of Porphyromonas gingivalis strains �

Synthetic inhibitors of benzamidine type have been found to have inhibiting effects on arginine specific cysteine proteinases of P. gingivalis. The purpose of our study was to assess the effects of these inhibitors on the virulence properties of two P. gingivalis strains, the reference strain ATCC 33277 and JH16-1, a clinical isolate obtained from a patient with severe periodontitis. The inhibitors tested were pentamidine, benzamidine, three bis-benzamidine derivatives with a pentamidine-related structure, one bis-benzamidine derivative with another structure, and one arginine derivative as a negative control, each in the concentrations of 2 microM and 20 microM. As virulence criteria the following parameters were determined: arginine-specific amidolytic activity, growth inhibition, hemagglutination of sheep erythrocytes, adherence to KB cells and immuno-phagocytosis including intracellular killing. Pentamidine and the bis-benzamidine derivatives with pentamidine-related structure showed the most remarkable effects on reduction of amidolytic activity by 35%, growth inhibition and reduced hemagglutination. Except for the arginine derivative all other inhibitors tested enhanced the phagocytosis capacities of granulocytes. No clear influence of the inhibitors on adherence of P. gingivalis to KB cells was seen. Although in vitro effects of the synthetic inhibitors of cysteine proteinases on virulence of P. gingivalis were observed further in vitro tests concerning immunomodulatory effects should be done before these substances are used for therapy in clinically controlled studies.

Porphyromonas gingivalis has been implicated as an important bacterial species in pathogenesis of severe forms of periodontitis.In the last few years a major focus of research on P. gingivalis has been on the cysteine proteinases with arginine and lysine cleavage specificity (Travis et al., 1997;Imamura, 2003).The so-called gingipains as a whole make up 85% of the proteolytic activity of this bacterium, and the molar concentration of Arg-gingipain is consistently 2-3-fold higher than that of Lys-gingipain (Potempa et al., 1997).While the cysteine proteinase activity with arginine specificity originates from two different genes, rgpA and rgpB, the lysine specific cysteine proteinase activity is derived from a single gene, kgp (Mikolajczyk-Pawlinska et al., 1998).rgpB mutants possess about 50% of the arginine specific cysteine activity (Aduse-Opoku et al., 1998;Tokuda et al., 1996), a rgpA mutant has 40% activity (Tokuda et al., 1998).The gingipains are associated with many of the virulence factors of this species, such as degradation of immunoglobulins G and A (Abe et al., 1998), degradation of complement factors (Grenier, 1992), enhancement of vascular permeability (Rubinstein et al., 2001), and formation of fimbriae (Tokuda et al., 1998).
Antibiotics are widely used as additive therapy in severe cases of P. gingivalis associated periodontitis.But this usage has limitations, e.g. because of the development of antibiotic resistance.Most P. gingivalis strains are carriers of the tet(Q) resistance gene (Chung et al., 2002).Therapy with tetracycline fibers is unable to eradicate P. gingivalis completely (Mombelli et al., 2002).New ideas to use inhibitors of major virulence properties have been discussed (Potempa et al., 2000).The purpose of this study was to evaluate the effect of inhibitors of benzamidine type on the virulence properties associated with arginine specific cysteine proteinase activity of P. gingivalis.Beside the influence on N-a-benzoyl-DL-arginine-p-nitroanilide (BApNA) activity the effect on growth, hemagglutination, adherence to epithelial cells, and immunophagocytosis were determined.

MATERIALS AND METHODS
Bacterial strains.P. gingivalis ATCC 33277 was obtained from the German strain collection DSMZ (Braunschweig, Germany).The JH16-1 strain was a clinical isolate obtained from a female patient with severe chronic periodontitis.Analysis for strain diversity confirmed that the clinical isolate had been persistent for two years in this patient (Eick et al., 2002).The strains were maintained on Schaedler agar enriched with 10% sheep blood and vitamin K. Bacterial cells used in assays were grown to late log phase (36 h), harvested, washed twice and resuspended in Medium 199 (Gibco) to an absorbance at 660 nm corresponding to 10 8 or 10 9 bacteria/ml.
Proteinase inhibitors.Seven inhibitors tested were chosen from a spectrum of substances with a benzamidine structure, they were named 1, 2, 3, 4, 6, 7 and 9 (pentamidine).The inhibitors 1, 4, and 6 were bis-benzamidines with a pentamidine related structure.Inhibitor 2 was a bis-benzamidine with another structure.Inhibitor 7 represented benzamidine."Inhibitor" 3 was an arginine derivative.The tests were carried out as a blind study.At the time of the study the investigators did not know either the group of benzamidines the inhibitors belonged to or that a negative control was included.
The inhibitors were synthesized by the group of J. Stürzebecher (Institute of Vascular Biology and Medicine, University Hospital of Jena, Germany).The inhibitors had been tested against purified Arg-gingipain RgpB (the enzyme was kindly provided by Jan Potempa, Jagiellonian University, (Kraków, Poland) before.The inhibition constant (K i ) was up to 0.45 mM (inhibitor 1).
Determination of the arginine specific cysteine amidolytic activity.Proteinase inhibitors were added to bacterial suspensions (10 9 /ml) in a final concentration of 2 mM and 20 mM.After mixing these suspensions were incubated at 37°C for 1 h.After that each 100 ml of suspension was placed into a well of a 96 well-microtitre plate.Each well contained 125 ml of reaction mixture for determination of the arginine specific cysteine amidolytic activity (0.5 mM BApNA, 10 mM L-cysteine, 10 mM CaCl 2 and 100 mM Tris/HCl, 1% agar, pH 7.6, according to Grenier & Turgeon (1994)).
The plate was incubated at 37°C for 1 h.Then the bacterial suspension was removed and the plates were washed three times with phosphate-buffered saline (PBS).The arginine specific amidolytic activity was measured by means of a spectrophotometer at 405 nm.
Growth inhibition by proteinase inhibitors.Bacteria (100 ml of a 10 8 /ml suspension) were added to tubes containing 10 ml of Schaedler broth with 10% sheep blood and vitamin K and a proteinase inhibitor (final concentration 0.05-20 mM), including a negative control without inhibitor.After incubation for 24 h in an anaerobic atmosphere at 37°C colony forming units (cfu) were determined.The MIC 50 and MIC 90 of the inhibitor were calculated.
Hemagglutination.For hemagglutination assays to bacterial suspensions (10 9 /ml) proteinase inhibitors were added (final concentration 2 mM and 20 mM) and incubated for 1 h.After that the mixtures were diluted in a twofold series with PBS.Aliquots (100 ml) of each dilution were mixed with an equal volume of a sheep erythrocyte suspension (2% in PBS) and incubated in a round-bottomed microtitre plate at room temperature for 3 h.Hemagglutination was evaluated visually.
Adherence assay.KB cells were seeded into two 24-well tissue culture plates at a density of about 10 4 cells per well, the cells were grown to confluent monolayer in EMEM (Gibco) enriched with 10% foetal bovine serum (FBS).The cells were washed twice with PBS.Bacterial suspension (10 8 /ml) in Medium 199 was mixed with EMEM 1:4.
Proteinase inhibitors were added to a final concentration of 2 mM and 20 mM and each well was inoculated with 1 ml of this suspension.Bacteria were co-cultured with KB cells at 37°C for 1 h.After that the infected monolayers were washed five times with PBS.The number of adherent P. gingivalis was counted as cfu after cell lysis.
Additionally the MTT tetrazolium salt colorimetric assay described by Mosmann (1983) was used to measure cytotoxicity.The method is based on the capacity of mitochondrial enzymes of viable cells to reduce the yellow soluble salt MTT to a blue insoluble formazan precipitate, which is quantified after dissolution in an organic solvent.The MTT-formazan production correlates directly with the count of viable cells.Proteinase inhibitors were mixed with EMEM + 10% FCS to a final concentration 2 and 20 mM and incubated with KB cells for 24 h in a 96-well microtiter plate.Then 10 ml of a MTT solution (5 mg/ml) was added to 100 ml of culture media.After incubation for 2 h at 37°C the suspension was removed and 20 ml of 3% SDS solution and 100 ml of isopropanole/0.04M HCl were added and incubated for 1 h at 37°C.Finally the MTT reaction was examined photometrically at 570 nm.
Freshly drawn blood obtained from 10 healthy subjects was layed on a cover slip.The cells adhered at 37°C with 5% CO 2 and 100% humidity for 25 min.Thereafter the soft blood clot was removed.
Bacterial suspensions were adjusted photometrically to 10 9 bacteria per ml with PBS.Bacterial suspension (0.2 ml) was mixed with 0.2 ml human Anti-AB-serum for opsonization for 30 min.These mixtures were centrifuged at 2000 ´g for 10 min and then the pellets were washed and resuspended in 1 ml PBS.Finally, 0.2 ml of foetal bovine serum and the proteinase inhibitor to be tested at a final concentration of 20 mM were added.The bacterial suspensions were transferred to the cover slips and the phagocytosis assay was incubated under the conditions mentioned above for 15 min.
Then the slips were stained with acridine orange (2.5 mg/7.5 ml PBS) for 1 min.After decanting the solution they were covered with 0.01% crystal violet in PBS to quench extracellular fluorescence, washed and fixed.A total of 100 polymorphonuclear leukocytes (PMN) were immediately examined by fluorescence microscopy.Viable and killed bacteria in granulocytes were distinguished by their uptake of acridine orange, viable bacteria appeared green and dead bacteria were red.The numbers of PMNs containing bacteria were counted.These phagocytosing cells were separated into groups by the following criteria: (i) cells with less than 10, cells with 10-20, and cells with more than 20 ingested bacteria and (ii) granulocytes with less than 50% and with more than 50% viable bacteria.
Statistical analysis.The statistical analysis was done using SPSS 10.0 software.The significance of the differences between each group with addition of an inhibitor (2 mM and 20 mM) and the controls (without inhibitor) was determined by the Mann-Whitney test.
The WILCOXON test was used to assess the significance of the results between paired samples in phagocytosis assays.

Amidolytic activity
Inhibitor 1 showed the highest reduction of the arginine cysteine amidolytic activity of the two P. gingivalis strains, the reduction being 35.5% for the ATCC strain at 2 mM.Also inhibitor 4, another pentamidine-related inhibitor, reduced the amidolytic activity of the two strains in the lower concentration of 2 mM.Inhibitor 6 in both concentrations showed effects on the JH16-1 strain, and pentamidine at 20 mM reduced the amidolytic activity of both strains (Table 1).

Growth curves
Only the inhibitors of the pentamidine-related group were able to inhibit the growth of the bacteria by 50% and 90% after 24 h.Pentamidine was most effective.No significant differences in growth inhibition was observed between the two strains (Table 2).

Table 1. Effect of proteases inhibitors on BApNA-reaction activity produced by P. gingivalis.
Activity of the non-treated suspension of bacterial cells was taken as 100%.

Hemagglutination
Without the addition of inhibitors hemagglutination was observed up to a dilution of 1:32 for the ATCC 33277 strain and up to 1:64 for the JH16-1 strain.Inhibitors 3 and 7 did not influence hemagglutination.Both tested concentrations of the pentamidine-related substances (1, 4, 6, and pentamidine) inhibited hemagglutination of the ATCC strain.The effects on the clinical isolate were not so clear.Substance 1 at concentrations of 2 mM and 20 mM of the substance 1 reduced hemagglutination of this strain, only the higher concentration of the other pentamidine-related inhibitors had a slight effect on hemagglutination.

Adherence to KB cells
Following incubation with KB cells in medium 2.2 ´10 5 cfu per well of adherent P. gingivalis ATCC 33277 and 1.5 ´10 4 cfu of adherent P. gingivalis JH16-1 were enumerated.Only pentamidine and inhibitor 7 at 20 mM were able to reduce the adherence of the ATCC 33277 strain, conversely, inhibitor 4 in both concentrations tested enhanced the number of adherent bacteria (P < 0.05).A reduced adherence of the clinical isolate JH16-1 strain was observed after addition of inhibitors 1, 2, 3, and 7 in both concentrations and of the pentamidine-related inhibitors 4 and 6 at 20 mM (P < 0.05, Fig. 1).
The MTT values showed no cytotoxicity for the inhibitors with a pentamidine-related structure.Reduced MTT values were measured for inhibitor 2 in both concentrations and the inhibitor 3 in the higher concentra-tion.So it can be concluded that these inhibitors are cytotoxic in the concentrations mentioned above.

Immunophagocytosis
No differences of phagocytosing capacities of peripheral blood PMNs were observed between the two bacterial strains without inhibitors added.So, in medium 59% of the PMNs had ingested more than 20 P. gingivalis ATCC 33277 or JH16-1.Inhibitor 3 (arginine derivative) was the only substance tested with no effect on phagocytosis.All other inhibitors enhanced the percentage of PMNs with more than 20 ingested P. gingivalis ATCC 33277 (P < 0.05).In contrast, the inhibitor 6 did not influence the phagocytotic capacity of granulocytes to the JH16-1 strain.Pentamidine and inhibitors 1, 2, 4 and 7 promoted the phagocytosis of this clinical isolate (P < 0.05).The results are presented in Fig. 2.

DISCUSSION
The purpose of the study was to determine the effects of inhibitors of benzamidine type on the bacteria themselves.The K i values were in the range from 10.4 to 0.45 mM.The effects were not so clear, if the bacteria as a whole and not the purified enzymes were assayed, but the inhibition of the amidolytic activity of the bacteria showed a good correlation with the K i values of the different inhibitors.Inhibition of the enzymatic activitity of the ATCC strain by 35% was found, the enzymatic activity of the clinical isolate was only reduced by 13% by one inhibitor.
Several tests have been performed to find an influence of the inhibitors on proteinases associated virulence properties.By cleaving a variety of host proteins, gingipains may provide small peptides and amino acids that meet the nutritional requirements of P. gingivalis.Consistently, Arg-gingipains play an important role in growth of this bacterium (Grenier et al., 2001).Bis-benzamidine derivatives, especially pentamidine, suppress the growth of P. gingivalis strains.But the effect of the The values are expressed as medians in per cent of the granulocytes with no, < 10, 10-20, and > 20 ingested bacteria.The values are expressed as medians (with the 25 and 75 percentiles) in per cent in relation to each P. gingivalis strain without the addition of an inhibitor.tested inhibitors on P. gingivalis growth was far more significant than inhibition of the bacterium-associated BApNA activity.This suggests a target for these compounds other than Arg-gingipain.
Both expression of fimbriae and hemagglutination are linked to the rgpA gene (Tokuda et al., 1998).A constructed rgpA, rgpB, kgp triple mutants exhibited no hemagglutinating properties using sheep erythrocytes and an rgpA rgpB mutant showed a reduced hemagglutinating activity (Shi et al., 1999).Although the inhibitors block only the arginine specific cysteine proteinases, the inhibition of hemagglutination was in accordance with the inhibition of enzymatic activity.
The results concerning the influence of the inhibitors on adhesion of bacteria to KB cells were contradictory, both an enhancement and a decrease were found.The arginine derivative (substance 3) and the benzamidine inhibitor 2 were cytotoxic to KB cells.In general it is believed that fimbriae represent the adhesins to receptors of epithelial cells.A strain with a double mutation of the rgp genes possessed very few fimbriae on the cell surface (Nakayama et al., 1996).Mutants of the fimA gene showed remarkably lower adherence and invasion in comparison with the parent strain (Njoroge et al., 1997;Weinberg et al., 1997).Recently it was reported that this mutant possesses only 5% of the Rgp and 13% of the Kgp total activities of the parent strain (Chen et al., 2001).Adherence to epithelial cells can occur in the absence of fimbriae.First, Tokuda et al. (1998) reported that an rgpA mutant expressed very few fimbriae but attached in high numbers.Later, Chen et al. (2001) described that the attachment level of an rgpA rgpB mutant was high.Detachment of P. gingivalis is caused by the degradation of receptors by Arg-gingipains (Chen et al., 2001).Gingipains are also able to degrade epithelial junctional proteins (Katz et al., 2002).The quantity of adherence to epithelial cells is not crucial for the virulence of P. gingivalis strains, as clinical isolates often adhere in a lower number than the reference strain (Eick et al., 2002).Contact of P. gingivalis with epithelial cells represses the secretion of Arg-gingipain and Lys-gingipain (Park & Lamont, 1998).But proteases of P. gingivalis may be involved in the invasion process (Lamont et al., 1995).P. gingivalis persists and multiplies within epithelial cells (Madianos et al., 1996).Persistent P. gingivalis strains can trigger immune response by release of proinflammatory interleukins (Eick et al., unpublished).Tests to determine the effects of the inhibitors on invasion, persistence and release of proinflammatory interleukins are in progress in this laboratory.
All the inhibitors tested enhanced the phagocytotic capacity of granulocytes to P. gingivalis.We used opsonized bacteria in the assays.So the benzamidines might block the degradation of immunoglobulins and factors of the complement.Gingipains cleave immunoglobulins (Abe et al., 1998).Previously it was decribed that gingipain degrades C3 and in this way eliminates the creation of C3-derived opsonins and renders P. gingivalis resistant to phagocytosis (Cutler et al., 1993;Schenkein et al., 1995).Furthermore, P. gingivalis alters expression of immunoglobulin G receptors on neutrophiles (Tai et al., 1993).Arg-gingipain reduces the respiratory burst in PMNs (Kadowaki et al., 1998;Abe et al., 1998).In our study the inhibitors did not enhance intracellular killing of P. gingivalis.Non-oxidative mechanisms are more important in killing periodontopathogenic bacteria in gingival sulcus (Miyasaki et al., 1994).Interestingly, we found a difference in killing of the strains used, the clinical isolate was more resistant than the ATCC strain.A strong case has been made for the development of synthetic inhibitors directed against specific pathogen-derived proteinases.Recently a report described an inhibitory effect on arginine-specific gingipains of tetracyclines and their analogues (Imamura et al., 2001).Agents, not in the first place antibiot-ics, were found to inhibit Arg-gingipains; malabaricone C, isolated from nutmeg, suppress growth of P. gingivalis (Shinohara et al., 1999), and protamines obtained from salmon or herring sperm inhibit fimbrial interaction with fibronectin (Kontani et al., 1999).Curtis et al. (2002) designed a very effective inhibitor of Lys-gingipains.We assayed several benzamidine derivatives.Although in vitro effects of the synthetic inhibitors of cysteine proteinases on virulence of P. gingivalis were observed further in vitro tests concerning immunomodulatory effects should be done before these substances are used for therapy in clinically controlled studies.These studies should focus on bis-benzamidine derivatives with pentamidine-related structure, especially inhibitor 1.