Anti-HmuY Antibodies Specifically Recognize Porphyromonas gingivalis HmuY Protein but Not Homologous Proteins in Other Periodontopathogens

Given the emerging evidence of an association between periodontal infections and systemic conditions, the search for specific methods to detect the presence of P. gingivalis, a principal etiologic agent in chronic periodontitis, is of high importance. The aim of this study was to characterize antibodies raised against purified P. gingivalis HmuY protein and selected epitopes of the HmuY molecule. Since other periodontopathogens produce homologs of HmuY, we also aimed to characterize responses of antibodies raised against the HmuY protein or its epitopes to the closest homologous proteins from Prevotella intermedia and Tannerella forsythia. Rabbits were immunized with purified HmuY protein or three synthetic, KLH-conjugated peptides, derived from the P. gingivalis HmuY protein. The reactivity of anti-HmuY antibodies with purified proteins or bacteria was determined using Western blotting and ELISA assay. First, we found homologs of P. gingivalis HmuY in P. intermedia (PinO and PinA proteins) and T. forsythia (Tfo protein) and identified corrected nucleotide and amino acid sequences of Tfo. All proteins were overexpressed in E. coli and purified using ion-exchange chromatography, hydrophobic chromatography and gel filtration. We demonstrated that antibodies raised against P. gingivalis HmuY are highly specific to purified HmuY protein and HmuY attached to P. gingivalis cells. No reactivity between P. intermedia and T. forsythia or between purified HmuY homologs from these bacteria and anti-HmuY antibodies was detected. The results obtained in this study demonstrate that P. gingivalis HmuY protein may serve as an antigen for specific determination of serum antibodies raised against this bacterium.


Introduction
Periodontitis is a group of multifactorial, inflammatory infectious diseases, initiated by an ecological shift in the composition of subgingival biofilm, resulting in inflammation and of HmuY, we also aimed to characterize responses of antibodies raised against the HmuY protein or its epitopes to the closest homologous proteins from P. intermedia and T. forsythia.

Ethics statement
All animal protocols and procedures used to produce custom designed polyclonal antibodies purchased from GenScript Corporation (GenScript USA Inc., Piscataway, NJ 08854) were approved by the GenScript Corporation Institutional Animal Care and Use Committee (IACUC, #ANT11-001). The Office of Laboratory Animal Welfare (OLAW) of the U.S. National Institutes of Health (NIH) provides guidance and interpretation of the Public Health Service (PHS) Policy on Humane Care and Use of Laboratory Animals. GenScript received OLAW's Animal Welfare Assurance. OLAW's Animal Welfare Assurance accentuates the responsibilities and procedures of GenScript regarding the care and use of laboratory animals. The GenScript developed partnership with the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International, which is a private, nonprofit organization that promotes the humane treatment of animals in science through voluntary accreditation and assessment programs. GenScript has earned AAALAC accreditation, demonstrating their commitment to responsible animal care and use through ongoing voluntary participation in AAALAC programs.

Construction of expression plasmids
To construct expression plasmids containing sequences encoding P. gingivalis HmuY homologs from P. intermedia and T. forsythia, DNA sequences encoding predicted mature proteins were PCR-amplified using the genomic DNAs as templates and primers listed in Table 1. For this purpose, P. intermedia DNA sequences encoding PinA (NCBI accession number WP_014709321) and PinO (NCBI accession number WP_014708291) were used. In the case of DNA encoding T. forsythia Tfo, some discrepancies in the gene sequence compared to the DNA sequence deposited in databases (NCBI accession number YP_005014932) were found. Therefore, in this study we determined the corrected DNA sequence encoding Tfo. All amplified PCR products were ligated into the pTriEx-4 vector (Merck), resulting in sequences encoding untagged proteins.
Overexpression and purification of proteins P. gingivalis HmuY lacking the first 25 amino acid residues (NCBI accession number CAM 31898) was overexpressed using a pHmuY11 plasmid and Escherichia coli ER2566 cells (New England Biolabs) and purified from a soluble fraction of the E. coli cell lysate as described previously [13]. As the soluble protein shed from the outer membrane, the purified HmuY lacked the signal peptide and first five amino acid residues (CGKKK) of the nascent secreted protein [14,15]. P. intermedia PinA, PinO and T. forsythia Tfo were overexpressed in E. coli ER2566 and Rosetta (DE3) strains (New England Biolabs or Merck), respectively. PinA, PinO and Tfo, lacking predicted signal peptides (Table 1), were prepared using purification procedures established in our laboratory. All proteins were purified from whole E. coli cell lysates obtained after sonication and centrifugation, as described for P. gingivalis HmuY [13] with some modifications. Briefly, PinO, PinA and Tfo purification was carried out using DEAE-Sephacel chromatography (50 mM Tris/ HCl buffer, pH 7.6, containing 30 mM NaCl). Then, hydrophobic chromatography using Phenyl Superose (HR 5/5, Pharmacia) was used. The proteins were bound to the chromatographic resin in 50 mM sodium phosphate buffer, pH 7.0, containing 2 M ammonium sulfate, and eluted using linear salt gradient 2.0-0.0 M. As the final step, gel filtration using Sephadex G-75 (50 mM Tris/HCl buffer, pH 7.6, containing 400 mM NaCl) was used.

Immunization of rabbits
Custom designed polyclonal antibodies were purchased from the commercial company (Gen-Script USA Inc.), which is a dedicated biology research service provider. P. gingivalis HmuY and three synthetic peptides derived from the amino acid sequence of HmuY, conjugated with keyhole limpet hemocyanin (KLH), were used to immunize rabbits. Each experimental animal had one ear tag with lot number and one animal was kept in one cage with cage card. Animals were maintained at 16-22°C and relative humidity 30-70%. Polyclonal antibodies were produced by subcutaneously injecting New Zealand white rabbits at weight 2.0-2.5 kg (2 rabbits per each antigen, 8 rabbits during the entire experiment). In addition, sera from 2 rabbits immunized with the purified HmuY protein in our previous study [14] were used. The rabbits received one injection with the antigen (0.2 mg per injection) emulsified 1:1 (0.25 ml per rabbit) with Freund's complete adjuvant, and the same amount of antigen with Freund's incomplete adjuvant on day14, 35 and 56. Before immunization and after immunization on day 21, 42, and 63 blood was collected (2.5 ml per animal) from ear margin veins. All animals were healthy trough the entire experiment and all serum samples were properly collected. Pre-immune sera (protocol #SC1088) and immune sera after test and final bleeds (protocols #SC1247 and # SC3031), containing anti-HmuY1 (purified protein), anti-HmuY1-1 (GKKKDEPNQPST-PEC), anti-HmuY1-2 (SKGEVVNVTDYKNDC), and anti-HmuY1-3 (CEMGPDGHQ-MEYEEQG) antibodies were used for all experiments.

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting
The reactivity of respective antibodies was investigated using purified proteins or bacterial cell lysates. Bacterial cultures were centrifuged for 20 min at 20,000×g at 4°C. Bacterial pellets were washed with 20 mM sodium phosphate buffer, pH 7.4, containing 140 mM NaCl (PBS) and suspended in PBS to OD 600 = 1.0. Samples corresponding to 20 μl of the bacterial culture or protein samples (1, 10 and 100 ng or 1 and 5 μg per lane) were denatured and separated on 12% SDS-PAGE gels, and subsequently transferred onto nitrocellulose membranes (Whatman). Nonspecific binding sites were blocked with 5% skim milk in PBS with addition of 0.1% Tween (PBST). HmuY or its epitopes were visualized with respective sera and secondary goat anti-rabbit IgG antibodies conjugated with horseradish peroxidase (HRP, Sigma). All sera were used at 1:10,000 dilution. The reaction was developed using chemiluminescence reagents (Western Lightning Plus-ECL, Perkin Elmer).

Results and Discussion
Infection with periodontopathogens leads to humoral immune responses with elevated serum antibodies to periodontal species in patients with periodontitis [29,[38][39][40][41][42][43][44]. In addition to P. gingivalis, antibody responses to several other periodontopathogens have been found, including P. intermedia [45,46] and T. forsythia [47]. More importantly, recent data demonstrated that an elevated antibody level against P. gingivalis indicated advanced periodontal disease and suggested progression of atherosclerosis, hypertension and rheumatoid arthritis [48][49][50][51][52]. Therefore, efficient screening of patients is essential for estimation of general health and disease stage as well as subsequent treatment efficiency. In our recent study, we presented an extensive phylogenetic analysis of P. gingivalis HmuY protein and the hmuY gene [36]. The HmuY from P. gingivalis was placed at the base of other Bacteroidia sequences represented by Bacteroides, Prevotella and Tannerella. In contrast, low identity to homologous sequences found in other bacteria was observed. We also demonstrated the presence of antibodies directed against P. gingivalis HmuY in sera of patients with chronic periodontitis [29]. This study extends our previously published data and presents characterization of the HmuY protein with regard to its future application to determine specific anti-P. gingivalis antibodies in serum. The best antigen targets are bacterial components, which are constitutively expressed and secreted from the bacterium or exposed on the pathogen's surface. The use of whole-sonicate antigens, which encompass many of the P. gingivalis proteins [53,54], would increase the possibility of detecting responses, but also increase the potential for false positive results because of cross-reacting antibodies to components of other bacteria. Therefore, to detect specific anti-P. gingivalis antibodies which do not cross react with components produced by other bacteria, we employed purified HmuY protein or its selected epitopes as antigens (Fig. 1). Since other periodontopathogens produce HmuY homologs, we aimed to characterize responses of antibodies raised against HmuY protein or its epitopes to the most homologous proteins from P. intermedia and T. forsythia. Taking into account the amino acid sequences and the theoretically modeled three-dimensional protein structures, we found the two closest HmuY homologs in P. intermedia 17, termed here PinA and PinO, and one HmuY homolog from T. forsythia ATCC 43037, termed here Tfo (Fig. 2). For the latter, we identified corrected DNA (Fig. 3A) and amino acid (Fig. 3B) sequences (EMBL accession number LN624459). It appears that previously in databases a nucleotide sequence possessing a 6-bp insertion and nucleotide substitutions has been deposited. In contrast, the amino acid sequence of Tfo identified in our laboratory contains both a two-amino-acid deletion and substitutions of nine amino acids (Fig. 3C). We did not identify the HmuY homolog in T. denticola, the third member of the "red complex" of periodontopathogens. In this study we overexpressed all proteins and established purification procedures for the PinA, PinO and Tfo proteins.
First we characterized all antibodies produced in rabbits against purified HmuY protein using Western blotting. For this purpose, we examined antibodies against the purified HmuY protein which were raised in 4 rabbits, and antibodies against each synthetic peptide which were raised in 3 pairs of rabbits. We did not observe any significant differences in reactivity between antigens tested and the sera of the different animals (data not shown). Therefore, in this study representative results from single rabbits only are shown. As shown in Fig. 4, the denatured HmuY protein reacted with antibodies directed against the HmuY protein as well as with antibodies directed against all three epitopes, with the highest affinity towards the purified protein and epitope 2. Further analyses showed that antibodies raised against the purified HmuY protein did not recognize HmuY homologs from P. intermedia and T. forsythia (Fig. 5A). This strongly demonstrates that antibodies raised against P. gingivalis HmuY recognize only this antigen, but not homologous proteins from other periodontopathogens. Also, whilst antibodies directed against epitopes 1, 2, and 3 reacted with purified HmuY protein, they did not recognize HmuY homologs  examined in this study (Fig. 5B, 5C, 5D). Similar experiments performed with denatured P. gingivalis cell lysates showed that antibodies raised against purified HmuY protein (anti-HmuY 1) exhibited the highest specificity and recognized the HmuY protein only (Fig. 6). Other antibodies were less specific, and in addition to detection of HmuY protein (Fig. 6, indicated with arrows), they cross reacted with other P. gingivalis antigens. This non-specific effect was corroborated when the reactivity of pre-immune serum was tested with P. gingivalis lysates (data not shown). Based on our data we concluded that antibodies raised against HmuY protein (most sensitive    Reactivity of anti-HmuY antibodies with P. gingivalis HmuY protein or its homologs from P. intermedia and T. forsythia. ELISA was carried out using antibodies raised against purified HmuY protein (anti-HmuY 1) (A) and epitope 2 (anti-HmuY 1-2) (B). No reactivity was observed for anti-HmuY 1-1 (epitope 1) and anti-HmuY 1-3 (epitope 3) antibodies. Overexpressed and purified HmuY (P. gingivalis), PinA, PinO (P. intermedia) and Tfo (T. forsythia) proteins were used as antigens.
doi:10.1371/journal.pone.0117508.g008 and specific) and epitope 2 (less sensitive and specific) detected HmuY protein compared to antibodies raised against epitope 1 and 3. We did also not observe significant differences between two different P. gingivalis strains used. The P. gingivalis strain A7436 was originally isolated from a refractory periodontitis patient. This strain is highly similar (both in its genetic background and phenotypic properties) to strain W83. Both strains belong to a group of more virulent P. gingivalis isolates and are often found in patients with chronic periodontitis [36]. Conversely, P. gingivalis ATCC 33277 strain was used because it belongs to a less virulent group of strains and is known to exhibit a different genetic background and attenuated virulence properties.
Characterization of antibodies was also carried out using native, non-denatured antigens. In agreement with Western blotting results, only antibodies raised against the purified HmuY protein or raised against epitope 2 reacted with HmuY protein used as an antigen, as determined by the ELISA assay (Fig. 7). When purified HmuY homologs from P. intermedia and T. forsythia were used as antigens, no reactivity with antibodies directed against purified HmuY protein (Fig. 8A) or epitope 2 (Fig. 8B) was observed. In addition, there was no reactivity when antibodies raised against epitope 1 and 3 were used (data not shown). When live P. gingivalis Reactivity of anti-HmuY antibodies with bacteria. ELISA was carried out using antibodies raised against purified HmuY protein (anti-HmuY 1) and P. gingivalis wild-type (A7436 and ATCC 33277) and hmuY mutant (TO4) strains. Bacteria were cultured in high iron/heme (A) or low-iron/heme (B) media. No reactivity was observed for anti-HmuY 1-1 (epitope 1), anti-HmuY 1-2 (epitope 2), or anti-HmuY 1-3 (epitope 3) antibodies. Live bacteria were used as antigens.
doi:10.1371/journal.pone.0117508.g009 bacteria were used as antigens in the ELISA assay, only reactivity with antibodies raised against purified HmuY protein was visible (Fig. 9), especially in bacteria which were starved of iron and heme (Fig. 9B), compared with bacteria cultured under high-iron/heme conditions (Fig. 9A). As a negative control we used the hmuY mutant strain (TO4), which does not produce the HmuY protein. Importantly, no reactivity was found when live cells of P. intermedia or T. forsythia were used as antigens (data not shown).
In conclusion, the results obtained in this study demonstrate that P. gingivalis HmuY protein may serve as an antigen for specific determination of antibodies raised against this bacterium. Therefore, purified HmuY protein will be used to determine levels of anti-P. gingivalis antibodies in sera of patients with chronic periodontitis and patients with other forms of periodontal diseases. These studies are underway in our laboratory.