J., Matthews,, D. A,, Hamlin, K. C. &. Kraut, J. ( 1 982) J. Biol. ('hem. 257, 13650- I3666. 4. Poet, R. & Milner-White, E. J. ( 1986) Cornpurer Graphics Forum.
578 studies 151 aimed at predicting the three-dimensional structure of proteins from sequence information. I . Richardson, J. S. ( 1 98 I ) Adv. Protein Chem. 34, 167-339 2. Milner-White, E. J. ( 1 988) J . Mol. B i d . 199,503-5 1 1 3. Milner-White, E. J. & Poet. R. (1987) Trends Biochrm. Sci. 12, 189- 192 4. Poet, R. & Milner-White, E. J. ( 1986) Cornpurer Graphics Forum 5.21 1-215
BIOCHEMICAL SOCIETY TRANSACTIONS 5. Milner-White, E. J. ( I 989) Trends I’harmucd S c i . 10, 70-74 6. Baker, E. N. & Hubbard, R. E. ( 1984) Prog Biophys. Mol. Hiol. 44, Y7- 179 7. Bolin, J. T., Filman, I). J., Matthews,, D. A,, Hamlin, K. C. & Kraut, J. ( 1 982) J. Biol. (’hem. 257, 13650- I3666
Received 14 November I989
Isolation and characterization of gingivain, a cysteine proteinase from Porphyromonas gingivulis strain W83 HAROUN N. SHAH,* SAHEER E. GHARBIA,* DEVANAND KOWLESSUR,t$ ELIZABETH WILKIEtS and KEITH BROCKLEHURSTtt *Department of Oral Microbiology, The London Hospital Medical College, University of London, Turner Street, London E l .?AD, arid tDepartmerit of Biochemistry, St Bartholomew’s Hospital Medical College, University of London, Charterhouse Sqiiare, Loridori E C l M 6BQ, U.K. The structural variation that exists within the cysteine proteinase family is proving to be valuable in the study of molecular recognition and enzyme catalytic mechanism [ 1-31 and reports of investigations on little-studied members of the family [4-81 continue to reveal behaviour considerably different from that of papain ( E C 3.4.22.2) which has often been regarded as the cysteine proteinase archetype. In the present communication we report evidence that compels the view that the extracellular proteinase secreted by the virulent Porphyromonas gingivalis strain W83 is a cysteine proteinase and suggest that it be known as gingivain. It is anticipated that characterization of this enzyme may both assist the understanding of the cause of periodontal disease and add to the general picture of structure-function relationships in the cysteine proteinase family. Among the large spectrum of bacterial species that are isolated from subgingival dental plaque, Racteroides gingivalis, recently reclassified as 1’. gingivalis [Y],is considered to play a central role in destructive periodontal disease [lo].It produces a variety of putative virulence factors of which an extracellular proteinase is regarded as one of the most important [ 111. Studies in vitro strongly suggest that, in addition to the probable role of this enzyme in microbial nutrition, it may disrupt a range of host defence mechanisms and connective tissue functions [ 101. The proteinase is often referred to as ‘trypsin-like’ [ 12, 131, although enhancement of catalytic activity by mercaptans and inhibition by iodoacetate and by N-ethylmaleimide suggest an essential role for a thiol group [lo]. In the present work the enzyme was isolated in fully active form by thiol-disulphide interchange covalent chromatography [ 141 and its cysteine proteinase nature was established by its behaviour towards the twohydronic-state thiol-specific inhibitor and reactivity probe, 2,2’-dipyridyl disulphide (2-Py-S-S-2-Py) [ 151. It was established also that the cysteine proteinase is associated with the vesicles (Fig. l ) , observed initially by Shah er al. [ 161 and more recently by others [ l o ] in the supernatant of 1’. gingivalis culture broth, but is removed from them during the covalent chromatography process. Abbreviations used: 2-Py-S-S-2-Py. 2.2‘-dipyridyl disulphide; LBAPNA, a-N-benzoyl-~-arginine-4-nitroanilide. $Present address: Department of Biochemistry, Queen Mary and Westfield College, University of London. Mile End Road, London E I 4NS. U.K.
Fig. 1. Electron micrograph of a negatively stained vesicle preparation from the cotzceritrated culture supertiatant of I? girigivalis strain W8.q The Figure shows a parent cell (approx. 500 nm diameter) and vesicles (approx. 30- 150 nm diameter), some of which appear to be associated with the cell surface. Original magnification x 1 15 000. All of the hydrolytic activity towards a-N-benzoyl-Larginine-4-nitroanilide ( L-BAPNA) in the vesicle-supernatant mixture is thiol-dependent, as evidenced by a thiol-specific inhibition-reactivation cycle [ 151 involving 2-Py-S-S-2-Py and 2-mercaptoethanol. The vesicles and all of the hydrolytic activity towards L-BAPNA were precipitated when the supernatant was brought to 70% saturation with ammonium sulphate. Isolation of the vesicles was achieved by centrifugation at 4°C for 22 h at 150 000 g. The new supernatant was devoid of both vesicles and of activity towards L-BAPNA and all of the thiol-dependent activity was present in the isolated vesicles resuspended in fresh buffer. Separation from the vesicles of the protein responsible for the thiol-dependent activity and purification in fully active form was achieved by thiol-disulphide interchange covalent chromatography [ 141. This involved admixture of the vesicle preparation and Sepharose-glutathione-2pyridyl disulphide gel in 0.1 M-Tris/HCI buffer pH 8.0, washing with the same buffer containing 0.3 M-NaCI, and release of the thiol-containing components bonded to the gel as mixed disulphides by elution with cysteine at pH 8.0. Elution with a linear gradient of cysteine (0.1-40 mM) produced two well-resolved peaks of thiol-containing protein. The first t o be eluted was devoid of activity towards L-BAPNA and the 1000
633rd MEETING, LONDON second contained the thiol-dependent activity towards this substrate. The active protein, separated from cysteine by gel filtration, was found to possess a key catalytic site characteristic that is typical of many cysteine proteinases [ 1, 2, IS]. Thus stopped-flow kinetic analysis of the reaction of its catalytically essential thiol group with 2-Py-S-S-2-Py showed that the reactivity is a minimum at a pH value of approx. 6 and increases at both lower and higher pH values (approx. 4 and X), behaviour characteristic of the presence of an - S - / -Im'H ion pair [2, IS]. We arc grateful for the receipt of an O.R.S. Award to S.E.G.. to the M.R.C. for a project grant. and to the S.E.R.C. for project grants and Earmarked Research Studentships for work o n the cysteine proteinase mechanism.
I . Brocklehurst. K. ( 1986) in Cjsteine Profeirimes cirid their 1rrhihifors.(Turk. V.. ed.), pp. 307-325. Walter d e Gruyter. Berlin 2. Hrocklehurst, K.. Willenbrock. F. & Salih. E. (1987) in /iydro/yrii. Drzymes (Neuberger. A. & Brocklehurst, K., eds.), pp. 39- 158. Elsevier. Amsterdam 3. Kowlessur. D., O'Driscoll. M.. Topham, C. H., Templeton, W.. Thomas. E. W. & Brocklehurst. K. ( 1089) Hiochem. J. 259.
579 5. Willenbrock. F. & Brocklehurst, K. ( 1 9x4) H i o c . h c w i . J. 222, 805-814 6. Baines, B. S., Brocklehurst, K., Carey, P. R., Jarvis, M., Salih, E. & Storer, A. C. ( 1986) Hiochem. J. 233, 1 19- 1 29 7. Baines, B. S., Brocklehurst. K., McKee. R.. O'l>riscoll. M.. Salih. H. & Smith. H. ( I 986) Hiochem. Soc. 7i.citr.s. 14. 1226- 1227 8. Laycock. M. V.. Hirama. T.. Hasnain. S., Watson, 11. & Storer. A. C. (1989) Riochem. 1. 263,439-444 9. Shah, H. N. & Collins, M. D: ( 1988) I n ( . J. Sy.st. /kicti,rio/. 38, 128-1 31 10. Mayrand. D. & Holt. S. C. ( 1988) Microhiol. Kc,i: 52, 134- I52 1 1. Slots, J. ( 19x2) in /lost-l'eireisitr Interrittiorrs in /'c,riodorrttrl Oisenses (Genco. R. J. & Mergenhapen, S. E., eds.). pp, 27-45. ASM Publications, Washington. D.C. 12. Grenier. D., Chao, G. & McBridc. U. C. ( 1089) Irrjiw. I m m i t r r . 57,95-99 13. Smalley, J. W., Birss, A. J., Kay. H. M.. McKce. A. S. & Marsh. P. D. (1989) O r d Microhio/. Immirriol. 4. 1 78- I8 I 14. Brocklehurst, K., Carlsson, J. & Kierstan. M. P.J. (1985) 7i)p. Enzyme Ferment. Hiorechnol. 10. 146- 1 88 IS. Brocklehurst, K. ( 1082) Merhods Enzyme/. 87C. 427-469 16. Shah, H. N. & Williams, R. A. D., Bowden, G. H. & Hardie, J. M. ( I 976) J . Appl. Hacreriol. 4 I , 473-492
443-452
1. Csoma. J. C. & Polgar. L. ( 1984) Wioc.hem.J. 222, 769-776
Received 23 November I989
Investigation of mechanistic consequences of natural structural variation within the cysteine proteinases by knowledge-based modelling and kinetic methods CHRISTOPHER M. TOPHAM,* JOHN OVERINGTON,* DEVENAND KOWLESSUR,tf MARK THOMAS,tt. EMRYS W. THOMAS5 and KEITH BROCKLEHURSTtS * Laboratory of Molecirlur RioloAy, Departmetit of Ctystullography, Rirkheck College, Utiiversity of Lotidoti, Muler Street. Lotidoti WC'IE 71M, t Depurttnetit of Hiochetnistty, SI Burtholomew's Hospital Medical College, Utiiversity of Lotidoti, Churterhoirse Sqiiure, Lotidoti EC'IM 6HQ, mid §Depurrmetit of Rioiogicul Sciences, Utiiver.yityof Sulford, Sulford M5 4 WT, U . K . The structural variation within the cysteine proteinases is being used in the study of molecular recognition and catalytic mechanism [ I , 21 and work on little-studied members of the family [ 3-51 continues to reveal behaviour considerably different from that of papain ( E C 3.4.22.2). The only highresolution X-ray structures of cysteine proteinases available are those of papain (1.65 A ) [6]and actinidin ( E C 3.4.22.14) ( 1.70 A) 171 which differ in active-centre behaviour [ 1, 21 despite having closely similar crystal structures. We are extending the availability of variants for mechanistic study by building three-dimensional structures of other members of the family for which complete amino acid sequences are available [8, 91. In the present work, structures of papaya proteinase R [XI and stem bromelain [ 9 ] were built (see Fig. 1 ) by using a knowledge-based approach incorporated in COMPOSER [ 10-12) and refinement by energy minimization (SYBIL, version 5.2, 1989, Tripos Associates Inc.). The approach is based on simultaneous rigid-body superposition of related known structures, building loops selected from homologous proteins or a wider database, and rule-based building of side-chain geometries. By using the co-ordinate data for papain (OPAP) and actinidin (2ACT) (Brookhaven Data Bank), the two strucAbbreviation used: Py.pyridyl. :Present address: Department of Biochemistry. Queen Mary and Westfield College. University of London, Mile End Road, London E I 4NS. U.K.
Vol. 18
tures were superposed by an iteratively reweighted least squares procedure [ 1 11 to define a new set of topologically equivalent residues from an initial set in the catalytic site region of each enzyme (Cys-25, His- 1 S9, Asn- 175, papain numbering [ 131). The weighted mean of each C a position comprises the framework for modelling. The alignment of the papaya proteinase 52 sequence with the consensus sequences (templates) of the framework is trivial in most regions due to the high sequence identity for the core region with papain (72.0%) and actinidin (53.4IK,) and the result is almost entirely consistent with the sequence alignment reported previously (81.The alignment o f the stem bromelain sequence to the framework differs in three regions from the multiple sequence alignment reported in [U]. The most striking difference occurs around a four residue delction (16%-d, papain numbering). The sequence alignment in [0] led to the conclusion that the highly conserved Asn- I75 and Ser-176 of papain are respectively depleted and mutated to Lys in stem bromelain. That alignment l(ii) and (iv)] permits conservation of Tyr- 170 in papain, actinidin and stem bromelain but cannot easily be rationalized in structural terms. Thus it would necessitate disruption o f the antiparallel P-sheet and hydrogen-bonding pattern common t o actinidin and papain. 170 Papain: GYGP----NY I L I KNSWGTGW Papain: GYGPN----Y I L I KNSWGTGW Actinidin: G Y G T E G G V D Y W I V K N S W D T T W Stem bromelain:GYGQDS I I - Y P K - - - K W G A K W ( v ) Stem brome1ain:GYGQ----DS I I Y PKKWGAKW
(i) (ii) (iii) (iv)
By contrast, our alignment [(i). (iii) and ( v ) ] allows hydrophobic residues Ile-17 1 and lle- 172 o f stern hromelain (papain numbering) to be aligned with buried hydrophobic residues in papain and actinidin, and preserves the hydrogen bonding network by placing the four-rcsidue deletion in a
