当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 感染与免疫杂志 > 2005年 > 第8期 > 正文
编号:11253993
Expression of Arg-Gingipain RgpB Is Required for Correct Glycosylation and Stability of Monomeric Arg-Gingipain RgpA from Porphyromonas ging
     MRC Molecular Pathogenesis Group, Centre for Infectious Disease, Institute of Cell and Molecular Science, Barts & The London, Queen Mary's School of Medicine and Dentistry, 4 Newark Street, London E1 2AT, United Kingdom

    School of Biological and Chemical Sciences, Birkbeck College, University of London, Gordon Square, London WC1 H0PP, United Kingdom

    ABSTRACT

    Arg-gingipains are extracellular cysteine proteases produced by the gram-negative periodontal pathogen Porphyromonas gingivalis and are encoded by rgpA and rgpB. Three Arg-gingipains, heterodimeric high-molecular-mass Arg-gingipain HRgpA comprising the -catalytic chain and the -adhesin chain, the monomeric soluble Arg-gingipain comprising only the -catalytic chain (RgpAcat), and the monomeric membrane-type heavily glycosylated Arg-gingipain comprising the -catalytic chain (mt-RgPAcat), are derived from rgpA. The monomeric enzymes contain between 14 and 30% carbohydrate by weight. rgpB encodes two monomeric enzymes, RgpB and mt-RgpB. Earlier work indicated that rgpB is involved in the glycosylation process, since inactivation of rgpB results in the loss of not only RgpB and mt-RgpB but also mt-RgpAcat. This work aims to confirm the role of RgpB in the posttranslational modification of RgpAcat and the effect of aberrant glycosylation on the properties of this enzyme. Two-dimensional gel electrophoresis of cellular proteins from W50 and an inactivated rgpB strain (D7) showed few differences, suggesting that loss of RgpB has a specific effect on RgpA maturation. Inactivation of genes immediately upstream and downstream of rgpB had no effect on rgpA-derived enzymes, suggesting that the phenotype of the rgpB mutant is not due to a polar effect on transcription at this locus. Matrix-assisted laser desorption ionization-time of flight analysis of purified RgpAcat from W50 and D7 strains gave identical peptide mass fingerprints, suggesting that they have identical polypeptide chains. However, RgpAcat from D7 strain had a higher isoelectric point and a dramatic decrease in thermostability and did not cross-react with a monoclonal antibody which recognizes a glycan epitope on the parent strain enzyme. Although it had the same total sugar content as the parent strain enzyme, there were significant differences in the monosaccharide composition and linking sugars. These data suggest that RgpB is required for the normal posttranslational glycosylation of Arg-gingipains derived from rgpA and that this process is required for enzyme stabilization.

    INTRODUCTION

    Porphyromonas gingivalis, a gram-negative anaerobic bacterium, is considered an important etiological agent in periodontal disease. The organism produces several extracellular proteolytic enzymes of which the Arg-X-specific and Lys-X-specific enzymes are predominant. These cysteine proteases cause the degradation of several physiologically important proteins, including collagens (6), fibrin and fibrinogen (28, 44), and fibronectin (47, 52). They also account for deregulatory effects on several host systems (21, 34, 43, 49, 55) and are therefore considered important virulence determinants. The adhesin domains of Arg-gingipains mediate the adherence of P. gingivalis to epithelial cells (8).

    The extracellular Arg-X-specific proteases of P. gingivalis W50 are encoded by two genes, rgpA and rgpB (9). Three enzymes, heterodimeric high-molecular-mass Arg-gingipain HRgpA comprising the -catalytic chain and the -adhesin chain, the monomeric soluble Arg-gingipain comprising only the -catalytic chain (RgpAcat), and the monomeric membrane-type heavily glycosylated Arg-gingipain comprising the -catalytic chain (mt-RgpAcat), are derived from rgpA (41) and two enzymes, RgpB and mt-RgpB, from rgpB (40). HRgpA occurs as a dimer (41) or multimer (39) composed of a catalytic chain (55 kDa) noncovalently associated with a polypeptide(s) derived from the long C-terminal extension of the initial full-length translation product. RgpAcat (55 kDa) and mt-RgpAcat (70 to 80 kDa) are monomers of the catalytic chain with different amounts of posttranslational modifications (10). The rgpB protease gene lacks the long C-terminal extension coding region of rgpA but is otherwise almost identical (40), and the two protease isoforms derived from this gene, RgpB and mt-RgpB, are structurally and kinetically similar to the monomeric RgpA isoforms. Mikolajczyk et al. (30) expressed the full-length precursor of RgpB and showed that three sequential autolytic processing steps at the N and C termini are required for full activity and that the N-terminal propeptide may serve as an intramolecular chaperone rather than as an inhibitory peptide.

    Biochemical analysis of the Rgps has established that the monomeric enzymes are glycoproteins containing between 14% (RgpAcat and RgpB) and 30% (mt-RgpAcat and mt-RgpB) carbohydrate by weight (10). The monosaccharide composition of RgpAcat has been examined in some detail and contains at least nine different sugars, including high levels of GalNAc and Neu5(Ac). Furthermore, the glycan additions to this isoform are immunologically related to a polysaccharide preparation of this organism, suggesting a common link between the maturation pathway of the Rgps and the synthesis of this macromolecule. The functional significance of the glycosylation of these proteases has not been determined, although the recognition of these glycan additions by periodontal serum immunoglobulin G antibody may indicate a role in immune subversion (46). Takii et al. (50) have recently isolated a cell-associated gingipain complex of a 660-kDa mass existing as a homodimer of two catalytically active monomers which comprise their catalytic and adhesin domains. Two-dimensional gel electrophoresis and immunoblot analyses revealed the association of lipopolysaccharide with the catalytic domains and a hemagglutinin domain of Rgp and Kgp in the complex. Takii et al. (50) also showed that the functional domains of lipopolysaccharide were structurally masked by the complex proteins, which suggested the importance of the complex in the evasion of host defense mechanisms as well as in host tissue breakdown.

    The molecular mechanism of glycosylation of these enzymes is not known. However, results obtained previously in this laboratory (2) have implicated the products of rgpB in the posttranslational events leading to the generation of the different enzyme isoforms derived from rgpA. Insertional inactivation of rgpB led not only to significantly reduced levels of total Arg-X protease activity but also to alterations to the chromatographic characteristics of the monomeric isoforms derived from rgpA (2). Most strikingly, the vesicle-associated monomer, mt-RgpAcat, was shown by active-site labeling and migration on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels to lack the extensive posttranslational modifications present on the wild-type enzyme. In addition, RgpAcat from the mutant appeared to be chemically different from the parent enzyme since the procedures routinely used for the purification of wild-type RgpAcat were unsuccessful in isolating this isoform from the D7 mutant. Conversely, HRgpA, which appears to show very low levels of glycosylation in the parent strain, was unaffected with respect to the amount of enzyme present, its subunit composition, and the ease of purification in this mutant (2). These data led us to suggest that the correct maturation/modification of the monomeric isoforms of rgpA was dependent on a functional rgpB.

    The possibility that the RgpA monomeric isoforms in the rgpB mutant are aberrantly posttranslationally modified with respect to carbohydrate additions provided us an opportunity to examine the influence of glycosylation on the enzymatic properties and immune recognition of these enzymes. In this paper, we first provide confirmation of the participation of RgpB in the maturation of rgpA-derived enzymes by showing that the loss of rgpB has a specific effect on rgpA protein maturation via proteomic analysis of the wild type and mutant strain. We also show that this effect is not due to a polar effect on transcription at the rgpB locus. Second, we describe the purification, properties, monosaccharide analysis, and identification of sugars linking oligosaccharides to Ser/Thr residues in RgpAcat from strain D7 and how these are different from RgpAcat from the parent strain. These data demonstrate the importance of posttranslational modifications to the stability of the monomeric RgpAcat from this organism.

    MATERIALS AND METHODS

    Materials. DEAE-Sephacel, Sephacryl S-200HR, and phenyl-Sepharose (high performance) were obtained from Pharmacia-Biotech (St. Albans, Hertfordshire, United Kingdom). L-Arg conjugated to 4% beaded agarose and N-benzoyl-DL-Arg-p-nitroanilide (DL-BRpNA) were purchased from Sigma Chemical Co. (Poole, Dorset, United Kingdom). N-acetyl-Lys-p-nitroanilide and leupeptin were obtained from Bachem Feinchemikalein AG (Bubendorf, Switzerland). Zwittergent 3-14 detergent and Dansyl-glutamyl-glycyl-arginyl-chloromethyl ketone (DNS-EGR-CK) were purchased from Calbiochem Novabiochem UK Ltd. (Nottingham, United Kingdom). All other chemicals and reagents were the purest grades available and were from BDH Chemicals (Poole, Dorset, United Kingdom) or Sigma Chemical Co. Restriction and DNA-modifying enzymes were purchased from either Amersham Biosciences (St. Albans, Hertfordshire, United Kingdom) or Roche (East Sussex, United Kingdom). Reagents for PCR (Reddy Load) were obtained from ABgene (Epsom, Surrey, United Kingdom). DNA isolation reagents were from either Flowgen (Lichfield, United Kingdom) (chromosomal) or QIAGEN (Crawley, United Kingdom) (plasmids, DNA fragments, and amplicons). The monoclonal antibody (MAb) 1B5 used in Western blotting has been described previously (10). Rabbit antiserum to the RgpA catalytic domain was prepared by immunizing rabbits with recombinant His6-tagged RgpA (R227-R719) which was prepared in Escherichia coli XL-1 Blue transformed with a derivative of pQE3010 described previously (2). Horseradish peroxidase-labeled anti-species antibodies were supplied by DAKO A/S, Denmark.

    Bacteria and growth conditions. Porphyromonas gingivalis W50 and isogenic mutants D7 (rgpB), DE1 (dppIV, PG0503), L1 (lipA, PG0504), H7 (ubiA, PG0509), and C7 (cicA, PG0508) were cultured anaerobically on blood agar or in brain heart infusion medium (Oxoid, Basingstoke, United Kingdom) supplemented with hemin (5 μg ml–1) as described previously (2). E. coli XL1 Blue and SCS110 (Stratagene) were propagated in Luria-Bertani medium at 37°C. Ampicillin was added to 100 μg ml–1 for plasmid selection. P. gingivalis allelic exchange mutants were selected on blood agar plates containing clindamycin hydrochloride at 5 μg ml–1.

    DNA manipulations. Manipulation of DNA, transformation of E. coli, and agarose gel electrophoresis were done as described by Sambrook et al. (42).

    PCRs. Three separate amplicons corresponding to the loci shown in Fig. 1 were used to clone DNA fragments for the inactivation of dipeptidyl peptidase (dppIV), lipoic acid synthetase (lipA), and ClpXP protease (cicA) or hydroxybenzoate octaprenyl transferase (ubiA). Primer pairs used were 5'-TTATGCTCGCAGTGCAGG-3' (DppIVF1) and 5'-TGCCTCTGTAAAAAGCATCG-3' (DppIVR1), 5'-GGAATACCCGCTATCATCTC-3' (LipAF1) and 5'-CTGCTCCATTATGCTTTTCC-3' (LipAR1), and 5'-TGCGGATTGAAAGACAAGAG-3' (UbiF1) and 5'-GAATCGACAATCTCCAAACG-3' (UbiR1), respectively.

    PCR (Omnigene thermal cycler; Hybaid) cycling parameters were 25 cycles at 94°C (1 min), 55°C (50°C for UbiF1/R1), and 72°C (4 min), with a final extension at 72°C for 5 min. The products were purified from agarose gels and cloned into a blunted EcoR1-HindIII site of pUC18not (19) using Sureclone (Amersham Biosciences). The resulting plasmids were propagated in E. coli XL1 Blue or SCS110.

    Allelic exchange. Suitable restriction enzyme sites were identified in the cloned amplicons (Fig. 1), and a blunted BamHI-KpnI fragment of pVA2198 containing an Erm cassette (14) was cloned into these sites to inactivate the genes in vitro. The cassette replaced a 900-bp ClaI-BclI fragment in the amplicon for dppIV, a 330-bp BclI-BclI fragment in ubiA, and a 131-bp EcoRV-HincII fragment in lipA. In the case of cicA, the Erm cassette was inserted at an NcoI site. These constructs were then retrieved from their pUC18not derivatives by NotI digestion and electroporated into P. gingivalis W50 as described previously (40).

    Enzyme assays. Arg-X protease activity was measured at 30°C with DL-BRpNA (500 μM) as the substrate. One unit of protease catalyzes the formation of 1 μmol of p-nitroaniline min–1 in this assay system (41). L-BRpNA (250 μM) was used as substrate for Km measurements. Lys-X protease activity was measured with N-acetyl-L-lysine-p-nitroanilide (Ac-LyspNA) (250 μM) as substrate in the same reaction buffer and under the same conditions as described above.

    Enzyme purification. Enzymes from the P. gingivalis D7 mutant were purified as follows. All steps were performed at 4°C. Cells from 4 liters of a 6-day culture were removed by centrifugation at 10,000 x g (60 min; Sorvall RC 5C Plus, SLA-3000), and solid ammonium sulfate (enzyme grade) was added to the supernatant to 85% saturation as previously described (41). Solubilization of enzyme activity from the ammonium sulfate precipitate, gel filtration, and affinity chromatography on Arg-agarose columns were performed essentially as described previously (41). Elution of HRgpA bound to the Arg-agarose column was performed as described previously (41) except that 0.1 M L-arginine was used. Enzyme fractions judged to be pure by constant specific activity and SDS-PAGE were stored at 4°C. This enzyme could be purified by rechromatography on Arg-agarose (41). This Arg-X protease is referred to herein as HRgpA/D7 or HRgpA(rgpB).

    Purification of RgpAcat(rgpB). Purification of enzymes present in the effluent from affinity column fractions (above) was achieved by dialysis against 20 mM Tris-acetate buffer, pH 7.5, followed by ion exchange chromatography on DEAE-Sephacel equilibrated in the same buffer. Enzyme was eluted by a linear gradient of NaCl to a final concentration of 0.15 M in equilibration buffer. Enzyme fractions were combined, and solid ammonium sulfate was added to make the solution 2 M in ammonium sulfate and applied to a column of phenyl-Sepharose equilibrated in 20 mM Tris-acetate buffer, pH 7.5, plus 0.1 M NaCl-2 M ammonium sulfate. Pure enzyme was eluted with linear gradients of ammonium sulfate from (i) 2 M to 0.6 M and (ii) 0.6 M to 0 M. Enzyme fractions judged to be pure by high constant specific activity and SDS-PAGE were stored at 4°C. This Arg-X protease is referred to herein as RgpAcat/D7 or RgpAcat(rgpB). The protein was essentially pure at this stage.

    Determination of protein concentration. Protein concentrations were determined by measurement of absorbance at 280 nm and 260 nm (11). Protein concentration (mg ml–1) was calculated according to the formula 1.55 A280 – 0.76 A260.

    Gel electrophoresis. SDS-PAGE was carried out at 5°C in 12.5% polyacrylamide slab gels (10 by 7 by 0.15 cm) (26). Samples of protease (10 to 20 μg) were first treated with 50 μl of leupeptin (1 mg ml–1) at 22°C for 20 min, heated at 100°C for 5 min, and dried in vacuo.

    Immunoblotting. Western blotting of purified enzymes (2 μg) was performed following inhibition of the proteases with leupeptin (1 mg ml–1). The samples were electroblotted from 12.5% polyacrylamide gels onto nitrocellulose membranes (Schleicher & Schuell, Germany) and probed using monoclonal antibody 1B5 (1:100) or rabbit antiserum to the recombinant catalytic domain of RgpA (1:250). Horseradish peroxidase-labeled anti-species antibodies were used at a 1:500 dilution with diamino-benzidine as the substrate.

    N-terminal sequencing of proteins. Arg-X proteases were subjected to N-terminal sequencing after either 8 M urea-PAGE (17) for HRgpA/D7 or SDS-PAGE for RgpAcat/D7 following transfer to Immobilon membranes (Millipore) took place. Sequencing was performed by the Haemostasis Research Group, Royal Postgraduate Medical School, London, United Kingdom, using an ABI 477A gas phase sequencer for 16 to 20 cycles.

    Fluorescent labeling of proteases with DNS-EGR-CK. Six-day culture supernatants (500 μl) of P. gingivalis W50 and isogenic mutants D7, DE1, L7, H7, and C7 were treated with 750 μl of ice-cold acetone in a bath of freezing mixture (–10°C) for 30 min. The solution was centrifuged at 13,000 x g (Heraeus Biofuge) for 15 min, and the pellet was subjected to reduction and labeling (2). Pure RgpAcat enzymes from W50 and D7 were labeled as described previously (17) and reprecipitated by the addition of 225 μl of ice-cold acetone in a bath of freezing mixture (–10°C) for 30 min. Labeled protein was centrifuged at 13,000 x g for 15 min, and the pellet was treated with SDS-PAGE sample buffer for gel electrophoresis.

    Two-dimensional electrophoresis. Cells from 50 ml of P. gingivalis cultures (24 h) were harvested by centrifugation and washed three times with equal volumes of 10 mM Tris-HCl buffer, pH 7.5, with the final wash containing 1 mM N-p-tosyl-L-lysine chloromethyl ketone (TLCK) and a cocktail of protease inhibitors (Roche, Germany). Cell pellets were resuspended in 3 ml of lysis solution (0.5 to 1.0% SDS, 1 mM TLCK, cocktail of protease inhibitors) and sonicated five times (10-s bursts) with cooling on ice for 1 min between bursts. DNase I (10 U/ml) and RNase A (1 U/ml) were added and incubated at 37°C for 1 h. The lysate was collected by centrifugation and precipitated with 5 volumes of ice-cold 10% trichloroacetic acid in acetone containing 0.2% dithiothreitol, and the pellet was washed twice with the same volume of ice-cold acetone. The precipitated protein pellets were solubilized in 1.5 ml of solubilization buffer (7 M urea, 2 M thiourea, 4% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}, 50 mM dithiothreitol, 0.002% bromophenol blue, 2% immobilized pH gradient [IPG] buffer, pH 3 to 10; Amersham Biosciences) at room temperature for 2 h with occasional mixing. The suspension was centrifuged, and the clear supernatant was stored in 100-μl aliquots at –70°C. Soluble and outer membrane proteins were separated according to the method of Murakami et al. (31).

    First dimension: isoelectric focusing. Isoelectric focusing with immobilized pH gradients (IPG strips) in the IPGphor Isoelectric Focusing System (Amersham Biosciences UK Ltd., Bucks, United Kingdom) was carried out according to the manufacturer's instructions using 7-cm or 13-cm IPG strips in the pH range 4 to 7 (linear) or 13-cm IPG strips in pH 3 to 5.6 (nonlinear).

    Second dimension: gel electrophoresis. IPG strips were incubated in 10 ml of equilibration buffer (18) with gentle shaking for 15 min to denature the proteins and reduce disulfide bridges, followed by alkylation of cysteine residues (18). SDS-PAGE was performed in either a SE260 mini-vertical system or a Hoefer SE600 standard vertical system without the use of stacking gel. Gels were stained with colloidal Coomassie brilliant blue according to the method of Neuhoff et al. (33).

    In-gel digestion and MALDI-TOF MS analysis. Protein spots were excised, destained, and digested with sequencing-grade trypsin (13 μg/ml) overnight (45). Peptide extracts were desalted and concentrated using C18 resin Zip-Tip. Peptides (0.5 μl) were mixed with an equal volume of saturated matrix solution (10 mg/ml -cyano-4-hydroxy cinnamic acid in 50% acetonitrile, 0.05% trifluoroacetic acid), applied to a matrix-assisted laser desorption ionization-time of flight mass spectrometry MALDI-TOF MS sample plate, and allowed to dry in air.

    Peptide mass fingerprinting was performed using MALDI-TOF MS. Identification of peptides was carried out using the RgpA sequence of P. gingivalis W50 (1). In some cases, quadrupole time of flight tandem mass spectrometry (Q-TOF MS-MS) was performed on selected high-intensity peptides to confirm the identity of the protein (MRC Clinical Sciences Centre, Imperial College Faculty of Medicine, London, United Kingdom).

    Measurement of stability of RgpAcat/D7. The stability of RgpAcat/D7 was measured as a function of pH in the presence or absence of 2-mercaptoethanol (10 mM) and in the presence or absence of CaCl 2 (10 mM) at 30°C. Buffers used were acetate (pH 4.5 to 6.0) and Tris-HCl (pH 7.3 to 8.3) buffers. Buffer (0.1 M) containing the desired additives was incubated in screw-cap Sarstedt tubes (Sarstedt Ltd., Leicester, United Kingdom) at 30°C in a water bath for a minimum of 1 h. A small volume of concentrated enzyme solution was added to the buffer to give a concentration of 12.5 to 15 μg ml–1. After thorough mixing, 2x 20-μl aliquots were withdrawn at various times and assayed immediately at pH 8.1, 30°C.

    Monosaccharide analysis of HRgpA/D7 and RgpAcat/D7. HRgpA/D7 (2 mg) and RgpAcat/D7 (0.2 mg) proteases were dialyzed against 5% (vol/vol) aqueous acetic acid to remove salts and detergents and freeze-dried. Monosaccharide analysis was essentially performed as described previously (10).

    Release of O-linked oligosaccharides by alkaline -elimination. RgpAcat (1.5 mg) and RgpAcat/D7 (3 mg) were dialyzed against 5% (vol/vol) aqueous acetic acid to remove salts and detergents and freeze-dried. -Elimination (28) and separation of released oligosaccharides was performed as described previously (10). Peak fractions were dried under a vacuum or by freeze-drying and the monosaccharide composition of oligosaccharides determined by methanolysis as described above.

    RESULTS

    P. gingivalis W50 produces five extracellular Arg-X-specific proteases, HRgpA, RgpAcat and mt-RgpAcat, encoded by rgpA, and RgpB and mt-RgpB, encoded by rgpB (Table 1). HRgpA is a dimer of the catalytic chain and a adhesin chain, whereas the other four enzymes are monomers containing the catalytic chain with different degrees of posttranslational modifications.

    Inactivation of rgpB also results in halving the Arg-X protease activity. In this mutant (2), HRgpA appeared to be unaffected whereas the monomeric enzyme mt-RgpAcat was not present, and RgpAcat showed altered chromatographic behavior compared to the wild-type enzyme. These results led us to suggest that the correct maturation/modification of the monomeric enzymes from rgpA was dependent on a functional rgpB. In order to confirm the specific involvement of rgpB in the maturation of rgpA-derived enzymes, we carried out targeted mutagenesis at the rgpB locus. Figure 1 shows the organization of genes around the rgpB locus in P. gingivalis W50. These are dipeptidyl peptidase (dppIV), lipoic acid synthetase (lipA), ClpXP protease (cicA), or hydroxybenzoate octaprenyl transferase (ubiA).

    Analysis of mutants in the rgpB locus. P. gingivalis W50, D7(rgpB), and mutants DE1 (dppIV), L7 (lipA), H7 (ubiA), and C7 (cicA) were grown in brain heart infusion supplemented with hemin as described in the Materials and Methods section. Whole cultures and supernatants were assayed for Arg-X and Lys-X enzyme activities after 24 h and 6 days. Mutants DE1, L7, H7, and C7 had enzyme activities comparable to those of the parent W50 strain (data not shown), whereas mutant D7 had approximately 50% of Arg-X activity and unaltered Lys-X activity in both whole culture and culture supernatant compared to the W50 strain. The 6-day culture supernatants of all six strains were treated with DNS-EGR-CK, the fluorescently labeled irreversible inhibitor of Arg-X activity. This method permits the separation and detection of mt-RgpAcat and mt-RgpB (a), Kgp (b), and HRgpA, RgpAcat, and RgpB (c) (Fig. 2).

    Gel electrophoresis on SDS-12.5% PAGE gels (Fig. 2) showed that, in accordance with the enzyme activity data, there were no discernible differences between the profiles obtained with strains W50, DE1, L7, H7, and C7. As shown previously, strain D7 lacked mt-RgpB and the mt-RgpAcat enzyme in the 70- to 80-kDa molecular mass range (2) and showed a slight reduction in the fluorescence intensity of band "c" due to the absence of RgpB.

    Two-dimensional electrophoresis of soluble cellular proteins. Two-dimensional gel electrophoresis of cellular proteins, soluble cytoplasmic proteins, and outer membrane preparations of P. gingivalis strains W50 and D7 were performed using Immobiline Dry IPG strips in the pH range 4 to 7 (linear) or 3 to 5.6 (nonlinear) essentially according to the instructions of the manufacturer (Amersham). Reproducible two-dimensional gel patterns were obtained in all cases (Fig. 3). Comparisons showed that the main difference in total cell proteins between W50 and D7 was the presence of mt-RgpAcat in W50, which was absent in D7 (Fig. 3A). The absence of mt-RgpAcat in D7 was also shown by fluorescent labeling of culture supernatants (Fig. 2).

    Two-dimensional electrophoresis of soluble cytoplasmic proteins of W50 and D7 in the pH range 3 to 5.6 showed several differences (Fig. 3B). The spot marked W373 corresponds to RgpB in W50 and, as expected, is absent in D7 as analyzed by peptide mass fingerprinting using MALDI-TOF and/or Q-TOF MS-MS. Another major difference between the two strains was the presence of enolase in W50 (spots W374 and W375), which was absent in D7. The intensity of spot W372 (phosphoribosylamine glycine ligase, PerD) was slightly higher in W50 than in D7 (D372). Spots D378 and D379 classified as conserved hypothetical protein and thiol protease tpr, respectively, were present in D7 but absent in W50. Figure 3C shows the two-dimensional electrophoresis of outer membrane proteins of W50 and D7 in the pH range 3 to 5.6. The significant difference between the two profiles is the shift of RgpAcat to a more basic pH in the case of D7 (boxed spots W414 and W415 versus boxed spots D417 and D418). W416 in the W50 two-dimensional gel map is a mixture of elongation factor Tu and the RagB protein. RagA protein (W413 and D413) and a proteolytic fragment of RagA (W413a and D413a) are marked for convenience.

    Purification of Arg-X proteases from P. gingivalis/D7 (rgpB). Arg-X activity was solubilized in aqueous buffer from the ammonium sulfate precipitation and contained both HRgpA and RgpAcat. Gel filtration chromatography of this fraction resulted in partial separation of HRgpA and RgpAcat. Selective purification of these two forms was achieved by affinity chromatography on Arg-agarose columns, to which only HRgpA bound under the conditions used. This behavior was identical to that of HRgpA from P. gingivalis W50 (41), which will henceforth be referred to as HRgpA/W50. Elution from the affinity column was performed first with 0.3 M L-Lys, which eluted predominantly Lys-X protease, followed by 0.1 M L-Arg, which eluted HRgpA activity. Further purification of HRgpA could be achieved by rechromatography on Arg-agarose columns after dialysis against affinity column equilibration buffer. Use of high-detergent (0.05%) buffers resulted in the solubilization of the bulk of the remaining Arg-X activity from the ammonium sulfate precipitate. HRgpA from this fraction was purified as described above but at 22°C and will henceforth be referred to as HRgpA/D7 or HRgpA(rgpB).

    Ion exchange chromatography of Arg-X activity not bound to the affinity column was performed at pH 7.5 in Tris-HCl buffer using DEAE-Sephacel columns (Fig. 4A). This yielded RgpAcat which appeared to contain trace amounts of Lys-X activity and other contaminants. However, hydrophobic chromatography of the DEAE-Sephacel fractions on phenyl-Sepharose columns yielded essentially pure RgpAcat of very high specific activity (Fig. 4B). The remaining Arg-X activity in ammonium sulfate precipitates was solubilized with buffer containing 0.05% detergent and was subjected to ion-exchange chromatography at 22°C, but the recovery and specific activity of RgpAcat were very low. This enzyme is referred to hereafter as RgpAcat/D7 or RgpAcat(rgpB). The following recovery of enzymes was obtained: 22.2% for HRgpA/D7 and 2.6% for RgpAcat/D7. The corresponding yields in the parent strain were 16% for HRgpA and 18% for RgpAcat. The specific activities of mutant enzymes were 6.1 U/mg for HRgpA/D7 and 20.6 U/mg for RgpAcat/D7, with DL-BRpNA as substrate at 30°C based on protein concentrations measured by absorbance at 280 nm and 260 nm. These values are comparable to the specific activities of the parent enzymes (41).

    The apparent subunit size of HRgpA/D7 and RgpAcat/D7 on a reducing SDS-PAGE gel (Fig. 5) was 53 ± 1 kDa. The molecular sizes obtained by gel filtration were 110 kDa for HRgpA/D7 and 55 kDa for RgpAcat/D7, similar to the enzymes from W50 (41). HRgpA/D7 could be dissociated into its component subunits (catalytic) and (adhesin) on 8 M urea-PAGE (2). N-terminal amino acid sequence analysis revealed that HRgpA/D7 and RgpAcat/D7 had N termini identical to the equivalent enzymes from P. gingivalis W50, as follows:

    in HRgpA/D7,

    chain, YTPVEEKQNGRMIVIVAKKY..., and

    chain, SGQAEIVLEAHDVWND...

    in RgpAcat/D7,

    YTPVEEKQNGRMIVIVAKKY... .

    Attempts were made to determine the C-terminal sequences of at least 4 residues using the PE-Applied Biosystems Procise-C C-terminal protein sequencing system at the School of Biochemistry and Molecular Biology, University of Leeds, Leeds, United Kingdom. These experiments were not successful.

    Physical properties of enzymes. Maximum enzyme activity for both HRgpA/D7 and RgpAcat/D7 proteases was obtained only in the presence of 10 mM L-cysteine and 0.02 to 10 mM CaCl2 at pH 7.9 to 8.1. No enzyme activity was detected in the absence of reducing agents. Ca2+ was not required for enzyme activity but improved enzyme stability, especially at higher temperatures and pHs (discussed later), as in the case of the parent enzymes.

    The Km for the chromogenic substrate L-BRpNA at pH 8.1 and 30°C for RgpAcat/D7 was 5.6 μM and compares favorably with the values obtained for RgpAcat/W50 (5 μM) and RgpB (4 μM) (40).

    The thermostability of RgpAcat/D7 was measured at 30°C under a variety of conditions, as shown in Table 2. First-order rate constants for the loss of enzyme activity at 30°C as a function of pH in the absence or presence of 10 mM 2-mercaptoethanol and in the absence or presence of 10 mM CaCl2 for RgpAcat/D7 showed that this enzyme is far less stable than wild-type RgpAcat, especially at higher pH values. The latter enzyme was stable at all pHs tested in the presence of 10 mM CaCl2, whereas RgpAcat/D7 lost activity at pH 8.3 even in the presence of 10 mM CaCl2 (half-life = 21.7 h). The (in)stability of RgpAcat/D7 at pH values >7.5 in the absence of 10 mM CaCl2 was even more dramatic, where the half-lives were between 40 and 80 times lower than the half-lives of the RgpAcat wild-type enzyme either in the absence or in the presence of 10 mM 2-mercaptoethanol (Table 2).

    Immunoreactivity. Figure 5 shows the results of Western blotting of rgpA-derived enzymes from P. gingivalis W50 and P. gingivalis W50/D7 developed using rabbit antiserum to recombinant RgpAcat as control, and monoclonal antibody 1B5, which recognizes a glycan addition to RgpAcat and also cross-reacts with a P. gingivalis polysaccharide. RgpAcat/D7 failed to react with MAb 1B5, suggesting that the glycan addition(s) present on RgpAcat was absent in this enzyme. However, polysaccharide prepared by proteinase K digestion of whole cells from both W50 and rgpB appeared identical on SDS-PAGE gels followed by silver staining (2) and both showed cross-reactivity with MAb 1B5 (data not shown), indicating that the glycan epitope is present on both polysaccharide molecules. Thus, P. gingivalis (D7) has the machinery necessary to synthesize the polysaccharide, which is indistinguishable from the polysaccharide from the parent W50 strain in its reactivity with MAb 1B5, but the glycan incorporating this epitope is not attached to the RgpAcat in this mutant.

    Monosaccharide analysis. The monosaccharide composition of proteases HRgpA/D7 and RgpAcat/D7 was determined after consecutive methanolysis, N-acetylation, and conversion of methyl glycosides to O-trimethylsilyl (O-TMS) ethers followed by gas chromatography-mass spectrometry (GC-MS). Table 3 shows the percentages and molar ratios of monosaccharides in the oligosaccharides of HRgpA, RgpAcat, HRgpA/D7, and RgpAcat/D7. These data are expressed both as a percentage and as an empirical formula of the monosaccharide constituents and help to predict the types of oligosaccharides present in the glycoprotein. Whereas HRgpA contained Rha, Man, Gal, Glc, GalNAc, and GlcNAc totaling 2.1% protein weight (10), HRgpA/D7 contained only small amounts of Man and Glc totaling <0.5% protein weight. RgpAcat/D7 contained Ara, Rha, Fuc, Gal, Glc, and GlcNAc totaling 14% protein weight. Although the total sugar content was similar to that present in RgpAcat (14.4%) (10), there were significant differences in composition (Table 3). GalNAc and Neu5(Ac), the most abundant sugars in RgpAcat/W50, were not present, and the level of GlcNAc was greatly reduced. Ara, Fuc, and Glc levels were significantly raised, whereas Rha and Gal were comparable in the two enzymes.

    O-linked oligosaccharides of RgpAcat/W50 and RgpAcat/D7. In order to characterize the oligosaccharides linked to RgpAcat and RgpAcat/D7, the chemical methods of -elimination (48) and hydrazinolysis (48) were used to release O-linked and both O- and N-linked oligosaccharides, respectively. In the case of hydrazinolysis, the recovery and yield of oligosaccharides was very poor. O-linked oligosaccharides of RgpAcat/W50 were released by -elimination and separated by high-pressure liquid chromatography (HPLC) using a porous graphitized column (Fig. 6A). Treatment with mild alkali/borohydride reduces the sugars involved in the O-glycosidic linkage with protein and releases oligosaccharides with the corresponding alditol at their reducing end. Six fractions labeled b, c, e, f, g, and h contained carbohydrate, of which fractions b, f, g, and h were most probably pure with respect to oligosaccharide since they contained only a single reduced monosaccharide, as shown in Table 4. Fractions c and e contained two oligosaccharides, each with GalNAc and Gal involved in O-glycosidic linkage to RgpAcat.

    Three of the pure oligosaccharides (b, f, and g) were attached to the protein through GalNAc, whereas Glc was the linking sugar in one oligosaccharide (h). There appeared to be at least eight O-linked oligosaccharides present in RgpAcat, with GalNAc being the linking sugar for five oligosaccharides (b, c, e, f, and g), Gal being the linking sugar in two cases (c and e), and Glc forming a linkage in one case (h).

    RgpAcat/D7 was also treated with alkaline borohydride to release O-linked oligosaccharides and analyzed as described for RgpAcat (Fig. 6B). Table 5, shows the monosaccharide composition of O-linked oligosaccharides obtained from RgpAcat/D7 which were purified by HPLC. Eight fractions labeled A through H contained carbohydrate, of which only B and C appeared to be mixtures. Fractions A, B, C, and E contained Ara as the linking sugar, whereas B, C, and D contained Gal in glycosidic linkage to Ser/Thr in RgpAcat/D7. C also contained an oligosaccharide linked via Fuc, whereas F, G, and H contained oligosaccharides with apparently increasing chain lengths linked via Glc. Although fraction B appeared to be a mixture, it contained only two reduced sugars, arabitol and galactitol, suggesting that only a single pentose and a single hexose residue were attached to two sites in RgpAcat/D7. Fraction A contained a disaccharide Ara-Ara in glycosidic linkage, whereas D, E, F, G, and H contained longer oligosaccharide chains. Hence, there were significant differences in the composition and nature of the linking sugars in the oligosaccharides present in RgpAcat/W50 and RgpAcat/D7.

    Two-dimensional electrophoresis of fluorescently labeled RgpAcat enzymes. Figure 7 shows the two-dimensional electrophoresis profiles obtained with fluorescently labeled RgpAcat and RgpAcat/D7. Whereas RgpAcat produces a train of spots with a pI range of 4.9 to 5.2 with four or five major charge-form species, RgpAcat/D7 shows a different spot pattern with a pI range of 5.25 to 5.7 with three or four major charge-form species. The train of spots could be due to different conformations of the proteases (53) or to the existence of multiple glycoforms. The difference in the pI range between RgpAcat and RgpAcat/D7 could also reflect the differences in their oligosaccharide profiles, the higher pI range of the latter being due to the lack of Neu5(Ac). This alteration in pI range of the purified enzymes is consistent with the observations seen in the two-dimensional electrophoresis of outer membrane proteins of the parent and mutant strains (Fig. 3C). However, both enzymes were determined to have identical molecular masses by SDS-PAGE.

    Peptide mass fingerprinting of RgpAcat and RgpAcat/D7. Pure RgpAcat and RgpAcat/D7 were subjected to two-dimensional gel electrophoresis, and the protein spots were excised from gels and digested with sequencing grade trypsin. Peptide mass fingerprints obtained by MALDI-TOF analysis were matched to the sequence of RgpA after accounting for carboxamido-methyl cysteine and the oxidation of methionine residues to methionine sulfoxide. Figure 8 shows the amino acid sequence of RgpA propeptide and -catalytic regions (1), and the numbers indicate the positions of residues in the polypeptide RgpA. The peptide mass fingerprints for both RgpAcat enzymes were almost identical except for Y247EG...DWK258, which is indicated by plus signs above the amino acid sequence, which was present only in D7. Peptides identified in both RgpAcat and RgpAcat/D7 are underlined. Very short peptides would not be detected. One striking observation is that both RgpAcat enzymes contained peptides from the propeptide region sequence T181LR... .PGR227. However, N-terminal amino acid sequence determination of both enzymes after SDS-PAGE and blotting onto polyvinylidene difluoride membranes shows the sequence Y228TPV... (see above and Rangarajan et al. [41]). This suggests that the propeptide may be loosely attached to RgpAcat enzymes and is probably dissociated from the mature enzymes upon boiling with SDS prior to SDS-PAGE. However, during two-dimensional PAGE, the IPG strips from the first dimension are incubated with SDS under gentler conditions prior to second-dimension PAGE, which is probably not strong enough to dissociate the propeptide from mature RgpAcat. The presence of the propeptide could also be due to the increased sensitivity of MS to trace components.

    Peptides in the regions from T355...K408, N421... .K492, H529... .K558, and T575...R719 were not detected for both enzymes. Although R360 and K367 are both susceptible to trypsin, there is a potential N-linked glycosylation site at N361ITT... (5) and a potential O-linked glycosylation site at PS383AD... (54) which, if glycosylated, would probably render the two cleavage sites resistant to tryptic attack. On the other hand, W368...K408 is a rather large peptide and may be difficult to detect due to poor ionization. The region N421...K492 contains a potential N-linked glycosylation site at N434YT... and is also a large peptide. In the region H529...K558, there are several K residues at 534, 551, 553, and 554, an R at residue 535 giving rise to short peptides which are of low molecular weight and would not be selected. In the C-terminal region T575...R719, there are trypsin-sensitive bonds at K580, K614, K648, K653, K678, K682, K 689, R698, and R704 which should in theory give rise to peptides which should be easily detected. However, there are at least three potential O glycosylation sites at PT579KMQV..., PS687TKT... ., and VS712DAP... . and seven potential N glycosylation sites at N590LT..., N597VS..., N623GT...,N629LT..., N635ES..., N669LT..., and N690AT... (54) which, if modified with oligosaccharides, would make some of the trypsin-sensitive sites resistant to hydrolysis by trypsin. Thus, apart from the region H529... .K558 which would generate very short peptides, a large part of the C-terminal region was not detected in both RgpAcat enzymes, suggesting that some of the potential trypsin-sensitive sites may have been modified with oligosaccharides, making them resistant to hydrolysis by trypsin. Since we have not been able to identify the C terminus of both W50 and D7 RgpAcat, another possibility is that both enzymes have their C termini at residue 574.

    DISCUSSION

    Now that there is a growing acceptance that bacteria have developed the necessary cellular machinery to perform glycosylation reactions, the location, mechanism, enzymology, and physiological consequences of these processes are coming under increasing scrutiny (51). We have previously shown that the monomeric RgpA proteases of P. gingivalis W50 are variably glycosylated (10). In these studies, we demonstrated that insertional inactivation of rgpB in P. gingivalis W50 influences the biochemical behavior of the monomeric RgpA protease isoforms (2), suggesting that RgpB may have a role in the maturation pathway of rgpA-derived enzymes. In the absence of a suitable autonomously replicating plasmid for genetic complementation studies in the Porphyromonads, in our earlier study, we were unable to exclude the possibility that polar effects on transcription at an adjacent locus could also have accounted for these findings. Hence, in the present investigation, we performed targeted insertional inactivation of the genes upstream and downstream of rgpB on the P. gingivalis chromosome in order to provide additional evidence for a specific involvement of RgpB in the maturation of rgpA-derived monomeric enzymes. In the present work, we used proteomic analysis of whole-cell proteins of wild-type W50 and mutant rgpB to confirm that the inactivation of rgpB leads to a complete loss of mt-RgpAcat, that this is a highly specific effect, and that there are no other major alterations to the cellular protein profile of this mutant.

    Four major open reading frames surround rgpB (Fig. 1). The gene immediately upstream encodes a 278-amino acid protein with 69% similarity over 275 amino acids with Rickettsia prowazekii lipoic acid synthase (LipA) (3) and also to LipA from a variety of organisms including bacteria, yeast, and plants. A dipeptidyl peptidase (DppIV) which shows extensive similarity to both prokaryotic and eukaryotic DppIV enzymes is encoded 73 bp upstream of lipA (24, 25). An orf 553 bp downstream of rgpB encodes a 200-amino acid protein which has 47% similarity to CicA from Caulobacter crescentus involved in ClpXP protease function (37) and is transcribed in the opposite direction to dppIV, lipA, and rgpB. Downstream of cicA and overlapping by 4 bp is ubiA, which encodes a 321-amino acid protein with strong similarity to 4-hydroxy benzoate octaprenyl transferases (4) of Mycobacterium tuberculosis (50% similarity over 281 amino acids). However, insertional inactivation of each of these flanking genes had no effect on either Arg-X enzyme activity or the appearance of fluorescently labeled extracellular proteases on SDS-PAGE compared to the parent strain. Moreover, a screen of eight independently isolated P. gingivalis rgpB mutants showed exactly the same phenotype as the original D7 mutant described previously: a 50% reduction in total Arg-X activity and loss of the extensively modified mt-RgpAcat isoform (not shown). Hence, it appears unlikely that the altered protease phenotype following the insertional inactivation of rgpB is a consequence of a polar effect on transcription at an adjacent locus and suggests a specific involvement of RgpB in the maturation of rgpA-derived enzymes.

    Analysis of purified protease revealed that one of the main differences between RgpAcat and RgpAcat/D7 was in their isoelectric points. The latter is far less acidic at pH 5.3, at which anion exchange chromatography is routinely performed during purification of RgpAcat and mt-RgpAcat. Consequently, RgpAcat/D7 was purified at pH 7.5 in order to accommodate the increase in its isoelectric point. Another important difference is that, unlike RgpAcat from the parent strain, the mutant enzyme also appears to have hydrophobic regions or patches which enable it to bind to phenyl-Sepharose in the presence of 2 M ammonium sulfate. Despite these differences, purified HRgpA and RgpAcat from D7 had almost identical specific activities and enzymatic properties compared to the enzymes from strain W50. RgpAcat/W50 and RgpAcat/D7 had identical N-terminal amino acid sequences and the same apparent molecular sizes as observed on gel filtration and on SDS-PAGE gels. In order to establish that both RgpAcat enzymes had similar C-terminal sequences, attempts were made to determine the C-terminal sequences of at least 4 residues using the PE-Applied Biosystems Procise-C C-terminal protein sequencing system. However neither protease yielded a sequence either due to the limitations of C-terminal sequencing or because of posttranslational modification. Hence, pure RgpAcat/W50 and RgpAcat/D7 were subjected to two-dimensional electrophoresis and peptide mass fingerprinting. Both enzymes showed differences in their isoelectric points but had identical mobilities in the second dimension (Fig. 7). Both proteins gave very similar peptide mass fingerprints suggesting that they had almost identical polypeptide chain lengths. However, peptides in the C-terminal region beyond residue 574 could not be detected for both RgpAcat enzymes. Thus, either both proteins had extensive posttranslational modifications in their C-terminal regions or both had residue 574 as their C terminus. The differences in isoelectric points of the two enzymes were also apparent in the two-dimensional electrophoresis gels (Fig. 3C) of outer membranes of 24-h cells, suggesting that charge differences are generated during the initial maturation of the enzyme prior to secretion into the supernatant.

    A striking difference in the enzymatic properties of RgpAcat and RgpAcat/D7 is the decrease in stability at 30°C at pHs >7.5 of the mutant enzyme. In the absence of 10 mM Ca2+, the half-lives of RgpAcat/D7 are between 40 and 80 times lower than those of RgpAcat at pHs 7.8 and 8.3 (Table 2). In the presence of 10 mM Ca2+, RgpAcat/D7 has a half-life of 21.7 h at pH 8.3 whereas RgpAcat is stable at all pHs tested under these conditions. Thus, aberrant glycosylation of RgpAcat appears to have a direct and dramatic influence on enzyme stability. The decreased stability does not appear to be caused by increased susceptibility to autolysis since the half-life of RgpAcat/D7 was unaffected in the presence of reducing agent, which is essential for enzyme activity. It is possible that the loss of certain types of oligosaccharide chains could cause minor changes in the local conformation of several regions of the RgpAcat/D7 molecule leading to instability. As indicated above, RgpAcat/D7 appeared to contain hydrophobic patches which enabled this protein to bind to phenyl-Sepharose columns, which could again be due to changes in local conformation. The prevalence of shorter oligosaccharide chains in this protease could result in differences in protein surfaces exposed to solvent compared to RgpAcat/W50.

    These properties were partially explained by the different sugar composition of RgpAcat/D7 compared to RgpAcat from strain W50. Monosaccharide analysis of RgpAcat/D7 showed that although it contained a similar amount of carbohydrate as the parent enzyme (14% versus 14.4% carbohydrate weight), the sugar composition was significantly different (Table 3). The most striking differences were that RgpAcat/D7 lacked Neu5(Ac) and GalNAc and the level of GlcNAc was greatly reduced whereas Ara, Fuc, and Glc were present in significantly larger amounts. The lack of Neu5(Ac) could account for the increase in the isoelectric point of RgpAcat/D7 relative to that of RgpAcat. Several O-linked oligosaccharides are present in RgpAcat. Two of these are linked through Glc, one through Gal, and at least five through GalNAc, analogous to mucin-type oligosaccharides (15). The absence of GalNAc in RgpAcat/D7 may indicate that the glycan chains normally linked through this sugar are absent. Furthermore, since Neu5(Ac) is often linked to nonreducing terminal GalNAc, the absence of this sugar may also explain the absence of sialylation of the mutant enzyme. Analysis of the O-linked oligosaccharides of RgpAcat/D7 showed dramatic differences compared to the parent enzyme. The oligosaccharides of RgpAcat (W50) were composed of between 7 and 35 monosaccharide residues (Table 4), whereas RgpAcat/D7 contained shorter chain oligosaccharides ranging from mono-, di-, and tetrasaccharides to a few longer (10- to 18-residue) chains. Oligosaccharide linkages were predominantly through Ara, Gal, Glc, and Fuc, unlike the parent strain enzyme, in which GalNAc and Gal were the main linking sugars. S-layer glycoproteins of Halobacterium halobium and Halobacterium salinarum have also been shown to contain oligosaccharides linked via Glc and Gal to Ser/Thr residues (29). In addition, the cellulase complexes of Bacteroides cellulosolvens and Clostridium thermocellum contain oligosaccharide moieties linked by Gal to Ser/Thr residues (16, 17). However, it is clear that inactivation of rgpB causes changes in both the composition and the linkage of oligosaccharides to RgpAcat/D7.

    Further confirmation that the glycan composition of the mutant enzyme is different from that of the parent came from immunochemical analysis with MAb 1B5. This monoclonal antibody was originally raised to RgpAcat but also recognizes a carbohydrate epitope(s) in mt-RgpAcat, mt-RgpB, and a polysaccharide of this organism (10). The polysaccharide preparations from P. gingivalis D7 retained immunoreactivity with MAb 1B5, indicating that while the epitope for this antibody is no longer present in RgpAcat/D7, the necessary machinery required to synthesize the epitope is still functioning in this mutant.

    Although it is well established that glycosylation of eukaryotic proteins is important for the maintenance of protein conformation and stability (38) and protection against proteolytic degradation (35), there are few examples of these functions in the bacterial glycoprotein literature. A -1,4-xylanase/exo--1,4-glucanase from Cellulomonas fimi was more resistant to proteolysis and had an increased affinity for microcrystalline cellulose (27, 36) when expressed in a glycosylated form. Finally, it has been suggested that glycosylation of a mycobacterial surface antigen acts to regulate the proteolysis of a linker region close to the N terminus of this molecule (20). Given the nature of the ecological niche occupied by P. gingivalis, it is possible that glycosylation of the RgpA proteases is necessary to ensure their stability and protection against proteolytic degradation in the hostile inflammatory conditions at periodontal sites undergoing destructive disease. There are other roles described for glycosylated bacterial proteins. Fischer and Haas (13) have shown that site-directed mutagenesis of a putative glycosylation site in Helicobacter pylori RecA results in the production of an unmodified RecA protein. This posttranslational modification is not involved in membrane targeting or cell division functions but is necessary for the full function of RecA in DNA repair. Karlyshev et al. (23) have described mutants of Campylobacter jejuni deficient in their ability to glycosylate a number of proteins, which reduces the ability of C. jejuni to adhere to and invade human epithelial cells and to colonize chicks.

    The precise role of rgpB in the correct glycosylation of the RgpA monomers is under continuing investigation. However, there are some interesting parallels involving the role of extracellular proteases of other pathogenic species in macromolecule biosynthesis. For example, it has recently been established that elastase, the product of lasB, in Pseudomonas aeruginosa plays an important role in the synthesis of alginate, the extracellular polysaccharide which is responsible for the mucoid phenotype of isolates of this organism in the cystic fibrosis lung (7, 22). In this case, cell-associated elastase is responsible for the proteolytic modification of the nucleoside diphosphate kinase (Ndk) of P. aeruginosa from a 16-kDa form to a 12-kDa moiety. Normally, Ndk is responsible for the synthesis of all nucleoside triphosphates or their deoxy derivatives. However, following elastase proteolysis, the truncated Ndk generates predominantly GTP, which is an important substrate in alginate biosynthesis. Thus, the inactivation of lasB in a mucoid strain leads to the abolition of alginate biosynthesis and the overexpression of elastase in a nonmucoid strain causes an increased synthesis of this polysaccharide. Similarly, it has recently been reported that inactivation of the extracellular cysteine protease, SpeB or streptopain, of Streptococcus pyogenes causes a decrease in the polysaccharide capsule expression in this organism although the molecular mechanism underlying this phenomenon has not been established (56). Hence, these two examples emphasize that proteolytic enzymes considered classical extracellular virulence factors may have additional cellular functions related to pathogenicity involving complex carbohydrate synthesis.

    In the case of the involvement of rgpB in the glycosylation of rgpA-derived enzymes, it is plausible that RgpB may be required for the activation of an enzyme(s) involved in the transfer of sugar residues to amino acid residues in proteins or in the extension of oligosaccharide structures. Loss of these activities in the rgpB mutant may disable this process. The immunoglobulin superfamily-like domain at the C terminus of RgpB with a presumptive role in protein recognition/binding (12) could conceivably be involved in such a targeted proteolytic action. Nakayama et al. (32) have shown that Rgps play an important role in the proteolytic maturation of precursors of the outer membrane and structural proteins on the surfaces of P. gingivalis cells, which supports an activation hypothesis. In the present work, two-dimensional electrophoresis of soluble cytoplasmic proteins of P. gingivalis strains W50 and rgpB in the pH range 3 to 5.6 showed several differences. The most striking was the absence of enolase and a slightly lower level of phosphoribosylamine glycine ligase in D7 and the absence of thiol protease tpr in W50. However, at present, it is uncertain whether these differences are related to the aberrant maturation of rgpA-derived enzymes in rgpB. An alternative hypothesis is that loss of the RgpB template for glycosylation could influence the nature of glycan additions to RgpA-derived enzymes. While the precise mechanism remains to be determined, the data in the present report emphasize that RgpB is an important accessory protein in the maturation pathway of the RgpA protease monomers leading to the enhanced stability of RgpAcat. Thus, the Rgps have multiple and overlapping functions in the pathogenicity of P. gingivalis.

    ACKNOWLEDGMENTS

    This research was funded by the Medical Research Council (United Kingdom) by grant no. PG9318173.

    REFERENCES

    1. Aduse-Opoku, J., J. Muir, J. M. Slaney, M. Rangarajan, and M. A. Curtis. 1995. Characterization, genetic analysis, and expression of a protease antigen (PrpRI) of Porphyromonas gingivalis W50. Infect. Immun. 63:4744-4754.

    2. Aduse-Opoku, J., M. Rangarajan, K. A. Young, and M. A. Curtis. 1998. Maturation of the arginine-specific proteases of Porphyromonas gingivalis W50 is dependent on a functional prR2 protease gene. Infect. Immun. 66:1594-1600.

    3. Andersson, J. O., and S. G. Andersson. 1997. Genomic rearrangements during evolution of the obligate intracellular parasite Rickettsia prowazekii as inferred from an analysis of 52015 bp nucleotide sequence. Microbiology 143:2783-2795.

    4. Andersson, S. G., A. Zomorodipour, J. O. Andersson, T. Sicheritz-Ponten, U. C. Alsmark, R. M. Podowski, A. K. Naslund, A. S. Eriksson, H. H. Winkler, and C. G. Kurland. 1998. The genome sequence of Rickettsia prowazekii and the origin of mitochondria. Nature 396:133-140.

    5. Bause, E., and H. Hettkamp. 1979. Primary structural requirements for N-glycosylation of peptides in rat liver. FEBS Lett. 108:341-344.

    6. Birkedal-Hansen, H., R. E. Taylor, J. J. Zambon, P. K. Barwa, and M. E. Neiders. 1988. Characterization of collagenolytic activity from strains of Bacteroides gingivalis. J. Periodontal Res. 23:258-264.

    7. Chakrabarty, A. M. 1998. Nucleoside diphosphate kinase: role in bacterial growth, virulence, cell signalling and polysaccharide synthesis. Mol. Microbiol. 28:875-882.

    8. Chen, T., and M. J. Duncan. 2004. Gingipain adhesin domains mediate Porphyromonas gingivalis adherence to epithelial cells. Microb. Pathog. 36:205-209.

    9. Curtis, M. A., H. K. Kuramitsu, M. Lantz, F. L. Macrina, K. Nakayama, J. Potempa, E. C. Reynolds, and J. Aduse-Opoku. 1999. Molecular genetics and nomenclature of proteases of Porphyromonas gingivalis. J. Periodontal Res. 34:464-472.

    10. Curtis, M. A., A. Thickett, J. M. Slaney, M. Rangarajan, J. Aduse-Opoku, P. Shepherd, N. Paramonov, and E. F. Hounsell. 1999. Variable carbohydrate modifications to the catalytic chains of the RgpA and RgpB proteases of Porphyromonas gingivalis W50. Infect. Immun. 67:3816-3823.

    11. Dawson, R., D. Elliot, W. Elliot, and K. Jones. 1986. Data for biochemical research. Clarendon Press, Oxford, United Kingdom.

    12. Eichinger, A., H. G. Beisel, U. Jacob, R. Huber, F. J. Medrano, A. Banbula, J. Potempa, J. Travis, and W. Bode. 1999. Crystal structure of gingipain R: an Arg-specific bacterial cysteine proteinase with a caspase-like fold. EMBO J. 18:5453-5462.

    13. Fischer, W., and R. Haas. 2004. The RecA protein of Helicobacter pylori requires a posttranslational modification for full activity. J. Bacteriol. 186:777-784.

    14. Fletcher, H. M., H. A. Schenkein, R. M. Morgan, K. A. Bailey, C. R. Berry, and F. L. Macrina. 1995. Virulence of a Porphyromonas gingivalis W83 mutant defective in the prtH gene. Infect. Immun. 63:1521-1528.

    15. Fukuda, M., and O. Hindsgaul. 1994. Molecular glycobiology: frontiers in molecular biology. Oxford University Press, Oxford, United Kingdom.

    16. Gerwig, G. J., J. P. Kamerling, J. F. Vliegenthart, E. Morag, R. Lamed, and E. A. Bayer. 1991. Primary structure of O-linked carbohydrate chains in the cellulosome of different Clostridium thermocellum strains. Eur. J. Biochem. 196:115-122.

    17. Gerwig, G. J., J. P. Kamerling, J. F. Vliegenthart, E. Morag, R. Lamed, and E. A. Bayer. 1992. Novel oligosaccharide constituents of the cellulase complex of Bacteroides cellulosolvens. Eur. J. Biochem. 205:799-808.

    18. Gorg, A., C. Obermaier, G. Boguth, A. Harder, B. Scheibe, R. Wildgruber, and W. Weiss. 2000. The current state of two-dimensional electrophoresis with immobilized pH gradients. Electrophoresis 21:1037-1053.

    19. Herrero, M., V. de Lorenzo, and K. N. Timmis. 1990. Transposon vectors containing non-antibiotic resistance selection markers for cloning and stable chromosomal insertion of foreign genes in gram-negative bacteria. J. Bacteriol. 172:6557-6567.

    20. Herrmann, J. L., P. O'Gaora, A. Gallagher, J. E. Thole, and D. B. Young. 1996. Bacterial glycoproteins: a link between glycosylation and proteolytic cleavage of a 19 kDa antigen from Mycobacterium tuberculosis. EMBO J. 15:3547-3554.

    21. Imamura, T., R. N. Pike, J. Potempa, and J. Travis. 1994. Pathogenesis of periodontitis: a major arginine-specific cysteine proteinase from Porphyromonas gingivalis induces vascular permeability enhancement through activation of the kallikrein/kinin pathway. J. Clin. Investig. 94:361-367.

    22. Kamath, S., V. Kapatral, and A. M. Chakrabarty. 1998. Cellular function of elastase in Pseudomonas aeruginosa: role in the cleavage of nucleoside diphosphate kinase and in alginate synthesis. Mol. Microbiol. 30:933-941.

    23. Karlyshev, A. V., P. Everest, D. Linton, S. Cawthraw, D. G. Newell, and B. W. Wren. 2004. The Campylobacter jejuni general glycosylation system is important for attachment to human epithelial cells and in the colonization of chicks. Microbiology 150:1957-1964.

    24. Kiyama, M., M. Hayakawa, T. Shiroza, S. Nakamura, A. Takeuchi, Y. Masamoto, and Y. Abiko. 1998. Sequence analysis of the Porphyromonas gingivalis dipeptidyl peptidase IV gene. Biochim. Biophys. Acta 1396:39-46.

    25. Kumagai, Y., K. Konishi, T. Gomi, H. Yagishita, A. Yajima, and M. Yoshikawa. 2000. Enzymatic properties of dipeptidyl aminopeptidase IV produced by the periodontal pathogen Porphyromonas gingivalis and its participation in virulence. Infect. Immun. 68:716-724.

    26. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685.

    27. Langsford, M. L., N. R. Gilkes, B. Singh, B. Moser, R. C. Miller, Jr., R. A. Warren, and D. G. Kilburn. 1987. Glycosylation of bacterial cellulases prevents proteolytic cleavage between functional domains. FEBS Lett. 225:163-167.

    28. Lantz, M. S., R. D. Allen, T. A. Vail, L. M. Switalski, and M. Hook. 1991. Specific cell components of Bacteroides gingivalis mediate binding and degradation of human fibrinogen. J. Bacteriol. 173:495-504.

    29. Lechner, J., and F. Wieland. 1989. Structure and biosynthesis of prokaryotic glycoproteins. Annu. Rev. Biochem. 58:173-194.

    30. Mikolajczyk, J., K. M. Boatright, H. R. Stennicke, T. Nazif, J. Potempa, M. Bogyo, and G. S. Salvesen. 2003. Sequential autolytic processing activates the zymogen of Arg-gingipain. J. Biol. Chem. 278:10458-10464.

    31. Murakami, Y., M. Imai, H. Nakamura, and F. Yoshimura. 2002. Separation of the outer membrane and identification of major outer membrane proteins from Porphyromonas gingivalis. Eur. J. Oral Sci. 110:157-162.

    32. Nakayama, K., F. Yoshimura, T. Kadowaki, and K. Yamamoto. 1996. Involvement of arginine-specific cysteine proteinase (Arg-gingipain) in fimbriation of Porphyromonas gingivalis. J. Bacteriol. 178:2818-2824.

    33. Neuhoff, V., N. Arold, D. Taube, and W. Ehrhardt. 1988. Improved staining of proteins in polyacrylamide gels including isoelectric focusing gels with clear background at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250. Electrophoresis 9:255-262.

    34. Nilsson, T., J. Carlsson, and G. Sundqvist. 1985. Inactivation of key factors of the plasma proteinase cascade systems by Bacteroides gingivalis. Infect. Immun. 50:467-471.

    35. Olden, K., B. A. Bernard, M. J. Humphries, T. K. Yeo, K. T. Yeo, S. L. White, S. A. Newton, H. C. Bauer, and J. B. Parent. 1985. Function of glycoprotein glycans. Trends Biochem. Sci. 10:78-82.

    36. Ong, E., D. G. Kilburn, R. C. Miller, Jr., and R. A. Warren. 1994. Streptomyces lividans glycosylates the linker region of a beta-1,4-glycanase from Cellulomonas fimi. J. Bacteriol. 176:999-1008.

    37. Osteras, M., A. Stotz, N. S. Schmid, and U. Jenal. 1999. Identification and transcriptional control of the genes encoding the Caulobacter crescentus ClpXP protease. J. Bacteriol. 181:3039-3050.

    38. Paulson, J. C. 1989. Glycoproteins: what are the sugar chains for Trends Biochem. Sci. 14:272-276.

    39. Pike, R., W. McGraw, J. Potempa, and J. Travis. 1994. Lysine- and arginine-specific proteinases from Porphyromonas gingivalis. Isolation, characterization, and evidence for the existence of complexes with hemagglutinins. J. Biol. Chem. 269:406-411.

    40. Rangarajan, M., J. Aduse-Opoku, J. M. Slaney, K. A. Young, and M. A. Curtis. 1997. The prpR1 and prR2 arginine-specific protease genes of Porphyromonas gingivalis W50 produce five biochemically distinct enzymes. Mol. Microbiol. 23:955-965.

    41. Rangarajan, M., S. J. Smith, S. U, and M. A. Curtis. 1997. Biochemical characterization of the arginine-specific proteases of Porphyromonas gingivalis W50 suggests a common precursor. Biochem. J. 323:701-709.

    42. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

    43. Schenkein, H. A., H. M. Fletcher, M. Bodnar, and F. L. Macrina. 1995. Increased opsonization of a prtH-defective mutant of Porphyromonas gingivalis W83 is caused by reduced degradation of complement-derived opsonins. J. Immunol. 154:5331-5337.

    44. Scott, C. F., E. J. Whitaker, B. F. Hammond, and R. W. Colman. 1993. Purification and characterization of a potent 70-kDa thiol lysyl-proteinase (Lys-gingivain) from Porphyromonas gingivalis that cleaves kininogens and fibrinogen. J. Biol. Chem. 268:7935-7942.

    45. Shevchenko, A., M. Wilm, O. Vorm, and M. Mann. 1996. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal. Chem. 68:850-858.

    46. Slaney, J. M., M. Rangarajan, J. Aduse-Opoku, S. Fawell, I. Darby, D. Kinane, and M. A. Curtis. 2002. Recognition of the carbohydrate modifications to the RgpA protease of Porphyromonas gingivalis by periodontal patient serum IgG. J. Periodontal Res. 37:215-222.

    47. Smalley, J. W., A. J. Birss, and C. A. Shuttleworth. 1988. The degradation of type I collagen and human plasma fibronectin by the trypsin-like enzyme and extracellular membrane vesicles of Bacteroides gingivalis W50. Arch. Oral Biol. 33:323-329.

    48. Spiro, R. G., and V. D. Bhoyroo. 1974. Structure of the O-glycosidically linked carbohydrate units of fetuin. J. Biol. Chem. 249:5704-5717.

    49. Sundqvist, G., J. Carlsson, B. Herrmann, and A. Tarnvik. 1985. Degradation of human immunoglobulins G and M and complement factors C3 and C5 by black-pigmented Bacteroides. J. Med. Microbiol. 19:85-94.

    50. Takii, R., T. Kadowaki, A. Baba, T. Tsukuba, and K. Yamamoto. 2005. A functional virulence complex composed of gingipains, adhesins, and lipopolysaccharide shows high affinity to host cells and matrix proteins and escapes recognition by host immune systems. Infect. Immun. 73:883-893.

    51. Tuomanen, E. I. 1996. Surprise Bacteria glycosylate proteins too. J. Clin. Investig. 98:2659-2660.

    52. Uitto, V. J., H. Larjava, J. Heino, and T. Sorsa. 1989. A protease of Bacteroides gingivalis degrades cell surface and matrix glycoproteins of cultured gingival fibroblasts and induces secretion of collagenase and plasminogen activator. Infect. Immun. 57:213-218.

    53. Veith, P. D., G. H. Talbo, N. Slakeski, and E. C. Reynolds. 2001. Identification of a novel heterodimeric outer membrane protein of Porphyromonas gingivalis by two-dimensional gel electrophoresis and peptide mass fingerprinting. Eur. J. Biochem. 268:4748-4757.

    54. Wilson, I. B., Y. Gavel, and G. von Heijne. 1991. Amino acid distributions around O-linked glycosylation sites. Biochem. J. 275:529-534.

    55. Wingrove, J. A., R. G. DiScipio, Z. Chen, J. Potempa, J. Travis, and T. E. Hugli. 1992. Activation of complement components C3 and C5 by a cysteine proteinase (gingipain-1) from Porphyromonas (Bacteroides) gingivalis. J. Biol. Chem. 267:18902-18907.

    56. Woischnik, M., B. A. Buttaro, and A. Podbielski. 2000. Inactivation of the cysteine protease SpeB affects hyaluronic acid capsule expression in group A streptococci. Microb. Pathog. 28:221-226.(Minnie Rangarajan, Ahmed )