Binding and Utilization of Human Transferrin by Prevotella nigrescens

Abstract

To survive and multiply within their hosts, pathogens must possess efficient iron-scavenging mechanisms. In the present study, we investigate the capacity of Prevotella nigrescens and Prevotella intermedia to use various sources of iron for growth and characterize the transferrin-binding activity of P. nigrescens. Iron-saturated human transferrin and lactoferrin, but not ferric chloride and the iron-free form of transferrin, could be used as sources of iron by P. nigrescens and P. intermedia. Neither siderophore activity nor ferric reductase activity could be detected in P. nigrescens and P. intermedia. However, both species showed transferrin-binding activity as well as the capacity to proteolytically cleave transferrin. To various extents, all strains of P. nigrescens and P. intermedia tested demonstrated transferrin-binding activity. The activity was heat and protease sensitive. The capacity of P. nigrescens to bind transferrin was decreased when cells were grown in the presence of hemin. Preincubation of bacterial cells with hemin, hemoglobin, lactoferrin, fibrinogen, immunoglobulin G, or laminin did not affect transferrin-binding activity. The transferrin-binding protein could be extracted from the cell surface of P. nigrescens by treatment with a zwitterionic detergent. Subjecting the cell surface extract to affinity chromatography on an agarose-transferrin column revealed that it contained a protein having an estimated molecular mass of 37 kDa and possessing transferrin-binding activity. The transferrin-binding activity of P. nigrescens and P. intermedia may permit the bacteria to obtain iron for survival and growth in periodontal pockets.

Periodontal diseases affect the tooth-supporting tissues and are initiated by an overgrowth of specific bacterial species found at the gingival margin. A number of research groups have reported associations between the presence of specific bacterial species in periodontal pockets and the different forms of periodontal diseases (reviewed in reference 16). Although the recent subdivision of strains of Prevotella intermedia into P. intermedia and Prevotella nigrescens makes earlier microbiological studies difficult to interpret, these two species have been suggested to play an etiologic role in gingivitis and destructive periodontitis (16). Recently, Paquet and Mouton (27) showed that strains typed as P. intermedia or P. nigrescens can be isolated from a variety of clinical situations, including gingival health, gingivitis, and periodontitis. This finding suggests that these strains may be opportunistic pathogens.

Iron is a constituent of important metabolic enzymes and is essential for the growth of almost all microorganisms (24). Consequently, a critical component of the virulence of microorganisms is their ability to obtain iron from their hosts. Little is known about iron sources in the periodontal environment. Iron-containing proteins such as hemoglobin, lactoferrin, and transferrin are known constituents of gingival crevicular fluid (GCF) (5, 7) and are likely to serve as sources of iron for the growth of periodontopathogens in vivo. In the course of periodontitis, transferrin may represent one of the most important sources of iron for periodontopathogens. To support that idea, Curtis et al. (7) showed that transferrin, along with albumin and immunoglobulin G, was the major protein in GCF from patients with gingivitis. They also reported that transferrin was present in large amounts in GCF from patients with destructive periodontitis (7).

There are several different mechanisms by which pathogenic bacteria can acquire iron from human transferrin, thus allowing their multiplication in the host. Extracellular low-molecular-mass iron-chelating molecules, also called siderophores, can sequester the iron bound to transferrin and transport it to a specific receptor present on the bacterial cell surface (13–15, 24, 33). Some bacterial species can obtain iron from transferrin via a siderophore-independent system which involves (i) production of cell surface receptors highly specific for transferrin (15, 24, 26, 32, 33); (ii) proteolytic cleavage of transferrin, resulting in disruption of the iron-binding sites, with the release of free iron (26); or (iii) reduction of exogenous Fe³⁺ and the consequent release of Fe²⁺ (15, 33).

Studies of sources of iron for and mechanisms of iron acquisition by periodontopathogens are crucial to a better understanding of the virulence of these bacteria. Although a number of research groups have investigated these aspects for Porphyromonas gingivalis (2, 30), to our knowledge nothing has been done concerning other black-pigmented anaerobic bacteria. The aims of this study were to investigate the capacity of P. nigrescens and P. intermedia to use various sources of iron and to study the transferrin-binding activity of P. nigrescens.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

P. nigrescens ATCC 33563, 5W2, Cg1265, R102, T2, and YD22-4; P. intermedia ATCC 25611, A5.4/6, BH20/30, BMH, G8-9K-3, and NY363; Actinomyces viscosus 54.2; Streptococcus mutans ATCC 10449; Actinobacillus actinomycetemcomitans ATCC 29522; Capnocytophaga ochracea 1956c; Fusobacterium nucleatum 102.3; Peptostreptococcus micros 89A; and Treponema denticola ATCC 35405 were used in this study. Most experiments were carried out with P. nigrescens ATCC 33563 and P. intermedia ATCC 25611. Bacteria were routinely grown in mycoplasma broth base (BBL Microbiology Systems, Cockeysville, Md.) supplemented with hemin (10 μg/ml), vitamin K (1 μg/ml), and glucose (20 mg/ml) (MBB-glucose).

Growth studies were carried out with the above medium not supplemented with hemin but treated with the chelating resin (3 g/100 ml) Chelex 100 (Sigma Chemical Co., St. Louis, Mo.) for 2 h at room temperature with constant agitation. This iron-poor medium was supplemented with either hemin, ferric chloride, human lactoferrin (iron-saturated form; 1,500 μg of iron/g), human apotransferrin (iron-free form; ≤30 μg of iron/g), or human holotransferrin (iron-saturated form; 1,200 to 1,600 μg of iron/g), all at 10 μM and obtained from Sigma. Human serum was also tested for its capacity to support bacterial growth. All cultures were incubated at 37°C in an anaerobic chamber (N₂-H₂-CO₂, 80:10:10). To evaluate the capacity of the bacteria to use the different iron sources, cultures were serially transferred into fresh media (1% inoculum) for up to 10 successive subcultures. An optical density of 0.5 at 660 nm was required before bacteria were transferred. A compound that could sustain bacterial growth for at least 10 subcultures was considered an efficient source of iron. Doubling times were estimated from the semilogarithmic plot of growth curve data.

Electrophoretic analysis of transferrin.

Proteolytic cleavage of transferrin during the growth of P. nigrescens ATCC 33563 and P. intermedia ATCC 25611 was evaluated by sodium dodecyl sulfate (SDS)–12.5% polyacrylamide gel electrophoresis (PAGE) analysis of culture supernatants obtained at various stages during culturing (mid-log, early stationary, and late stationary growth phases). Electrophoresis was carried out by the procedure of Laemmli (22), and gels were stained for proteins with Coomassie brilliant blue.

Detection of siderophore activity.

The universal siderophore assay of Schwyn and Neilands (28) was used to evaluate the production of siderophores by P. nigrescens ATCC 33563 and P. intermedia ATCC 25611. Culture supernatants from bacteria grown (36 h) in iron-chelated MBB-glucose medium (hemin free; third subculture) were mixed with Chrome Azurol S solution, and the absorbance at 630 nm was measured. Culture supernatants concentrated 10-fold by freeze-drying were also tested. Serial dilutions (1:2) of ferrichrome (100 μg/ml in MBB-glucose; Sigma), a siderophore produced by Ustilago sphaerogena (11), were used to establish the sensitivity of the colorimetric method.

Determination of ferric reductase activity.

The procedure of Morrissey et al. (25) was used to detect ferric reductase activity in supernatants and whole cells of P. nigrescens ATCC 33563 and P. intermedia ATCC 25611 from cultures grown (36 h) in iron-chelated MBB-glucose medium (hemin free; third subculture). An aliquot of 500 μl of a culture was centrifuged, and cells were suspended in 1 ml of assay buffer (50 mM sodium citrate [pH 6.5], 5% glucose) containing 1 mM ferric chloride and 1 mM bathophenanthroline disulfonate. The supernatant (pH adjusted to 6.5) of the culture was mixed with 500 μl of assay buffer (twofold concentrated). Samples were incubated at 30°C for 60 min in the dark, and the absorbance at 520 nm of the assay mixture supernatant was measured. When cells were tested, the assay mixture was centrifuged prior to measurement of the absorbance. The level of ferrous ions produced was estimated from a calibration curve constructed from a solution of known ion concentrations. An uninoculated culture medium served as a negative control. Cells of Candida albicans LAM-1 were previously reported as possessing ferric reductase activity (25) and were used as a positive control.

Determination of transferrin-binding activity.

The binding of human transferrin by whole cells of the strains listed above was evaluated by a solid-phase dot blot enzyme procedure. A nitrocellulose membrane was spotted with 5 μl of a cell suspension (optical density at 660 nm in 50 mM phosphate-buffered saline [pH 7.2] [PBS], 1.0) of bacteria grown (36 h) in iron-chelated MBB-glucose medium (hemin free; third subculture). This quantity corresponded to the application of approximately 5 × 10⁶ to 15 × 10⁶ cells, as determined with a Petroff-Hausser counting chamber. The membrane was incubated in 20 mM Tris buffer (pH 7.5)–0.5 M NaCl (TBS) supplemented with 3% bovine serum albumin for 1 h at room temperature with shaking. The membrane was transferred to TBS containing 1.5% bovine serum albumin and 1 μg of horseradish peroxidase (HRP)-conjugated human transferrin (Bio/Can Scientific, Mississauga, Ontario, Canada) per ml and incubated for 4 h at room temperature with shaking. The membrane was washed (four times for 15 min each time) in TBS containing 0.05% Tween 20 and stained with a color development kit (Bio-Rad Laboratories, Mississauga, Ontario, Canada) in accordance with the manufacturer’s instructions. Positive transferrin-binding activity was indicated by appearance of a purple spot. Cells of Moraxella catarrhalis PD were used as a positive control (34).

Effects of growth conditions and treatments on transferrin-binding activity.

The effects of growth conditions on the transferrin-binding activity of P. nigrescens ATCC 33563 were investigated by comparing in the solid-phase dot blot enzyme assay the activities of cells obtained after growth in (i) iron-chelated MBB-glucose medium and (ii) iron-chelated MBB-glucose medium supplemented with 10 μM hemin. In order to determine the nature of the molecules involved in the transferrin-binding activity, cells of P. nigrescens ATCC 33563 were submitted to various treatments prior to the dot blot enzyme assay. The heat stability of the transferrin-binding activity was tested by incubation (10 min) of whole cells at 60, 70, or 80°C. Bacteria were also incubated for 4 h at 37°C with either pancreatic trypsin, pancreatic chymotrypsin, or proteinase K at a final concentration of 1.0 mg/ml. Lastly, the effects of putative inhibitors of transferrin-binding activity were evaluated by preincubating the cells in the presence of selected substances. The molecules included in this experiment were hemin, human holotransferrin, human apotransferrin, human hemoglobin, bovine lactoferrin, human fibrinogen, human immunoglobulin G, and human laminin, all at 1.0 mg/ml.

Identification of transferrin-binding proteins.

Transferrin-binding proteins of P. nigrescens ATCC 33563 were extracted by suspending whole cells from a 500-ml culture (36 h; third subculture in iron-chelated hemin-free MBB-glucose medium) in 40 ml of 50 mM PBS (pH 8.0) containing 10 mM EDTA and 0.5% Zwittergent 3-14 (Calbiochem, La Jolla, Calif.). After gentle shaking for 18 h at 4°C, the suspension was centrifuged (10,000 × g for 30 min), and the supernatant was collected. This extract was submitted to affinity chromatography to isolate molecules with transferrin-binding activity as described by Ferron et al. (12). Briefly, holotransferrin in PBS (pH 8.0) (50 mg in 10 ml) was incubated with agarose beads (Affi-Gel 15; Bio-Rad) for 18 h at 4°C. Thereafter, the beads were harvested and further incubated (1 h at 4°C) in PBS (pH 8.0) containing 1 M glycine to block the nonreactive ester groups. After the samples were washed in PBS (pH 8.0), the agarose-transferrin was placed in a column (0.7 by 12 cm) equilibrated with PBS (pH 8.0) containing 0.5% Zwittergent 3-14. The bacterial extract was loaded, and the column was washed with PBS. Bound proteins were eluted with 100 mM glycine-HCl (pH 3.2), and the fractions (2 ml) obtained were brought to neutrality with 1 N NaOH. The fractions were analyzed for transferrin-binding activity with the solid-phase dot blot enzyme assay. Fractions showing activity were pooled and concentrated (10-fold) by ultrafiltration through a 10,000-Da (nominal molecular mass cutoff) membrane filter. Proteins contained in this final fraction were separated by SDS–12% PAGE by the procedure of Laemmli (22). After electrophoresis, proteins were visualized by staining with silver nitrate. The fraction was also electrophoretically transferred to a nitrocellulose membrane at a constant voltage of 60 V for 2 h. The presence of protein bands with transferrin-binding activity was determined with HRP-conjugated transferrin as described for the dot blot enzyme assay.

RESULTS

In the first part of this study, the growth of P. nigrescens ATCC 33563 and P. intermedia ATCC 25611 was evaluated by use of an iron-chelated MBB-glucose medium supplemented with different sources of iron. The results obtained are summarized in Table 1. Although the bacteria could grow significantly for the first three subcultures, the iron-chelated medium without supplements was unable to support the long-term growth of either species. The initial growth obtained in the absence of a source of iron was likely related to cellular reserves of hemin or to a carryover of iron during inoculation. Supplementing the medium with either hemin, lactoferrin, or holotransferrin allowed growth for at least 10 successive subcultures. The minimal amount of holotransferrin required to support bacterial growth was found to be 2.5 μM. On the other hand, ferric chloride and apotransferrin, the iron-free form of transferrin, were not able to support the long-term growth of either species. When human serum was used as a growth medium, P. nigrescens could be cultivated for 10 subcultures, whereas P. intermedia did not grow for more than 1 subculture. Comparison of the growth of P. nigrescens in a medium containing either holotransferrin or hemin (10 μM) as the source of iron revealed similar growth rates (doubling time, 5.75 h). However, the final optical density at 660 nm obtained with hemin was higher than that obtained with holotransferrin (1.12 compared to 0.82, respectively).

TABLE 1.

Growth responses of P. nigrescens and P. intermedia under different iron conditions

Strain	Growth response under the following iron conditiona:
	None	Hemin	FeCl3	Lactoferrin	Apo-Tf	Holo-Tf	Serum
P. nigrescens ATCC 33563	−	+	−	+	−	+	+
P. intermedia ATCC 25611	−	+	−	+	−	+	−

^aHemin, FeCl₃, lactoferrin, apotransferrin (Apo-Tf), and holotransferrin (Holo-Tf) were added at 10 μM in iron-chelated MBB-glucose medium. None, no iron was added. +, growth was obtained for at least 10 successive subcultures; −, no long-term growth. 

SDS-PAGE analysis of supernatants obtained at various stages during the growth of P. nigrescens revealed partial proteolytic degradation of transferrin (Fig. 1). Most of the degradation seemed to occur once the culture had reached the stationary growth phase. Initial cleavage of the transferrin molecule was associated with the generation of a fragment which had a molecular mass of approximately 40 kDa and which was further degraded into peptides too small (<15 kDa) to be detected by the electrophoretic procedure used. The possibility that the binding of transferrin to bacterial cells also might have been partly responsible for the decreased intensity of the transferrin band in the culture supernatant should not be excluded. Similar results for the proteolytic degradation of transferrin by P. intermedia were also obtained (data not shown).

FIG. 1.

SDS-PAGE analysis of transferrin in the culture supernatant of P. nigrescens ATCC 33563 grown in iron-chelated MBB-glucose medium (hemin-free) supplemented with holotransferrin. Lane 1, uninoculated culture medium; lane 2, mid-log growth phase (12-h culture); lane 3, early stationary growth phase (24-h culture); lane 4, late stationary growth phase (36-h culture); and lane 5, late stationary growth phase (48-h culture). Molecular mass markers were, from top to bottom, phosphorylase b (97.4 kDa), bovine serum albumin (68 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa), and β-lactoglobulin (19 kDa).

Siderophore activity in the culture supernatants of P. nigrescens ATCC 33563 and P. intermedia ATCC 25611 grown under iron-restricted conditions could not be detected by the universal Chrome Azurol S assay (28). Concentrated culture supernatants (10-fold) were also devoid of activity. The minimal concentration of ferrichrome required to yield a positive reaction was 5 μg/ml. Ferric reductase activity could not be detected in the supernatants of P. nigrescens and P. intermedia grown in the iron-restricted culture medium. Bacterial cells were also devoid of ferric reductase activity. The reduction of Fe³⁺ was observed with cells of C. albicans, which served as the positive control.

Transferrin-binding activity was tested by a solid-phase dot blot enzyme assay in which whole cells were immobilized on a nitrocellulose membrane, which was then probed with HRP-conjugated human transferrin (Fig. 2). With this binding assay, a positive reaction was obtained with M. catarrhalis, P. nigrescens ATCC 33563, and P. intermedia ATCC 25611. Additional strains of P. nigrescens and P. intermedia as well as a number of gram-positive and gram-negative bacterial species (A. viscosus, S. mutans, A. actinomycetemcomitans, C. ochracea, F. nucleatum, P. micros, and T. denticola) were tested in this assay. All strains of P. nigrescens and P. intermedia were found to bind transferrin to various extents, whereas the other bacterial species under investigation did not show any transferrin-binding activity. Strains of P. nigrescens and P. intermedia could be categorized as reacting strongly (ATCC 33563, ATCC 25611, G8-9K-3, T2, NY363, R102, BH20/30, and Cg1265) or weakly (5W2, YD22-4, BMH, and A5.4/6) in the assay.

FIG. 2.

Demonstration of transferrin-binding activity of oral bacteria by a solid-phase dot blot enzyme assay. (A) M. catarrhalis PD. (B) P. nigrescens ATCC 33563. (C) P. intermedia ATCC 25611. (D) P. nigrescens 5W2. (E) P. intermedia BMH. (F) A. actinomycetemcomitans ATCC 29522.

The transferrin-binding activity of P. nigrescens ATCC 33563 was further investigated. Growth conditions appeared to modulate the level of transferrin-binding activity (Fig. 3). Cells cultivated in the presence of hemin showed much less transferrin-binding activity than cells grown in the hemin-free iron-chelated MBB-glucose medium. The effects of various treatments on the binding of transferrin by P. nigrescens ATCC 33563 are shown in Fig. 3. Heat treatment (10 min) of whole cells at 70°C completely inhibited the binding of transferrin, whereas no effect was observed after treatment at 60°C. Treatment of cells with proteolytic enzymes (trypsin, chymotrypsin, or proteinase K) was associated with a strong decrease in transferrin-binding activity. The effect of putative inhibitors on the transferrin-binding activity of P. nigrescens was also investigated by the dot blot enzyme assay. Preincubation of bacteria with hemin, hemoglobin, lactoferrin, fibrinogen, immunoglobulin G, or laminin did not affect transferrin-binding activity, whereas complete inhibition was obtained when cells were preincubated with either the iron-free or the iron-saturated form of transferrin.

FIG. 3.

Effect of various growth conditions or treatments on transferrin-binding activity of P. nigrescens ATCC 33563, as determined by the solid-phase dot blot enzyme assay. (A) Cells grown in the presence of hemin. (B) Cells grown in the iron-chelated medium. (C) Cells treated with trypsin. (D) Cells treated at 70°C. (E) Cells preincubated with immunoglobulin G. (F) Cells preincubated with iron-free transferrin.

Cells of P. nigrescens ATCC 33563 were treated with a zwitterionic detergent (Zwittergent 3-14) to solubilize transferrin-binding proteins from the outer cell envelope. This extract demonstrated strong transferrin-binding activity, as revealed by the solid-phase dot blot enzyme assay (data not shown). In order to identify the transferrin-binding proteins of P. nigrescens, the extract was submitted to affinity chromatography on an agarose-transferrin column. The fractions showing transferrin-binding activity were pooled and analyzed by SDS-PAGE (Fig. 4A). Two bands, 37 and 80 kDa, were visualized by silver nitrate staining. The 80-kDa band was found to react with an antitransferrin antibody and likely represented transferrin molecules that got loose from the agarose beads (data not shown). The pooled fractions were also analyzed by SDS-PAGE, Western blotting, and reactivity with HRP-conjugated transferrin (Fig. 4B). These procedures revealed that only the 37-kDa band possessed transferrin-binding activity.

FIG. 4.

SDS-PAGE analysis of transferrin-binding proteins obtained by affinity chromatography of a zwitterionic detergent extract from P. nigrescens ATCC 33563. (Left gel) Silver nitrate staining. Lane 1, molecular mass standards (from top to bottom) phosphorylase b (97.4 kDa), bovine serum albumin (68 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa), and β-lactoglobulin (19 kDa); lane 2, zwitterionic detergent extract (initial material); lane 3, pooled fractions, which showed transferrin-binding activity in the solid-phase dot blot enzyme assay. (Right gel) Western blot showing reactivity with HRP-conjugated transferrin. Pooled fractions which showed transferrin-binding activity in the solid-phase dot blot enzyme assay were tested.

DISCUSSION

Since iron plays significant roles in metabolic reactions, the ability of bacterial pathogens to obtain this growth-essential nutrient from their hosts is a major virulence determinant. Transferrin, whose major physiological role is the solubilization of Fe³⁺ and its delivery from sites of absorption and storage to sites of utilization, plays an important role in host defense by rendering the iron nonavailable for microorganisms. Several different mechanisms for iron acquisition from transferrin have been demonstrated for pathogenic bacteria (15, 24, 26, 32, 33).

In this study, we showed the capacity of P. nigrescens and P. intermedia to grow in the presence of holotransferrin but not apotransferrin. This finding suggests that both bacterial species can obtain iron from the iron-loaded form of the plasma protein transferrin. The absence of growth in the presence of apotransferrin, in addition to being related to the lack of iron, may have resulted from the antibacterial activity of the molecule. Indeed, Ellison et al. (10) previously reported that iron-binding proteins (lactoferrin and transferrin) can alter the outer membrane permeability of Gram-negative bacteria. This alteration was thought to be related to the chelating property of the molecules. We also found that P. nigrescens could grow in human serum, a condition that more closely mimics an in vivo situation. This result indicates that P. nigrescens is resistant to serum bactericidal activity and supports the idea that this bacterial species may be able to utilize iron-bound transferrin present in the GCF found in periodontal pockets.

Neither siderophore activity nor ferric reductase activity could be detected in P. nigrescens and P. intermedia. However, both species showed transferrin-binding activity as well as proteolytic activity toward transferrin. The transferrin-binding capacity appeared not to be a characteristic common in oral bacteria, since among the nine bacterial species tested, only strains of P. nigrescens and P. intermedia were found to attach to transferrin. The transferrin-binding activity was decreased when cells were grown under hemin-plentiful conditions compared to iron-restricted conditions. Since preincubation of bacterial cells with hemin did not show any inhibitory effect on the transferrin-binding activity (in the dot blot enzyme assay), the activity may be hemin regulated. Regulation of transferrin receptor expression by hemin has been reported for Haemophilus influenzae (17).

The binding of transferrin to P. nigrescens did not involve electrostatic interactions, since sodium chloride was included at 0.5 M during the incubation with HRP-conjugated transferrin. Preincubation of cells with either iron-free or iron-saturated transferrin prevented the binding of HRP-conjugated transferrin, indicating that the level of iron saturation of the transferrin molecule had no effect on the binding to bacteria. Proteins with transferrin-binding activity could be extracted from the cell surface of P. nigrescens by treatment with a zwitterionic detergent. Affinity chromatography on a transferrin-agarose column allowed the isolation of a transferrin-binding protein with an estimated molecular mass of 37 kDa.

Receptors for human transferrin have been demonstrated for a variety of pathogenic bacteria (13, 14, 32, 33). Their molecular masses appear to be variable, as a receptor of approximately 102 kDa has been reported for Neisseria gonorrhoeae (13), whereas in Borrelia burgdorferi, the binding of transferrin involves a protein of 28 kDa (3). The best-characterized mechanism of iron acquisition involving cell surface transferrin-binding proteins concerns N. gonorrhoeae (4, 6, 13, 24). It has been demonstrated that two proteins (TbpA [∼102 kDa] and TbpB [∼85 kDa]) of the outer membrane are involved in the binding of transferrin and that the iron is removed from the transferrin in an energy-dependent process. A third protein (FbpA [∼33.5 kDa]) acts as a shuttle vector, transporting the iron through the periplasm to the cytoplasmic membrane. Transferrin-binding activity has been also observed for the oral bacterial species P. gingivalis (31) and Streptococcus oralis (1). Tazaki et al. (31) reported the capacity of P. intermedia to bind human transferrin but did not investigate further the activity.

We observed that P. nigrescens and P. intermedia could degrade human transferrin. This finding is in agreement with a previous study by Jansen et al. (18), who reported the degradation of various serum proteins, including transferrin, haptoglobin, albumin, and immunoglobulins. Moreover, P. intermedia has been shown to possess on its cell envelope a 31-kDa serine protease with elastase-like activity (29). Since the degradation of transferrin by P. nigrescens and P. intermedia occurred mostly at the end of growth (stationary growth phase), it is likely that this activity plays only a minor role in the ability of the bacteria to grow in the presence of this plasma protein as the source of iron.

Our study also revealed that lactoferrin could serve as a source of iron for P. nigrescens and P. intermedia. Lactoferrin-binding activity was previously demonstrated for both species (8, 9, 20). However, the relationship between the presence of this activity in the bacteria and their ability to grow in the presence of lactoferrin as a source of iron has not been established. The lactoferrin-binding protein present in the outer membrane of P. nigrescens was reported to have a molecular mass of 40 kDa (9), close to the 37-kDa transferrin-binding protein demonstrated in the present study. We found that no inhibition of transferrin-binding activity occurred in the presence of lactoferrin, suggesting the involvement of two different proteins. P. nigrescens is also known to possess receptors for immunoglobulin G, laminin, and fibrinogen (19, 21, 23). Once again, no inhibition of binding of transferrin was observed when bacteria were preincubated with these molecules, suggesting that different receptors participate in binding.

In summary, we showed that P. nigrescens and P. intermedia have the capacity to use human transferrin as a source of iron and possess transferrin-binding activity. A 37-kDa protein with transferrin-binding activity was identified on the surface of P. nigrescens. However, further studies are required to demonstrate that this protein is a receptor specific for human transferrin. P. nigrescens and P. intermedia also demonstrated the capacity to proteolytically cleave transferrin. Both activities may permit the bacteria to obtain iron for their survival and growth in periodontal pockets.