Cleavage of Human Transferrin by Porphyromonas gingivalis Gingipains Promotes Growth and Formation of Hydroxyl Radicals

Abstract

Porphyromonas gingivalis, a gram-negative anaerobic bacterium associated with active lesions of chronic periodontitis, produces several proteinases which are presumably involved in host colonization, perturbation of the immune system, and tissue destruction. The aims of this study were to investigate the degradation of human transferrin by gingipain cysteine proteinases of P. gingivalis and to demonstrate the production of toxic hydroxyl radicals (HO^·) catalyzed by the iron-containing transferrin fragments generated or by release of iron itself. Analysis by polyacrylamide gel electrophoresis and Western immunoblotting showed that preparations of Arg- and Lys-gingipains of P. gingivalis cleave transferrin (iron-free and iron-saturated forms) into fragments of various sizes. Interestingly, gingival crevicular fluid samples from diseased periodontal sites but not samples from healthy periodontal sites contained fragments of transferrin. By using ⁵⁵Fe-transferrin, it was found that degradation by P. gingivalis gingipains resulted in the production of free iron, as well as iron bound to lower-molecular-mass fragments. Subsequent to the degradation of transferrin, bacterial cells assimilated intracellularly the radiolabeled iron. Growth of P. gingivalis ATCC 33277, but not growth of an Arg-gingipain- and Lys-gingipain-deficient mutant, was possible in a chemically defined medium containing 30% iron-saturated transferrin as the only source of iron and peptides, suggesting that gingipains play a critical role in the acquisition of essential growth nutrients. Finally, the transferrin degradation products generated by Arg-gingipains A and B were capable of catalyzing the formation of HO^·, as determined by a hypoxanthine/xanthine oxidase system and spin trapping-electron paramagnetic resonance spectrometry. Our study indicates that P. gingivalis gingipains degrade human transferrin, providing sources of iron and peptides. The iron-containing transferrin fragments or the release of iron itself may contribute to tissue destruction by catalyzing the formation of toxic HO^·.

Periodontitis is a chronic inflammatory disorder of the periodontium initiated by an overgrowth of specific bacterial species and characterized by the destruction of tooth-supporting connective tissues, including the alveolar bone. Porphyromonas gingivalis, a gram-negative anaerobic bacterium, has been implicated as a major etiological agent in the onset and progression of chronic periodontitis (33). Arg- and Lys-gingipain cysteine proteinases, which occur in multiple molecular forms due to proteolytic processing of the initially translated polypeptides, are the main endopeptidases produced by P. gingivalis (12, 15, 16). Two genes code for Arg-gingipains (rgpA and rgpB), and one gene codes for Lys-gingipain (kgp) (12, 15, 16). An increasing number of reports have stressed the potential roles of Arg- and Lys-gingipains of P. gingivalis in the pathogenesis of this disease, as they are able to degrade a variety of tissue and plasma proteins and have the potential to elicit the host defense system (12, 15, 16). The critical contribution of gingipains in the pathogenicity of P. gingivalis is supported by the fact that immunization with purified Arg-gingipain A or B protects against colonization and invasion by P. gingivalis in a mouse chamber model (11).

Gingipains of P. gingivalis can participate in tissue destruction and invasion by (i) degrading extracellular matrix proteins, such as fibronectin and type IV collagen (34), (ii) activating latent host matrix metalloproteinases (7), (iii) inactivating the tissue inhibitor of metalloproteinase 1 (13), and (iv) inducing the secretion of collagenase by human gingival fibroblasts (38). Since P. gingivalis gingipains can cleave human transferrin (6, 36), we hypothesized that the iron-containing peptide fragments generated may catalyze the reaction between superoxide (O₂^−·) and hydrogen peroxide (H₂O₂) to form toxic hydroxyl radicals (HO^·) via the Haber-Weiss reaction.

The formation of HO^·, an important reactive oxygen species involved in in vitro and in vivo cell injury through its reactivity with lipids, proteins, and DNA (4), may represent an additional mechanism promoting tissue destruction during periodontitis.

The present study was aimed at investigating the degradation of human transferrin by P. gingivalis gingipains and demonstrating the production of toxic HO^· catalyzed by the iron-containing transferrin fragments or by the release of iron itself.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

P. gingivalis ATCC 33277 and three derivative gingipain-deficient mutants (KDP112 [rgpA rgpB], KDP129 [kgp], and KDP128 [rgpA rgpB kgp]) were used in the study. The mutants were constructed by allelic replacement mutagenesis or integration of a suicide plasmid, as described previously (24, 31). Bacteria were grown anaerobically (N₂-CO₂-H₂, 80:10:10) at 37°C for 24 h in Todd-Hewitt broth (THB) (BBL Microbiology Systems, Cockeysville, Md.) supplemented with hemin (10 μg/ml) and vitamin K (1 μg/ml). To maintain selective pressure and prevent the appearance of revertants, tetracycline (0.7 μg/ml) and erythromycin (10 μg/ml) were added when mutants KDP112 and KDP128 were grown on agar plates. Before the mutants were used, their phenotypes were confirmed by testing their ability to cleave the chromogenic synthetic substrates for Arg-gingipain (benzoyl-Arg-p-nitroanilide) and Lys-gingipain (N-p-tosyl-Gly-Pro-Lys-p-nitroanilide), as described previously (6).

Purification of gingipains.

Mutant KDP129 (kgp) was used to prepare a purified fraction containing both Arg-gingipains (Arg-gingipains A and B), whereas mutant KDP112 (rgpA rgpB) was used to isolate Lys-gingipain. Use of these mutants facilitated purification of the gingipains, which are difficult to separate from one another. Gingipains were purified from cell envelope extracts by affinity chromatography on arginine Sepharose 4B (Amersham Pharmacia Biotech, Baie d'Urfé, Québec, Canada) as previously described (10, 27). Linear concentrations (0 to 500 mM) of l-arginine and l-lysine in Tris-HCl buffer (pH 7.4) were used for elution of Arg-gingipain and Lys-gingipain, respectively. Arg- and Lys-gingipain activities were quantified by using benzoyl-Arg-p-nitroanilide and N-p-tosyl-Gly-Pro-Lys-p-nitroanilide, respectively (6). One unit of enzyme activity was defined as the amount of enzyme required to release 1 nmol of p-nitroaniline per h.

Degradation of human transferrin by gingipains.

Equal volumes of apotransferrin (iron-free form) or holotransferrin (iron-saturated form) (0.75 mg/ml; Sigma Chemical Co., St. Louis, Mo.), dithiothreitol (30 mM), and gingipains (300 U/ml) were incubated at 37°C for 2 h. Degradation of transferrin was evaluated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) with the buffer system of Laemmli (17), followed by Western immunoblotting with an alkaline phosphatase-conjugated goat anti-human transferrin antibody (1/3,000 dilution; Cedarlane, Hornby, Ontario, Canada). Undegraded transferrin and transferrin fragments were visualized following development in 100 mM carbonate buffer (pH 9.8) containing 0.3 mg of nitroblue tetrazolium chloride per ml and 0.15 mg of 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt per ml.

Autoradiography analysis for detection of ⁵⁵Fe following degradation of transferrin.

⁵⁵Fe-transferrin was prepared by a method based on the protocols of Pintor et al. (28) and Simonson et al. (32) by using human apotransferrin and ⁵⁵FeCl₃ (NEN Life Science Products Inc., Boston, Mass.). Apotransferrin at a concentration of 1 mg/ml was mixed with 0.0075 μmol of ⁵⁵FeCl₃ and 0.075 μmol of sodium citrate in 40 mM Tris-HCl buffer (pH 7.4) containing 2 mM sodium carbonate. After incubation at room temperature for 30 min, six consecutive dialyses (molecular mass cutoff, 12 to 14 kDa) were performed at 4°C for 16 h against 40 mM Tris-HCl buffer (pH 7.4) containing 2 mM sodium carbonate until no significant radioactivity was detected in the dialysate with a gamma counter. After this, the final concentration of transferrin was determined by the method of Lowry et al. (18), whereas the percentage of iron saturation of transferrin was evaluated by a colorimetric assay (Sigma Chemical Co.) by using ferrozine. ⁵⁵Fe-transferrin was subjected to degradation by preparations of gingipains. Briefly, equal volumes of ⁵⁵Fe-transferrin (1 mg/ml), dithiothreitol (40 mM), Tris-HCl buffer (200 mM, pH 7.2), and gingipains (300 U/ml) were incubated at 37°C for 2 h. To preserve the interaction between iron and transferrin or transferrin fragments, samples were analyzed by nondenaturing nonreducing electrophoresis on a 3 to 20% polyacrylamide gradient gel in the presence of Triton X-100, as previously described by Vyoral et al. (39). Following electrophoresis, the gel was dried and exposed at −80°C for 24 h to Kodak Biomax film (Kodak, Rochester, N.Y.) by using an intensifying screen (Transcreen-Low Energy; Kodak).

Uptake of iron from transferrin by cells of P. gingivalis.

Uptake of ⁵⁵Fe from ⁵⁵Fe-transferrin by P. gingivalis ATCC 33277 cells was determined as follows. The bacteria were grown to the late exponential growth phase in THB containing vitamin K but no hemin. In the absence of hemin, growth of P. gingivalis is possible once since bacteria use endogenous hemin stores present in the cell envelope. Iron-restricted bacterial cells were harvested by centrifugation (8,000 × g for 15 min) and suspended in either oxygen-free THB or THB containing 20 μM ferric citrate to an optical density at 660 nm (OD₆₆₀) of 1. This corresponded to a concentration of 2.5 × 10⁹ bacteria per ml, as determined with a Petroff-Hausser counting chamber. Heat-inactivated bacteria (60°C for 30 min) were used as a control. Equal volumes of ⁵⁵Fe-transferrin (13 μM) and bacterial suspensions were incubated at 37°C under anaerobiosis for up to 25 h. After various times of incubation, cells were harvested by centrifugation (8,000 × g for 15 min), washed twice in phosphate-buffered saline (PBS), and resuspended in the same buffer to an OD₆₆₀ of 1. The radioactivity associated with the bacteria was quantified with a gamma counter. Assays were performed in triplicate, and the means ± standard deviations were calculated.

Growth of P. gingivalis in the presence of transferrin.

Growth studies with P. gingivalis ATCC 33277, KDP112 (rgpA rgpB), KDP129 (kgp), and KDP128 (rgpA rgpB kgp) were carried out in the chemically defined medium previously described by Milner et al. (22). This basal medium, which contains NaH₂PO₄ (10 mM), KCl (10 mM), MgCl₂ (1.2 mM), ZnCl₂ (25 mM), CaCl₂ (20 mM), CoCl₂ (10 mM), CuCl₂ (5 mM), NaMoO₄ (0.1 mM), boric acid (5 mM), citric acid (2 mM), α-ketoglutarate (50 mM), and vitamin K (3 μM), was supplemented with either hemin (10 μg/ml), tryptone (pancreatic digest of casein; BBL Microbiology Systems) (2%), hemin plus tryptone (2%), 30% iron-saturated transferrin (2%), or apotransferrin (2%). The final maximum OD₆₆₀ reached by the cultures were recorded after 3 days of incubation at 37°C under anaerobiosis.

Formation of hydroxyl radicals.

Hydroxyl radical formation was determined by using a lumazine/xanthine oxidase system and spin trapping-electron paramagnetic resonance (EPR) spectrometry, as previously described (8). Transferrin fragments were generated by incubating equal volumes of holotransferrin (2 mg/ml in PBS), dithiothreitol (2.5 mM), and the preparation containing Arg-gingipains A and B (300 U/ml) for 2 h at 37°C. In one control, holotransferrin was replaced by apotransferrin, the iron-free form, or by deionized distilled water. An additional control consisted of replacing the enzyme preparation by PBS. All buffers, as well as the gingipain preparation, were treated with Chelex-100 (Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada) to remove adventitious iron. In addition, all glassware was washed with 6 N HCl and extensively rinsed in deionized distilled water. To the reaction mixtures were added 4-pyridinyl-N-oxide (4-POBN) (10 mM), ethanol (170 mM), and lumazine (100 mM). This was followed by addition of xanthine oxidase (40 mU/ml). The final volume of all reaction mixtures was 0.5 ml. Each reaction mixture was then transferred to a flat quartz EPR cell and placed into the cavity of an EPR spectrometer (model ESP 300; Brüker Instrument Co., Karlsruhe, Germany). EPR measurements were then recorded at 25°C. The instrument settings were as follows: modulation frequency, 100 kHz; modulation amplitude, 0.5 G; time constant, 327.68 ms; sweep time, 335.544 s; and gain, 4 × 10⁵. In the 4-POBN-ethanol spin trapping system, the formation of HO^· leads to generation of the 4-POBN-^·CH₂(OH)CH₃ spin adduct. Preliminary studies revealed that at the concentrations employed, the presence of dithiothreitol and the gingipains had no effect on the magnitude or stability of the EPR spectrum of the 4-POBN-^·CH₂(OH)CH₃ spin. Two independent experiments were carried out to demonstrate the reproducibility of the results.

Detection of transferrin fragments in gingival crevicular fluid.

Gingival crevicular fluid samples were obtained from 36 patients attending the dental clinic at Université Laval. The pocket depth of each site was measured with a Michigan periodontal probe. Patients were distributed into four categories: healthy (pocket depth, ≤3 mm), mild periodontitis (pocket depth, 4 to 5 mm), moderate periodontitis (pocket depth, 6 to 8 mm), and severe periodontitis (pocket depth, ≥9 mm). Paper strips (2 by 8 mm; 3MM; Whatman, Clifton, N.J.) were inserted into the subgingival site for 30 s. The strips were then placed in 250 μl of 100 mM PBS (pH 7.2) containing a protease inhibitor cocktail with a broad inhibitory spectrum (Complete Mini tablets; Roche Diagnostics, Laval, Quebec, Canada) to prevent proteolysis and were stored at −20°C. To recover the absorbed proteins, the samples were rapidly thawed and shaken at 4°C for 2 h. The paper strips were then removed, and the samples were kept frozen until they were used. A 10-μl portion of each gingival crevicular fluid sample was mixed with 10 μl of electrophoresis sample buffer (0.06 M Tris-HCl, 0.1% glycerol, 2% SDS, 0.015% bromophenol blue, 5% β-mercaptoethanol; pH 6.8) and boiled for 5 min. Transferrin and transferrin fragments were visualized by Western immunoblotting, as described above.

RESULTS

The degradation of human apotransferrin (iron-free form) and holotransferrin (iron-saturated form) by the purified preparations of Arg-gingipains A and B and Lys-gingipain was analyzed by SDS-PAGE and Western immunoblotting (Fig. 1). Apotransferrin appeared to be more susceptible than holotransferrin to degradation by either preparation of gingipains since no lower-molecular-mass fragments accumulated, suggesting that there was production of fragments too small to be detected by the immunological analysis used. This indicated that the presence of iron may stabilize the protein and partially protect it from degradation by gingipains. Although some fragments may have been similar, the degradation profile of holotransferrin produced by the Arg-gingipain A-Arg-gingipain B preparation differed to some extent from that produced by the preparation of Lys-gingipain. Several fragments with molecular masses between 40 and 80 kDa were produced. Incubating transferrin (iron-free or iron-saturated form) with both preparations of gingipains resulted in a profile comparable to that generated with the Lys-gingipain preparation. When human serum was used instead of commercial transferrin, degradation was also observed (data not shown).

FIG. 1.

Degradation of human apotransferrin (Apo-TF) and holotransferrin (Holo-TF) by preparations containing Lys-gingipain and/or Arg-gingipains A and B, as determined by SDS-PAGE and Western immunoblotting. Lanes 1 and 10, molecular weight markers; lanes 2 and 9, control transferrin; lanes 3 and 6, Arg-gingipains A and B; lanes 4 and 7, Lys-gingipain; lanes 5 and 8, Lys-gingipain and Arg-gingipains A and B.

In order to localize iron in the degradation products generated with the preparation containing Arg-gingipains A and B or Lys-gingipain, human apotransferrin was radiolabeled with ⁵⁵Fe. By using an iron colorimetric assay with ferrozine, the ⁵⁵Fe-transferrin preparation was estimated to be 80% iron saturated. Following incubation with gingipains, the degradation products of ⁵⁵Fe-transferrin were analyzed by nondenaturing nonreducing electrophoresis and autoradiography. As shown in Fig. 2, iron was localized in transferrin fragments with an intermediate molecular mass and a low molecular mass and also under a free form migrating with the front dye. The ⁵⁵Fe-containing band detected on top of the gel, which was generated following treatment of transferrin with gingipains, may represent nonspecific aggregates of transferrin fragments. An attempt to determine the amino acid sequence of the small peptide containing iron was unsuccessful, probably due to the small amount recovered (data not shown).

FIG. 2.

Degradation of ⁵⁵Fe-transferrin by preparations containing Lys-gingipain and/or Arg-gingipains A and B, as determined by nondenaturing nonreducing PAGE and autoradiography. (A) Lane 1, control transferrin; lane 2, uncomplexed ⁵⁵Fe. (B) Lane 3, control transferrin; lane 4, transferrin plus Arg-gingipains A and B; lane 5, transferrin plus Lys-gingipain; lane 6, transferrin plus Arg-gingipains A and B and Lys-gingipain.

The intracellular uptake of ⁵⁵Fe from ⁵⁵Fe-transferrin by cells of P. gingivalis was investigated. Figure 3 indicates that the uptake of radioactive iron was time dependent. In the presence of ferric citrate as an alternative source of iron, the uptake of ⁵⁵Fe was significantly reduced, whereas heat-inactivated cells did not incorporate any ⁵⁵Fe. Interestingly, mutant KDP128 deficient for both Arg- and Lys-gingipain activities and unable to cleave transferrin did not take up ⁵⁵Fe (data not shown).

FIG. 3.

Uptake of ⁵⁵Fe from ⁵⁵Fe-transferrin by live and heat-treated cells of P. gingivalis ATCC 33277 following various periods of incubation. An assay in the presence of ferric citrate as an alternative source of iron was also carried out. Assays were performed in triplicate, and the means and standard deviations are indicated.

The capacities of P. gingivalis ATCC 33277 and the gingipain-deficient mutants to use the physiological form of transferrin (30% iron saturated) as a source of iron and peptides to promote bacterial growth were investigated. Figure 4 shows that all strains grew in the control medium, which consisted of the chemically defined medium supplemented with hemin and tryptone as sources of iron and peptides, respectively. In the absence of these supplements, no growth was observed. The wild-type strain and to a lesser extent the Arg-gingipain-deficient mutant (KDP112) and the Lys-gingipain-deficient mutant (KDP129) were able to grow in the chemically defined medium supplemented with 30% iron-saturated transferrin. Under these conditions, mutant KDP128, deficient for all three gingipains, did not grow. None of the strains could grow in the presence of the iron-free form of transferrin.

FIG. 4.

Growth of P. gingivalis ATCC 33277 and three derivative gingipain-deficient mutants (KDP112 [rgpA rgpB], KDP129 [kgp], and KDP128 [rgpA rgpB kgp]) in a chemically defined medium containing various iron and peptide supplements.

The capacity of the transferrin degradation products generated by the preparation containing Arg-gingipains A and B to catalyze the production of HO^· is indicated by the EPR spectra shown in Fig. 5. In the absence of added Fe, the reaction of xanthine oxidase and lumazine did not lead to the formation of HO^· as detected by spin trapping (Fig. 5, trace A). Significant production of HO^· was observed when the degradation products of holotransferrin were included in the lumazine/xanthine oxidase system (Fig. 5, trace D). This production was far greater than that observed with holotransferrin that had not been previously exposed to Arg-gingipains A and B (Fig. 5, trace C). No HO^· formation was detected when apotransferrin cleaved with Arg-gingipains A and B (Fig. 5, trace B) was employed or when the gingipain preparation alone was added to the lumazine/xanthine oxidase reaction mixture (data not shown). Comparable results were obtained in two independent experiments.

FIG. 5.

Degradation products of transferrin generated by the action of Arg-gingipains A and B can catalyze the Haber-Weiss reaction. The spectra are EPR spectra, representative of two separate experiments, resulting from the reaction of xanthine oxidase (40 mU/ml) with lumazine (100 mM) in the presence of ethanol (170 mM) and 4-POBN (170 mM) in PBS (trace A), apotransferrin treated with Arg-gingipains A and B (trace B), holotransferrin that had not been incubated with Arg-gingipains A and B (trace C), or holotransferrin that had been incubated with Arg-gingipains A and B (trace D). The final concentration of transferrin in each case was 1.5 μM. The EPR peaks in traces C and D are peaks of 4-POBN-^·CH₂(OH)CH₃, which in this spin trapping system reflected the formation of OH^·. The magnitude of each peak corresponds to the amount of hydroxyl radical detected.

Finally, to support the hypothesis that there is in vivo degradation of transferrin during periodontitis, gingival crevicular fluid samples obtained from patients with different periodontal conditions were analyzed by SDS-PAGE and Western immunoblotting. As shown in Fig. 6, some samples recovered from healthy sites showed the presence of a weak band corresponding to intact transferrin. None of the samples showed the presence of transferrin fragments. On the other hand, most gingival crevicular samples obtained from patients suffering from periodontitis at various levels of severity showed the presence of both transferrin and transferrin fragments of various sizes.

FIG. 6.

Detection of transferrin and transferrin fragments in gingival crevicular fluid samples by SDS-PAGE and Western immunoblotting. Lane 1, molecular mass markers (myosin [206 kDa], β-galactosidase [117 kDa], bovine serum albumin [79 kDa], ovalbumin [48 kDa], carbonic anhydrase [34.7 kDa], soybean trypsin inhibitor [29 kDa], lysozyme [21 kDa], and aprotinin [7.5 kDa]); lane 2, control human serum diluted 1/1,000; lanes 3 to 12, gingival crevicular fluid samples. (A) Samples from healthy periodontal sites. (B) Samples from periodontal sites with a pocket depth of 4 to 5 mm. (C) Samples from periodontal sites with a pocket depth of 6 to 8 mm. (D) Samples from periodontal sites with a pocket depth of ≥9 mm.

DISCUSSION

Transferrin is a glycoprotein possessing two iron-binding sites that is present in biological fluids, including gingival crevicular fluid. Transferrin plays an important role in host defense by sequestering iron and making it unavailable for pathogens (26). The iron-binding property of transferrin also allows this protein to function as an antioxidant in vivo since the bound iron cannot act as an HO^· catalyst (2). However, it has been demonstrated previously that proteolytic fragments of transferrin can bind iron in a catalytic form and are able to catalyze the formation of HO^· (5, 20, 21), an important reactive oxygen species involved in in vitro and in vivo cell injury by reacting with lipids, proteins, and DNA (4). In this study, we proposed that P. gingivalis gingipains cleave human transferrin to produce iron-containing fragments or to release free iron that may in turn allow bacterial growth and promote the formation of toxic HO^·.

In the first part of the study, SDS-PAGE-Western immunoblotting analysis showed that transferrin and transferrin fragments were present in gingival crevicular fluid samples, particularly samples obtained from patients suffering from periodontitis. These results support those of Adonogianaki et al. (1), who demonstrated by immunological analysis that the transferrin concentrations in gingival crevicular fluid samples were significantly higher when they were obtained from diseased sites than when they were obtained from healthy periodontal sites. The presence of transferrin fragments in samples from periodontitis patients suggests that there is in vivo proteolytic degradation. Proteases responsible for this degradation may be host and/or bacterium derived. On the one hand, host cells during the inflammatory process of periodontitis produce a variety of matrix metalloproteinases which accumulate in periodontal sites (37). On the other hand, diseased periodontal sites harbor numerous bacterial species, including P. gingivalis, that possess high proteolytic potential (9).

Preparations of Arg- and Lys-gingipains were used to verify the contribution of each P. gingivalis gingipain activity to the degradation of human transferrin. A greater variety of fragments were produced by the preparation containing Lys-gingipain incubated in the presence of iron-saturated transferrin. This may be explained by the fact that transferrin possesses many more lysine residues than arginine residues in its amino acid sequence (56 and 23 residues, respectively) (29). Degradation also occurred when human serum was used as the transferrin source. This suggests that gingipain activity is not affected by plasma proteinase inhibitors and supports the notion that degradation of transferrin in vivo is possible. The degree of iron saturation of transferrin was found to affect the gingipain degradation profile. The iron-free form was more susceptible to degradation, suggesting that the presence of iron may render the cleavage site less accessible to gingipains. Greater stability of the iron-saturated form of transferrin with Pseudomonas aeruginosa elastase has been reported previously (41). In a recent study, Sroka et al. (36) reported that the lysine-specific gingipain of P. gingivalis was more effective than the arginine-specific gingipain in degrading transferrin present in human serum. Interestingly, these workers also found that growth of P. gingivalis in the presence of 10% human serum was associated with a significant increase (>10-fold) in Lys-gingipain activity.

The use of ⁵⁵Fe-labeled transferrin allowed us to localize iron following its degradation by preparations containing gingipains. Iron was found to be associated with low- and intermediate-molecular-mass fragments and to be in a free form, which appeared to be predominant. ⁵⁵Fe was found to be assimilated by whole cells of P. gingivalis. The fact that no uptake was observed with heat-inactivated cells, as well as with the gingipain-deficient mutant (KDP128), suggests that there is specific intracellular uptake rather than nonspecific binding of iron or iron-containing fragments to the bacterial cell surface. These results are in agreement with the previous finding (6) that proteolytic cleavage of transferrin is the sole mechanism of iron acquisition in P. gingivalis. In the presence of a source of iron that is more easily assimilable, such as ferric citrate, much less ⁵⁵Fe was taken up by P. gingivalis cells. Uptake of iron from transferrin mediated by proteolytic degradation was previously reported for Leishmania chagasi (40) and Pseudomonas pseudomallei (42).

The ability of transferrin to serve as a source of peptides and iron to support growth of P. gingivalis was also tested. It was found that 30% iron-saturated transferrin, the physiological form, could serve as an efficient source of both iron and peptides for growth of the wild-type strain. The Arg-gingipain-deficient mutant (KDP112) and the Lys-gingipain-deficient mutant (KDP129) were also capable of growth, although to a lesser extent, in the presence of 30% iron-saturated transferrin. Mutant KDP128, deficient for all three gingipains, grew only in the basal medium containing hemin as a source of iron and casein hydrolysate as a source of peptides. The results described above support the hypothesis that both gingipains play a critical role in the acquisition of essential nutrients required for growth of P. gingivalis. Other periodontopathogens found in periodontal sites may also take advantage of the production of iron and peptides by P. gingivalis. In a previous study (14), evidence that the ability of P. gingivalis to multiply in vitro under iron-limiting conditions may be correlated with its ability to induce infections in an animal model was obtained. It was suggested that strains capable of causing infections have either a more efficient iron acquisition system or a lower iron requirement for growth than nonpathogenic strains.

Degradation products of transferrin generated by the action of Arg-gingipains A and B were found to catalyze via the Haber-Weiss reaction the formation of cytotoxic HO^·, which may contribute to periodontal tissue injury. Indeed, HO^· can interact with a variety of biological molecules, resulting in breakdown of DNA, oxidation of proteins, and stimulation of peroxidation of membrane lipids. The production of HO^· mediated by the action of gingipains on transferrin may also (i) activate latent collagenases produced by human neutrophils (35) and (ii) cause depolymerization of gingival proteoglycan and hyaluronic acid (3), two mechanisms that favor periodontal tissue destruction. In vivo, O₂^−· and H₂O₂ are likely produced by the membrane-associated NADPH-dependent oxidase system of neutrophils following contact with bacteria (23, 30). This novel mechanism of tissue destruction mediated by iron-containing transferrin fragments or the release of iron itself could take place during periodontitis since the underlying tissue itself has an adequate supply of oxygen via the microvasculature that allows the formation of reactive oxygen species. Reactive oxygen species, including HO^·, may also have detrimental effects on bacteria. In P. gingivalis, the genome encodes a superoxide dismutase, an alkyl hydroperoxide reductase, a thiol peroxidase, and a Dps that is involved in the repair of oxidatively damaged nucleic acid (19, 25). Altogether, these proteins function to protect bacteria from oxidative stresses.

In summary, gingipains of P. gingivalis can cleave transferrin found in gingival crevicular fluid to produce iron and peptides that support the growth of asaccharolytic bacteria present in the periodontal pocket. Transferrin fragments containing iron or the release of iron itself may also promote the production of cytotoxic HO^·, resulting in periodontal tissue degradation.