1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Arg-gingipains (RgpsA and B) of Porphyromonas gingivalis are a family of extracellular cysteine proteases and are important virulence determinants of this periodontal bacterium. A monoclonal antibody, MAb1B5, which recognizes an epitope on glycosylated monomeric RgpAs also cross-reacts with a cell-surface polysaccharide of P. gingivalis W50 suggesting that the maturation pathway of the Arg-gingipains may be linked to the biosynthesis of a surface carbohydrate. We report the purification and structural characterization of the cross-reacting anionic polysaccharide (APS), which is distinct from both the lipopolysaccharide and serotype capsule polysaccharide of P. gingivalis W50. The structure of APS was determined by 1D and 2D NMR spectroscopy and methylation analysis, which showed it to be a phosphorylated branched mannan. The backbone is built up of α-1,6-linked mannose residues and the side-chains contain α-1,2-linked mannose oligosaccharides of different lengths (one to two sugar residues) attached to the backbone via 1,2-linkage. One of the side-chains in the repeating unit contains Manα1-2Manα1-phosphate linked via phosphorus to a backbone mannose at position 2. De-O-phosphorylation of APS abolished cross-reactivity suggesting that Manα1-2Manα1-phosphate fragment forms part of the epitope recognized by MAb1B5. This phosphorylated branched mannan represents a novel polysaccharide that is immunologically related to the post-translational additions of Arg-gingipains.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Porphyromonas gingivalis is a Gram-negative anaerobe and is an important aetiologic agent in adult periodontal disease, a chronic inflammatory condition of the periodontal tissues. The organism produces several virulence factors including extracellular cysteine proteases with specificities for Arg-X and Lys-X peptide bonds and lipopolysaccharide (LPS), which are believed to play important roles in the deregulation of innate and inflammatory systems in the host (Aduse-Opoku et al., 1995; Curtis et al., 2001). These are important antigens in periodontal patients and may account for a significant proportion of the immune response directed towards this organism (Ogawa et al., 1990; Slaney et al., 2002).

Porphyromonas gingivalis W50 produces five extracellular Arg-X specific cysteine proteases, Arg-gingipains, which comprise HRgpA (dimer of an α-catalytic chain and a β-adhesin chain), RgpAcat and mt-RgpAcat (both monomers of α-catalytic chain) encoded by rgpA, and RgpB and mt-RgpB (both monomers of α-catalytic chain) encoded by rgpB (Rangarajan et al., 1997). The four monomeric Arg-gingipains are post-translationally modified with glycan additions and contain between 15% and 30% by weight of carbohydrate (Gallagher et al., 2003). A monoclonal antibody, 1B5, raised to RgpAcat also cross-reacts with mt-RgpAcat, mt-RgpB and with a crude polysaccharide (PS) preparation of P. gingivalis W50 (Curtis et al., 1999). Chemical deglycosylation of RgpAcat and mt-RgpAcat with anhydrous trifluoromethane sulphonic acid abolishes their cross-reactivity with MAb1B5 indicating that this antibody recognizes a carbohydrate epitope that is also present in the PS preparation of P. gingivalis W50 (Curtis et al., 1999).

Only limited information is available on the nature of the PSs produced by P. gingivalis. Although the structure of the lipid A component of the LPS of P. gingivalis SU 63 has been solved, there is some suggestion that the same P. gingivalis strain may produce a population of lipid As with different levels of acylation depending on the growth conditions (Kumada et al., 1995). Three serotypes of the O-antigen PS have been reported (Schifferle and Beanan, 1999) and we recently solved the structure of one of these: the O-PS of P. gingivalis W50 [[RIGHTWARDS ARROW]6)-α-d-Glcp-(1[RIGHTWARDS ARROW]4)-α-l-Rhap-(1[RIGHTWARDS ARROW]3)-β-d-GalNAc-(1[RIGHTWARDS ARROW]3)-α-d-Galp-(1[RIGHTWARDS ARROW]], which is modified by ethanolamine phosphate at position 2 of Rha in a non-stoichiometric manner (Paramonov et al., 2001).

Laine et al. (1997) and Sims et al. (2001) showed the prevalence and distribution of six capsular serotypes (K antigens, K1–K6) of P. gingivalis in periodontitis patients and linked the presence of K antigen capsules with virulence in animal models. The antigens were thermostable, negatively charged, sensitive to periodate degradation and resistant to proteinase K treatment (Laine et al., 1996). We recently identified and characterized the K-antigen locus of P. gingivalis W50 (K1 serotype) and compared the sequence of this region, PG0106-PG0120 with the corresponding region in P. gingivalis 381 (K strain) (J. Aduse-Opoku et al., submitted). However, there is only limited structural information on the capsule PS of this organism.

In the current investigation, we aimed to establish the identity and structure of the PS from P. gingivalis W50, which is cross-reactive with MAb1B5 and shares an epitope with the glycosylated Arg-gingipains, as part of a larger investigation into the mechanism and pathway of post-translational modifications of this group of enzymes.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Immunochemical analysis of P. gingivalis phenol extracts and purified LPS

Western blot analysis of P. gingivalis W50 whole cells treated with Proteinase K using MAb1B5 revealed a ladder of cross-reacting bands in the molecular weight range 95 000 to 14 000 Da (Fig. 1A and B). However, LPS purified from P. gingivalis by the method of Darveau and Hancock (1983) showed no cross-reactivity with this antibody (Fig. 1C and D). We have previously shown that inactivation of porR in P. gingivalis leads to the complete loss of cross-reactivity in the mutant with MAb1B5 (Shoji et al., 2002) (Fig. 1B). In order to determine whether inactivation of porR leads to structural alteration to the O-antigen of LPS, we analysed the structure of O-antigen preparations from P. gingivalis W50 and porR-defective mutant strain by nuclear magnetic resonance (NMR)(Fig. 2). Comparison of the 1H-NMR spectra in the anomeric region revealed no differences indicating that the O-antigens are identical. Hence, the PS that is cross-reactive with MAb1B5 is distinct from the LPS of this organism.


Figure 1. SDS-PAGE and Western blotting of LPS and Proteinase K digests of cells of Porphyromonas gingivalis W50 and porR-defective mutant strains. Cells of P. gingivalis were grown for 48 h in BHI and cultures equivalent to an OD540 of 1.0 were centrifuged and washed twice with PBS, resuspended in 75 µl of SDS-sample buffer and heated at 100°C for 15 min followed by digestion with 50 µg of Proteinase K at 37°C for 16–18 h. The sample was heated at 100°C for 5 min and aliquots (10 µl) were used for SDS-PAGE followed by (A) silver staining and (B) Western blotting with MAb1B5. P. gingivalis W50 LPS purified by the Darveau and Hancock (1983) procedure was subjected to SDS-PAGE followed by (C) silver staining or (D) Western blotting with MAb1B5. The molecular masses of marker proteins (kDa) are given alongside the gels and blots.

Download figure to PowerPoint


Figure 2. 1H-NMR spectra of the anomeric regions of O-polysaccharide of LPS of (A) P. gingivalis W50 and (B) porR-defective mutant strains. O-PS purified from the LPS of P. gingivalis W50 and porR-defective mutant strain was analysed by NMR spectroscopy. Unit A refers to [RIGHTWARDS ARROW]4)-α-Rhap (with phosphoethanolamine attached at position 2); B refers to [RIGHTWARDS ARROW]6)-α-Glcp; C refers to [RIGHTWARDS ARROW]3)-α-Galp; D refers to [RIGHTWARDS ARROW]4)-α-Rhap and E refers to [RIGHTWARDS ARROW]3)-β-GalNAcp.

Download figure to PowerPoint

Characterization of P. gingivalis W50 (porR)

Analysis of porR-defective mutant strain suggested that the cells were more prone to lysis than the parent strain. Growth profiles of W50 and porR-defective mutant strains in BHI supplemented with haemin indicated that loss of porR function did not influence the growth maximum in this medium (Fig. 3A). However, in stationary phase there was a rapid decline in optical density at 540 nm of the porR-defective mutant culture, which fell to < 50% of levels of W50 after 6 days suggesting increased cell lysis in the mutant strain. Figure 3B shows electron micrographs of P. gingivalis W50 and porR-defective mutant cells from 48 h liquid cultures, which demonstrates a significant reduction in the electron-dense surface layer of the mutant cells.


Figure 3. Comparison of the growth and electron micrographs of P. gingivalis W50 and porR-defective strains. A. P. gingivalis W50 and porR mutant strain were grown in an anaerobic cabinet in BHI supplemented with haemin. Samples were withdrawn at different time points and the optical density was measured at 540 nm. B. Electron micrographs of P. gingivalis W50 and porR cells grown in liquid culture (48 h).

Download figure to PowerPoint

Analysis of K-capsular serotypes of P. gingivalis

In order to determine whether the cross-reacting polymer represents the capsular PS of P. gingivalis, we examined the reactivity of a range of K-capsular serotypes and strains devoid of capsule: P. gingivalis 381 (natural K strain) and a deletion mutant of P. gingivalis W50 (Δgpc) at the K-antigen locus (J. Aduse-Opoku et al., submitted). Δgpc is a deletion mutant of P. gingivalis W50(K1) in genes PG0110–PG0118 (Nelson et al., 2003) and encompasses three glycosyl transferases (PG0110, PG0111 and PG0118), two other transferases (PG0106 and PG0113), three conserved proteins (PG0113, PG0114, PG0116), an acetyl transferase (PG0113), a dehydrogenase (PG0108) and a flippase (PG0117). The Δgpc mutant strain no longer produces a capsular antigen based on staining with India ink and fuchsine and loss of cross-reactivity with antibody specific to K1 (J. Aduse-Opoku et al., submitted). P. gingivalis strains (capsular serotypes K, K1–K6) and mutant Δgpc were grown for 48 h and cells were treated with Proteinase K and subjected to SDS-PAGE and Western blotting with MAb1B5 (Fig. 4). All preparations, including those from the natural and mutant K strains showed an identical immunoreactive ladder of bands with MAb1B5. Hence the cross-reacting polymer is distinct from the serotype capsular PS of this organism.


Figure 4. SDS-PAGE of Proteinase K-treated P. gingivalis capsular serotype cells. Cells of P. gingivalis were grown for 48 h in BHI and cultures were treated as described in Fig. 1 for SDS-PAGE followed by Western blotting with MAb1B5. Lane 1, phenol extract of P. gingivalis W50; lane 2, serotype K; lane 3, K1; lane 4, K2; lane 5, K3; lane 6, K4; lane 7, K5; lane 8, K6; lane 9, capsule mutant Δgpc of P. gingivalis W50.

Download figure to PowerPoint

Isolation of anionic polysaccharide (APS) from P. gingivalis W50 cells

The results obtained above led us to purify the novel PS, distinct from LPS and capsular polysaccharide (CPS), which is immunoreactive with MAb1B5. Extraction of 14 g of freeze-dried P. gingivalis W50 cells with hot aqueous phenol (Westphal and Jann, 1965) followed by DNAse, RNAse A and proteinase treatment gave 690 mg of crude material (PS). Gel filtration chromatography of this material on a Sephacryl S-300 HR column in the presence of 1.5% (w/v) sodium deoxycholate gave the profile shown in Fig. 5. Fractions that eluted soon after the void volume of the column (high molecular weight PS) showed cross-reactivity with MAb1B5. These were combined and the PS recovered by precipitation with ethanol.


Figure 5. Gel-filtration chromatography of crude polysaccharides. A. 45% (v/v) phenol extract of P. gingivalis W50 polysaccharide mixture was applied to a column of Sephacryl S-300 HR (2.6 cm I.D. × 95 cm) in 50 mM Tris-HCl−1 mM EDTA−1.5% (w/v) sodium deoxycholate pH 9.5 buffer. Column effluent was monitored using a refractive index detector. Fractions (9 ml) were collected every 5 min. B. 12.5% SDS-PAGE/Western blot of column fractions from A. Fractions (10 µl) were used in the Western blotting experiment with MAb1B5. Molecular weights of markers are indicated alongside blot. Fraction numbers are indicated along the bottom. C refers to the crude phenol extract. Fractions that were cross-reactive were pooled and precipitated with four volumes of 95% (v/v) aqueous ethanol at −20°C.

Download figure to PowerPoint

The purified cross-reactive PS was subjected to anion-exchange chromatography on DEAE-Sephacel at pH 6.5, which yielded two components. Neutral polysaccharide (NPS, eluted during the sample loading and washing steps) showed no cross-reactivity with MAb1B5 and was not used for further analysis. Acidic material bound to DEAE-Sephacel was eluted with a gradient of NaCl from 0 to 2 M and showed a low level of absorbance at 280 nm. Fractions that showed cross-reactivity with MAb1B5 were combined and subjected to an additional purification step by gel-filtration chromatography on the Sephacryl S-300 HR column as described above. This purification procedure yielded 14 mg of APS, which was used in all the analyses performed in this study.

1D and 2D NMR analysis

Initial analysis of 1D 1H NMR spectrum of APS sample was complex because of broad signals that could be attributed to its high viscosity and/or the presence of oligosaccharides of different lengths in the repeating units of APS in a non-stoichiometric manner. The 1H, 13C (Table 1) and 31P chemical shifts, the linkage position and sequence of the glycosidic residues in the repeating units of APS were established on the basis of the sequence of 1H–1H, 1H–31P and 1H–13C correlation experiments.

Table 1.  1H NMR aand13C NMR data for APS and dephosphorylated APS (DPS) bfrom P. gingivalis W50 (δH,C in p.p.m.).
ResidueH-1/C-1H-2/C-2H-3/C-3H-4/C-4H-5/C-5H-6, H-6′/C-6
  • a

    . The coupling constants are not reported, but when measured, were in agreement with expected values.

  • b

    . Chemical shifts for DPS are given in brackets.

  • c

    . Position of substituted carbons is given in bold.

  • d

    . Assignments could be interchanged.

Unit A  5.46 4.05 3.86 3.78 3.84 3.91; 3.75
[RIGHTWARDS ARROW]2)-α-Man-P 96.71 79.47 c 74.7968.2872.9362.03
Unit B  5.30 (5.29) 4.13 (4.14) 3.96 3.70 3.75 3.91; 3.75
[RIGHTWARDS ARROW]2)-α-Man101.39(101.8) 79.98 b (78.82) 71.0567.073.5462.04 (61.5)
Unit C  5.18 4.10 3.86 3.69 3.76 3.91; 3.78
Unit D  5.15 4.09d 3.90d 3.67 3.80 3.91; 3.75
Unit E  5.12 (5.09) 4.03 (4.02) 3.92d 3.80 3.82d 4.02; 3.68
[RIGHTWARDS ARROW]2,6)-α-Man 99.35 (99.2) 80.02 (79.92) 70.8068.2672.0 66.21 d (65.5)
Unit F  5.09 (5.06) 4.04 (4.01) 3.92d 3.78 3.82d 4.02; 3.68
[RIGHTWARDS ARROW]2,6)-α-Man 99.32(99.2) 80.02 (79.92) 70.8068.2772.0 66.21 d (65.5)
Unit G  5.08 (4.89) 4.16 (3.98) 3.90 (3.82) 3.80 (3.82) 3.79 4.01; 3.67 (3.96; 3.76)
[RIGHTWARDS ARROW]2,6)-α-Man 99.57(100.6) 79.83 (71.12)71.5 (71.7)68.2 (68.5)72.42 66.42 (66.8)
Unit H  5.06 (5.05) 4.08d (4.06) 3.93 3.65 3.75 3.91; 3.75
t-α-Man103.05(103.2)71.05 (71.22)71.3167.2274.4562.04

Monosaccharide analysis of APS showed Man to be the only constituent with traces of Glc, Gal and GlcNAc probably because of the presence of a minor amount of CPS (Farquharson et al., 2000).

The residues in the repeating unit were designated A–H in order of decreasing chemical shifts of their anomeric protons. The 1H NMR spectrum of APS showed main signals for H-1 of α-linked mannopyranosyl residues at δH 5.46 (3JH,P 7.33 Hz), 5.30, 5.18, 5.15, 5.12, 5.09, 5.08 and 5.06. 2D 1H-{13C} heteronuclear multiple quantum coherence (HMQC) experiment allowed us to assign the corresponding anomeric carbons at δC 96.71, 101.39, 103.35, 103.60, 99.35, 99.32, 99.57 and 103.05 respectively. The resonances for secondary linkage carbons were positioned at 79.47, 79.98, 80.02 and 79.83 p.p.m and for primary linkage carbons (C-6) at 66.21 and 66.42 p.p.m. Broad signals for H-5 resonances were found in the region 72.0–74.48 p.p.m. The 31P NMR spectrum of APS (Fig. 6) showed a single resonance for phosphodiester at δ−1.267, which indicated the absence of phospho-monoester groups (Bretthauer et al., 1973; Slodki et al., 1973).


Figure 6. Partial contour plot of 1H-{31P} HMQC-TOCSY data for APS of P. gingivalis W50. The proton–phosphorous correlations marked H-1A, H-2A and H-2G shown on the track of the phosphorous nucleus resonance at −1.267 p.p.m. indicate the presence of Man units indicated in the text.

Download figure to PowerPoint

Methylation analysis of APS (Table 2) gave molar ratios of terminal and substituted units different from those expected based on NMR data. In particular, the amount of 2-substituted mannose and 2,6-di-substituted mannose were very low and their yield was ∼12% of expected values. This could be low because of the acid lability of 3,4-di-O-methyl mannose (Baretto-Bergter et al., 1980) or steric reasons.

Table 2.  Methylation analysis data.
ComponentsMolar ratios
APS (W50)DPS (W50)
  1. NF, not found.

Term – Mannopyranose12.85.8
[RIGHTWARDS ARROW]2) Mannnopyranose 0.50.7
[RIGHTWARDS ARROW]6) MannopyranoseNF0.5
[RIGHTWARDS ARROW]2,6) Mannopyranose 11

The downfield position of H-1 for unit A at 5.46 p.p.m. (Table 1) together with heteronuclear 2D 1H-{31P} heteronuclear multiple quantum coherence – total correlation spectroscopy (HMQC-TOCSY) data (Fig. 6), which showed correlation with the phosphorous resonance at −1.267 p.p.m. (specific for phosphodiester linkages), clearly indicated that this mannose residue is linked to the phosphate group at C-1 of unit A.

The position of H-2 and C-2 of unit A at 4.05 and 79.47 p.p.m., respectively (Table 1), were assigned from TOCSY data together with the assignment of cross-peaks in 1H-{13C} HMQC experiment.

As judged from the 1H-{13C} HMQC data, the chemical shift for C-2 of unit B at 79.98 p.p.m., C-2 of units E, F, and G at 80.02, 80.02 and 79.83 p.p.m., respectively, and the positions of C-6 of units E, F and G at 66.2, 66.2 and 66.4 p.p.m., respectively, were in accordance with the linkage sites of α-mannopyranosyl residues and established on the basis of 13C NMR reference data for α-mannosides (Bock and Thogersen, 1982). These results together with the methylation analysis data indicated that units E, F and G represent the α-(1[RIGHTWARDS ARROW]6) backbone of the repeating unit of APS. As judged from methylation analysis and 2D heteronuclear 1H-{13C} HMQC NMR data, the side-chains containing mono- and disaccharides are linked to the backbone by α-(1[RIGHTWARDS ARROW]2)-linkages.

1H–1H NOESY data suggested the sequence pattern of residues in APS (Table 3). The presence of intra-residue nuclear overhauser effect (NOE) contacts between the anomeric protons, H-1 resonances of which are located downfield in the region 4.90 p.p.m and H-2 of the same mannopyranosyl units showed the absence of β-mannopyranosyl residues (Vinogradov et al., 2000). Inter-residue type NOE contacts were observed under pre-irradiation of H-1 of residue B at 5.30 p.p.m. as follows: 4.04 and 5.12 p.p.m., which were assigned as H-2F and H-1E respectively. The sequence of cross-peaks on the tracks of H-1 of units E, F and G at 5.12/4.02, 5.09/4.01 and 5.08/4.02 p.p.m. was interpreted as H-1E/H-6F, H-1F/H-6G and H-1G/H-6E respectively. The correlation between corresponding interresidue NOEs H-1E/H-6′F, H-1F/H-6′G and H-1G/H-6′E are shown in Table 3. The above data together with the 13C NMR and methylation analysis results suggested that units E, F and G were α(1[RIGHTWARDS ARROW]6)-linked mannopyranosyl units.

Table 3.  NOEs observed under pre-irradiation of the anomeric protons of APS (δH,C in p.p.m.).
  • a

    . Assignment can be interchanged.

Unit B 5.305.124.134.04
Unit C 
t-α-Man H-2E H-2C 
Unit D 5.154.09 4.05 4.03
t-α-ManH-2D H-2A H-2G
Unit E 5.124.03a 4.02 3.68
Unit F 5.094.04a 4.01 3.67
Unit G 5.084.09 4.03 4.02; 3.67
[RIGHTWARDS ARROW]2,6) α-ManH-2D H-2E, H-2G H-6E, H-6′E
Unit H 5.064.13  

Because of partial overlapping of H-2 for units E and F at 4.03 and 4.04 p.p.m., the NOEs at 5.09/4.04 p.p.m. were assigned as H-1F/H-2E or H-1F/H-2F. However, the presence of NOE at 5.30/5.09 p.p.m. showed the close proximity of H-1 of units C and F, which indicated that unit C is attached to unit F at position 2. As judged from NOE at 5.06/4.13 p.p.m., the terminal α-mannosyl unit H is attached to unit B at position 2.

On pre-irradiation of H-1 of unit G at 5.08 p.p.m., the cross-peaks at 4.03 and 4.09 p.p.m. were assigned as H-1G/H-2G and H-1G/H-2D NOE contacts respectively. The inter-NOE contacts observed between H-1 of unit H at 5.06 p.p.m and H-2 of unit B at 4.13 p.p.m. unambiguously proved that unit H is linked to unit B at position 2. In addition, the NOE contact on the track of residue D at 5.15/4.05 p.p.m. was assigned as H-1D/H-2A. Therefore, the sequence D-(1[RIGHTWARDS ARROW]2)-A[RIGHTWARDS ARROW]P is also shown to be present as a side-chain in the repeating unit.

1H-{31P} HMQC-TOCSY data (Fig. 6) showed strong correlation at 4.03 and 4.16 p.p.m. on the line of the phosphorous nucleus. The former could be assigned either as H-2 of unit E or H-2 of unit F. From 1H–1H TOCSY experimental data, the latter was assigned as H-2 of unit G.

Thus, the repeating unit of APS is built up of a backbone of sequentially linked α-(1[RIGHTWARDS ARROW]6) mannopyranosyl units E, F and G with side-chains of oligo-mannopyranosyl residues of different lengths attached to each of them at position 2.

NOE suggested that unit C glycosylates unit F at position 2, unit D substitutes unit A at position 2 and unit H glycosylates unit B at position 2 respectively. The fragment D-(1[RIGHTWARDS ARROW]2)-A[RIGHTWARDS ARROW]P also substitutes unit G at position 2.

Analysis of de-O-phosphorylated APS (DPS)

De-O-phosphorylation of the APS (14 mg) with 48% aq. hydrofluoric acid (HF) (4°C, overnight) followed by lyophilization with a NaOH trap gave 9 mg of the de-O-phosphorylated polysaccharide (DPS), which was subjected to gel-filtration chromatography on Fractogel TSK HW-40(S) column. The high-molecular weight peak (DPS) eluting at the void volume of the column contained carbohydrate and was used in all analyses. The presence of DPS in the void volume indicated that de-O-phosphorylation did not destroy the polymeric structure of APS and suggested that the phosphate group is probably located in the side-chain(s) of APS rather than in the α-(1[RIGHTWARDS ARROW]6)-linked backbone.

De-O-phosphorylated polysaccharide was subjected to SDS-PAGE and Western blotting with MAb1B5 and the results obtained are shown in Fig. 7. DPS on SDS-PAGE shows the same pattern of bands as seen for untreated APS, which agrees with the gel-filtration results. The most striking observation is that all cross-reactivity to MAb1B5 has been lost on de-O-phosphorylation. This strongly implies that Manα1-phosphate [RIGHTWARDS ARROW]2)-Man motif in the repeating unit of APS is part of the epitope recognized by MAb1B5.


Figure 7. SDS-PAGE and Western blotting of APS, acid-treated APS and dephosphorylated APS (DPS). APS, acid-treated APS and DPS (10–20 µg) were subjected to SDS-PAGE as described in Methods and either subjected to silver staining (A) or blotted on to nitrocellulose membranes for Western blotting with MAb1B5 (B). 1 = untreated APS; 2 = APS heated with 10 mM HCl at 100°C for 30 min; 3 = DPS.

Download figure to PowerPoint

During the purification of APS it was observed that at acid pH (< pH 4) APS appeared to lose cross-reactivity to MAb1B5. The result of treatment of APS with 10 mM HCl (100°C, 30 min) followed by SDS-PAGE silver staining and Western blotting with MAb1B5 showed that acid-treated APS had retained its polymeric structure but lost its ability to cross-react with MAb1B5 (Fig. 7).

Methylation analysis of DPS showed the presence of 6-substituted mannose (Table 2) unlike APS, which did not contain this component. These data agreed with the preliminary structural data for the PS. In the case of DPS, the values for [RIGHTWARDS ARROW]2)-mannose and [RIGHTWARDS ARROW]2,6)-mannose were ∼16% of the theoretical values.

Analysis of 1H and 13C NMR data of DPS was simpler than that for the native acidic polymer and partly confirmed the structure of the repeating units of APS (Table 1). 1H and 13C NMR chemical shifts of residues, which did not contain phosphate-linked side-chains in APS, resembled those in DPS.

The 1H–13C HMQC data of DPS showed the disappearance of cross-peaks at 5.46/96.71 and 5.15/103.60 p.p.m., which corresponded to H-1/C-1 of units A and D respectively. There was a significant increase in intensity for a group of proton signals in the 1H NMR spectrum of DPS in the region 4.08–4.10 p.p.m. compared with the intensity for those in the 1H NMR spectrum of APS. These were assigned as resonances for H-2s of units C and H and confirmed the structural difference between the repeating units of the DPS and APS.

The high-field shift of H-1 and H-2 of unit G at 4.89 and 3.98 p.p.m. in DPS compared with those at 5.08 and 4.16 p.p.m. in APS, respectively, together with an up-field shift of its C-2 from 79.83 p.p.m. in APS to 71.12 p.p.m. in DPS were in accordance with the phosphorylation pattern of APS described above, i.e. that the D-(1[RIGHTWARDS ARROW]2)A[RIGHTWARDS ARROW]P motif is linked to unit G at position 2.

The resonances at 3.80 and 3.78 (Table 1) were assigned to H-4 of units E and F, respectively, in DPS. Reversing these assignments would imply that DPS is composed of two different block structures (two types of repeating units) of approximately equal size. One block would consist of repeating α(1,6)-linked units E-F-G in which side-chains H[RIGHTWARDS ARROW]B and C would be linked to units E and F, respectively, and the other block would consist of a backbone of α-1,6-linked units F-E-G with side-chains H[RIGHTWARDS ARROW]B and C attached to units F and E respectively.

Similarly, an interpretation of inter-residual NOE contacts observed for APS leads to the same conclusion for the structural features of the repeating units of APS as suggested above for DPS. Under pre-irradiation of H-1 of unit B and H-1 of unit C there are NOEs at 4.03 and 4.04, which were assigned (Table 1) as H-2 of unit E and H-2 of unit F respectively. The NOE contact observed between H-1 of B and H-1 of E finally allowed us to confirm that the structure of the repeating unit for APS is that shown in Fig. 8. The structure consists of repeating α-(1,6)-linked units E-F-G to which side-chains H[RIGHTWARDS ARROW]B and C are linked to units E and F respectively.


Figure 8. Proposed structure of APS of P. gingivalis W50. Proposed structure of APS. The letters adjacent to the Man residues represent Man units as indicated in the text.

Download figure to PowerPoint

In order to confirm the structure of APS deduced from NMR experiments, acetolysis studies were carried out. Acetolysis selectively cleaves the α1[RIGHTWARDS ARROW]6-bonds of the backbone of the PS and the released side-chain fragments were separated by high-performance liquid chromatography (HPLC) and analysed by MS (Kocourek and Ballou, 1969).

The electrospray – mass spectroscopy (ES-MS) analysis in negative mode of the monosccharide-containing low-molecular weight fraction isolated by HPLC on Glycopac N reversed phased column showed the presence of pseudo molecular ion peaks at m/z 259 (M-H) and m/z 97 (M-H) corresponding to mono negatively charged monosaccharide linked to a phosphate group and a phosphate group respectively (Fig. 9). This confirmed the presence of a mannose-phosphate-linked moiety among the products of acetolysis of APS. The elimination of a molecule of ketene (mass 42) and one of water (mass 18) leads to the fragmentation peak at m/z 199.1. Glucose-6-phosphate used as standard gave a pseudo molecular ion peak at m/z 259.0 (M-H) in the negative mode and a fragmentation peak at m/z 199.0.


Figure 9. Electrospray mass-spectrum in negative mode of low-molecular-weight fraction of products of acetolysis of APS. Acetolysis products of APS were fractionated by HPLC on Glycopac N column and the material corresponding to low-molecular-weight fractions were collected and used for ES-MS analysis. The peaks at m/z 97 and m/z 259.0 correspond to pseudo molecular ions of free phosphate group and the residue of mannose-2-phosphate respectively.

Download figure to PowerPoint

A representative matrix-assisted laser desorption/ionization – time of flight mass spectrometry (MALDI-TOF) spectrum of the oligosaccharide-containing fraction obtained by HPLC (Fig. 10) showed two major pseudo molecular ions [M+Na]+ at m/z 365.6 and m/z 527.5, which corresponded to a disaccharide and tri-saccharide respectively. The minor molecular ions [M+Na]+ at m/z 689.6 and m/z 851.5 could be attributed to the presence of an extra residue(s) of mannose and/or glucose residue (unit X) in the side-chain probably occurring once every six to 10 repeating units attached to residue H or unit X could be linked to unit D. Such an interpretation of the above data may correspond to the structure of the biological repeating unit of APS or may represent a heterogeneity in the population of PS repeating units as trace amounts of Glu, Gal and GlcNAc were detected during methanolysis.


Figure 10. MALDI-TOF spectrum in positive mode of purified oligosaccharide fraction obtained by acetolysis of APS. Acetolysis products of APS were fractionated by HPLC as described in Fig. 9 and the fractions containing oligosaccharides were used for MALDI-TOF analysis. The pseudo molecular ions [M+Na]+ at m/z 365.6 and m/z 527.5 correspond to di- and tri-saccharide respectively. Minor molecular ions at m/z 689.6 and m/z 851.5 corresponding to tetra- and penta-saccharide, respectively, are also present.

Download figure to PowerPoint

On the basis of the above data, the repeating units of APS is built up of branched α-d-mannan the backbone of which consists of α-1,6-linked mannose residues. The side-chains contain a number of α-1,2-linked mannose-containing oligosaccharides of different lengths (from one to two sugar residues), which are attached to the backbone via α-1,2-linkages. The unique feature of this structure is that one of the side-chains is Manα1[RIGHTWARDS ARROW]2Manα1-phosphate (P), which is directly attached to one of the backbone mannoses at position 2. These results indicate that APS is distinct from the capsular PS and LPS O-antigen of P. gingivalis and represents a novel PS produced by P. gingivalis W50.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

These investigations began with the observation that a monoclonal antibody (MAb 1B5) raised to the cysteine protease Arg-gingipain RgpAcat does not cross-react with the recombinant RgpA α-catalytic chain heterologously expressed in Escherichia coli but was immunoreactive with a crude PS preparation of P. gingivalis W50 and also recognizes a carbohydrate epitope on the membrane type Arg-gingipains mt-RgpAcat and mt-RgpB (Curtis et al., 1999). Along with other evidence, this led us to suggest that some members of the Arg-gingipain family of enzyme isoforms are post-translationally modified with glycan chains, which are immunologically and structurally related to a cell surface polymer of this Gram-negative bacterium. Monosaccharide analysis of RgpAcat indicated the presence of Ara, Rha, Fuc, Man, Gal, Glc, GalN(Ac), GlcN(Ac) and Neu5(Ac) totalling 14–15% by weight of protein (Rangarajan et al., 2005). Oligosaccharides attached to Ser/Thr residues in RgpAcat were released by alkaline borohydride treatment and separated by HPLC (Rangarajan et al., 2005). However, because of their heterogeneity and low yields obtained, it was not possible to characterize them individually. In the present work, we therefore concentrated on the characterization of cross-reacting PS from whole cells of P. gingivalis W50.

Glycosylation of bacterial surface proteins with glycans related to cellular PSs is not unprecedented. For example, the pilin of Pseudomonas aeruginosa 1244 is glycosylated with a single trisaccharide-containing moiety, α-5N-(β-hydroxybutyryl)-7-N-formyl-pseudaminic acid-(α2[RIGHTWARDS ARROW]4)-Xyl-(β1[RIGHTWARDS ARROW]3)-FucNAc-(β1[RIGHTWARDS ARROW]3)-β-Ser in the polypeptide chain (Castric et al., 2001). This structure has the same sugar composition and sequence as the O-antigen repeating unit of P. aeruginosa 170046, a strain which also belongs to LPS O5 serotype (Knirel et al., 1987). The pilin glycan differs from the O-antigen only in that the FucNAc residue is not O-acetylated. Mutations in the O-antigen biosynthetic gene cluster of P. aeruginosa 1244 lead not only to the absence of O-antigen but also to the production of non-glycosylated pili. This suggests that the pilin glycan of P. aeruginosa 1244 is a product of the O-antigen biosynthetic pathway.

In the current work, we established that the polymer cross-reactive with MAb1B5 was distinct from LPS. LPS purified from P. gingivalis W50 using the Darveau and Hancock (1983) procedure does not cross-react with MAb1B5. Furthermore, the O-antigen of a mutant of P. gingivalis W50 in porR has the same sugar composition and sequence as the O-antigen of the wild-type strain (Paramonov et al., 2001). PorR is a putative transaminase and is orthologous to RfbE, which belongs to the DegT clusters of orthologous groups (COGs). DegT of Bacillus stearothermophilus (Takagi et al., 1990) is involved in a range of biochemical functions including glycan synthesis, regulation of protease activity and pilin, flagellin and protein glycosylation. In Vibrio cholerae O1 and E. coli 0157, rfbE encodes the enzyme perosamine synthetase, which catalyses the biosynthesis of 4-amino-4,6-di-deoxy mannose (perosamine) (Bilge et al., 1996; Albermann and Piepersberg, 2001) and the rfbE orthologue in Caulobacter crescentus (per) (Awram and Smit, 2001) encodes a perosamine synthetase. Perosamine is found as a major constituent of LPS in Brucella spp. (Wu and Mackenzie, 1987), E. coli 0157 (Bilge et al., 1996), V. cholerae O1 (Villeneuve et al., 2000; Albermann and Piepersberg, 2001) and C. crescentus (Awram and Smit, 2001). As neither P. gingivalis W50 LPS nor LPS from porR-defective mutant strain cross-reacts with MAb1B5, we concluded that O-antigen and core region of LPS from the P. gingivalis W50 and porR-defective strains do not contain the epitope recognized by MAb1B5.

Examination of cells of P. gingivalis serotypes K1–K6 by electron microscopy revealed an extracellular capsule-like structure that varied among different P. gingivalis strains (Laine et al., 1996). For example, strains HG1690(K5), HG1691(K6) as well as HG184(K2) and A7A1-28(K3) appeared to possess a thick layer of extra-cellular material. However, strains W83 (K1) and ATCC 49417 (K4) contained smaller amounts of the extra-cellular layer. P. gingivalis 381(K) revealed a thin extra-cellular layer under electron microscopy, did not show encapsulation by phase-contrast microscopy and did not contain K-antigen when serologically tested (van Winkelhoff et al., 1993). In the present study, cells of the porR-defective mutant were also shown to have a reduced extra-cellular surface layer by electron microscopy suggesting that the cross-reactive polymer may play a structural role. Farquharson et al. (2000) described the isolation and partial characterization of a CPS with gel-like viscoelastic properties from P. gingivalis ATCC 53978. Although the structure of the CPS was not determined because of poor solubility and the presence of significant impurities of non-carbohydrate origin, the monosaccharide composition showed that it contained ManA, GlcA, GalA, Gal and GlcNAc in relative molar proportions of 0.60:0.90:0.44:0.51:1.00 respectively (Farquharson et al., 2000). SDS-PAGE of proteinase K-treated cell extracts of P. gingivalis capsular type strains K, K1–K6 and mutant Δgpc of P. gingivalis W50 followed by Western blotting showed that the PS reactive with MAb 1B5 was present in all capsular serotypes and capsule negative strains (Fig. 4). Consequently, the polymer cross-reactive with MAb1B5 is distinct from both LPS and the CPS of this organism.

We therefore purified and structurally characterized this molecule and showed it to be an APS comprising a phosphorylated branched α-mannan. 1D and 2D NMR spectroscopy together with methylation analysis of both native and de-O-phosphorylated polysacccharide showed that the repeating unit of APS was a branched α-d-mannan composed of α(1[RIGHTWARDS ARROW]6)-linked mannose residues whereas the side-chains contained a number of α-1,2-linked mannose-containing oligosaccharides of one to two Man residues attached to the backbone via α-1,2-linkage (Fig. 8). 1D and 2D homo- and heteronuclear NMR spectroscopy showed the presence of a phosphate residue linked via a diester bond forming Manα1-2Manα1-P- in the repeating unit of APS. De-O-phosphorylation of APS abolished its cross-reactivity with MAb1B5 strongly indicating that the presence of the oligosaccharide fragment Manα1-2Manα1-P was necessary for the recognition of APS by MAb1B5 (Fig. 7) and may therefore be part of the epitope.

Analysis of the P. gingivalis genome reveals the presence of all the genes required for the synthesis of GDP-Man from Fructose-6-phosphate in this organism. These include Mannose-6-phosphate isomerase (PG0486), Phosphomannomutase (PG1094, PG2010) and Mannose-1-phosphate guanyltransferase (PG2215). However, the polymerase/transferase necessary for the synthesis of α-(1,6) backbone and α-(1,2) side-chains are not readily identifiable.

Branched mannans are unusual in bacteria but have been described in the Gram-negative bacterium Pseudomonas syringae pv. ciccaronei, the pathogenic agent responsible for the leaf spots of carob plants. The exopolysaccharide purified from P. syringae is a branched mannan similar in structure to the APS synthesized by P. gingivalis. It consists of a backbone of α-(1[RIGHTWARDS ARROW]6)-linked mannopyranose units with 80% of them substituted at position 2 by mono-, di- and tri-saccharide side-chains. In addition, in some branches an extra terminal glucose residue is present and a phosphate group is attached to the terminal non-reducing mannose residues at position 6 (Corsaro et al., 2001).

Cell wall phosphorylated mannans of yeast Saccharomyces cerevisiae (Hernandez et al., 1989) are acid labile and this was also the case with P. gingivalis APS (Fig. 7). APS treated with mild acid (10 mM HCl, 100°C, 30 min) lost all cross-reactivity with MAb1B5 adding credence to the idea that it is part of the epitope recognized by MAb1B5. In yeast, mannoproteins located in the outermost layer of the cell wall determine the porosity of the cell wall and thereby regulate leakage of proteins from the cell and entrance of macromolecules from the environment (Jigami and Odani, 1999). Interestingly, analysis of the phenotype of the porR-defective mutant of P. gingivalis, which does not synthesize the polymer cross-reactive with MAb1B5, suggests that the branched phosphorylated mannan of P. gingivalis may play a similar role. In the porR-defective strain, the extracellular surface layer is of reduced thickness compared with the parent strain as seen by electron microscopy and the cells appear more fragile based on the rate of decrease of the culture optical density in stationary phase. As mannosyl-phosphate residues confer a net negative charge on the cell wall, it has the potential to change the properties and environment of the cell surface. In addition, transfer of Man-P from GDP-Man by the action of a mannosyl-phosphate transferase will result in the generation of GMP, which could inhibit the synthesis of charged sugar nucleotides involved in PS synthesis.

Vanterpool et al. (2005) have described an isogenic mutant of P. gingivalis W83, vimF, a putative glycosyltransferase gene, which fails to cross-react with MAb1B5 and hence may also be defective in APS biosynthesis. As with the porR-defective mutant, the vimF-defective strain has lower Arg- and Lys-gingipain activities and reduced haemolysis and haemagglutination compared with the parent W83 strain. However, there is no information relating to the cell-wall structure nor susceptibility to cell lysis.

It is well recognized that mannan PSs have the potential to exert multiple biological effects in host–pathogen interactions. For example, the ability of Candida albicans to establish an infection involves multiple components of this human fungal pathogen, but its ability to persist in host tissue may involve primarily the immunosuppressive property of a major cell-wall mannan. In this instance, mannan and oligosaccharide fragments of mannan are potent inhibitors of cell-mediated immunity (Nelson et al., 1991). The branched mannan produced by P. syringae has a phytotoxic effect on tobacco plants (Corsaro et al., 2001). The contribution of APS to the overall pathogenicity of P. gingivalis in periodontal disease is therefore worthy of further investigation particularly in relation to the potential immunomodulatory effects of mannans described in other systems, the interplay of this polymer with the pattern recognition systems of the innate response and the biological consequences of modification of the Arg-gingipains with APS derived epitopes.

Slaney et al. (2002) conducted a study to determine whether the post-translational additions to RgpAcat were significant targets of the immune response of periodontal patients. The data showed that a significant proportion of the serum IgG antibody response to RgpAcat is directed towards the glycans present on RgpAcat. Chemical deglycosylation of RgpAcat not only abolishes immune recognition by MAb1B5 but also by patient serum IgG. Hence, it is possible that glycosylation of Rgps with determinants derived from APS may provide a novel means of immune evasion as has been suggested for the glycosylation of surface proteins of Campylobacter (Stintzi et al., 2005).

Alternatively, glycosylation may be important for enzyme stabilization. Analysis of Rgps in the porR-defective mutant of P. gingivalis demonstrates that the protease isoforms synthesized by this mutant are those that do not acquire the MAb1B5 reactive glycan during their normal maturation in the parent strain, namely HRgpA and RgpB (A. Gallagher, J. Aduse-Opoku, M. Rangarajan, J.M. Slaney and M.A. Curtis, unpubl. results). The remaining isoforms (the mt-enzymes and RgpAcat), all of which routinely cross-react with MAb1B5, are not produced in this mutant. Therefore, the inability to synthesize the MAb1B5 glycan epitope in porR-defective mutant may lead to instability and subsequent loss of these isoforms in this strain. Previous work has shown that aberrant glycosylation of RgpAcat in an rgpB mutant of P. gingivalis also leads to enzyme instability (Rangarajan et al., 2005). In this case, inactivation of rgpB leads to synthesis of an RgpAcat, which does not cross-react with MAb1B5 although it still contains ∼14% carbohydrate by weight. The enzyme from mutant strain is less stable than the parent strain enzyme particularly at pHs > 7.5. For example, at pH 8.3 in the absence of Ca++, RgpAcat from the mutant strain has a half-life of ∼8 min compared with ∼7 h for the parent W50 enzyme (Rangarajan et al., 2005).

As described previously, monosaccharide analysis of RgpAcat from P. gingivalis W50 showed the presence of not only Man but also several other sugars. These are contained in at least eight oligosaccharides linked to Ser/Thr residues in RgpAcat via Gal, GalNAc and Glc (Rangarajan et al., 2005) resulting in a complex glycosylation pattern in this protease. Incorporation of the Manα1-2Manα1-P epitope into the glycans of the Arg-gingipains is likely to form just one aspect of the glycosylation process in this organism.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References


Sephacryl S-300HR, Sephadex G-50 Superfine and DEAE-Sephacel were purchased from Amersham Pharmacia Biotech, UK. All chemicals were obtained from Merck, Poole, Dorset, UK, or from Sigma-Aldrich Company, Poole, Dorset, UK and were the purest grades available. 30% Acrylamide/Bis solution, 37.5:1 was from Bio-Rad Laboratories (Hercules, CA, USA). HRP-labelled mouse immunoglobulin was purchased from Dako A/S, High Wycombe, Bucks., UK.


Bacterial growth conditions.  Porphyromonas gingivalis W50 was grown either on blood agar plates containing 5% defibrinated horse blood or brain heart infusion broth supplemented with haemin (5 mg ml−1) in an anaerobic atmosphere of 80% N2, 10% H2 and 10% CO2 (Aduse-Opoku et al., 1995) for 24 h. The cells were harvested, washed once with water and freeze-dried.

The seven capsular serotypes of P. gingivalis[K (HG91), K1(HG66), K2(HG1644), K3(HG1701), K4(HG1660), K5(HG1690) and K6(HG1705)] were obtained from Dr M.L. Laine (Laine and van Winkelhoff, 1998). Δgpc[deletion mutant of strain  P. gingivalis W50 in genes PG0109–PG0118 (TIGR:; J. Aduse-Opoku et al., submitted] and porR (PG1138, inactivated by insertion of erm cassette by allelic exchange; A. Gallagher, unpublished) were constructed in this laboratory.

Electron microscopy.  Cells of P. gingivalis W50 and porR-defective mutant strain grown in liquid culture (48 h) were first fixed in 0.5% glutaraldehyde in 0.1 M sodium cacodylate buffer pH 7.4 followed by fixing in 1% osmium tetroxide and embedded in Taab Embedding Resin (Taab, Aldermaston, UK). Sections were collected onto 400 mesh copper grids (Agar Scientific, Stansted, UK) and stained successively with a saturated solution of uranyl acetate and Reynold's lead citrate. Electron micrographs were taken using a Jeol 1200 EX11 transmission electron microscope operating at 80 kV.

Isolation of PS.  Polysaccharide was extracted from freeze-dried P. gingivalis W50 cells (14 g) using the modified phenol-water procedure (Westphal and Jann, 1965) followed by exhaustive dialysis against distilled water. The water-insoluble material was removed by centrifugation at 10 000 g and the supernatant was freeze-dried.

The freeze-dried residue (1.62 g) was dissolved in 75 ml of 50 mM Tris-HCl−10 mM MgCl2 pH 7.5 and incubated with DNAse I (32 µg ml−1) and RNAse A (40 µg ml−1) at 37°C for 4 h, followed by digestion with proteinase (70 µg ml−1) overnight at 37°C. The reaction mixture was dialysed against distilled water and freeze-dried.

Gel-filtration chromatography (GFC).  The freeze-dried crude PS (690 mg) was dissolved in 50 ml of 50 mM Tris-HCl−1 mM EDTA−0.15 M NaCl−1.5% sodium deoxycholate (pH 9.5) and purified by precipitation with four volumes of 95% (by vol.) ice-cold ethanol at −20°C overnight. The crude PS was recovered by centrifugation followed by dialysis against distilled water and freeze-dried (650 mg). It was dissolved in 25 ml of 50 mM Tris-HCl−1 mM EDTA−1.5% sodium deoxycholate pH 9.5 and applied to a column of Sephacryl S-300 HR column (2.6 cm I.D. × 95 cm) equilibrated in the same buffer, at 22°C. The column was eluted with buffer at 80 ml h−1 and fractions (9 ml) were collected. The refractive index of the column was monitored [Knauer Wellchrom K-2400 RI detector (Wissenschaftliche Geratebau, Dr Ing. Herbert KNAUER GmbH, Hegauer Weg 38, 14163 Berlin, Germany)] and A280 of the column fractions was measured. Aliquots (10 µl) of column fractions were subjected to SDS-PAGE and Western blotting with MAb1B5 (see below).

Fractions that showed immunoreactivity were combined and the PS was precipitated with ethanol at −20°C as described above. The suspension was centrifuged at 10 000 g for 60 min at 4°C. The pellet was resuspended in distilled water, dialysed against distilled water at 4°C and freeze-dried.

Ion-exchange chromatography.  The freeze-dried residue (275 mg) was dissolved in 55 ml of 50 mM Tris-HCl (pH 6.5) and applied to a DEAE-Sephacel column (2.6 cm I.D. × 8 cm) equilibrated in the same buffer at 22°C. The refractive index (RI) of column effluent was monitored and the column washed with equilibrating buffer till there was no change in RI values in the column effluent. Any NPS present will not bind to the DEAE-Sephacel column and will be present in the column wash. Negatively charged PS (Anionic PS) was eluted with a gradient of 0–2 M NaCl in buffer (total vol. 300 ml) at a flow rate of 60 ml h−1. A280 of column fractions (3 ml) was measured and aliquots (10 µl) were subjected to SDS-PAGE and Western blotting with MAb1B5. NPS was not cross-reactive with MAb1B5 whereas fractions eluted with NaCl showed strong immunoreactivity with MAb1B5.

Final purification of APS.  Fractions that showed cross-reactivity with MAb1B5 were pooled, dialysed against distilled water and freeze-dried. The residue of crude APS was subjected to a final GFC step on Sephacryl S-300HR as described above. This final purification yielded 14 mg of purified APS that was cross-reactive with MAb1B5 and used for further analysis.

SDS-PAGE and Western blotting.  SDS-PAGE was performed according to Laemmli (1970) using 12.5% (w/v) acrylamide gels. Samples of PS (5–10 µl each) were treated with 10 µl of SDS-sample buffer at 100°C for 3 min. P. gingivalis K serotypes (K-, K1–K6) and Δgpc mutant strain were grown for 48 h and cultures equivalent to an O.D.540 of 1.0 were centrifuged at 13 000 g and washed with 2 × 1 ml of PBS. Cell pellets were resuspended in 75 µl of SDS sample buffer and heated at 100°C for 15 min. After cooling to room temperature, 50 µg of ProteinaseK was added and incubated at 50°C for 16–18 h. A 5–10 µl aliquot was used for SDS-PAGE. Silver staining was performed using the Sigma Silver Staining kit according to the manufacturer's instructions. PS samples were transferred to nitrocellulose membranes and probed with MAb1B5.

Acetolysis of APS.  Acetolysis of APS was carried out using the modified procedure described by Kocourek and Ballou (1969).

Deacetylation of the acetolysis products.  The products of acetolysis were dissolved in 4 ml of dry methanol and 0.5 M solution of methanolic sodium methoxide was added dropwise (5–10 drops). After 20 min at 22°C, excess of sodium methoxide was decomposed by the addition of Dowex ion-exchange resin (H+ form) with vigorous stirring. Excess of methanol was removed by evaporation after the solution attained a pH of 5–6.

Isolation of the de-O-acetylated products of acetolysis of APS.  De-O-acetylated products of acetolysis were isolated by gel-filtration chromatography on a Biogel P-2 column (2 cm I.D. × 80 cm) and was eluted with 50 mM ammonium bicarbonate at 65 ml h−1 and fractions (2 ml) were collected. The RI of column effluent was monitored. Fractions containing low-molecular-weight material (from di- to pentasaccharides) were subjected to HPLC (HPLC Beckman Coulter Bioresearch) hardware system [Beckman Coulter (U.K) Limited] operated by Gold 32 Karat software (Beckman Coulter, USA) on a Glycopac N column using 32 Karat software workstation. The column was eluted in isocratic mode at a flow rate 1 ml min−1 with 10% solution of methanol in HPLC grade water. The absorbance of the column effluent at 200 and 280 nm was monitored with a Gold 168 UV detector. Fractions containing material corresponding to monosaccharide were collected and used for ES-MS analysis and the effluent containing material of size from di- to pentasaccharides was analysed by MALDI-TOF MS.

NMR spectroscopy.  Deuterium exchange of PS samples was performed by lyophilizing solutions in 99.6% D2O at least twice followed by dissolving the residue in 0.55 ml of 99.96% D2O. Acetone (δH 2.225; δC 31.45) was used as the internal standard and spectra were recorded at 40°C on either a BRUKER DAX-500 using XWINNMR software (V.2.5 and 3.1), or a Varian 500 Unity Plus Spectrometer using VNMR software (V.6.0 and 6.1). The 2D pulse programmes were as follows: TOCSY (HOHAHA) with presaturation during relaxation delay and MLEV-17 pulse sequence for mixing (Bax and Davis, 1985); NOESY using TPPI with presaturation during relaxation delay (Wagner and Wuthrich, 1982); 1H-{13C} HMQC using TPPI with presaturation during relaxation delay and GARP decoupling during acquisition (Bax et al., 1983); HMBC (Bax and Summers, 1986); 1H-{13C} HSQC-TOCSY (Bax and Subramanian, 1986) and 1H-{31P} HMQC-TOCSY (Kover et al., 1997) with or without WALTZ 16 decoupling during acquisition. For NOESY and TOCSY experiments, mixing delays of 0.12 and 0.075 s, respectively, were used.

MALDI-TOF MS.  MALDI-TOF MS was performed according to Forno et al. (2004) using a Kratos Axima CFR instrument (Kratos Analytical, Manchester, UK) in positive linear mode. Dextran was used for calibration. A solution of 10 mg ml−1 of 2,5-dihydroxybenzoic acid (Sigma Chemical, Poole, Dorset, U.K) in 4:1 [0.1% aqueous trifluoroacetic acid (by vol):acetonitrile] (by vol.) was used as matrix solution. Sample (0.5 µl) at a concentration of 50–100 pmol µl−1 was applied to the MALDI target plate followed immediately by the application of 0.5 µl of matrix solution. The mixture was rapidly dried under vacuo and ethanol (0.3 µl) was applied to each sample spot and air dried.

ES-MS.  ES-MS analysis of the products of acetolysis of APS was carried out according to Jiang and Cole (2005) on a Micromass Q-ToF micro spectrometer using Masslynx software. MS and MS/MS analyses were performed in negative mode. Samples were dissolved in 100 µl of a 5% aqueous methanol solution. A 20 µl aliquot was diluted to 1 ml with a 1:1 mixture of acetonitrile: water (v/v).

Composition and methylation analysis.  The monosaccharide composition of APS was determined by methanolysis as described (Altman et al., 1989). Methyl glycosides were converted to O-trimethylsilyl ethers and analysed by gas chromatography – mass spectometry (GC-MS) (Kakehi and Honda, 1989). APS was methylated according to Kvernheim (1987) followed by hydrolysis with 0.5 M trifluoroacetic acid for 1.5 h at 120°C. Methyl ethers were reduced with NaBD4 (22°C, 4 h) and acetylated with pyridine:acetic anhydride (1:1, by vol.) at 100°C for 1 h. Alditol acetates were analysed by GC-MS. The absolute configurations of the monosaccharides were determined as described by Gerwig et al. (1978).

De-O-phosphorylation.  De-O-phosphorylation of APS was performed according to Di Fabio et al. (1990). Usually, 6 mg of APS was dissolved in 0.1 ml of cold (4°C) 48% aqueous hydrofluoric acid HF (by vol.) and the mixture was allowed to stand at 4°C for 24 h followed by freeze-drying with a NaOH trap. GPC of de-O-phosphorylated APS (DPS) was performed on a Fractogel TSK HW-40(S) and the column effluent was monitored using a RI detector. Three fractions were obtained (DPS1, DPS2 and DPS3). Fraction DPS1(DPS) that contained polymeric material was subjected to SDS-PAGE/Western blotting with MAb1B5 and used for further analysis.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

These investigations were supported by the Medical Research Council (PG9318173). The authors wish to thank Dr E. Tarelli (Medical Biomics Centre, Jenner Wing St Georges Hospital Medical School) for running MALDI-TOF spectra, Dr G. Lord (MRC Bioanalytical Group at Birkbeck College of London) for electrospray MS analysis and Keith Pell (Queen Mary's School of Medicine and Dentistry) for EM.


  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Aduse-Opoku, J., Muir, J., Slaney, J.M., Rangarajan, M., Curtis, M.A. (1995) Characterization, genetic analysis, and expression of a protease antigen (PrpRI) of Porphyromonas gingivalis W50. Infect Immun 63: 47444754.
  • Albermann, C., Piepersberg, W. (2001) Expression and identification of the RfbE protein from Vibrio cholerae O1 and its use for the enzymatic synthesis of GDP-d-perosamine. Glycobiology 11: 655661.
  • Altman, E., Perry, M.B., Brisson, J.R. (1989) Structure of the lipopolysaccharide antigenic O-chain produced by Actinobacillus pleuropneumoniae serotype 4 (ATCC 33 378). Carbohydr Res 191: 295303.
  • Awram, P., Smit, J. (2001) Identification of lipopolysaccharide O antigen synthesis genes required for attachment of the S-layer of Caulobacter crescentus. Microbiology 147: 14511460.
  • Baretto-Bergter, E.M., Travassos, L.R., Gorin, P.A.J. (1980) Chemical structure of the d-galacto-d-mannan component from hyphae of Aspergillus niger and other Aspergillus spp. Carbohydrate Res 86: 273285.
  • Bax, A., Davis, D.G. (1985) MLEV-17-based two dimensional homonuclear magnetisation transfer spectroscopy. J Magn Reson 65: 355360.
  • Bax, A., Subramanian, S. (1986) Sensitivity-enhanced two-dimensional heteronuclear shift correlation NMR-spectroscopy. J Magn Reson 67: 565569.
  • Bax, A., Summers, M.F. (1986) H-1 and C-13 assignments from sensitivity-enhanced detection of heteronuclear multiple-bond connectivity by 2D multiple quantum NMR. J Am Chem Soc 108: 20932094.
  • Bax, A., Griffey, R.H., Hawkins, B.L. (1983) Correlation of proton and N15 chemical-shifts by multiple quantum NMR. J Magn Reson 55: 301315.
  • Bilge, S.S., Vary, J.C., Jr, Dowell, S.F., Tarr, P.I. (1996) Role of the Escherichia coli O157:H7 O side chain in adherence and analysis of an rfb locus. Infect Immun 64: 47954801.
  • Bock, K., Thogersen, H. (1982) Nuclear magnetic resonance spectroscopy in the study of mono- and oligosaccharides. Annu Rep NMR Spectr 13: 157.
  • Bretthauer, R.K., Kaczorowski, G.J., Weise, M.J. (1973) Characterization of a phosphorylated pentasaccharide isolated from Hansenula holstii NRRL Y-2448 phosphomannan. Biochemistry 12: 12511256.
  • Castric, P., Cassels, F.J., Carlson, R.W. (2001) Structural characterization of the Pseudomonas aeruginosa 1244 pilin glycan. J Biol Chem 276: 2647926485.
  • Corsaro, M.M., Evidente, A., Lanzetta, R., Lavermicocca, P., Molinaro, A. (2001) Structural determination of the phytotoxic mannan exopolysaccharide from Pseudomonas syringae pv. ciccaronei. Carbohydr Res 330: 271277.
  • Curtis, M.A., Thickett, A., Slaney, J.M., Rangarajan, M., Aduse-Opoku, J., Shepherd, P., et al. (1999) Variable carbohydrate modifications to the catalytic chains of the RgpA and RgpB proteases of Porphyromonas gingivalis W50. Infect Immun 67: 38163823.
  • Curtis, M.A., Aduse-Opoku, J., Rangarajan, M. (2001) Cysteine proteases of Porphyromonas gingivalis. Crit Rev Oral Biol Med 12: 192216.
  • Darveau, R.P., Hancock, R.E. (1983) Procedure for isolation of bacterial lipopolysaccharides from both smooth and rough Pseudomonas aeruginosa and Salmonella typhimurium strains. J Bacteriol 155: 831838.
  • Di Fabio, J.L., Michon, F., Brisson, J., Jennings, H.J. (1990) Structures of the L1 and L6 core oligosaccharide epitopes of Neisseria meningitidis. Can J Chem 68: 10291034.
  • Farquharson, S.I., Germaine, G.R., Gray, G.R. (2000) Isolation and characterization of the cell-surface polysaccharides of Porphyromonas gingivalis ATCC 53978. Oral Microbiol Immunol 15: 151157.
  • Forno, G., Bollati, F.M., Oggero, M., Kratje, R., Etcheverrigaray, M., Conradt, H.S., Nimtz, M. (2004) N- and O-linked carbohydrates and glycosylation site occupancy in recombinant human granulocyte-macrophage colony-stimulating factor secreted by a Chinese hamster ovary cell line. Eur J Biochem 271: 907919.
  • Gallagher, A., Aduse-Opoku, J., Rangarajan, M., Slaney, J.M., Curtis, M.A. (2003) Glycosylation of the Arg-gingipains of Porphyromonas gingivalis and comparison with glycoconjugate structure and synthesis in other bacteria. Curr Protein Pept Sci 4: 427441.
  • Gerwig, G.J., Kamerling, J.P., Vliegenthart, J.F.G. (1978) Determination of the D and L configuration of neutral monosaccharides by high-resolution capillary g.l.c. Carbohydrate Res 62: 349357.
  • Hernandez, L.M., Ballou, L., Alvarado, E., Tsai, P.K., Ballou, C.E. (1989) Structure of the phosphorylated N-linked oligosaccharides from the mnn9 and mnn10 mutants of Saccharomyces cerevisiae. J Biol Chem 264: 1364813659.
  • Jiang, Y., Cole, R.B. (2005) Oligosaccharide analysis using anion attachment in negative mode electrospray mass spectrometry. J Am Soc Mass Spectrom 16: 6070.
  • Jigami, Y., Odani, T. (1999) Mannosylphosphate transfer to yeast mannan. Biochim Biophys Acta 1426: 335345.
  • Kakehi, K., Honda, S. (1988) Silyl ethers of carbohydrates. In Analysis of Carbohydrates by GLC and MS. Biermann, C.J., MacGinnis, G.D. (eds). Boca Raton, FL: CRC Press, pp. 4385.
  • Knirel, Y.A., Kocharova, N.A., Shashkov, A.S., Dmitriev, B.A., Kochetkov, N.K., Stanislavsky, E.S., Mashilova, G.M. (1987) Somatic antigens of Pseudomonas aeruginosa. The structure of O-specific polysaccharide chains of the lipopolysaccharides from P. aeruginosa O5 (Lanyi) and immunotype 6 (Fisher). Eur J Biochem 163: 639652.
  • Kocourek, J., Ballou, C.E. (1969) Method for fingerprinting yeast cell wall mannans. J Bacteriol 100: 11751181.
  • Kover, K.E., Hruby, V.J., Uhrin, D. (1997) Sensitivity- and gradient-enhanced heteronuclear coupled/decoupled HSQC-TOCSY experiments for measuring long-range heteronuclear coupling constants. J Magn Reson 129: 125129.
  • Kumada, H., Haishima, Y., Umemoto, T., Tanamoto, K. (1995) Structural study on the free lipid A isolated from lipopolysaccharide of Porphyromonas gingivalis. J Bacteriol 177: 20982106.
  • Kvernheim, A.L. (1987) Methylation analysis of polysaccharides with butyllithium in dimethylsulfoxide. Acta Chem Scand 41: 150152.
  • Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680685.
  • Laine, M.L., van Winkelhoff, A.J. (1998) Virulence of six capsular serotypes of Porphyromonas gingivalis in a mouse model. Oral Microbiol Immunol 13: 322325.
  • Laine, M.L., Appelmelk, B.J., van Winkelhoff, A.J. (1996) Novel polysaccharide capsular serotypes in Porphyromonas gingivalis. J Periodontal Res 31: 278284.
  • Laine, M.L., Appelmelk, B.J., van Winkelhoff, A.J. (1997) Prevalence and distribution of six capsular serotypes of Porphyromonas gingivalis in periodontitis patients. J Dent Res 76: 18401844.
  • Nelson, R.D., Shibata, N., Podzorski, R.P., Herron, M.J. (1991) Candida mannan: chemistry, suppression of cell-mediated immunity, and possible mechanisms of action. Clin Microbiol Rev 4: 119.
  • Nelson, K.E., Fleischmann, R.D., DeBoy, R.T., Paulsen, I.T., Fouts, D.E., Eisen, J.A., et al. (2003) Complete genome sequence of the oral pathogenic bacterium Porphyromonas gingivalis strain W83. J Bacteriol 185: 55915601.
  • Ogawa, T., Kusumoto, Y., Hamada, S., McGhee, J.R., Kiyono, H. (1990) Bacteroides gingivalis-specific serum IgG and IgA subclass antibodies in periodontal diseases. Clin Exp Immunol 82: 318325.
  • Paramonov, N., Bailey, D., Rangarajan, M., Hashim, A., Kelly, G., Curtis, M.A., Hounsell, E.F. (2001) Structural analysis of the polysaccharide from the lipopolysaccharide of Porphyromonas gingivalis strain W50. Eur J Biochem 268: 46984707.
  • Rangarajan, M., Aduse-Opoku, J., Slaney, J.M., Young, K.A., Curtis, M.A. (1997) The prpR1 and prR2 arginine-specific protease genes of Porphyromonas gingivalis W50 produce five biochemically distinct enzymes. Mol Microbiol 23: 955965.
  • Rangarajan, M., Hashim, A., Aduse-Opoku, J., Paramonov, N., Hounsell, E.F., Curtis, M.A. (2005) Expression of Arg-Gingipain RgpB is required for correct glycosylation and stability of monomeric Arg-gingipain RgpA from Porphyromonas gingivalis W50. Infect Immun 73: 48644878.
  • Schifferle, R., Beanan, J.M. (1999) O-antigenic serotypes of Porphyromonas gingivalis lipopolysaccaharide J Dent Res 78: 507.
  • Shoji, M., Ratnayake, D.B., Shi, Y., Kadowaki, T., Yamamoto, K., Yoshimura, F., et al. (2002) Construction and characterization of a nonpigmented mutant of Porphyromonas gingivalis: cell surface polysaccharide as an anchorage for gingipains. Microbiology 148: 11831191.
  • Sims, T.J., Schifferle, R.E., Ali, R.W., Skaug, N., Page, R.C. (2001) Immunoglobulin G response of periodontitis patients to Porphyromonas gingivalis capsular carbohydrate and lipopolysaccharide antigens. Oral Microbiol Immunol 16: 193201.
  • Slaney, J.M., Rangarajan, M., Aduse-Opoku, J., Fawell, S., Darby, I., Kinane, D., Curtis, M.A. (2002) Recognition of the carbohydrate modifications to the RgpA protease of Porphyromonas gingivalis by periodontal patient serum IgG. J Periodontal Res 37: 215222.
  • Slodki, M.E., Ward, R.M., Boundy, J.A. (1973) Concanavalin A as a probe of phosphomannan molecular structure. Biochim Biophys Acta 304: 449456.
  • Stintzi, A., Marlow, D., Palyada, K., Naikare, H., Panciera, R., Whitworth, L., Clarke, C. (2005) Use of genome-wide expression profiling and mutagenesis to study the intestinal lifestyle of Campylobacter jejuni. Infect Immun 73: 17971810.
  • Takagi, M., Takada, H., Imanaka, T. (1990) Nucleotide sequence and cloning in Bacillus subtilis of the Bacillus stearothermophilus pleiotropic regulatory gene degT. J Bacteriol 172: 411418.
  • Vanterpool, E., Roy, F., Fletcher, H.M. (2005) Inactivation of vimF, a putative glycosyltransferase gene downstream of vimE, alters glycosylation and activation of the gingipains in Porphyromonas gingivalis W83. Infect Immun 73: 39713982.
  • Villeneuve, S., Souchon, H., Riottot, M.M., Mazie, J.C., Lei, P., Glaudemans, C.P., et al. (2000) Crystal structure of an anti-carbohydrate antibody directed against Vibrio cholerae O1 in complex with antigen: molecular basis for serotype specificity. Proc Natl Acad Sci USA 97: 84338438.
  • Vinogradov, E., Petersen, B.O., Duus, J.O. (2000) Isolation and characterization of non-labeled and 13C-labeled mannans from Pichia pastoris yeast. Carbohydr Res 325: 216221.
  • Wagner, G., Wuthrich, K. (1982) Sequential resonance assignments in protein 1H nuclear magnetic resonance spectra. Basic pancreatic trypsin inhibitor. J Mol Biol 155: 347366.
  • Westphal, O., Jann, K. (1965) Bacterial lipopolysaccharides: extraction with phenol-water and further applications of the procedure. In Methods in Carbohydrate Chemistry. Whistler, R.L. (ed.). New York: Academic Press, pp. 8391.
  • van Winkelhoff, A.J., Appelmelk, B.J., Kippuw, N., de Graaff, J. (1993) K-antigens in Porphyromonas gingivalis are associated with virulence. Oral Microbiol Immunol 8: 259265.
  • Wu, A.M., Mackenzie, N.E. (1987) Structural and immunochemical characterization of the O-haptens of Brucella abortus lipopolysaccharides from strains 19 and 2308. Mol Cell Biochem 75: 103111.