Proc. Natl. Acad. Sci. USA Vol. 90, pp. 6796-6800, July 1993 Biochemustry Crystal structures of native and inhibited forms of human cathepsin D: Implications for lysosomal targeting and drug design (aspartic protcase/N-linked oligosaccharide/pepstatin A) ERic T. BALDWIN*, T. NARAYANA BHAT*, SERGEI GULNIK*, MADHUSOODAN V. HOSUR*t, RAYMOND C. SOWDER Il, RAUL E. CACHAU*, JACK COLLINS*, ABELARDO M. SILVA*, AND JOHN W. ERICKSON*§ *Structural Biochemistry Program, Frederick Biomedical Supercomputing Center and tAIDS Vaccine Program, Program Resources Inc./DynCorp, National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, MD 21702 Communicated by David R. Davies, March 24, 1993 (receivedfor review February 4, 1993) ABSTRACT Cathepsin D (EC 3.4.23.5) is a lysosomal duced in the vicinity of the growing tumor, may degrade the protease suspected to play important roles in protein catabo- extracellular matrix and thereby promote the escape of lism, antigen processing, degenerative diseases, and breast cancer cells to the lymphatic and circulatory systems and cancer progresson. Determination of the crystal structures of enhance the invasion of new tissues (17, 18). The design of cathepsin D and a complex with pepstatin at 2.5 A resolution potent and specific inhibitors of cathepsin D will aid the provides insights into inhibitor binding and lysosomal targeting further elucidation of the roles of this enzyme in human for this two-chain, N-glycosylated aspartic protease. Compar- disease. We previously described the purification and crys- ison with the structures of a complex of pepstatin bound to tallization ofhuman cathepsin D from liver (3); similar studies rhizopuspepsin and with a human renin-bihbitor complex have been reported recently for cathepsin D isolated from revealed differences in subsite structures and inhibitor-enzyme bovine liver (19) and human spleen (20). We now describe the interactions that are consistent with affnity differences and complete three-dimensional structure of native and pepsta- structure-activity relationships and suggest strategies for fme- tin-inhibited forms of human cathepsin D.W tuning the specificity of cathepsin D inhibitors. Mutagenesis studies have identified a phosphotransferase recognition region that is required for oligosaccharide phosphorylation but is 32 MATERIALS AND METHODS A distant from the N-domain glycosylation site at Asn-70. Crystals of native human liver cathepsin D (space group P65; Electron density for the crystal structure of cathepsin D a = b = 125.9 A, c = 104.1 A) were prepared as described (3). indicated the presence of an N-linked oligosaccharide that The structure was solved by molecular replacement using the extends from Asn-70 toward Lys-203, which is a key component program XPLOR-3.0 (21). The crystal structure of porcine of the phosphotransferase recognition region, and thus pro- pepsin was used as the search molecule (22). The final R vides a structural explanation for how the phosphotransferase factor was 18.8% for data from 10.0 to 2.5 A resolution for can recognize apparently distnt sites on the protein surface. 28,077 unique reflections. The rms deviations from ideality for bond lengths, andes, dihedral torsions, and improper Cathepsin D (EC 3.4.23.5) is an aspartic protease that is torsions were 0.012 A, 3.10, 27.30, and 1.140, respectively. normally found in the lysosomes of higher eukaryotes where Cocrystals of cathepsin D with pepstatin were prepared by it functions in protein catabolism (1). This enzyme is distin- using conditions similar to those used for the crystallization guished from other members of the pepsin family (2) by two of native enzyme. The structure of the complex was solved features that are characteristic of lysosomal hydrolases. by using the refined native structure of cathepsin D and First, mature cathepsin D is found predominantly in a two- difference Fourier methods. After a single round of refine- chain form due to a posttranslational cleavage event (1, 3). ment, the resulting 2F. - F, and F. - F, maps showed Second, it contains phosphorylated, N-linked oligosaccha- substantial electron density in the active-site region. The rides that target the enzyme to lysosomes via mannose inhibitor was built and the complex was refined to a final R 6-phosphate (M6P) receptors (4, 5). Phosphorylation in- factor of 17.9% with XPLOR. The rms deviations from ideality volves recognition of both sugar and protein structural de- for bond lengths, an.gles, dihedral torsions, and improper terminants by a phosphotransferase enzyme (6, 7). torsions were 0.012 A, 3.20, 26.60, and 1.140, respectively. Interest in cathepsin D as a target for drug design results The inhibition constants for pepstatin were determined by from its association with several biological processes of using a fluorometric assay at 370C with the substrate Ac-Glu- therapeutic significance including lysosomal biogenesis and Glu(Edans)-Lys-Pro-Ile-Cys-Phe-Phe-Arg-Leu-Gly-Lys(Dab- protein targeting (4, 5), antigen processing and the presen- cyl)-Glu-NH2, which was kindly supplied by Grant Krafft tation of peptide fragments to class II major histocompati- (Abbott) {Edans is 5-[(2-aminoethyl)amino]naphthalene-1- bility complexes (8-10), connective tissue disease pathology sulfonic acid and Dabcyl is p-(dimethylaminophenylazo)ben- (11), muscular dystrophy (12), degenerative brain changes zoic acid}. The Km for this substrate was 5 ± 0.6 ,uM and 0.3 (13, 14), and cleavage of amyloid precursor protein within ± 0.07 ,uM for cathepsin D and rhizopuspepsin, respectively. senile plaques of Alzheimer brain (15). Recent studies of In the case of cathepsin D, the inhibitor concentration was primary breast cancers demonstrated that elevated levels of cathepsin D were correlated with an increased risk of me- Abbreviations: M6P, mannose 6-phosphate; Nag, N-acetylglu- tastasis and shorter relapse-free survival (16). High levels of cosamine. cathepsin D and other proteases such as collagenase, pro- tPresent address: Solid State Physics Division, Bhabha Atomic Research Centre, Trombay, Bombay, India. §To whom reprint requests should be addressed. The publication costs ofthis article were defrayed in part by page charge 1The atomic coordinates and structure factors have been deposited payment. This article must therefore be hereby marked "advertisement" in the Protein Data Bank, Chemistry Department, Brookhaven in accordance with 18 U.S.C. §1734 solely to indicate this fact. National Laboratory, Upton NY 11973 (reference 1LYA, 1LYB). 6796 Downloaded by guest on September 24, 2021 Biochemistry: Baldwin et al. Proc. Natl. Acad. Sci. USA 90 (1993) 6797 less than or equal to the enzyme concentration, and kinetic This disulfide bridge, which had been predicted from mod- data were analyzed by using a model for tight-binding inhib- eling studies of porcine cathepsin D (26), is within the light itors (23). Michaelis-Menten kinetics were assumed with chain region and may stabilize the C terminus of the light rhizopuspepsin. The experimental data were fitted by non- chain against further proteolytic degradation. There are three linear regression analysis. proline residues in cathepsin D that refined in a cis-peptide conformation. A cis-peptide bond involving Pro-24 is con- RESULTS AND DISCUSSION served among all the aspartic proteases except for penicil- lopepsin, which has a deletion in this region. The other The 2.5 A resolution structures of native cathepsin D and of proline residues with cis-peptide bonds are localized within a its isomorphous complex with pepstatin, a naturally occur- proline-rich segment, termed the "proline loop," that con- ring aspartic protease inhibitor, were solved by molecular sists of residues Pro-312, Pro-313, cis-peptide Pro-314, Ser- replacement. The crystallographic asymmetric unit ofthe P65 315, Gly-316, and cis-peptide Pro-317. An analogous struc- unit cell contains two molecules that are identical and related ture has been described recently for human renin (24). by a pseudo-twofold rotation, K = 160.90, and by a relative Binding of pepstatin to cathepsin D induced small struc- translation of 36 A. Cathepsin D contains three topologically tural changes in the "flap" region (the (-hairpin structure distinct regions that are typical of aspartic proteases (Fig. 1): composed of residues 72-87; Fig. 1). Residues 79 and 80 at an N-terminal domain (residues 1-188), a C-terminal domain the tip of the flap moved in toward the inhibitor by about 1.7 (residues 189-346), and an interdomain, anti-parallel /3sheet A. The flexibility of this (bend decreased upon inhibitor composed of the N terminus (residues 1-7), the C terminus binding as indicated by a drop in the mean B factor from 64 (residues 330-346), and the interdomain-linking residues A2 to 25 A2 for the 16 main chain atoms of residues 78-81. (160-200). The latter region links the pseudo-twofold-related Small flap movements due to inhibitor binding have been N and C domains (K = 179.30 for the superposition of 19 Ca observed for other aspartic proteases (2). No subdomain atom pairs with a rms of 1.5 A), each ofwhich contributes an displacements were observed in cathepsin D upon inhibitor aspartic acid, Asp-33 and Asp-231, to the active site. The binding, unlike the case of pepsin (27) and endothiapepsin overall structural agreement for the main chain of cathepsin (28). D with the known structures of other mammalian aspartic The pepstatin-cathepsin D complex is stabilized by nu- proteases, human renin (24), porcine pepsin (Protein Data merous hydrogen bonds between backbone atoms of the Bank identification code, 4PEP; ref. 25), and bovine chy- inhibitor and both main chain and side chain atoms of the mosin (1CMS; ref. 25), ranged between 0.9 and 1.0 A rms. enzyme (Fig. 2A). The central statine hydroxyl group occu- Several structural features distinguish cathepsin D from pies the position of a water molecule that interacts with the other aspartic proteases.