TetR Family Regulator brpT Modulates Biofilm Formation in Streptococcus sanguinis

Biofilms are a key component in bacterial communities providing protection and contributing to infectious diseases. However, mechanisms involved in S. sanguinis biofilm formation have not been clearly elucidated. Here, we report the identification of a novel S. sanguinis TetR repressor, brpT (Biofilm Regulatory Protein TetR), involved in biofilm formation. Deletion of brpT resulted in a significant increase in biofilm formation. Interestingly, the mutant accumulated more water soluble and water insoluble glucans in its biofilm compared to the wild-type and the complemented mutant. The brpT mutation led to an altered biofilm morphology and structure exhibiting a rougher appearance, uneven distribution with more filaments bound to the chains. RNA-sequencing revealed that gtfP, the only glucosyltransferase present in S. sanguinis, was significantly up-regulated. In agreement with these findings, we independently observed that deletion of gtfP in S. sanguinis led to reduced biofilm and low levels of water soluble and insoluble glucans. These results suggest that brpT is involved in the regulation of the gtfP-mediated exopolysaccharide synthesis and controls S. sanguinis biofilm formation. The deletion of brpT may have a potential therapeutic application in regulating S. sanguinis colonization in the oral cavity and the prevention of dental caries.


Introduction
Biofilms represent a major health concern as they are ubiquitous in nature and frequently attach to biotic and abiotic surfaces [1]. Bacteria within these communities create an extracellular matrix composed of exopolysaccharides (EPS), proteins, lipids, DNA and ions [2,3]. Attachment is first established through the adherence of bacteria to the surface through the interaction of membrane associated proteins, followed by the accumulation of bacteria and the a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 upstream of the gene ORF, a promoterless kanamycin cassette (aphA-3) to replace the ORF and the 1-kb sequence downstream of gene ORF. The three fragments were combined through a second round of PCR. The final recombinant PCR product was transformed into S. sanguinis SK36 then selected by kanamycin resistance and confirmed by PCR analysis.
For the construction of ΔbrpT/ΔgtfP, three sets of primers were used to independently PCR amplify the 1-kb sequence upstream of gtfP, an erythromycin cassette isolated from plasmid pVA838 [22] to replace the gtfP ORF and the 1-kb sequence downstream of gtfP. The three fragments were then combined through a second round of PCR. The final recombinant PCR product was transformed into the ΔbrpT mutant, selected by kanamycin and erythromycin resistance then confirmed by PCR analysis.
For complementation of the brpT mutant, a similar PCR-based method was employed [23]. Briefly, three DNA fragments were independently amplified using primer sets 0144F1/0144R1, 0144F2/0144R2 and 0144F3/0144R3, for the 1-kb sequence upstream plus the coding sequence of brpT, the erythromycin resistance cassette (pVA838) [22] and the 1-kb sequence downstream of brpT, respectively. The final recombinant PCR product containing these three fragments was generated by overlapping PCR. It was then introduced into the brpT mutant to replace the kanamycin resistance cassette with the brpT ORF and the erythromycin resistance cassette. An erythromycin resistant and kanamycin sensitive transformant was selected and confirmed by PCR analysis.

Quantification of glucans in biofilm
Wild-type S. sanguinis and mutants were grown in BM supplemented with 1% sucrose for 24 h in 24-well plates. The following day, the supernatant was removed and biofilms were re-suspended with an equal volume (500 μl) of distilled water, transferred into 2 ml tubes and centrifuged. The precipitate was re-suspended again with 500 μl of distilled water and centrifuged. Supernatants from these two washes were mixed and used for extraction of water soluble glucans (WSG). The sediment was dissolved in the same volume of 1 N NaOH for 3 h, and centrifuged. The supernatant of this extraction was used for the quantification of alkali-soluble glucans (water insoluble glucan, WIG). These two fractions containing WSG and WIG were precipitated separately by 3 volumes of isopropanol overnight. The precipitates obtained by centrifugation were then air dried, and dissolved in 100 μl of sterile dH 2 O (WSG) or 1 N NaOH (WIG). The amount of glucans in each fraction was quantified by the phenol-sulfuric acid method as previously described [24]. Briefly, 50 μl of each sample was pipetted into 96-well plates, and 25 μl of 5% phenol was added, then 125 μl of concentrated sulfuric acid was added rapidly and mixed by pipette. The plates were allowed to stand 30 min at room temperature. The absorbance was then measured at 490 nm. H 2 O and 1 N NaOH were used as solvents for detection of WSG and WIG, respectively, as well as standard curves, which were generated by using glucose as a reference carbohydrate. The amounts of WSG and WIG were expressed as glucose equivalent in each well. To determine the efficiency of glucan accumulation, the concentration of glucans was normalized to the concentration of genomic DNA extracted using a QIAamp DNA Mini kit (Qiagen, Hilden, Germany) and quantified by a Nanodrop 2000 spectrophotometer (Thomas Scientific).

Confocal laser scanning microscopy (CLSM) analysis of biofilm
Wild-type S. sanguinis and mutants were grown in BM supplemented with 1% sucrose in 24-well plates as described above. The following day, the wells were gently washed three times with phosphate buffer saline (PBS, pH7.4), followed by staining with 1.5 μM SYTO 9 (Invitrogen, Grand Island, NY) for 25 min. After removing the stain, the wells were gently washed by PBS. Biofilm images were visualized and collected by CLSM using a LSM 710 confocal laser scanning system (Zeiss, Thornwood, NY). SYTO 9 fluorescence was detected by excitation 488 nm, and collected with a 493-to 559-nm bandpass filter. All z-sections were collected at 5.8μm intervals by using a Plan-Neofluar ×10/0.3 objective lens. ImageJ software was used for image processing.

Scanning electronic microscopy (SEM) analysis of biofilm
Wild-type S. sanguinis and mutants were cultured overnight. The following day, cells were diluted 1:100 in BM supplemented with 1% sucrose. Biofilms were grown on small sterile polystyrene coverslips within 24-well flat-bottom plates for 24 h at 37˚C under anaerobic conditions. Biofilms on the coverslips were washed once with PBS, fixed by 2% glutaraldehyde in 0.1 M sodium cacodylate buffer for 1 h. Following dehydration through a graded series of ethanol, the coverslips were air dried and sputter coated with gold. Samples were then scoped by a SEM machine (Zeiss EVO 50 XVP, Jena, Germany).

RNA-sequencing preparation
Wild-type S. sanguinis SK36 and the brpT mutant were grown in BHI medium to OD 600~0 .9. Cells were collected and treated with RNAprotect bacteria reagent (Qiagen, Valencia, CA) to stabilize RNA then broken by mechanical disruption using FastPrep lysing matrix B (Qbiogene, Irvine, CA). Total RNA was treated with DNase I (Qiagen) and prepared by using RNA easy mini kits (Qiagen) according to the manufacturer's instructions. Ribo-Zero Magnetic Kit for Bacteria (Illumina) was used to deplete ribosomal RNA from 2 μg of total RNA. NEBNext Ultra Directional RNA Library Prep Kit for Illumina (New England BioLabs) was used for the following RNA-seq library preparation according to the manufacturer's protocol. Briefly, ribosomal-depleted RNA was fragmented followed by first-strand cDNA synthesis from random primers using ProtoScript II Reverse Transcriptase (New England BioLabs). Second strand cDNA was synthesized and purified. End repair was performed on the double-stranded cDNA and primed with the addition of 5'-phosphorylated dA-tailed ends using T4 DNA polymerase, Klenow DNA polymerase and T4 polynucleotide kinase (New England BioLabs). This was immediately followed by adaptor ligation (New England BioLabs) and purified. Samples were PCR-amplified for 12 cycles with Phusion HiFi polymerase (New England Biolabs) with paired-end primers and a randomly chosen unique barcode (Illumina). Agencourt AMPure XP Beads (Beckman Coulter) were used for all purification steps. Library sequencing was performed by the Nucleic Acids Research Facilities at Virginia Commonwealth University using Illumina HiSeq2000. The raw RNA-seq data are available in the NCBI Gene Expression Omnibus (GEO) (www.ncbi.nlm.gov/geo/query) under the accession number: GSE89964.

Mapping and analysis of RNA-sequencing data
Reads obtained from RNA-sequencing were aligned against the S. sanguinis SK36 genome using Rockhopper v. 2.03 [25]. Analyses were run on default parameter settings to obtain expression data of brpT mutant compared to wild-type SK36. Significance was determined by a q-value 0.01 adjusted for a false discovery rate of 1%. Transcriptome profiles were analyzed for enriched pathways and functionally related genes using DAVID v. 6.8 Beta [26].

Quantitative RT-PCR (qRT-PCR) analysis
qRT-PCR was performed as described previously [23]. Wild-type S. sanguinis SK36 and mutants were grown in BHI medium with or without sucrose to OD 600 0.9-1.0. Cells were collected and broken by mechanical disruption using FastPrep lysing matrix B (Qbiogene, Irvine, CA). Total RNA was treated with DNase I (Qiagen) and prepared by using RNA easy mini kits (Qiagen, Valencia, CA) according to the manufacturer's instructions. First-strand cDNA synthesis was performed in a 20 μl system containing 100 ng Total RNA, 0.2 μl Random Primer strand buffer and RNase free water to a 20 μl volume. Reactions lacking reverse transcriptase were performed in parallel as control for possible DNA contamination. First-strand cDNA from each reaction was subjected to 10-fold dilution and used in subsequent qRT-PCR. The qRT-PCR was prepared in reactions containing 5 μl 2X SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA), 0.1 μl each PCR primer (20 μM), 0.5 μl diluted firststrand cDNA and distilled water to a 10 μl volume. The reaction was performed on an ABI 7500 fast real-time PCR system. The housekeeping gene gyrA was used as a normalization control. The 2 -ΔΔCt method was employed for calculation of relative expression levels of target genes. The data were collected and statistically analyzed from triplicates.

S. sanguinis brpT affects biofilm formation
Through the continued study of our S. sanguinis genome-wide gene mutant library [18], we identified a new biofilm-related gene, brpT, a TetR family transcriptional regulator. The preliminary screening indicated that the brpT deletion mutant displayed an increased biofilm phenotype compared to the wild-type SK36, when grown in either BM or trypticase soy broth (data not shown).
To confirm that the increase in biofilm was a result of the brpT mutation, a complemented mutant was constructed. The growth rates of the wild-type, the brpT mutant and the complemented mutant were first examined and we found no significant change in the bacterial overnight growth of brpT mutant (S2 Fig). Biofilms for the brpT mutant, the wild-type and the complemented mutant were then quantified by CV staining. The brpT mutant showed a significant increase in biofilm formation compared to the wild-type and the complemented mutant (P <0.01), while no significant difference between the biofilms of the wild-type and the complemented mutant was observed (Fig 1).
BrpT is a small 22 kDa protein that contains a helix-turn-helix domain in the N-terminal region, typical of the XRE-family transcriptional regulators (Xenobiotic Response Element, a prokaryotic transcriptional DNA binding regulator family), and a tetracycline repressor domain in the C-terminal region. Previous studies indicate that the brpT homolog, SMU.1349, in S. mutans, is involved in the transcriptional regulation of itself and adjacent sequences [19]. To examine whether any neighboring genes were potential effector(s) responsible for the observed biofilm increase, eight mutants with deletions either upstream or downstream of brpT were selected from our S. sanguinis mutant library [18] and screened. We found that mutations in these neighboring genes did not significantly alter biofilm formation compared to the wild-type (S3 Fig), indicating that the observed biofilm increase was not caused by a polar effect and only the deletion of brpT was responsible for the observed increase in biofilm production.

Deletion of brpT alters S. sanguinis biofilm properties
During our screening, we noticed that deletion of brpT resulted in biofilms with an uneven appearance compared to the wild-type SK36 and the complemented mutant. To better understand these morphological differences, biofilms for the wild-type, the brpT mutant and the complemented mutant were grown under anaerobic conditions in flat-bottom polystyrene microtiter plates and examined by confocal laser scanning microscopy (CLSM). The biofilm of the wild-type and the complemented mutant showed uniform green fluorescence intensity whereas bright fluorescence clusters surrounded by dark (blank) areas were observed for the brpT deletion mutant (Fig 2A). A quantitative analysis conferred with what was visually observed. The average thickness of the mutant biofilm (119.9±3.3 μm) was approximately 1.5-fold more than that of the wild-type (81.2±5.8 μm) and the complemented mutant (79.3 ±3.3 μm) (Fig 2B) and there was a large increase in the biofilm roughness corresponding to the uneven structural morphology (Fig 2C).
SEM analysis was employed to further assess the changes in biofilm morphology. As shown in Fig 3A, when scanned under a low magnification (1000×), the brpT mutant biofilm showed a noticeably different morphology from that of the wild type. There were many peaks and dents (like a corrugated surface) compared to the relative uniform distribution of the wildtype and complemented mutant. Interestingly, the streptococcal chains of the brpT mutant were surrounded by numerous fine filamentous substances when scanned under a higher magnification (20,000×) ( Fig 3B). These filamentous substances were also observed in SK36 and the complemented mutant, but were less abundant than that seen in the ΔbrpT biofilms. Though further investigations should be taken towards the nature of these filaments, we hypothesize they may be water insoluble polysaccharides or glucans. These data suggest that brpT is involved in the regulation of the biofilm composition and structure.

Deletion of brpT increases the quantity of glucans in S. sanguinis biofilm
We noticed that the brpT mutant easily formed large clusters in tubes with broth supplemented with sucrose and these clusters were difficult to dissolve in water by vigorous vortex. Studies performed in other oral streptococci have demonstrated that glucans are important contributors to the bulk physical integrity and stability of biofilms and act as a key component in the matrix of oral cariogenic biofilms [27][28][29]. Corresponding to the increase in filamentous structures we noticed in the SEM analysis, we hypothesized that the mutant biofilm accumulated more exopolysaccharides (glucans). To examine this, static biofilms for the wild-type, brpT mutant and complemented mutant were scraped from polystyrene wells. Glucans were detected using a phenol-sulfuric acid method. The resulting solutions from the biofilms consisted of a water-soluble fraction for the water-soluble (WSG) glucans and an alkali-soluble fraction for the water-insoluble glucans (WIG). As shown in Table 2, S. sanguinis formed a more robust biofilm when supplied with higher sucrose concentrations and the concentration of WIG was at least 10-times more than the WSG for all of the biofilms examined. More importantly, the brpT mutant formed significantly more biofilm than the wild-type or the complemented with a higher accumulation of both WIG and WSG regardless of the sucrose concentration ( S4 Fig). To compare the efficiency of glucan accumulation, the amount of glucans in each biofilm was normalized to the genomic DNA. As shown in Fig 4, the brpT mutant was more efficient at accumulating both WSG and WIG than either the wild type or the complemented mutant. These results suggest that S. sanguinis glucan synthesis is sucrose-dependent and biofilm development is influenced by brpT regulated glucan accumulation.

Transcriptome analysis of brpT regulated genes
To further investigate genes that may be influenced or regulated by brpT, RNA-seq was performed for a genome-wide transcriptome analysis. The brpT mutant and the wild-type SK36 were grown in BHI and cells were harvested at mid-log growth phase. RNA-seq data revealed 1074 genes with a significant change in expression compared to the wild-type; 594 genes were down-regulated and 480 genes up-regulated (S1 Table). Approximately 27% of the down-regulated and 29% of the up-regulated genes encoded hypothetical proteins, with a majority of those processing transmembrane domains and therefore membrane related. This was somewhat expected as membrane proteins are thought to play a crucial role in biofilm formation through cell-cell interaction, surface attachment, protein binding, nutrient transport and enzymatic activity [30]. A pathway analysis through DAVID showed a statistically significant enrichment for ribosomal proteins, translation initiation factors, tRNA synthetases and   [31]. This was followed by a significant enrichment of genes involved in fatty acid and lipid metabolism. Of the up-regulated genes there were a number of membrane associated proteins such as lipoproteins, penicillin-binding proteins, histidine transport permeases and genes for ethanolamine metabolism. We noted several multidrug ABC transport systems within the down-regulated genes. The transcriptome analysis suggested many genes are regulated by brpT.
Notably, RNA-sequencing for the brpT mutant showed more than a 9-fold increase in gtfP expression compared to the wild-type. This was verified by qRT-PCR where an 8-fold increase was observed. When BHI was supplemented with 1% sucrose and gtfP expression was re-analyzed by qRT-PCR, we observed a 15-fold increase in expression (Fig 5). This indicated the expression level of gtfP in the brpT mutant was significantly stimulated by sucrose (1.9-fold, brpT Modulates Biofilm in Streptococcus sanguinis P <0.01), whereas neither the wild-type nor the complemented mutant showed a significant increase in gtfP levels after sucrose stimulation. These results suggest that brpT has a role in repressing the expression of gtfP and de-repressed gtfP expression in the brpT mutant is stimulated by sucrose.
To further link the expression of glucosyltransferase to our increased biofilm phenotype in ΔbrpT, we next examined the single deletion mutant, ΔgtfP [18] and the double mutant ΔbrpTΔgtfP. While both mutants could form pellicles on the bottom of the well similar to the wild type, the pellicles were loosely attached to the polystyrene surface and were easily washed away (Fig 6A). In addition, the biofilm biomass ( Fig 6B) and the ability to synthesize glucans (Fig 6C and 6D) were significantly less than the wild type.
As ΔgtfP and ΔbrpT/ΔgtfP still had the ability to form biofilm pellicles but showed reduced biomass and a significantly lost in attachment, we hypothesize that brpT influences the ability of the S. sanguinis biofilm to strongly adhere to surfaces through the regulation of WSG and WIG accumulation within the biofilm.

Discussion
S. sanguinis is a pioneering colonizer within the oral cavity, initiating the establishment of disease-contributing plaque biofilm. Therefore, identifying and understanding genes that are involved in biofilm formation are critical to develop novel therapeutic strategies against oral diseases. Here, we present the first TetR family transcriptional regulator shown to influence biofilm formation in S. sanguinis. These transcriptional regulators are particularly interesting as they not only function as a repressor in controlling tetracycline resistance, but are shown to play key roles in regulating genes involved in a diverse range of adaptive responses [13,14]. From our continued studies with the S. sanguinis genome-wide mutant library [18], we noted that deletion of the TetR gene SSA_0144, which we named brpT, increased biofilm formation and altered the biofilm structure. We also observed an increase in the quantity of water soluble and insoluble glucans in the biofilms. This led us to examine the glucan synthesizing enzyme, glucosyltransferase (gtfP). Interestingly, RNA-seq analysis showed the expression of gtfP was significantly up-regulated in the brpT mutant and deletion of gtfP in S. sanguinis significantly Bacterial exopolysaccharides are a primary component of extracellular polymeric substances (EPS) or matrices in biofilms, and their role in the development and stabilization of biofilms have been extensively studied [2,[32][33][34]. Sucrose obtained from the diet and utilized by oral streptococci is considered a major factor in dental caries as it is fermentable and can serve as a substrate for extracellular and intracellular polysaccharides [35,36]. It has been noted that S. mutans, a key contributor to dental caries, utilizes three glucosyltransferases for the conversion of sucrose into polysaccharides. The glucosyltransferases, GtfB and GtfC, catalyze the synthesis of water-insoluble glucans, and GtfD, is responsible for water-soluble glucan synthesis [37,38]. GtfP, the sole S. mutans Gtf homolog in S. sanguinis, exhibits 59%, 49% and 49% protein sequence similarity to S. mutans GtfD, GtfB and GtfC, respectively, and has been revealed to play a role in the synthesis of WSG for S. sanguinis [39]. Our RNA-seq and subsequent qRT-PCR analyses for the brpT mutant showed a significant increase in expression of gtfP, indicating a possible relationship between brpT and gtfP in exopolysaccharide production and biofilm formation. When gtfP was knocked out from either the wild-type or the ΔbrpT strain it formed a weak biofilm that could only form pellicles loosely attached to a polystyrene surface (Fig 6A). While the brpT mutant accumulated significantly more WSG and WIG than the wild type or the complemented mutant ( Table 2), deletion of gtfP reduced the ability of S. sanguinis to synthesize both WSG and WIG by about 2-fold and 20-fold, respectively (Fig 6C  and 6D). The sole glucosyltransferase in S. sanguinis not only promotes WSG synthesis, but more importantly, determines the biogenesis of WIG in the biofilm, which may be the predominant contributor to biofilm adhesion on abiotic surfaces. This correlates with previous findings that formation of S. mutans microcolonies on saliva-coated hydroxyapatite surfaces was determined largely by the gtfB (associated with WIG) or gtfC (associated with WIG and WSG) and gtfB [40]. While these results indicate that brpT influences biofilm formation in a glucan dependent manner, mediated through the expression of gtfP, it is not clear whether BrpT directly or indirectly regulates gtfP. A bioinformatic prediction presented few known TF binding sites at 77 bp-(argR2), 75 bp-(ihf), and 36 bp-(rpoD16) upstream of the GtfP coding sequence but it is unclear as to whether BrpT recognizes these sites.
The contribution of exopolysaccharides to the biofilm three-dimensional architecture and modulation of inter-biofilm interactions has been documented in S. mutans [41,42]. Here we also observed that alteration in exopolysaccharide content led to changes in the S. sanguinis biofilm structure. The brpT mutant produced high levels of glucans and exhibited a rough and uneven biofilm surface. As shown by the CLSM analysis, an increase in glucan synthesis led to an increase in biofilm thickness, indicating the importance of glucans in the maintenance of S. sanguinis biofilm spatial structure. Furthermore, SEM analysis gave us possible explanations for these structural changes. The bacterial chains of the brpT mutant biofilm were bound by numerous filaments. Since the biofilm was washed by PBS prior to fixation, we hypothesized that these filamentous substances were composed mainly of water insoluble glucans. And the wild-type S. sanguinis biofilm with normal quantities of filaments showed a relatively flat and thin biofilm. Collectively, these observations implicate brpT as a novel regulator of S. sanguinis biofilm structure through glucan biosynthesis.
Although many factors have been characterized to modulate the expression of Gtf and biofilm formation, the mechanisms involved in these processes still need further illustrations. In S. mutans, the expression of gtf depends on quorum sensing [43], carbohydrate sources, pH [44], vicRK [45], frp [46], vicX [47] as well as chemical agents [48][49][50]. Unlike S. mutans, S. sanguinis biofilm formation was independent of AI-2 quorum sensing [51], and vicRK exhibited a negative regulation of gtfP transcription at certain growth stages [52], making the regulation S. sanguinis gtfP more complicated. Here, our data suggest that gtfP in S. sanguinis is negatively regulated by the novel TetR family repressor, brpT. Considering that exopolysaccharides can bind to the bacterial chain and maintain the spatial structure, which is essential for the biofilm integrity, the right expression levels of exopolysaccharides might be critical for biofilm development and maturation. If exopolysaccharides on the cell surface are low, like the gtfP mutant, the ability to attach to an abiotic surface would be lost and biofilm development impaired. However, bacteria that produce too much exopolysaccharides might have a negative influence. This would hamper bacteria movement and biofilm diffusion, as with the brpT mutant biofilm, where many empty areas were observed surrounding bright green clusters (Fig 2). Were bacteria bound tightly by large amounts of sticky glucans on the cell surface, restricting biofilm diffusion? Interestingly, a similar concern was raised previously. A P. aeruginosa mutant, ΔsadC, produced less Pel polysaccharides and exhibited increased swarming ability, even in high-viscosity medium [53]. In light of the findings described here, the possibility should be considered that brpT and gtfP cooperate, modulating S. sanguinis biofilm formation and dispersion. Further studies may obtain more evidence to elucidate this theory.
One of the primary findings of the present study is that the GtfP-catalyzed glucan synthesis in S. sanguinis biofilm is controlled by the TetR repressor BrpT. As the BrpT homologs are widely spread in oral Streptococci (S1 Fig), it is possible that the role of TetR in exopolysaccharides production and biofilm formation is conserved. Given that exopolysaccharides can facilitate the expression of virulence factors in mixed-species oral biofilms [42], further investigations into the relationship between brpT and gtfP may generate new insight into oral biofilm development and provide new targets for the design of effective anti-caries therapeutics. For example, methods for increasing the expression or activity of brpT homologs in oral pathogens may aid in inhibiting biofilm formation and accelerate pathogen clearance.