An Extracytoplasmic Function Sigma/Anti-Sigma Factor System Regulates β-Glucanase Expression in Tannerella forsythia in Response to Fusobacterium nucleatum Sensing

ABSTRACT

The periodontal pathogen Tannerella forsythia expresses a β-glucanase (TfGlcA) whose expression is induced in response to Fusobacterium nucleatum, a bridge bacterium of the oral cavity. TfGlcA cleaves β-glucans to release glucose, which can serve as a carbon source for F. nucleatum and other cohabiting organisms. A two-gene cluster encoding a putative extracytoplasmic function (ECF) sigma factor and a FecR-like anti-sigma factor has been recognized upstream of a TfGlcA operon. We characterized and analyzed the role of these putative ECF sigma and anti-sigma factors in the regulation of TfGlcA expression. For this purpose, deletion mutants were constructed and analyzed for β-glucanase expression. In addition, an Escherichia coli-produced ECF sigma factor recombinant protein was evaluated for transcriptional and DNA binding activities. The results showed that the recombinant protein promoted transcription by the RNA polymerase core enzyme from the glcA promoter. Furthermore, in comparison to those in the parental strain, the β-glucanase expression levels were significantly reduced in the ECF sigma-factor deletion mutant and increased significantly in the FecR anti-sigma factor deletion mutant. The levels did not change in the mutants following coincubation with the F. nucleatum whole cells or cell extracts. Finally, the levels of β-glucanase produced by T. forsythia strains paralleled F. nucleatum biomass in cobiofilms. In conclusion, we identified a β-glucanase operon regulatory system in T. forsythia comprising an ECF sigma factor (TfSigG) and a cognate FecR-like anti-sigma factor responsive to F. nucleatum and potentially other stimuli.

IMPORTANCE Previous studies have shown that F. nucleatum forms robust biofilms with T. forsythia utilizing glucose from the hydrolysis of β-glucans by T. forsythia β-glucanase, induced by F. nucleatum. In this study, we showed that a regulatory system comprising of an ECF sigma factor, TfSigG, and a FecR-like anti-sigma factor, TfFecR, is responsible for the β-glucanase induction in response to F. nucleatum, suggesting that this system plays roles in the mutualistic interactions of T. forsythia and F. nucleatum. The findings suggest the development and potential utility of small-molecule inhibitors targeting the β-glucanase activity or the TfSigG/TfFecR system as therapeutic drugs against dental plaque formation and periodontitis.

INTRODUCTION

Periodontitis, a chronic inflammatory disease of the tooth-supporting tissues that often leads to tooth loss, is initiated by a subgingival multispecies biofilm. Tannerella forsythia and Fusobacterium nucleatum, found frequently in subgingival biofilms, are strongly implicated in the development of periodontitis (1, 2). T. forsythia is known to express several putative virulence factors that can either promote colonization or trigger inflammation (3). F. nucleatum is considered a bridge bacterium that facilitates plaque biofilm development due to its ability to coaggregate with both the early- and late-colonizing species (4). T. forsythia and F. nucleatum are also considered strong risk factors for the progression of extraoral diseases such as esophageal and colorectal cancers, respectively (5, 6). In vitro, T. forsythia has been shown to coaggregate and form synergistic cobiofilm with F. nucleatum (7). In human dental plaque, these species are found in a close physical proximity of each other (8), thus pointing to mutualistic relationship of the two species.

We have shown that F. nucleatum or its cell-free sonicated F. nucleatum extract can induce β-glucanase expression in T. forsythia, resulting in the hydrolysis of β-glucans into glucose as a nutrient for F. nucleatum growth (9). The T. forsythia β-glucanase-coding gene, TfglcA (BFO_0186), is present in a 3-gene operon located immediately downstream of two open reading frames (ORFs) coding for a putative extracytoplasmic function (ECF) sigma factor (BFO_0190) and an anti-sigma factor, (BFO_0189) (9). The ECF sigma factors are often associated with gene regulatory networks involved in sensing environmental cues to regulate the activity of core RNA polymerase (RNAP) complex for growth and survival in bacteria (10). They are involved in sensing environmental stresses in both Gram-positive and Gram-negative bacteria, such as in response to antibiotics (11), heavy metals (12), oxidative stresses (13), osmolytes (14), or temperature shock (15). In the resting state, ECF sigma factors often remain in a locked, inactive state by the binding of cognate anti-sigma factors (16). Extracytoplasmic function (ECF) sigma factors, also known as group 4 sigma factors, play key roles in the regulation of many vital functions in response to an environmental stimulus. They are identified based on the presence of two conserved regions: σ2 and σ4, spaced by approximately 50 amino acids, and by the lack of the σ3 region of the σ70 protein (17). Similar to the two-component systems (TCSs) comprising a sensor kinase and response regulator, an ECF σ factor and its cognate anti-σ factor form a functional regulatory unit. ECF sigma and anti-sigma genes are often present close to the genes they regulate, although this not always the case. In Gram-negative bacteria, the activity of an ECF sigma factor is often regulated by a cytoplasmic membrane anti-sigma factor, whose activity, in turn, is regulated by a cell surface signaling (CSS) cascade mediated by an outer membrane receptor (18). Thus, CSS enables Gram-negative bacteria to transduce an environmental signal into a transcription response via mediation of an ECF (19).

In this study, we sought to confirm the functional identity of the BFO_0190 and BFO_0189 proteins by constructing and analyzing mutants with deletions of genes encoding the respective proteins, and by functional analysis of a recombinantly expressed BFO_0190 protein. Our data showed that BFO_0190 is a bona fide sigma factor which together with BFO_0189 forms an F. nucleatum-responsive ECF sigma/anti-sigma factor system for glucanase regulation in T. forsythia.

RESULTS

BFO_0190 codes for an ECF sigma factor.

Analysis of the BF0_0190 protein at InterProScan (www.ebi.ac.uk/interpro/search/sequence/) revealed the presence of subdomains r2 (residues 50 to 115) and r4 (residues 148 to 197) (see Fig. S1 in the supplemental material) within sigma factor 70. The predicted r4 subdomain of BFO_0190 contains a characteristic helix-turn-helix (HTH) DNA binding domain, and the r2 subdomain contains helical domains potentially involved in core RNA polymerase recognition. In silico analysis of the BFO_0189 protein predicted the presence of an N-terminal cytoplasmic domain (residues 1 to 96 [Fig. S2]), a transmembrane domain (transmembrane helix residues 97 to 118 [Fig. S2]), a FecR-like domain (residues 129 to 223 [Fig. S2]), and a C-terminal FecR_C domain (residues 267 to 333 [Fig. S2]) likely to be periplasmically localized. In order to investigate the conformation of the BFO_0190 and BFO_0189 protein, the Phyre2 protein fold server was employed to generate its three-dimensional (3D) structure (Fig. S3). The top 3 homologous proteins with corresponding protein models resulting from this analysis are listed in Table S1. Template c6in7B (based on the crystal structure of RNA polymerase sigma-h factor) with the highest identity percentage, and 100% confidence, was selected for BFO_0190 structural analysis and the generated PDB file was reconstructed using Chimera Molecular Graphics and Analysis software (www.rbvi.ucsf.edu/chimera) to visualize the BFO_0190 3D structure (Fig. S3A). This 3D model indicates a conserved sigma factor, subdomain r2, in the N terminus and an r4 domain with HTH motif in the C terminus, similar to sigma-h factor domains. To confirm that the BFO_0190 protein functions as a sigma factor, we tested the DNA binding and sigma factor activity of a recombinant BFO_0190 protein expressed in Escherichia coli. Briefly, BFO_0190 ORF was cloned into the pET-30a (+) cloning vector and protein expression was induced in the E. coli BL21 strain. The expressed protein purified by nickel-nitrilotriacetic acid (Ni-NTA) resin migrated as a single band on an SDS-PAGE gel with a molecular weight of 23 kDa (Fig. S4). The purified recombinant BFO_0190 protein (recombinant T. forsythia SigG [rTfSigG]) was then used in an in vitro transcription runoff assay with E. coli RNAP as a core enzyme, ribonucleotides, and a DNA fragment encompassing the glcA operon promoter as a template. This reaction led to a transcription product of the expected size (~1,144 nucleotides [nt]) in the case of the reaction mixture containing rTfSigG as judged by electrophoresis (Fig. 1, lane 1). DNase I treatment did not affect the intensity of the expected band, but treatment with RNase A did (Fig. 1, lane 2 and 3). The DNA binding ability of the recombinant BFO_0190 (rTfSigG) protein was confirmed by an electrophoretic mobility shift assay (EMSA) using a Cy5.5-labeled 492-bp DNA fragment encompassing the promoter region of the glcA operon. As shown in Fig. 2, incubation of the rTfSigG protein retarded the mobility of the DNA fragment. Together, the data demonstrated that rTfSigG, by binding to the promoter of the glcA operon, promotes the transcription activity of core RNAP. The data confirmed that BFO_0190 is a bona fide sigma factor, here referred to as TfSigG.

FIG 1.

In vitro runoff transcription assay with rTfSigG. The in vitro transcription runoff assay was performed with a DNA fragment encompassing the putative promoter region along with the first gene of the glcA operon (2 pM) in the presence of increasing amounts of rTfSigG protein (10 pM) with E. coli core RNA polymerase. Transcription products were resolved on an 1.2% denaturing formaldehyde agarose gel and stained with RedSafe nucleic acid staining dye. This gel is representative of three independent experiments.

FIG 2.

Electrophoretic mobile shift assay (EMSA) of the binding rTfSigG protein and TfragA promoter region DNA. The EMSA was carried out with an rTfSigG protein (+, 6 pM; ++, 15 pM; +++, 30 pM; ++++, 36 pM [per reaction]) mixed with Cy5.5-labeled glcA operon promoter region DNA fragment (0.25 pM/reaction) with or without an unlabeled DNA fragment (+, 0.25 pM; ++, 1.25 pM; +++, 2.5 pM [per reaction]). Mixed samples were loaded onto a 6% native acrylamide gel and electrophoresed at 200 V for 25 min.

TfSigG ECF sigma factor and FecR-like anti-sigma factor homolog BFO_0189 regulate F. nucleatum-dependent T. forsythia β-glucanase expression.

Next, we investigated the role of TfSigG and its putative FecR-like anti-sigma factor BFO_0189 in β-glucanase regulation. Anti-sigma factors typically have a small cytoplasmic domain which is often linked by one or sometimes more transmembrane regions (TMR) to a C-terminal periplasmic domain. FecR-like proteins function as anti-sigma factors for FecI-like ECF sigma factors (20). These proteins harbor a conserved periplasmic FecR domain, which is important for stimulus perception. The FecR-like anti-sigma factor homolog, BFO_0189, has a characteristic FecR-like domain, as well as a predicted periplasmic domain and a transmembrane domain typical of anti-sigma factors. To identify the roles of TfSigG and the putative anti-sigma factor BFO_0189 in glucanase expression, we constructed mutants with deletions of respective proteins and analyzed the expression of β-glucanase in bacterial cell extracts by zymography using the β-glucan lichenin as the substrate. Our results showed that compared to that of the T. forsythia parent strain ATCC 43037 (Fig. 3A, lane 1), the β-glucanase activity was significantly reduced in the TfΔsigG BFO_0190 deletion mutant (Fig. 3A, lane 2), and increased significantly in the TfΔfecR BFO_0189 deletion mutant (Fig. 3A, lane 3).

FIG 3.

(A) Detection of β-glucanase activity in T. forsythia cells by gel zymography. Bacterial cell lysate (100 μg/lane) from each strain was separated by 10% SDS-PAGE with 0.1% lichenin, followed by Congo red staining. Lanes: 1, T. forsythia 43037; 2, T. forsythia BFO_0190 sigG deletion mutant; 3, T. forsythia BFO_0189 fecR deletion mutant. (B) Detection of β-glucanase activity in T. forsythia in response to F. nucleatum cell extract by microplate zymography assay. T. forsythia cells were incubated with F. nucleatum cell extracts (10 mg/mL), and β-glucanase activity in cell lysates of T. forsythia was assayed by lichenin-agar zymography assay in a 96-well microplate. Untreated lichenin-agar wells were used as background, and β-glucanase enzyme-treated wells were used as positive controls. Bars represent the percent hydrolysis, calculated by the equation in Materials and Methods from a representative experiment repeated 3 times yielding similar results. *, P < 0.05. Fn−, without Fn cell extract; Fn+, with Fn cell extract; WT, wild type.

We have previously shown that the expression of β-glucanase is induced in T. forsythia when the bacterium senses whole cells or cell extracts of F. nucleatum (9). In the present study, 96-well microplate zymography assays showed that compared to the parent T. forsythia strain, β-glucanase activity in a TfΔsigG deletion mutant significantly reduced and was not induced by F. nucleatum cell extracts. On the other hand, the β-glucanase activity was significantly elevated in the TfΔfecR deletion mutant in comparison to that in the wild-type strain, and the glucanase activity did not change in the presence of F. nucleatum extracts (Fig. 3B).

The above results showed that BFO_0190 (TfSigG) can promote glcA operon transcription to express β-glucanase and that BFO_0189 acts as a negative regulator of TfSigG ECF sigma factor. BFO_0189, therefore, functions as the anti-sigma factor, here referred to as TfFecR.

Effect of TfsigG and TffecR inactivation on T. forsythia-F. nucleatum cobiofilm formation in the presence of β-glucan.

We evaluated the biofilm formation ability of the T. forsythia mutant strains with F. nucleatum in the presence of β-glucan (lichenin), as we have described previously (9). Biofilm formation was evaluated by in situ hybridization with bacterium-specific probes and biofilm; biomass was quantified by Comstat 2.0 software (9). In lichenin-containing cobiofilm, F. nucleatum showed the highest increase in its biomass (15.908 μm³/μm²) with the TfΔfecR mutant; this increase was significantly higher than with the parent strain (12.443 μm³/μm²) or the TfΔsigG mutant (9.014 μm³/μm²) (Fig. 4A). The increase of F. nucleatum biomass in cobiofilm paralleled the levels of glucanase secretion by T. forsythia strains and hence the availability of glucose from lichenin hydrolysis. In the absence of lichenin, the F. nucleatum biomass in cobiofilm did not show significant differences (data not shown). Under similar conditions, the biomass of T. forsythia, whose growth is not supported by glucose (9), did not differ among T. forsythia strains (Fig. 5A). The total biomass (T. forsythia plus F. nucleatum) showed the same trend as the F. nucleatum biomass; the TfΔfecR mutant with F. nucleatum showed a higher biomass (19.076 μm³/μm²) than the T. forsythia parent strain with F. nucleatum (15.542 μm³/μm²) or the TfΔsigG mutant with F. nucleatum (11.371 μm³/μm²) (Fig. 6A). The F. nucleatum biofilm biomass in monospecies growth was significantly lower than in the cobiofilm growth with the wild-type and fecR deletion mutant of T. forsythia, supporting the above results showing that glucose production in biofilms is responsible for F. nucleatum growth (Fig. S6). On the other hand, no significant differences were observed for T. forsythia biomasses under monospecies and cobiofim conditions (Fig. S6 and S7). Interestingly, a trend toward reducing T. forsythia biomass was observed in the monospecies growth compared to that in the cobiofim growth. These data validate our previous findings that T. forsythia and F. nucleatum form robust biofilms under coculture growth (7).

FIG 4.

Biomass of F. nucleatum in T-medium with 0.2% lichenin determined by in situ staining using bacterium-specific probes and confocal laser scanning microscopy. F. nucleatum was stained with an F. nucleatum-specific Cy3-labeled DNA probe. Confocal images were taken at ×200, and F. nucleatum volumes were estimated from confocal images with Comstat 2 software. The biomass volume from one representative experiment (mean ± SD) (A) and representative confocal images of T. forsythia WT plus F. nucleatum (B), T. forsythia ΔsigG mutant plus F. nucleatum (C), and T. forsythia ΔfecR mutant plus F. nucleatum (D) are shown.

FIG 5.

Biomass calculation and confocal laser scanning microscope pictures of T. forsythia in cobiofilm with T. forsythia and F. nucleatum. The biomass of T. forsythia in T-medium with 0.2% lichenin was calculated by in situ staining using bacterium-specific probes followed by confocal laser scanning microscopy as described in Materials and Methods and the legend to Fig. 4. T. forsythia was stained with FITC-labeled specific DNA probe. Confocal images were taken at ×200. T. forsythia volumes were estimated from confocal images with Comstat 2 software and biomass volume calculation from one representative experiment (mean ± SD) (A), and representative confocal images of T. forsythia WT plus F. nucleatum (B), T. forsythia ΔsigG mutant plus F. nucleatum (C), and T. forsythia ΔfecR mutant plus F. nucleatum (D) are shown.

FIG 6.

Total bacteria (F. nucleatum plus T. forsythia) in cobiofilms in T-medium with 0.2% lichenin were calculated by in situ staining using bacterium-specific probes followed by confocal laser scanning microscopy as described above. T. forsythia was stained with FITC-labeled DNA probe (green), and F. nucleatum was stained with Cy5-labeled DNA probe (red). Both confocal images were taken at ×200. F. nucleatum and T. forsythia biomass volumes were estimated from confocal images with Comstat 2 software. For each panel, graphs of biomass volume were calculated from one representative experiment (mean ± SD) (A), and representative confocal images of T. forsythia WT plus F. nucleatum (B), T. forsythia ΔsigG mutant plus F. nucleatum (C), and T. forsythia ΔfecR mutant plus F. nucleatum (D) are shown.

DISCUSSION

We identified a two-gene cluster coding for a putative extracytoplasmic function (ECF) sigma factor (BFO_0190) and a FecR-like anti-sigma factor (BFO_0189) upstream of an operon that expresses glucanase enzyme TfGlcA in T. forsythia. It has been shown previously that the expression of TfGlcA is induced in response to F. nucleatum recognition during cobiofilm setting (9). In this study, the functionality of the BFO_0190 gene product as a sigma factor (TfSigG) was established by confirming the transcriptional and promoter binding activities of a recombinant TfSigG protein. Moreover, the current study investigated the role of the TfSigG sigma factor and FecR-like homolog TfFecR in the regulation of β-glucanase expression. The data showed that the β-glucanase activity was significantly reduced in a TfSigG-deficient mutant compared to that of the T. forsythia parent strain. On the other hand, a TfFecR mutant exhibited significantly higher levels of β-glucanase activity than did the wild-type strain. Moreover, the expression levels of TfGlcA in the mutants were not altered by F. nucleatum whole cells or F. nucleatum cell extracts, indicating that the glcA operon is regulated by the upstream SigG-FecR regulatory proteins.

The number of ECF sigma factors encoded by a bacterial genome can vary from 0 to over 100 (21). In T. forsythia strain ATCC 43037, approximately 20 putative ECF sigma factors have been predicted based on genome analysis (Integrated Microbial Genome [https://img.jgi.doe.gov/]) (Table S2). It would be interesting to determine what roles each of the ECF factors play in the biology of T. forsythia. We hypothesize that the TfSigG/TfFecR regulators play a critical role in β-glucanase regulation and thus the pathogenicity of T. forsythia. This notion is supported by the fact that the sigG/fecR-glcA genetic locus is well conserved in the genomes of the clinical isolates of T. forsythia we analyzed (Fig. S5). Thus, there seems to be a strong selective advantage to maintain this locus in the genome.

While our study focused on the role of TfSigG/TfFecR regulatory systems in the regulation of F. nucleatum-dependent β-glucanase expression, it is likely that this system is involved in sensing other environmental factors and bacteria and in regulating the expression of other key proteins as well in T. forsythia. It has been shown that a single ECF factor can sometime regulate multiple downstream genes in bacteria (22). For example, in the periodontal pathogen Porphyromonas gingivalis, it has shown that the ECF-sigma-anti-sigma factor can regulate multiple downstream targets (23). Biological functions performed by ECF factors include modulation of envelope stress response (e.g., E. coli σ^E and Bacillus subtilis σ^w), iron transport (E. coli σ^FecI and Pseudomonas aeruginosa σ^PvdS), oxidative stress (Streptomyces coelicolor σ^R and Rhodobacter sphaeroides σ^E), and general stress response (e.g., Methylobacterium extorquens σ^EcfG1) (17).

With regard to the two ORFs located upstream of the glcA gene in the operon, BFO_0188 and BFO_0187 are predicted to code for SusC (RagA)-type and SusD (RagB)-type outer membrane proteins, respectively. RagA and RagB proteins form outer membrane protein complexes with diverse functions, playing roles in either nutrient acquisition, environmental sensing, or immune regulation, in Gram-negative bacteria (24–26). BFO_0188 (SusC/RagA) is a predicted TonB-dependent outer membrane receptor, and BFO_0187 (SusD/RagB) is a predicted glycan binding protein. T. forsythia genome encodes a plethora of RagA/RagB orthologs, and the definitive functions of these proteins are yet to be elucidated. We plan to determine if BFO_0187 and BFO_0188 proteins might be involved in the regulation of glucanase activity in T. forsythia.

In summary, our study identified a regulatory system comprising a sigma factor, TfSigG, and a cognate anti-sigma factor, TfFecR, in the sensing and regulation of F. nucleatum-dependent β-glucanase expression in T. forsythia. It would be supporting to determine if the same system is involved in the modulation of other pathways and virulence factors in T. forsythia.

MATERIALS AND METHODS

Bacterial strains, genes, and plasmids.

T. forsythia strains (T. forsythia ATCC 43037 and T. forsythia BFO_0190 and BFO_0189 inactivation mutants generated in this study) were maintained in BF broth or agar with added antibiotics as needed (27); F. nucleatum ATCC 25586 was maintained in Trypticase soy agar or broth (TS agar or TSB; Difco BD, Franklin Lakes, NJ) containing 5% horse blood. E. coli strains were cultured in LB-Miller medium or agar (Difco, BD, Franklin Lakes, NJ) with appropriate antibiotics. For gene cloning, pET-30a (Millipore Sigma, St. Louis, MO) was used for recombinant protein expression.

The gene identifiers (IDs) used in this report are based on T. forsythia strain 92A2, whose genome has been fully sequenced and assembled (GenBank accession number NC_016610.1). The genome sequence of T. forsythia strain ATCC 43037 used in this study is available in a draft form (GenBank accession number JUET00000000.1). BFO_0190 and BFO_0189 are identified as Tanf_09195 (98% amino acid sequence identity [99% similarity] to the BFO_0190 protein) and Tanf_09200 (95% amino acid sequence identity [97.5% similarity] to the BFO_0189 protein), respectively, in ATCC 43037.

Cloning, expression, and purification of BFO_0190 (rTfSigG) protein.

An expression plasmid, pET-rTfsigG, was generated by cloning the BFO_0190 ORF in frame in the pET-30a(+) expression vector. Briefly, a PCR fragment amplified with primers rTfsigGF and rTfsigGR (Table S3) from T. forsythia ATCC 43037 genomic DNA was cloned into the NdeI and XhoI region of the pET-30a(+) vector to obtain the plasmid pET-rTfsigG. For protein expression, the Escherichia coli BL21(DE3) strain carrying pET-rTfsigG was grown in 500 mL of LB medium with kanamycin (30 μg/mL) at 37°C to an optical density at 600 nm (OD₆₀₀) of 0.5. Protein expression was induced with isopropyl β-d-1-thiogalactopyranoside (IPTG; final concentration of 0.5 mM) for an additional 3 h at 37°C. Bacteria were collected by centrifugation at 7,000 × g for 10 min, washed with phosphate-buffered saline (PBS) twice, resuspended in 15 mL of lysis buffer (50 mM NaH₂PO₄H₂O, 300 mM NaCl, 10 mM imidazole [pH 8.0], 1 mg/mL of lysozyme, and 1% Triton X-100), and incubated for 2 h at 37°C. After incubation, bacteria were homogenized by a French press (G-M; Glen Mills, Clifton, NJ) at 2,500 lb/in². E. coli lysate was centrifuged at 7,500 × g for 20 min, and supernatants were bound to 3 mL of Ni-NTA His Bind Superflow resin (EMD Millipore) equilibrated with lysis buffer (50 mM NaH₂PO₄H₂O, 300 mM NaCl, and 10 mM imidazole [pH 8.0]) for 1 h at 4°C. The resin was washed with wash buffer (50 mM NaH₂PO₄H₂O, 300 mM NaCl, and 20 mM imidazole [pH 8.0]), followed by an elution with an elution buffer (50 mM NaH₂PO₄H₂O, 300 mM NaCl, and 250 mM imidazole [pH 8.0]). Samples were collected in 1.5-mL fractions, and the rTfSigG protein content in the eluted fractions was confirmed by SDS-PAGE (Fig. S4).

Runoff transcription assay.

The runoff transcription assay was performed as per standard protocol (23, 28), in the presence or absence of the purified rTfSigG protein. Briefly, the DNA fragment encompassing a region from 300 nucleotides to the transcription start site along with the first gene (BFO_ 0188) of the glcA operon was amplified from the T. forsythia genomic DNA with the primers TfSig_Runoff_F and TfSig_Runoff_R (Table S3) to generate a 1,436-bp DNA template. The reaction mixture contained an E. coli RNA polymerase core enzyme(s) (Epicentre, WI), the DNA template (2 pM), and the purified rTfSigG protein (10pM) in 50 mM Tris-HCl (pH 7.5), 150 mM KCl, 10 mM MgCl₂, 0.1 mM dithiothreitol (DTT), 0.01% Triton X-100, and 2.5 mM each nucleoside triphosphate (NTP). The transcription mixture was incubated at 37°C for 2 h, and the products were analyzed on denaturing formaldehyde agarose gels visualized with RedSafe nucleic acid staining dye (iNtRon Biotechnology) as per the manufacturer’s recommendations. In some cases, the reaction products were treated with either RNase A or DNase I prior to electrophoresis to validate that the product was indeed an RNA transcript.

EMSA.

To confirm the ability of rTfSigG to bind to the T. forsythia β-glucanase promoter, an electrophoretic mobility shift assay (EMSA) was performed. Briefly, a Cy5.5-labeled 492-bp DNA probe containing the promoter region (nt position −166 bp to the transcription start site [TSS] of the glcA operon) was PCR amplified from genomic DNA with Cy5.5-labeled primers TfragA-promF and TfragA-promR (Table S3). The promoter region of the TfglcA operon was selected from the result of the 5′ rapid amplification of cDNA ends (RACE)-PCR that was carried out in our previous study (9). The standard protein-DNA binding reaction mixture contained 0.25 pM Cy5.5-labeled DNA and 6 pM to 36 pM purified rTfSigG in the binding buffer [10 mM Tris-HCl, 50 mM KCl, 1 mM DTT, 100 mg of bovine serum albumin (BSA) per ml, 100 mg of poly(dI-dC) per ml, and 5% (vol/vol) glycerol (pH 7.0)]. For competitive inhibition, 0.25 pM to 2.5 pM unlabeled DNA fragments was used. After incubation for 20 min at room temperature, protein-DNA complexes were separated on 6% native polyacrylamide gels run in TBE buffer (22 mM Tris base, 22 mM boric acid, and 0.5 mM EDTA [pH 8.0]) at 200 V for 45 min. The unbound DNA and protein-bound DNA were visualized by a Cy5.5 detection filter using a ChemiDoc system (Bio-Rad).

Construction of T. forsythia BFO_0190 and BFO_0189 inactivation strains.

Deletion mutants were constructed by an allelic-replacement strategy with an erythromycin (EM) cassette, as we have described previously (29). Briefly, a DNA fragment containing the ermF gene flanked by upstream and downstream DNA regions of BFO_0190 were generated by an overlap extension PCR strategy. DNA sequences of PCR primers used for the construction are listed in Table S3. The 5′ and 3′ regions of BFO_0190 were amplified with primer sets SigG 1F/SigG 3R and SigG 6F/SigG 2R, respectively, from genomic DNA. The ermF fragment was amplified by PCR with primer set SigG 4F/SigG 5R using the pVA2198 plasmid (30) as a template. All three DNA fragments were combined, and an overlap PCR was carried out with primer set SigG 1F/SigG 2R to yield a 2,175-bp fragment. The DNA fragment for inactivating BFO_0189 was constructed and generated similarly. The 5′ and 3′ regions of BFO_0189 were amplified with primer sets FecR 1F/FecR 3R and FecR 6F/FecR 2R from the T. forsythia ATCC 43037 genomic DNA and ermF fragment, with a T. forsythia genomic DNA overlap region amplified by PCR with primers FecR 4F/FecR 5R from pVA2198, followed by overlap PCR to obtain a 2,684-bp DNA fragment with FecR 1F/FecR 2R. The purified PCR fragments were then used for the transformation of T. forsythia ATCC 43037 by electroporation as described previously (31). Transformants were plated onto BF agar plates containing 5 μg/mL of erythromycin and incubated at 37°C under anaerobic conditions for 10 to 14 days. Erythromycin-resistant colonies were screened by PCR and DNA sequences around the EM resistance cassette insertion site.

Positive clones were identified as BFO_0190 and BFO_0189 gene-inactivated clones and were subjected to β-glucanase analysis following confirmation that no polar effects occurred due to the BFO_0190 inactivation.

Detection of β-glucanase activity.

β-Glucanase activity of T. forsythia strains was assessed by agar microplate and PAGE zymography assays using lichenin as the β-glucan substrate as we have described previously (9, 32). Briefly, for microplate assays, 0.1% lichenin (MP Biomedicals, Irvine, CA) was included in melted 1.5% Bacto agar (BD), and 100 μL of lichenin-agar solution was aliquoted into 96-well plates, cooled to room temperature, and stored at 4°C until needed. T. forsythia cells grown to an OD₆₀₀ of 1.0 were harvested, washed twice with PBS, and adjusted to an OD₆₀₀ of 2.0. F. nucleatum cell-free crude extract was prepared by ultrasonic disruption at 26 W for 3 min on ice, followed by centrifugation (15,000 × g for 10 min at 4°C) and filtration through a 0.45-μm filter. For β-glucanase assay, 3 mL of T. forsythia cell suspension prepared as described above was incubated with F. nucleatum cell-free crude extract (10 mg/mL) for 18 h at 37°C anaerobically. After incubation, the T. forsythia cells were retrieved by centrifugation and washed twice with PBS, and bacterial suspension was adjusted to an OD₆₀₀ of 2.0. Cell pellets from 1 mL of cell suspensions were collected by centrifugation and lysed by incubation in 150 μL of 5% Triton X-100 at 25°C for 10 min. Lysed cells were centrifuged at 13,000 × g for 15 min at 4°C, and supernatants were collected by centrifugation. Ten-microliter volumes of supernatants from noninduced or F. nucleatum-induced T. forsythia cells were applied in quadruplicates to the lichenin-agar wells described above. Wells with only lichenin in agar were used as untreated background controls. Plates were incubated at 37°C overnight, washed twice with distilled water, and stained with 0.1% Congo red dye solution at room temperature for 20 min. Staining solution was removed and wells were washed twice with water, followed by a wash with 1 M NaCl for 30 min at room temperature. Lichenin hydrolysis due to β-glucanase activity was assessed by measuring absorbance at 500 nm (33). Glucan hydrolysis was presented as a percentage of the value obtained by the following equation: (OD₅₀₀ of negative-control well – OD₅₀₀ of T. forsythia extract-treated well)/(OD₅₀₀ of negative-control well).

PAGE zymography was carried out by the previously described protocol (9). Briefly, T. forsythia cells harvested from mid-log-phase cultures and washed with PBS were lysed in 5% Triton X-100–PBS on ice for 30 min. Cell supernatants were collected by centrifugation at 12,000 × g for 10 min at 4°C and subjected to SDS-PAGE on 10% acrylamide gels (10 by 7. 5 cm) containing 0.1% lichenin after mixing with equal amount of 2× SDS-PAGE Laemmli sample buffer without a reducing agent (50 μg of total protein per well). After electrophoresis, gels were incubated, with rocking, in 100 mM sodium phosphate (pH 5.5) containing 20% isopropanol for 20 min at room temperature. This step was repeated one more time to remove SDS, and gels were then incubated in 100 mM sodium phosphate (pH 5.5) buffer containing 10 mM 2-mercaptoethanol (2-ME) and 1 mM EDTA for 60 min at room temperature. Finally, gels were placed in a fresh phosphate (pH 6.0)–10 mM 2-ME–1 mM EDTA buffer for 45 min at 37°C to promote enzyme activity. Lichenin hydrolysis was visualized by staining with 0.1% Congo red as described above for agar plate assays.

Biofilm formation of T. forsythia strains with F. nucleatum.

The impact of ECF sigma and anti-sigma factor inactivation for T. forsythia glucan hydrolysis on the dual-species biofilms was assessed by fluorescent in situ hybridization (FISH) as described in the work of Thurnheer et al. (34) and our previous publication (9).

For the calculation of T. forsythia/F. nucleatum biomass in biofilms, confocal scanning laser microscopy (CSLM) analysis was performed with a Zeiss LSM 510 Meta NLO confocal microscope attached to a Zeiss Axioimager Z1 and Axiovert 200M. Briefly, T. forsythia ATCC 43037, T. forsythia sigG and fecR inactivation strains, and F. nucleatum cells were resuspended in T-L medium (1.7% tryptone, 0.5% sodium chloride, and 0.25% dipotassium phosphate), with N-acetylmuramic acid (NAM; 5 μg/mL; Chem-Impex International Inc., Wood Dale, IL), with 0.2% lichenin mixed T. forsythia to F. nucleatum at a 20:1 ratio, and then loaded into a chambered cell culture slide (1 mL/well; BD Falcon, Thermo Fisher) and incubated at 37°C for 48 h anaerobically (5% CO₂, 5% H₂, 90% N₂). The biofilms in each chamber were stained by the FISH protocol described above. Briefly, wells were washed twice with distilled water after incubation to remove unbound cells, and biofilm cells were fixed in 4% paraformaldehyde (PFA) for 30 min at room temperature. Fixed biofilms were treated with 100 μL/well of hybridization buffer (0.9 M NaCl, 20 mM Tris-HCl [pH 7.5], 0.01% SDS, and 20% formamide) at 46°C for 20 min for prehybridization blocking. Biofilms were then incubated with 100 μL/well of 0.1 nM (each) fluorescein isothiocyanate (FITC)-labeled T. forsythia or Cy3-labeled F. nucleatum-specific DNA probes (9) in hybridization buffer for 3 h at 46°C. After hybridization, wells were washed with 100 μL/well of wash buffer (20 mM Tris-HCl [pH 7.5], 5 mM EDTA, 0.01% SDS, and 200 mM NaCl) for 20 min at 49°C, followed by 200 μL/well of distilled water twice at 37°C for 15 min. The biofilm pictures were taken at a magnification of ×200 on a CSLM (excitation of 494 nm/emission of 522 nm for FITC and excitation 550 nm/emission of 570 nm for Cy3) and analyzed with Zeiss Zen 2 core software. T. forsythia (FITC) and F. nucleatum (Cy5) fluorescence intensities were calculated and exported to Comstat 2 software (http://www.comstat.dk/) (35) to determine the biofilm mass of each species.

Statistical analysis.

Prism 9 software (GraphPad Software) was used for all statistical analyses. Statistical significance was determined by a two-tailed paired or unpaired Student t test for two groups or one-way analysis of variance (ANOVA) with appropriate post hoc tests for multiple-group comparisons. A P value of less than 0.05 was considered statistically significant.