|Home | About | Journals | Submit | Contact Us | Français|
Although the Gram-negative, anaerobic periodontopathogen Porphyromonas gingivalis must withstand nitrosative stress, which is particularly high in the oral cavity, the mechanisms allowing for protection against such stress are not known in this organism. In this study, microarray analysis of P. gingivalis transcriptional response to nitrite and nitric oxide showed drastic upregulation of the PG0893 gene coding for hybrid cluster protein (Hcp), which is a putative hydroxylamine reductase. Although regulation of hcp has been shown to be OxyR dependent in Escherichia coli, here we show that in P. gingivalis its expression is dependent on the Fnr-like regulator designated HcpR. Growth of the isogenic mutant V2807, containing an ermF-ermAM insertion within the hcpR (PG1053) gene, was significantly reduced in the presence of nitrite (P < 0.002) and nitric oxide-generating nitrosoglutathione (GSNO) (P < 0.001), compared to that of the wild-type W83 strain. Furthermore, the upregulation of PG0893 (hcp) was abrogated in V2807 exposed to nitrosative stress. In addition, recombinant HcpR bound DNA containing the hcp promoter sequence, and the binding was hemin dependent. Finally, V2807 was not able to survive with host cells, demonstrating that HcpR plays an important role in P. gingivalis virulence. This work gives insight into the molecular mechanisms of protection against nitrosative stress in P. gingivalis and shows that the regulatory mechanisms differ from those in E. coli.
The oral cavity has nitrite concentrations ranging from 10 μM to greater than 1 mM (34). Such high nitrite concentrations are due to high dietary nitrate intake, which is then reduced by oral nitrate-reducing bacteria (25, 59, 60). Furthermore, following intake of foods rich in sucrose, the pH in the oral cavity drops drastically due to metabolic conversion of the sugar to acid by oral microbiota, such as streptococci and lactobacilli, thus creating conditions that are favorable for the chemical generation of nitric oxide from nitrite (44). Nitric oxide may also be generated from nitrite by oral microbiota during metabolism through the respiratory nitrite reductase system, NrfHA (51). Another source of nitrite and nitric oxide in the oral cavity, as well as in other sites of human hosts, is eukaryotic cells, which respond to microbial infection by producing nitric oxide and O2-related species (16). This response is part of the innate immune response and is mediated by inducible nitric oxide synthases (iNOS) by both immune and nonimmune cells (35, 40).
Despite the high concentrations of nitrite/nitric oxide in the oral cavity, the mechanisms used by oral bacteria to tolerate nitrosative stress are poorly understood. One of the major pathogens in the oral cavity is Porphyromonas gingivalis (24). This is a Gram-negative anaerobic bacterium that plays a role in the development and progression of chronic adult periodontitis (15, 31). It is also found in other parts of the body, such as the cardiovascular system and umbilical cord (5, 13, 14). Despite studies regarding the virulence aspects present in this organism, mechanisms of nitrite/nitric oxide detoxification in P. gingivalis are poorly understood.
Examination of the genomic sequence of this bacterium has allowed us to predict possible mechanisms involved in nitrite/nitric oxide detoxification processes (42). P. gingivalis is an anaerobic organism, and like other anaerobic bacteria, it codes for putative genes playing a role in respiration using alternative electron acceptors such as in nitrite ammonification. Thus, the cytochrome c nitrite reductase system, NrfAH, which converts nitrite into ammonia, has been identified on the genome of this bacterium (PG1820 and PG1821 [PG1820-1]) (Oralgen) (37). Also, a small protein with putative nitrite reductase activity is encoded by PG2213 (Oralgen). These mechanisms may have a role not only in metabolism but also in detoxification of nitrosative stress (37), but so far those for P. gingivalis remain to be established.
A redox enzyme known as hybrid cluster protein, Hcp, is present in anaerobic bacteria (Desulfovibrio spp.) and facultative anaerobic bacteria (Escherichia coli, Salmonella enterica serovar Typhimurium, Acidothiobacillus ferrooxidans, Rhodobacter capsulatus, and Shewanella oneidensis) (reviewed in reference 46). Hcp is induced by nitrite, indicating that it has a role in nitrogen metabolism. The proposed role of Hcp is to reduce hydroxylamine generated from nitrite or nitric oxide into ammonia and water (8, 46, 58). As such, it plays a significant role in toxic nitrogen compound detoxification. Indeed, recent studies have shown that Hcp plays a role in the protection of a variety of bacteria from nitrosative stress as well as from nitrosative stress-based killing by macrophages (48). Hcp is also encoded on the genome of P. gingivalis by PG0893 and protects the bacterium from exposure to nitric oxide (7). However, the biological function and the regulatory mechanisms governing expression of this protein remain unknown.
Hcp expression in E. coli under anaerobic conditions in the presence of nitrate is regulated by OxyR (48). However, the response to nitrate/nitrite stress is also regulated by fumarate-nitrate regulator (Fnr)-like proteins both in E. coli and in other bacteria (17, 46, 51). This family of broad-spectrum regulators mediates adaptation to a variety of stimuli, including lack of oxygen (Fnr), toxin products, and exposure to nitric oxide (Dnr) (30, 36, 45). In addition, a novel Fnr-like regulator, designated HcpR, has been identified using in silico analysis and is predicted to regulate the expression of hcp and other genes involved in response to nitrite (9, 46). This putative regulator was proposed to be present in gammaproteobacteria (Desulfovibrio spp., Geobacter spp.), clostridia, the Cytophaga-Flavobacteria-Bacteroides (CFB) group of bacteria (Bacteroidetes), Fusobacterium nucleatum, Treponema denticola, and Thermotogales (e.g., Thermotoga maritima). The presence of the novel regulator HcpR in these anaerobic bacteria suggests that distinct mechanisms may regulate expression of the critical nitrosative stress defense player hcp in those organisms, and the role of OxyR in this process marks just the beginning of our understanding of hcp regulation, which is possibly limited to facultatively anaerobic bacteria. Such an assumption prompts experimental verification of the role of HcpR in the regulation of hcp expression and in the response to nitrite.
In this study, microarray and bioinformatics analyses of P. gingivalis transcriptional response to nitrite and nitric oxide were performed. In addition, an Fnr-like protein involved in the adaptation of P. gingivalis to nitrosative stress was identified. Our data show that this novel protein is indispensable for bacterial growth with nitrite and nitric oxide, and this adaptation is mediated by its ability to activate hcp expression. Thus, we designated the regulator HcpR. This work gives insight into the molecular mechanisms of protection against nitrosative stress in P. gingivalis and marks the beginning of experimental characterization of the HcpR family of regulators.
The bacterial strains used in this study are listed in Table S1 in the supplemental material. Primarily, P. gingivalis strain W83 and its derivatives were used in this study. P. gingivalis strain W83 is a virulent, encapsulated strain (49). The bacteria were grown anaerobically (80% N2, 10% H2, and 10% CO2) at 37°C in an anaerobic chamber (Coy Manufacturing, Ann Arbor, MI). Blood agar plates (TSA II plus 5% sheep blood; BBL, Cockeysville, MD) or brain heart infusion (BHI) broth containing hemin (5 μg/ml; Sigma, St. Louis, MO) were used to maintain the anaerobic bacteria. One Shot Top10 chemically competent Escherichia coli cells (Invitrogen, Life Technologies, Grand Island, NY) and E. coli BL21(DE3 pLys) (Novagen, EMD4Biosciences, Merck KGaA, Darmstadt, Germany) were used for cloning and protein expression, respectively. Kanamycin, erythromycin, and clindamycin were used to select E. coli transformants (50 μg/ml of kanamycin sulfate in Luria-Bertani medium [LB; Gibco, BRL Inc., Gaithersburg, MD] with or without 1.5% agar or 300 μg/ml erythromycin in LB broth or agar) and P. gingivalis mutants (0.5 μg/ml in BHI agar medium).
A single colony from a blood agar plate (BBL, BD, Franklin Lakes, NJ) served to inoculate 3 ml of mycoplasma medium (BD, Franklin Lakes, NJ) broth. The cultures were grown to confluence (approximately 48 h) under anaerobic conditions. The cultures were then diluted 1:10 with fresh mycoplasma broth and incubated for 18 h under anaerobic conditions. These cultures were used for our growth studies.
Cultures were prepared by inoculating mycoplasma broth (kept under anaerobic conditions or equilibrated with 6% oxygen overnight) to an optical density at 660 nm (OD660) of 0.05. Growth studies were conducted under anaerobic and aerobic conditions in the presence of 6% oxygen (microaerophilic atmosphere consisting of 6% O2, 80% N2, 7% CO2, and 7% H2 generated using Anoxomat Mark II [Mart Microbiology B.V., Netherlands]). Aliquots (500 μl) were removed at various time points, and growth was monitored by measuring the OD660. Three independent cultures grown under both conditions were prepared on different days to ensure biological significance of the results.
Cultures were prepared using mycoplasma broth and grown in the presence of various concentrations of nitrosative stress-generating species. The growth studies were also done in mycoplasma broth supplemented with 5 μg/ml hemin.
A fragment of P. gingivalis W83 genomic DNA coding for the entire HcpR protein was PCR amplified using forward (PG1053F) and reverse (PG1053R) primers (see Table S2 in the supplemental material) and cloned into the pCR2.1 vector according to the manufacturer's instructions (Invitrogen, Life Technologies, Grand Island, NY). The ermF-ermAM cassette was inserted into the SspI site of the cloned hcpR gene. The DNA fragment containing the mutagenized hcpR was released by digestion with EcoRI and used to electroporate P. gingivalis W83 according to a previously published protocol (19). Transformants were selected on BHI agar plates supplemented with 0.5 μg/ml of clindamycin. The expected insertion of the ermF-ermAM cassette on the genome of the mutant strains was examined by PCR with PG1053F and PG1053R primers and confirmed by PCR fragment sequencing. The mutant was designated V2807. The OxyR-deficient strain was generated similarly: the ermF-ermAM cassette was inserted into the HincII site present at 556 bp of the oxyR gene. The OxyR-deficient mutant was designated V2798.
The mutation in V2807 was reverted as described previously (27). Briefly, the DNA fragment containing an intact copy of the hcpR gene was electroporated into the HcpR mutant V2807 strain. The mixture was anaerobically grown in the presence of 1.5 ml of BHI supplemented with 5 μg/ml hemin for 5 h. Revertants were then selected by plating on mycoplasma agar containing 500 nM nitrite. As a negative control, V2807 was electroporated without DNA addition. Reversion of the hcpR mutation was verified by the absence of growth on BHI plates supplemented with clindamycin (0.5 μg/ml) and PCR amplification of hcpR followed by sequencing. One of the nitrite-resistant revertants, designated V2835, was chosen for further studies.
RNA was isolated from bacteria grown to an early logarithmic phase (OD660, 0.3 to 0.6) using an RNeasy minikit (Qiagen, Valencia, CA) according to the manufacturer's instructions. Residual DNA was removed using the DNA-free kit (Ambion, Austin, TX) following the manufacturer's instructions.
P. gingivalis cultures were grown to the midlogarithmic phase of growth in mycoplasma broth and then supplemented with 0.2 mM NaNO2 or 20 nM GSNO, and the cultures were grown for an additional 60 min. Cells were harvested, and total RNA was isolated as described above. Total RNA (10 μg) was reverse transcribed with ArrayScript reverse transcriptase (Applied Biosystems/Ambion, Austin, TX). The cDNA was labeled with dyes (Cy3 and Cy5; GE Healthcare, Piscataway, NJ) using the Array 900 MPX kit from Genisphere (Genisphere, Hatfield, PA). Differentially labeled cDNAs were hybridized to glass genomic microarrays (obtained from the J. Craig Venter Institute [JCVI]; formerly The Institute for Genomic Research [TIGR]) containing 70-mer oligonucleotide probes for all predicted open reading frames (ORFs) present on the P. gingivalis W83 genome. Microarray hybridization and washing were done as described by Genisphere (Genisphere, Hatfield, PA). Dithiothreitol (DTT) (0.1 mM) was added to washes following the final hybridization, and the slides were protected from the impact of ozone by the application of DyeSaver (Genisphere, Hatfield, PA). The microarrays were scanned with an Axon 4200A microarray scanner (Molecular Devices, Downingtown, PA) to detect the hybridized cDNA. The obtained images were inspected for quality and quantified using GenePix v6.0 software (Molecular Devices, Downingtown, PA), and the .gpr files were analyzed for significant differences using the Significance Analysis for Oral Pathogen Microarrays (SAOPMD) tools available at the Bioinformatics Resource for Oral Pathogens (BROP) website provided by The Forsyth Institute (http://www.brop.org). Thus, results for four technical replicates for nitrosative stress exposure tested were averaged. Ratios of the average values of samples exposed to nitrosative stress versus controls were generated. Genes with fold change of >1.5 at P < 0.05 were considered to be significantly regulated in our study.
Real-time quantitative reverse transcriptase PCR (qRT-PCR) was done with the SYBR green-based detection system on an Applied Biosystems 7500 fast real-time PCR system (Applied Biosystems, Austin, TX). Primers were designed using Primer3 software (http://frodo.wi.mit.edu). The primers used in this study are listed in Table S2 in the supplemental material. The cDNA was generated with ArrayScript (Ambion, Austin, TX) using 10 ng of total RNA and random hexamers (for 16S rRNA, 0.02 ng of total RNA was used). qPCR was done using the SYBR green qPCR mix (Applied Biosystems, Austin, TX). Experimental samples were tested in triplicate using gene-specific primers, and samples in which reverse transcriptase was omitted served as negative controls. Quantitative PCR was done using the standard curve mode protocol (a calibration curve was constructed using serial 5-fold dilutions of cDNA obtained using 50 ng of total RNA) and a thermal profile consisting of one cycle of 10 min at 95°C and 40 cycles of 15 s at 95°C and 1 min at 60°C. The amounts of RNA used for the analysis were normalized using a probe specific for 16S rRNA.
Human umbilical vein endothelial cells (HUVECs) (Lifeline Cell Technology, Frederick, MD) and human oral keratinocytes (HOKs) (ScienCell, Carlsbad, CA) were grown to confluence in 6-well tissue culture plates in the presence of gamma interferon (2 ng/ml) (Peprotech). The cells were infected with the parental W83 and hcpR mutant V2807 P. gingivalis strains at a multiplicity of infection (MOI) of 100. Infection was conducted for 30 min under anaerobic conditions at 37°C. For association of bacteria with HUVECs, the infection medium was removed and the cells were washed twice with phosphate-buffered saline (PBS) and lysed to release intracellular bacteria. For microbial invasion studies, washed cells were incubated for 60 min with a cell-specific medium (e.g., endothelial growth medium [EGM] for HUVECs) supplemented with 300 μg/ml gentamicin and 400 μg/ml metronidazole to kill extracellular bacteria. Next, the cells were washed with PBS and lysed by the addition of 1 ml of BHI supplemented with 1% saponin. Serial dilutions of lysed eukaryotic cells prepared in anaerobic BHI were plated to count colony-forming units (CFU), which are representative of a single surviving cell. Two biological and two technical repeats were used to examine survival of bacterial cells in each cell type.
The coding sequence of the hcpR gene was obtained from http://www.oralgen.lanl.gov. The gene was PCR amplified from the chromosomal DNA of P. gingivalis W83 with primers FnF (with an NcoI site) and FnR (with an XhoI site) (see Table S2 in the supplemental material). The amplified DNA was cloned into the pET30a vector (Novagen, EMD4Biosciences, Merck KGaA, Darmstadt, Germany), which attached a tail of six histidines to the 3′ end of the gene. The resulting construct, pET30-hcpR, was introduced into E. coli BL21(DE3) cells (Novagen, EMD4Biosciences, Merck KGaA, Darmstadt, Germany). To purify the recombinant protein, an overnight culture of E. coli-pET30-hcpR was diluted 1:100 in LB broth supplemented with 50 μg/ml of kanamycin, and the culture was grown until midlog phase (OD660 = 0.6). For generation of cell lysate, the cells were suspended in phosphate-buffered saline (PBS), disrupted by sonication, and centrifuged to remove the cellular debris. The supernatant was aspirated, and protein concentration was determined with the Bradford assay (Bio-Rad). For protein purification the cells were suspended in a buffer consisting of 50 mM NaH2PO4, pH 8.0, 300 mM NaCl, and 20 mM imidazole. Recombinant HcpR (rHcpR) was purified under native conditions using Ni-agarose (Qiagen) as described in the manufacturer's protocol. The purity of rHcpR was assessed using SDS-PAGE, and the protein concentration was determined with a Bradford assay (Bio-Rad).
A 200-bp DNA fragment containing the promoter region of hcp was amplified by PCR using primers IGSF and IGSR (see Table S2 in the supplemental material). The fragment was labeled with biotin using the biotin 3′ end DNA labeling kit (Thermo Scientific, Rockford, IL), and then the labeled DNA was used in an electrophoretic mobility gel shift assay (EMSA) using the LightShift chemiluminescent EMSA kit (Thermo Scientific, Rockford, IL). Briefly, a 20-μl reaction mixture consisting of cell lysate proteins or rHcpR, 100 mM NaCl, 10% glycerol, and 2 μg of dI-dC was incubated for 10 min at ambient temperature. The samples were subjected to electrophoresis using a 2% agarose gel in Tris-HCl buffer containing 2 mM EDTA and 5% glycerol. Following electrophoresis, DNA was transferred to nylon membranes (GE Healthcare, Piscataway, NJ), and the membrane was developed with the LightShift chemiluminescent EMSA kit (Thermo Scientific, Rockford, IL).
The P. gingivalis W83 transcriptome was determined as described previously (61). Reads were aligned to the reference genome using the CLC Genomics Workbench (CLC Bio). The transcriptional start site was further verified using qRT-PCR analysis with the primers shown (see Fig. 6B and Table S2 in the supplemental material).
The Basic Local Alignment Search Tool (BLAST) was used to find protein sequences similar to HcpR, and domain organization was examined using the National Center for Biotechnology Information (NCBI) search tools (2) (http://www.ncbi.nlm.nih.gov). An HcpR three-dimensional (3-D) structural model was generated using the SWISS-MODEL homology modeling server (4, 6, 47). The probable structure was visualized and aligned with other similar structures using the PyMOL molecular graphics system, version 1.r1 (http://www.pymol.org). Search for HcpR binding sites was done using PRODORIC (38).
Microarray data were deposited in the GEO database; the accession number is GSE38220.
The oral cavity contains high nitrite concentrations. To determine the susceptibility of P. gingivalis to nitrosative stress, we compared its ability to grow in various concentrations of nitrate (NaNO3), nitrite (NaNO2), and S-nitrosogluthathione (GSNO). As shown in Fig. 1A, 40 mM nitrate reduced bacterial growth. However, P. gingivalis was still able to grow with such a high nitrate concentration. No growth was observed in the presence of 8 mM or 4 mM nitrite, although P. gingivalis did grow with 1 mM nitrite (Fig. 1A). In addition, no growth inhibition was observed in the presence of 0.2 mM nitrite. These results show that P. gingivalis is highly tolerant to nitrate and nitrite stress. Next, we examined the ability of P. gingivalis to grow with nitric oxide stress. A nitric oxide donor, GSNO, inhibited P. gingivalis growth when 0.7 × 10−3 mM was used, and reduced bacterial growth was observed even in the presence of low concentrations (0.07 × 10−3 mM) of GSNO (Fig. 1B), thus indicating that the pathogen is highly sensitive to nitric oxide stress.
To gain insight into the adaptive response of P. gingivalis to nitrosative stress, we examined its transcriptional response to nitrite or GSNO exposure. For exposure to nitrite, we used 0.2 mM NaNO2, which was shown in our studies to have no effect on bacterial growth (Fig. 1A). Using a 1.5-fold change in expression level we found that 36 genes were upregulated and 105 genes were downregulated (Table 1; see Table S3 in the supplemental material). The most drastically upregulated gene was hcp (PG0893), which codes for hydroxylamine reductase (Table 1). Other regulated genes included PG0616, coding for thioredoxin, PG1227, encoding LysR-type regulator, PG0776 and PG0900, coding for oxidoreductases, PG1318, encoding ECF sigma protein, PG1129, encoding ribonucleotide reductase, and PG1321, coding for a formate-tetrahydrofolate ligase. Several genes coding for hypothetical putative proteins were also upregulated. Downregulated genes included ones coding for putative nitrite reductase (PG2213), fumarate reductase (PG1614), and alkyl hydroperoxide reductase (PG0618-9) (see Table S3 in the supplemental material).
Exposure to GSNO resulted in a much more altered transcriptional response; 538 genes were upregulated 1.5-fold, and 628 genes were downregulated 1.5-fold (Table 1; see Table S4 in the supplemental material). This response could be predicted based on the reduced growth of P. gingivalis in the presence of GSNO. As observed upon exposure to nitrite, the most drastically upregulated gene was hcp (PG0893) encoding putative hydroxylamine reductase (Tables 1 and and2).2). However, there was no significant overlap between other nitrite and GSNO-regulated genes. Upregulated by GSNO were ones coding for reactive oxygen detoxification proteins, such as PG0195, coding for rubrerythrin, PG0275 and PG0616, encoding putative thioredoxins, and PG0090, coding for Dps (Table 2; see Table S4 in the supplemental material) (54, 55). The upregulation of rbr expression is consistent with the protective role of rubrerythrin against reactive nitrogen species (39).
Several genes coding for regulators, such as PG1203, PG1205, PG0173, PG1431-2, PG0162, PG0121, PG0465, and PG1240, were upregulated. Also, a large portion of the upregulated genes included ones coding for hypothetical proteins, indicating that novel aspects may be detected when considering nitric oxide protection mechanisms in P. gingivalis. Interestingly, upregulated genes included PG1551, coding for HmuY protein, which has been shown to mediate hemin uptake in P. gingivalis (33, 43, 50), and PG2213, encoding putative nitrite reductase (Oralgen). In addition, the gene coding for fumarate reductase (PG1614-15) was downregulated (see Table S4 in the supplemental material). This result is similar to what was observed after challenge with nitrite. Furthermore, several rRNA-encoding genes and an alkyl hydroperoxide reductase-encoding gene (PG0619) were downregulated (see Table S4).
The microarray results were verified and extended by using the qRT-PCR results (Table 3). Significant upregulation (over 150-fold) of PG0893 (hcp) encoding prismane (hydroxylamine reductase) was observed when this assay was used. This was by far the most upregulated gene after exposure to both nitrite and GSNO. In addition, we tested gene expression in P. gingivalis grown in mycoplasma broth as well as in BHI supplemented with hemin, and under both conditions we noted significant upregulation of PG0893 (hcp) (Table 3).
The 684-bp PG1053 (here designated the hcpR gene) is the last gene located in the six-gene genomic locus coding for cell envelope biogenesis proteins (Fig. 2A). The first gene of the locus, amiA, codes for a peptidoglycan biosynthesis protein (N-acetylmuramoyl-l-alanine amidase); two genes, PG1049 and PG1050, encode proteins with unknown function; and PG1051 codes for O-antigen polymerase. The last two genes in the locus code for regulatory proteins. PG1052, a 354-bp gene, codes for a 118-amino-acid protein with similarity to MerR-like regulators (Fig. 2A). Residues 11 to 106 of the protein encoded by PG1052 are 33% similar to the negative regulator of heat shock response (HspR) protein, which is a regulator from Mycobacterium tuberculosis (H37RV) (12, 53). PG1053 codes for an Fnr-like regulator named here HcpR. Comparison of the P. gingivalis hcpR genomic locus to hcpR-like genomic loci in other bacteria showed that the P. gingivalis locus is distinct in that it does not include the hcp gene (data not shown). Sequence comparisons of the P. gingivalis HcpR to other proteins were done using an NCBI BLAST search. The highest similarity was to a putative transcriptional regulator from P. endodontalis (68% identity), and the similarity dropped drastically for putative regulators from other members of the order Bacteroidales (41% for Bacteroides coprosuis and 36% for both Parabacteroides johnsonii and Prevotella multiformis) (see Table S5 in the supplemental material). Also, similarity was detected with proteins of known function. For example, DnrS, a regulator from Pseudomonas stutzeri, was 24% similar for residues 28 to 225 (57); Fnr-like protein from Lactococcus lactis was 24% similar for residues 29 to 224 (22); a Crp-like transcriptional regulator from Thermatoga maritima (putative HcpR) was 28% similar for residues 51 to 220 (52). Comparison of the latter protein sequences with P. gingivalis HcpR is shown in Fig. 2B. We noted differences in two DNA specificity-conferring amino acids, 180 and 181 (residues underlined in Fig. 2B). These amino acids bind DNA; R180 correlates with G3, and Q181 correlates with G6 in HcpR.
The 684-bp hcpR gene codes for a 228-amino-acid protein (Fig. 2). We built a model of full-length HcpR using Dnr from Pseudomonas aeruginosa (structure 3dkw) as a template (21). Several functional regions, as identified using bioinformatics analysis, included amino acids 25 to 135, which code for the CAP-ED superfamily region that encompasses the ligand binding site and the flexible hinge region; the region including amino acids 170 to 210 codes for a helix-turn-helix (HTH) family, which is an HTH DNA-binding site (Fig. 2C). The region between the ligand-binding domain and the HTH domain comprises a dimerization domain (also known as a dimerization helix). Next, we compared the predicted structure of P. gingivalis HcpR with that of P. aeruginosa Dnr (Fig. 2D). Most of the protein could be overlaid with the structure of Dnr; however, differences in the HTH domain and the ligand-binding domain were noted, indicating that the DNA-binding sites may differ in the case of P. gingivalis HcpR compared to P. aeruginosa Dnr. Also, the differences in the ligand-binding domain may indicate differences in the nature of the substrate to which the regulator responds. In addition, we compared the predicted structure of P. gingivalis HcpR with that of a putative HcpR from T. maritima (the structure of the C-terminal truncated TM1171 is available; the disordered HTH domain was cleaved to promote crystallization of the protein) (52) (results not shown). Again, extensive structural similarity between the two protein structures was noted; however, differences in the ligand-binding domain were observed, indicating that the substrate specificity for both regulators may differ.
The 684-bp hcpR gene encodes a 26-kDa protein. In this study, we disrupted the hcpR gene at the SspI site located 291 bp within the gene (Fig. 2A). The mutant strain, designated V2807, grew in a manner similar to that of the parental W83 strain in BHI medium. In addition, it formed black-pigmented colonies when grown on blood agar plates (results not shown). The growth characteristics of the mutant and parental P. gingivalis strains were investigated under anaerobic and aerobic (6% oxygen) conditions. This oxygen concentration was used as P. gingivalis does not grow under higher oxygen concentrations (32). Growth studies were done in mycoplasma broth without cysteine. The growth characteristics of the parental and mutant strains were indistinguishable under anaerobic conditions (data not shown). Also, under aerobic conditions both strains grew in a similar manner (data not shown). Thus, we concluded that HcpR is not required for P. gingivalis growth under either aerobic or anaerobic conditions.
HcpR-like regulators are implicated in bacterial sensitivity to nitrosative stress (9). To gain insight into the role of P. gingivalis HcpR in response to nitrosative stress, growth studies were done using broth cultures. Our studies regarding the response of the parental strain to nitrate and nitrite verified the results we obtained earlier, and they showed that P. gingivalis is highly tolerant to nitrate (grows with 40 mM nitrate) and also could grow in the presence of 1 mM nitrite (Fig. 3A). The V2807 strain also grew in the presence of 40 mM nitrate (Fig. 3B). However, significant growth reduction was observed when 0.2 mM nitrite was used (P < 0.05, W83 versus V2807 under 0.2 mM nitrite, Student's t test). Furthermore, V2807 did not grow in the presence of 1 mM nitrite (Fig. 3B) (*, P < 0.002, W83 versus V2807 under 1 mM nitrite, Student's t test). Similar results were obtained using a disk diffusion assay; while no growth inhibition was observed for the wild-type W83 strain using 2 M nitrite, a significant zone of inhibition was noted for the V2807 strain (data not shown). This result significantly differed from the growth inhibition observed in the presence of 1% peroxide, where a very sharp and clear inhibition zone of similar size was observed for the two strains examined, indicating that HcpR plays no role in the response of P. gingivalis to peroxide (data not shown). These results indicate that HcpR plays a role in adaptation to nitrite but is not required for tolerance to oxidative stress.
To further verify the role of HcpR in P. gingivalis growth in the presence of nitrite, we constructed the hcpR mutant revertant strain, V2835. Growth studies were done to examine the susceptibility of the strains to nitrate and nitrite. Significant growth inhibition was noted for the V2807 strain grown in medium supplemented with 1 mM nitrite, compared to the parental W83 strain (data not shown). However, restoration of the hcpR gene in V2835 reduced the nitrite inhibition to the level seen in the parental strain, thus confirming that the nitrite susceptibility observed in V2807 is due to disruption of the hcpR gene. Similarly low levels of growth inhibition were observed for all strains tested when mycoplasma medium was supplemented with 40 mM nitrate (data not shown).
The studies described above were done using mycoplasma broth without hemin to simulate conditions in healthy sites, such as nonbleeding periodontal pockets. However, P. gingivalis also encounters environments rich in hemin sources, such as bleeding periodontal pockets, and thus we also examined the role of hemin on bacterial susceptibility to nitrite and nitrate. When high hemin concentrations (5 μg/ml) were present, both strains were able to grow with 40 mM nitrate or 1 mM nitrite (results not shown). However, reduced growth of the mutant strain V2807 compared to the W83 strain was observed when higher concentrations of nitrite were used (2 mM or higher) (Fig. 3C and andD).D). Growth of the V2835 strain containing restored hcpR gene with both 2 mM and 4 mM nitrite was similar to that of the parental W83 strain (Fig. 3C and andDD).
As nitrite can be readily generated from nitric oxide as well as to test the susceptibility of the HcpR mutant to other types of nitrosative stress encountered under physiological conditions, we compared the susceptibility of both strains to nitrite and GSNO. As shown in Fig. 4, the V2807 strain is more susceptible not only to nitrite but also to GSNO (*, P < 0.001, W83 versus V2807 under 30 nM GSNO, Student's t test).
In order to determine which genes are regulated by HcpR, we compared transcript levels from the parental W83 strain and the mutant V2807 strain grown in mycoplasma media with and without nitrosative stress for several genes regulated upon exposure to nitrite or GSNO (Tables 1 and and2).2). As shown in Table 3, the major gene regulated by HcpR was hcp. qRT-PCR analysis showed that hcp was drastically upregulated by nitrite and GSNO exposure; however, no such upregulation was detected in the HcpR-deficient V2807 strain (Table 3). The induction of expression of hcp was also observed in the HcpR revertant strain, V2835, thus verifying that the reduced expression of hcp in V2807 is indeed due to the mutation in the hcpR gene. Finally, we also tested the effect of OxyR mutation using strain V2798 (see Table S1 in the supplemental material) on expression of hcp; again nitrite exposure resulted in significant overexpression of the gene, indicating that OxyR plays no role in regulation of hcp in P. gingivalis. These results indicate that HcpR upregulates hcp expression upon exposure to nitrite. Other genes with reduced expression included PG0616, PG1820, PG2213, and PG1551 (Table 3). The reduced gene expression indicates that HcpR may also be an activator of these genes in the presence of nitrite. All regulators detected as upregulated in our array study were also upregulated using qRT-PCR analysis (Table 3), indicating that some of this regulation of the above-described gene may be indirect. Interestingly, PG1181, reported to be the most upregulated gene in the presence of nitric oxide (7), was slightly downregulated in our study.
To examine the molecular basis of this regulation, we scrutinized the DNA sequence of the hcp promoter sequence for the HcpR-binding site. The hcp gene is flanked by intergenic sequences (IGS) and is not part of an operon (Fig. 5A). This was further verified by determination of the hcp transcriptional start site using high-throughput sequencing of the P. gingivalis transcriptome (Fig. 5B and data not shown). The transcriptional start site was observed to be located 29 bp upstream of the hcp translational start site (Fig. 5B and data not shown). qRT-PCR with primers originating within the predicted transcript (1R and 2R) showed the presence of a template whereas no product was detected using a primer upstream of the transcriptional start site (3R) (results not shown), thus further reinforcing the results obtained from the transcriptome analysis. We identified a direct repeat sequence, characteristic of the Fnr binding site (TGTCGCnnnnGCGACA), which was 81 bp upstream of the translational start site of hcp (Fig. 5B). To verify binding of HcpR to the predicted DNA sequence, we performed an electrophoretic mobility shift assay with both E. coli cell lysates and purified rHcpR (data not shown). As shown in Fig. 5C, reduced migration of a DNA fragment containing the HcpR binding site was observed in E. coli cell lysates expressing HcpR (lane 2). No reduced mobility was observed in the lane containing E. coli lysate grown in the absence of IPTG (Fig. 5C, lane 3). Furthermore, the purified rHcpR bound the hcp promoter DNA. The binding was observed only when rHcpR reconstituted with hemin was used (lanes 6 to 9). Also, the binding was reduced by unlabeled hcp promoter DNA (lanes 7 and 9), indicating that the binding is specific. These results indicate that HcpR binds to the hcp promoter sequence and directly regulates hcp expression.
P. gingivalis adheres, invades, and survives for prolonged periods of time in a variety of host cells. Thus, we investigated the role of HcpR in the survival of P. gingivalis in host cells using a HUVEC and HOK survival assay as described by Ueshima et al. (55). As shown in Table 4, the parental strain W83 was recovered from HUVECs following total microbial challenge (total interaction) and from the intracellular environment (intracellular survival). However, we recovered no live V2807 from HUVECs. Next, we examined the ability of the bacterium to survive with HOKs. While live bacteria were recovered from cells challenged with wild-type W83 P. gingivalis, drastically reduced numbers of microbial cells were recovered when the cells were challenged with V2807. Finally, we also tested the ability of the revertant V2835 strain to survive with HUVECs. We observed that the ability of V2835 to survive with the eukaryotic cells was comparable to that of W83, thus verifying that the reduced survival of V2807 described above is due to mutation in HcpR (Table 4). These results indicate that HcpR plays a significant role in sustaining bacterial cell viability when bacteria are exposed to eukaryotic cells.
We show that P. gingivalis is capable of withstanding high concentrations of nitrite. It can grow with nitrite concentrations as high as 2 mM, the level found in the oral cavity (34). Previous work has shown the sensitivity of oral bacteria to acidified nitrite, including the ability of P. gingivalis to grow with 0.2 mM nitrite (1). Knowledge of the adaptive potential to nitrite has practical applications for nitrate-rich foods, such as vegetables (spinach, beetroot, and lettuce) or fruits, which are readily available and are recommended for daily intake. Believed to be rich sources of nutrients, these foods may also have the potential to reduce pathogenic bacterial growth by providing large amounts of nitrate that could be metabolized to toxic nitrite and consequently promote oral health.
In addition, P. gingivalis can withstand nitric oxide exposure. Thus, it can survive the innate immune response, which is especially elevated in inflamed periodontal pockets. Our data verify the results recently reported by Boutrin et al. (7), who have shown that P. gingivalis can grow in the presence of low nitric oxide concentrations.
With exposure to nitrite or nitric oxide (GSNO), the hcp gene (PG0893) was the most upregulated in our microarray study. Such results indicate that the gene product plays an indispensable role in the protection of P. gingivalis against nitrosative stress. Indeed, previous studies have shown that hcp is required for bacterial growth with nitric oxide (7). The upregulation of hcp was quantified using qRT-PCR to be more than 150-fold higher upon exposure to nitrosative stress. This result is much higher than that observed in a previous study that used NONOate as a nitric oxide donor (7). Such significant upregulation of hcp is in agreement with previous reports demonstrating drastic upregulation of this gene in other anaerobic and facultatively anaerobic bacteria exposed to nitrite (26, 46, 48, 56). In addition, array examinations of the response of Desulfovibrio vulgaris to nitrite have shown hcp2 to be upregulated 254-fold upon exposure to nitrite (26). Furthermore, significant upregulation of hcp in E. coli grown anaerobically in the presence of nitrate has been recently reported (48).
Despite the significant upregulation of hcp in response to nitrite, indicating that the gene product plays a major role in nitrosative stress protection, the mechanisms of gene regulation remained poorly characterized for a long time. The NsrR and NorR repressors were proposed to mediate regulation of hcp in facultative anaerobic bacteria (enterobacteria, beta- and alphaproteobacteria) (17) and Vibrionales, respectively (46). Very recently, the regulation of hcp in E. coli was reported to be mediated by OxyR (48). However, we demonstrate that the Fnr-like regulator, designated HcpR, activates hcp expression in the anaerobic bacterium P. gingivalis. Likewise, we observe that the P. gingivalis OxyR plays no role in regulation of the gene. Such results are in agreement with the previous report of Fnr also playing a role in regulation of hcp in anaerobically grown E. coli (17). Other differences in regulation could be speculated to be based on different metabolisms of the bacteria; P. gingivalis is an anaerobic organism, and E. coli is a facultatively anaerobic bacterium. Despite extensive biochemical and structural studies, the physiological role of Hcp is still not well defined (3, 11, 18, 56) and its contribution to microbial metabolism is unknown. Its role in nitrosative stress may indicate that it acts through both common intermediates generated during nitrite, nitric oxide, and hydroxylamine metabolism/detoxification (Fig. 6). Another intriguing finding was the genomic location/organization of the gene coding for Hcp. In facultative anaerobes, hcp is followed by a gene encoding a putative NADH oxidoreductase, while in obligate anaerobes, hcp loci lack the NADH oxidoreductase-coding gene (46, 56). In addition, in some bacteria, such as Desulfovibrio vulgaris, two genes code for Hcp, hcp1 and hcp2. While hcp1 expression is unaffected by nitrite, there is significant upregulation of hcp2 in the presence of nitrite (26). These results are consistent with the different roles of Hcp proteins in various bacteria. Thus, possibly different regulation mechanisms are required to activate expression of these genes.
HcpR is part of a multigene locus. Two intriguing observations were noted. First, HcpR is encoded downstream of another regulator with similarity to HspR (12). Second, the other upstream genes code for cell envelope biogenesis. In addition, Hsp regulates cell surface biogenesis in response to heat (12). Thus, it is possible that the P. gingivalis locus is involved in a multifaceted response to environmental stress. Furthermore, the genomic locus coding for HcpR differs in P. gingivalis compared to other Bacteroides bacteria (data not shown). In Prevotella intermedia 17, HcpR is encoded downstream of Hcp, whereas in B. fragilis and B. thetaiotamicron it is encoded upstream of the Hcp-coding gene. Interestingly, in T. maritima, a locus encoding functionally similar proteins to that found in Bacteroidetes is present (data not shown). As this is an ancient anaerobic Gram-negative bacterium (28), it is probable that the hcp-hcpR locus has been derived from this bacterium. Otherwise, it may have been acquired from the archaea due to the fact that T. maritima has multiple genes of archaeal origin (41). Although many of these genes confer the ability of the bacterium to live at high temperatures, it is likely that genes required for adaptation to other stress conditions were acquired as well.
Our studies suggest hemin as an important player in defense against nitrosative stress. First, we found upregulation of the gene hmuY (PG1551), coding for the major hemin uptake protein (33, 50), in the presence of the nitric oxide-generating agent GSNO. As hemin plays an antioxidative role, such upregulation may also be a mechanism used by P. gingivalis in defense against nitrosative stress. However, P. gingivalis encodes the nitrite reductase system NrfHA, which is a multiheme protein system. NrfA is a periplasmic pentaheme cytochrome c nitrite reductase that can reduce nitrite, nitric oxide, or hydroxylamine to ammonia, and membrane-anchored NrfH is a tetraheme cytochrome c menaquinol dehydrogenase (Fig. 6) (37). This system plays a crucial role in defense against nitrite, nitric oxide, and hydroxylamine stress in Wolinella succinogenes (29, 37); thus, it is probable that NrfHA plays a similar role in P. gingivalis. Lack of hemin would be expected to render the NrfHA system inactive. Therefore, upregulation of the hemin uptake system may be a way to ensure sufficient provision of hemin to keep the system in its active form. It is likely that the system gets oxidized rapidly as it is located in the periplasm; thus, high levels of hemin are required to keep it functional.
A second connection of hemin to the response to nitrosative stress was seen when examining regulatory mechanisms; interestingly, HcpR appeared to be an activator of PG1551, although this regulation may be indirect as no binding of HcpR to DNA containing the PG1551 promoter sequence was observed (results not shown). Hcp is an iron protein, and higher hemin intake may ensure that enough iron is available to keep the elevated levels of Hcp in its active form. Of note, hemin is required for HcpR binding to DNA. As hcpR expression is not affected by nitrosative stress, mechanisms other than transcriptional control govern the transition of the regulator to its active form. We show that hemin is required for HcpR binding to the hcp promoter. Hemin has also been proposed to be involved in the conversion of Dnr to an active DNA-binding form (10). P. gingivalis HcpR is a Dnr-like regulator that also is required for tolerance to nitrosative stress (45). Although it can adopt the typical structure of Fnr-like or Crp-like regulators (20, 23), differences exist in the substrate-binding and DNA-binding domains, indicating that it may regulate different genes and respond to different stimuli. These results are consistent with the Dnr-regulating response of genes coding for denitrification mechanisms (45) and P. gingivalis HcpR primarily regulating hcp. However, based on the structural similarity between Dnr and HcpR as well as their roles in response to nitrosative stress, it is probable that the mechanism of activation is similar for both proteins. Thus, a higher provision of hemin would also ensure HcpR activation. The hemin-based activation mechanism of HcpR is consistent with the protein lacking the conserved cysteines required for binding of the Fe-S cluster, as seen in the Fnr-like regulators, and also ruling out Cys switch or Cys nitrosylation, which are the molecular basis for activation of the OxyR regulator (48). Finally, in silico analysis has shown that the Dnr-like regulator may regulate hcp expression in A. ferrooxidans and Thermochromatium tepidum (46), thus further bridging the gap between the roles, and likely molecular mechanism, of the two regulators.
We show that the growth of the P. gingivalis HcpR mutant V2807 in the presence of nitrite and GSNO is reduced relative to the parental strain. The ability of P. gingivalis to adhere, invade, survive, and multiply in a variety of host cells is well documented (13, 62). As release of oxidative and nitrosative reactive species is part of the immune defense against invading pathogens, P. gingivalis would be expected to have protective mechanisms against such stress. Our data indicate that HcpR is required for survival in the presence of nitrosative stress. We also showed that survival of V2807 in host cells is drastically reduced. The nearly complete killing of the mutant bacteria by the host cells indicates that nitrosative stress is the major effector against P. gingivalis. These studies, combined with the inhibition of growth of the strain in the presence of nitrite, which is characteristic of the oral cavity, establish the biological significance of the regulator HcpR.
In summary, this work identifies HcpR as a novel regulator of P. gingivalis adaptation to nitrosative stress. In addition, it verifies the role of the HcpR family of regulators in the regulation of hcp expression. The important role of HcpR in nitrosative stress warrants further investigation of the protein, including its regulatory mechanism, through biochemical and structural studies.
This research was supported by USPHS grants 5K22DE14180, R01DE016124 and R01DE018039 from the National Institute of Dental and Craniofacial Research awarded to Janina P. Lewis.
The determination of the genomic sequence of P. gingivalis was carried out collaboratively by The Institute for Genomic Research (TIGR) and The Forsyth Dental Center database with support from NIDCR. The genomic sequence of P. gingivalis W83 was obtained from JCVI (http://www.jcvi.org) and the Los Alamos Oral Pathogen Sequence Database (http://www.oralgen.lanl.gov). We thank Anuya Paranjapee for help with the EMSA studies, Karina Martinez for carrying out the qRT-PCR analyses, and Hiroshi Miyazaki for his help preparing cell cultures for this study.
Published ahead of print 9 July 2012
Supplemental material for this article may be found at http://iai.asm.org/.