Role of VicRKX and GlnR in pH-Dependent Regulation of the Streptococcus salivarius 57.I Urease Operon.

Ureolysis by Streptococcus salivarius is critical for pH homeostasis of dental plaque and prevention of dental caries. The expression of S. salivarius urease is induced by acidic pH and carbohydrate excess. The differential expression is mainly controlled at the transcriptional level from the promoter 5' to ureI (p ureI ). Our previous study demonstrates that CodY represses p ureI by binding to a CodY box 5' to p ureI , and the repression is more pronounced in cells grown at pH 7 than in cells grown at pH 5.5. Recent sequence analysis revealed a putative VicR consensus and two GlnR boxes 5' to the CodY box. The results of DNA affinity precipitation assay, electrophoretic mobility shift assay, and chromatin immunoprecipitation-PCR analysis confirmed that both GlnR and VicR interact with the predicted binding sites in p ureI . Isogenic mutant strains (vicRKX null and glnR null) and their derivatives (harboring S. salivarius vicRKX and glnR, respectively) were generated in a recombinant Streptococcus gordonii strain harboring a p ureI-chloramphenicol acetyltransferase gene fusion on gtfG to investigate the regulation of VicR and GlnR. The results indicated that GlnR activates, whereas VicR represses, p ureI expression. The repression by VicR is more pronounced at pH 7, whereas GlnR is more active at pH 5.5. Furthermore, the VicR box acts as an upstream element to enhance p ureI expression in the absence of the cognate regulator. The overall regulation by CodY, VicR, and GlnR in response to pH ensures an optimal expression of urease in S. salivarius when the enzyme is most needed. IMPORTANCE Dental plaque rich in alkali-producing bacteria is less cariogenic, and thus, urease-producing Streptococcus salivarius has been considered as a therapeutic agent for dental caries control. Being one of the few ureolytic microbes in the oral cavity, S. salivarius strain 57.I promotes its competitiveness by mass-producing urease only at acidic growth pH. Here, we demonstrated that the downregulation of the transcription of the ure operon at neutral pH is controlled by a two-component system, VicRKX, whereas the upregulation at acidic pH is mediated by the global transcription regulator of nitrogen metabolism, GlnR. In the absence of VicR-mediated repression, the α subunit of RNA polymerase gains access to interact with the AT-rich sequence within the operator of VicR, leading to further activation of transcription. The overall regulation provides an advantage for S. salivarius to cope with the fluctuation of environmental pH, allowing it to persist in the mouth successfully.

INTRODUCTION

Urease is a Ni2+-dependent metalloenzyme that generally consists of three subunits, α, β, and γ, encoded by ureC, -B, and -A, respectively (1). An exception is found in Helicobacter pylori, in which the α subunit (UreA) is encoded by a fusion of the ureA and ureB genes seen in other bacterial urease systems (2). The assembly of a catalytically active urease requires the products of ureE, -F, -G, and -D, known as accessory genes that encode proteins required for the incorporation of nickel into the metallocenter within the active site. Some of the bacterial urease operons contain genes encoding uptake systems for urea and Ni2+. For example, a H+-gated urea transporter is encoded by ureI in H. pylori (3). The Streptococcus salivarius strain 57.I ureMQO genes encode a Ni2+-specific ATP binding cassette transporter (4). Although bacterial ureases are highly conserved, the expression of bacterial urease operons is regulated by various mechanisms. For instance, Bacillus pasteurii and Sporosarcina ureae express urease constitutively, whereas the urease expression in Proteus mirabilis is activated by a urease-specific activator, UreR, in the presence of urea (5). On the other hand, the expression of the urease operon in Bacillus subtilis is regulated by GlnR, TnrA, CodY, and PucR in response to nitrogen availability (6, 7).

Urea is present abundantly in the saliva and crevicular fluid in healthy individuals (8). Thus, ureolysis by bacterial ureases to produce ammonia and CO2 is the primary alkali generation machinery in the oral cavity, which plays a key role in plaque pH homeostasis and dental caries prevention (9). Among the oral microflora, S. salivarius is the most dominant and highly ureolytic species (10). Genes encoding a functional urease are arranged as an operon in S. salivarius 57.I (ureIABCEFGDMQO) (4, 11, 12). A previous study using the chemostat culture system indicates that the expression of S. salivarius urease is enhanced by acidic growth pH, excess amounts of carbohydrates, and high growth rates (13). Expression analyses demonstrate that the differential expression of the urease operon in response to growth conditions is regulated mainly at the transcriptional level via a σ70-dependent promoter located 5′ to ureI (p) (12). Analysis of the cis elements of p reveals that the 21-bp region immediately 5′ to the −35 element of p is responsible for the repression of p, whereas the 40-bp region further upstream participates in the positive regulation of p (14). Furthermore, the regulation of p in response to pH is also present in the recombinant nonureolytic Streptococcus gordonii strain CH1, which harbors a p-chloramphenicol acetyltransferase (CAT) gene (cat) fusion, suggesting that the expression of p is regulated by a global regulatory circuit (14). By using the chemostat culture system and various molecular analyses, we found that CodY inhibits p expression by binding to the CodY box located 2 bases 5′ to the −35 element of p, and the repression is more evident during growth at pH 7 than at pH 5.5. Furthermore, in the absence of CodY, the AT tract in the CodY box also acts as an upstream (UP) element to enhance urease expression (15).

Recent sequence analysis revealed a VicR binding consensus element (5′-TGTWAH-N5-TGTWAH) (16, 17) and two GlnR binding consensus elements (5′-TGTNA-N7-TNACA) (18, 19) located 5′ to the CodY box in p, raising the possibility that both VicR and GlnR participate in the urease regulatory circuit. VicR is the response regulator of the VicRKX two-component system (TCS). This system is highly conserved in low-GC-content Gram-positive bacteria (20). Different from the typical TCS, this operon also encodes a metallohydrolase, VicX, in addition to VicR and the sensor kinase VicK. VicRKX has been implicated as the master regulatory system for cell wall metabolism, cell viability, biofilm formation, genetic competence, acid stress response, and oxidative stress response (16, 21–23). This system is essential for the viability of several streptococcal species (16, 23), with the exception of S. gordonii CH1 (22).

GlnR, a member of the MerR family of regulators, is the key regulator for nitrogen metabolism in most Gram-positive bacteria (24). The optimal DNA binding activity of GlnR requires feedback-inhibited glutamine synthetase (FBI-GS) in B. subtilis (25). Generally, GlnR represses the expression of the GlnR regulon under nitrogen excess (26). A recent study by Chen and colleagues demonstrates that GlnR is activated at acidic growth pH in Streptococcus mutans, and the repression of the GlnR regulon at acidic pH shifts the metabolism from glutamine synthesis to ATP generation to enhance acid tolerance (19).

The regulation by VicR and GlnR of p expression under different growth conditions was investigated in this study. We found that the regulation by VicR and GlnR of p is modulated by the growth pH. GlnR activates the expression of p at acidic pH, whereas VicR represses p activity. In the absence of VicR, the AT tract within the VicR box of p acts as an UP element to further enhance p expression.

RESULTS

Both VicR and GlnR bind directly to the 5′ flanking region of p.

Recent sequence analysis identified a VicR box and two putative GlnR boxes in p (Fig. 1). The VicR box, 5′-TGTAAATGTTGcaaAAT, differs by 3 bases (indicated by lowercase letters) from the consensus derived from S. mutans (16). The 3′ end of the VicR box overlaps the 5′ end of the CodY box by 4 bp. GlnR box 1 (5′-TGTTAGCTTGACTAAtA) and GlnR box 2 (5′-TGTCATTTTTTGaCACc) are 3 bases apart and differ from the GlnR box consensus of S. mutans by 1 and 2 bases (indicated by lowercase letters), respectively.

FIG 1 

Nucleotide sequence of p and the 5′ flanking region. The transcription initiation site (+1) of the urease operon is indicated. The −10 and −35 elements of p are underlined. The translation start codon of ureI is indicated by a bent arrow. The ribosome binding site of ureI is in italics. The CodY box is shaded. The VicR box and the two GlnR boxes are overlined. Bases of the putative VicR box and GlnR boxes that are different from the consensus sequence are in lowercase.

A DNA affinity precipitation assay (DAPA) was performed to investigate whether the endogenous VicR interacts with p. A VicR-specific signal was detected with the probe containing the putative VicR box, and the signal was abolished completely when a probe with mutations in the VicR box was used, confirming the binding specificity of VicR (Fig. 2A). As both direct and indirect interaction with the target DNA could lead to a positive result in the DAPA, an electrophoretic mobility shift assay (EMSA) was performed to verify whether VicR binds directly to the predicted VicR box in p. A shift of the VicR box-specific probe was observed with 0.8 µM MalE-VicR, and a dose-dependent increase in the intensity of the signal was observed with this probe (Fig. 2B), indicating that MalE-VicR bound directly to the target. When an unlabeled probe in 300-fold excess was used in the reaction mixture, no shift was observed, confirming the binding specificity of VicR to p. The recombinant MalE protein alone failed to bind to the probe (data not shown). Finally, the in vivo interaction between VicR and p was verified by chromatin immunoprecipitation (ChIP)-PCR assay, and the result confirmed the interaction between VicR and p (Fig. 2C).

FIG 2 

DAPA, EMSA, and ChIP-PCR analyses demonstrating in vitro and in vivo interaction of VicR with the VicR box in p. (A) DAPA demonstrating specific interaction between VicR and the putative VicR box. A 100-µg amount of the total cell lysate of S. salivarius 57.I was used as the input control (I). Amounts of 1 mg of the total lysate were incubated with a biotin-labeled, VicR box-specific probe (an annealing product of VicR_box_S and VicR_box_AS) (II) and with a probe with mutations in the VicR box (an annealing product of mVicR_box_S and mVicR_box_AS) (III). The VicR protein in the reaction mixtures was detected by immunoblotting with anti-VicR antiserum. (B) EMSA of VicR binding to the putative VicR box in p. Lane 1, probe only; lanes 2 to 6, 0.2 to 3.2 µM MalE-VicR in 2-fold increments; lane 7, 3.2 µM MalE-VicR with a specific competitor. All reactions were carried out with 0.01 pmol biotin-labeled probe. (C) The in vivo interaction between VicR and p was examined by ChIP-PCR assay. (I) Input control; (II) PCR product obtained from a reaction with anti-VicR antiserum; (III) final result from a reaction mixture with the preimmune rabbit serum.

The same approach was used to investigate the interaction between GlnR and the putative GlnR boxes described above. Similarly, specific interactions between endogenous GlnR and the probes containing the putative GlnR box 1 and GlnR box 2, respectively, were observed in the DAPA (Fig. 3A). The binding of GlnR to the putative GlnR boxes was further confirmed by EMSA (Fig. 3B). A shift was seen with probes specific for each of the GlnR boxes, and a dose-dependent enhancement was seen with the probe specific for GlnR box 1 with increasing amounts of MalE-GlnR. Although a signal with high intensity was seen with the probe specific for GlnR box 2 in the presence of 0.8 µM MalE-GlnR, no signal was detected in the reaction mixtures with smaller amounts of MalE-GlnR. Thus, the relative degrees of affinity of GlnR for these two targets remain unclear. When an unlabeled probe in 300-fold excess was used in the reaction mixture, no shifted band was detected. As above, the recombinant MalE protein failed to interact with both probes, confirming the binding specificity of GlnR (data not shown). Notably, the addition of glutamine and glutamine synthetase to the EMSA reaction mixture did not enhance the binding (data not shown), indicating that, unlike in B. subtilis, FBI-GS was not required for the binding activity of S. salivarius GlnR. Finally, the in vivo interaction between GlnR and p was confirmed by the ChIP-PCR assay (Fig. 3C).

FIG 3 

DAPA, EMSA, and ChIP-PCR analyses demonstrating in vitro and in vivo interaction of GlnR with the GlnR boxes in p. (A) DAPA demonstrating specific interaction between GlnR and the putative GlnR boxes. A 1-µg amount of the total cell lysate of S. salivarius 57.I was used in the input control (I). Amounts of 500 µg of the total cell lysate were incubated with biotin-labeled probes specific for GlnR box 1 (an annealing product of GlnR_box-1_S and GlnR_box-1_AS) and box 2 (an annealing product of GlnR_box-2_S and GlnR_box-2_AS), respectively (II), and with probes with mutated bases in the putative GlnR boxes (annealing products of mGlnR_box-1_S plus mGlnR_box-1_AS and mGlnR_box-2_S plus mGlnR_box-2_AS) (III). The GlnR protein in the reaction mixtures was detected by immunoblotting with anti-GlnR antiserum. (B) EMSA demonstrating interaction between GlnR and the putative GlnR box 1 (lanes 1 to 5) and GlnR box 2 (lanes 6 to 10) of p. Lanes 1 and 6, probe only; lanes 2 to 4 and 7 to 9, 0.2 to 0.8 µM MalE-GlnR in 2-fold increments; lanes 5 and 10, 0.8 µM MalE-GlnR with a specific competitor. All reactions were carried out with 0.01 pmol biotin-labeled probe. (C) ChIP-PCR assay demonstrating in vivo interaction between GlnR and p. (I) Input control; (II) PCR product obtained from a reaction mixture containing anti-GlnR antiserum; (III) final result from a reaction with preimmune rabbit serum.

VicRKX represses the transcription of p.

A vicR-deficient strain is needed to investigate whether VicR is involved in the urease regulatory circuit. However, we failed to generate a vicR-null recombinant strain in S. salivarius 57.I. Several studies have shown that VicR is essential for the viability of several oral streptococci (16, 23), and thus, it is possible that VicR plays a similar role in S. salivarius 57.I. To circumvent this difficulty, we first investigated the regulatory function of VicRKX in p expression by using a promoter fusion with mutations in the predicted VicR box. A derivative of S. salivarius strain MC308 was constructed, strain MC308_mVicR_box, in which the sequence of −64 to −59 of p (5′-TGTAAA) in the p-cat fusion was mutated to 5′-GTCGAC. Notably, the CodY box in this fusion remains untouched. Both strain MC308 and strain MC308_mVicR_box were cultivated in brain heart infusion (BHI) at pH 7.5 and pH 5.5. A 4-fold increase in CAT activity was observed in strain MC308_mVicR_box compared to its activity in strain MC308 in cells grown at pH 7.5, whereas only a 2.1-fold increase was detected in cells grown at pH 5.5 (Fig. 4A), indicating that the putative VicR box was involved in the negative regulation of p, and the repression was more evident at pH 7.5.

FIG 4 

The VicRKX two-component system represses p expression. (A) Effect of the putative VicR box on p activity. The p-cat activities in batch-grown S. salivarius MC308 and MC308_mVicR_box were determined by the CAT assay. Both strains were grown in BHI containing 50 mM KPO4 at pH 7.5 or BHI-HCl at pH 5.5. (B) Effect of VicRKX on p activity. The CAT activities in batch-grown S. gordonii SL17, SL17_ΔvicRKX, and SL17_CΔvicRKX at pH 7.5 and pH 5.5 were examined. (C) Effect of growth pH and glucose concentration on p activity. The CAT activities in chemostat-grown SL17_CΔvicRKX and SL17_ΔvicRKX were examined in cells grown at pH 7 and pH 5.5, with 20 mM and 100 mM glucose (Glc). The specific activity (Sp. Act.) was calculated as nmol Cm acetylated min−1 mg−1 total protein. Values are the mean results and standard deviations from three independent experiments. Significant differences between strains were analyzed using one-way ANOVA. ***, P < 0.001.

Since a vicRKX-null strain is available in S. gordonii CH1 (22), we examined p expression in a vicRKX-null derivative of S. gordonii strain SL17 (strain SL17_ΔvicRKX) and its derivative that harbors the vicRKX operon of S. salivarius 57.I (strain SL17_CΔvicRKX). Notably, S. gordonii SL17 harbors a single copy of the p-cat fusion on gtfG. In agreement with the cis element analysis, elevated CAT activity was detected in the vicRKX-null background at both pH 7.5 and pH 5.5. A wild-type level of CAT activity was observed in strain SL17_CΔvicRKX, confirming that VicR repressed the transcription of p (Fig. 4B). It was also noted that the CAT activity of strain SL17_CΔvicRKX was consistently lower (approximately 30%) than that of strain SL17, suggesting that S. salivarius VicR represses p expression more efficiently than S. gordonii VicR does.

To investigate the potential effects of growth pH and glucose concentration on the regulation of VicR, strains SL17_CΔvicRKX and SL17_ΔvicRKX were grown in a chemostat, a culture system that allows tight control of the growth pH and glucose concentration. The CAT activities in cells grown at pH 7 and pH 5.5 with 20 and 100 mM glucose were examined. At pH 7, a 4.7-fold increase in CAT activity was seen in SL17_ΔvicRKX compared to the activity in SL17_CΔvicRKX in the presence of 20 mM glucose, but comparable levels of CAT activity were detected between these two strains when cells were grown with 100 mM glucose (Fig. 4C). On the other hand, 1.8- and 1.5-fold increases in activity were seen in cells grown at pH 5.5 with 20 mM and 100 mM glucose, respectively (Fig. 4C). Collectively, the activity of VicR was modulated by both pH and carbohydrate concentration and VicRKX repressed p most effectively at neutral pH under glucose limitation.

GlnR activates p more strongly at pH 5.5.

A glnR-deficient strain is essential to investigate how GlnR regulates urease expression. Unfortunately, multiple attempts failed to generate a glnR-null mutant strain in S. salivarius 57.I, indicating that mutations in glnR are also lethal in S. salivarius 57.I. Thus, we initiated the study by using p-cat fusions with mutations in the putative GlnR boxes, since the interaction between GlnR and the putative GlnR boxes 5′ to p has been confirmed (Fig. 3). Previous promoter deletion analysis indicates that a 40-bp region 22 bases 5′ to the −35 element of p exhibits a positive effect on p transcription (14). This region contains the 3′ portion of GlnR box 2 and the entire GlnR box 1, suggesting that GlnR activates p expression. As a positive effect would be more evident in a repressor-free host, i.e., strain S. salivarius ΔcodY, the activity of all promoter derivatives was examined in the codY-deficient background. In agreement with the hypothesis, mutations in the putative GlnR boxes reduced CAT activity in batch-grown cells at both pH 7.5 and pH 5.5 and the reduction is slightly more evident at pH 5.5 (Fig. 5A), suggesting that GlnR acts as an activator for p expression and its regulatory activity on p is modulated by the growth pH.

FIG 5 

GlnR positively regulates p. (A) The p-cat activity in S. salivarius MC308_ΔcodY, ΔcodY_mGlnR_box-1, and ΔcodY_mGlnR_box-2. All strains were grown in BHI at pH 7.5 and pH 5.5. (B) The p-cat activity in S. gordonii SL17, SL17_ΔglnR, and SL17_CΔglnR. Cells were grown in BHI at pH 7.5 and pH 5.5. (C) Levels of p activity in chemostat-grown S. gordonii SL17_CΔglnR and SL17_ΔglnR. The specific activities (Sp. Act.) were expressed as indicated in the legend to Fig. 4. The values are the mean results and standard deviations from three independent experiments. Significant differences between strains were analyzed using one-way ANOVA. **, P < 0.01; ***, P < 0.001.

As stated above, the nonureolytic S. gordonii strain seems to be a logical alternative host for studying p regulation, and luckily, a glnR-null S. gordonii derivative could be obtained. Thus, the effect of GlnR on p expression was examined in S. gordonii by using a similar approach as for analyzing VicR regulation. The CAT activities in the glnR-null S. gordonii SL17 (strain SL17_ΔglnR) and its derivative harboring the glnR gene of S. salivarius 57.I (strain SL17_CΔglnR) were examined in batch-grown cells at pH 7.5 and pH 5.5. In agreement with the cis element analysis, the CAT activity in strain SL17_ΔglnR was lower than the levels in strains SL17 and SL17_CΔglnR at both pH 7.5 and pH 5.5, indicating that GlnR activates p expression (Fig. 5B). It was also noticed that the CAT activity of strain SL17_CΔglnR was consistently higher (approximately 40%) than that of strain SL17, suggesting that S. salivarius GlnR works more efficiently on p than S. gordonii GlnR does.

To further investigate the effects of growth pH and carbohydrate concentration on the regulation of GlnR on p, strains SL17_CΔglnR and SL17_ΔglnR were cultivated in a chemostat at pH 7 or pH 5.5 with 20 or 100 mM glucose. At pH 7, comparable expression levels were observed in strains SL17_CΔglnR and SL17_ΔglnR under glucose limitation (20 mM), whereas a 1.8-fold increase in CAT activity was detected in strain SL17_CΔglnR compared to that in strain SL17_ΔglnR under glucose excess (100 mM). At pH 5.5, 4-fold and 3.2-fold increases in CAT activity were seen in strain SL17_CΔglnR compared to the levels in strain SL17_ΔglnR under 20 mM and 100 mM glucose, respectively (Fig. 5C). These results indicated that the activation by GlnR of p was modulated by both carbohydrate concentration and growth pH, and the activation was most pronounced at pH 5.5.

The VicR box acts as a UP element of p.

As the VicR box is rich in AT, it is hypothesized that this region could act as a UP element to enhance the activity of RNA polymerase in the absence of the cognate repressor. Thus, the VicR box in the 5′ flanking region of p was mutated in strain SL17_ΔvicRKX to verify the possibility. Notably, the CodY box remains intact in this mutant strain. Mutations in the VicR box downregulated the CAT activity in strain SL17_ΔvicRKX at both pH 7.5 and pH 5.5 (Fig. 6A), suggesting that this region exhibited a positive effect on p expression in the absence of VicR. To further investigate whether this region could interact with the C-terminal domain (CTD) of the RNA polymerase α subunit (α-CTD), EMSA was carried out with a probe of 21 bp covering the entire VicR box. The result indicated that the MalE-tagged α-CTD recombinant protein interacted with the probe, and the shift was abolished when an unlabeled probe in 300-fold excess was included in the reaction mixture (Fig. 6B), confirming that the VicR box acted as a UP element to enhance p expression.

FIG 6 

Functional analysis of the VicR box as a UP element to enhance p expression. (A) The expression of p in S. gordonii SL17, SL17_ΔvicRKX, and ΔvicRKX_mVicR_box strains was detected in cells grown in BHI at pH 7.5 and pH 5.5. The specific activity (Sp. Act.) was expressed as indicated in the legend to Fig. 4. Values are the mean results and standard deviations from three independent experiments. Significant differences between the results for strains SL17_ΔvicRKX and ΔvicRKX_mVicR_box were analyzed using one-way ANOVA. ***, P < 0.001. (B) EMSA of α-CTD binding to the VicR box in p. Lane 1, probe only; lanes 2 to 6, 0.25 to 4 µM MalE–α-CTD in 2-fold increments; lane 7, 4 µM MalE–α-CTD with a specific competitor. All reactions were carried out with 0.01 pmol biotin-labeled probe.

DISCUSSION

Urease expression in both H. pylori and S. salivarius is upregulated during growth at acidic pH; however, H. pylori activates the expression at acidic pH by the activity of NikR and ArsR (27–29), which is different from the repression by CodY and VicR at neutral pH in S. salivarius. It seems most cost-effective for H. pylori to activate urease expression at acidic pH, as the pH of the stomach is generally below pH 5. On the other hand, the pH of oral mucosal pH is close to neutral normally (30), and therefore, S. salivarius gains the greatest advantage by repressing urease expression at neutral pH. The repression of S. salivarius p by the VicRKX system was most evident in cells cultivated at pH 7 with 20 mM glucose and less evident at pH 5.5, but surprisingly, the repressive effect was absent when cells were cultivated at pH 7 with 100 mM glucose, raising the possibility that the VicRKX system is insensitive to environmental pH under glucose excess. An attempt was made using ChIP-quantitative PCR to verify the DNA binding activity of VicR under different growth pH conditions in batch-grown S. salivarius 57.I. More VicR binding was detected in cells grown at pH 7.5 than at pH 5.5 (data not shown), suggesting that VicR is more active at neutral pH. Thus, the absence of repression of p by VicR at pH 7 under 100 mM glucose may result from the repression of an additional regulatory protein that exerts regulation mainly at neutral pH under glucose excess, and/or the repression by this regulatory protein is augmented in the absence of VicR under this growth condition. Our recent observations suggested that the catabolite control protein A (CcpA) also participates in the regulation of p expression. Inactivation of ccpA led to upregulation of p in cells grown at pH 7 with 100 mM glucose, but only marginal upregulation was seen in cells grown at pH 7 with 20 mM glucose (unpublished data). The regulation by CcpA of p is not understood currently, as no CcpA binding consensus element is found in the flanking region of p. However, the effect of CcpA in response to carbohydrate concentration at pH 7 may explain the absence of upregulation in SL17_ΔvicRKX grown at pH 7 with 100 mM glucose.

It is intriguing that p was also repressed by the VicRKX system at pH 5.5 regardless of the glucose concentration, if neutral pH is required to activate the system. A study in S. gordonii indicates that inactivation of VicR reduces the tolerance of S. gordonii for oxidative stresses (22). Furthermore, studies in S. mutans have suggested that acidic growth pH could induce the oxidative stress response (31, 32). Specifically, the expression levels of genes encoding enzymes metabolizing reactive oxygen species, e.g., sod, ahpC, and ahpF, are upregulated in chemostat-grown S. mutans at pH 5 compared to their expression levels in cells grown at pH 7 (31). Thus, the activity of VicR at pH 5.5 may be part of the oxidative stress response.

The results of site-directed mutagenesis of the GlnR boxes and p expression in chemostat cultures (Fig. 5) suggest that growth pH modulates the activity of GlnR. Although it is not understood how the acidic pH activates the DNA binding activity of GlnR, the activation by acidic pH is not unique to S. salivarius. A recent study in S. mutans demonstrates that the repression of GlnR is activated at pH 5.5 (19). Thus, GlnR of oral streptococci is likely to be activated by both excess amounts of the nutrient nitrogen and acidic growth pH. Additionally, the presence of more than one GlnR box in the promoter region is also not unique to p. For instance, the glnRA operon, the ureABC operon, and tnrA of B. subtilis all possess two GlnR boxes in the promoter regions (7, 18, 33), and cooperative binding of GlnR to the two GlnR boxes, 6 bp apart, has been demonstrated in the promoter of glnRA (33). As both GlnR box 1 and GlnR box 2 are required for GlnR-dependent activation (Fig. 5A), these two sites may participate in cooperative binding of GlnR. Furthermore, the positive regulation of GlnR in p transcription is similar to the activity of TnrA in B. subtilis (26). As a tnrA homolog is absent in most streptococcal species with known genomes (8, 34–36) and studies in Lactococcus lactis and Streptococcus pneumoniae have shown that GlnR carries out some of the functions that are exerted by TnrA in B. subtilis (37, 38), GlnR of S. salivarius is likely to function in the same way.

Studies in S. mutans have demonstrated that chemostat-grown cells under glucose limitation have an excess of amino acid nutrients at a high growth rate (D = 0.4−1) (39). Reduced expression of glnRA was observed in chemostat-grown S. salivarius 57.I supplemented with 20 mM glucose compared to that with 100 mM glucose (data not shown), suggesting that the nutrient nitrogen is likely to be in excess under glucose limitation. If this is true, it is expected that the activation of p transcription by GlnR should be more pronounced under glucose limitation than under glucose excess. However, we did not observe downregulation of p in strain SL17_ΔglnR cultivated at pH 7 under 20 mM glucose, suggesting that repressor(s) which are most active at pH 7 with limited sugar supply could mask the activation by GlnR. Based on what we have learned, the repression is likely to be governed by CodY (15) and VicR (Fig. 4B).

A working model for the regulatory network governing urease expression is proposed (Fig. 7). CodY and VicR inhibit urease expression via binding to the cognate operators in the 5′ flanking region of p at neutral pH to avoid overalkalinization of the oral cavity. GlnR activates p transcription via binding to the GlnR boxes at acidic pH to enhance acid tolerance. In the absence of the cognate repressors, both the CodY and VicR boxes could act as UP elements to enhance p expression. The complex regulation of the urease operon by these regulators links nitrogen metabolism and the acid stress response, which could ensure optimal fitness of S. salivarius against environmental stresses.

FIG 7 

Model for urease regulation in S. salivarius. The relative locations of the −10 and −35 elements are indicated by triangles. The CodY box, VicR box, and GlnR boxes are indicated by black, gray, and white squares, respectively. The model suggests that CodY and VicR repress p expression at pH 7.0, whereas GlnR activates the expression at pH 5.5. In the absence of CodY and VicR, the AT tract in both the CodY box and the VicR box could act as a UP element to enhance p expression.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and general genetic manipulations.

The bacterial strains used in this study are listed in Table 1. S. salivarius 57.I, S. gordonii CH1, and their derivatives were grown routinely in BHI (Difco) at 37°C under 10% CO2 atmosphere. When necessary, kanamycin (Km) was added to the culture medium at 1,000 and 250 µg · ml−1 for recombinant S. salivarius and S. gordonii strains, respectively. When necessary, 750 µg · ml−1 of spectinomycin (Sp) and 5 µg · ml−1 of erythromycin (Em) were used for other recombinant streptococcal strains. Recombinant Escherichia coli strains were maintained in L broth supplemented, where indicated, with Sp at 50 µg · ml−1. To obtain batch cultures grown under neutral or acidic pH conditions, cells were cultivated to mid-exponential phase (optical density at 600 nm [OD600] of ≈0.6) in BHI containing 50 mM KPO4 at pH 7.5 and in BHI that was adjusted to pH 5.5 by the addition of 2 N HCl, respectively. For continuous-culture studies, recombinant S. gordonii strains were grown in a Biostat B plus bioreactor (Sartorius Stedim Biotech) at a dilution rate (D) of 0.3 h−1 (generation time, 2.3 h) in medium containing 3% tryptone and 0.5% yeast extract (TY) (13). Cultures were kept at a specific growth condition for at least 10 generations to reach steady state. Furthermore, when 20 mM glucose was included in the medium, glucose was undetectable in the culture supernatant, whereas approximately 50 mM glucose remained in the culture supernatant when 100 mM glucose was used. Thus, 20 and 100 mM glucose present glucose limitation and excess, respectively.

TABLE 1 

Bacterial strains used in this study

Strain	Resistance	Description	Reference or source
S. salivarius strains
57.I		Wild-type strain	47
MC308	Sp	57.I harboring spe-pureI-cat at lacZ	15
MC308_ΔcodY	Sp, Em	MC308 codY::erm	15
MC308_mVicR_box	Sp	Nucleotides −59 to −64 of pureI in MC308 are mutated	This study
ΔcodY_mGlnR_box-1	Sp, Em	GlnR box 1 of pureI in MC308_ΔcodY is mutated	This study
ΔcodY_mGlnR_box-2	Sp, Em	GlnR box 2 of pureI in MC308_ΔcodY is mutated	This study
S. gordonii strains
CH1		Wild-type strain	42
SL17	Sp	CH1 harboring spe-pureI-cat at gtfG	This study
SL17_ΔvicRKX	Sp, Km	SL17 vicR::Ωkan	This study
SL17_CΔvicRKX	Sp, Km, Em	SL17_ΔvicRKX harboring vicRKX (of 57.I) at gtfG	This study
SL17_ΔglnR	Sp, Em	SL17 glnR::erm	This study
SL17_CΔglnR	Sp, Km	glnR::erm in SL17 is replaced by glnR (of 57.I)-kan	This study
ΔvicRKX_mVicR_box	Sp, Km	Nucleotides −52 to −64 of pureI in SL17_CΔvicRKX are mutated	This study

The oligonucleotides used in this study are listed in Table 2. Synthetic DNA oligonucleotides were purchased from Genomics BioSci & Tech (Taiwan) and Integrated DNA Technologies (Singapore). Restriction enzymes and DNA-modifying enzymes were purchased from New England Biolabs (NEB). PCRs were performed with the high-fidelity DNA polymerase Blend Taq plus (Toyobo).

TABLE 2 

Primers used in this study

Primer	Sequencea	Purpose
57.I_VicR_XhoI_S	GTAACAACTCGAGTGTATTTTACT	PCR for SL17_CΔvicRKX construction
57.I_VicX_SphI_AS	GAGGAATTCCTAGCATGCCTAA
CH1_GtfG_S	GCTAATCAAGTGACCAATG
CH1_GtfG_BamHI_AS	CCAGTTGTTTCGGATCCTGTCTT
CH1_GtfG_XhoI_S	GATAAGACATCTCGAGCCAATTCT
CH1_Spec_AS	CTCTCCAAGATAACTACGAACTGCT
CH1_VicR_S	TTTCAACCATGGAACGCTTCA	PCR for SL17_ΔvicRKX construction
CH1_VicR_XhoI_AS	CATCTCGAGAGCTTCACGACCATCAA
CH1_VicR_BamHI_S	ATCGGATCCGACGACTATGTGACTAAG
CH1_VicR_AS	TGCATATCCAGAACCTCAGA
57.I_GlnR_NcoI_S	AGGCCATGGATGGCAAGAGAGAAAGA	PCR for SL17_CΔglnR construction
57.I_GlnR_BamHI_AS	AAAGGATCCTTAGATACGCAAGTTACCAAA
CH1_GlnR_S	GGAAATGTAACGTATATCTATCCAA
CH1_GlnR_NcoI_AS	CATCCATGGCCTCCTTTCGTAAGATATG
CH1_GlnA_XbaI_S	TGTTCTAGACCATAAGGAGATCACCATGCC
CH1_GlnA2_AS	GCGAGTTCCCATGCCGCGAGATGCTG
CH1_SGO_0212_S	CAAGATAGGATATAGAGGGTGA	PCR for SL17_ΔglnR construction
CH1_GlnR_XhoI_AS	TGCTCGAGTACGACGCAGTT
CH1_GlnR_SphI_S	CTGCATGCAAGCGTAAGTATG
CH1_GlnA1_AS	ACATGCACGTAGAACTTCATCA
EMSA_CTD_21_S	CCTGTAAATGTTGCAAAATCC	EMSA probe for analyzing the binding of α-CTD to VicR box
EMSA_CTS_21_AS	GGATTTTGCAACATTTACAGG
GlnR_PstI_S	AGTTCCTGCAGTTATTAGATACGCAAG	PCR for His-GlnR construction
GlnR_BamHI_AS	GAGGGGATCCATGGCAAGAGAGAAAGA
GlnR_box 1_SalI_S	TGAGTCGACTGTAAATGTTGCAAAATTTC	Inverse PCR to mutate GlnR box 1
GlnR_box 1_SalI_AS	ACAGTCGACTCAAGC GCCGCT GTGGTGTCAA
GlnR_box 2_NsiI_S	TTTATGCATACATGTTAGCTTGACTAATATGTAAATG	Inverse PCR to mutate GlnR box 2
GlnR_box 2_NsiI_AS	TGTATGCATAAAAAAGCTAGTTACATTTGGTAATTCT
GlnR_box-1_S	ACCACATGTTAGCTTGACTAATATGTAAAT	DAPA and EMSA probe for analyzing the binding of GlnR to the GlnR box
GlnR_box-1_AS	ATTTACATATTAGTCAAGCTAACATGTGGT
GlnR_box-2_S	AATGTAATGTCATTTTTTGACACCACATGT
GlnR_box-2_AS	ACATGTGGTGTCAAAAAATGACATTACATT
mGlnR_box-1_S	ACCACAGTCTTATGTAAAGTCATTGTAAAT
mGlnR_box-1_AS	ATTTACAATGACTTTACATAAGACTGTGGT
mGlnR_box-2_S	AATGTAAGTCTTATCTATCCATTGACATGT
mGlnR_box-2_AS	ACATGTCAATGGATAGATAAGACTTACATT
pMAL_VicR_EcoRI_S	AAGGAATTCCTAGTCATAATTCTTCATATA	PCR for MalE-VicR construction
pMAL_VicR_PstI_AS	AAGCTGCAGTGAAATCTCCTGCTTACCAGA
pMAL_GlnR_EcoRI_S	AGGGAATTCATGGCAAGAGAGAAAGAATTAAGGCG	PCR for MalE-GlnR construction
pMAL_GlnR_PstI_AS	TTCCTGCAGTTATTAGATACGCAAGTTACCAAATTG
pureI_4870_S	CGGACTATATTGTCAGAAACAGTC	Used in ChIP-PCR
pureI_5090_AS	CACCTAACATAAGAACCTCCTAAG
pureI_VicR_box_SalI_S	ATAGTCGACTGTTGCAAAATTTCTGAAA	Inverse PCR to mutate VicR box in 57.I
pureI_VicR_box_SalI_AS	ACAGTCGACTATTAGTCAAGCTAACATG
pureI_320_SalI_S	AAAGTCGACCAAATTTCTGAAAATTCGTTG	Inverse PCR to mutate VicR box in CH1
pureI_320_SmaI_AS	TTTGTCGACCCCGGGTATTAGTCAAGCTAA
VicR_BamHI_S	GGTGGATCCATGAAAAAAATTCTAGTAGTT	PCR for His-VicR construction
VicR_PstI_AS	AAGCTGCAGTGAAATCTCCTGCTTACCAGA
VicR_box_S	CTAATATGTAAATGTTGCAAAATTTCTGAA	DAPA and EMSA probe for analyzing the binding of VicR to the VicR box
VicR_box_AS	TTCAGAAATTTTGCAACATTTACATATTAG
mVicR_box_S	CTAATACTAGTGTAATAATAGTATTCTGAA
mVicR_box_AS	TTCAGAATACTATTATTACACTAGTATTAG

Inserted restriction recognition sites and mutated sequences are underlined.

Purification of recombinant proteins and generation of polyclonal antisera.

The coding region of VicR was PCR amplified from S. salivarius 57.I using primers VicR_BamHI_S and VicR_PstI_AS and cloned into pQE30 (Qiagen) in E. coli M15. The identity of the recombinant plasmid was confirmed by sequencing analysis. The recombinant His-tagged VicR (His-VicR) was induced and purified under denaturing conditions using the standard methods. The identity of the recombinant protein was confirmed by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) analysis. The concentration of the purified protein was determined by Bio-Rad protein assay based on the method of Bradford (40). Using a similar approach, the coding region of GlnR was amplified from S. salivarius 57.I by using primers GlnR_PstI_S and GlnR_BamHI_AS, and a His-tagged GlnR protein (His-GlnR) was prepared. The purified His-VicR and His-GlnR were used to generate polyclonal antisera in rabbits by LTK BioLaboratories (Taiwan). The titers and specificities of all antisera were tested by immunoblotting.

To purify recombinant VicR and GlnR under the native condition, constructs to produce maltose binding protein-tagged VicR (MalE-VicR) and GlnR (MalE-GlnR) were generated by using pMAL-c2X (NEB). The recombinant proteins were purified by amylose affinity chromatography (NEB) and verified by MALDI-TOF analysis prior to performing the EMSA analysis.

DAPA and Western blot analysis.

All probes used in DAPA are generated by annealing two biotin-labeled, complementary oligonucleotides. The oligonucleotides were labeled by using the Pierce biotin 3′-end DNA labeling kit. Mid-exponential phase (OD600 of ≈0.6) cultures of S. salivarius 57.I were harvested, washed once with an equal volume of 10 mM Tris (pH 7.6), and then resuspended in 1/100 of the original culture volume in the DAPA binding buffer (60 mM KCl, 10 mM Tris-HCl [pH 7.6], 5% glycerol, 0.5 mM EDTA, 1 mM dithiothreitol [DTT]). Concentrated cell suspensions were subjected to mechanical disruption in the presence of an equal volume of glass beads (0.1 mm in diameter) by homogenization in a BeadBeater (BioSpec Products) for a total of 120 s at 4°C. Amounts of 1 mg and 500 µg of the total lysate were incubated with 20 nM biotin-labeled DNA probes specific for the VicR box and for the GlnR boxes, respectively, in the DAPA binding buffer. The binding reaction was carried out under rotation at 4°C for 1 h. The DNA-protein complexes were captured by using 50 µl of streptavidin MagneSphere paramagnetic particles (Promega). The mixture was incubated at 4°C for 1 h, followed by five washes with the binding buffer. Finally, the proteins of the DNA-protein complexes were eluted in electrophoresis sample buffer, separated on 12% SDS–PAGE, and then detected by immunoblotting. Anti-VicR and anti-GlnR antisera were used at 1:2,000 and 1:20,000 dilution, respectively. The signals were detected by horseradish peroxidase-conjugated anti-rabbit IgG (GeneTex) and luminol-based reagents (Merck Millipore).

EMSA and ChIP-PCR.

Two biotin-labeled, complementary oligonucleotides containing the target site were annealed and used in EMSA. An amount of 0.01 pmol of the annealed probe was incubated with increasing amounts of recombinant MalE-VicR, MalE-GlnR, and MalE–α-CTD in the EMSA binding buffer [50 mM Tris-HCl (pH 7.4), 50 mM KCl, 2 mM MgCl2, 100 ng poly(dI-dC), and 0.5 µg bovine serum albumin (BSA)]. All reactions were carried out at 4°C for 30 min, and the products were resolved on 6% nondenaturing polyacrylamide gels. Specific competition was carried out by including the same probe without labeling in a 300-fold excess. The DNA-protein complex was electrotransferred to a piece of Hybond blotting membrane (Amersham), and the signal was detected by the chemiluminescent nucleic acid detection module kit (Pierce).

ChIP-PCR assay was performed as previously described (15). The PCR was carried out by using primers pureI_4870_S and pureI_5090_AS.

Construction of a p-cat fusion and its derivatives in S. gordonii CH1 and S. salivarius 57.I.

The p-cat fusion was tagged with an Sp resistance gene (spe) (41) and cloned into the integration vector for S. gordonii CH1, pMJB6 (14), which allows the integration of the promoter fusion at gtfG. The resulting plasmid, pSL16, was introduced into S. gordonii CH1 by transformation (42), and the double-crossover recombination was verified by colony PCR. The resulting strain was designated S. gordonii SL17.

The putative VicR box in the p-cat fusion was mutated by site-directed mutagenesis. Briefly, the primer pairs pureI_VicR_box_SalI_S plus pureI_VicR_box_SalI_AS and pureI_320_SalI_S plus pureI_320_SmaI_AS were used in an inverse PCR using pMC300 (15) (for S. salivarius) and pSL16 (for S. gordonii) as the template, respectively. The PCR products were digested, ligated, and established in E. coli. The resulting plasmids were confirmed by sequencing analysis, and the correct constructs were introduced into S. salivarius 57.I by electroporation (12) and into S. gordonii ΔvicRKX by transformation (42). The double-crossover recombination was verified by colony PCR, and the resulting strains are designated S. salivarius MC308_mVicR_box and S. gordonii ΔvicRKX_mVicR_box, respectively.

The GlnR boxes in the p-cat fusion were mutated by a similar approach. Briefly, the primer pairs GlnR_box 1_SalI_S plus GlnR_box 1_SalI_AS and GlnR_box 2_NsiI_S plus GlnR_box 2_NsiI_AS were used in PCR to introduce mutations into GlnR box 1 and 2, respectively.

Construction of the S. gordonii vicRKX-deficient strain and its derivative.

All recombinant S. gordonii strains were generated by using PCR ligation mutagenesis (43). To inactivate vicRKX in S. gordonii SL17, the 5′ and 3′ flanking fragments of vicR were generated from S. gordonii CH1 by using the primer pairs CH1_VicR_S plus CH1_VicR_XhoI_AS and CH1_VicR_BamHI_S plus CH1_VicR_AS, respectively. The PCR products were digested and then ligated to the 5′ and 3′ ends of an Ωkan (44) fragment. The ligation mixture was used to transform S. gordonii SL17, with selection for Km resistance. The double-crossover recombination was verified by colony PCR, and the resulting strain was designated SL17_ΔvicRKX. The region encoding the 41st to 92nd amino acids of VicR in strain SL17_ΔvicRKX was replaced by Ωkan.

To generate a vicRKX recombinant strain, two fragments for integrating the vicRKX gene of S. salivarius at gtfG were generated from SL17_ΔvicRKX by primer pairs CH1_GtfG_S plus CH1_GtfG_BamHI_AS and CH1_GtfG_XhoI_S plus CH1_Spec_AS, respectively. A DNA fragment containing the vicRKX operon was amplified from S. salivarius 57.I by PCR using primers 57I_VicR_XhoI_S and 57.I_VicX_SphI_AS. A DNA fragment containing the nonpolar Em resistance gene (erm) from Tn916ΔE (45), which does not possess a promoter or a transcription terminator, was also prepared by PCR. All PCR products were digested and used in a ligation reaction. The ligation mixture was used to transform strain SL17_ΔvicRKX, and the allelic exchange event in the Em-resistant transformants was verified by colony PCR. The resulting strain, SL17_CΔvicRKX, harbors a copy of erm-tagged vicRKX on gtfG.

Construction of the S. gordonii glnR-deficient and derivative strains.

The glnR gene of S. gordonii CH1 was inactivated by the method described above. Briefly, the 5′ and 3′ flanking fragments of glnR were generated from S. gordonii CH1 by primer pairs CH1_SGO_0212_S plus CH1_GlnR_XhoI_AS and CH1_GlnR_SphI_S plus CH1_GlnA1_AS, respectively. These two fragments were ligated to the 5′ and 3′ ends of a nonpolar erm fragment, and the ligation mixture was used to transform S. gordonii SL17. The double-crossover recombination was verified by colony PCR, and the resulting strain was designed strain SL17_ΔglnR. The region encoding the 12th to 76th amino acids of GlnR was replaced by a nonpolar erm in this strain.

An S. gordonii glnR-derived strain was generated as described above. Briefly, the 5′ and 3′ flanking fragments were generated from SL17_ΔglnR by PCR with the primer pairs CH1_GlnR_S plus CH1_GlnR_NcoI_AS and CH1_GlnA_XbaI_S plus CH1_GlnA2_AS, respectively. A DNA fragment containing glnR was generated from S. salivarius 57.I by PCR using primers 57I_GlnR_NcoI_S and 57.I_GlnR_BamHI_AS. A ligation mixture of all three fragments and a DNA fragment containing a nonpolar kan cassette (44) that lacks a promoter and a transcription terminator was prepared and used to transform strain SL17_ΔglnR. The ligation was prepared to favor the formation of a construct comprising the 5′ flanking fragment followed by glnR of S. salivarius, the kan fragment, and the 3′ flanking fragment. The correct allelic exchange event in the Km-resistant transformants was verified by colony PCR, and the resulting strain was designed SL17_CΔglnR.

CAT assay.

Total protein lysates from the concentrated cell suspensions were subjected to mechanical disruption as described previously (14). The CAT activities were determined by the method of Shaw (46), and the specific activities were calculated as nmol of Cm acetylated min−1 mg−1.

Statistical analysis.

Statistical analysis was performed using one-way analysis of variance (ANOVA) with GraphPad Prism (version 5) software.