Identification of a Novel Bacteriocin Regulatory System in Streptococcus mutans

Summary

Recently, we described the function of an uncharacterized two-gene regulatory system consisting of a LytTR family transcription regulator and a putative membrane protein, which we referred to as the hdrRM operon. In this study, we determined that the HdrRM system controls the expression of an analogous uncharacterized regulatory system annotated as SMU.2080 and SMU.2081. Like hdrRM, the SMU.2080–2081 operon encodes a LytTR family transcription regulator and putative membrane protein, which we now refer to as BrsR and BrsM, respectively. Examination of the regulatory mechanism of the BrsRM system suggests that BrsM serves to antagonize the function of the transcription regulator BrsR. Further analyses of the regulatory role of BrsR determined that it functions as a transcription activator for a variety of bacteriocins and bacteriocin-related genes. In vitro electromobility shift assays confirmed that BrsR binds to the promoter regions of several bacteriocin genes and requires the presence of a LytTR family consensus direct repeat in order to stably bind DNA. In addition, we identified a novel regulatory scheme in which both the HdrRM and BrsRM systems coregulate each other and ultimately determine whether bacteriocin production will inhibit competitor organisms or result in lethality to the producer.

Introduction

Streptococcus mutans is a Gram positive oral commensal species that is a typical inhabitant of the human oral bioflm (dental plaque) and frequently associated with the development of dental caries (tooth decay) (Barsamian-Wunsch et al., 2004, Bowden, 1990, Munson et al., 2004, Nobre dos Santos et al., 2002, Thenisch et al., 2006, van Houte, 1993, Zambon & Kasprzak, 1995). Normally, fierce interspecies competition limits the proportion of S. mutans within the mixed species oral biofilm community and prevents pathogenicity. Competitors such as the Mitis group streptococci colonize the oral biofilm earlier, grow faster, and can generate hydrogen peroxide, which is a potent inhibitor of S. mutans growth (Kreth et al., 2008, Diaz et al., 2006, Caufield et al., 2000, Kreth et al., 2005b). However, under certain environmental conditions, S. mutans can outcompete other organisms leading to its overrepresentation within the plaque population and caries initiation. Abilities such as aciduricity/acidogenicity, biofilm formation, natural competence, and bacteriocin production are all crucial for S. mutans virulence (Banas, 2004, Kuramitsu, 1993).

The bacteriocins produced by S. mutans vary widely between strains, but virtually all strains seem to produce at least one type and likely encode several more (Bekal-Si Ali et al., 2002, Balakrishnan et al., 2002, Kamiya et al., 2005, Kamiya et al., 2008, Hamada & Slade, 1980). A significant proportion of the bacteriocin-like genes found in strains of S. mutans are currently classified as putative bacteriocins, as they are frequently not expressed under typical laboratory conditions and/or the appropriate indicator organisms have not been identified to detect their production. Thus, little is known about the regulation and function of many of these genes.

Multiple studies of the S. mutans competence system have demonstrated that the upstream competence regulators ComCDE are intimately involved in controlling the expression of a group of putative bacteriocins and related genes in addition to regulating natural competence (van der Ploeg, 2005, Hale et al., 2005a, Kreth et al., 2007, Kreth et al., 2006, Yonezawa & Kuramitsu, 2005, Perry et al., 2009). Activation of the competence system begins with the gene product of comC, which is a secreted cell signaling peptide referred to as the competence stimulating peptide (CSP) (Cheng et al., 1997, Li et al., 2002, Lunsford & London, 1996, Pestova et al., 1996). When the optimal CSP concentration has accumulated in the environment, it will then be sensed by a two-component regulatory system composed of the ComD sensor kinase and the ComE response regulator. Upon activation by CSP, the ComD sensor kinase will phosphorylate ComE, which will subsequently stimulate its transcription factor activity by increasing its affinity for downstream bacteriocin promoters (van der Ploeg, 2005, Yonezawa & Kuramitsu, 2005, Cheng et al., 1997, Havarstein et al., 1996, Pestova et al., 1996, Kreth et al., 2005a, Perry et al., 2009). Interestingly, the promoter regions of bacteriocin genes controlled by ComE each contain a set of direct repeats that match to the consensus pattern for the LytTR family of transcription regulators (Nikolskaya & Galperin, 2002, Kreth et al., 2006, van der Ploeg, 2005). Likewise, ComE is a member of the LytTR family and has been shown to bind directly to DNA containing this repeat sequence (Kreth et al., 2007).

Recently, we described the function of the hdrRM regulatory system, which is encoded by a two-gene operon and regulates the same competence and bacteriocin genes as ComCDE (Merritt et al., 2007, Okinaga et al., 2010b). Likewise, we demonstrated that regulation of the bacteriocin mutacin IV by HdrR requires the same direct repeat sequences as ComE, which suggests that the HdrRM system may recognize the same LytTR family consensus sequences as well (Okinaga et al., 2010a). Previous functional studies of the HdrRM system suggest that the membrane protein encoding gene hdrM negatively regulates the activity of HdrR, a LytTR family transcription regulator. For example, an hdrM mutant strain has increased genetic competence and mutacin IV production compared to the wild-type (Okinaga et al., 2010a). In contrast, an in-frame deletion of hdrR or a double deletion of hdrRM both result in wild-type phenotypes (Merritt et al., 2007). Furthermore, overexpressing hdrR recreates the phenotypes of the hdrM mutant, whereas overexpressing both hdrR and hdrM together suppresses these phenotypes near to wild-type levels (Okinaga et al., 2010a). All of these data suggest that HdrR is an active transcription activator in the absence of HdrM activity.

Here, we describe the function of a novel, uncharacterized two-gene regulatory system in S. mutans originally identified through studies of the HdrRM regulon (Okinaga et al., 2010b). This operon, annotated as SMU.2080–2081, is structurally analogous to the hdrRM operon. It encodes a LytTR family transcription regulator (SMU.2080) and membrane protein (SMU.2081), and our data suggest that SMU.2081 serves as a negative regulator of SMU.2080 transcription factor activity. Surprisingly, overexpression of SMU.2080 was lethal in multiple strains of S. mutans. Our results suggest that the source of lethality was due to bacteriocin overexpression. Thus, our data suggest that the primary function of the SMU.2080–2081 system is to serve as a dedicated regulatory system for the induction of S. mutans bacteriocin genes containing conserved LytTR family regulatory elements within their promoters. Under certain conditions, this system may also function together with the HdrRM system to trigger a cell death pathway mediated by bacteriocins.

Results

Identification of the SMU.2080–2081 operon

In our previous studies of the HdrRM system, we had determined that hdrR was a critical mediator of the competence phenotype of the hdrM mutant (Merritt et al., 2007). In fact, it was possible to reproduce this phenotype in a wild-type hdrM background simply by overexpressing hdrR (Okinaga et al., 2010a). Likewise, a microarray study of both the hdrM mutant and hdrR overexpression strain determined that both strains possess nearly identical gene expression profiles, with the late competence genes exhibiting the largest differences from the wild-type (Okinaga et al., 2010b). Further analysis attributed this result to the highly increased gene expression of the competence-specific sigma factor comX. Besides the late competence genes, we had also observed a similarly large increase in gene expression (~30-fold) from an uncharacterized two-gene operon annotated as SMU.2080 and SMU.2081. Upon further inspection, we noticed that this operon is highly reminiscent of the hdrRM operon. Both operons are arranged similarly with the first gene encoding a predicted LytTR family transcription regulator and the second encoding a putative membrane protein with two to three membrane spanning segments. In addition, both hdrM and SMU.2081 exhibit no homology to typical two-component sensor kinases nor do they have any obvious domain matches that might indicate some enzymatic function, such as ATPase or kinase domains. In order to confirm our previous microarray results, we measured the expression of the SMU.2080–2081 operon in the hdrM background via real-time RT-PCR. Given the large number of genes affected by ComX in the microarray data set, we were also interested to determine whether the operon was part of the ComX regulon. Thus, we compared operon expression in the comX and hdrM/comX backgrounds as well. As shown in figure 1, the expression of the SMU.2080–2081 operon was nearly 60-fold higher than the wild-type in the hdrM background with only a slight reduction in the hdrM/comX double mutant strain. This suggested that the increased SMU.2080–2081 expression is likely to be independent of ComX. As expected, we also found greatly increased SMU.2080–2081 expression in the hdrR overexpression strain (Fig. S1), which further confirmed the role of the HdrRM system as a regulator of SMU.2080–2081 expression.

Fig. 1. Effect hdrM and comX mutations upon SMU.2080–2801 operon expression.

The expression of the SMU.2080–2081 operon was measured by real-time RT-PCR in the wild-type (UA140), hdrM (M), hdrM/comX (M/X), and comX (X) backgrounds. The results are the average of three independent experiments and presented relative to the transcript abundance of the SMU.2080–2081 operon in UA140. All experiments measuring transcription via real-time RT-PCR were normalized using the gyrA gene as a housekeeping control.

Overexpression of SMU.2080 causes lethality

Since we had previously determined that the hdrRM operon functioned as an upstream regulator of the competence system, we were also curious to determine whether SMU.2080–2081 might play a role in mediating the competence phenotype of the hdrM mutant. Thus, we created a series of mutations and performed an epistatic analysis of hdrM and SMU.2080–2081. Consistent with the comX transcription data, we saw no evidence that suggested the increased competence phenotype of the hdrM mutation required either SMU.2080 or the SMU.2080–2081 operon (Fig. 2A). Surprisingly, we were not able to test the competence of the hdrM/SMU.2081 double mutant, as this particular combination was apparently lethal. In contrast, we had not observed any obvious toxicity associated with a double mutation of hdrM and SMU.2080 or SMU.2080 SMU.2081. From our previous experience with the HdrRM system, we reasoned that this result might indicate that SMU.2081 serves as a negative regulator of SMU.2080. To test this, we created two overexpression shuttle plasmids. The first constitutively expressed SMU.2080 and was predicted to recreate the hdrM/SMU.2081 lethal phenotype even in a wild-type background. The other constitutively expressed the entire SMU.2080–2081 operon, which we predicted to behave similarly to the wild-type, due to the presence of high SMU.2081 expression. We transformed each of these vectors as well as an empty vector control into the wild-type UA140. The transformations of both the empty vector and the constitutive SMU.2080–2081 plasmid yielded >2000 CFU/ml, whereas a similar transformation of the constitutive SMU.2080 construct produced few, if any, colonies (Fig. 2B). These results seemed to corroborate the lethality we had observed from the hdrM/SMU.2081 double mutation and suggested that SMU.2081 likely functions as an inhibitor of SMU.2080.

Fig. 2. Transformation of various SMU.2080–2081 mutants.

A) Transformation efficiency was measured as described in Experimental Procedures and values were normalized to the wild-type UA140 value (6.2 × 10⁻⁵). The strains are listed as follows: wild-type (UA140), SMU.2080 in-frame deletion (2080), SMU.2081 mutant (2081), operon deletion (2080–2081), hdrM mutant (M), hdrM/SMU.2080 double mutant (M/2080), hdrM/SMU.2081 double mutant (M/2081), and hdrM/SMU.2080–2081 triple mutant (M/2080–2081). The transformation efficiency of the hdrM/SMU.2081 strain was not determined and is listed as (ND). B) The shuttle vector pDL278 (Vector), the constitutive SMU.2080 plasmid (Ldh2080), and the constitutive SMU.2080–2081 plasmid (Ldh2080–2081) were all transformed into UA140. 10 μL of the transformation reactions were spotted in triplicate on selective medium in successive 10-fold dilutions. The numbers on the right side of the figure indicate the level of dilution from 10⁰ –10⁻².

Examination of the SMU.2080 regulon

Based upon the predicted function of SMU.2080 as a transcription regulator, we suspected that the source of lethality from SMU.2080 overexpression was likely due to toxicity generated from the altered expression of downstream genes. Thus, we examined the regulatory role of SMU.2080 with microarray analysis in hopes of identifying potential sources of toxicity. Given that the constitutive SMU.2080 construct was not viable, we decided to create an inducible SMU.2080 construct in order to circumvent the problem of lethality. We had previously determined that the S. mutans lacA promoter worked well for induction with lactose (unpublished results). Therefore, we placed the SMU.2080 open reading frame (ORF) under the control of the lacA promoter. For tighter regulatory control, we also cloned the lacR repressor onto the construct to prevent any titration effects of the chromosomally encoded lac repressor. This construct was then transformed into the SMU.2080–2081 deletion mutant strain and integrated onto the chromosome at the lac locus by single crossover homologous recombination. To confirm the functionality of the system, we compared the growth characteristics of the inducible SMU.2080 strain with that of a mock inducible strain carrying the same construct minus the SMU.2080 ORF. As shown in figure 3A, when grown under noninducing conditions (glucose), both strains had identical growth patterns. However, in the presence of lactose, the inducible SMU.2080 strain exhibited a marked reduction in growth rate (Fig. 3B), which suggested the construct was functioning as expected. As detailed in Experimental Procedures, we next grew both of these strains under inducing conditions and extracted RNA for microarray analysis. In total, we identified nearly 100 genes that were affected 2-fold or greater (Table S1). Of these, approximately 2/3 was due to increased expression. These genes also accounted for the vast majority of the large expression changes (>4-fold). Interestingly, the data suggested that SMU.2080 regulates a gene set with a strong overlap to hdrR. Both SMU.2080 and hdrR primarily regulate genes related to competence and bacteriocin production. However, SMU.2080 exhibited a considerably larger effect upon bacteriocin-related genes (Table 3), whereas hdrR had a much greater impact upon competence genes (Okinaga et al., 2010b). Thus, our current data indicate that hdrRM exhibits a strong bias for competence gene regulation, whereas SMU.2080–2081 is heavily biased towards bacteriocin regulation. SMU.2080 also had a significant impact upon hdrRM gene expression suggesting that the two systems can regulate each other (Table 3 and Table S1). In order to further confirm the role of SMU.2080 as a bacteriocin regulator, we repeated the lactose induction experiment and used real-time RT-PCR to measure the expression of SMU.150 (nlmA), SMU.423, and SMU.1906, which are all known or predicted bacteriocins (Kreth et al., 2007, Kreth et al., 2006, Qi et al., 2001, van der Ploeg, 2005). Consistent with the microarray data, we saw large increases in the expression of all three genes under SMU.2080 inducing conditions with SMU.423 exhibiting the greatest change. In the presence of lactose, both SMU.150 and SMU.1906 were expressed approximately 15–20-fold higher, whereas SMU.423 was >60-fold higher expressed (Fig. 4).

Fig. 3. Growth characteristics of the inducible SMU.2080 strain.

The growth curves of the mock inducible strain (black) and the SMU.2080 inducible strain (grey) are shown when grown in the presence of A) glucose and B) lactose. Shown here are the average optical density values of three replicate wells for each time point. This experiment was performed three times with similar results.

Table 3.

^* Selected genes from themicroarray data set.

NCBI Locus Tag	Fold Change	Functional Class
SMU.423	27.9	Putative bacteriocin
SMU.1905c	18.4	Putative bacteriocin secretion protein
SMU.925	12.1	Bacteriocin immunity protein
SMU.150	9.8	nlmA; mutacin IV
SMU.1854	7.5	hdrR; Competence and bacteriocin regulator
SMU.654	7.0	Putative bacteriocin associated ABC transporter
SMU.1906c	4.6	Putative bacteriocin
SMU.1997c	3.5	comX; Competence-specific sigma factor
SMU.1001	3.0	dprA; Late competence protein
SMU.1987	2.6	comYA; Late competence protein
SMU.625	2.3	comEA; Late competence protein
SMU.836	2.0	Late competence protein

^*Only one gene per operon is listed.

Fig. 4. Expression of bacteriocin genes after SMU.2080 induction.

Both the mock inducible strain (black bars) and the SMU.2080 inducible strain (grey bars) were grown in the presence of lactose and subsequently tested for the expression of SMU.150, SMU.423, and SMU.1906. The data are presented relative to the expression values for each gene in the mock inducible strain. Shown here are the averages of three independent experiments.

SMU.2080 is a transcription activator of genes containing LytTR family consensus direct repeats

It has been previously noted that multiple bacteriocins and bacteriocin-associated genes in S. mutans contain a direct repeat sequence in their upstream intergenic regions that is essential for their expression (Kreth et al., 2007, Kreth et al., 2006, van der Ploeg, 2005). These repeats also match to the predicted consensus pattern recognized by LytTR family transcription regulators (Nikolskaya & Galperin, 2002). With the exception of SMU.1914, which is not present in UA140, all of these previously identified genes were induced by SMU.2080. Given that SMU.2080 is predicted to be a LytTR family transcription regulator, this suggested that SMU.2080 could be a direct regulator of these genes. Furthermore, as a LytTR family transcription regulator, we would also predict that SMU.2080 requires these direct repeat sequences in order to stably interact with DNA. To test this, we PCR amplified approximately 300 bp fragments of the upstream intergenic regions of SMU.150, SMU.423, and SMU.1906 to use for electrophoretic mobility shift assays (EMSA). Using each of these DNA fragments as a template, we also created another set of similar DNA fragments that contained a rearrangement of the direct repeat sequences (Fig. 5A). Next, we expressed and purified recombinant SMU.2080 from E. coli and performed EMSA with both the wild-type and mutant intergenic region DNA for each gene. As shown in figures 5BD, a distinct mobility shift was observed with each of the three DNA fragments containing the wild-type direct repeat sequences. In contrast, no mobility shifts were observed in any of the reactions containing the mutant direct repeats, which indicated that SMU.2080 does in fact require these direct repeats to stably interact with DNA in vitro. As further evidence for the function of SMU.2080 in vivo, we also determined the transcription start sites for SMU.150, SMU.423, and SMU.1906 using rapid amplification of cDNA ends (RACE) PCR. All currently described LytTR family transcription regulators function in transcription activation (Galperin, 2008) with the exception of ComE from S. mutans, which functions both as an activator and a repressor (Kreth et al., 2007). Based upon the RACE PCR data, we determined that for all three genes the direct repeat sequences terminate 6 bp upstream of the predicted −35 sequences (Fig. 5E). This result coupled with the transcription and EMSA data all suggest that SMU.2080 functions in transcription activation. It is also apparent from the RACE PCR data of SMU.150, SMU.423, and SMU.1906 that all three genes share nearly identical promoters. They each possess similar extended −10 sequences and identical −35 sequences. Also, the spacing between the −10 and −35 is identical for SMU.150 and SMU.1906 and 1 nucleotide longer for SMU.423.

Fig. 5. Analysis of SMU.2080 transcription factor activity.

A) Illustrated here are the sequences of the DNA fragments used for electrophoretic mobility shift assays (EMSA) of SMU.2080 and the promoter regions of SMU.150, SMU.423, and SMU.1906. Sequences labeled DR+ contain the wild-type direct repeat sequence, whereas those labeled DR- contain a rearrangement of the direct repeat nucleotides. All other bases in the DNA fragments are identical. EMSA analysis was performed using purified recombinant SMU.2080 and both wild-type (DR+) and mutant (DR-) direct repeat DNA for B) SMU.150, C) SMU.423, and D) SMU.1906. As described in Experimental Procedures, increasing concentrations of SMU.2080 were incubated with 10 ng of the indicated DNA before running the reactions on a nondenaturing polyacrylamide gel. For each series of four EMSA reactions, the first lane contained no protein. E) Rapid amplification of cDNA ends (RACE) PCR was used to determine the transcription start sites of SMU.150, SMU.423, and SMU.1906. The LytTR family direct repeats are shown in bold, while the -35 and extended -10 sequences are underlined. The transcription start sites are indicated with an arrow over the +1 site.

Induction of SMU.2080 increases bacteriocin production

Given the direct role of SMU.2080 as a transcription regulator of multiple bacteriocin-like genes, next we were curious to determine whether the increased transcription of SMU.150, SMU.423, and SMU.1906 also resulted in increased bacteriocin production. Thus, we transformed UA140 with either a lactose-inducible SMU.2080 shuttle vector construct or the empty shuttle vector and measured bacteriocin production on chemically defined medium plates containing lactose. We first examined the production of mutacin IV (encoded by SMU.150), as its production and inhibition spectra are both well characterized (Kreth et al., 2005a, Qi et al., 2001). As shown in figure 6A, there was a distinct growth inhibition halo of Streptococcus sanguinis surrounding the S. mutans strain containing the empty vector, while this halo was considerably larger in the inducible SMU.2080 strain. As expected, the large growth inhibition halo surrounding the inducible SMU.2080 strain disappeared almost completely in the mutacin IV mutant background. Next, we were interested to determine whether the expression of SMU.2080 resulted in a similar response from the gene products of SMU.423 and SMU.1906. Since it was unknown whether either of these genes actually encodes active bacteriocins, it was first necessary to screen a panel of Gram positive organisms to find potential indicator strains (data not shown). After repeated attempts, we were unsuccessful at attributing any specific inhibitory activity to SMU.1906, but we were able to correlate inhibitory activity with SMU.423. Both Lactococcus lactis and vancomycin resistant Enteroccoccus faecium (FQ1VRE) were found to be suitable indicators for the SMU.423 bacteriocin. We chose to use FQ1VRE as our indicator because we had observed that our strain of L. lactis was susceptible to mutacin IV as well (data not shown). As shown in figure 6B, FQ1VRE was inhibited to a greater extent with SMU.2080 induction, but there was also apparently another bacteriocin produced that partially obscured our results. Thus, we repeated the same experiment in another strain of S. mutans (UA159) and found that the SMU.423 bacteriocin was only produced in the inducible SMU.2080 strain (Fig. 6C). The empty vector control strain exhibited no obvious inhibition towards FQ1VRE. Thus, UA159 is likely to express little or no SMU.423 under standard bacteriocin assay conditions. Based upon the activity of the SMU.423 bacteriocin, we propose that this gene be referred to as nlmD (mutacin VI) to be consistent with the previous naming convention of similar S. mutans bacteriocins (Hale et al., 2005b, Qi et al., 2001). In addition, we are now referring to SMU.2080 as brsR (bacteriocin regulatory system Regulator) and SMU.2081 as brsM (bacteriocin regulatory system Membrane protein), due to their direct role in the regulation of bacteriocin gene expression.

Fig. 6. Deferred antagonism assay of the inducible brsR strain.

A) The production of mutacin IV was tested by the deferred antagonism assay as described in Experimental Procedures. The development of a growth inhibition halo is indicative of the presence of mutacin IV inhibiting the sensitive strain Streptococcus sanguinis NY101. S. mutans strains are listed as follows: UA140 + shuttle plasmid (Vector), UA140 + inducible brsR plasmid (Lac2080), and UA140 ΔSMU.150 + inducible brsR plasmid (Lac2080Δ150). B) The development of a growth inhibition halo is indicative of the presence of mutacin VI inhibiting the sensitive strain Enterococcus faecium (FQ1VRE). S. mutans strains are listed as follows: UA140 + shuttle plasmid (Vector), UA140 + inducible brsR plasmid (Lac2080), and UA140 ΔSMU.423 + inducible brsR plasmid (Lac2080Δ423). C) Enterococcus faecium (FQ1VRE) was used as indicator organism to detect bacteriocin production in the following strains: UA159 + shuttle plasmid (Vector), UA159 + inducible brsR plasmid (Lac2080), and UA159 ΔSMU.423 + inducible brsR plasmid (Lac2080Δ423). Deferred antagonism assays were performed at least 3 times with similar results. Shown here are the results from a representative experiment.

Bacteriocin overproduction can result in lethality

Given the potent transcription activation of multiple bacteriocins by BrsR, we suspected that bacteriocin overproduction could be the source of toxicity associated with brsR constitutive expression. From our transcription data, it seemed the most likely culprits would be SMU.150 (nlmA), SMU.423 (nlmD), or SMU.1906, as these bacteriocins were the most strongly affected. Thus, we deleted each of these genes and then transformed them with either the empty shuttle vector or the constitutive brsR construct. As shown in figure 7A, each of the mutant strains exhibited similar transformation efficiencies as the wild-type when they were transformed with the empty shuttle vector. In contrast, all but the SMU.150 mutant yielded few, if any colonies when transformed with the constitutive brsR plasmid (Fig. 7B). This indicated that the major source of toxicity was likely to be due to the overexpression of SMU.150. As mentioned previously, the bacteriocins produced by S. mutans vary widely between strains and not all strains possess SMU.150. Therefore, we were curious as to whether the toxicity of the constitutive brsR construct would be similar in other strains of S. mutans. After transforming a panel of common laboratory strains and clinical isolates, we found that each was highly transformable with the empty shuttle vector, whereas we rarely observed any colonies from a transformation of the constitutive brsR construct (Table 4). Hence, a transformation of the constitutive brsR construct resulted in at least a 100-fold reduction in the number of antibiotic resistant colonies from each strain, which suggests that these strains all likely suffer from a similar lethal effect from brsR. However, this lethality is unlikely to be solely attributable to SMU.150 because we determined that some of the tested strains do not encode this gene (data not shown). Likewise, we mutated SMU.150 in UA159 and saw no evidence for reduced toxicity from the constitutive brsR plasmid (data not shown). This suggests that the toxicity associated with brsR overexpression is likely to be similar among S. mutans strains, whereas the source of toxicity is probably strain-specific.

Fig. 7. Transformation of the constitutive brsR plasmid into bacteriocin mutant strains.

A) The shuttle vector pDL278 was transformed into the following strains: wild-type (UA140), SMU.150 mutant (Δ150), SMU.423 mutant (Δ423), and SMU.1906 mutant (Δ1906). 10 μL of the transformation reactions were spotted onto selective medium at the indicated dilutions (10⁰ –10⁻²). B) The same strains were transformed with the constitutive brsR construct. 10 μL of the transformation reactions were spotted onto selective medium at the indicated dilutions (10⁰ –10⁻²). This experiment was performed three times with similar results. Shown here is one representative experiment.

Table 4.

Transformation of the constitutive brsR plasmid into various S. mutans strains

Strains	Serotype	* Transformation Efficiency	Reference
GS-5	c	<1%	(Gibbons, 1972)
Ingbritt	c	<1%	(Coykendall, 1971)
L18	c	<1%	(Waterhouse & Russell, 2006)
NG8	c	<1%	(Lee et al., 1989)
UA159	c	<1%	(Ajdic et al., 2002)
CL1	k	<1%	Clinical isolate
JF3	c	<1%	Clinical isolate
JF243	c	<1%	Clinical isolate

^*Transformation efficiency is presented relative to that of the empty shuttle vector, which was arbitrarily set at 100%.

Discussion

In the current paper, we describe the function a previously uncharacterized two-gene operon annotated as SMU.2080 and SMU.2081, which we now refer to as the brsRM operon. This operon bears a strong resemblance to the recently described hdrRM regulatory system, but unlike hdrRM, brsRM only seems to possess a modest regulatory control over the competence system. The brsRM system is, however, heavily biased for the regulation of bacteriocins. Functionally, the brsRM system is the mirror image of hdrRM, which is strongly biased for competence gene regulation and a more modest regulator of bacteriocins. Interestingly, the two systems seem to be strong inducers of each other as well. When one system is active, the other operon will be induced. However, it is apparently lethal for both systems to be fully activated simultaneously, as we were unable to generate an hdrM/brsM double mutant. Based upon our data, this lethality can be primarily attributed to the regulatory function of BrsR, since it was not possible to transform a constitutive brsR expression plasmid. In the case of UA140, the toxic effect from brsR overexpression required the presence of SMU.150 (mutacin IV), which indicates that bacteriocin overproduction could be the source of toxicity. This is consistent with a recent study by Perry et al., which found that the exogenous addition of high levels of CSP could trigger cell death in a subpopulation of producer cells via the induction of the bacteriocin SMU.1914 (cipB) (Perry et al., 2009).

After HdrRM, the BrsRM regulatory system is now the second example of what appears to be a novel class of signal transduction system in S. mutans. Both systems seem to function quite similarly and studies are underway to determine the mechanism by which these two systems regulate the activity of their transcription factors. Thus far, the evidence suggests that the membrane proteins for both systems have the ability to antagonize the transcription regulators. For example, in both systems, a mutation of the membrane protein leads to constitutive activity from the transcription regulator. Likewise, the phenotypes associated with constitutive expression of the transcription regulator can be negated simply by coexpressing the membrane protein (Fig. 2B). This type of regulation is highly analogous to the regulation of ECF sigma factors by membrane bound anti-sigma factors (Raivio & Silhavy, 2001). Although, it seems highly unlikely that BrsR functions as a sigma factor, since nlmA, nlmD, and SMU.1906 all start transcription downstream from typical sigma-70 type promoters (Fig. 5E). However, it is conceivable that the mechanisms involved in controlling the HdrRM and BrsRM systems could function similarly as ECF sigma factor systems. It is also intriguing that both HdrRM and BrsRM are potent activators of each other. Currently, it is unknown which environmental cues are involved in triggering the two systems. Should they respond to the same signal, we would predict that this signal could elicit severe toxicity to the cell because of a positive feedback regulation between both HdrRM and BrsRM (Fig. 8). This situation would mimic that of the hdrM/brsM double mutation. However, if they respond to separate signals, then we would predict that individual stimulation of either system would result in the induction of both operons without significant toxicity, since one of the two systems would remain inactive. In this case, toxicity would only occur upon simultaneous stimulation of both systems. In either scenario, it seems that the toxic effect specifically requires the overexpression of brsR, since we could recreate the lethal phenotype simply by forcing the constitutive expression of brsR (Fig. 2B). In contrast, a similar hdrR construct elicited no obvious toxicity, which is consistent with its weaker impact upon bacteriocin regulation (Okinaga et al., 2010a). Though it should be noted that HdrR is still a critical component of the pathway leading to lethality, since it is essential for the positive feedback regulation that results in brsR overexpression. This would explain why the brsM mutant strain is fully viable even though it can no longer inhibit BrsR. In the brsM background, HdrM is still available to antagonize HdrR and prevent brsR overexpression (Fig. 8). While the ecological function of the HdrRM and BrsRM systems remains to be determined, the genetic data indicate that the cross regulation between the two systems could constitute a mechanism to induce cell death via the overexpression of bacteriocins. Similarly, it may also be possible to subvert this regulation as a novel strategy to selectively inhibit the growth of S. mutans.

Fig. 8. Model of the S. mutans competence and bacteriocin regulatory network.

In S. mutans, competence development and bacteriocin production are coordinately regulated by multiple systems. In response to CSP, the ComCDE system induces comC and comED gene expression as well as competence and bacteriocin genes. In response to unknown signals, the HdrRM system primarily induces the expression of competence genes and the brsRM operon, while the BrsRM system activates bacteriocin gene expression as well as the hdrRM operon. Both HdrM and BrsM antagonize the activity of HdrR and BrsR respectively. If the inhibitory activity of either HdrM or BrsM is intact, the cell will express bacteriocin genes normally and inhibit the growth of competitor organisms. However, if both membrane proteins are simultaneously inactivated, a strong positive feedback regulation will cause the overexpression of brsR and result in toxic levels of bacteriocin production that inhibit the producer.

Even though in UA140 SMU.150 mutant strains were viable in the brsR overexpression background, it should be noted that bacteriocin genes are quite variable among S. mutans strains and our results suggest that SMU.150 is highly unlikely to be the only bacteriocin capable of causing toxicity when overexpressed. Additionally, BrsR probably regulates a variety of bacteriocins in other strains due to the presence of highly conserved promoter elements among numerous S. mutans bacteriocin genes. Interestingly, among this group of bacteriocins considerable variability can occur within the structural genes, despite the conservation of promoter elements. For example, the bacteriocin Smb produced in the S. mutans strain GS-5 (Yonezawa & Kuramitsu, 2005) is an entirely different class of bacteriocin from both mutacin IV (SMU.150) and mutacin VI (SMU.423). Yet, as noted by van der Ploeg, the upstream intergenic region of the smbM1–smbG operon contains a set of the same type of direct repeats found in nlmA (SMU.150), nlmD (SMU.423), and SMU.1906 (van der Ploeg, 2005). This region also contains identical putative extended −10 and −35 sequences to those of nlmA and SMU.1906. Data from a transposon insertion into the predicted extended −10 sequence of the smbM1 smbG operon further suggests that this sequence is likely to function as the operon promoter (Yonezawa & Kuramitsu, 2005). Such high conservation of the regulatory elements in the promoters of S. mutans bacteriocins might also explain the widespread toxicity of the constitutive brsR construct (Table 4). Presumably, this is because many or all strains of S. mutans possess bacteriocin genes that can be activated by BrsR. Thus, we speculate that there are likely to be a variety of bacteriocins capable of generating toxicity to the producer when overexpressed. This would also explain why strains that do not encode SMU.150 are still subject to severe growth defects in the brsR overexpression background.

Initially, it was surprising that brsR induction consistently resulted in a much greater transcriptional response from nlmD compared to nlmA and SMU.1906. For example, after the induction of brsR, nlmA and SMU.1906 both increased in expression by approximately 15–20-fold, whereas nlmD was >60-fold increased (Fig. 4). All three genes contain nearly identical promoter sequences and mRNA leader regions (Fig. 5A and E). However, there are two subtle differences in the nlmD promoter that may account for its greater response. The first difference is in the direct repeats. Of the three genes, nlmD is the only one that encodes an identical direct repeat, whereas the repeats of both nlmA and SMU.1906 differ by 1 nucleotide (Fig. 5A and E). The second difference is the spacing between the −10 and −35 sequences. Compared to nlmA and SMU.1906, the nlmD promoter has 1 additional base in between its −10 and −35 sequences (Fig. 5E). For a typical promoter, this wider spacing would be predicted to actually weaken its strength, as both the nlmA and SMU.1906 −10 and −35 sequences are already spaced slightly wider than a consensus promoter. However, Sidote et al. recently reported the first crystal structure of a LytTR family transcription regulator (Sidote et al., 2008). In their cocrystal structure of the DNA binding domain of AgrA with DNA, it is apparent that dimers of AgrA induce significant bends in the target DNA. While it has yet to be determined what role DNA bending plays in transcription activation from LytTR family transcription factors, it may be that the slightly wider spacing of the nlmD promoter actually improves transcription initiation once BrsR has bound the direct repeats.

Experimental Procedures

Bacterial strains and growth conditions

The bacterial strains and plasmids used in this study are listed in Table 1. S. mutans UA140 and its derivatives were grown in brain heart infusion broth (BHI; Difco) or on BHI agar plates. When performing transformation assays, cells were grown in Todd Hewitt medium (Difco) with 0.3% (wt/vol) yeast extract (THYE). When testing for bacteriocin inducing production or performing growth curves, cells were grown in a chemically defined medium (CDM) (He et al., 2008) containing either glucose or lactose as the carbon source. For the selection of antibiotic-resistant colonies, BHI plates were supplemented with 800 μg ml⁻¹ kanamycin (Sigma), 15 μg ml⁻¹ erythromycin (MP Biomedicals), 15 μg ml⁻¹ tetracycline (Sigma), or 1,000 μg ml⁻¹ spectinomycin (Sigma). S. mutans strains were grown anaerobically (85% N₂, 10% CO₂, 5% H₂) at 37°C. Escherichia coli cells were grown in Luria-Bertani (LB; Difco) medium with aeration at 37°C. E. coli strains carrying plasmids were grown in LB containing 100 μg ml⁻¹ ampicillin (Fluka), 100 μg ml⁻¹ spectinomycin, or 100μg ml⁻¹ kanamycin.

Table 1.

Bacterial strains and plasmid used in this study

Strains and plasmids	Characteristics*	Reference	Purpose
Strain
E.coli DH5α	Cloning strain		Plasmid cloning
E. coli BL21(DE3) pLysS	Protein expression strain		BrsR production
UA159	Genome reference wild-type Streptococcus mutans	(Ajdic et al., 2002)	Bacteriocin assay; viability assay
UA140	Wild-type Streptococcus mutans	(Qi et al., 2001)	Principal wild-type recipient strain
Ingbritt	Streptococcus mutans	(Coykendall, 1971)	Viability assay
GS5	Streptococcus mutans	(Gibbons, 1972)	Viability assay
NG8	Streptococcus mutans	(Lee et al., 1989)	Viability assay
CL1	Streptococcus mutans	This work	Viability assay
JF3	Streptococcus mutans	This work	Viability assay
JF243	Streptococcus mutans	This work	Viability assay
NY101	Wild-type Streptococcus sanguinis	(Weerkamp et al., 1977)	Indicator strain
FQ1VRE	Vancomycin-resistant Enterococcus faecium	This work	Indicator strain
TO-m	UA140::pHdrM-i; hdrM-; Spr	(Okinaga et al., 2010a)	Epistatic analysis
ZX-r	UA140 ΔbrsR; Kmr	This work	Epistatic analysis
ZX-m	UA140 ΔbrsM; Kmr	This work	Epistatic analysis
ZX-rm	UA140 ΔbrsRM; Kmr	This work	Epistatic analysis
ZX-rTOm	UA140::pHdrM-i; ΔbrsR Spr; Kmr	This work	Epistatic analysis
ZX-rmTOm	UA140::pHdrM-i; ΔbrsRM Spr; Kmr	This work	Epistatic analysis
TO-x	UA140::pComX-KO; comX−; Emr	(Li et al., 2002)	Epistatic analysis
TO-mx	UA140::pHdrM-i and pComX-KO; Sper; Emr	(Okinaga et al., 2010a)	Epistatic analysis
UA140 I− IV+	UA140 ΔmutC; Tcr	(Kreth et al., 2005b)	Mutacin IV assay
UA140 I+ IV−	UA140 ΔnlmAB; Kmr	(Kreth et al., 2005b)	Mutacin IV assay
UA140 I− IV−	UA140 ΔmutC ΔnlmAB; Tcr; Kmr	(Kreth et al., 2005b)	Mutacin IV assay
ZX-d	UA140 ΔnlmD; Kmr	This work	Mutacin IV assay
ZX-cd	UA140 ΔmutC ΔnlmD; Tcr; Kmr		Mutacin IV assay
ZX-um	UA140 ΔSMU.1906c 1905c Emr	This work	Bacteriocin assay
ZX-lac	UA140 ΔbrsRM::pFW-C; Kanr; Spr	This work	Microarray
ZX-lacR	UA140 brsRM::pFW-IER; Kanr; Spr	This work	Microarray
ZX-ab	UA159 ΔnlmAB; Kmr	This work	Mutacin IV assay
ZX-abd	UA159 ΔnlmAB ΔnlmD; Kmr Tcr	This work	Mutacin IV assay
ZX-ct	UA140 ΔmutC + pDL278; Tcr; Spr	This work	Mutacin IV assay
ZX-ir	UA140 ΔmutC + pDL-IER; Tcr; Spr	This work	Mutacin IV assay
ZX-irab	UA140 mutC nlmAB + pDL-IER; Tcr; Kmr; Spr	This work	Mutacin IV assay
ZX-ird	UA140 ΔmutC ΔnlmD + pDL-IER; Tcr; Kmr; Spr	This work	Mutacin VI assay
ZX-ct2	UA159 ΔnlmAB + pDL278; Kmr; Spr	This work	Mutacin IV assay
ZX-ir2	UA159 ΔnlmAB + pDL-IER; Kmr; Spr	This work	Mutacin IV assay
ZX-ird2	UA159 ΔnlmAB ΔnlmD + pDL-IER; Kmr; Tcr; Spr	This work	Mutacin IV assay
Plasmids
pGNaa3	pBluescript with aphAIII in MCS; Apr; Kmr	(Niu et al., 2008)	Mutant construction
pDL278	E. coli–Streptococcus shuttle vector; Spr	(Chen & LeBlanc, 1992)	Bacteriocin assay
pOER	pDL278::Φ(ldhp -brsR); Spr	This work	Viability assay
pRM	pGEM + brsRM; Apr	This work	Viability assay
pOERM	pDL278::Φ(ldhp-brsRM); Spr	This work	Viability assay
pDL-IER	pDL278::Φ(lacAp -brsR) + lacR; Spr	This work	Bacteriocin assay
pFW5	Suicide vector; Spr	(Podbielski et al., 1996)	Microarray
pFW-C	pFW5 + lacAp + lacR; Spr	This work	Microarray
pFW-IER	pFW5::Φ(lacAp -brsR) + lacR; Spr	This work	Microarray
pET-29b(+)	His6 fusion protein expression vector; Kmr	Novagen	BrsR production
pETR	pET29b + brsR; Kmr	This work	BrsR production
pWT150	pGEM + nlmAp; Apr	This work	BrsR EMSA
pWT423	pGEM + nlmDp; Apr	This work	BrsR EMSA
pWT1906c	pGEM + SMU.1906cp; Apr	This work	BrsR EMSA
pDR150	pGEM + nlmAp DR−; Apr	This work	BrsR EMSA
pDR423	pGEM + nlmDp DR−; Apr	This work	BrsR EMSA
pDR1906c	pGEM + SMU.1906cp DR−; Apr	This work	BrsR EMSA

^*Ap^r: ampicillin resistance, Sp^r: spectinomycin resistance,Km^r: kanamycin resistance,Em^r: erythromycin resistance, and Tc^r: tetracycline resistance

General DNA manipulation and strain construction

Phusion DNA polymerase, restriction enzymes, T4 DNA ligase, and other DNA modifying enzymes were all purchased from New England Biolabs. All bacterial strains and plasmids are listed in Table 1, while the PCR primers used in this study are listed in Table. 2.

Table 2.

Primers used in this study.

Primer	Sequence*	Purpose
2080upF	TGATAAGATCATGCTCTTG	brsR deletion, brsRM deletion
2080upR	GTGCCTCCATTACATCAGTTATCATCCTTTCTTG	brsR deletion
2081F	GATAACTGATGTAATGGAGGCACATCATGAAA	brsR deletion,
2081R-kan	TAATCTTACCTATCACCTGCTAGAGAAAATTTTGTGGG	brsR deletion,brsRM overexpression
kanF	AGGTGATAGGTAAGATTATACCG	aphAIII cassette amplification
kanR	CCCTATCTAGCGAACTTTTAGA	aphAIII cassette amplification
2081dnF-kan	AAAGTTCGCTAGATAGGGCTGACAAGTCCCTTTCAACA	brsR deletion
2081dnR	GTTCGAAGTAATGAAAAGTC	brsR deletion, brsM deletion, brsRM deletion
2081upF	AAACTCTCCTTTAGCTTCA	brsM deletion
2081upR-kan	TAATCTTACCTATCACCTGAGTATTTTCATGATGTGC	brsM deletion
2080upR2-kan	TAATCTTACCTATCACCTCTTAAAGTTTGAATCCTTG	brsRM deletion
423upF	TCATCGTATTGAACGTGTTG	smu.423 deletion
423upR-kan	TAATCTTACCTATCACCTGCTTGTGTATTCATATGATAG	smu.423 deletion
423dnF-kan	AAAGTTCGCTAGATAGGGGTGCTGCTACTTTTTGTTGA	smu.423 deletion
423dnR	TCATTTTGATGTCACCTCCA	smu.423 deletion
423upR-tet	ATTCAAAAGCCCAAAAGGGCTTGTGTATTCATATGATAG	smu.423 deletion
423dnF-tet	GGTTTCTTGTTAAATAAGGTGCTGCTACTTTTTGTTGA	smu.423 deletion
tetF	CCTTTTGGGCTTTTGAATGGAGGAA	tetM cassette amplification
tetR	CTTATTTAACAAGAAACCATATTTATATAACAAC	tetM cassette amplification
1906cupF	CAATGTAGAAGTCTCAAGTA	smu.1906c-1905c deletion
1906cupR-erm	CAACCACCCGACTTTGAAGTTCAAATGCTTG TGTATTC	smu.1906c-1905c deletion
1905cdnF-erm	TATACTACTGACAGCTTCAAGTAATGAGAA CTAAAGTAGG	smu.1906c-1905c deletion
1905cdnR	TGTATTCTTGCTATCTGATG	smu.1906c-1905c deletion
ermF	TTCAAAGTCGGGTGGTTGTC	ermAM cassette amplification
ermR	GAAGCTGTCAGTAGTATACC	ermAM cassette amplification
ldhF	GCCGGATCCCCGAGCAACAATAAC	ldh promoter amplification
ldhR	CGCACTAGTAACATCTCCTTATAATTT	ldh promoter amplification
2080F	GCACTAGTCTATATACTAAAATCAAGAAAGG	Overexpression and induced expresson of brsR
2080R	CCCAAGCTTTGAGTATTTTCATGATGTGC	Overexpression and induced expresson of brsR
lacAF	GCGGATCCGACACTTTATGCTCAAGCTTTG	lacA promoter amplification
lacAR	GCCACTAGTTCTCCTTGAAATGATTTTTTAGTC	lacA promoter amplification
2080RTF	CCATCCGAGGCAGCATTG	brsR RT-PCR
2080RTR	CGCCTAGCCACGAAACTGA	brsR RT-PCR
150RTF	GGACAGCCAAACACTTTCAACTG	nlmA RT-PCR
150RTR	TGCCGCGACTGCTTCTC	nlmA RT-PCR
423RTF	GGTGGTATGATTAGATGTGCACTTG	smu.423 RT- PCR
423RTR	TCCCATACCCCCTACAAATCC	smu.423 RT-PCR
1906RTF	CACAAGCATTTGAACAATTTAACGT	smu.1906c RT-PCR
1906RTR	TCCAACTACATCCCCCACCTT	smu.1906c RT-PCR
gyrARTF	ACCCGGCCCCGATTT	gyrA RT-PCR
gyrARTR	ATGTATACCCGATTTTCCCATGA	gyrA RT-PCR
2080-29bF	CCGCATATGAAATTAAAAATTATCAAGGAT	His6 tagged expression of BrsR
2080-29bR	GCGCTCGAGAAGATTTAATTGTTCTTTGAAG	His6 tagged expression of BrsR
P150F	GTGTAAAACTTCTATTAAACAG	DNA preparation for EMSA
P150R	TTGCACATATTCCAATTGCC	DNA preparation for EMSA
P423F	TATGTTTGTAGTCAGTTGCG	DNA preparation for EMSA
P423R	AATCCTAAACCTGCAGAACC	DNA preparation for EMSA
P1906cF	GAAAAAATCATGGATTTTCTTG	DNA preparation for EMSA
P1906cR	CTGAAAAACCAGCTTTATCG	DNA preparation for EMSA
150DR-F	ATAGCTCTGTCAATAGATATTTTGCTCCATTTTGAA	Direct repeat mutagenesis
150DR-R	TTTGTCTAGTCGAACCATTTTTGATGTTCTGAAAC	Direct repeat mutagenesis
423DR-F	ATAGCTCTGTCAATAGATATTTTGCTCCATTTTGAA	Direct repeat mutagenesis
423DR-R	TTTGTCTAGTCGAACTAAAAATGACGTTATAATGCCTG	Direct repeat mutagenesis
1906cDR-F	ATAGCTCTGTCAATAGATATTTTGCTCCATTTTGAA	Direct repeat mutagenesis
1906cDR-R	TTTGTCTAGTCGAACCATTTTTGATGTTCTGAAACTATC	Direct repeat mutagenesis
150outerR	ATCGAATGAGTCCCCAAGTGCC	5′RACE
423outerR	AGCCTCCTAAAGCCGCTCCAG	5′RACE
1906couterR	CGCACCGTATGCAGCTAAGCCAC	5′RACE
150innerR	GAGTCGCACCTGCCAGTCCTCCAATTGCTGCTG	5′RACE
423innerR	GAGAGTAACTGTACCAGCTCCCATACCCCCTAC	5′RACE
1906cinnerR	CCACCAACGGGGTTACCACCTGCACCAATG	5′RACE

^*Sequences complementary to the aphAIII cassette are italicized, sequences complementary to the ermAM gene are in bold, and sequences complementary to the tetM gene are italicized in bold. Underlined sequences indicate restriction sites.

Construction of brsR in-frame deletion

Since brsR is the first gene of the brsRM operon, we constructed an in-frame deletion mutant. Approximately 1kb of upstream sequence of brsR was generated by PCR using the primer pair 2080upF/2080upR, while a 500 bp fragment encompassing brsM was PCR amplified with the primer pair 2081F/2081R-kan. The two amplicons contained a complementary region, which allowed for a subsequent overlapping PCR reaction using the primer pair 2080upF/2081R-kan and the two amplicons as template. The resulting 1.5 Kb amplicon contained the 1Kb brsR upstream region, an in-frame-deletion of brsR (only start and stop codons remain) and an intact brsM. In order to introduce this fragment into S. mutans, a kanamycin resistance cassette aphAIII was PCR amplified from pGNaa3 with the primer pair kanF/R and the brsM downstream fragment was PCR amplified with the primer pair 2081dnF-kan/2081dnR. The two amplicons contained a complementary region, which allowed for a subsequent overlapping PCR reaction using the primer pair kanF/2081dnR. The resulting 2 kb amplicon contained the aphAIII cassette and the brsM downstream region. In order to ligate the two fragments, the 1.5 Kb and 2 Kb amplicons were mixed and used as a template for overlapping PCR using the primer pair 2080upF/2081dnR. The final 3.5 Kb amplicon was transformed into UA140 and mutants were selected for kanamycin resistance. Mutant isolates were confirmed by sequencing to ensure the integrity of the expected brsR in-frame deletion. Confirmed mutant strains were designated as ZX-r.

Construction of single and double mutations of brsM, brsRM, and hdrM

The brsM deletion mutant strain ZX-m was constructed by an overlapping PCR mutagenesis strategy. Briefly, about a 1Kb region upstream of brsM was PCR amplified with primer pair 2081upF/2081upR-kan. Another 1 Kb region downstream of brsM was PCR amplified with the primer pair 2081dnF-kan/2081dnR. The aphAIII cassette was PCR amplified with primer pair kanF/R. The three amplicons were mixed and used as a template for a subsequent PCR using the primer pair 2081upF/2081dnR. The resulting 3 Kb amplicon was transformed into UA140 and transformants were selected with kanamycin. The brsRM operon deletion mutant strain ZX-rm was constructed by using a similar overlapping PCR strategy. The 1 Kb upstream and downstream fragments were generated by PCR using the primer pairs 2080upF/2080upR2-kan and 2081dnF-kan/R. Both amplicons have overlapping regions with the aphAIII cassette. Construction of the hdrM mutant strain TO-m has been described previously (Okinaga et al., 2010a). In this study, the hdrM mutation was introduced into ZX-r and ZX-rm by transforming genomic DNA to generate ZX-rTOm and ZX-rmTOm respectively.

Construction of SMU.423 and SMU.1906c-SMU.1905c mutants

Both the SMU.423 mutant strain ZX-d and SMU.1906c-1905c mutant strain ZX-um were created by transforming overlapping PCR products. The SMU.423 upstream and downstream fragments were generated by PCR using the primer pairs 423upF/423upR-kan and 423dnF-kan/423dnR. The upstream and downstream amplicons were mixed with PCR amplified aphAIII and used as a template for overlapping PCR with the primer pair 423upF/423dnR. The SMU.1906c-1905c upstream and downstream fragments were generated by PCR using the primer pairs 1906cupF/906cupR-erm and 1905cdnF-erm/1905cdnR. The upstream and downstream amplicons were mixed with ermAM amplified with the primer pair ermF/ermR and used as a template for overlapping PCR with the primer pair 1906cupF/1905cdnR.

Construction of mutC, nlmAB, and SMU.423 single and double mutants

The mutacin I deficient mutant strain UA140 I⁻IV⁺, mutacin IV deficient mutant strain UA140 I⁺IV⁻, and the corresponding double mutant UA140 I⁻IV⁻ were constructed previously (Kreth et al., 2005). The ΔmutC ΔSMU.423 double mutant strain ZX-cd was constructed by transforming the mutC genomic DNA into the SMU.423 strain ZX-d.

Construction nlmAB and SMU.423 single and double mutants in UA159

The mutacin IV mutation was moved into UA159 by transforming the genomic DNA of UA140 I⁺IV⁻ to generate the strain ZX-ab ( nlmAB). To construct the SMU.423 nlmAB double mutant strain ZX-abd, the tetracycline resistance cassette tetM and the upstream/downstream regions of SMU.423 were amplified with the primer pairs tetF/tetR, 423upF/423upR-tet, and 423dnF-tet/423dnR. Using the three amplicons as template, overlapping PCR was performed with the primer pair 423upF/423dnR. Next, the SMU.423 PCR product was transformed into the UA159 nlmAB background and transformants were selected with tetracycline.

Construction of lactose inducible strains for microarray

The mock inducible strain ZX-lac and the brsR inducible strain ZX-lacR used in the microarray studies were constructed by transforming the mock inducible and brsR inducible plasmids pFW-C and pFW-IER (Table 1) into the brsRM deletion mutant strain ZX-rm.

Construction of strains used for the deferred antagonism assay

The empty vector control strains ZX-ct and ZX-ct2 were generated by transforming the shuttle vector pDL278 (Table 1) into a mutacin I deficient strain of UA140 (ΔmutC) and a mutacin IV deficient strain of UA159 (ΔnlmAB) respectively. The inducible brsR shuttle plasmid pDL-IER (Table 1) was transformed into the following strains: UA140 ΔmutC to create the strain ZX-ir, UA140 ΔmutC ΔnlmAB to create ZX-irab, UA140 ΔmutC ΔnlmD to create ZX-ird, UA159 ΔnlmAB to create ZX-ir2, and UA159 ΔnlmAB ΔnlmD to create ZX-ird2. Assays measuring the production of mutacin IV utilized the strains ZX-ct, ZX-ir, and ZX-irab. The production of mutacin VI was tested with ZX-ct, ZX-ir, ZX-ird, ZX-ct2, ZX-ir2, and ZX-ird2.

Constructions of plasmids

pOER and pOERM

The promoter region of the constitutive lactate dehydrogenase gene ldh was generated by PCR with the primer pair ldhF/R and the brsR fragment was amplified with 2080F/R. BamHI and SpeI was used to digest the ldh promoter amplicon and the brsR amplicon was digested with SpeI and HindIII. The two fragments were then ligated into the BamHI and HindIII sites of the shuttle vector pDL278 to generate the brsR constitutive expression plasmid pOER. To create the brsRM constitutive expression plasmid pOREM, the brsRM fragment was first PCR amplified with the primer pair 2080F/2081R and cloned into the pGEM-T easy vector (Promega) to generate the plasmid pRM. The plasmid pRM was then digested with SpeI and EcoRI to release the brsRM insert and agarose gel purified. This fragment was mixed with same BamHI/SpeI digested ldh promoter amplicon used for the pOER construct and both inserts were ligated into pDL278 cut with BamHI and EcoRI to generate pOERM.

pDL-IER and pFW-IER

A fragment encompassing the promoter region of the lacA gene and the entire lactose operon repressor open reading frame was PCR amplified with the primers lacAF/R, while the brsR fragment was amplified with the primers 2080F/R. The lacA amplicon was digested with BamHI and SpeI and the brsR amplicon was digested with SpeI and HindIII. The two fragments were then ligated into pDL278 digested with BamHI and HindIII to generate pDL-IER. The same fragments were also ligated to the suicide vector pFW5 cut with BamHI and HindIII to generate pFW-IER. The control plasmid pFW-C was constructed by inserting the BamHI/SpeI lacA promoter fragment into the BamHI and XbaI sites of pFW5.

Transformation assay

Determination of transformation efficiency was performed using a previously described methodology (Niu et al., 2008). UA140 and its derivatives were diluted 1:30 from overnight cultures and grown to an optical density OD₆₀₀ 0.2–0.3 in THYE before the addition of transforming DNA. Genomic DNA containing a tetracycline marker was added at a final concentration of 10 μg/mL for each reaction followed by additional 2 h of incubation. In some cases, plasmids containing a spectinomycin resistance marker were used for transformation efficiency determination at a final concentration of 20 ng/mL. After the incubation period, cultures were briefly sonicated to disperse cell chains and plated on tetracycline or spectinomycin-containing BHI agar plates, as well as on nonselective BHI plates. Successful transformation was scored based on acquired antibiotic resistance following transformation, and the total number of viable cells was determined by counting the colonies growing on nonselective plates. Transformation efficiency was determined by calculating the ratio of transformants to the total number of viable cells.

RNA extraction

For lactose inducing conditions, UA140 and its derivatives were cultivated overnight at 37°C in CDM containing 0.5% (wt/vol) glucose. The overnight cultures were diluted 1:30 in CDM with 1% (wt/vol) lactose in a total volume of 30 ml. The cells were allowed to grow to an optical density at 600 nm of 0.3 and collected by centrifugation. If not otherwise mentioned, UA140 and its derivatives were cultivated overnight at 37°C in BHI. The overnight cultures were then diluted 1: 30 in 30 ml BHI and harvested when the OD₆₀₀ reached 0.3. All cultures were centrifuged at 4°C and stored at -80°C prior to RNA extraction. RNA was extracted from cell pellets using a previously described methodology (Niu et al., 2008) and used for all real-time RT-PCR analyses, microarray assays, and 5′ rapid amplification of cDNA ends (RACE) PCR analyses.

Real-time PCR analysis

Total RNA (300 ng) was used for complementary DNA (cDNA) synthesis using Stratascript reverse transcriptase (Stratagene) according to the manufacturer’s protocol. For real-time reverse transcription polymerase chain reaction (RT-PCR), primers were designed using Primer Express 3.0 software (Applied Biosystems), the reactions were prepared using Bio-Rad iTaq SYBR Green Supermix with ROX (Bio-Rad) and an Applied Biosystems 7300 was used for detection (Applied Biosystems). Relative changes in gene expression were calculated using the ΔΔCT method described previously (Niu et al., 2008). Total cDNA abundance between samples was normalized using primers specific to the gyrA gene. All primers used for real-time RT-PCR are listed in Table. 2.

Microarray

The microarray protocol has been previously described (Okinaga et al., 2010b, Ajdic & Pham, 2007). Briefly 15 μg of total RNA was used for cDNA synthesis with SuperScript II reverse transcriptase (Invitrogen) and fragmented with Roche DNase I (Roche). Hybridization, washing and scanning of the custom S. mutans GeneChip expression microarrays (Affymetrix) were performed according to the procedures described by Affymetrix. Microarray data processing and analysis employed the Genechip operating software (GCOS) version 1.4. Samples with a signal detection P-value <0.01 were analyzed for expression differences. Microarray data are available at the National Center for Biotechnology Information Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo) under the accession number GSE22902.

Bacterial growth curves

The test strains were first grown overnight in CDM medium containing 0.5% (wt/vol) glucose under anaerobic conditions. The overnight cultures were then diluted 1:30 in CDM containing either 1% (wt/vol) lactose or 0.5% (wt/vol) glucose. 200 μl of cell suspension was added to each well of clear, sterile 96-well microplate, which was loaded and sealed with TopSeal^™ microplate sealing film (VWR) under anaerobic conditions. The sealed plate was then placed inside a FLUO Star Optima plate reader (BMG) and optical density was measured over a 20-hour period. Negative control wells containing medium only were used as blanks.

Electrophoretic mobility shift assays (EMSA)

Expression and Purification of recombinant BrsR

The coding region of brsR was amplified by PCR using the primer pair 2080-29bF/2080-29bR and subsequently digested with NdeI and XhoI. The resulting fragment was cloned into the corresponding sites of the expression plasmid pET-29b to generate pETR. The recombinant plasmid was introduced into E. coli BL21(DE3) pLys for protein expression. BrsR protein was expressed as a C-terminal His₆-tagged fusion protein. BrsR protein was purified using a Probond^™ (Invitrogen) resin column according to the manufacturer’s protocol. Desalting and protein concentration were performed with an Amicon Ultra-15 Centrifugal Filter Unit (Millipore).

Preparation of DNA fragments

Three fragments corresponding to the promoter regions of SMU.150, SMU.423, and SMU.1906c were generated by PCR using Accuprime Pfx (Invitrogen) with the primer pairs P150F/R, P423F/R and P1906cF/R. In order to generate the corresponding promoter fragments containing direct repeat mutations, the three wild-type promoter fragments were first cloned into the pGEM-T easy vector (Promega) to generate pWT150, pWT423, and pWT1906c respectively. Each plasmid was then used as a template for inverse PCR using the following phosphorylated primers pairs 150DR-F/R, 423DR-F/R, 1906cDR-F/R. The PCR fragments were then ligated and transformed into E. coli. Plasmids containing the expected mutant direct repeat sequences were confirmed by sequencing and designated as pDR150, pDR423, and pDR1906c respectively. Using the resulting plasmids as templates, the mutated promoter regions were amplified by PCR with the primer pairs P150F/R, P423F/R, and P1906cF/R.

Electrophoretic mobility shift assays

Binding reactions were performed in a total volume of 20 μl and contained 10 ng DNA with purified His-tagged BrsR protein at the following concentrations: 0 nM, 10 nM, 20 nM, and 30 nM. The reaction buffer was composed of 10 mM Tris-HCl (pH 7.5), 2 mM dithiothreitol (DTT), 5 mM MgCl₂, 0.5 mg ml⁻¹ calf BSA, and 5% (vol/vol) glycerol. After incubation on ice for 30 min, protein-bound and free DNA were separated by electrophoresis on non-denaturing 5% polyacrylamide gels with a running buffer composed of 0.5x TBE containing 40 mM Tris-HCl (pH 7.8) and 20 mM boric acid at 6V cm⁻¹ for 1.5 hours at 4 °C. After electrophoresis, gels were stained with SYBR Green I (Sigma) at a 1:10,000 dilution for 30 minutes. Gel images were obtained using an Alphaimager gel imager (Alpha Innotech) equipped with a 302 nm UV transilluminator and the appropriate filter.

Determination of transcription start sites

Transcription start sites were determined using the FirstChoice RLM-RACE kit (Ambion). 5′ rapid amplification of cDNA ends (RACE) PCR was performed according to the manufacturer’s protocol. Total RNA (10 μg) was ligated with the 5′ RACE adaptor and reversed transcribed. PCR amplification was performed in two successive reactions, first using the 5′ RACE outer and gene-specific outer primers followed by a second PCR using the 5′ RACE inner and gene-specific inner primers. The amplicon was gel-purified, cloned into the pGEM-T Easy vector (Promega) and sequenced. The gene-specific outer primers for SMU.150, SMU.423, and SMU.1906c are 150outerR, 423outerR, and 1906couterR respectively. The gene-specific inner primers for SMU.150, SMU.423, and SMU.1906c are 150innerR, 423innerR, and 1906cinnerR respectively (Table 2)

Deferred antagonism assay for mutacin production

UA140, UA159, and their derivatives were first grown overnight in CDM liquid medium containing 0.5% (wt/vol) glucose under anaerobic conditions. A 5 μl volume of each overnight culture was spotted onto CDM agar plates containing 1% (wt/vol) lactose and incubated anaerobically at 37 °C overnight. Next, the plates were overlaid with soft agar suspension of the indicator strain (Streptococcus sanguinis NY101 for mutacin IV and Enterococcus faecium FQ1VRE for mutacin VI). The presence of a growth inhibition halo after overnight incubation was indicative of mutacin production. All assays measuring mutacin production in UA140 were performed in a mutacin I deficient strain. Assays measuring mutacin VI production in UA159 were performed in a mutacin IV deficient strain.