WO2000071682A2 - Crystal structure data of cysteinprotease gingipain r - Google Patents

Abstract

The objects of the present invention are a crystal structure of gingipain R as shown in (Fig. 1) and a crystal structure of a gingipain R in complex with H-D-Phe-Phe-Arg-chloromethylketone as shown in (Fig. 4). A further object is the use of such crystal structure to obtain crystal structure data for the design and/or identification of inhibitors of gingipain R, which again form part of the present invention, as well as pharmaceutical compositions containing such inhibitors.

Description

Crystal structure data of Cysteinprotease Gingipain R

SPECIFICATION

The present invention is concerned with crystal structure information obtained from crystalline preparations of gingipain R and its use for the design and/or identification of inhibitors of gingipain R. A further subject of the invention are gingipain R-inhibitor molecules and their use in pharmaceutical composititions for the treatment and/or prevention of periodontital diseases.

Periodontital diseases represent infections that are associated with inflammation of the gingiva, increased crevicular fluid, massive polymorphonuclear leukocyte infiltration of the gingival tissues, destruction and pocket formation of the periodontal tissue, and alveolar bone resorption. If untreated, they eventually lead to exfoliation of teeth, and are the major cause of tooth loss in industrial countries. In addition, severe periodontitis may predispose to more serious systemic conditions such as cardiovascular diseases and to the delivery of preterm infants (Page R.C. ( 1 998) Ann. Periodontol 3: 1 08-1 20). The severity of periodontitis correlates with the presence of specific bacteria, which trigger inflammatory host responses that, together with the bacterial virulence factors, cause the majority of tissue destruction (for ref. see Lamont, R.J. and Jenkinson, H.F. ( 1 998) Microb. Mol. Biol. Rev. 62, 1 244-1 263; Genco et al.,(1 998) Infect. Immun. 66, 41 08-41 14).

Porphyromonas gingivalis, a Gram-negative, anaerobic, black-pigmented bacterium, has been implicated as the major aetiologic agent in the initiation and progression of adult periodontitis (Holt et al., ( 1 988) Science 239, 55- 57) . It possesses a repertoir of virulence factors such as proteinases, fimbriae, lectin-type adhesins and hemagglutinating factors, which enable it to colonize in the gingival sulcus or the periodontal pockets (reviewed by Cutler et al., (1 995) Trends Microbiol. 3, 45-51 ; Lamont and Jenkinson, supra) . It is generally accepted that the proteolytic enzymes of this organism play a central role in pathogenesis of periodontitis. Although several different types of proteinases are expressed by P. gingivalis, cysteine proteinases with Lys-Xaa and Arg-Xaa i.e. "trypsin-like" specificity referred to as gingipains R (EC 3.4.22.37) and K, respectively, are responsible for at least 85% of the overall proteolytic activity and are recognized as the major virulence factors of this periodontal pathogen (Potempa and Travis, ( 1 996) Acta Biochim. Pol. 43, 455-466; Kuramitsu, H. K. (1 998) Oral Microbiol. Immunol. 13, 263-270).

Arg-gingipains (Rgps) occur either in soluble (RgpA, HRgpA, RgpB) or in membrane associated forms (mt-RgpA and mt-RgpB) and are products of two related genes, rgpA and rgpB (for a review, see Potempa et al., (1 998) J. Biol. Chem. 273, 21 648-21 657) . The rgpA gene-encoded gingipains are released into the medium as the single chain 50 kDa RgpA proteinase (alias RGP-1 or Rl-A) or as the high molecular mass HRgpA (alias HRGP or Rl), which is a non-covalent 95 kDA complex of a 50 kDa catalytic domain with a hemagglutinin/adhesin domain(s) derived from the initial rgpA gene product via proteolytic processing (Pavloff et al., (1 995) J. Biol. Chem. 270, 1007-1010; Rangarajan et al., ( 1 997) Mol. Microbiol. 23, 955-965). The related rgpB gene, in contrast, lacks the coding region of the hemagglutinin/adhesin domain and codes for a 507 amino acid residue protein (Mikolajczyk-Pawiinska et al., ( 1 998) Biol. Chem. 379, 205-21 1 ) ; its mature C-terminally truncated product RgpB (alias RGP-2) is a single chain protein of 48.3 kDa (of at least 435 residues) essentially identical to RgpA, with the C-terminal part from position 363 onwards showing a significant divergence, however (Potempa et al., ( 1 998), supra) .

The Rgps exhibit a strong specificity for endogenous Arg-Xaa peptide bonds, but are also capable of hydrolysing N-terminal Arg-Xaa and C-terminal Arg-Arg peptide bonds (Rangarajan et al., (1997) Biochem. J. 323, 701 -709) . Their hydrolytic activity requires activation by reducing agents such as cysteine and they are stabilized by calcium (Chen et al. , ( 1 992) J. Biol. Chem. 267, 1 8896-1 8901 ; Nakayama et al., ( 1 995) J. Biol. Chem. 270, 2361 9-23626), while Gly-Gly although stimulating the amidolytic activity enhances the inactivation rate (Potempa et al., (1998), supra). RgpB has a broad pH activity profile, with a broad pH optimum of pH 9 (Potempa et al., ( 1 998), supra; Rangarajan et al., ( 1 997a), supra) .

Besides providing nutrients for bacterial growth, the Rgps are implicated in the processing of their own pro-forms as well as of pro-Kgp and other housekeeping functions (reviewed by Potempa and Travis, (1 996), supra). They have been shown to be involved in the processing of the 75kDa major outer membrane protein and (via activation of the constituent pro-fimbrilin) in the maturation and assembly of the endogenous P. gingivalis fimbriae considered as another important virulence factor (Onoe et al., ( 1 996) Microbiol. Pathogenesis 1 9, 351 -364; Kadowaki et al., (1 998) J. Biol. Chem. 273, 29072-29076) . Furthermore, they can trigger bradykinin release through prekallikrein activation (Imamura et al., ( 1997) J. Biol. Chem. 272, 1 6062-1 6067), degrade complement factors such as C3 and C5 attracting neutrophils to the gingival lesion site (reviewed by DiScripio et al., (1 996) Immunology 87, 660-667), inactivate TNF-σ (Calkins et al., ( 1 998) J. Biol. Chem. 273, 661 1 -6614), and are capable of activating coagulation factors X and prothrombin (Imamura et al., (1 997), supra). Therefore, the Rgps affect host defense elements, attenuate neutrophil bactericidal activity, dysregulate coagulation, complement, and kallikrein-kinin cascades, and disrupt the local cytokine network (Lamont and Jenkinson, (1 998), supra).

Thus, the gingipains represent an unique family of cysteine proteinases, whose members play pivotal roles in the aetiology of periodontitis. In spite of a wealth of sequential and functional data, their fold and catalytic mechanism were completely unknown, however.

In view of the importance of gingipain R for the bacterial metabolism and in diseases caused by the bacteria due to the influence gingipain R exerts on the immune system of the patient, the conformation of the protein and the sites of interaction with its substrate are of particular interest. It was therefore an object of the present invention to identify the exact three-dimensional structure of the molecule to be able to gain conclusions on the substrate interaction sites, from which to derive for example highly potent inhibitors.

This object is solved by providing a crystal and molecular structure of gingipain R as shown in Fig. 1 .

The crystal structure is further described in the following disclosure of the invention.

Fig. 1 shows a ribbon plot of gingipain R (RgpB) (front view) . The molecule consists of the catalytic domain (top) subdivided into subdomains A (right) and B (left) and the IgSF (immuno globulin super family) domain (bottom). Strands are shown as yellow arrows, helices as red spirals, and the connecting segments as blue ropes. The residues of the catalytic triad on top are shown as stick models, and the three putative calcium as spheres.

A further subject of the invention is a crystal structure of gingipain R in complex with H-D-Phe-Phe-Arg-chloromethylketone as shown in Fig. 1 , whereby however H-D-Phe-Phe-Arg chloromethylketone is bound to the active site. A schematic drawing of the H-D-Phe-Phe-Arg chloromethylketone interaction with RgpB's active site is shown in fig. 2. Only by the present invention it was conceived that the chloromethylketone binds to this substrate specific active site of gingipain R and the active site could be examined in detail.

In the course of the present invention, the RgpB structure has first been determined by MAD techniques and refined with 1 .5 Λ data for a P2τ cell, with three para-chloro-mercury benzamidine molecules situated in the active site and at two other surface-located Cys residues (pCMBA, defined from amino acid residues Tyr1 to Ser435). With this model, the 2.1 6 A structure of the corresponding ligand-free RgpB was determined (NATI) . It furthermore served to solve the 2.0 A P2₁2_l2₁ structure (one copy per asymmetric unit) and the 2.9 A R3 structure (two copies per asymmetric unit) of a covalent RgpB complex with H-D-Phe-Phe-Arg- chloromethylketone, with only the first (FFRCMK, defined from Tyr1 to Glu432) being fully refined (see Table 1 ) . Upon optimal superposition, the monoclinic and orthorombic structures show rms deviations of up to 0.4 A for 428 alpha-carbon atoms. More significant deviations occur in the first three and the last three residues, respectively; the smaller C-terminal (immunoglobulin-like) domain deviates as a whole, probably reflecting a slight relative rigid body rotation between both domains presumably caused by different cell contacts.

The RgpB molecule has the shape of a crooked one-root "tooth", made up of an almost spherical "crown" of average diameter 45 A, and a pointed about 40 A long and 20 A thick "root" attached at an angle of about 30 deg (Fig. 1 ) . Its polypeptide chain forms (in consecutive order) the right- and the left-hand side of the crown part, before it turns down to the root. The crown formed by the N-terminal 351 residues represents the "catalytic domain", while the root made by the last 84 residues resembles an immunoglobulin superfamily (IgSF) domain. Both molecular parts are characterized by a high ratio of regular secondary structure elements, with the catalytic domain exhibiting the structural motif of a typical σ/β protein, and the IgSF domain having an all-β conformation. The seven Cys residues of RgpB are unpaired, with all but three (Cys 1 85, the catalytic Cys244 and Cys 299) exhibiting a buried side chain. With 29 Glu and 31 Asp residues opposing 31 Lys, 1 0 Arg and 8 His residues, RgpB shares a comparatively large ratio of charged residues, several of which are buried (see below); due to a large excess of negative residues, RgpB has a low isoelectric point. The molecular surface, in particular that of the catalytic domain, is relatively densely covered with charged residues. Noteworthy are two extended patches of negative electrostatic potential at the subdomain A front side and around the active site.

The catalytic domain of RgpB can be subdivided into the two subdomains A and B. Each subdomain comprises a central β-sheet flanked by helices on either side as characteristic for open σ/β sheets (Fig.3). The subdomain A sheet consisting of four parallel strands of order s2, s1 , s3 and s4, is regularly twisted in an (as usual) right-handed manner (if looked in strand direction) with the two outermost 7 strands, s2 and s4, arranged at an angle of about 45 deg to one another (Fig. 1 ). The flanking helices placed in front of (h i and h4) and behind (h2 and h3) this sheet run approximately parallel to one another, but antiparallel to the sheet strands. Most of these helices are quite regular σ-helices; only helices h3 and h4 represent a short 3₁₀ helix and exhibit a //-helical last turn, respectively. The two protein cores between the central β-sheet and the four sandwiching helices are mainly hydrophobic in nature; both cores are traversed by salt bridges/clusters, however, namely by the Arg 1 1 2...Asp93 bridge (back compartment) and the the salt bridge cluster Arg34...Glu 1 32...Lys31 (front part) . Between s3 and h3, a short hairpin loop is inserted, which projects from the top of the catalytic domain (Fig. 1 ).

Via a short extended surface-located segment connecting helix h4 and strand s5, the RgpB polypeptide chain runs over to the adjacent subdomain B (Fig. 3) . Central to this subdomain is the relatively flat β-sheet , made up of six strands arranged in the order (from the molecular periphery towards the center) s6, sS, s7, s8, s9, s10. While the first five are arranged parallel and show almost no twist to one another, the last innermost (s 10) strand runs antiparallel and is rotated for about 20 deg against the adjacent strand s9; to the other side of the sheet , strand s1 0 crosses the s4 edge strand of subdomain A sheet, making two hydrogen bonds at Phe339...Arg1 1 2 (under formal formation of a short parallel two-rung β-sheet) . Subsequent to strand s10, the polypeptide chain, after passing a short double loop between Val338 and Val344, traverses to the left side, where it enters the IgSF domain C (Fig. 1 ).

Subdomain B contains the seven helices h5 to hi 1 (Fig. 3). Except the short 3₁₀ helix h7, all of these helices are essentially cr-helical. Helix h9 is special in that its N-terminal Tyr283-Trp284-Ala285-Pro286-Pro287 part, positioned close to the active center (see below), adopts a principally σ-helix-like but more open conformation due to the incapability of Pro287 to form a hydrogen bond to the Tyr283 carbonyl group; furthermore, the last turn of this helix is somewhat wider approaching the geometry of a /7-helix. The central sheet is flanked by three helices (h6, h7 and h8, placed on the front side of the sheet) and four helices (hS, h9, h10 and h1 1 , positioned in the back of the sheet), with all helices but hi 0 running parallel to one another and antiparallel to the sheet strands, respectively. Between s7 and h7 an additional but short 3-stranded β-sheet with simple up-down-up geometry is inserted, which starts with the active-site His21 1 and ends at Thr224. Segment Ala263-Thr270 connecting h8 and s9 makes an extra hairpin loop. The active center is (as usual in such open β-sheet structures, see Branden and Tooze, (1 991 ), Introduction of Protein Structure, Garland Publishing, New York and London, p. 49) in a crevice outside of the carboxyl end of β-sheet B, with the active-site residues Glu1 52, His21 1 and Cys244 presented by the loops s5-h5, s7-h7 and s8-h8, which divirge to alternative sides (see Fig. 3a) . The majority of internal residues in the front side core is hydrophobic. However, there are a few internal water molecules and a number of internal Asn and Gin residues, with their carboxamide groups involved in hydrogen bonds. This front compartment also harbours an internal Glu1 67 with its carboxylate group, engaged in fully buried hydrogen bonds with Thr209 Oy and Cys147 Sy, and (arranged close to the active Cys244) an internal Asp241 forming a buried hydrogen bond to the carboxamide oxygen of Asn246, with both acidic side chains lacking any charge compensating group in their vicinity. The back compartment of this subdomain B lacks any internal water molecule and has a very hydrophobic core, which besides other hydrophobic residues contains five spatially adjacent Met side chains (of Mets294, 31 5, 31 8, 321 , 332) .

The interface between both RgpB subdomains A and B contains hydrophobic side chains, but also a number of buried water molecules, polar side chains and (partially) buried charged groups such as Glu6, and a cluster of charged residues (including the partially buried Arg262, the internal Glu258 and Asp78, and the surface located Glu1 1 6), which clamp both subdomains together via a central metal (probably calcium) ion KAL002.

This calcium KAL002 (Fig. 1 ) is in a roughly octahedral manner surrounded by eight oxygen ligands of two carboxγlates, one carbonyl group and three water molecules, as characteristic for calcium coordination spheres. Two further spheres of high electron density are interpreted to be calcium ions. These are KAL001 located in the back of subdomain B and KAL003 on the back surface of subdomain A (see Fig. 1 ) . KAL001 is particularly noteworthy in that it is placed just below the S1 specificity pocket, clamping helices h5 and h9 together, in this way presumably stabilizing this specificity determining binding site. It is well conceivable that removal of this metal/calcium ion causes a shift or disordering of Asp1 63, in this way affecting hydrolytic activity. In the structure of gingipain R in complex with the inhibitor H-D-Phe-Phe- Arg-chloromethylketone (FFRCMK), that is based on a crystallisation condition with 1 0OmM zinc in the crystallisation buffer, but not in the other two structures (NATI and pCMBA), a number of zinc sites are observed. ZIN001 is on the back of subdomain B, ZIN002 is within the three-stranded extra sheet, close to the catalytic His21 1 (see Fig. 3a), ZIN004 is bridging the catalytic Glu 1 52 carboxylate with His 21 1 Ne2 (see Fig. 4), and ZIN005 is in front of subdomain A. Ligands from symmetry related molecules contribute to the coordination of ZIN002 and ZIN004.

All bound metal ions fulfill the criteria of high electron density in combination with short distances to coordinating ligands and high coordination numbers. Except for KAL002, all metal sites are located on the surface of the catalytic domain.

Figs. 4 and 5 allow a view into the active center of RgpB shown in standard orientation, i.e. with a bound peptide substrate running from left to right. The active site and its immediate environment placed on the almost flat "masticating surface" of the crown is demarcated by strand His21 1 -Glu214 (bottom, Fig. 4), the (intervening) rising segment Ala243-Val245 (right), the (parallel aligned) long s9-h9 connecting segment (top), and the perpendicular running straight s5-h5 double-stranded loop (left). In the center resides the exposed Cys244, which in the MAD phased map immediately attracted attention due to the presence of the highest of the three mercury-related peaks accounting for pCMBA. Fromthe bottom strand ejects the His21 1 imidazole side chain, with its N<51 atom in all RgpB structures placed around 5.5 A distant from the Cys244 Sy but (in NATI and FFRCMK) linked through an ordered water molecule.

Towards the left of this Cys244 the deep S 1 pocket opens, which is covered by the indole moiety of Trp284 and harbours the Asp1 63 carboxylate group at its bottom as a charge anchoring point. The amidinophenyl part of the Cys244-Hg bound pHgBA extends into this pocket, with its amidino group shifted against the Asp1 63 carboxylate group, under formation of one N...O hydrogen bond. On its upper and lower sides, this pocket is lined by the side chain of Met288 and by segment Gly21 0-His21 1 and the His21 1 side chain, respectively. In the NATI and pCMBA crystals, the side chain of Glu1 52 points away from His21 1 . However, in the FFRCMK structure the Glu1 52 side chain points towards His21 1 , because both Glu 1 52 and His21 1 coordinate the bound ZIN004 from the crystallisation buffer.

In the FFRCMK structure, the peptidic inhibitor is bound to Cys244 forming through its methylene group a covalent bond with Cys244 Sy (Figs. 4 and 6) . Obviously, this bond results from a direct methylene attack or a preceding C(O)-C rearrangement occuring upon chlorine release after initial Sy-carbonyl bond (hemithioketal) formation. The ketone group oxygen forms favourable hydrogen bonds to Gly21 2 N and Cys244 N, which together function as an oxyanion hole. The CMK (chloromethyl ketone) peptidyl moiety juxtaposes RgpB segment Gln282-Trp284 in a nearly extended manner, forming a two-stranded twisted β-pleated sheet, with a favourable hydrogen bond between Gln282 0 and P1 -Arg3l N (3.0 A), and a longer one between Trp284 N and P3-Phe1 l O (4.1 Λ)(Fig. 6). The Arg3l side chain inserts into the S1 pocket with a kink, with its terminal guanidyl group sandwiched between the Trp284 indole moiety (lid) and the Val 242 and Thr209 Cy (base), with the Ne-H directed toward the Tyr283 carbonyl, and the two Uζ nitrogens pointing to the carbonyl groups of Trp 284 and Gly 21 0, besides making a frontal 2N....20 salt bridge with the carboxylate group of Asp 1 63. The S1 "slot" covered by hydrophobic lids and bordered by in-plane hydrogen bond acceptors certainly accounts for RgpB's strong preference for Arg residues at P1 /S1 . One of both guanidinium Nf nitrogens is close enough to Met288 Sδ to make a hydrogen bond interaction with one of its lone pair orbitals. In Kgp, this Met288 is probably replaced by a Tyr residue (according to sequence and alignment Pavloff et al., (1 997) J. Biol. Chem. 272, 1 595-1 600; Potempa and Travis, (1 996) supra), whose phenole might be rotated into the S1 pocket from its base presenting an extra hydrogen bond donor to the ammonium group of an inserted P1 -Lys; residue 288 thus might be a key element in discriminating between RgpB and Kgp. The L-Phe2l and the DPhel I side chain of the inhibitor extend out of the molecular surface, with their benzyl groups making edge-on-face contacts with one another and with the phenole group of the adjacent Tyr283 (Fig. 4).

Except for the entrance "hole" to the S 1 pocket, the molecular surface around the active site is relatively flat and characterized by a negative electrostatic potential (Fig. 5). Modeling studies show that a productively bound contiguous polypeptide substrate consisting of L-amino acids could bind in an FFRCMK-similar i.e. overall extended manner, stretching from the (pseudo-helical) N-terminal part of helix h9 (Ala285 N, Trp284 N) along segment Asp281 -Tyr283 to strand His21 1 -Glu214, between P5-S5 and P3'-S3' clamped to the proteinase by up to seven inter main-chain hydrogen bonds (P5 O...Ala285 N; P3 O...Trp284 N; P1 N...GIn282 0; P1 O...Cys244 N and Gly21 2 N; P2' N...GIy21 2 0; P3' O...GIu214 N; see Fig. 6). The P4 and P2 side chains could slot into hydrophobic-acidic surface depressions (S4 subsite shaped by the Tyr283, Pro286 and Asp328 side chains; S2 formed by the Tyr283 and Asp281 side chains), while the P3 side chain would be placed on top of the Trp284 indole moiety and the P1 ' side chain could make surface contacts with the Asp281 carboxylate. Thus, gingipain seems to bind its peptide substrates primarily via main chain interactions, with the exact cleavage site determined, however, by the quite selective P1 -Arg...S1 interaction assisted by P4...S4, P2...S2 and PI'...ST.

In this way, the carbonyl group of the scissile Arg-Xaa peptide bond of a bound substrate is presented in a rigid and stereochemically favorable manner to Cys244 for nucleophilic attack by Sy (Figure 5) . Assisted by the polarization of the P1 carbonyl in the oxyanion hole, Cys244 Sy could bind to the carbonyl carbon of the Arg-Xaa scissile peptide bond toward its Re face, under approach of the tetrahedral intermediate state. In this reaction, the attacking Sy lone pair orbital might be oriented toward the carbonyl by hydrogen bonding from the Gln282 N-H. The thiolate anion would certainly be a better nucleophile than the uncharged thiol group in forming the tetrahedral intermediate. However, at neutral pH the active centre exhibits quite a negative electrostatic surface potential (Figure 4) , and the Cys244 side chain is placed adjacent to the exposed carboxylate group of Asp281 , so that its thiol group presumably has a normal if not high pK. A putative acidic protonated sulfur of the hemimercaptal part created upon nucleophilic attack by the thiol group could easily transfer its proton to the spatially adjacent carboxylate group of Asp281 or to bulk water.

Simultaneously, the imidazole group of His21 1 positioned on the opposite side of the scissile bond could (even in a concerted move together with the

Glu1 52 carboxylate group) turn towards the pyramidalizing Xaa leaving group nitrogen. The negative electrostatic surface potential together with the properly placed Glu1 52 will probably stabilize this His21 1 imidazole in its protonated form, enabling it without previous proton transfer to donate a proton to the leaving group nitrogen, thus promoting the C-N break in the bound substrate and the release of the C-terminal substrate portion (see Fig.

5) . The remaining thiol ester, in turn, could be hydrolyzed in an inverse manner, i.e. with a water molecule positioned at the site of the former leaving nitrogen attacking the ester carbonyl and hydrolysing the ester bond, leading to the release of the amino-terminal part of the substrate under simultaneous transfer of both protons to the His21 1 imidazole and the Cys244 Sy, respectively.

Using the crystal structure data obtained from the crystalline preparation of gingipain R (RgpB) or the gingipain R-complex and described above in detail it became possible to detect the topology of active sites and the conformation of the gingipain R molecule and to further elucidate enzyme- substrate interaction sites. It is only based on this knowlegde of the topology and conformation of the gingipain R-moIecule that it became possible to design and produce substances that specifically interact with gingipain R to modulate and especially to suppress its ability to bind to substrate and/or its enzymatic activity.

In a preferred embodiment of the use according to the invention a library of chemical compounds is searched for molecules that fit the crystal structure of the active site of gingipain R disclosed in the present invention.

In another preferred embodiment of the use according to the invention, a computer aided modelling program is used to either design inhibitors that bind to the active site of gingipain R or to improve molecules that have been identified in a previous library screening step (see above). In many cases, molecules identified in the library screening step will have an arginine side chain or an arginine analogue that interacts with the S1 pocket of the enzyme, and in particular with the acidic Asp 1 63 at its bottom. Molecules that do not already have this feature can be improved by adding an arginine side chain or an analogue in the appropriate position. This position can easily be identified from the crystal structure of the enzyme and the inhibitor docking mode identified by computer-aided modelling. An example for this strategy will be given below, in the description of the design of the oxalic acid-1 -ethyl-ester-2-(3-amidino)anilide gingipain R inhibitor.

In yet another preferred embodiment of the use according to the invention, a computer aided modelling program is used to design inhibitors based on the binding mode of the H-D-Phe-Phe-Arg-chloromethylketone inhibitor, for which the binding mode is disclosed in the crystal structure of the present invention. 5A and B show computer simulations of the surface of gingipain R, to which H-D-Phe-Phe-Arg-chloromethylketone (fig. 5A) and a modelled artificial substrate (fig. 5B), -respectively, are bound. It can be seen that the arginine residues immerse deeply into the S1 pocket whereas other parts of the molecules remain on the surface and can interact with the surface molecules specifically or unspecifically . In one aspect therefore the inhibitor molecule according to the invention is designed to mimick the natural substrate and contain a group that becomes immersed into the S1 pocket thereby interacting with Asp1 63. According to this embodiment, a competitive binding takes place between natural substrate and inhibitor.

Such structure based inhibitor search and design allows to generate inhibitors that are superior to the known chloromethylketone inhibitors H-D- Phe-Pro-Arg-CH₂CI and H-D-Phe-Phe-Arg-CH₂CI (Potempa et al, Biol. Chem. 378, 223-230) . The known inhibitors, that are designed simply by combining a substrate moiety with a reactive group that binds irreversibly to the enzyme, suffer from cross-reactivity with human enzymes, e.g. trypsin and from susceptibility of the purely peptidic moiety to degradation. Rational drug design allows to generate inhibitors that overcome these problems.

A still further subject of the present invention are inhibitor molecules that interact with gingipain R in such a way that the proteolytic activity of the enzyme is at least substantially impaired. Such inhibitors can interact either covalently or noncovalently with the enzyme. Covalent binding occurs either reversibly or irreversibly, and is very efficient in inhibiting the activity of gingipain R.

Irreversible bonds to the sulfur in Cys 244 or other reactive groups in its vicinity can for example be formed by nitrites (formation of isothioamide), halomethylketones, diazomethylketones, acyloxymethylketones, methylsulfonium salts, epoxysuccinyl derivatives, vinylsulfones, O- acylhydroxamates, aziridines or activated disulfides.

Formation of covalent hydrolytically labile bonds between an inhibitor and gingipain R leads to reversible inhibition and can be accomplished by using for example aldehydes with thioacetal formation, methylketones and trifluoromethylketones under formation of tetrahedral hemiketals, or a- ketoacids, -esters or -amides or diketones. In the context of the present invention, such reversibly binding inhibitors are preferred over irreversibly binding inhibitors, because side reactions with other molecules that occur in vivo are expected to be less dramatic.

It is, however, equally possible that the inhibitor molecule is designed to non-covalently bind to gingipain R. Such non-covalent binding can be effected through reactive groups that are capable of forming salt bridges or hydrogen bonds with the anchor groups of gingipain R.

In one aspect of the invention, the inhibitor is designed to mimick the natural substrate and contains a group that becomes immersed into the S1 pocket, thereby interacting with Asp1 63. This group that occupies the S1 pocket can be an arginine residue as in the natural substrate, or it can be an argininomimetic, an unflexible aromatic residue or a saturated or unsaturated cyclic compound. Examples for argininomimetics are molecules that contain a basic head group linked to a spacer that has a similar extension as the -CH₂-CH₂-CH₂- in arginine, although other hydrogen donors can also be used. Suitable basic groups are

R being an organic residue.

Suitable spacers are (CH₂)_n groups with n = 2-7, saturated or unsaturated alkyl spacers, aryl spacers or cyclic or heterocyclic, saturated or unsaturated spacers. The cationic or basic group can be present in ω-position, but it also useful in side positions as for example in

or in a ring position as for example in pyridine, aminopyridine or

In the absence of a crystal structure, due to the high flexibility of the arginine side chain with three consecutive methylene groups, choice of proper arginine analogues and suitable spacers to bridge the distance to other anchoring sites on the gingipain R molecule such as the thiol/thiolate group of Cys 244 and the oxyanion hole is an extremely difficult task. The crystal structure solves this probelm, because it specifies these distances between anchoring groups and, in addition, indicates the required relative orientation of the respective funcional groups in an inhibitor molecule.

Fig. 4 shows a ball-and-stick model of the amino acids in the environment of the S1 pocket. Fig. 6 shows a schematic two dimensional view of the respective amino acids indicating the S1 pocket to which a substrate containing an arginine residue is bound. Figs. 5A and B show computer simulations of the surface of gingipain R, to which H-D-Phe-Phe-Arg- chloromethylketone (fig. 5A) and a modelled artificial substrate (fig. 5B), - respectively, are bound. It can be seen that the arginine residues immerses deeply into the S1 pocket.

In another aspect of the invention, the inhibitor molecule is designed in such a way that the opening of the S1 pocket is blocked without immersion of the inhibitor into the S1 pocket. This possibility is for the purposes of the present invention also to be subsumed under the description "bind to and/or interact with the active site at position Asp163". Such an inhibitor molecule would bind to and/or interact with groups neighbouring the S1 pocket and containing residues that more or less shield the S1 pocket, so that the natural substrate cannot bind and/or immerse anymore.

In a preferred embodiment of the invention, the inhibitor molecule does not only interact with and/or bind to Asp163, but also with one or more of Cys244, Asp281 , His21 1 , Trp284, Val242, Gly210, Tyr283 and Gly212. Some of these amino acids contribute to the S1 pocket. The carboxylate group of Asp163 is located at the very bottom of the pocket, Val242 forms a hydrophobic base via it's Cy, the indol group of Trp284 forms the cover and the carbonyl groups of Gly 210, Tyr283 and Trp284 line this pocket.

Important anchor groups for the inhibitor molecules are also the thiol group of Cys244, the imidazol group of His21 1 , the carboxylate groups of Asp 281 and Glu 214, and the oxyanion hole, there mainly the NH groups of Gly 212 and Cys 244.

Preferred interaction sites, especially for peptidic or peptidomimetic inhibitor molecules are Asp281 -Ala285, Gly210 to Glu214 and the specifitγ pockets S2 (for hydrophobic interactions) and S4 (for acidic interactions). Fig. 2 shows the location of the respective anchor groups in the gingipain R molecule to which a hypothetic substrate is bound. Fig. 2 also shows the respective distances of the preferred anchor groups for the inhibitor molecule and for substrates. An inhibitor molecule according to the invention preferably is designed to be able to bind to several of the above preferred anchor groups, however additional hydrophobic groups or groups interacting via salt bridges can also be present and might even enhance the binding capability of the inhibitor.

The inhibitor molecule according to the invention is preferably designed in such a way to be able to interact with several anchor groups and therefor has to have a structure that allows for such interactions in view of length and breadth. For this purpose the inhibitor preferably contains a backbone chain, to which one or more specific reactive groups are attached that mediate the binding and/or interaction. Additionally, the backbone chain itself can interact with the active site and its environment, for example via hydrogen bonds thereby further strengthening the binding/interaction.

Such backbone chain can be of any nature as long as it is possible that the reactive groups arrange in a specifically oriented manner. The backbone chain is preferably a peptide or a peptidomimetic chain, but can also be any other organic molecule. A polypeptidic or peptidomimetic backbone chain has the advantage that it can be formed close to the natural substrate, however for example much shorter, containing reactive groups that bind stronger or to more anchor groups of gingipain R etc.

Organic molecules can also be favourably present as backbone chain, for example in view of cost, ease of production, versatility for attaching reactive groups thereto etc.

In a preferred embodiment of the invention, the gingipain R inhibitor contains a polypeptide chain backbone with 1 to 10 amino acids and more preferred 2 to 8 amino acids, or a synthetic organic backbone chain that has a length corresponding to such polypeptide chain length.

If a non peptidic organic backbone chain or a combination of peptidic and non- peptidic residues is present, it is preferred that the flexibility of the chain is reduced to decrease the entropic price for binding of the inhibitor molecule to the enzyme, but only to an extent that allows the functional groups of an inhibitor molecule to adjust to the appropriate anchor groups on gingipain R. It is therefore preferred that such a chain is a hydrocarbon chain optionally containing heteroatoms, substitutions, double or triple bonds, and/or ring systems, optionally containing heteroatoms, to adjust the flexibility of the chain. As already mentioned above, the backbone chain itself can also form bonds with the gingipain R anchor groups or active site environment, thereby stabilizing the position of the specifically binding reactive groups and/or the whole inhibitor molecule.

All the embodiments of inhibitor molecules containing groups specifically binding to one or more or the described anchor groups of gingipain R can be designed by the man in the art upon the knowledge of the nature of the anchor groups and their respective distances as shown in Fig. 2. Using the given data it is well within the skill of the man in the art to design inhibitor molecules that interact specifically with gingipain R. The following description is a non-limiting example for the use of the crystal structure of gingipain R for the design of inhibitor molecules. As explained above, and as shown with para-chloromercury benzaminidine, benzamidine (I) can bind to the S1 pocket of gingipain R. The crystal structure shows that the salt bridge between the amidino group of benzamidine and the carboxylate group of the aspartate is the essential element for binding.

Binding is competitive with a binding constant of 86μM (65μM for HRGP and > 1000μM for the lysin specific gingipain K). Screening of a database of available chemicals (ACD library, MDL Information Systems, Inc) and subsequent docking with FlexX (GMD Sankt Augustin, Germany) shows that the molecule (II) can dock to the active site.

^'oxyanion hole ( J)

The interaction of the keto group oxygen with the oxyanion hole, formed by residues Gly21 2N and Cys244N, has been indicated. The fragments (I) and (II) can now be combined to yield inhibitors as

S 1 pocket (III)

that are capable of interactions both with the S 1 pocket and with the oxyanion hole. With reference to (III), the compound where R1 = H, R2 = H and R3 is CO-OC₂H₅ , i. e. oxalic acid-1 -ethyl-ester-2-(3-amidino)anilide, is referred to as the lead compound. It has been synthesised as described in Example 2. It is a competitive inhibitor of gingipain R with a binding constant of 52μM. In spite of the inverted orientation of the peptide bond, the molecule may still be slowly hydrolysed.

Docking of the lead compound to the active site of ginigpain R reveals a number of ways to improve the molecule. In the lead compound, the phenyl moiety can rotate freely around its bond to the amidino group. This rotational degree of freedom can be blocked by the interaction of the substituents R1 and R2 with the carbonyl oxygens of residues Trp284, Tyr283 (Ser 283 in some strains) or Gln282, and also with the -amino group of Gln282. Preferably, R1 is a hydrogen bond donor, and most preferably, R1 = OH. Preferably, R2 is also chosen to be a hydrogen bond donor, most preferably R2 = OH. An isosteric replacement for R2 is a possibility, and R2 = F could lead to an interaction of the fluorine atom with Gln282N.

The choice of R3, that faces solvent or the surface of the molecule, allows to modulate affinity of the inhibitor for gingipain R and other properties of the inhibitor like lipophilicity, membrane permeability etc. Molecules with various R3 substituents are easily prepared by reacting appropriate activated carboxylic acids with the aromatic amino group. If R3 = NH-R4, then these urea derivatives can be conveniently generated by reacting the aromatic amino group with appropriate isocyanates. Alternatively, 3- isocyanatobenzamidines can be reacted with suitable amines. In both cases, it is necessary to protect the amidino group of benzamidine against reactions with the isocyanates. Another way to avoid side reactions is to react 3-aminobenzonitriles with isocyanates or 3-isocyanatobenzonitriles with amines and to subsequently convert the resulting benzonitrile derivatives into the corresponding benzamidine derivates, e.g. via an amide oxime intermediate.

To avoid potential hydrolysis of the amide bond, replacement of the -NH- CO-R3 group in (III) with another, preferably planar group, e.g. a phenyl moiety, can be considered. For the molecule to interact with the oxyanion hole of the enzyme, a negatively charged group or a hydrogen bond acceptor should be present in ortho position to the bond that joins the group to the benzamidine moiety. As hydrogen bond acceptors, an amino group or fluorine, more preferred a hydroxyl group and most preferred a carbonyl oxygen can be used. In favorable cases, such compounds can be prepared from commercially available biphenyl precursors. If this is not the case, such compounds can for example be prepared by the Ullmann reaction.

Such exemplary compounds as described above are preferred inhibitors according to the present invention. The above disclosure enables the man in the art to design and produce inhibitor molecules that by interaction or binding to the active center of gingipain R block binding of the natural substrate and its cleavage. Such inhibitor molecules, therefore, can be advantageosly used in pharmaceutical compositions which constitute a further subject of the present invention. Any gingipain R inhibitors can be used advantageously in pharmaceutical compositions which constitute a further subject of the present invention. Such pharmaceutical compositions according to the invention contain the inhibitor optionally together with pharmaceutically acceptable carrier and adjuvant substances.

The inhibitor and the optional adjuvants are used in such amounts that at least a considerable inhibition of gingipain R is achieved. The pharmaceutical compositions can be used either to prevent or to treat peridontal diseases and other illnesses that are caused or affected by primary infections with Porphyromonas gingivalis. Such use of the pharmaceutical composition for the treatment and/or prevention of periodontal diseases, especially periodontitis, and/or cardiovascular diseases that occur after a primary periodontal disease is a further subject of the present invention.

Suitable amounts of inhibitor, necessary to achieve the desired treatment or prevention, can easily be determined by the skilled artisan. The form of application will normally be oral, an application into periodontal pockets will be most preferable. Such application can for example take place in form of a gel, a salve, impregnated fibres (like silk fibres) or any other suitable formulation that can be introduced into the periodontal pockets. Sticky formulations are preferred for a prolonged persistency of the active substance in the infected area.

Since primary peridontal diseases may lead to potentially life-threatening secondary diseases, the inhibitors according to the invention and especially the pharmaceutical compositions are important agents that can help avoid or cure the nasty periodontal diseases thereby avoiding manifestation of the more dangerous secondary diseases. The inhibitors according to the invention and pharmaceutical compositions containing such inhibitors therefore represent a considerable progress in pharmaceutics that can help avoid inter alia high cost for the treatment of periodontitis and secondary diseases. Examples and figures further illustrate the invention:

Examples:

1 a) Data Collection and Processing

A NATI dataset of crystals of gingipain A was measured using a 300 mm MAR-Research image plate detector mounted on a Rigaku RU 200 rotating anode X-ray generator with graphite monochromatized CuKσ radiation to a resolution of 2.16 A.

Crystals of native protein without bound inhibitor were grown in Limbro plates by the vapour diffusion method. For preparation of the monoclinic crystals NATI and pCMBA, drops of 1 .5 μ\ of an 8 mg/ml protein solution and 1 .5 μ\ of reservoir solution (3.6 M 1 ,6-hexanediol, 300 mM MgCI₂, 100 mM Tris-HCI pH 8.6) were mixed and equilibrated at 6°C for 1 year and then for 6 months at 21 °C. For faster crystal growth, microseeding was applied under slightly optimized conditions (3.4 M 1 ,6-hexanediol, 200 mM MgCI₂, 100 mM Tris-HCI pH 8.5) at 21 °C, yielding crystals of up to 0.5 x 0.2 x 0.02 mm within several weeks. The pCMBA derivative was obtained by soaking overnight in a solution containing 3 mM self-made p- chloromercurγ benzamidine and 5 mM cysteine. Before harvesting, the crystal was washed for 3 min in the precipitant solution to remove unbound mercury atoms. At BW6, DESY (Deutsches Elektronensynchrotron), Hamburg, these monoclinic crystals (Table I) diffract to beyond 1 .5 A and contain one molecule per asymmetric unit (V_M = 2.1 5 A³/Da), corresponding to a solvent content (v/v) of 43%.

The pCMBA crystals served for phase determination by the multiple anomalous dispersion technique. The MAD measurements were performed at the BW6 beam line at DESY in Hamburg, Germany. From a crystal flash frozen in liquid nitrogen, diffraction data to 1 .5 A resolution were collected at cryo temperatures using a MAR-Research CCD detector. MAD data were measured at the three wavelengths Λ- (remote, f = -8.7 e, f" = 4.3 e), λ₂ (peak, f = -1 2.8 e, f" = 10.0 e) and Λ₃ (edge, f = -1 5.1 e, f" = 9.5 e) by continously collecting 260 (remote), 590 (peak) and 259 (edge) frames of 0.7 deg each (see Table 2). Because the signal of the mercury atoms in the small protein crystal used was too weak, the fluorescence scan to determine optimal wavelengths for data collection was carried out using crystalline chloromercury benzoic acid (Fluka).

X-ray diffraction data of FFRCMK were collected to 2.0 and 2.9 A resolution from orthorhombic and rhombohedral crystals (Banbula et al., (1998), Prot. Sci. 7, 1259-1261 ) mounted in glass capillaries at 16 " C on a 300 mm MAR-Research image plate detector attached to a Rigaku RU200 rotating anode X-ray generator providing graphite monochromatized CuKσ radiation.

The intensities of all datasets were integrated with MOSFLM (Leslie, A.G.W. (1991 ) Crystrallographic Computing 5, Oxford University Press, Oxford, UK, pp. 50-61 ), scaled with SCALA (CCP4, Collaborative Computational Project, Number 4 (1 994), The CCP4 Suite: Programs for Protein Crystallography. Acta Cryst.,D50, 760-763), and converted to amplitudes with TRUNCATE (CCP4, (1994), supra).

1 b) MAD phasing

The mercury atoms in pCMBA (bound to Cys 244, Cys299 and Cys1 85 with decreasing occupancy) were localized in an anomalous difference Patterson map of the peak wavelength data set using CCP4 programs. The refinement of heavy atom parameters and calculation of MAD phases were carried out with SHARP (La Fortelle and Bricogne, (1997) In Sweet, R.M. and Carter, C.W., Jr (eds.), Methods Enzymol., 276, Academic Press, New York, USA, pp. 472-494). The final parameters are given in Table 3. The initial MAD phases were improved with SOLOMON (Abrahams and Leslie, ( 1 996) Acta. Cryst. D52,30-42), resulting in a 1 .5 A electron intensity map that was interpretable.

1 c) Molecular replacement

The NATI model was determined by molecular replacement with MOLREP using 3 A data and the pCMBA model as determined by MAD. The initial R-factor and correlation coefficient were 37.6 % and 66.9 %.

Both FFRCMK structures were solved by molecular replacement using the refined pCMBA model, 3 A data and MOLREP (Vagin and Teplyakov, (1 997) J. Appl. Cryst. 30, 1022-1025). The initial R-factor and correlation coefficient for the orthorhombic structure were 35.1 % and 70.5 %, respectively.

1 d) Model building and refinement

A first pCMBA model_was built on an Evans and Sutherland graphic workstation with FRODO (Jones,( 1 978) J. Appl. Cryst., 1 1 , 268-272) against a 1 .5 A solvent flattened electron density map calculated with the MAD phases. This and later models were subjected to crystallographic refinement cycles with CNS (Brϋnger et al., (1 998) Acta Cryst. D54, 905- 921 ) using the conjugate gradient method with an amplitude based maximum likelihood target function, Engh and Huber parameters for geometric restraints (Engh and Huber, (1 991 ) Acta Crystallog. sect A, 47, 392-400). In early stages, improved phases obtained with SIGMAA (Read, J.M. (1986) Acta crystallogr. sect A 42, 140-149) by a combination of the initial MAD phases with phases calculated from the current partial model and solvent flattening (SOLOMON, Abrahams and Leslie, (1 996), supra) were used for structure factor and density calculations. In the first step of the refinement process each initial model was subjected to a simulated annealing calculation (starting temperature: 2500 K, dropping in 25 K steps to 0 K and a final step at 300 K). Later, a water model was calculated using ARP (Lamzin and Wilson, (1993) Acta Cryst., D49, 129-147), and individual isotropic B-factors were refined using refinement protocols employing an amplitude based maximum likelihood target function until convergence was reached.

For the refinement of the orthorhombic FFRCMK model, the appropriately placed pCMBA model with all residues around the active site cysteine omitted served to calculate a simulated annealing omit map. The corresponding 2.0 A density map allowed tracing of all omitted residues and of the complete inhibitor using usual Arg-methylene parameter constraining the C-Sy distance to 1 .8 A. After remodeling and completely new insertion of solvent molecules, this model was refined as done for pCMBA.

The final refinement statistics is given in Table 3.

1 e) Quality of the models

In the final NATI and pCMBA models, the polypeptide chain is (except for peptide Gly67-Asn68) continuously defined from Tyr1 to Ser435, but could accomodate some more residues beyond Ser435. The final FFRCMK model lacks the last three residues in addition. Besides the protein chains and the inhibitors, the models contain a few ions interpreted as calcium and zinc ions (Table 3). The final electron densities are of high quality, in agreement with the reasonable R-f actors of around 16% (see Table 3). According to a search done with PROCHECK (Laskowski et al., (1993) PROCHECK, J. Appl. Crystallog. 26, 283-291 ), almost all non-glycine residues fall into the allowed or additionally allowed regions of the Ramachandran plot (Table 3); in all three structures Val245, Ser220 and Lys326 are clearly in the generously allowed region, however. The preceding peptide groups of Pro49 and Pro 1 88 have cis conformation.

1 f) Topology search.

Topology searches of the catalytic domain and of the IgSF domain were done against the PDB (November 1 998) employing the topological search facilities of TOP3D (Lu G. ( 1 996) Protein Data Bank quaterly Newsletters, #78, 10-1 1 )). This search revealed high scores only for the four caspase structures deposited RgpB (catalytic domain), but many IgSF structures of comparable score for the IgSF domain.

2) Synthesis of Oxalic-acid- 1-ethyl-ether-ester-2(3-amidino)anilide

The chemicals were purchased from Aldrich (Steinheim), Fluka (Neu-Ulm) and Merck (Darmstadt) and used without further purification. The solvents for HPLC were gradient grade.

2a) Oxalic acid-1 -ethyl-ester-2-(3-cyano)anilide (1)

A solution of oxalic acid monoethyl ester chloride (620 μ\; 5.4 mmol) in 5 ml of dry tetrahydrofuran (THF) was dropped slowly into an icecooled solution of 3-amino-benzonitrile (640mg; 5.4mmol) in 1 5ml of THF and triethylamine (1 .1 ml; 8.1 mmol).

The solution was stirred at 0 " C for 2 hours and then warmed to room temperature. The white precipitate was filtered off and the solvent was evaporated under reduced pressure. The resulting yellowish solid was used without further purification; yield 81 5 mg (70 %); HPLC (reversed phase silica C₁₈; eluents A: 5 % acetonitrile in 2% phosphoric acid, B: 10 % of 2% phosphoric acid in 90 % acetonitrile; flow 1 .5 rrrf/min; gradient: 0-100 % B in 12 min): R_f 7.2 min; FAB-MS: m/z = 219.2 [M + H⁺]; M_r = 218.2 for

2b) Oxalic acid-1 -ethyl-ester-2-(3-hydroxyamidino)anilide (2)

To a solution of 1 (240 mg; 1 .1 mmol) in 10 ml of ethanol KOH (90 mg; 1 .6 mmol) hydroxylamine hydrochloride (1 1 2 mg; 1 .6 mmol) was added. The resulting suspension was refluxed for 30 hours, cooled to room temperature and filtered. The solvent was evaporated under reduced pressure. The resulting white solid was purified by preparative HPLC (column: reversed phase silica gel C₁₈; eluent A: 0.1 % TFA in water; eluent B: 0.08 % TFA in acetonitrile; flow 10 ml/min; gradient 3-50 % B in 60 min; UV detection at 210 nm); yield: 1 10 mg (40 %); HPLC (reversed phase silica C₁₈; eluents A: 5 % acetonitrile in 2% phosphoric acid, B: 10 % of 2% phosphoric acid in 90 % acetonitrile; flow 1 .5 ml/min; gradient: 0-100 % B in 12 min): R_f 3.5 min; FAB-MS: m/z = 252.2 [M + H⁺]; M_r = 251 .2 for C₁₁H₁₃N₃O₄

2c) Oxalic acid-1 -ethyl-ester- 2-(3-amidino)anilide (3)

A solution of 2 (40 mg; 0.16 mmol) in 20 ml of methanol was acidified to pH 4 with acetic acid. 10% Pd/C (5 mg) were added and hydrogen was bubbled through the solution at room temperature for 3 days. The catalyst was filtered off and the solvent evaporated under reduced pressure. The crude product was purified by preparative HPLC (column: reversed phase silica gel C₁₈; eluent A: 0.1 % TFA in water; eluent B: 0.08 % TFA in acetonitrile; flow 10 ml/min; gradient 3-25% B in 60 min; UV detection at 210 nm); yield: 7 mg (19 %); HPLC (reversed phase silica C₁₈; eluents A: 5 % acetonitrile in 2% phosphoric acid, B: 10 % of 2% phosphoric acid in 90 % acetonitrile; flow 1 .5 ml/min; gradient: 0-100 % B in 1 2 min): R_f 3.5 min; FAB-MS: m/z = 236.2 [M + H⁺]; M_r = 235.2 for C₁₁H₁₃N₃O₃ 3) Determination of Inhibition Constants

The measurements were carried out on a microplate reader (MR 5000, Dynatech, Denkendorf, D) at 25 ° C. The test medium consisted of 200 μ\ buffer (0.05 M Tris, 0.1 M NaCI, 5 mM CaCl₂, 10 mM cystein, 50 mM H- Gly-Gly-OH, 5% ethanol, pH 8.0), 25 μ\ aqueous substrate solution and 50 μ\ enzyme solution. Two concentrations of the substrate and five concentrations of the inhibitor were used. Three min after the addition of the enzyme 25 μ\ acetic acid (50%) was added to quench the reaction and the optical density was measured at 405 nm. The K_rvalues were calculated according to Dixon [Biochem. J. 1953, 55, 170-171 ) using a linear regression programme. With the different gingipains the following enzyme concentrations and the respective substrates were used: HRGP (final concentration 0.52 nM), substrate CH₃SO₂-D-hexahydrotyrosyl-Gly-Arg-pNA (final concentrations 0.36 and 0.18 mM); RGP-2 (final concentration 1 .7 nM), substrate CH₃SO₂-D-hexahydrotyrosyl-Gly-Arg-pNA (final concentrations 0.36 and 0.1 8 mM); KGP (final concentration 4.1 nM), substrate Tos-Gly-Pro-Lys-pNA (final concentrations 0.18 and 0.091 mM). The substrates were supplied by Pentapharm Ltd., Basel, CH.

Figure legends.

Fig. 1 : Ribbon plot of RgpB (front view). The molecule consists of the catalytic domain (top) subdivided into subdomains A (right) and B (left), and the IgSF domain (bottom). Strands are shown as arrows, helices as spirals, and the connecting segments as ropes. The catalytic residues of the catalytic triad on top are shown as stick models, and the 3 putative calcium as spheres. Figure made with MOLSCRIPT (Kauiis, P.J., (1 991 ) J.Appl. Cryst., 1 1 ,268-272) and rendered with POVRay (POVRay programm, version 3.02 (1 997). URL: http:Wwww.povray.org)

Fig. 2: Schematic drawing of the FFRCMK interaction with RgpB's active site (shown in strand orientation). Covalent bonds are given as continuous lines, important intramolecular hydrogen bonds as dashed lines, and interesting spacings are indicated as bold arrows with distances given in Angstroms. The S1 , S2 and S4 pockets are shown as U-shaped curves. Cβ in both phenylalanines is not explicitly shown.

Fig. 3: Topology and sequence of RgpB.

A: Topological diagram of RgpB. Arrows denote strands s1 to s10 (catalytic domain) and sA to sH (IgSF domain), and cylinders indicate helices hi to hi 1 (catalytic domain).

B: RgpB sequence from Tyr1 to Ser435. Arrows and braces indicate β- strands and helices in RgpB. 1 18 residues of caspase-1 (Thomberry et al. (1992) Nature 356, 768-774; Cerretti et al. (1992) Science 256, 97-100) have been aligned to subdomain B according to topological equivalency with caspase-1 as deposited by Rano et al in the PDB under accession code 1 IBC. The numbering is for RgpB (top, SWISS-PROT entry code P95493) and caspase-1 (bottom). Figure made with ALSCRIPT (Barton, G.J. (1 993) Prot.Engng. 6, 37-40).

Fig. 4: Interaction of the H-D-Phe-L-Phe-L-Arg methylene inhibitor with the RgpB active site. The active-site region of RgpB, besides a few important residues mainly represented by the ribbon-like backbone, is shown in standard orientation (obtained from the front view, Fig. 1 , upon a 90° rotation about a horizontal axis). The inhibitor chain (light stick model) covalently linked via its methylene group to Cys244 Sy (center, right) runs from left to right, with its Arg-P1 side chain reaching back into the S1 pocket. The imidazole side chain of His21 1 and the carboxylate of Glu1 52 are arranged on the molecular surface (bottom) opposite to Cys244. For clarity, ZIN004 that has been captured from the crystallisation buffer and bridges Glu1 52 and His21 1 in the crystal structure has been omitted. Figure made with Insight II (Molecular Simulations Inc., San Diego). Fig. 5: View toward the solid surface of the RgpB active site (standard orientation), colored according to the electrostatic surface potential. Figures made with GRASP (Nicholls et al. (1991 ) Proteins: Structure, Function and

Genetics, 1 1 , 281 -296).

A: Close-up view toward the active center of the inhibited RgpB, with the H-

D-Phe-L-Phe-L-Arg methylene moiety (stick model) covalently bound. The

Arg side chain is partially buried in the S1 pocket.

B: View toward the active site of RgpB, with the modeled all-L-heptapeptide non-covalently attached (stick model).

Fig. 6: Schematic drawing of the probable peptide substrate-active center interaction of RgpB. The view is in the standard orientation, so that the modeled substrate (thick connections) runs from left to right. Probable intermolecular hydrogen bonds are shown by dashed lines, while the routes of attack of the Cys244 S on the Arg-P1 carbonyl and transfer of the His21 1 NcS hydrogen toward the leaving group are indicated by arrows. In this figure, a putative catalytic triade including Glu1 52 is shown.

Tables.

Table 1: Data collection statistics

NATI FFRCMK model pCMBA spacegroup P . 2 ^„ ^P2ι².². cell constants 55.57_, 59.36, 62.70 54.67, 59.92^, 63.21 51.93^, 79.92^, 99^.82. a, b_, c_, α, β, γ 90.00_, 94.69_, 90.0 90.00, 95.20, 90^.00 90.00^, 90^.00^, 90.00

wave leng_th λ ( A) 1.0697 1.0060 1.0100 1.5418 1.5418

Resolution range 17.96 - 1.49 27.84 -2, .16 15.64 - 2.03

Unique reflections 62,420 63,274 50,402 19,470 25,583

Multiplicity 3.3 6.6 3^.2 2.4 3.3

Completeness (%) 95.3 95.7 92^.3 85.2 99.0

R (%)* 5.5 7.4 4.9 12.0 7.8

Table 2: MAD phasing statistics

Phasing power

^ lhs λ, λ. , X λ. λ.

0.52 0.73 - 1.97 1.21 cen_tric reflections -

0.48 0.68 - 2.67 1.55 acentric reflections iso. ano. 0.93 0.83 0.S3 0.31 1.36 1.20

_{• r}, • -₇__iP^Λ number of acentric reflections: 61334 Number of centric reflections: 24Cv.^, numoer υ mean figure of merit: 0.452 (16.0 - 1.5 A)

= <phase integrated lack of closure>/lF._H - F,l

^Cullis phas_ing power = _I I_PF_{Hιeale l}l/_/^<nphnaassee i _nπtιe^cg_βrated lack of closure>

Table 3: Refinement statistics

Model pCMBA NAT! FFRCMK

Rcsolutioα range (A) 17.0-1.5 17.0-2.16 15.0-2.0

Reflections in working set li 0,620-$ 16,891 25,146

Reflections in test set 5,610 1,8-iS 1,865

R-_r-W 16.6 \1A 16.3

R_*_. W 19.1 2^-0 20.7

Protein atoms (noπ-H) 3,372 3,357 3,346

Ligand atoms (non-H) 21 - £D

Water molecules 561 237 322

Ion atoms 3Ca 3 Ca 3 Ca, 4 Zn

Average B -factor (A^:) 16.1 20.2 22.3 r. m. s. Δ3 (A^:) 3.0 3.0 l.S

Dcviacions from ideality r. m. s. bonds (A) 0.014 0.010 0.009 r. m. s. angels (^c) 1.79 I.6^J- 1.55

Model quality due to Ramachaπdran analysis allowed (%) 89.8 89.3 88.7 additionally allowed (%) 9.2 9.2 10.6 generously allowed (%) <X=h 0'- 0.7

*R«^-. = ∑(IF_Λ-F y∑IF..

not merged for refinement. Therefore, the number of

Biivoet pairs were nox meiyc .-.

S ons in the wo -r,,k.;i-n,„g s ,e_tt i •s_<; l laargeerr t thhaann t thhee p prreevviuously stated number ot unique reflections.

Claims

C l a i m s

1 . Crystal structure of gingipain R as shown in Fig. 1 .

2. Crystal structure of gingipain R in complex with the H-D-Phe-Phe-Arg- chloromethylketone as shown in Fig. 2 and Fig. 4.

3. Use of a crystal structure according to claim 1 or 2 to obtain crystal structure data for the design and/or identification of inhibitors of gingipain R.

4. Use according to claim 3, wherein a computer aided modelling programm is used for the design of inhibitor molecules.

5. Gingipain R-inhibitor, characterized in that it is able to specifically bind to and/or interact with D1 63 in the P1 -pocket of gingipain R and in that it has a nitrile, diazomethylketone, acyloxymethylketone, methylsulfonium salt, epoxysuccinyl derivative, vinylsulfone, O-acylhydroxamate, aziridine or activated disulfide group that forms a covalent, hydrolytically stabile bond to the enzyme.

6. Gingipain R-inhibitor, characterized in that it binds non-covalently to the enzyme or forms a covalent, hydrolytically labile bond to the enzyme, and in that it specifically binds to and/or interacts with D163 in the P1 -pocket of gingipain R.

7. Gingipain R-inhibitor according to claim 6, that has an aldehyde, methylketone or trifluoromethylketone, σ-ketoacid, -ester or amide or diketone group that forms a covalent, hydrolytically labile bond to the enzyme.

8. Gingipain R-inhibitor according to anyone of claims 5 to 7, characterized in that it contains an anchor group specific for Asp1 63 which is arginine, an argininomimetic, an unflexible aromatic residue or a saturated or unsaturated cyclic compound.

9. Gingipain R-inhibitor according to anyone of claims 5 to 8, characterized in that it further binds to and/or interacts with one or more amino acids of gingipain R selected from the group consisting of Cys244, Asp281 , His21 1 , Trp284, Gly 21 2, Thr 242, Gly 210, and Tyr 283.

10. Gingipain R-inhibitor according to anyone of claims 5 to 8, characterized in that it binds at least to Asp163, Cys244 and His21 1 .

1 1 . Gingipain R-inhibitor according to anyone of claims 5 to 8, characterized in that it contains groups that effect interaction with His21 1 and/or Cys244 and/or Trp284 via hydrogen bonds.

1 2. Gingipain R-inhibitor according to anyone of claims 5 to 1 1 , characterized in that it contains a backbone chain to which one or more groups that mediate the binding or interaction are attached.

13. Gingipain R-inhibitor according to claim 1 2, characterized in that the backbone chain is a polypeptide or a synthetic organic molecule.

14. Gingipain R-inhibitor according to claim 13, characterized in that a polypeptide backbone chain contains 2 to 8 amino acids and a synthetic organic backbone chain has a length corresponding to such polypeptide chain length.

1 5. Gingipain R-inhibitor according to anyone of claims 1 2 to 14, characterized in that the synthetic organic backbone chain is an alkyl chain optionally containing heteroatoms, substitutions, double or triple bonds and/or ring systems optionally containing heteroatoms, to reduce the flexibility of the chain, but only to an extent that allows the functional groups of an inhibitor molecule to adjust to the appropriate anchor groups of gingipain R.

16. Gingipain R-inhibitor according to claim 15, characterized in that it contains basic side groups other than the one that projects into the S1 pocket.

17. Gingipain R-inhibitor according to formula (III)

(III)

18. Gingipain R-inhibitor according to claim 17, wherein R1 is a hydrogen bond donor or hydrogen, R2 is a hydrogen bond donor or fluorine or hydrogen, R3 is alkyl or aryl .

1 9. Gingipain R inhibitor according to claim 1 8, wherein the benzene ring is replaced with a heterocyclic, aromatic compound.

20. Gingipain R-inhibitor according to claim 1 8 or claim 1 9, wherein -NH- CO-R3 is replaced with a substituted or unsubstituted ring system optionally containing heteroatoms.

21 . Pharmaceutical composition, characterized in that it contains a gingipain R-inhibitor according to anyone of claims 5 to 20 as active agent.

22. Pharmaceutical composition according to claim 21 , characterized in that it contains additionally pharmaceutically acceptable carrier or adjuvant substances.

23. Use of a pharmaceutical composition according to claim 21 or 22 for the treatment and/or prevention of periodontal diseases, especially periodontitis, and/or cardiovascular diseases that occur after a primary periodontal disease.

24. Use of a gingipain R-inhibitor for the production of a pharmaceutical composition for the treatment and/or prevention of periodontal diseases, especially periodontitis, and/or cardiovascular diseases that occur after a primary periodontal disease.