Novel Two-Component System of Streptococcus sanguinis Affecting Functions Associated with Viability in Saliva and Biofilm Formation

ABSTRACT

Streptococcus sanguinis is a pioneer species of teeth and a common opportunistic pathogen of infective endocarditis. In this study, we identified a two-component system, S. sanguinis SptRS (SptRS_Ss), affecting S. sanguinis survival in saliva and biofilm formation. Isogenic mutants of sptR_Ss (SKsptR) and sptS_Ss (SKsptS) showed reduced cell counts in ex vivo assays of viability in saliva compared to those of parent strain SK36 and complemented mutants. Reduced counts of the mutants in saliva were associated with reduced growth rates in nutrient-poor medium (RPMI) and increased susceptibility to the deposition of C3b and the membrane attach complex (MAC) of the complement system, a defense component of saliva and serum. Conversely, sptR_Ss and sptS_Ss mutants showed increased biofilm formation associated with higher levels of production of H₂O₂ and extracellular DNA. Reverse transcription-quantitative PCR (RT-qPCR) comparisons of strains indicated a global role of SptRS_Ss in repressing genes for H₂O₂ production (2.5- to 15-fold upregulation of spxB, spxR, vicR, tpk, and ackA in sptR_Ss and sptS_Ss mutants), biofilm formation, and/or evasion of host immunity (2.1- to 11.4-fold upregulation of srtA, pcsB, cwdP, iga, and nt5e). Compatible with the homology of SptR_Ss with AraC-type regulators, duplicate to multiple conserved repeats were identified in 1,000-bp regulatory regions of downstream genes, suggesting that SptR_Ss regulates transcription by DNA looping. Significant transcriptional changes in the regulatory genes vicR, spxR, comE, comX, and mecA in the sptR_Ss and sptS_Ss mutants further indicated that SptRS_Ss is part of a regulatory network that coordinates cell wall homeostasis, H₂O₂ production, and competence. This study reveals that SptRS_Ss is involved in the regulation of crucial functions for S. sanguinis persistence in the oral cavity.

INTRODUCTION

Streptococcus sanguinis is a major commensal pioneer colonizer of tooth surfaces and an important opportunistic pathogen of infective endocarditis (1–4). The establishment of S. sanguinis in the oral cavity occurs with the eruption of primary teeth (1, 5), which are bathed in saliva and gingival crevicular fluid (GCF), a serum transudate of the gingival crevice. There is clinical evidence that robust tooth colonization by S. sanguinis inhibits the establishment of pathogenic competitor species, e.g., the caries pathogen Streptococcus mutans (1, 5). These findings are compatible with the capacity of S. sanguinis to inhibit S. mutans growth through the production of hydrogen peroxide (6). However, the molecular mechanisms by which S. sanguinis competitively initiates tooth colonization are not understood. As a pioneer colonizer of teeth, S. sanguinis needs to interact with components of the dental pellicle (7), an acellular layer formed on tooth surfaces composed of salivary, epithelial, and serum components (8). Pellicle-adhered cells then initiate their sessile growth, affecting local oxidative conditions and producing H₂O₂, which in turn inhibits the growth of competitor species and promotes the release of bacterial DNA, an important component of the extracellular matrix of maturing biofilms (9–11). Therefore, to become established in the oral cavity, S. sanguinis needs to switch the expression of genes required for persistence in oral fluids (whole saliva and GCF) with the expression of genes for biofilm formation during cyclical transitions from planktonic to biofilm modes of growth.

Two-component systems (TCSs) are major bacterial sensory systems required for efficient physiological responses to environmental and host stimuli (12–14). Classical TCSs are generally formed by a transmembrane histidine kinase (HK) sensor protein, which undergoes autophosphorylation in response to an environmental stimulus and transmits the input signal to a cognate intracellular response regulator (R) through phosphoryl transfer reactions. Afterwards, activated R directly interacts with the regulatory regions of the downstream genes required for physiological adaptation to the sensed stimuli (12, 13). In this study, we investigate the role of a TCS orthologous to the TCS SptRS_Spy (Spt for salivary persistence) of Streptococcus pyogenes (15) in the capacity of S. sanguinis to survive in human saliva and to form biofilms by using phenotypic comparisons of isogenic mutants of S. sanguinis SptRS (SptRS_Ss)-encoding genes (sptS_Ss and/or sptR_Ss) with parent strain SK36 and the respective complemented mutants. Functions potentially involved in the altered phenotypes were then screened by assessing planktonic growth under different nutritional and oxygen availabilities, sensitivity to oxidative stress, autolysis, the production of H₂O₂ and extracellular DNA (eDNA), and susceptibility to the deposition of C3b and C5b-C9 (membrane attack complex [MAC]) of the complement system, an immune component of saliva, GCF, and serum (16, 17). A comparative analysis of the transcriptional profiles between mutant, parent, and complemented strains was then performed to identify SptRS_Ss downstream genes. Finally, DNA-protein binding assays and bioinformatic analyses of the promoter regions of the transcriptionally affected genes revealed potential mechanisms of gene regulation by the TCS SptRS_Ss.

RESULTS

Analyses of sptRS_Ss locus structure, SptRS_Ss domain architecture, and sptRS_Ss inactivation.

BLASTP analysis against the genome of S. sanguinis strain SK36 revealed an orthologue of the TCS SptRS_Spy encoded by Spy_0874 and Spy_0875, which regulates S. pyogenes virulence and persistence in human saliva (15). As shown in Fig. 1, SSA_1974 and SSA_1973 encode proteins with 81% and 72% similarities with SptR_Spy and SptS_Spy, respectively. SptR_Ss and SptS_Ss also show 83% and 74% similarities with proteins encoded by SMU.927 and SMU.928 of S. mutans strain UA159, respectively (Fig. 1). SMU.927 and SMU.928 were designated RelR and RelS, respectively, in a previous study (18). No SptRS_Spy orthologues were found in Streptococcus pneumoniae strains (data not shown). As for sptR_Spy (Spy_0874) and S. mutans relR (relR_Sm) (SMU_927), sptR_Ss is downstream and in the same orientation as that of a gene encoding a putative RelA/SpoT-like ppGpp synthetase (SSA_1795) (Fig. 1A). Therefore, to investigate if sptR_Ss and sptS_Ss were transcribed as part of an operon with SSA_1795, we performed reverse transcription-PCR (RT-PCR) analysis using primer sets to amplify sequences spanning sptR_Ss and sptS_Ss sequences. Amplicons of 1,555 bp obtained from SK36 cDNA with primers sptR/P1 and sptR/P4 (see Table S1 in the supplemental material) indicated that sptR_Ss and sptS_Ss are cotranscribed with SSA_1795. Negative and positive controls were the same as those used for RT-quantitative PCR (RT-qPCR) analysis.

FIG 1.

Genomic context, protein similarities, and domain architecture of SptRS_Ss of S. sanguinis. (A) Schematic representation of the sptRS chromosomal loci in S. sanguinis, S. pyogenes, and S. mutans strains. Gene open reading frames (ORFs) are represented by arrows to indicate the direction of transcription; locus tags are indicated above each respective arrow. Dark gray arrows represent genes encoding SptR and SptS orthologues. Light gray arrows represent genes encoding putative RelA/SpoT-like ppGpp synthetases. (B) Results of BLASTP analyses of SptS_Ss and SptR_Ss. Percent amino acid (aa) identities and similarities were determined by using SK36 sequences as a reference. (C) Schematic representations of domain architectures of SptS_Ss (left) and SptR_Ss (right) obtained with the SMART research tool (http://smart.embl-heidelberg.de/). Black rectangles at the N-terminal part of SptS_Ss represent transmembrane domains. Intracytoplasmic domains at the SptS_Ss C-terminal part include the dimerization/phosphoacceptor domain (HisKA) and the catalytic/ATP binding domain (HATPase_c). The schematic representation of SptR_Ss shows the relative positions of the N-terminal receiver (REC) domain and of the C-terminal transcriptional regulatory domain (Trans reg_C).

Analysis of the domain architecture of SptS_Ss suggested two transmembrane (TM) domains at the N-terminal region (positions 7 to 29 and 160 to 182), which likely delimitate an extracellular sensor region of 130 amino acids (Fig. 1C). The cytoplasmic HisKA domain (dimerization/histidine phosphotransfer domain) is located at positions 196 to 258 of SptS_Ss. This conserved domain is linked to the catalytic and ATP-binding C-terminal domain (HATPase_c [histidine kinase-like ATPase C-terminal domain]) (at positions 302 to 411) by a linker of 43 amino acids. No additional accessory domains were detected in SptS_Ss. Domain architecture analysis of SptR_Ss indicated a receiver (REC) domain (positions 2 to 113) linked to a C-terminal DNA-binding output domain (Trans_reg_C) (positions 145 to 222) (Fig. 1C). Conserved-domain analysis further revealed that SptR_Ss is within the family of AraC-type regulators.

Isogenic mutants of sptR_Ss, sptS_Ss, and sptRS_Ss could be generated in SK36 via double-crossover recombination with the sptR-, sptS-, or sptRS-null alleles, respectively, indicating that the TCS SptRS_Ss is not essential for the viability of SK36. However, phenotypic screening of these mutants indicated some inconsistencies between the double mutant and the single mutants (data not shown). Transcriptional profiling of the strains revealed several genes affected in the sptR_Ss and in the sptS_Ss single mutants, which were not affected in the double mutant, in a way compatible with their respective phenotypes (data not shown). These findings indicate that the nonpolar deletion of the entire TCS SptRS_Ss might induce compensatory activities of other transcriptional regulators. Therefore, the double mutant was not further analyzed. The transcriptional activity of SSA_1795 was not affected in the three mutants.

Inactivation of sptS_Ss or sptR_Ss reduces survival of SK36 in human saliva.

The SK36 strain was able to growth in the reference saliva or in pooled saliva (Fig. 2) to counts of >10¹⁰ CFU/ml during the initial 48 h of incubation. Cell counts of SK36 remained relatively constant during the following 2 to 4 days of incubation in samples of whole saliva (Fig. 2A and C), indicating a good capacity of SK36 to persist in this host fluid. Differently, the two isogenic mutants (SKsptR and SKsptS) showed significant decreases in viable counts (Fig. 2A and C) compared to those of the parent strain (P < 0.05). Importantly, the capacities to grow or to survive in saliva suspensions were restored in the complemented mutants (Fig. 2B and D). In addition, light microscopy analysis of Gram-stained saliva suspensions confirmed that reductions in cell counts observed in the mutant strains were not a result of changes in chain formation or in bacterial aggregation mediated by saliva (data not shown). Therefore, the deletion of sptS_Ss and sptR_Ss significantly reduces viable counts of S. sanguinis bacteria in human saliva.

FIG 2.

Comparisons of ex vivo survival in human saliva between parent strain SK36 and sptR_Ss and sptS_Ss mutants (A and C) or complemented strains (+) (B and D). Numbers of viable bacteria in suspensions obtained with a reference saliva sample (A and B) or with pools of saliva from six volunteers (C and D) were analyzed during 6 to 14 days. Dots represent means of triplicates from one representative experiment. Bars indicate standard deviations. Symbols indicate significant differences in cell counts of SKsptS (*), SKsptR (†), SKsptRS (‡), SKsptS⁺ (#), or SKsptR⁺ (§) compared to SK36 (P < 0.05 by a Kruskal-Wallis test with Dunnet's post hoc test).

Deletion of sptR_Ss or sptS_Ss affects planktonic growth in nutrient-poor medium and chain formation under specific atmospheric conditions.

During initial colonization of tooth surfaces, saliva is likely the major nutrient source supporting S. sanguinis growth. Saliva has low concentrations of simple sugars and a limited availability of other free compounds, including trace metals (19). Thus, we hypothesized that reduced viable counts of sptR_Ss and sptS_Ss mutants in saliva could be associated with their reduced capacity to adapt to nutrient-limited conditions. Therefore, we compared the growth curves of sptR_Ss- and sptS_Ss-defective strains with those of the parent and complemented strains in complex medium (brain heart infusion [BHI] medium) and in nutrient-poor medium (RPMI) under conditions of aerobiosis (rotational aeration) or 10% CO₂ (static incubation). When grown in BHI medium, the two mutant strains showed similar growth curves compared to that of the parent strain under either aerobiosis (Fig. 3A) or 10% CO₂ (Fig. 3C); slight reductions in the growth rates of sptR_Ss and sptS_Ss mutants were observed only under 10% CO₂ (Fig. 3C). Small reductions in the growth rates of the complemented strains SKsptS⁺ and/or SKsptR⁺ compared to that of the parent strain were also observed under the two atmospheric conditions (Fig. 3A and C) (P < 0.05), likely because BHI medium was supplemented with spectinomycin for the maintenance of plasmids used for complementation. On the other hand, significant reductions in the growth rates of sptR_Ss and sptS_Ss mutants were observed in RPMI under aerobiosis (Fig. 3B) and, more evidently, under 10% CO₂ (Fig. 3D). Therefore, the deletion of sptS_Ss or sptR_Ss reduces S. sanguinis growth rates under nutrient-limited conditions.

FIG 3.

Comparisons of growth curves and chain formation between parent strain SK36 and sptR_Ss and sptS_Ss mutants or complemented strains (+). (A to D) Growth curves were obtained in complex BHI medium (A and C) and in RPMI nutrient-poor medium (B and D) under aerobic shaking (A and B) or 10% CO₂ (C and D). Dots represent means of triplicates from one representative experiment. Bars indicate standard deviations. Statistically significant differences in relation to SK36 at each time point are indicated with symbols for SKsptS (*), SKptR (†), SKsptS⁺ (#), or SKsptR⁺ (‡) (P < 0.05 by a Kruskal-Wallis test). (E) Representative micrographs of strains in the late log phase of growth (BHI medium with 10% CO₂). The top panels show Gram-stained cells under a light microscope (total magnification, ×1,000). The bottom panels show cells stained with the Live/Dead BacLight system under a confocal microscope (Leica system; objective HCX PL APO lambda 63.0× 1.4 oil).

We also analyzed cell chaining and morphogenesis of the strains at mid-log and late log phases of growth in BHI medium under aerobiosis or 10% CO₂. No significant changes in chain formation were observed among strains grown aerobically (data not shown). Under 10% CO₂, significant changes in cell chaining were observed in the sptR_Ss and sptS_Ss mutants only when these strains reached the late log phase of growth. These phenotypes were further restored in the complemented mutant SKsptR⁺ or SKsptS⁺ (Fig. 3E). Quantitative comparisons of chain formation between strains at this growth phase under 10% CO₂ were performed by determining the number of cocci per chain in a total of 50 chains of each strain (10 chains/field; 5 fields analyzed). This analysis revealed increased numbers of cocci per chain for SKsptS (mean, 20.2 ± 6.1 cocci per chain) and SKsptR (mean, 38.7 ± 5.2) compared to those for the parent strain (mean, 9.2 ± 3.0) (P < 0.05 by a Mann-Whitney test) or to the complemented mutants SKsptS⁺ (mean, 8.0 ± 1.26) and SKsptR⁺ (mean, 10.1 ± 2.37) (P < 0.05 by a Mann-Whitney test). Fluorescence microscopy analysis of strains stained by using the Live/Dead BacLight system showed that the long chains of SKsptR and SKsptS were formed by viable cells, as indicated by the absence of staining by propidium iodide, a red fluorescent nucleic acid stain which penetrates only cells with a damaged membrane (Fig. 3E). Therefore, the deletion of sptR_Ss or sptS_Ss affects chain formation only in the late log phase of growth under specific atmospheric conditions (10% CO₂ and static incubation).

Deletion of sptS_Ss or sptR_Ss does not significantly influence S. sanguinis sensitivity to oxidative stress but increases its susceptibility to complement deposition.

As a pioneer colonizer of teeth, S. sanguinis needs to cope with oxidative stresses at saliva-bathed surfaces (20). We thus investigated the effects of the sptR_Ss or sptS_Ss deletion on the sensitivity of S. sanguinis to oxidative stress by comparing reductions in viability of strains after exposure to high doses of H₂O₂. No significant changes were detected in the sensitivities of the sptR_Ss and sptS_Ss mutants to H₂O₂ compared to that of the parent strain when strains were grown aerobically (data not shown). Because S. sanguinis does not produce H₂O₂ in the absence of O₂, we further assessed the sensitivity of strains grown anaerobically to oxidative stress. No changes in sensitivity to oxidative stress were detected in the mutant strains under these atmospheric conditions as well (data not shown).

The complement system plays important innate defense functions in host tissues and fluids, including saliva (16) and gingival crevicular fluid (21, 22). In addition, the deletion of genes involved in complement evasion impaired S. pyogenes persistence in whole saliva (15). Therefore, we investigated the effects of sptR_Ss and sptS_Ss inactivation on the susceptibility of S. sanguinis to the deposition of C3b, a central effector molecule derived from complement activation (23), and the MAC, a pore-like protein complex which binds to bacterial membranes (24, 25). Significant increases in C3b and MAC deposition were detected in SKsptR (136.9% and 118.5% increases, respectively) and SKsptS (49.9% and 99.8% increases, respectively) compared to that of the parent strain (P < 0.05) (Fig. 4). Additionally, levels of C3b and MAC binding were restored in the complemented mutants. Thus, the TCS SptRS_Ss influences the susceptibility of S. sanguinis to complement deposition.

FIG 4.

Effects of inactivation of sptR_Ss and sptS_Ss on S. sanguinis susceptibility to complement deposition. Intensities of C3b (A) and the MAC (B) bound to the bacterial surface were determined by flow cytometry (MFI) in strains previously treated with 20% human serum. Columns represent means of data from three independent experiments; bars indicate standard deviations. Asterisks indicate significant differences in relation to the parent strain (P < 0.05 by a Kruskal-Wallis test with Dunnet's post hoc test).

Deletion of sptS_Ss or sptR_Ss increases biofilm formation on saliva-coated surfaces.

The switch from planktonic to biofilm growth is a major mechanism by which bacteria face host defenses and/or environmental stresses (26). Thus, we investigated the influence of the TCS SptRS_Ss on the capacity of S. sanguinis to form biofilms. sptR_Ss and sptS_Ss mutants were compared with parent or complemented strains regarding their ability to form biofilms on surfaces previously conditioned with saliva. After 18 h of biofilm formation, 77.2% and 40% increases in biofilm biomasses were observed in the sptS- and sptR-defective strains, respectively, compared to the biofilm biomass of SK36 (Fig. 5). In addition, amounts of biofilm formed by the complemented sptS_Ss⁺ and sptR_Ss⁺ mutants did not significantly differ from that of SK36 biofilms (Fig. 5). The absorbances (A₅₅₀) of biofilm culture fluids were similar among the tested strains (data not shown), confirming that differences in biofilm biomasses were not a result of differences in growth yields.

FIG 5.

Biofilm formation by S. sanguinis strains. Shown are mean biomasses of biofilms (A₅₇₅) formed during 18 h on saliva-coated surfaces in the absence or presence of DNase I. Columns represent means of data from three independent experiments performed in triplicate. Bars indicate standard deviations. Asterisks indicate significant differences in relation to the parent strain under the same conditions (P < 0.05 by a Kruskal-Wallis test with Dunn's post hoc test).

S. sanguinis depends on the production of eDNA to form biofilms (9, 11). Therefore, we assessed the effect of medium supplementation with DNase I on the formation of biofilms by the tested strains. The biomasses of biofilms formed in medium supplemented with DNase I by all the tested strains were reduced compared to those of biofilms formed in medium not supplemented with this enzyme (P < 0.05 by a Kruskal-Wallis test with Dunn's post hoc test) (Fig. 5). However, medium supplementation with DNase I did not significantly reduce differences in biofilm formation between the sptR_Ss and sptS_Ss mutants and the parent strain (Fig. 5), suggesting that increased biofilm formation by the sptR_Ss and sptS_Ss mutants does not rely exclusively on the production of eDNA. As expected, treatment with heat-inactivated DNase I did not significantly affect biofilm formation (data not shown).

Deletion of sptS_Ss or sptR_Ss augments production of eDNA and hydrogen peroxide but does not affect autolysis.

Extracellular DNA is a major structural component of S. sanguinis biofilms (11), and its release depends on the production of H₂O₂ (9). Because the sptR_Ss and sptS_Ss mutants showed an enhanced capacity to form biofilms, amounts of eDNA and H₂O₂ in culture fluids were compared between mutant and parent or complemented strains. As shown in Fig. 6A, the SKsptS and SKsptR mutants produced 2- and 2.7-fold more eDNA, respectively, than did the parent strain (P < 0.05 by a Kruskal-Wallis test with Dunn's post hoc test). Consistently, significantly higher concentrations of H₂O₂ were detected in culture supernatants of these mutants (4.4- and 5.6-fold increases in SKsptS and SKsptR, respectively; P < 0.05) (Fig. 6B). In addition, the productions of eDNA and H₂O₂ were restored in the complemented strains SKsptS⁺ and SKsptR⁺ (Fig. 6B). To investigate whether the increased production of eDNA and H₂O₂ in these mutants was associated with altered autolytic activity, temperature-induced autolysis was compared between strains. No clear changes in autolysis were observed in these mutants compared to SK36 during 72 h of incubation (data not shown).

FIG 6.

Comparative analysis of the production of eDNA and H₂O₂ between sptR_Ss and sptS_Ss mutants or complemented (+) strains and parent strain SK36. (A) Amounts of eDNA in culture supernatants of strains grown in BHI medium (aerobiosis) until the late log phase (A₅₅₀ of 0.7) were determined by qPCR. (B) Amounts of H₂O₂ in cultures of strains grown under the same conditions were determined. Columns represent means of data from three independent experiments performed in triplicate. Bars indicate standard deviations. Asterisks indicate significant differences in comparison to SK36 (P < 0.05 by a Kruskal-Wallis test).

Deletion of sptR_Ss or sptS_Ss affects transcription of genes required for biofilm formation, interaction with host components, and/or evasion of host immune factors.

By comparing the transcript levels of genes potentially involved in the sptR_Ss and sptS_Ss mutant phenotypes, we detected significant changes in the activities of a large set of genes implicated in biofilm formation in sptS- and sptR-defective strains (P < 0.01 to 0.05), whereas these genes were not affected in the complemented strains (P > 0.05) (Table 1). The sptS_Ss and sptR_Ss mutants showed a strong upregulation of spxB, encoding pyruvate oxidase (SpxB) (15.4- and 7-fold increases in transcript levels in SKsptS and SKsptR, respectively; P < 0.01). SpxB converts pyruvate, inorganic phosphate, and O₂ to H₂O₂, CO₂, and acetyl phosphate (AcP) (27). In addition, genes associated with the induction of spxB transcription (spxR, tpk, ackA, and vicR) (11, 28) were upregulated in both mutant strains (P < 0.01) (Table 1). spxR encodes a putative direct regulator of spxB, and ackA encodes the acetate kinase for the synthesis of ATP from AcP (27, 28). tpkA encodes a thiamine pyrophosphokinase, which is likely required for providing a cofactor (thiamine pyrophosphate) for SpxB (28). The sptR_Ss and sptS_Ss mutants also showed a significant upregulation of vicR, encoding the VicR response regulator of the TCS VicRK_Ss, which is required for biofilm formation in S. sanguinis (11). Consistently, besides spxB, additional genes downstream of VicR_Ss were also upregulated in the sptR_Ss- and sptS_Ss-defective strains, including SSA_0094, cwdP, and pcsB (P < 0.05) (Table 1). Both CwdP and PcsB are putative murein hydrolases of S. sanguinis potentially associated with the autolysis-independent production of eDNA (11). Interestingly, significant downregulations of genes involved in early (comE) and late (comX, encoding the sigma factor σ^X) competence development were detected in sptR_Ss and sptS_Ss mutants (Table 1). comX downregulation in SKsptS was further associated with the upregulation of mecA, which encodes an adaptor protein (MecA) involved in the degradation of σ^X (29). The sptR_Ss and sptS_Ss mutants did not show significant changes in transcriptional levels of SSA_1795 or of other ppGpp synthetase-encoding genes present in the SK36 genome (SSA_1210 and relA), except for a 1.54-fold upregulation of SSA_1210 in SKsptS (Table 1). In addition, limited changes in the expression levels of metabolic genes (SSA_2265 and SSA_0886) were detected in the mutant strains (Table 1). Therefore, the deletion of the sptR_Ss and sptS_Ss genes has a broad effect on the activity of genes associated with the production of H₂O₂ and eDNA.

TABLE 1.

Transcription profiles of the sptR and sptS isogenic mutants and complemented mutants relative to parental strain SK36 at the late log phase of growth

Function and locus tag (NCBI)	Encoded protein	Mean fold change in transcript levels (SD)a
		SKsptS	SKsptR	SKsptS+	SKsptR+
Production of hydrogen peroxide
SSA_0192	AckA	+4.90 (0.58)†	+1.35 (0.08)	+1.32 (0.44)	+1.40 (0.43)
SSA_2118	TpK	+5.71 (0.42)†	+4.46 (0.33)†	+1.18 (0.45)	+1.24 (0.55)
SSA_0391	SpxB	+15.36 (4.66)†	+6.98 (1.87)†	+2.97 (3.14)	+1.50 (1.41)
SSA_1492	SpxR	+5.81 (0.80)†	+3.64 (0.86)‡	+1.26 (0.66)	+1.22 (0.64)
SSA_1576	CcpA	+1.10 (0.20)	+1.27 (0.12)	+1.09 (0.28)	+1.14 (0.26)
Murein hydrolases
SSA_0094	LysM	+1.31 (0.37)	+3.63 (0.53)†	+1.29 (0.39)	+1.45 (0.77)
SSA_0304	CwdP	+2.64 (0.28)†	+2.23 (0.05)†	+1.09 (0.30)	+1.23 (0.42)
SSA_0019	PcsB	+3.84 (1.11)†	+2.05 (0.47)‡	+1.38 (0.75)	+1.42 (0.61)
Synthesis of extracellular glucan
SSA_0613	GtfP	+1.16 (0.05)	+1.18 (0.37)	0.98 (0.21)	+1.01 (0.24)
Proteases
SSA_1882	PrtS	+1.16 (0.05)	+1.18 (0.37)	−1.08 (0.29)	−1.15 (0.42)
SSA_0263	PepO	−1.03 (0.18)	+1.31 (0.30)	+1.06 (0.30)	+1.28 (0.23)
SSA_0331	CppA	+1.59 (0.07)‡	+1.17 (0.38)	−1.19 (0.09)	+1.27 (0.07)
SSA_1106	IgA1 protease	+11.42 (2.93)†	+2.14 (1.26)	+1.77 (1.37)	+2.19 (1.26)
SSA_2381	HtrA	+1.13 (0.12)	−1.78 (0.10)‡	+1.05 (0.21)	+1.15 (0.12)
Nucleases and nucleotidases
SSA_1234	Nt5e	+3.82 (0.61)†	+4.73 (0.57)†	+1.16 (0.23)	+1.67 (0.50)
SSA_1750	NucH	+1.16 (0.04)	+1.05 (0.19)	−1.05 (0.05)	+1.22 (0.03)
SSA_2116	Cbf	+2.04 (0.25)†	+1.38 (0.26)	+1.22 (0.07)	+1.05 (0.19)
Binding to host and other components
SSA_1219	SrtA	+7.24 (0.67)†	+4.47 (0.52)†	+1.07 (0.45)	−1.06 (0.06)
SSA_1663	CpbA	+2.94 (0.16)†	+1.48 (0.28)	+1.16 (0.08)	+1.18 (0.13)
SSA_0904	CshA-like surface protein A	−1.86 (0.26)‡	−1.76 (0.16)‡	−1.11 (0.16)	−1.23 (0.15)
Metabolic enzymes and transporters
SSA_2265	MalP	+1.08 (0.17)	+1.42 (0.13)	0.95 (0.19)	+1.05 (0.14)
SSA_0287	GldA	+6.68 (2.24)†	+1.73 (0.51)	+1.21 (0.42)	+1.15 (0.42)
SSA_0886	Enolase	+1.74 (0.07)‡	+1.62 (0.27)	+1.10 (0.28)	+1.21 (0.17)
SSA_0260	SsaB	+1.92 (0.12)‡	+3.10 (0.24)†	+1.21 (0.22)	+1.10 (0.33)
Competence
SSA_0016	ComX	−4.76 (0.21)†	−33.55 (2.34)†	−1.71 (0.58)	−2.21 (1.64)
SSA_2378	ComE	−1.46 (0.02)‡	−3.70 (0.18)†	−1.28 (0.16)	−1.02 (0.11)
SSA_1958	MecA	+4.01 (0.38)†	+1.58 (0.31)	+1.33 (0.13)	+1.45 (0.29)
Components of the sptRS locus and ppGpp synthetases
SSA_1793	SptS		+1.23 (0.35)	+1.59 (0.71)	+1.26 (0.06)
SSA_1794	SptR	+1.34 (0.27)		+1.44 (0.27)	+1.61 (0.47)
SSA_1795	ppGpp synthetase	+1.29 (0.38)	+1.11 (0.20)	−1.43 (0.30)	−1.33 (0.18)
SSA_1210	ppGpp synthetase	+1.54 (0.09)‡	+1.39 (0.11)	−1.11 (0.09)	−1.08 (0.07)
SSA_0250	RelA	+1.26 (0.13)	−1.02 (0.15)	−1.16 (0.13)	−1.07 (0.09)
Two-component system VicRKX and CovR
SSA_1563	VicX	+1.54 (0.17)‡	+1.35 (0.14)	−1.26 (0.10)	+1.01 (0.07)
SSA_1564	VicK	+1.65 (0.17)‡	+1.06 (0.06)	+1.26 (0.07)	−1.08 (0.04)
SSA_1565	VicR	+2.33 (0.11)†	+1.86 (0.09)‡	+1.07 (0.07)	+1.18 (0.08)
SSA_1810	CovR	+1.55 (0.16)‡	+1.46 (0.14)‡	−1.05 (0.06)	+1.15 (0.08)

^aValues represent means (standard deviations) from four independent experiments performed in duplicate. †, P < 0.01 by ANOVA with a Dunnett post hoc test; ‡, P < 0.05 by ANOVA with a Dunnett post hoc test.

Besides genes required for H₂O₂ and eDNA production, the gene encoding sortase A (srtA) was 4.4- and 7.2-fold upregulated in SKsptS and SKsptR, respectively (P < 0.01) (Table 1). In addition, a >10-fold upregulation of iga (SSA_1106) was observed in the sptS-defective strain (P < 0.01) (Table 1). iga encodes immunoglobulin A1 (IgA1) protease, which has the sortase A recognition motif LPxTG. Another sortase A-anchored protein, the Nt5e ecto-5′-nucleotidase (SSA_1234), which is likely involved in the evasion of cellular immune responses through ATP hydrolysis (30), was transcriptionally upregulated in sptR_Ss and sptS_Ss mutants (3.8- and 4.7-fold upregulation, respectively; P < 0.05). On the other hand, significant reductions in transcript levels of SSA_0904, encoding an LPxTG-containing CshA-like protein, were observed for the mutant strains (1.8- to 1.9-fold reductions; P < 0.05). CshA contributes to the surface hydrophobicity of S. sanguinis (31, 32). Proteases potentially involved in the evasion of complement immunity by S. sanguinis were also analyzed, including pepO (encoding a putative endopeptidase O), prtS (encoding a putative serine protease with a domain of streptococcal C5a peptidase), and cppA (encoding a putative C3-degrading proteinase) (SSA_0331). No significant changes in the activities of pepO and prtS were detected in the mutant strains, but cppA was upregulated 1.6-fold (P < 0.05) in SKsptS. These findings reveal that the TCS SptRS is involved in the transcriptional regulation of several cell surface proteins involved in S. sanguinis evasion of host immune components.

Finally, to investigate mechanisms by which the deletion of sptR_Ss or sptS_Ss affects the transcription of downstream genes, we applied electrophoretic mobility shift assays (EMSAs) and in silico analyses of the regulatory regions of genes downstream of SptRS_Ss. EMSAs were performed with DNA fragments amplified from the promoter regions of potential target genes (spxB, spxR, cwdP, ssaB, and comX) (approximately 250 to 400 bp upstream of the translation start site [TSS]) (see Table S1 in the supplemental material) and with rSptR_Ss purified under denaturing and hybrid conditions. However, we could not detect shifts in the migration of the tested DNA fragments incubated with a wide range of protein concentrations under several binding conditions (data not shown), whereas the control protein recombinant CovR_Sm (rCovR_Sm) promoted clear shifts in the migration of the covR_Sm DNA fragments (used as a positive control) (33). Based on the homology of SptR_Ss with AraC-type regulators and on transcriptional data indicating that SptR_Ss functions mostly as a negative regulator (Table 1), we examined the possibility that SptR_Ss interacts with conserved binding sites located at a position more distant from the TSS of potential target genes. Bacterial transcriptional repressors (including AraC- and DeoR-type regulators) typically bind to duplicate (or multiple) regions at regions more distant from the TSS, promoting DNA loops that affect the binding of RNA polymerase and/or transcription elongation (34–36). In silico analysis of 1,000-bp sequences upstream of the TSSs of genes significantly upregulated in the sptR_Ss and sptS_Ss mutants (spxB, spxR, pcsB, iga, and nt5e) revealed conserved direct repeats of 8 nucleotides [(A/T)GAA(A/T)(A/T)N(A/T/G)], which resemble binding sites identified for AraC-type regulators (37) (Fig. 7). In addition, these putative binding sites were located at positions compatible with DNA loop formation, i.e., spaced by sequences ranging from 150 to 425 bp (34) (Fig. 7).

FIG 7.

Conserved direct repeats identified in the regulatory regions of genes significantly upregulated in the sptR_Ss and sptS_Ss mutants. (A) Sequences and positions of the conserved repeats in relation to the translation start site of the respective genes. (B) Sequence logo for the putative SptR binding sites generated by using WebLogo (http://weblogo.berkeley.edu/).

DISCUSSION

The prominent roles of S. sanguinis as an initiator of tooth colonization (5, 10) and an opportunistic pathogen of systemic infections (3, 4, 38) imply the expression of functions to cope with innate and adaptive immune host factors of saliva and/or serum. In this study, we established that the TCS SptRS_Ss modulates the viability of S. sanguinis in saliva and biofilm formation. Although SptRS_Ss shows homology to the TCS RelRS of S. mutans (18), we have preliminary results showing that the deletion of relR_Sm in S. mutans UA159 does not affect viable counts or biofilm formation of S. mutans in saliva but promotes the significant upregulation of ppGpp synthetase-encoding genes (relP_Sm, relA_Sm, and relQ_Sm) (39), which is consistent with data from a previous report (18). As for relRS, sptR_Ss and sptS_Ss are associated with SSA_1795, which is likely involved in the stringent response, because it encodes a putative ppGpp synthetase. Here we show that SSA_1795 is cotranscribed with sptRS_Ss, but the deletion of sptR_Ss and sptS_Ss does not affect the transcription of ppGpp synthetase genes (SSA_1795, relA, and SSA_1210), except for a 1.5-fold upregulation of SSA_1210 in SKsptS. Therefore, the sptR_Ss-sptS_Ss-SSA_1795 operon might be activated in response to nutrient limitations capable of inducing stringent responses, although SptRS_Ss does not regulate the synthesis of ppGpp alarmones.

The role of SptRS_Spy in bacterial persistence in whole saliva was established in a S. pyogenes serotype M1 strain and involved the coordination of multiple gene functions for nutrient acquisition, the response to oxidative stress, and the evasion of host immune components (15). The deletion of sptR_Ss or sptS_Ss in S. sanguinis SK36 not only reduces viable counts in saliva but also enhances biofilm formation, suggesting that SptRS_Ss modulates the transition from the planktonic state in saliva to a biofilm state. The reduced viability of the mutant strains in saliva was associated with reduced growth under nutrient-limited conditions (RPMI) but not with increased sensitivity to oxidative stress. On the other hand, the deletion of sptR_Ss and sptS_Ss increased the susceptibility of S. sanguinis to the deposition of effector molecules of the complement system (C3b and the MAC), suggesting that S. sanguinis expresses functions for complement evasion regulated by the TCS SptRS_Ss.

In S. pyogenes, the deletion of sptR_Spy significantly affects the expression of speB (for streptococcal pyrogenic exotoxin B) and sic (for streptococcal inhibitor of complement-mediated lysis) (15). Both speB and sic are required for the persistence of S. pyogenes in whole saliva and are involved in complement evasion (40). Serum is classically considered the major source of complement proteins synthesized in the liver, but these proteins are also produced in other host tissues and are found to be active in saliva and dental tissues (16, 41, 42) and to be incorporated into dental pellicles (43). The genome of S. sanguinis SK36 does not harbor speB or sic orthologues (44), and there is limited information about mechanisms by which S. sanguinis copes with innate defense components present in saliva or serum. Although Gram-positive bacteria are not lysed by the MAC of the complement, deposition of the MAC in S. pyogenes occurs at regions of the division septum (24), and there is evidence indicating that the surface-bound MAC cooperates with antimicrobial peptides present in host fluids for controlling Gram-positive bacteria (24, 45, 46). An in silico screening of the genome of SK36 revealed three genes encoding putative proteases of the complement (prtS, pepO, and cppA). However, none of these genes were significantly affected in the sptR_Ss- and sptS_Ss-defective strains at the late log phase of growth, suggesting that S. sanguinis expresses additional functions for complement evasion.

Streptococcal pathogens typically express surface proteins that bind host components for colonization and/or evasion of host immune functions (47, 48). The sptR_Ss and sptS_Ss mutants show a significant upregulation of sortase A (srtA), which catalyzes the covalent linkage of LPxTG-containing precursor proteins to the cell wall, implying a role of SptRS_Ss in cell surface decoration. Because srtA is required for the binding of S. sanguinis to saliva-coated surfaces (49, 50), its upregulation might contribute to the increased capacity of S. sanguinis to initiate biofilms. There is evidence that srtA also influences the susceptibility of S. sanguinis to opsonophagocytosis (49), although no clear role of srtA in the systemic virulence of S. sanguinis could be detected in a rabbit model of infective endocarditis (50). The genome of SK36 includes at least 33 genes encoding LPxTG-containing proteins (50), but mechanisms regulating their expression are poorly understood. Among these genes, nt5e and iga were clearly upregulated in the sptR_Ss and sptS_Ss mutants, whereas cshA (which accounts for surface hydrophobicity) (32) was downregulated. Therefore, cell wall-anchored proteins are independently regulated by SptRS_Ss in a fashion not directly associated with srtA expression.

The regulation of IgA1 protease and N5te (cell wall proteins involved in the evasion of host immune functions) by SptRS_Ss strengthens the potential role of this TCS in the persistence of S. sanguinis at oral sites. The production of IgA1 proteases is a prominent trait of early colonizers of teeth (51, 52) and of pathogens of the oropharyngeal mucosa (53, 54), but there is limited information about the mechanisms regulating these functions. The IgA1 subclass (including its secretory form, S-IgA1) accounts for 60 to 90% of the total IgA present in serum, upper respiratory secretions, and saliva (55, 56). S-IgA protects mucosal and tooth surfaces through bacterial agglutination, reductions of bacterial hydrophobicity, blockage of surface adhesins or glucan binding proteins, and modulation of antimicrobial effects of salivary innate immune components, e.g., peroxidase and lysozyme (53, 56–58). Thus, we speculate that defects in sptR_Ss and sptS_Ss functions would trigger the upregulation of iga to enhance survival in saliva as well as the capacity of S. sanguinis to initiate biofilms. Although not associated with biofilm formation, Nt5e significantly contributes to the systemic virulence of S. sanguinis in a rabbit model of infective endocarditis (30). Nt5e downregulates proinflammatory host responses by hydrolyzing ATP (a proinflammatory molecule) and generating adenosine, which has anti-inflammatory functions (30). Thus, Nt5e could also be involved in S. sanguinis immune evasion for persistence in its oral sites. Compared to the latter colonizers of dental biofilms, S. sanguinis is a poor inducer of proinflammatory adaptive responses (59).

Biofilm formation by S. sanguinis is highly dependent on the production of eDNA (11), which is released in an autolysis-independent way, apparently through the tight regulation of cell wall-lytic activities in a process dependent on the production of H₂O₂ (9, 11). Pyruvate oxidase (SpxB) is the major source of H₂O₂ produced by S. sanguinis (9). spxB is directly induced by VicR_Ss, which also regulates murein hydrolase genes (lysM_Ss, cwdP_Ss, and pcsB_Ss) (11) that were upregulated in sptR_Ss- and sptS_Ss-defective strains. These transcriptional changes explain, at least in part, the enhanced capacity of sptR_Ss and sptS_Ss mutants to produce H₂O₂ and eDNA and to form biofilms. S. pneumoniae spxB (spxB_Spn) is more strongly upregulated in O₂- and CO₂-rich atmospheres (60). Thus, the altered chain phenotypes of SKsptR and SKsptS at the late log phase of growth under 10% CO₂ could be associated with altered activities of murein hydrolases affected by H₂O₂ production. Additional genes associated with the production of H₂O₂ (spxR, ackA, and tpk) (28) but not regulated by VicR_Ss (11) were also upregulated in sptR_Ss and sptS_Ss mutants, suggesting a global role of SptRS_Ss in regulating the production of H₂O₂ and DNA release. On the other hand, transcript levels of ccpA (involved in carbon catabolite repression of spxB) (61) were not affected in sptR_Ss and sptS_Ss mutants, indicating that the SptRS_Ss regulatory circuit is not responsive to glucose availability. It is, however, possible that the regulation of H₂O₂ production by SptRS_Ss is linked to the adaptation of S. sanguinis to nutrient-limited conditions, because ackA encodes the acetate kinase for the synthesis of ATP from AcP, a product of SpxB activity.

The mechanisms underlying the regulation of downstream genes by SptRS_Ss remains to be investigated in detail. Except for comX and comE, all genes affected in the sptR_Ss and sptS_Ss mutants were upregulated, compatible with the function of SptR_Ss as a negative regulator. No DNA binding or consensus binding site was established for SptR/RelR proteins in S. pyogenes or S. mutans (15, 18). Consistently, we could not detect binding of rSptR_Ss to the promoter regions of genes downstream of SptRS_Ss (located approximately 250 to 400 bp from the TSS) in our EMSAs. However, analyses of longer regions upstream of these genes revealed conserved repeats that resemble DNA binding sites of transcriptional repressors known to function through the generation of DNA loops (34, 35). In vitro and in vivo analyses of the interactions of SptR_Ss with these conserved regions will be required to understand the SptR_Ss-mediated mechanisms of gene regulation. In addition, downregulation of comX and comE in the sptR_Ss and sptS_Ss mutants remains to be further analyzed. In S. pneumoniae, comX_Spn is required for bacterial persistence in animal models of acute pneumonia and bacteremia (62) and seems to be controlled by additional regulators not involved in quorum sensing (63, 64). Although comX_Ss could not be associated with the capacity of S. sanguinis to promote infective endocarditis in a competitive animal model (65), further studies are required to define the roles of comX_Ss in S. sanguinis biology.

In summary, in this study, we identify a TCS of S. sanguinis, SptRS_Ss, which regulates functions associated with survival in saliva and biofilm formation. By combining phenotypic and transcription profiling of sptR_Ss and sptS_Ss isogenic mutants, we identify major gene functions involved in the SptRS_Ss regulatory circuit and provide evidence that this TCS modulates S. sanguinis cell wall homeostasis, the production of H₂O₂ and eDNA, as well as the evasion of host immune components. Comparisons of the promoter regions of genes downstream of SptRS_Ss further indicated potential mechanisms of SptR_Ss-mediated regulation. These findings imply that deciphering the complex SptRS_Ss regulatory network will be important for understanding the mechanisms by which S. sanguinis persists in the oral environments and perhaps evades host defenses to promote systemic infections.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

Strains used in this study are described in Table 2. S. sanguinis strains were grown from frozen stocks in BHI agar (Difco) (37°C with 10% CO₂ for 24 h). Colonies were then inoculated in BHI medium and incubated for 18 h. Adjusted inocula were then transferred to fresh BHI medium, Mueller-Hinton broth (MHB; Difco), or RPMI 1640 nutrient-poor medium (Gibco Life Technologies) and incubated (37°C) under 10% CO₂ (90% air) or aerobiosis (rotational aeration at 150 rpm). Erythromycin (Erm) (10 μg/ml) and/or spectinomycin (Spec) (200 μg/ml) was added to media for the selection and maintenance of the respective mutant and complemented strains. The Escherichia coli DH5α strain was used for plasmid propagation. All reagents were purchased from Sigma-Aldrich unless otherwise specified.

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s) or purpose	Source or reference
Strains
SK36	Erms	ATCC
SKsptS	ΔsptS::Ermr	This study
SKsptR	ΔsptR::Ermr	This study
SKsptRS	ΔsptS ΔsptR::Ermr	This study
SKsptS+	ΔsptS::Ermr; pDL278::sptS; Specr	This study
SKsptR+	ΔsptR::Ermr; pDL278::sptR; Specr	This study
E. coli DH5α	General cloning and plasmid amplification	Invitrogen
E. coli BL21	Expression of pET22B[27]::sptS and pET22B[27]::sptR	Novagen
Plasmids
pVA838	Ermr source	66
pDL278	Specr cassette; low-copy-no. vector for construction of complemented strains	67
pET22B+	Ampr; empty vector for construction and expression of His-tagged SptR protein	Novagen

Construction of sptRS_Ss nonpolar isogenic mutants and complemented strains.

Plasmids and oligonucleotides used for the construction of isogenic mutants of sptR_Ss (SSA_1794) and sptS_Ss (SSA_1793) or of both the sptR_Ss and sptS_Ss genes are shown in Table S1 in the supplemental material. The sptS_Ss nonpolar deletion mutant was obtained from strain SK36 (SKsptS) by double-crossover recombination with a null allele constructed by PCR ligation (11, 68). In the recombinant allele (2,037 bp), an internal sequence of 1,240 bp of the sptS_Ss encoding region (SSA_1793) was replaced by an erythromycin resistance (Erm^r) cassette obtained from plasmid pVA838. For sptR inactivation, a 1,555-bp recombinant allele was constructed by replacing a 705-bp internal sequence of sptR (SSA_1794) by the Erm^r cassette. The sptS_Ss sptR_Ss double mutant was obtained with a recombinant allele (of 2,988 bp) in which a 1,840-bp sequence of the sptR_Ss and sptS_Ss genes (from position 481 of sptR_Ss to position of 1068 of sptS_Ss) was replaced by the Erm^r cassette. Complemented sptS_Ss and sptR_Ss mutants (SKsptS⁺ and SKsptR⁺, respectively) were obtained by transforming the mutant strain with plasmid pDL278 (which harbors a spectinomycin resistance gene) containing an intact copy of the respective deleted gene.

Analyses of chain formation and cell viability.

The morphologies and chain lengths of S. sanguinis strains were examined by light microscopy (Axio Scope.A1 light microscope; Carl Zeiss) of Gram-stained cells at the mid-log (A₅₅₀ of 0.3) and late log (A₅₅₀ of 0.7) phases of growth in BHI cultures obtained by incubation (37°C) in 10% CO₂ or aerobic shaking. In parallel, cell viability was analyzed by using cells stained with the Live/Dead BacLight system (ThermoFisher) as recommended by the manufacturer, using a confocal laser scanning microscope (Leica TCS SP5 microsystem) (HCX PL APO lambda 63.0× 1.4 oil). The numbers of cocci per chain in the Gram-stained samples were determined by a blind examiner, who analyzed (at a total magnification of ×1,000) a total of 50 chains from five predetermined fields (10 chains per field) for each strain.

Ex vivo growth in human saliva.

Saliva samples were collected from six healthy volunteers as previously described (69), with some modifications (11, 40), according to a protocol approved by the Ethics Committee of the Piracicaba Dental School, University of Campinas (CEP/FOP-UNICAMP) (protocol no. 013/2013). The six volunteers (3 males and 3 females) were 20 to 38 years of age and showed good oral health (including the absence of active dental caries or periodontal disease), as clinically determined by a dental professional (M.P.M.R.M.) of FOP-UNICAMP. The volunteers were nonsmokers and did not use drugs for at least 2 weeks before saliva collection. Collection of stimulated whole saliva was performed at least 2 h after food ingestion. Briefly, salivation was stimulated by chewing on paraffin wax, and saliva was collected during 1 min on glass tubes maintained on ice. Samples were then added to 2.5 mM dithiothreitol (DTT) to allow subsequent filtration (40). Samples were then clarified by centrifugation (16,000 × g for 15 min at 4°C) and filter sterilized in 0.20-μm-pore-size filters (Nalgene Labware, Rochester, NY, USA). Aliquots of saliva obtained from a reference volunteer and from pools of saliva obtained from six volunteers were maintained at −70°C until use.

The analysis of the ex vivo growth of S. sanguinis strains in whole saliva was adapted from methods reported in a previous study (40). Briefly, to obtain bacterial suspensions in saliva samples, the same numbers of colonies of each strain collected from cultures grown for 24 h in BHI agar (with the addition of selective antibiotic when required) were inoculated in MHB (with or without the addition of selective antibiotic) and incubated (10% CO₂ at 37°C) during 18 h. Cultures were then diluted 1:100 in fresh MHB medium and incubated under the same conditions until the mid-log phase of growth (A₅₅₀ of 0.3). These cultures were then diluted 1:50 in 5 ml of saliva samples, and incubation continued during 6 to 14 days. Aliquots of these samples were collected just after culture dilution (time zero) and every 24 or 48 h for the determination of the number of viable cells on BHI agar cultures (CFU per milliliter). Time points for counting of bacteria were determined in preliminary experiments with the tested saliva samples. Additionally, similar experiments were previously performed with saliva from a reference volunteer to assess the influence of possible interstrain variations of chain formation or saliva-mediated aggregation on bacterial counts. In these assays, bacterial counts were determined for parallel samples collected at each time point, which were previously sonicated to break chains or were not sonicated before plating onto BHI agar. Microscopy analyses of Gram-stained samples were also performed. These analyses revealed that the observed differences in viable counts in saliva between the mutants and the parent or complemented strains were not influenced by the formation of chains or saliva-mediated aggregation (data not shown). Thus, the final experiments were performed by using nonsonicated suspensions to exclude the possibility of sonication effects on cell wall integrity in the mutant strains. At least three experiments were performed in duplicate with the reference saliva and with pooled saliva.

Sensitivity to oxidative stress.

Sensitivity to oxidative stress was assessed for strains exposed to high doses of H₂O₂ as detailed previously (11), with minor modifications. Briefly, strains were grown in BHI medium (supplemented with the selective antibiotic when necessary) and incubated under aerobiosis (aerobic shaking) or anaerobiosis (10% H₂, 10% CO₂, and 80% N₂) until an A₅₅₀ of 0.2 was reached. H₂O₂ (10 μM) was then added to the cultures, and the cultures were incubated at room temperature for 1 h. Afterwards, a lethal dose of H₂O₂ (100 μM) was added, and incubation continued for an additional 30 min. Serial dilutions of cells were grown in BHI agar (37°C with 10% CO₂ for 48 h) for the determination of bacterial counts (CFU per milliliter). Sensitivity to oxidative stress was expressed as the percentage of viable bacteria (log CFU per milliliter) recovered after H₂O₂ exposure in relation to bacterial counts determined before exposure to H₂O₂ (set as 100%). Three independent experiments were performed in triplicate.

Binding of C3b and the MAC to S. sanguinis strains.

Binding of C3b and the MAC to the surface of serum-treated strains was determined as previously described (70), with some modifications. Briefly, approximately 10⁷ CFU of strains at the mid-log phase of growth (A₅₅₀ of 0.3) were harvested by centrifugation (10,000 × g at 4°C), washed twice with phosphate-buffered saline (PBS) (pH 7.4), and incubated (37°C for 30 min) in 20 μl of 20% serum in PBS. After two washes with PBS–0.05% Tween (PBST), cells were then incubated on ice (40 min) with fluorescein isothiocyanate (FITC)-conjugated polyclonal goat anti-human C3 IgG antibody (ICN, CA, USA) (1:300 in PBST) or monoclonal anti-human MAC antibody (1:1,000 in PBST). Samples probed with anti-MAC antibodies were then washed twice with PBST and incubated on ice (40 min) with FITC-conjugated anti-mouse IgG antibody (1:1,000). Afterwards, bacterial cells were washed twice with PBST, fixed in 3% paraformaldehyde in PBS, and analyzed on a FACSCalibur flow cytometer (BD Biosciences) using forward- and side-scatter parameters to gate at least 25,000 bacteria. Levels of surface-bound C3b or the MAC on C3b/MAC-positive cells were expressed as the geometric mean fluorescence intensity (MFI). Control samples included bacteria treated only with PBS instead of serum. Heat-inactivated sera (56°C for 20 min) were also used as negative controls in preliminary experiments.

Biofilm formation assays.

Biofilms were formed on saliva-coated polystyrene surfaces and quantified as previously described (11), with minor modifications. Briefly, BHI cultures grown for 18 h with adjusted absorbances (A₅₅₀ of 0.3) were diluted (1:10) in fresh medium supplemented with 1% sucrose. Aliquots of the culture dilutions (200 μl) were then transferred to wells of saliva-treated (18 h at 4°C) polystyrene 96-well plates (CralPlast). Plates were incubated (37°C) under gentle agitation (80 rpm) during 18 h. Afterwards, aliquots of biofilm fluids were transferred to a new microtiter plate to monitor planktonic growth (A₅₅₀), and nonadherent cells were removed from biofilm plates by gentle washing (three washes by immersion in water). Biofilms were then stained with 1% crystal violet during 30 min (room temperature). After this, the stain was eluted in ethanol (30 min at room temperature), and absorbances (A₅₇₅) of the eluates were expressed as indirect measures of biofilm biomass. To assess the effects of eDNA on biofilm formation, similar assays were performed by using culture dilutions supplemented with 50 μg/ml of DNase I (DN-25) or with heat-inactivated DNase I as negative controls, as previously described (11). Three independent experiments were performed in triplicate.

Production of hydrogen peroxide by S. sanguinis strains.

Concentrations of H₂O₂ in the supernatants of planktonic cultures were determined as described previously (11). Culture supernatants were obtained from 1 ml of BHI cultures (aerobiosis at 37°C) at the late log phase of growth (A₅₅₀ of 0.7) by centrifugation (twice at 16,000 × g for 5 min at 4°C). Volumes of 40 μl of these samples were then transferred to wells of 96-well microplates (CralPlast, Brazil) containing 160 μl of a fresh solution of sodium acetate (0.1 M; pH 5.0) with 0.1 μg of horseradish peroxidase and 10 μl of an o-dianisidine solution (1 mg/ml in methanol). Microplates were incubated with protection from light (room temperature for 10 min), and the absorbance (A₅₇₀) of the reaction mixtures was then measured. The concentrations of H₂O₂ in cultures were calculated from standard curves prepared with BHI medium (0.29 to 4.68 mM H₂O₂).

Autolysis assay.

The autolytic activities of strains were determined as previously described (11). Briefly, strains were grown in BHI medium (37°C with 10% CO₂) until the mid-log phase of growth (A₅₅₀ of 0.3). Cells were then harvested (16,000 × g for 5 min at 4°C) and resuspended to an A₅₅₀ of 0.9 in autolysis buffer (20 mM KH₂PO₄, 1 M KCl, 1 mM CaCl₂, 1 mM MgCl₂, 0.4% sodium azide [pH 6.5]). Suspensions were incubated aerobically at 44°C, and autolysis was monitored spectrophotometrically (A₅₅₀) at 24, 48, and 72 h. Three independent experiments were performed in duplicate.

Quantification of eDNA.

Amounts of eDNA in planktonic cultures were measured by quantitative PCR (qPCR) as previously described (11), with modifications. Briefly, culture supernatants were obtained from 1 ml of cultures in BHI medium (aerobiosis at 37°C) at the late log phase of growth (A₅₅₀ of 0.7) by centrifugation (twice at 16,000 × g at 4°C for 10 min), followed by filtration through polyethersulfone membranes (0.22-μm pore diameter; Millipore) for the removal of remaining cells. qPCR was performed in a StepOne real-time PCR system (Life Technologies) with volumes of 1 μl of culture supernatants mixed with 3.4 μl molecular-grade water, 5 μl of Power SYBR green PCR master mix (Thermo Fischer Scientific, USA), and 0.3 μl of a 10 mM stock solution of each primer for 16S rRNA (see Table S1 in the supplemental material). qPCR cycling conditions were 95°C for 10 min followed by 40 cycles of 95°C for 15 s, 58°C for 15 s, and 72°C for 30 s. The DNA concentration was calculated based on average threshold cycle values against a 10-fold dilution series of purified SK36 genomic DNA in the same medium. Sterile culture medium was used as a negative control in qPCRs. Three independent experiments were performed in duplicate.

RNA isolation, reverse transcription, and qPCR.

RNA was purified by using the RNeasy minikit (Qiagen, GE) from cells at the late log phase of growth, which were harvested from BHI cultures (37°C with aerobiosis), mechanically disrupted (bead beating during 3 cycles of 30 s with 1 min of rest on ice), and treated with Turbo DNase (Ambion, USA) (33). cDNA was obtained from 1 μg of RNA samples by using random primers (71) and SuperScript III (Life Technologies, USA), according to the manufacturer's instructions (33). qPCR was performed with 1 μl of cDNA samples, 300 μM each primer (Table S1), and 1× Power SYBR green PCR master mix (Life Technologies, USA) in a total volume of 10 μl. Results were normalized against S. sanguinis 16S rRNA gene expression, which was invariant under the experimental conditions tested. Assays were performed in duplicate with four independent RNA samples.

Production of recombinant SptR_Ss protein and electrophoretic mobility shift assays.

The production of the recombinant SptR_Ss protein and EMSAs were performed as described previously (11), with modifications. To obtain the SptR_Ss-His fusion protein (rSptR_Ss), the sptR_Ss encoding region was amplified from SK36 genomic DNA by using primers rSSA1794F and rSSA1794R (Table S1). Amplicons were then restricted with NcoI and XhoI, and the purified product was cloned into NcoI/XhoI-digested pET-22b (Novagen), yielding pET-SptR. Plasmid pET-SptR was transformed into E. coli BL21 cells, and the recombinant protein was isolated from 1 liter of culture (A₅₅₀ of 0.5) after 3 h of induction with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG). After cell lysis, rSptR was purified under denaturing or hybrid conditions by affinity chromatography on Ni²-nitrilotriacetic acid (NTA) agarose (Qiagen, GE). The eluted proteins were dialyzed overnight (4°C) in PBS and stored at −70°C. Protein samples were analyzed in SDS-PAGE gels stained with Coomassie blue, and the rSptR protein concentration was determined by using the Bradford reagent.

For EMSAs, amplicons of the promoter regions of candidate (spxB, spxR, ssaB, cwdP, and comX) and control (gtfP) genes were generated with specific primers (Table S1), purified, and labeled with digoxigenin (DIG) by using the DIG Gel Shift kit (Roche). Binding reactions of labeled DNA (∼3 fmol) with rSptR_Ss at several concentrations (0, 6, 15, 20, 30, 40, 44, 50, 100, 160, 200, 400, 600, and 800 pmol) were performed with volumes of 25 μl containing 1× DIG Gel Shift buffer [20 mM HEPES, 1 mM EDTA, 10 mM (NH₄)₂SO₄, 1 mM dithiothreitol (DTT), 0.2% Tween 20, 30 mM KCl (pH 7.6)], poly-l-lysine (5 ng/μl), and unspecific competitor poly(dI-dC) or poly(dA-dT). Samples were incubated (25°C for 60 min), and DNA-protein complexes were separated in nondenaturant 6% acrylamide gels (120 V for 1.5 h) in 0.5× Tris-borate-EDTA (TBE) buffer (pH 8.0). As a positive control for DNA-protein binding, similar binding reactions with rCovR_Sm (15 pmol) and ∼3 fmol of a DIG-labeled fragment of the covR_Sm promoter region were applied (33). Protein-DNA complexes were electrotransferred to positively charged nylon membranes (Amersham, GE) and detected by using anti-DIG antibodies conjugated with alkaline phosphatase and a CDP Star system (Roche, Switzerland), according to the manufacturer's protocol. Alternatively, in some experiments, purified DNA fragments were labeled with biotin by using the Biotin 3′ End DNA labeling kit (Thermo Scientific), and protein-DNA complexes previously transferred to nylon membranes were probed with a stabilized streptavidin-horseradish peroxidase (HRP) conjugate and detected by using the LightShift chemiluminescent EMSA kit (Thermo Fischer Scientific, USA), according to the manufacturer's protocol. To assess the specificity of binding, a 200-fold excess of an unlabeled test fragment (cold DNA) was incubated with rSptR_Ss in control reaction mixtures.

In silico analyses of SptRS_Ss structure and of promoter regions of potential SptR_Ss target genes.

The identification of SptRS_Ss and protein sequence alignments were performed by using BLAST and Clustal W (http://www.genome.jp/tools/clustalw/) analyses. The domain architecture of SptRS_Ss was analyzed by using the SMART tool (http://smart.embl-heidelberg.de/) (72, 73). Conserved-domain analysis was also performed by using the NCBI Conserved Domains Database (CDD) (https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml) (74). To investigate the promoter regions of potential SptR_Ss target genes, sequences of 1,000 bp upstream of the TSSs of genes transcriptionally affected in the sptR_Ss and sptS_Ss mutants were retrieved from GenBank (https://www.ncbi.nlm.nih.gov/gene/). These sequences were then analyzed for the detection of potential binding sites of conserved transcription factors by using Clustal W, BLAST, and regulatory sequence analysis tools (RSAT) (75). Predictions of promoters and transcription factor binding sites were also performed by using the Bacterial Promoter Prediction SoftBerry tool (http://linux1.softberry.com/berry.phtml?topic=case_study_bacteria). Promoter sequences and transcription start sites were further analyzed by using Neural Network Promoter Prediction tool (http://www.fruitfly.org/seq_tools/promoter.html). A graphical representation of the identified conserved sequences was obtained by using WebLogo (http://weblogo.berkeley.edu/).

Data analysis.

Phenotypic comparisons were performed by using a nonparametric Kruskal-Wallis test with Dunn's post hoc multiple-comparisons or Mann-Whitney tests. Ex vivo survival in saliva and growth curves were compared by using a Kruskal-Wallis test with Dunn's post hoc test with correction for repeated measures. Parametric analysis of variance (ANOVA) with Dunnett's post hoc multiple-comparisons test was used to compare transcriptional data. Differences were considered significant when a P value of <0.05 was obtained.