A two-component system regulates gene expression of the type IX secretion component proteins via an ECF sigma factor

The periodontopathogen Porphyromonas gingivalis secretes potent pathogenic proteases, gingipains, via the type IX secretion system (T9SS). This system comprises at least 11 components; however, the regulatory mechanism of their expression has not yet been elucidated. Here, we found that the PorY (PGN_2001)-PorX (PGN_1019)-SigP (PGN_0274) cascade is involved in the regulation of T9SS. Surface plasmon resonance (SPR) analysis revealed a direct interaction between a recombinant PorY (rPorY) and a recombinant PorX (rPorX). rPorY autophosphorylated and transferred a phosphoryl group to rPorX in the presence of Mn2+. These results demonstrate that PorX and PorY act as a response regulator and a histidine kinase, respectively, of a two component system (TCS), although they are separately encoded on the chromosome. T9SS component-encoding genes were down-regulated in a mutant deficient in a putative extracytoplasmic function (ECF) sigma factor, PGN_0274 (SigP), similar to the porX mutant. Electrophoretic gel shift assays showed that rSigP bound to the putative promoter regions of T9SS component-encoding genes. The SigP protein was lacking in the porX mutant. Co-immunoprecipitation and SPR analysis revealed the direct interaction between SigP and PorX. Together, these results indicate that the PorXY TCS regulates T9SS-mediated protein secretion via the SigP ECF sigma factor.

Scientific RepoRts | 6:23288 | DOI: 10.1038/srep23288 TCS is one of the most common signal transduction mechanisms in bacteria to sense and respond to environmental stresses. In many pathogens, the system is also known to modulate the transcription of genes associated with multiple pathogenic events such as invasion into host cells, biofilm formation, chemotaxis, lipopolysaccharide modification, and resistance to antibiotics [13][14][15][16] . TCS typically consists of two protein components: an HK, which autophosphorylates in response to environmental stimuli, and an RR, which receives the phosphate group from HK and controls the regulation of target gene expression.
Tiling DNA microarray analysis using a custom tiling DNA array chip with the genome sequence of P. gingivalis ATCC 33277 previously revealed that 20 genes were down-regulated in the porX deletion mutant to less than 60% of the wild-type parental strain 11 . These 20 genes included porT, sov, porK, porL, porM, porN, and porP of T9SS. Reverse transcriptase (RT)-PCR analysis confirmed that these genes were down-regulated in the porY mutant, suggesting that PorX and PorY have roles in regulating the expression of T9SS component-encoding genes. Usually, HK-and RR-encoding genes are located tandemly, or nearby on a chromosome. However, porX and porY are 'orphan' genes because they are located separately on the P. gingivalis chromosome 17 .
To mediate adaptation processes in response to environmental signals, bacteria possess multiple signalling pathways. Extracytoplasmic function (ECF) sigma factors are involved in one of the major bacterial signalling pathways as well as TCS 18 . ECF sigma factors, which belong to the sigma-70 family of proteins, facilitate the alteration of gene expression by binding to and guiding the core RNA polymerase. Six ECF sigma factors, PGN_0274, PGN_0319, PGN_0450, PGN_0970, PGN_1108, and PGN_1740, are encoded by the P. gingivalis ATCC 33277 genome. Previous studies have indicated that PGN_0274 and PGN_0450 in strain ATCC 33277 (PG0162 and PG1660 in strain W83, respectively) are involved in the expression of gingipains 19,20 , that PGN_1740 (SigH, PG1827 in W83) is required for the adaptation to oxygen stress 21 , that PGN_1108 (PG1318 in W83) has a role in the regulation of mutation frequency 22 , and that PGN_0274 and PGN_1740 are involved in biofilm formation 23 .
In this study, we examined whether PorX and PorY function as a cognate pair of the TCS. We also investigated the relationship of the ECF sigma factor PGN_0274, which we designated SigP (sigma factor for expression of por genes), with T9SS in association with the TCS.

Results
Subcellular localization of PorX and PorY in P. gingivalis. The predicted amino acid sequences and expected functional domains of PorX and PorY are shown in Fig. 1A. PorX has a putative RR receiver domain at the N-terminus, which is predicted to receive a phosphoryl group from its cognate HK partner in the TCS. At the C-terminal region of this RR domain is a PglZ domain, which is a member of the alkaline phosphatase clan. PorY is predicted to have two transmembrane domains following to a typical signal peptide at the N-terminus. Two HK-associated sequences are located in the C-terminal half of PorY: a phosphoacceptor domain (HK1) and an ATPase domain (HK2).
Localization of these proteins was determined by subcellular fractionation and Western blot analysis (Fig. 1B). A 55-kDa protein band immunoreactive to the anti-PorX antibody was detected in the cytoplasmic/periplasmic fraction of the wild-type protein, and was absent from the whole cell lysate of a porX mutant. Additionally, a 37-kDa protein reactive to the anti-PorY antibody was mainly detected in the inner membrane fraction and partially in the outer membrane fraction, and was absent from the whole cell lysate of a porY mutant. Thus, PorX appears to localize in the cytoplasm/periplasm and PorY appears to span the inner membrane, which are the typical locations of RRs and HKs. The computational prediction with TMHMM (http://www.cbs.dtu.dk/services/TMHMM-2.0/) and DAS (http://www.sbc.su.se/~miklos/DAS/) also suggest that PorY includes two helical lipid-interfacing domains, which is preferentially observed in inner membrane spanning domains.

Direct interaction between PorX and PorY. To investigate the direct interaction between PorX and
PorY, the binding activity of recombinant PorX (rPorX) and recombinant PorY (rPorY) was determined by a real-time surface plasmon resonance (SPR) assay using BIAcore apparatus ( Fig. 2A). Full-length rPorX was dose-dependently and stably bound to rPorY immobilized on CM5 sensor chips under physiological conditions [10 mM Hepes (pH 7.5) and 0.15 M NaCl]. The dissociation constant (K D ) was calculated as 1.409 × 10 −6 M with 1.573 × 10 −4 (Ms) −1 of ka and 2.215 × 10 −2 s −1 of kd. In contrast, binding of bovine serum albumin (BSA) to rPorY was not detected under the same conditions.
To examine the interaction specificity between rPorX and rPorY, we determined whether rPorX could bind to another HK of P. gingivalis, FimS (Supplemental Fig. 1). FimS/FimR is one of the TCSs in P. gingivalis which regulates the expression of a fimbrial component, FimA 24,25 . rPorX showed no significant interaction with recombinant FimS (rFimS), although recombinant FimR (rFimR) was bound to rFimS in a dose-dependent manner. The results clearly indicate that rPorX binds specifically to rPorY.  Autophosphorylation of PorY and its in vitro phosphorelay to PorX. Signal transduction through TCS typically involves the autophosphorylation of HK at a histidine residue followed by transfer of the HK histidine residue phosphoryl group to an aspartate residue of RR. The resulting phosphorylated RR further modulates the transcription of effector genes 26 . Because the SPR assay demonstrated a direct interaction between PorX and PorY, we next examined the autophosphorylation of PorY and the transfer of a phosphoryl group to PorX using [γ -32 P]ATP. As shown in Fig. 2B, rPorY with a molecular mass of 26 kDa was radiolabelled within 1 min, indicating rapid autophosphorylation. Additional incubation of phosphorylated rPorY with rPorX resulted in the transfer of the radiolabelled signal to a 55-kDa protein band corresponding to rPorX. The radioactivity of the 55-kDa rPorX band was increased in a time-dependent manner as revealed by the densitometrical analysis. These results strongly indicate that PorY and PorX act as an HK and RR of a cognate TCS, respectively, and regulate the secretion of virulence factors through the T9SS in P. gingivalis.

Mn 2+ requirement in the activation of PorXY TCS. A divalent cation around the phosphorylation site
is considered necessary to add or remove phosphoryl groups in RRs. Mg 2+ is preferred for many RRs, but other cations such as Mn 2+ can be required for the reaction in some RRs. Therefore, we examined the effects of Mg 2+ and Mn 2+ on the phosphorylation of rPorY and rPorX. rPorY was autophosphorylated in the presence of either Mg 2+ or Mn 2+ , whereas rPorY-mediated phosphorylation of rPorX required Mn 2+ , but not Mg 2+ (Fig. 2C).
In vivo phosphorylation of PorX by PorY. Phosphorylation of a protein leads to a change in its isoelectric point (pI), revealed by a pI shift in two-dimensional (2D) Western blotting. As shown in Supplemental Fig. 2a, three protein spots reacting with anti-PorX were detected in a 2D blot of P. gingivalis ATCC 33277 (wild-type) whole cell lysate. In the porY mutant, the spot with the higher pI value (solid arrowhead) was intensified compared with the wild type, whereas the intensities of the middle and lower pI spots (arrow and open arrowhead) were decreased. To confirm that the pI shift was caused by a phosphorylation state, whole cell lysate from the porY mutant was treated with lambda protein phosphatase (PP), or the direct phosphor donor acetyl phosphate   (AcP), and then subjected to 2D-Western blotting (Supplemental Fig. 2b). The spot with the higher pI value was increased by PP treatment (solid arrowheads), while the spots with lower pI values and a higher molecular mass were increased by AcP treatment (open arrowheads). These results were not inconsistent with the in vitro result that PorX was phosphorylated by PorY.

Involvement of SigP in the transcription of T9SS components.
The 193-amino acid protein PGN_0274 in strain ATCC 33277 (PG0162 in strain W83) encoded by sigP is predicted to be an ECF sigma factor with sigma-70 subdomains, the highly conserved region 2, and a helix-turn-helix r4 (Fig. 3A). Recently, mutants deficient in sigP have been reported to show decreased gingipain activity 19 and increased biofilm formation 23 . To confirm the relationship between sigP and colonial pigmentation associated with gingipain activity, we constructed sigP deletion mutants (KDP391 and KDP392) from strains ATCC 33277 and W83, respectively, and a complement strain, KDP393. The sigP mutants exhibited non-pigmented colonies, while the complement strain (KDP393) and wild type strains (ATCC 33277 and W83) showed vigorous black pigmentation (Fig. 3).
To determine which genes are influenced by sigP, microarray analysis of a sigP mutant (KDP314) 23 using a custom tiling DNA array chip was performed. The 40 genes most down-regulated by the disruption of sigP are compared with those from the porX deletion mutant in Table 1. The expression of porP, sov, porR, porK, porT, porM, and porL, which are involved in the T9SS, decreased to less than 60% in the sigP mutant, which was the same as in the porX mutant. The decreased transcription of these genes was also confirmed by quantitative real time RT-PCR (Supplemental Fig. 3). Furthermore, 26 of the 40 down-regulated genes were common to sigP and porX mutants. A gene expression profile from the custom tiling DNA array also revealed the similarity in gene expression between sigP and porX mutants (correlation coefficient, r = 0.48) (Fig. 4A). These results indicate that the expression of a set of genes is regulated by both PorX and SigP.
Binding of rSigP to the promoters of T9SS component genes. We next examined the ability of rSigP to bind to the promoter regions of genes listed in Table 1. Previous tiling array analysis revealed that ruvA-sov and porP-porK-porL-porM-porN regions produce polycistronic transcripts at least in part 11 , and that their expression was significantly downregulated in porX and sigP mutants (Supplemental Fig. 4). Since ruvA and sov seem to have no functional relationship; ruvA encodes a Holliday junction DNA helicase and sov encodes a T9SS component protein, and since Fjoh_3477 and FP2412, the orthologs of porP in Flavobacterium johnsoniae and Flavobacterium psychrophilum respectively, are separately located from porK-porL-porM-porN in their genomes, we determined whether rSigP binds to DNA regions 5′ -adjacent to sov and porK in addition to porT, ruvA, porV and porP. The mobilities of DNA probes corresponding to the putative promoters of porT (encoding PGN_0788), ruvA (encoding PGN_0833), porV (encoding PGN_0023), and porP (encoding PGN_1677) shifted more slowly in the presence of rSigP (Fig. 4B). Addition of an excess amount (100-fold) of unlabelled probes significantly reduced the mobility shifts, confirming the specificity of the binding. On the other hand, rSigP showed no binding to the DNA regions 5′ -adjacent to sov and porK.
DNA sequences of the probes for porV, porT, ruvA and porP in electrophoretic mobility shift assays (EMSA) were analyzed by the computational Motif Discovery scan, which examines DNA sequence motifs representing protein-DNA interaction sites (Supplemental Fig. 5). The consensus promoter motif, 5′ -C(T)AAA(G/C/T)A(T) A(C)T(A/G)A-3′ , was suggested to be a SigP target promoter motif. The predicted sequence was also found in the upstream regions of PGN_1353, PGN_1534, PGN_1639, PGN_1047 and PGN_0460, which were the coding sequences markedly downregulated in the sigP deletion mutant.

Lack of SigP in the porX deletion mutant. To demonstrate the relationship between SigP and PorX,
we examined the presence of SigP in the porX deletion mutant by Western blot analysis using an anti-SigP antibody (Fig. 5A). An approximately 30-kDa protein band corresponding to SigP disappeared in the porX deletion mutant, whereas the band was clearly detected in the wild-type and complemented porX/porX + strains. To determine whether the lack of SigP in the porX mutant was caused at the transcriptional level, quantitative real time RT-PCR was performed (Fig. 5B). sigP mRNA levels in the porX mutant and the complemented porX/porX + strain were shown to be decreased to 41% and 60% of that in wild type, respectively.

Direct interaction between PorX and SigP. To investigate the direct interaction of PorX and SigP,
binding of rSigP to rPorX were determined by SPR analysis with BIAcore (Fig. 6A). rSigP bound to rPorX immobilized on a CM5 sensor chip at the dissociation constant (K D ) of 7.828 × 10 −8 M. To examine the in vivo interaction between PorX and SigP, co-immunoprecipitation on P. gingivalis lysates was performed. Using anti-PorX antibody-conjugated protein G sepharose beads, SigP protein was co-precipitated with PorX in a sigP/sigP-Myc + complement strain as revealed by immunoblot detection with anti-SigP antibody, whereas SigP protein band was not detected in the precipitated fraction from a porX mutant (Fig. 6B middle panel). The co-immunoprecipitation of SigP was confirmed by use of anti-c-Myc antibody (Fig. 6B right panel). These results indicated that PorX and SigP physically interacted with each other in P. gingivalis cells.

Discussion
We previously demonstrated that the T9SS was involved in the secretion of virulence factors responsible for colony pigmentation, haemagglutination, adherence, the modification of bacterial surface proteins, and the degradation and/or activation of host proteins. T9SS substrate proteins are translocated across the inner membrane via Sec machinery, and then secreted through the outer membrane by the T9SS in association with their C-terminal domains 27 . Thus, the T9SS machinery seems to be crucial for the determination of bacterial pathogenicity. In the present study, we identified a regulatory cascade system for the T9SS of P. gingivalis, namely, the TCS composed of PorX and PorY and its downstream ECF sigma factor, SigP.
A divalent cation is considered necessary to add or remove phosphoryl groups in RRs. PhoPQ TCS is known to be important for virulence regulation in Salmonella typhimurium, in which PhoP acts as an RR that is activated under low Mg 2+ concentrations 31 . The chemotaxis protein Y (CheY) from Escherichia coli is an RR that regulates chemotactic flagellar rotation, which also requires the presence of Mg 2+ 32 . Cyanobacterial phytochrome Cph1, a light regulated HK, and Chlamydia trachomatis TCS, CtcB-CtcC, were previously reported to be highly activated in the presence of Mn 2+ 33,34 . In this study, we found that Mg 2+ or Mn 2+ was required for autophosphorylation of PorY and that transphosphorylation from PorY to PorX required Mn 2+ , but not Mg 2+ (Fig. 2C). Requirement of the divalent cation for the reactions suggests that phosphorylation of the PorXY TCS is similar to that of other known TCSs.
Besides the PorXY TCS, four TCS have been characterized in P. gingivalis to date: (1) the FimRS system comprising FimR (PGN_0903) and FimS (PGN_0904), which are involved in fimbriation 24 ; (2) the predicted hybrid HK with an RR domain, GppX (PGN_1768), which is associated with black pigmentation and the secretion of gingipains 35 ; (3) the stress response RR, RprY (PGN_1186), whose cognate kinase has not yet been identified 36 ; and (4) the PGN_0752-PGN_0753 system, which participates in bacterial haemin acquisition 37 . Wild-type gingipains were previously shown to mainly associate with cells, although some are released into the culture supernatant; this differs from the gingipains of gppX-deficient mutants, which are typically found in the supernatant 35 . Therefore, GppX is thought to be involved in the regulation of biosynthesis of an anionic lipopolysaccharide, A-LPS, that anchors gingipains to the cell surface.
RRs are typically composed of two domains, a phosphor-receiver domain and an effector domain. Functions of the latter domain can be classified as DNA binding (63%), enzymatic activity (13%), protein binding (3%), RNA binding (1%), and no functional domain (17%) 38 . PorX lacks a typical DNA binding motif such as a Scientific RepoRts | 6:23288 | DOI: 10.1038/srep23288 helix-turn-helix, but contains a PglZ domain at its C terminus (Fig. 1). In fact, EMSA revealed that rPorX did not bind the upstream regions of genes encoding T9SS components which are downregulated in the porX-deficient mutant (data not shown).
A similar set of genes including T9SS components were found to be controlled by both PorX and SigP in the present study (Fig. 4, Table 1), while the sigP mutant phenotype was similar to those of porX and porY mutants 19 . EMSA revealed that rSigP could bind to predicted promoter regions of porT (PGN_0788), ruvA (PGN_0833), porV (PGN_0023), and porP (PGN_1677), but not to the upstream regions of sov (PGN_0832) and porK (PGN_1676). These results indicate that genes encoding T9SS components are directly regulated by the SigP sigma factor.
We found that SigP protein levels were barely detectable in the porX mutant, and that sigP mRNA expression was 40% that of wild type. In contrast, the amount of SigP in the sigP/sigP + complemented strain was similar to wild type, although sigP mRNA levels in the complemented strain were 60% that of wild type (Fig. 5). This suggests that SigP sigma factor is unstable in the absence of PorX.
The direct interaction between PorX and SigP in vitro and in vivo were demonstrated by SPR analysis and co-immunoprecipitaion, respectively, in this study (Fig. 6). Interaction between RRs and sigma factors was previously reported in E. coli RssB adaptor, a member of the RR family and an RpoS sigma factor 39 . In this case, a complex of RpoS and RssB is rapidly degraded by the ATP-dependent protease ClpXP. Under stress conditions, the small anti-adaptor proteins IraP, IraM, and IraD, which form under different stress conditions, bind to the RssB adaptor and block RpoS degradation, resulting in the activation of RpoS-dependent genes. The RssB C terminus does not contain a DNA-binding domain and has similarities to PP2C Ser/Thr phosphatases 40,41 . Although the interaction between PorX and SigP in P. gingivalis is similar to that between RssB and RpoS in E. coli, the consequence of the PorX-SigP interaction appears to differ from that of RssB-RpoS because the SigP sigma factor is present and active in the presence of PorX and lacking in its absence.
The regulatory system of T9SS in P. gingivalis could be a novel therapeutic target for periodontal diseases caused by this bacterium because the machinery contributes to the secretion of various virulence factors, including gingipains. Functional T9SSs have been experimentally demonstrated in Cytophaga hutchinsonii 42 , Flavobacterium johnsoniae 43 , Tannerella forsythia 44,45 , and Capnocytophaga ochracea 46 , in addition to P. gingivalis. Moreover, genome analysis has revealed that many bacteria belonging to the phylum Bacteroidetes possess T9SS component and substrate genes 47 . An understanding of T9SS expression will help in the development of disease control of infections associated with T9SS-possessing bacteria.

Methods
Bacterial strains and culture conditions. Bacterial strains and plasmids used in this study are listed in Supplemental Table 1. P. gingivalis cells were grown anaerobically (80% N 2 , 10% CO 2 , 10% H 2 ) in enriched brain heart infusion broth (Becton Dickinson, Franklin Lakes, NJ) and on enriched tryptic soy agar (Nissui, Tokyo, Japan) supplemented with 5 μg/ml haemin (Sigma, St. Louis, MO), and 0.5 μg/ml menadione (Sigma). For blood agar plates, defibrinated laked sheep blood was added to enriched tryptic soy agar at 5% 8 . Luria-Bertani (LB) broth (Sigma) and LB agar plates were used to grow E. coli strains. Antibiotics were added to the medium for the selection and maintenance of antibiotic-resistant strains at the following concentrations: ampicillin (100 μg/ml for E. coli), erythromycin (10 μg/ml for P. gingivalis).

Construction of P. gingivalis deletion mutants. P. gingivalis deletion mutants (KDP391 and KDP392
from ATCC 33277 and W83, respectively) were generated by double recombination of the targeted genes and the introduction of erythromycin resistance genes, as previously described 11 . Primers used in this study are listed in Supplemental Table 2, and experimental details are described in Supplemental Text.
Construction of complemented sigP-deficient mutant strain. The entire sigP containing its promoter was PCR-amplified from chromosomal DNA using primers N0274-U-F-K and N0274-D-R-X, digested with KpnI and XhoI, and then inserted into the KpnI-NotI site of the pTCB plasmid carrying Myc-and His 6 -tags, and the transcriptional terminator DNA region (pKD911). After mating of E. coli S17-1 containing pKD911 and the P. gingivalis sigP deletion mutant, an Em r Tc r transconjugant (KDP393) was obtained. Subcellular fractionation. Subcellular fractionation of P. gingivalis cells was performed as previously described 48 . Further details are given in Supplemental Text.
The pET28a(+ ) plasmids containing the N-terminal His 6 -tagged porX or the 3′ -region of porY, and the pET32b(+ ) plasmid containing N-terminal thioredoxin-and C-terminal His 6  Preparation of antisera. Antisera against PorX, PorY, and SigP protein were raised in rabbits using recombinant proteins. Full-length PorX and SigP, and the C-terminal region of PorY lacking transmembrane domains were expressed with tags of thioredoxin and His 6 in the pET32b system, followed by purification with Ni-NTA agarose. Rabbit antibodies against PorX or SigP were obtained by binding to Protein A sepharose. Monoclonal anti-c-Myc antibody was purchased from Sigma.
Immunoprecipitation. Antibody against PorX was chemically crosslinked to Protein G sepharose with 50 mM dimethyl pimelimidate (Sigma). P. gingivalis porX-deficient and sigP/sigP-Myc + complement strains were lysed in BugBuster Protein Extraction Reagent (Merk Novagen) and insoluble debris was removed by centrifugation at 27,000 × g for 30 min. The supernatants were incubated with anti-PorX antibody-conjugated protein G sepharose and washed 5 times with 50 mM Tris-HCl buffer containing 0.15 M NaCl and 0.05% NP-40, pH7.5. Associated proteins were eluted by boiling in a sample buffer.
Gel electrophoresis and Western blot analysis. Western blots were performed to detect PorX, PorY and SigP in extracts of wild-type and mutant cells of P. gingivalis. A total of 10 μg protein, as determined by a BCA assay (Thermo Fisher), was loaded per lane and proteins were separated on 12% polyacrylamide gels and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred onto polyvinylidene difluoride (PVDF) membranes, then immunodetected with anti-PorX or anti-PorY antiserum, and anti-SigP or anti-c-Myc antibodies obtained in this study.
Binding affinity analysis with BIAcore surface plasmon resonance. The binding affinity of rPorX to rPorY and rSigP to rPorX was investigated by SPR using a BIAcore X100 biosensor (GE Healthcare, Little Chalfont, UK). For the experiment, rPorY or rPorX proteins were immobilized onto a carboxylated dextran matrix sensor chip (CM5) resulting in an average bound density of 3000 response units. The series diluents of the recombinant rPorX were used as analytes in a running buffer [10 mM Hepes (pH 7.5) and 0.15 M NaCl] at a 30-μl/min flow rate. The association was recorded in the form of sensorgrams which were used to calculate association and dissociation constants. Following each cycle, the flowcell surfaces were regenerated with 30 μl of 0.1 M Tris-HCl (pH 8.0) containing 1 M NaCl at a 10-μl/min flow rate and re-equilibrated in the running buffer.
BiaEvalution software (version X100 from GE Healthcare) was used to determine the simultaneous dissociation (kd) and association (ka) rates by non-linear fitting assuming a 1:1 complex model. The equilibrium dissociation constants (K D ) were then calculated as kd/ka. Binding data are summarized in Figs 2A and 6A from the average of three complete sets of data.
In vitro phosphorylation and phosphorelay assays. For the autophosphorylation assay, rPorY was incubated in a kinase buffer [25 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM MnCl 2 , 2 mM DTT, 5 mM β -glycerophosphate, 0.1 mM Na 3 Vo 4 and 0.5 μCi of [γ -32 P]ATP ] for 1, 5, 10, or 20 min at 0 °C. Phosphotransfer assays were conducted by allowing rPorY to autophosphorylation for 5 min, followed by the addition of purified rPorX with excess cold ATP (1mM). Reactions were stopped at indicated times. Samples were mixed with SDS loading buffer and subjected to SDS-PAGE for separation in 12% polyacrylamide gels. After electrophoresis, gels were dried and exposed to Imaging Plates (GE Healthcare).
Tiling microarray analysis. Custom tiling microarrays spanning the whole genome of P. gingivalis ATCC33277 with 25-mer probes (each of which was eight-bases shifted on the genome sequence) were purchased from Affymetrix (Santa Clara, CA). The analytical procedure was described previously 11 . Details are given in Supplemental Text.
Electrophoretic mobility shift assays (EMSA). EMSAs were performed using the DIG Gel Shift Kit (Roche) according to the manufacturer's instructions. Probes used in this assay were designed from the possible transcription starting sites in the tiling microarray, and PCR-amplified from P. gingivalis ATCC 33277 DNA using primers shown in Supplemental Table 2. Labelled probes (4 ng) were incubated with rSigP (1 μg) in 20 μl binding buffer for 30 min at room temperature. The reactions were run on 6% Novex DNA gel Retardation gels in 0.5 × TBE buffer (44.5 mM Tris, 44.5 mM boric acid, 1 mM EDTA, pH 8.0) at 4 °C. DNA-protein complexes were electrically transferred to Hybond-N + membranes (GE Healthcare), and DIG-labelled reactions detected with chemiluminescence using an alkaline phosphatase-conjugated anti-DIG antibody (Roche) and CDP-star detection reagent (Roche).
Quantitative real-time RT-PCR. Total RNA was transcribed into cDNA with SuperScript III first-strand synthesis system (Thermo Fisher). Quantitative real-time RT-PCR was performed with the Brilliant II Fast SYBR green QPCR master mix (Agilent Technologies, Santa Clara, CA) and Mx3005P real-time PCR system (Agilent Technologies) according to the manufacturer's instructions. Primers for real-time analysis (Supplemental Table  2) were designed using Primer3 software (http://primer3.sourceforge.net/). Quantitative real-time PCR conditions were as follows: 1 cycle at 95 °C for 2 min, then 35 cycles of 95 °C for 5 s and 60 °C for 20 s. At each cycle, the accumulation of PCR products was detected by the reporter dye from the double-stranded DNA binding SYBR green. To confirm that a single PCR product was amplified, a dissociation curve was constructed in the range of 55-95 °C after the PCR. All data were analysed using Mx3005P software. The expression level of each targeted gene was normalized to that of 16S rRNA. All PCRs were carried out in triplicate. The efficiency of primer binding was determined by linear regression by plotting the cycle threshold (C T ) value versus the log of the cDNA dilution. Relative quantification of the transcript was determined using the comparative C T method (2 −ΔΔCT ) calibrated to 16S rRNA. Experiments were independently performed three times with comparable results.