Molecular characterization of the dextran-binding lectin B gene dblB of Streptococcus criceti in Streptococcus mutans strain GS-5 with mutations in both gbpC and spaP genes

Streptococcus mutans, a cariogenic agent, has a glucan-binding protein gene, gbpC, and S. criceti possesses four gbpC homologs, including dblA and dblB, as does S. sobrinus. The S. criceti dblB gene encodes a 1,717-amino-acid protein having two repetitive alanine-rich and proline-rich regions and an LPXTG motif, which is recognized by the sortase SrtA, near the C terminus. Reverse transcription-PCR analysis indicated no cotranscription of the dblA and dblB genes of S. criceti. As we could not obtain a dblB mutant of S. criceti, the dblB gene was characterized in S. mutans strain GS-5, which has genetic mutations in both gbpC and spaP genes and shows an inability to agglutinate triggered by dextran. A dextran-induced agglutination assay showed that S. mutans cells carrying dblB agglutinated in the presence of dextran. A hydrophobicity assay showed that the cells containing dblB were hydrophobic. A biofilm formation assay showed that the dblB gene was associated with biofilm formation by cells cultivated in brain heart infusion broth supplemented with glucose and maltose, but not sucrose. Nucleotide sequence analysis of the S. criceti strains studied revealed a frameshift mutation in the srtA gene encoding sortase, but intact dblA and dblB genes were found in dextran-induced agglutination-negative strains, whereas intact dblA, dblB and srtA genes were found in dextran-induced agglutination-positive strains. These results suggest the cell-surface localization of dblA and dblB gene products by SrtA and the responsibility of dblB for dextran-induced agglutination, cellsurface hydrophobicity and biofilm formation in S. criceti.


INTRODUCTION
Streptococcus mutans is a cariogenic agent in humans (for review, see Colby and Russell, 1997) and possesses cell-surface proteins such as glucan-binding protein C (GbpC) (Sato et al., 1997) and SpaP (also known as antigen I/II, PAc, P1 and antigen B) (for review, see Brady et al., 2010), which contain an LPXTG motif recognized by sortase SrtA for anchorage to the cell wall (Igarashi, 2004).SpaP contains two repetitive domains, an alanine-rich A region and a proline-rich P region, and is involved in binding to salivary glycoproteins (Sciotti et al., 1997) and in cell-surface hydrophobicity (Murakami et al., 1997).GbpC is responsible for dextran-dependent agglutination in S. mutans grown under stress conditions such as at a subinhibitory concentration of tetracycline, 4% ethanol, or 42°C incubation (Sato et al., 1997).Agglutination also occurs through the transcriptional regulator IrvA, which promotes agglutination via not only GbpC but also an increase of SpaP (Zhu et al., 2009).S. mutans GS-5 fails to agglutinate in the presence of dextran owing to a nonsense mutation in the gbpC gene (Sato et al., 2002).Furthermore, a frameshift mutation in the spaP gene has been found in one strain of GS-5 (Murakami et al., 1997), but intact spaP has been identified in other strains of GS-5 (Sato et al., 2002;Biswas and Biswas, 2012).
Streptococcus criceti is a cariogenic organism rarely found in humans and is a member of the mutans streptococci, including S. mutans and S. sobrinus (for review, see Colby and Russell, 1997).S. criceti strain HS-6 T contains four gbpC homologs encoding LPXTG-containing proteins, namely gbpC1, gbpC2, dblA and dblB genes; the gbpC1 and gbpC2 genes are tandemly located in one locus and the dblA and dblB genes in another, as occurs in S. sobrinus strain 6715 (Kojima et al., 2012).Dextraninduced agglutination is a characteristic of S. criceti, similar to S. sobrinus (Drake et al., 1988).Among the four gbpC homologs in S. sobrinus strain 6715, the dextranbinding lectin DblB is most responsible for dextrandependent cell agglutination (Sato et al., 2009).However, it is not known how dextran-induced agglutination occurs in S. criceti.Furthermore, dextran-induced agglutination ability varies in S. criceti strains: agglutination-positive strains are HS-6 T and OMZ 61, and agglutination-negative strains are E49 and HS-1 (Tamura et al., 2012).dblA and dblB genes encoding 1,093-and 1,717-amino-acid proteins, respectively, have been identified in strain HS-6 T (Kojima et al., 2012), but nucleotide sequences of the dblA and dblB genes have not been determined in strains E49, HS-1 or OMZ 61.
As we could not obtain a dblB mutant of S. criceti, we characterized the dblB gene of S. criceti in S. mutans using a dextran-induced agglutination assay, a cell-surface hydrophobicity assay and a biofilm formation assay.Presuming that gene mutations of dblA, dblB and/or srtA might occur in dextran agglutination-negative strains of S. criceti, we sequenced all three genes in three strains of S. criceti and found an srtA gene mutation in the two agglutination-negative strains.

MATERIALS AND METHODS
Bacterial strains, plasmids and growth conditions All strains and plasmids are listed in Table 1.Escherichia coli was grown in Luria-Bertani broth (Difco Laboratories) at 37°C.Streptococcal cells were maintained in brain heart infusion (BHI) broth (Difco Laboratories) in an anaerobic jar at 37°C.To select for transformants, spectinomycin was added to the broth for E. coli (50 μg/ml) and streptococci (500 μg/ml) as needed.For the biofilm formation assay, streptococcal cells were grown in a 5% CO 2 incubator at 37°C.DNA preparation, manipulation and transformation Bacterial DNA was purified using the Wizard Genomic DNA Purification Kit (Promega).Plasmid DNA was isolated using NucleoSpin Plasmid (Macherey-Nagel GmbH).To obtain dblA, dblB and srtA genes from S. criceti and gbpC, spaP and srtA genes from S. mutans, PCR was performed using TaKaRa LA Taq DNA polymerase (Takara Bio) with the primers listed in Table 2.The PCR conditions were as follows: 94°C for 1 min; 35 cycles of 94°C for 10 sec, 51 to 57°C for 15 sec and 72°C for 5 to 8 min; 72°C for 10 min.Agarose gel electrophoresis was performed to separate PCR products.Competent cells of E. coli strain Stellar were transformed in accordance with the supplier's protocol (Clontech Laboratories).S. mutans was transformed using heatinactivated horse serum as described previously (Perry et al., 1983).
DNA sequence analysis DNA sequencing of PCR products was performed using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) and analyzed on a 3130 Genetic Analyzer (Applied Biosystems).DNA sequence analysis and signal peptide prediction were performed with the computational package GENETYX ver.11 (Genetyx).Sequence alignments were generated using Clustal W2 (Larkin et al., 2007).Nucleotide sequence data  a Primers were designed from the nucleotide sequences of S. criceti strain HS-6 T (DDBJ accession numbers AEUV02000002 and AB557645; Kojima et al., 2012) and S. mutans strain GS-5 (DDBJ accession number CP003686; Biswas and Biswas, 2012).
The sequence homologous to the shuttle vector pDL278 (LeBlanc et al., 1992) for in-fusion cloning is shown in lower case.HindIII sites are underlined.
pDL278B.The nucleotide sequence of the dblB gene of the plasmid was verified by DNA sequencing.S. mutans strain GS-5 was transformed with plasmids pDL278 (LeBlanc et al., 1992) and pDL278B, resulting in strains GS-5P and GS-5B, respectively.The existence in these strains of a region containing the plasmid-borne spectinomycin resistance gene was confirmed by PCR.

Dextran-induced agglutination assay
The agglutination assay was performed by modifying a method described previously (Tamura et al., 2012).Bacterial cells cultured anaerobically at 37°C for 16 h in BHI broth were washed three times with phosphate-buffered saline (PBS).Cell suspensions prepared in polypropylene tubes were adjusted to an optical density at 600 nm (OD 600 ) of 1.0.For the assay, cell suspensions transferred to glass tubes were or were not supplemented with dextran (200 μg/ml) and mixed for 20 sec.After the indicated time, agglutination was assessed visually.
Hydrophobicity assay Surface hydrophobicity of streptococcal cells was determined by their adsorption on hexadecane as described by Koga et al. (1990) with some modifications.Briefly, the cells were grown at 37°C for 16 h in BHI broth, and then washed three times and suspended in PBS to an OD 600 of 1.0.Hexadecane was added and the suspension was mixed for 1 min.After incubation for 15 min, turbidity of the aqueous phase was measured at 600 nm.Hydrophobicity is represented as the percentage of adsorbed cells.
Biofilm formation assay Biofilms produced in flatbottomed 96-well polystyrene cell culture plates (TPP Techno Plastic Products) were evaluated by slightly modifying the method described previously (Yamada et al., 2009).Briefly, bacterial cells (1.0 × 10 4 CFU) were cultured in BHI broth, with or without 1% glucose, sucrose or maltose, at 37°C for 16 h under 5% CO 2 in air.Biofilms were stained with 1% crystal violet solution and then washed three times and air-dried.The adhered crystal violet was solubilized with ethanol and the optical density at 595 nm (OD 595 ) was determined.

RESULTS
RT-PCR analysis of dblA and dblB genes Nucleotide sequences of the dblA and dblB genes of S. criceti HS-6 T were retrieved from DDBJ/EMBL/GenBank databases (accession number AB557645; Kojima et al., 2012).As dblA and dblB genes are arranged in tandem (Fig. 1A), their expression was examined by RT-PCR.PCR products generated with primers for the dblA (197 bp) and dblB (295 bp) genes were obtained with reverse transcription and genomic DNA of S. criceti HS-6 T (Fig. 1B), whereas products for the dblA/dblB region (807 bp) were obtained only with genomic DNA.Thus, expression of the dblA and dblB genes was found, but their cotranscription was not observed.

Structural characterization of DblB
The predicted translational product of the dblB gene, DblB, is a protein of 1,717 amino acids with an estimated molecular weight of 182,703.A signal peptide sequence containing a YSIRK-G/S motif (Bae and Schneewind, 2003) was found in the N-terminal region (residues 1-36: MEKNAQRF-SIRKYSFGAASVLLGTAVFVLSAPTVLA, underlined letters indicating the motif sequence).An LPXTG motif for sorting on the cell surface by sortase (Schneewind et al., 1993), LPQTG, was found in the C-terminal region (residues 1,681-1,685).These results suggest that DblB is a cell-surface protein sorted by sortase SrtA.DblB contains two repetitive regions, termed the A and P regions (Fig. 2A).The A region is rich in alanine (ranging from 16 to 18 per repeat unit, namely, 32.7- 36.7% occupied) and contains ten repeating units (A1-A10) of a 49-residue sequence with 26 residues completely conserved in all the repeats (Fig. 2B).The P region is rich in proline (8 to 11 per repeat unit; 33.3-45.8%occupied) and includes 11 repeat units (P1-P11) of a 24-residue sequence with 9 residues completely conserved in all the repeats (Fig. 2C).Two sequences similar to the repeat unit were found around the P region and were designated as PreP and PostP.
Sequence analysis of gbpC, spaP and srtA genes of S. mutans strain GS-5 As transformants of S. criceti cells could not be obtained, we attempted to characterize the dblB gene in S. mutans.To minimize the effects of S. mutans cell-surface proteins such as GbpC and SpaP on experiments, we chose strain GS-5, a dextran agglutination-negative strain (Sato et al., 2002).As three laboratories have reported two variants of strain GS-5, a spaP-mutant strain (Murakami et al., 1997) and a gbpCmutant and spaP-intact strain (Sato et al., 2002;Biswas and Biswas, 2012), we verified the gbpC, spaP and srtA genes of strain GS-5 in our laboratory.Nucleotide sequence analysis of gbpC and spaP genes (assigned DDBJ accession numbers AB742519 and AB819737, respectively) showed that a nonsense mutation in gbpC occurs at codon 65 and an insertion mutation in spaP  (Larkin et al., 2007).Asterisks and periods indicate fully conserved and weakly similar residues, respectively.(C) Amino acid sequence of DblB around the P region.Eleven tandem repeats, P1 to P11, of a 24-residue proline-rich sequence were found in the P region and are flanked with sequences PreP and PostP, which are similar to the repeat unit.Alignment of the 11 repeat units (P1-P11) was performed using Clustal W2 (Larkin et al., 2007).Asterisks and periods indicate completely conserved and weakly similar residues, respectively.occurs at codon 1,157, indicating that our strain is a gbpC-and spaP-mutant strain.The predicted translational products of the gbpC and spaP genes, consisting of 64 and 1,158 residues, respectively, are truncated and lack an LPXTG motif (Schneewind et al., 1993).The nucleotide sequence of the 5,173-bp region containing the srtA gene in our strain is identical to that of the genomesequenced strain (DDBJ accession number CP003686) (Biswas and Biswas, 2012).

Dextran-induced agglutination assay of S. mutans cells
The result of the agglutination assay showed that GS-5 and GS-5P (GS-5 harboring an empty vector) cells did not agglutinate in the presence of dextran (Fig. 3A).In contrast, GS-5B, containing the dblB gene, was rapidly agglutinated by dextran and clumped cells adhered to the glass tube (Fig. 3A).GS-5B cells without dextran adhered to the glass surface above the water phase, whereas the cells with dextran agglutinated and clumped (Fig. 3B).These results confirmed that the agglutination was induced by dextran and was dblB-dependent.

Hydrophobicity assay of S. mutans cells
Cell-surface hydrophobicity was analyzed using hexadecane.GS-5B cells carrying the dblB gene showed the highest hydrophobicity among the three cells studied (Fig. 4).This observation indicated that the dblB gene is involved in cell-surface hydrophobicity in S. mutans.

Biofilm formation assay of S. mutans cells Biofilms
formed by GS-5, GS-5P and GS-5B cells were evaluated (Fig. 5).Although BHI broth is an enriched medium, all three cells formed less biofilm in BHI medium alone.GS-5 and GS-5P cells formed poor biofilm in BHI broth with glucose, sucrose or maltose.In contrast, GS-5B cells formed more biofilm in BHI broth with glucose or maltose, but poor biofilm with sucrose.These findings suggest that the dblB gene product contributes to biofilm formation in BHI broth with glucose and maltose.
Sequence analysis of the dblA, dblB and srtA flanking regions of S. criceti strains To test the hypothesis that mutations in the dblA and/or dblB gene(s) might be present in the dextran agglutination-negative strains Fig. 3. Dextran-induced agglutination assay.(A) Cell suspensions adjusted to an OD 600 of 1.0 were supplemented with dextran (200 μg/ml), mixed, and then incubated at ambient temperature for 30 min.The wild-type strain GS-5, GS-5P (GS-5 harboring an empty vector) and GS-5B (GS-5 harboring pDL278B carrying the dblB gene) were used.(B) GS-5B cell suspensions adjusted to an OD 600 of 1.0 were or were not supplemented with dextran (200 μg/ml), mixed, and then incubated at ambient temperature for 5 min.From the left: cell suspensions before the addition of dextran and vortexing, cell suspensions without dextran after vortexing, and cell suspensions after the addition of dextran and vortexing.Data are representative of three separate experiments.E49 and HS-1, the 9,554-bp sequences of the dblA and dblB genes of strains E49, HS-1 and OMZ 61 were determined (Fig. 6A).In all three strains, the dblA and dblB genes were arranged in tandem, as they are in S. criceti strain HS-6 T (accession number AEUV02000002).The nucleotide sequences of the dblA and dblB genes of strains HS-1 and OMZ61 (accession numbers AB751199 and AB751200, respectively) were identical to those of strain HS-6 T .Regarding strain E49, a substitution was found (AB751198): codon 1,244 of dblB in E49 was CTT (leucine), and those in HS-6 T , HS-1 and OMZ 61 were CTC (leucine).These results show that the dblA and dblB genes in strains E49 and HS-1 are intact.Next, we conjectured that mutations in the srtA gene might be present in the dextran agglutination-negative strains, as the responsibility of the srtA gene for anchoring cell-surface proteins containing an LPXTG motif in S. mutans has been reported (Igarashi, 2004).To test this, the srtA genes and flanking regions were sequenced in strains E49, HS-1 and OMZ 61 (Fig. 6B).The 3,837-bp sequence of the srtA region of strain OMZ 61 (accession number AB910263) was identical to that of strain HS-6 T (AEUV02000002).A comparison of the 3,836-bp sequences of the srtA regions of strains E49 and HS-1 (AB910261 and AB910262, respectively) with those of strains OMZ 61 and HS-6 T showed that a single cytosine deletion in the srtA gene at position 392 had occurred in strains E49 and HS-1 (Fig. 6C).This deletion causes a frameshift at codon 131 and introduces a premature stop codon into the srtA gene (Ala131Valfs*2).SrtA of strains HS-6 T and OMZ 61 is a 258-amino-acid protein containing a cysteine-217 residue, which was predicted as necessary for sortase activity as cysteine-184 of Staphylococcus aureus SrtA (Ton-That et al., 2002) (Fig. 6D).SrtA of strains E49 and HS-1 is, thus, a 133-amino-acid protein lacking a conserved cysteine residue that may be required for sortase activity, suggesting that LPXTG-containing proteins cannot anchor to the cell wall in strains E49 and HS-1.

DISCUSSION
Little is known about the molecules associated with dextran-induced agglutination in S. criceti.Furthermore, the roles of the dblA and dblB gene products in S. criceti are not understood.As both genes are tandemly situated in strain HS-6 T (Kojima et al., 2012), their expression was examined.In our transcriptional analysis, cotranscription of dblA and dblB genes was not detected, whereas both gene transcripts were observed.These findings suggest that their expression is differently regulated by unidentified molecules.
Our results revealed that the primary sequences of DblB in S. criceti strain HS-6 T contained two sequence motifs, the YSIRK-G/S motif for signal peptide processing (Bae and Schneewind, 2003) and the LPXTG motif for cell-wall anchoring, which is conserved in cell-surface proteins (Schneewind et al., 1993).This suggests that DblB is anchored to the cell wall by SrtA, as are other LPXTGcontaining proteins such as DblA, GbpC1 and GbpC2.Furthermore, DblB exhibited 48.3% identity with DblA, suggesting that DblA and DblB are functionally distinct cell-surface proteins.It is of note that DblB contained two repeating regions, the A and P regions, as in the S. mutans cell-surface protein SpaP (for review, see Brady et al., 2010).A and P regions are also found in DblBs of S. sobrinus strain 6715 and S. downei ATCC 33748 T .DblB of S. sobrinus contains six repeating units (A1-A6) of a 49-residue sequence in the A region (residues 266-559) and seven repeat units (P1-P7) of a 24-residue sequence in the P region (residues 1,129-1,296).DblB of S. downei contains eight repeating units (A1-A8) of a 49-residue sequence in the A region (residues 266-657) and nine repeat units (P1-P9) of a 24-residue sequence in the P region (residues 1,227-1,442).Of the three, the largest DblB of S. criceti contains ten repeating units of a 49-residue sequence in the A region and eleven repeat units of the size of ORFs and intergenic regions, respectively, in base pairs.The gyrA and pgm genes encode DNA gyrase subunit A and phosphoglucomutase, respectively.The locus of srtA originated from sequence data of strains HS-6 T (nucleotides 1347059 to 1350895 of AEUV02000002), E49 (AB910261), HS-1 (AB910262) and OMZ 61 (AB910263).(C) Deletion of a single cytosine nucleotide in the srtA genes of strains E49 and HS-1.The deletion at position 392 and the resultant premature termination codon, TGA, on nucleotide sequence alignments of the four srtA genes are shown by dashes and underlining, respectively.Amino acids corresponding to codons in the srtA genes of HS-6 T and HS-1 are represented above and below the corresponding codons, respectively.Sequence data are from strains HS-6 T (AEUV02000002), E49 (AB910261), HS-1 (AB910262) and OMZ 61 (AB910263).(D) Multiple sequence alignment of the deduced amino acid sequences of sortase enzymes.Sequence alignments were generated using Clustal W2 (Larkin et al., 2007).Dashes represent gaps introduced for alignment.Asterisks, colons and dots denote fully, strongly and weakly conserved residues, respectively, in the four.Scri, S. criceti HS-6 T (WP_004226711); Smut, S. mutans GS-5 (AFM81499); Spyo, S. pyogenes MGAS315 (NP_664615); Saur, S. aureus 8325-4 (AAD48437).A frameshift position at codon 131 in SrtA of S. criceti strains E49 and HS-1 is double-underlined.The catalytic site residue cysteine 184 of SrtA of S. aureus (Ton-That et al., 2002) is in bold.
a 24-residue sequence in the P region.
In this study, dblB of S. criceti was expressed in S. mutans strain GS-5, which was confirmed as a gbpC-and spaP-mutant strain.In two laboratories, GS-5 was reported as a gbpC-mutant and spaP-intact strain (Sato et al., 2002;Biswas and Biswas, 2012).As at least two genotypes exist in strain GS-5, care should be taken in experiments using this strain.Use of the gbpCand spaP-mutant strains is appropriate to eliminate the effects of GbpC and SpaP on experiments of dextraninduced agglutination, cell-surface hydrophobicity, and biofilm formation.Indeed, loss of their functions in S. mutans has demonstrated that the gbpC mutant was dextran-induced agglutination-negative (Sato et al., 1997), and the spaP mutant was hydrophilic regarding the cell surface (Koga et al., 1990).In accordance with a previous observation of strain GS-5 (Sato et al., 2002), our GS-5 cells were unable to agglutinate.Furthermore, we showed that the dblB gene is responsible for dextraninduced agglutination in S. mutans.It is of note that cells harboring dblB adhered to the glass surface above the water phase when cells were mixed without dextran, suggesting that dblB gene products act as adhesins on abiotic surfaces in S. mutans cells.Interestingly, the cell-surface hydrophobicity of cells containing dblB increased markedly.This is similar to SpaP, which is important in surface hydrophobicity in S. mutans (Koga et al., 1990).In this study, we showed that S. mutans strain GS-5 cells formed less biofilm in BHI medium alone.The magnitude of biofilm formation in S. mutans strain GS-5 in this study was lower than that in S. mutans strain Xc (Yamada et al., 2009).S. mutans Xc cells possess a SpaP of approximately 190 kDa (Koga et al., 1989).Sequence analysis of the spaP gene in this study predicted that GS-5 cells produce a 1,158-aminoacid protein lacking the cell-wall-anchoring motif LPXTG.Indeed, it has been demonstrated that the spaP gene of GS-5 encodes a 1,158-amino-acid protein and that GS-5 produces extracellularly a 155-kDa SpaP (Murakami et al., 1997).Therefore, the poor biofilm formation of GS-5 cells may be attributable to the loss of SpaP on the cell surface.Furthermore, we showed that cells carrying the dblB gene could form biofilms in BHI broth supplemented with glucose and maltose, suggesting the involvement of dblB gene products in biofilm formation in S. mutans.In contrast, biofilm formation was poor in BHI with sucrose.In S. mutans, glucosyltransferases encoded by the gtfB and gtfC genes synthesize primary water-insoluble glucans from sucrose (Yamashita et al., 1993) and the gtfD gene product synthesizes water-soluble glucans (Hanada and Kuramitsu, 1989).Furthermore, it has been reported that the expression of gtfB and gtfC decreases in the presence of 2% sucrose in BHI broth, whereas gtfD expression increases (Fujiwara et al., 2002).This implies that the expression of genes involved in biofilm formation may alter in the presence of carbohydrates.Therefore, the elimination of cells from biofilms by the wash process in the biofilm formation assay might account for the decreased expression of gtfB and gtfC and the enhanced gtfD expression.The mechanism(s) of the diverse effects of carbohydrates on biofilm formation remain to be studied.Taken together, these findings suggest that the dblB gene product acts as agglutinin and plays a role in biofilm formation in S. criceti.
Our previous work showed that dextran-induced agglutination-negative strains were S. criceti strains E49 and HS-1 and S. sobrinus OMZ 65 (Tamura et al., 2012).In other non-agglutinating strains of S. sobrinus, changes in the dblB gene have been identified: strain OMZ 176, which possesses truncated DblA (584 residues) and DblB (254 residues), has mutations comprising a single adenine nucleotide insertion in the dblA gene and a 736-nucleotide deletion in dblB; and strain B-13N, which produces DblA (1,270 residues) but lacks DblB, shows replacement of the dblB gene by an IS1548-like sequence (Sato et al., 2009).Conversely, these findings support the responsibility of the dblB gene for dextran-induced agglutination in S. sobrinus.In contrast, it remains to be elucidated why some strains of S. criceti could not exhibit dextraninduced agglutination.Our results showed that srtA was mutated in dextran agglutination-negative strains, whereas the dblA and dblB genes were intact.This finding suggests that SrtA was responsible for the ability of dextran-induced agglutination via LPXTG-containing proteins such as DblB, an explanation which is supported by a report that inactivation of the srtA gene inhibited dextran-induced agglutination by GbpC in S. mutans (Igarashi et al., 2004).However, we cannot exclude the possibility of causative genes of dextran-induced agglutination other than dblB.It has been reported that in S. sobrinus strain 6715, dextran-binding activities are observed in GbpC2 and DblA as DblB, but there is no activity in GbpC1 (Sato et al., 2007).Thus, it is presumed that GbpC2 and DblA, but not GbpC1, of S. sobrinus and probably S. criceti, are independently involved in dextraninduced agglutination.
In S. sobrinus, strain 6715 possesses dblA and dblB genes (Sato et al., 2009) and strains OM55d and SL1 T harbor dblA and a dblA/dblB homologous gene, dblC, instead of the dblB gene (Kojima et al., 2012).This suggests variations of dblA/dblB/dblC gene arrangements in S. sobrinus.In contrast, the four S. criceti strains studied possessed dblA and dblB genes, suggesting that the dblA and dblB regions are conserved in S. criceti.Furthermore, tandemly arranged genes, such as gtfB and gtfC in S. mutans, are considered to have replicated by tandem duplication (Hoshino et al., 2012).Regarding the dblA and dblB genes, orthologous genes were found in S. sobrinus, S. downei and S. criceti (Kojima et al., 2012).Thus, tandem duplication might have occurred in the ancestor of the three species.
In conclusion, the dblB gene of S. criceti was characterized.Our results suggest that the cell-surface localization of dblA and dblB gene products is governed by SrtA and that the dblB gene is responsible for agglutination induced by dextran, cell-surface hydrophobicity and biofilm formation in S. criceti.

Fig. 1 .
Fig. 1.Transcriptional analysis of the dblA/dblB genes of S. criceti.(A) Schematic representation of the dblA and dblB region of S. criceti strain HS-6 T .Arrows and arrowheads indicate open reading frames (ORFs) and the positions of primers used for RT-PCR analysis, respectively.Numbers inside and between arrows indicate the size of ORFs and intergenic regions, respectively, in base pairs.Sequence data are from DDBJ accession number AB557645 (Kojima et al., 2012).(B) RT-PCR analysis in a 1.5% agarose gel.The sizes of PCR products and representative DNA standards are shown on the right and left, respectively.Lane M, size marker; lane 1, RT-PCR mixture; lane 2, negative control without RT; lane 3, positive control of PCR with genomic DNA of HS-6 T .The results are representative of three separate experiments.

Fig. 2 .
Fig. 2. Structural characteristics of DblB of S. criceti.(A) Schematic representation of DblB.The signal peptide sequence (residues 1-36), alanine-rich repeat region (A region, residues 265-754) and proline-rich repeat region (P region, residues 1,325-1,588) are indicated.Locations of PreP and PostP flanking the P region and an LPXTG motif (residues 1,681-1,685) are shown.(B) The A region of DblB.Ten repeating units, A1 to A10, of a 49-residue alanine-rich sequence were found.Multiple sequence alignment was constructed by Clustal W2(Larkin et al., 2007).Asterisks and periods indicate fully conserved and weakly similar residues, respectively.(C) Amino acid sequence of DblB around the P region.Eleven tandem repeats, P1 to P11, of a 24-residue proline-rich sequence were found in the P region and are flanked with sequences PreP and PostP, which are similar to the repeat unit.Alignment of the 11 repeat units (P1-P11) was performed using Clustal W2(Larkin et al., 2007).Asterisks and periods indicate completely conserved and weakly similar residues, respectively.

Fig. 4 .
Fig. 4. Cell-surface hydrophobicity assay.Surface hydrophobicity of GS-5, GS-5P and GS-5B cells was determined by their adsorption on hexadecane.The results express the means and SDs (error bars) of triplicate samples, and are representative of three separate experiments.GS-5GS-5P GS-5B

Fig. 5 .Fig. 6 .
Fig. 5. Biofilm formation assay.GS-5, GS-5P and GS-5B cells were cultivated on polystyrene cell culture plates and in BHI broth with or without 1% additional glucose, sucrose or maltose.Biofilms were stained with 1% crystal violet and OD 595 values were measured.The results express the means and SDs (error bars) of quadruplicate samples in three independent experiments.

Table 1 .
Bacterial strains and plasmids used in this study

Table 2 .
Primers used in this study For amplification of flanking region of srtA gene of S. mutans srtASmF TTAACGAGAGCGAAAGCGTAA srtASmR CGGATTGTTAGCCTTAACCTACT