Vol. 51 No. 1/2004 253–262 QUARTERLY

We have previously reported on the identification and characterization of the Porphyromonas gingivalis A7436 strain outer membrane receptor HmuR, which is involved in the acquisition of hemin and hemoglobin. We demonstrated that HmuR interacts with the lysine- (Kgp) and arginine- (HRgpA) specific proteases (gingipains) and that Kgp and HRgpA can bind and degrade hemoglobin. Here, we report on the physiological significance of the HmuR-Kgp complex in heme utilization in P. gingivalis through the construction and characterization of a defined kgp mutant and a hmuR kgp double mutant in P. gingivalis A7436. The P. gingivalis kgp mutant exhibited a decreased ability to bind both hemin and hemoglobin. Growth of this strain with hemoglobin was delayed and its ability to utilize hemin as a sole iron source was diminished as compared to the wild type strain. Inactivation of both the hmuR and kgp genes resulted in further decreased ability of P. gingivalis to bind hemoglobin and hemin, as well as diminished ability to utilize either hemin or hemoglobin as a sole iron source. Collectively, these in vivo results further confirmed that both HmuR and Kgp are involved in the utilization of hemin and hemoglobin in P. gingivalis A7436.

P. gingivalis to acquire heme are still not well understood.Several reports have described P. gingivalis genes (ihtA, hemR, ragA, tla and tlr) that exhibit homology to genes encoding putative TonB-dependent heme/hemoglobin receptors (Aduse-Opoku et al., 1997;Karunakaran et al., 1997;Hanley et al., 1999;Dashper et al., 2000;Slakeski et al., 2000).We have recently reported on the identification and characterization of the outer membrane heme/hemoglobin receptor, HmuR (Simpson et al., 2000;Olczak et al., 2001).P. gingivalis hmuR mutant was shown to be defective in growth in the presence of both hemin and hemoglobin and exhibited a decreased ability to bind hemin and hemoglobin (Simpson et al., 2000).Furthermore, the recombinant HmuR protein itself and Escherichia coli cells expressing outer membrane-associated recombinant HmuR were capable of binding hemin and hemoglobin (Olczak et al., 2001).
Several reports documenting the binding of hemin and/or hemoglobin by Kgp have employed kgp mutants.However, in these studies, the ability of the defined kgp mutants to utilize various iron sources was not examined.In addition, previous studies in our laboratory, which revealed the interaction of Kgp with HmuR, were performed using purified proteins (Olczak et al., 2001).Thus, the physiological significance of the binding and degradation of hemoglobin by Kgp and the formation of the HmuR-Kgp complex in intact P. gingivalis cells have not been established.In this study, we report on the significance of this interaction to P. gingivalis by the construction and characterization of defined mutants in the kgp and hmuR genes in the A7436 strain.

SDS/PAGE and Western blot analysis.
The absence of the Kgp protein in WS10 and WS15 was verified by Western blot analysis.P. gingivalis A7436, WS1, WS10, and WS15 cultures were harvested at A 660 = 2.0 and adjusted to an A 660 of 1.0.One milliliter of each culture was then removed and centrifuged (1 min, 14 000 r.p.m., room temp.).The cell lysates were examined by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS/PAGE) on 12% gels.The proteins were transferred onto nitrocellulose membranes (BioRad, Hercules, CA, U.S.A.) and Western blotting with anti-Kgp serum (kindly provided by Dr. Jan Potempa, Jagiellonian University, Cracow, Poland) was performed as described previously (Olczak et al., 2001).Polyclonal antibodies were raised against purified Kgp and IgG fraction was isolated from the resulted antiserum (Olczak et al., 2001).
Arginine-X and lysine-X enzymatic assays.The arginine-and lysine-specific protease activities of P. gingivalis A7436, WS1, WS10, and WS15 in whole cell cultures (bacterial suspension was previously adjusted to an A 660 of 1.0), containing cell-associated and secreted gingipains, were determined with either N-benzoyl-L-arginine-p-nitroanilide, or N-p-tosyl-glycine-proline-lysine-p-nitroanilide as substrates.Samples were preincubated in 200 mM Tris/HCl buffer, pH 7.6, containing 100 mM NaCl, 5 mM CaCl 2 , and 10 mM cysteine for 5 min at 37°C and assayed for amidase activity with 0.5 mM substrates.The formation of p-nitroanilide was monitored spectrophotometrically at 405 nm.

Hemoglobin and hemin binding assays.
To determine if WS10 and WS15 were affected in their ability to bind different iron sources, the binding of human hemoglobin or hemin to the cells was examined by a spectrophotometric assay.P. gingivalis wild type (A7436) and mutant (WS1, WS10 and WS15) cells were prepared as described previously (Simpson et al., 2000;Olczak et al., 2001).The percent binding of hemin and hemoglobin to P. gingivalis whole cells was determined spectrophotometrically by the decrease of absorbance at 400 nm of the supernatant of the mutant cells compared to the wild type (Simpson et al., 2000;Olczak et al., 2001), which was arbitrarily set at 100%.
Growth experiments.To study the ability of P. gingivalis to grow with different iron sources, wild type A7436 and mutant WS1, WS10, and WS15 strains were prepared and examined as described previously (Simpson et al., 2000).
Statistical analysis.Data expressed as mean ± standard deviation (±S.D.) were analyzed using the Student's t-test; P values below 0.05 were considered significant.

RESULTS AND DISCUSSION
Production of an isogenic kgp mutant and a hmuR kgp double mutant Previously (Olczak et al., 2001) we showed in vitro using ELISA assay that recombinant HmuR interacted with purified Kgp and HRgpA.To test the importance of P. gingivalis HmuR-Kgp interaction in vivo, first we constructed and characterized a kgp mutant in A7436 strain.To date, kgp mutants were constructed in P. gingivalis strains other than A7436 (Nakayama et al., 1998;Okamoto et al., 1998;Aduse-Opoku et al., 2000).This allowed us to compare the single hmuR mutant constructed previously (Simpson et al., 2000) with the single kgp and double hmuR kgp mutants all constructed in P. gingivalis A7436.
The specific probe used in Southern blot analysis hybridized to a 3.3 kb fragment in PstI-digested genomic DNA of A7436 and WS1 strains, which contain intact kgp genes (not shown).In WS10 and WS15 strains, the probe hybridized to a 4.8 kb fragment encompassing the region into which the two cat cassettes had been inserted (not shown).In the case of the second specific probe hybridization bands of 8 kb were observed in strains A7436 and WS10.Further verification showed that the hmuR-specific primers amplified an 855 bp fragment from the intact hmuR gene of A7436 and WS10 and a 3.2 kb fragment from the disrupted hmuR gene in strains WS1 and WS15 (not shown).The kgp-specific primers amplified a 2.4 kb fragment from the intact kgp gene of A7436 and WS1 and a 3.9 kb fragment from the disrupted kgp gene of strains WS10 and WS15 (not shown).

P. gingivalis kgp and hmuR kgp mutant cells are devoid of Kgp expression and lysine-specific protease activity
To confirm that the disruption of the kgp gene resulted in translational effects, we tested WS10 (kgp mutant) and WS15 (hmuR kgp mutant) strains for Kgp protein production and lysine-specific proteinase activity.We did not observe reactivity to the hemagglutinin domains of Kgp (39, 27, 17, and 15 kDa) in WS10 and WS15 strains (Fig. 1).We did, however, detect reactivity with additional proteins in these strains, which most likely represents the cross reactive epitopes present in HRgpA and HagA (Han et al., 1996;Potempa et al., 1998).Interestingly, we did not observe reactivity of the Kgp-specific antibodies with the 48 kDa catalytic domain of Kgp.This finding is in accordance with previous reports (Genco et al., 1995;Potempa et al., 1997), which demonstrated that immunization with purified gingipains generated a major IgG response targeted to epitopes within the hemagglu-tinin/adhesion domains, but only a very weak response against the catalytic domain.As expected, the P. gingivalis wild type and WS1 strains expressed proteins corresponding to the intact Kgp protein (Fig. 1).
As shown in Table 1, WS10 and WS15 strains were found to be essentially devoid of lysine-specific proteinase activity as compared to the parental A7436 strain and to the isogenic hmuR mutant (WS1).Examination of WS10 and WS15 for arginine-specific ac-tivity revealed that these strains exhibited activities similar to the wild type strain.This is in agreement with results presented by Aduse-Opoku et al. (2000) demonstrating that a P. gingivalis kgp mutant made in W50 strain produced arginine-specific activity at a level comparable to the wild type strain.Our previous examination of a different kgp mutant, MSM-3, revealed that it exhibited increased transcription of both rgpA and rgpB (Genco et al., 1995).As the kgp mutation in MSM-3 arose due to spontaneous mobilization of an endogenous insertion sequence element (IS1126) (Simpson et al., 1999), it could not be definitively stated whether this increase in Rgp activity was a compensatory mechanism brought about due to the absence of Kgp from the protease population, or if a regulator of the rgpA and rgpB genes had been unknowingly affected.In contrast, studies conducted by Tokuda et al. (1998) showed that mutation of the rgpA gene resulted in a decreased transcription of the kgp gene.

Hemoglobin and hemin binding is diminished in P. gingivalis strains WS10 and WS15
Studies performed in our laboratory (Olczak et al., 2001) have demonstrated that HmuR is capable of interacting in vitro with purified Kgp.In the current study, we have examined the biological significance of the interaction After SDS/PAGE of whole P. gingivalis cell lysates (bacterial cells were adjusted to the same A 660 ), proteins were transferred onto nitrocellulose membrane and probed with anti-Kgp antibodies.Kgp hemagglutinin domains (HA) are shown in kDa on the right.

Table 1. Arginine-and lysine-specific protease activities* of P. gingivalis
*Activity was determined in 1 ml of whole bacterial culture adjusted to the same absorbance and expressed as milli-A units per minute for each substrate.Results are representative of two separate experiments, each performed in triplicate.A7436, wild type strain; WS1, hmuR mutant; WS10, kgp mutant; WS15, hmuR kgp mutant.
of Kgp with HmuR with regard to heme utilization by P. gingivalis.Examination of the hemoglobin binding activities of the isogenic hmuR (WS1) and kgp (WS10) mutants and the hmuR kgp double mutant (WS15) revealed that the removal of these proteins, either singly or in combination, significantly diminished the ability of the organism to bind both hemin and hemoglobin (Fig. 2).When the hemin binding capabilities were examined, it was revealed that all the mutant strains were significantly diminished in their ability to bind this iron source as compared to the wild type strain.Interestingly, the hmuR kgp double mutant was not completely diminished in its ability to bind hemoglobin and hemin.This indicates that the ability of P. gingivalis to bind hemoglobin and/or hemin is also attributed to proteins other than HmuR and Kgp.Studies in our laboratory (Olczak et al., 2001;Sroka et al., 2001) as well as others (Okamoto et al., 1998) have indicated a role for HRgpA in this binding.Thus, HRgpA present on the outer membranes of WS10 and WS15 cells, as well as other putative hemoglobin and/or hemin binding proteins, may be involved in binding of these iron sources with lower efficiency as compared to Kgp.The probability that multiple proteins comprise the hemoglobin-binding machinery of P. gingivalis correlates well with the involvement of several hemoglobin binding proteins reported for other Gram-negative microorganisms (Lewis & Dyer, 1995;Lewis et al., 1997;Morton et al., 1999).

Growth analysis of P. gingivalis strains WS10 and WS15
The A7436 strain cultured in Schaedler broth plus dipyridyl supplemented with hemin, hemoglobin, and ferric chloride exhibited a typical growth pattern (Fig. 3A and 3B).In contrast, the initial growth of WS10 with hemoglobin as a sole iron source was delayed when compared to the wild type strain; however, WS10 was capable of growth at late time points.Interestingly, WS10 exhibited a diminished ability to utilize hemin as a sole iron source (Fig. 3A).Growth of the hmuR kgp double mutant WS15 was diminished with either hemin or hemoglobin as sole iron sources (Fig. 3B).This finding correlates well with the growth pattern of the isogenic hmuR mutant, strain WS1, which is deficient in its ability to utilize either hemin or hemoglobin as sole iron sources (Simpson et al., 2000).These results are also in agreement with the diminished ability of another kgp mutant, MSM-3, to utilize hemin (Genco et al., 1995).Similarly to P. gingivalis strain WS10, the MSM-3 strain was capable of utilizing hemoglobin as a sole iron source (Genco et al., 1995).We demonstrated that both WS10 and WS15 strains grew with ferric chloride, similarly to the wild type and WS1 mutant strains (Fig. 3A and B).This indicates that the uptake of this iron source in P. gingivalis occurs via a mechanism(s) which is independent of either Kgp or HmuR.In conclusion, our results indicate that the formation of the Kgp-HmuR complex might increase the overall efficiency of heme acquisition in P. gingivalis.We observed that following prolonged growth on blood agar plates P. gingivalis hmuR mutant is characterized by higher pigmentation than the wild type A7436 strain (Simpson et al., 2000).This effect may be due to excessive heme storage on the cell surface, likely trough Kgp, and an inability to uptake the heme moiety into the cell due to the absence of HmuR.P. gingivalis kgp mutant was non-pigmented, as it was shown for the kgp mutants constructed in other P. gingivalis strains (Nakayama et al., 1998;Okamoto et al., 1998;Aduse-Opoku et al., 2000).P. gingivalis hmuR kgp double mutant was also non-pigmented, suggesting that these cells cannot store and/or use heme due to the absence of Kgp and HmuR.Based on the results presented here it is likely that Kgp is mostly involved in hemoglobin and/or hemin binding and delivering heme to HmuR, and HmuR is engaged in heme transport into the cell.

CONCLUDING REMARKS
Based on the results presented in this report we propose that HmuR alone might be sufficient for the binding and internalization of hemin and hemoglobin.Kgp has been demonstrated to bind (Kuboniwa et al., 1998;Nakayama et al., 1998;Okamoto et al., 1998;DeCarlo et al., 1999;Lewis et al., 1999;Shi et al., 1999;Olczak et al., 2001) and degrade hemoglobin (Lewis et al., 1999;Sroka et al., 2001), a process by which heme can be liberated.We assume that the uptake of heme in P. gingivalis may require the binding of heme to Kgp, and the subsequent binding of Kgp by the receptor HmuR.These collective findings led us to hypothesize that soluble Kgp and also in lesser degree HRgpA could function as heme scavengers or hemophore-like proteins, similarly to the hemophore HasA protein secreted by Serratia marcescens (Letoffe et al., 1994;Ghigo et al., 1997).HasA is capable of

Figure 2 .
Figure 2. Binding of hemoglobin and hemin to P. gingivalis A7436 wild type (WT) and WS1 (hmuR mutant), WS10 (kgp mutant), and WS15 (hmuR kgp double mutant) strains.P. gingivalis cells were resuspended in phosphate buffered saline, adjusted to an A 660 of 1.0, and incubated for 1 h at room temp.with hemin (30 mM) or hemoglobin (3 mM).The percent binding of hemin and hemoglobin to P. gingivalis whole cells was determined by the decrease of absorbance at 400 nm of the supernatant of the mutant cells compared to the wild type strain, which was arbitrarily set at 100%.Data were analyzed using the Student's t-test (P values below 0.05 were considered significant) and are shown as the mean ±S.D. from three independent experiments, each performed in triplicate.*P < 0.05, **P < 0.005 (WS1, WS10, and WS15 mutants versus WT); Hb, hemoglobin; Hm, hemin; WT, black bars; WS1, striped bars; WS10, white bars; WS15, grey bars.