT (3.2 × 10⁻⁹ to 4.8 × 10⁻¹⁰). Similar to the effects of sCSP in complex medium, only a twofold increase in transformation efficiency was observed with addition of sXIP (2.0 × 10⁻⁴ to 1.0 × 10⁻⁴).

Fig. 8. XrpA can decrease competence and ComX levels.

A. Comparison of transformation efficiency between UA159 and comX::T162C (ΔxrpA). Strains were grown in either BHI with or without 1 μM sCSP or in FMC with either 1% dimethyl sulfoxide (vehicle) or 2 μM sXIP. Differences in transformation efficiency between strains were evaluated for statistical significance in all medium conditions by Student’s t-test. *P value 0.015.

B. Comparison of transformation efficiency between pIB184 (vector only) and pIB184xrpA (xrpA overexpressing), grown in either BHI with or without 1 μM sCSP or in FMC with either 1% dimethyl sulfoxide (vehicle) or 2 μM sXIP. Differences in transformation efficiency between strains were evaluated for statistical significance in all medium conditions by Student’s t-test. *P value 0.035 and 0.014 respectively.

C. Immunoblot against ComX in either pIB184 (V) or pIB184xrpA (XrpA). One micromolar sCSP or 2 μM sXIP was added at OD₆₀₀ = 0.2, and cells were harvested at OD₆₀₀ = 0.5. Ten microgram of protein from clarified whole cell lysates was loaded and separated on a 12.5% SDS-PAGE gel before transfer onto a polyvinylidene difluoride (PDVF) membrane. ComX was detected by a 1:1000 dilution of primary antisera against full-length recombinant ComX from Streptococcus mutans. Molecular mass standards (in kDa) are shown to the left. The calculated molecular mass of ComX is 19 kDa.

To test whether increasing the expression of xrpA could negatively impact the ability of the organism to become competent, the xrpA coding region was cloned into the streptococcal shuttle vector pIB184, such that xrpA transcription was under the control of the constitutively strong P₂₃ promoter (Biswas et al., 2008; Ahn et al., 2014; Guo et al., 2014) and translation was dependent on the cognate xrpA RBS. To confirm overexpression of xrpA, xrpA mRNA levels were compared in the overexpression strain (pIB184xrpA) and the vector-only control strain (pIB184) by qRT-PCR. When expressed from the P₂₃ promoter, xrpA mRNA levels were nearly 700-fold higher in the overex-pression strain compared with the vector-only control (Supporting Information Fig. S9). Also, it is important to note that the expression level of xrpA in pIB184xrpA is a log higher than in the ΔrcrR-NP strain (1.41 × 10⁷ to 1.29 × 10⁶ copies × μg RNA⁻¹).

A 6.8-fold decrease in transformation efficiency was observed in the pIB184xrpA strain compared with the vector-only control strain in the absence of exogenous sCSP (Fig. 8B). When sCSP was added, a threefold lower transformation efficiency was noted in pIB184xrpA compared with pIB184. Similar to the ΔxrpA strain above, stronger inhibition of transformation was seen in FMC medium in cells treated with sXIP, as the xrpA-overexpressed strain had a 33-fold decrease in transformation efficiency compared with the vector-only control. Additionally, overexpression of xrpA resulted in decreased levels of ComX protein in cells grown in either complex or defined medium and treated with sCSP or sXIP respectively (Fig. 8C). These data show that xrpA decreases transformation efficiency by reducing the levels of ComX. Of note, simply overexpressing xrpA in a strain that had an intact comX and xrpA gene did not eradicate transformability, so the absence of ComX and other changes elicited by the loss of RcrR, overproduction of RcrPQ and the production of the rcrQ peptides contribute significantly to the non-transformable phenotype of ΔrcrR-NP.

Overexpression of xrpA affects oxidative stress tolerance

By examining the XrpA protein coding sequence, we noticed that XrpA contains six cysteine residues spaced over the entire length of the protein (Supporting Information Fig. S4B), raising the possibility that the function of XrpA may be influenced by redox. In fact, several genes surrounding the rcrRPQ operon (SMU.924-SMU.928) are differentially expressed under aerobic conditions, including the bacteriocin immunity protein cipI, tpx encoding a thiol peroxidase and the small (p)ppGpp synthetase relP (Ahn et al., 2007). To see if XrpA could impact oxidative stress tolerance, growth of pIB184, pIB184comX and pIB184xrpA strains was monitored in aerobic conditions (Fig. 9A). Interestingly, the xrpA-overexpressing strain displayed significantly slower growth than both the WT and comX-overexpressing strains (doubling time = 124 ± 12 versus 58 ± 2 and 68 ± 2 min, respectively) and a lower final yield (OD₆₀₀ = 0.35 ± 0.0 versus 0.49 ± 0.1 and 0.45 ± 0.0, respectively) when cells were grown in aerobic conditions. The enhanced sensitivity of the xrpA-overexpressing strain in the presence of air was also observed when cells were cultured on agar plates for 5 days (Supporting Information Fig. S10). The observable growth defect of the pIB184xrpA strain under aerobic conditions became even more apparent upon addition of sCSP (Fig. 9B). In this condition, the xrpA-overexpressing strain failed to grow at all, compared with the WT and comX-overexpressing strains, which displayed minimal effects on growth in the presence of sCSP. Growth of these strains was also monitored under acidic conditions (pH 5.5), under nutritional stress (removal of isoleucine from the chemically defined medium FMC) and during activation of the stringent response by addition of mupi-rocin to the growth medium. Both the pIB184comX and pIB184xrpA strains showed no significant changes in growth compared with the pIB184 control.

Fig. 9. Growth of xrpA-overexpressing strain in aerobic conditions.

A. Growth curves of the pIB184 (circles), pIB184comX (squares) and pIB184xrpA (triangles) strains in BHI without a 50 μl sterile mineral oil overlay to examine growth inhibitory effects of air.

B. Growth curves of the pIB184 (circles), pIB184comX (squares) and pIB184xrpA (triangles) strains in BHI containing 1 μM sCSP. Strains were grown without a mineral oil overlay. Growth was monitored using a Bioscreen C lab system every 30 min for 24 h. Growth is displayed as an average of three biological replicates monitored in three individual wells.

Discussion

Stress tolerance and the development of genetic competence in the dental pathogen S. mutans are intertwined to a remarkable degree (Senadheera et al., 2005; Banu et al., 2010; Okinaga et al., 2010; Kajfasz et al., 2011; Kim et al., 2013; Guo et al., 2014). One example of this integration is the heat shock-inducible, surface-associated serine protease HtrA that is induced by addition of CSP, while deletion of htrA results in decreased transformation efficiency (Ahn et al., 2005). A second example is the CiaRH TCS, which is required for acid and oxidative stress tolerance, but also integrates environmental signals to coordinate induction of multiple com genes (Ahn et al., 2006). Despite clear connections between competence and stress tolerance in S. mutans, there has been very little progress to define how the regulatory systems that control these phenomena cross-regulate expression of competence and stress-related genes. Multiple gene products encoded by the rcrRPQ operon of S. mutans have profound effects on acid and oxidative stress tolerance, (p)ppGpp metabolism and genetic competence, and thus these products may constitute a central regulatory system that allows cells to integrate physiologic status with competence development (Seaton et al., 2011). One key issue that needs to be addressed is to understand the molecular basis for how rcrRPQ-encoded proteins and peptides accomplish the integration of environmental signals with competence. Although various mutations in the rcrRPQ genes can influence competence and growth in the presence of stressors, the profoundly different phenotypes of the polar (ΔrcrR-P) and non-polar (ΔrcrR-NP) rcrR mutants provided the most logical point to initiate studies on how rcr gene products govern the decisions for cells to commit to the competent state. A major breakthrough in understanding how the polar and non-polar rcrR mutants differentially impact competence was made when RNA-Seq was employed to gain a broader view of how the transcriptome of S. mutans changes in the two mutant backgrounds. It was with this information that we discovered that transcription of comX was affected in a very unusual way in these two strains.

First, it was surprising that the expression of the 5′ portion of the comX mRNA was essentially absent in the non-polar rcrR mutant, with RNA-Seq read levels in the rcrR-NP strain being even lower than in the WT organism growing in the absence of sCSP. The discovery of differential expression in the 5′ region of comX between the two strains answers one key question as to how the loss of RcrR and resultant overproduction of RcrPQ and two peptide effector molecules (Ahn et al., 2014) in the non-polar mutant translate into an inability of cells to become non-transformable, regardless of whether exogenous sCSP or sXIP is provided. The answer is that an intact comX gene is not expressed (Figs 1 and 3), which leads to lack of production of detectable levels of the alternative sigma factor ComX (Fig. 2). Because a comX mutant has been shown to be non-transformable under all conditions tested and ComX is required for sensitivity to sCSP, the loss of ComX would be sufficient to account for the phenotypes of the rcrR non-polar mutant. Using qRT-PCR, we have shown that mutations in rcrR do not cause downregu-lation of the genes for CSP and its cognate TCS (comCDE) or the ComR-XIP (comRS) pathway for comX induction, nor do they influence the ability of CSP to activate bacte-riocin gene expression via ComDE (Ahn et al., 2014). However, the rcrR mutations studied here dominantly influ-ence the expression of comX, with constitutive activation of comX in the polar mutant and no full-length comX transcript present in the non-polar mutant.

There are two mechanisms that are likely to account for the loss of the comX 5′ region and subsequent high level of expression of the xrpA transcript. The first mechanism relies on the presence of a previously unrecognized promoter within the comX coding sequence that expresses a de novo xrpA transcript. This proposed intragenic promoter would be ‘turned on’ in the non-polar mutant, and the comX promoter would concurrently be ‘turned off’ or silenced. Silencing of the comX promoter could result from failure of ComRS to activate the comX promoter in the rcrR non-polar mutant. In fact, we were not able to detect PcomX activity in the ΔrcrR-NP strain in FMC medium (Supporting Information Fig. S2). However, PcomX expression could be detected after addition of low concentrations of sXIP (100 nM). Therefore, ComR is functional and able to interact with XIP in the ΔrcrR-NP strain; so if a block in activation of comX occurs, it is probably associated with maturation of ComS to XIP, or possibly some yet to be defined function of ComS. We are currently exploring these possibilities. A second, and we believe more likely, mechanism is that the comX promoter transcribes a single 0.5 kb transcript that is processed in the non-polar strain to the observed 0.35 kb transcript (Fig. 3), with rapid degradation of the 5′ portion of the comX mRNA. Indeed, we have been able to show that deletion of either comR or comS, or mutation of the comX promoter, leads to disappearance of all transcripts that hybridize to a comX probe, verifying that PcomX activity is essential for xrpA mRNA production. Also, we were not able to detect promoter activity from the 5′ portion of the comX coding sequence (bp 1–160) using LacZ transcriptional fusions, providing further evidence that there is not an intragenic comX promoter upstream of xrpA. Future studies are aimed at reconciling the molecular basis for the dependence of xrpA transcription on PcomX with the concomitant disappearance of the 5′ portion of the comX mRNA.

The second novel observation is the discovery of a functional and apparently unique 69-aa protein arising from the internal xrpA transcript. While there is some precedent for this gene-within-a-gene organization in bacteriophage, such as Rz/Rz1 of bacteriophage lambda (Berry et al., 2008), we are unaware of a similar example within a bacterial genome and have found no evidence for such an arrangement in the comX genes of related bacteria (Supporting Information Fig. S5). To our knowledge, we believe these are the only genes to date described in bacteria that share the same DNA in different reading frames and are in turn associated with the same phenotype. The evidence supporting that xrpA does, in fact, encode a protein product includes the behaviors arising from two different base changes within comX that preserve ComX protein sequence while preventing the production of XrpA: inactivation of the xrpA start codon (comX::T162C) and an insertion of a stop codon leading to premature termination of xrpA translation (comX::T210A) (Fig. 6). While both mutations in a WT background increase basal transformation efficiency by several fold, the changes in phenotypes are very profound in the presence of the non-polar rcrR mutation, where ComX expression and transformation efficiency are not only restored, but reach levels seen in the hyper-transformable rcrR polar strain. Similarly, overexpression of xrpA leads to substantial decreases in the ability of cells to be transformed. In all cases, the changes in transformation efficiency associated with either the mutation of xrpA or the overexpression of xrpA were strongly correlated with the levels of ComX protein (Fig. 8C). We were also able to demonstrate that the xrpA RBS and start codon can drive the production of GFP (Fig. 7). In addition, we added a FLAG epitope to the N- or C-terminus of XrpA in the ΔrcrR-NP background and tried to detect the protein using anti-FLAG antibodies using various methods of cell lysate preparation. Although we were not able to detect the FLAG-tagged XrpA, it appears that the FLAG epitope rendered the XrpA protein nonfunctional. Specifically, FLAG tagging of XrpA in the rcrR non-polar genetic background yielded cells with transformation and sCSP/sXIP sensitivity profiles that were essentially identical to cells carrying mutations in the coding sequence for XrpA. In fact, we found that xrpA mRNA was not observed in these FLAG-tagged strains by northern blotting (Supporting Information Fig. S11) and that the ComX levels resembled those found in the rcrR-P hyper-transformable strain, again similar to the xrpA start codon mutant. The same results were obtained when N- or C-terminal 6X His tags were added to XrpA (data not shown). The behavior of epitope-tagged strains provides further evidence that a functional XrpA protein is indeed an effector molecule.

All of the data available to date support that xrpA encodes a small protein that exerts its effects on competence by either altering PcomX activity, affecting the processing of the comX mRNA and/or promoting the degradation of ComX. Nevertheless, we cannot exclude that the 3′ transcript expressed in the ΔrcrR-NP strain may also function as a small RNA (sRNA). The transcript fits the profile of an sRNA from the perspectives that it is between 50 and 500 nt and is believed to be expressed under particular environmental conditions (Waters and Storz, 2009). A limited number of sRNAs have been identified and studied in streptococci (Le Rhun and Charpentier, 2012). Most of these sRNAs function as antisense RNAs through base-pairing with their target mRNAs to influence mRNA stability or block translation by preventing ribosome binding (Gottesman and Storz, 2011). If the xrpA transcript were to act as an sRNA, it would most certainly be transacting, as no known mRNA is transcribed from the anti-sense direction of the comX coding region. Recently, strand-specific RNA-Seq showed there were virtually no antisense transcripts generated within the comX gene (Zeng and Burne, unpublished data). Despite the potential for the xrpA mRNA to play some role in competence development, we are proceeding with a working model that treats xrpA as a coding mRNA that produces the XrpA effector protein.

Recently, two peptides were found in the 3′ region of the rcrQ gene that function to antagonize competence, possibly by increasing ComX susceptibility to MecA/Clp-dependent degradation (Ahn et al., 2014). There are several parallels between these peptides and XrpA, as both are highly expressed and display the strongest phenotypes in the non-polar rcrR mutant. Similar to XrpA, homologs of these peptides could not be found in other streptococci, even though the rcrRPQ operon appears widely distributed. Thus, a model was developed in which these peptides regulate ComX stability and thus have similarities with anti-sigma factors in other organisms that modulate ComX activity (Liu and Zuber, 1998; Luo et al., 2004; Sung and Morrison, 2005; Ahn et al., 2014). These peptides, and the XrpA protein described herein, may function to increase the efficiency of MecA-ComX delivery to the ClpC protease (Wahl et al., 2014). While the rcrR mutants have been invaluable for the discovery of these novel competence antagonists, it is important to consider how they may function in WT cells. For example, transcription of comX mRNA in a cell that is growing exponentially or undergoing competence development likely leads to the production of both ComX and XrpA. Consequently, ComX:XrpA ratios may dictate whether cells respond to CSP or XIP by progressing to competence, ignoring the signal or undergoing programmed cell death (Perry et al., 2009; Wenderska et al., 2012). Because rcrRPQ transcription is responsive to (p)ppGpp and oxidative stress, a primary role for the RcrRPQ system and the recently discovered peptide effector molecules encoded at the end of rcrQ may be to allow cells to integrate physiologic status with the decision network that chooses (i) to ignore responses to competence signals (bistability, Son et al., 2012), (ii) to commit to competence or (iii) to progress toward cell death (Lemme et al., 2011; Wei and Håvarstein, 2012). Importantly, very small changes in rcrRPQ promoter activity can lead to dramatic changes in transformation efficiency (Seaton et al., 2014). We therefore propose that changes in redox, pH (Guo et al., 2014), (p)ppGpp levels or other inputs trigger changes in rcrRPQ expression that in turn lead to changes in the amount of comX transcription and/or in ComX stability, with a resultant imbalance in ComX:XrpA ratios. In this manner, cells, or subpopulations of cells, could integrate stress sensing with the decision to progress to competence or undergo lytic death.

One final intriguing result is that a strain that overex-presses xrpA becomes sensitive to growth in an aerobic environment (Fig. 9). The protein sequence of XrpA contains six cysteine residues that may participate in disulfide bond formation, impacting the folding and function of the protein as the reducing environment in the cytosol changes. Also, ComX of S. mutans has a single cysteine residue, so disulfide bond formation between ComX and XrpA cannot be excluded as a mechanism to integrate redox conditions with the decision to progress to competence. In this regard, it is noteworthy that genes surrounding the rcrRPQ operon, including tpx, cipI and the relPRS operon, were found to be upregulated during growth in aerobic conditions (Ahn et al., 2007), showing that the presence of oxygen has the potential be a signal that regulates the expression of genes in this region.

In summary, we have shown that the rcrRPQ operon of S. mutans modulates competence development through differential expression of comX and of a novel gene internal to comX, xrpA. Coupled with recent studies showing the presence of peptide effectors and the importance of proper expression levels of the RcrPQ efflux pumps (Seaton et al., 2011; 2014; Ahn et al., 2014), the results support that the regulation of competence by S. mutans is substantially more complex than previously appreciated. We are particularly intrigued by the intertwining of competence with (p)ppGpp metabolism, especially by RelP and RelA, and with oxidative stress. In early dental plaque, S. mutans is exposed to a higher oxygen and higher redox environment and must compete with a variety of commensals, many of which use hydrogen peroxide as a primary defense mechanism against S. mutans (Kreth et al., 2008). The ability of S. mutans to optimize growth and gene expression at a global level as dental plaque matures, and in certain cases as the environment becomes favorable to the development of dental caries, is critical for persistence and virulence of the organism. It appears that RcrRPQ serves a central function in helping the organisms to balance growth and survival decisions depending on the environmental conditions. While much remains to be learned about the interaction of factors such as XrpA and the signaling pathways for competence and stress tolerance, the discovery of this unusual pathway and its genetic organization has opened a new area for investigation. Should XrpA indeed prove unique to S. mutans, it may also serve as an attractive target for anti-caries therapeutics given its central role in homeostasis.

Experimental procedures

Culture conditions

Escherichia coli strain DH10B was grown in Luria broth that was supplemented with 300 μg ml⁻¹ erythromycin and 50μg ml⁻¹ of kanamycin or spectinomycin, when needed. S. mutans WT strain UA159 and its derivatives (Table 4) were grown in either BHI broth (Difco) or the chemically defined medium FMC (Terleckyj and Shockman, 1975) supplemented with 10 μg ml⁻¹ erythromycin and 1 mg ml⁻¹ of kanamycin or spectinomycin, when necessary. Unless noted otherwise, cultures were grown overnight in BHI medium with antibiotics when needed at 37°C in a 5% CO₂ aerobic atmosphere. To compare growth characteristics of strains, overnight cultures were diluted 1:50 and grown to mid-exponential phase (OD₆₀₀ = 0.5). Cultures were then re-diluted 1:100 into 300 μl of BHI medium in the absence or presence of 1 μM sCSP and added to multi-well plates. To reduce the growth inhibitory effects of air, each well was overlaid with 50 μl of sterile mineral oil. Wells were not covered in mineral oil when assessing growth in aerobic conditions. The optical density at 600 nm (OD₆₀₀) was monitored at 30 min intervals for 24 h using a Bioscreen C lab system (Helsinki, Finland) at 37°C, with shaking for 10 s before each reading. sCSP, corresponding to the 21-aa peptide, was synthesized by the Interdisciplinary Center for Biotechnology Research (ICBR) facility at University of Florida and its purity was confirmed by high performance liquid chromatography (HPLC). The sXIP peptide GLDWWSL, corresponding to residues 11–17 of ComS, was synthesized and purified to 96% by NeoBioSci (Cambridge, MA, USA).

Table 4.

List of strains.

Strain or plasmid	Relevant characteristicsa	Source or reference
Streptococcus mutans strains
UA159	Wild-type	ATCC 700610
ΔrcrR-P	rcrR::ΩKmR	Seaton et al., 2011
ΔrcrR-NP	rcrR::NPKmR	Seaton et al., 2011
PcomX::gfp	UA159 harboring PcomX::gfp	Son et al., 2012
PcomX::gfp/ΔrcrR-P	ΔrcrR-P harboring PcomX::gfp	This study
PcomX::gfp/ΔrcrR-NP	ΔrcrR-NP harboring PcomX::gfp	This study
ΔcomX	comX::NPErmR	Ahn et al., 2006
ΔcomSΔrcrR-NP	comS::NPErmR, rcrR::NPKmR	This study
ΔcomRΔrcrR-NP	comS::NPErmR, rcrR::NPKmR	This study
ΔPcomXΔrcrR-NP	comX -10 TATAAT::TCCCCT, rcrR::NPKmR	This study
comX::T162C	162nd bp comX T::C	This study
comX::T162C/ΔrcrR-NP	162nd bp comX T::C in ΔrcrR-NP background	This study
comX::T210A	210th bp comX T::A	This study
comX::T210A/ΔrcrR-NP	210th bp comX T::A in ΔrcrR-NP background	This study
XrpA::gfp	UA159 harboring XrpA::gfp	This study
XrpA::gfp/ΔrcrR-NP	ΔrcrR-NP harboring XrpA::gfp	This study
PcomXXrpA::gfp	UA159 harboring PcomXXrpA::gfp	This study
PcomXXrpA::gfp/ΔrcrR-NP	ΔrcrR-NP harboring PcomXXrpA::gfp	This study
pIB184	UA159 harboring pIB184 vector	Guo et al., 2014
pIB184xrpA	UA159 harboring pIB184-xrpA	This study
pIB184xrpA/ΔrcrR-NP	ΔrcrR-NP harboring pIB184-xrpA	This study
Plasmids
pDL278	Escherichia coli–Streptococcus shuttle vector, SpR	LeBlanc et al., 1992
pIB184	Shuttle expression plasmid with the constitutive P23 promoter, EmR	Biswas et al., 2008

^aP, polar; NP, non-polar; Km, kanamycin; Em, erythromycin; Sp, spectinomycin.

Construction of strains and DNA manipulation

Mutant strains of S. mutans (Table 4) were created using a PCR ligation mutagenesis approach, as previously described (Lau et al., 2002). Splice overlap extension was utilized to create single-base change mutations within the comX gene (Ho et al., 1989). Briefly, primers with the desired mutation(s) and primers that matched sequences located 0.6 kbp upstream and downstream of the sequence of interest (Supporting Information Table S1) were used to generate two PCR products. Selected pairs of PCR products with about 20 bp of homology, which included the desired mutations, were subjected to PCR for five cycles in the absence of added primers. Then, a second PCR of 30 cycles using the outer primers was performed to generate a 1.2 kbp fragment that contained the desired mutation. The final PCR product was run on a 1% agarose gel, purified and transformed into competent S. mutans along with a suicide plasmid carrying an internal fragment of the lacG gene and an erythromycin resistance (Em^R) determinant (Zeng et al., 2010). A Mismatch Amplifica-tion Mutation Assay (Cha et al., 1992) was used to screen Em^R isolates for the desired mutations, then PCR and DNA sequencing were used to confirm that the desired mutation had been introduced into the chromosome. Strains that had lost the suicide plasmid could be identified by their ability to grow on lactose and sensitivity to erythromycin. Strains containing translation fusions to xrpA regulatory elements were constructed by PCR amplification of the desired regions from the chromosome with primers incorporating HindIII and SpeI sites. Both PCR products and parent vector PcomX::gfp (Son et al., 2012) were digested with HindIII and SpeI, and ligated with T4 DNA ligase (Invitrogen). Overexpression of genes was achieved by amplifying the structural genes of interest from S. mutans UA159 and cloning into the expression plasmid pIB184 (Biswas et al., 2008). Transformants were confirmed by PCR and sequencing after selection on BHI agar with appropriate antibiotics. Plasmid DNA was isolated from E. coli by using QIAGEN (Chatsworth, CA, USA) columns, and restriction and DNA-modifying enzymes were obtained from Invitrogen (Gaithersburg, MD, USA) or New England Biolabs (Beverly, MA, USA). PCRs were carried out with 100 ng of chromosomal DNA by using Taq DNA poly-merase, and PCR products were purified with the QIAquick kit (QIAGEN). DNA was introduced into S. mutans by natural transformation and into E. coli by the calcium chloride method (Cosloy and Oishi, 1973).

RNA extraction and quantitative real-time PCR (qRT-PCR)

To measure the expression of genes using qRT-PCR, three colonies of the desired strain were grown overnight then diluted 1:50 into fresh BHI medium. All cultures were harvested by centrifugation after reaching an OD₆₀₀ of 0.5. Cell lysis was achieved through mechanical disruption (bead beating) and RNA was extracted by acidic phenol phase separation. The RNA was further purified with an RNeasy minikit (QIAGEN) according to the provided protocol and treatment with DNase 1 (QIAGEN). Purified RNA (1 μg) was used to generate cDNA from gene-specific primers (Supporting Information Table S1) using the Superscript III first-strand synthesis (Invitrogen) reverse transcription protocol. Real-time PCRs were carried out using an iCyclerQ real-time PCR detection system (Bio-Rad) and iQSYBR green supermix (Bio-Rad) according to the supplier’s protocol. 16S rRNA was used as an internal reference. All assays were performed in triplicate.

Production and purification of ComX antisera

Cloning, expression and purification of a His-tagged recombinant ComX for production of a ComX-specific antisera followed a previously described protocol (Ahn and Burne, 2006). Briefly, the comX coding region was amplified by PCR and cloned into the vector pET-45b(+) (Novagen, Darmstadt, Germany) so that the coding sequence of ComX was in frame with an N-terminal 6-histidine tag. Expression of the recombinant protein was induced with 0.4 mM isopropyl-β-D-thiogalactopyranoside. Cells were harvested 3 h after induction and recombinant ComX was purified using a Ni-nitrilotriacetic acid affinity chromatography under denaturing conditions. The purified protein (400 μg) was separated by SDS-PAGE and the gel slices containing the desired protein were excised and sent to Lampire Biological Laboratories (Pipersville, PA, USA), where an antiserum was elicited in a rabbit.

Immunoblotting

Overnight cultures of S. mutans were diluted 1:50 into 35 ml of fresh medium and harvested by centrifugation when the cultures reached an OD₆₀₀ = 0.5. When desired, 1 μM sCSP or 2μM sXIP was added when the cultures reached an OD₆₀₀ value of 0.2. Cell pellets were washed once with buffer A (0.5 M sucrose; 10 mM Tris-HCl, pH 6.8; 10 mM MgSO₄) containing 10 μg ml⁻¹ of phenylmethanesulfonyl fluoride (ICN Biomedicals), collected by centrifugation, and resuspended in 0.5 ml Tris-buffered saline (50 mM Tris-HCl, pH 7.5; 150 mM NaCl). Cells were lysed using a Mini Bead Beater (Biospec Products) in the presence of 1 volume of glass beads (average diameter 0.1 mm) for 30 s intervals, three times, with incubation on ice between homogenizations. Lysates were then centrifuged at 3000 × g for 10 min at 4°C. Protein concentrations of the resulting supernates were determined using the bicinchoninic acid assay (Thermo Scientific) with purified bovine serum albumin as the standard. ComX production was detected by immunoblotting with a protocol detailed elsewhere (Ahn and Burne, 2006). Briefly, 10 μg aliquots of proteins was mixed with 5× SDS sample buffer [200 mM Tris-HCl, pH 6.8; 10% (v/v) SDS; 20% glycerol (v/v); 10% (v/v) β-mercaptoethanol; 0.02% (v/v) bromophenol blue], loaded on a 12.5% polyacrylamide gel with a 5% stacking gel and separated by SDS-PAGE. Proteins were transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore) using a Trans-Blot Turbo transfer system and a protocol provided by the supplier (BioRad). The membranes were treated with primary polyclonal anti-ComX antisera at a 1:1000 dilution and a secondary peroxidase-labeled, goat anti-rabbit immunoglobulin G (IgG) antibody (1:5000 dilution; Kirkegaard & Perry Laboratories, USA). For GFP detection, 5 μg of protein from a whole cell lysate was loaded on a 12.5% polyacryla-mide gel with a 5% stacking gel and separated by SDS-PAGE. Anti-6X His tag^® antibody (ab81663, Abcam^®) was added at a final concentration of 0.5 ng μl⁻¹ for primary antibody incubation, and a secondary peroxidase-labeled, goat anti-mouse IgG antibody was later added at a 1:5000 dilution (Kirkegaard & Perry Laboratories). Detection was performed using a SuperSignal West Pico Chemiluminescent Substrate kit (Thermo Scientific) and visualized with a FluorChem 8900 imaging system (Alpha Innotech, USA).

Northern blots

Analysis of mRNA was also performed with a NorthernMax kit (Ambion) according to the supplier’s protocol. Briefly, 5 μg of purified S. mutans RNA was loaded onto a 1% agarose-formaldehyde gel in 1× MOPS (morpholinepropanesulfonic acid) buffer after denaturation at 65°C for 15 min. The RNA was then transferred to a BrightStar-Plus membrane (Ambion) through capillary transfer and UV cross-linked. The membrane was then hybridized with a 352 bp comX specific probe (Supporting Information Table S1) that was randomly labeled with biotin (BrightStar Psoralen-Biotin Nonisotopic Labeling Kit, Ambion) at 42°C overnight in ULTRAhyb buffer (Ambion). The hybridized membrane was washed in low- and high-stringency wash solutions for 5 and 15 min, respectively, and hybridization signals were detected using a BrightStar BioDetect kit (Ambion), following the supplier’s protocol.

5′-RACE

The ΔrcrR-P and ΔrcrR-NP strains to be analyzed were grown in BHI medium to an OD₆₀₀ of 0.5 before harvesting by centrifugation. 5′-RACE was conducted according to a protocol provided by Life Technologies (Carlsbad, CA, USA). Briefly, total RNA was isolated and purified as above, then cDNA was generated with 5′RACE-GSP1 (Supporting Information Table S1) using the Superscript III first-strand synthesis (Inv-itrogen) reverse transcription protocol. After the reaction was complete, RNase H was added to the sample and incubated at 37°C for 30 min. The remaining product was purified (QIAGEN) and used in a homopolymeric tailing reaction with 2 mM dCTP and 15 units of TdT (Invitrogen). The tailing reaction product was then further amplified using several nested gene-specific primers and anchor primers as specified in the supplier’s protocol. PCR products were analyzed on 1% agarose gels by electrophoresis. When a suitable amount of product could be visualized after electrophoresis, products were excised from the gel and purified using a gel extraction kit (QIAGEN). Products were then submitted to the ICBR facility at the University of Florida for sequencing with a nested comX-specific primer. TSSs were determined from the sequence results by locating the beginning of the homopoly-meric tailing sequence.

Transformation assays

Strains of S. mutans were grown overnight and then diluted 1:50 in 200 μl of the desired medium. When the cultures reached an OD₆₀₀ of 0.2, a final concentration of either 1 μM sCSP or 2 μM sXIP was added, if desired, and cultures were incubated at 37°C for 10 min. Then, 0.5 ng of purified plasmid pDL278, which harbors a spectinomycin resistance (Sp^R) gene, was added. Following 3 h of incubation at 37°C in a 5% CO₂ aerobic atmosphere, dilutions of the cultures were plated on BHI agar plates with or without 1 mg ml⁻¹ spectinomycin. colony forming units (CFUs) were enumerated after 48 h of incubation at 37°C in 5% CO₂ and transformation efficiency was determined by dividing the number of transformants (BHI-Sp^R) by the total number of viable bacteria from the cultures (BHI only) and is expressed as a percentage. All assays were performed in triplicate from three independent colonies.

Measurements of GFP fluorescence

For measurements of GFP fluorescence, cells were grown overnight, washed and diluted 1:50 into fresh FMC medium, with or without 2 μM sXIP. Two hundred microliters of the dilution was then loaded on a Costar™ 96-well assay plate (black plate with clear bottom; Corning Incorporated) and incubated at 37°C in a 5% CO₂ aerobic atmosphere. When cells reached an OD₆₀₀ = 0.5, the OD₆₀₀ and fluorescence (excitation 485/20 nm, emission 528/20 nm) were measured with a Synergy 2 multimode microplate reader (BioTek). Relative expression was calculated by subtracting the background fluorescence of UA159 (mean from six replicates) from raw fluorescence units of the reporter strains and then dividing by OD_600.

Transcriptome analysis (RNA-Seq)

Total RNA was isolated and purified from the various strains as described above. To remove 16S and 23S rRNAs, 10 μg of high-quality total RNA was processed using the MICROBEx-press^TM Bacterial mRNA Enrichment Kit (Ambion of Life Technologies, Grand Island, NY, USA), twice, before precipitating with ethanol and resuspending in 25 μl of nuclease-free water. The final quality of enriched mRNA samples was analyzed using an Agilent Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). cDNA libraries were generated from the enriched mRNA samples using the TruSeq Illumina kit (Illumina, San Diego, CA, USA), following instructions from the supplier. Deep sequencing was performed at the Cornell University Life Sciences Core Laboratories Center (Ithaca, NY, USA). Treatment of the Illumina sequence data and statistical analyses were preformed as described in detail elsewhere (Zeng et al., 2013). For analysis of the comX coding region, counts were extracted from the UCSC Table Browser (http://strep-genome.cshl.edu) (experimental tracks RNA-seq BAM FASTQ029-31 for strain ΔrcrR-NP and RNA-seq BAM FASTQ032-34 for strain ΔrcrR-P) based on the BAM FASTQ files for S. mutans UA159 (assembly: March 2012) and normalized by reads per kilobase per million mapped reads.