View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Research Article 51

Crystal structure of porcine H determined at 2.1 Å resolution: location of the mini-chain C-terminal carboxyl group defines aminopeptidase function Gregor Guncar*, Marjetka Podobnik, Joze Pungercar, Borut Štrukelj, Vito Turk and Dušan Turk

Background: Cathepsin H is a lysosomal cysteine , involved in Address: Department of Biochemistry and intracellular degradation. It is the only known mono-aminopeptidase in Molecular Biology, Jozef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia. the -like family and is reported to be involved in tumor metastasis. The cathepsin H structure was determined in order to investigate the structural basis *Corresponding author. for its aminopeptidase activity and thus to provide the basis for structure-based E-mail: [email protected] design of synthetic inhibitors. Key words: aminopeptidase, cathepsin H, crystal structure, , papain Results: The crystal structure of native porcine cathepsin H was determined at 2.1 Å resolution. The structure has the typical papain-family fold. The so-called Received: 9 October 1997 mini-chain, the octapeptide EPQNCSAT, is attached via a disulfide bond to the Revisions requested: 12 November 1997 Revisions received: 19 November 1997 body of the and bound in a narrowed active-site cleft, in the substrate- Accepted: 19 November 1997 binding direction. The mini-chain fills the region that in related comprises the non-primed substrate-binding sites from S2 backwards. Structure 15 January 1998, 6:51–61 http://biomednet.com/elecref/0969212600600051 Conclusions: The crystal structure of cathepsin H reveals that the mini-chain © Current Biology Ltd ISSN 0969-2126 has a definitive role in substrate recognition and that carbohydrate residues attached to the body of the enzyme are involved in positioning the mini-chain in the active-site cleft. Modeling of a substrate into the active-site cleft suggests that the negatively charged carboxyl group of the C terminus of the mini-chain acts as an anchor for the positively charged N-terminal amino group of a substrate. The observed displacements of the residues within the active-site cleft from their equivalent positions in the papain-like endopeptidases suggest that they form the structural basis for the positioning of both the mini-chain and the substrate, resulting in exopeptidase activity.

Introduction enzyme between the heavy and light chains is partial and Cathepsin H (EC 3.4.22.16) is a relatively abundant pro- can occur between residues Asn168B and Gly168C (papain tease and, like other lysosomal cysteine , is numbering). The ratio of cleaved to uncleaved forms involved in intracellular protein degradation [1]. There is a varies between different species and organs, but in general growing body of evidence that the level of cathepsin H or the two-chain form is present in larger amounts than the cathepsin H-like enzymes [2] is increased in disease states, three-chain form [7–9]. including breast carcinoma [3], melanoma and tumor meta- stasis [4]. In the condition mucolipidosis II [5], cathepsin The amino-acid sequences of rat [8], human [10] and H activity is diminished, together with that of other mouse cathepsin H [11] are in agreement with those . Cysteine proteases are no longer seen as lyso- derived from the rat [12] and human [13] cDNA sequences. somal enzymes only, but are involved in numerous From these sequences, it is evident that the mini-chain processes outside the [6]. For example, a cathep- originates from the cathepsin H propeptide and that it is sin H-like enzyme has been found in cells from a human located between propeptide residues Glu76P and Thr83P melanoma cell line [2]. This enzyme can degrade fibrino- (propeptide numbering). shows that gen and fibronectin and thus may, together with other cathepsin H belongs to the superfamily of papain-like cys- proteases, be involved in the destruction of extracellular teine proteases, which includes the mammalian lysosomal matrix components, leading to proliferation, migra- cathepsins B, L, S, K and N [14] and the plant enzymes tion and metastasis [2]. papain, actinidin, aleurain and oryzains. The closest plant enzymes are aleurain [15] and oryzain γ [16]. Mature cathepsin H consists of three fragments, the N- terminal heavy and C-terminal light chains, and an Cathepsin H is inhibited by the three groups of intracellu- octapeptide called the mini-chain [7]. Cleavage of the lar and extracellular protein inhibitors — stefins [17–19], 52 Structure 1998, Vol 6 No 1

[20] and kininogens [7]. Cathepsin H interacts as shown in the standard view (Figure 1), with the addition with these inhibitors more strongly than does of an octapeptide bound to the R-domain. but less strongly than do the endopeptidases papain and . An exception is stefin B, which is a stronger The predominant secondary structure of the L-domain is inhibitor of cathepsin H than of the endopeptidases [17,19]. α helix, whereas the scaffold of the R-domain is based on a β barrel. The two domains interact through an extended In contrast to cathepsins B, L [21], S [22] and K [23], amphipathic interface stabilized by numerous hydrogen cathepsin H is only slowly inhibited by irreversible epoxy- bonds as well as by hydrophobic contacts. The two-domain succinyl-based inhibitors derived from E-64 [21]. It is not interface opens at the top into a V-shaped active-site cleft inhibited by leupeptin [24]. It is, however, strongly inhib- in which the two catalytic residues, Cys25 and His159, and ited by substrate analogues composed of a single amino- the mini-chain are located. acid residue bound to a diazomethane or fluoromethane group, which react with the active-site cysteine [25]. Superposition of the Cα atoms of cathepsin H on the Cα atoms of papain, actinidin and cathepsin B shows that Cathepsins H, B and C are the only known papain-like 180 amino-acid residues of cathepsin H and actinidin lysosomal cysteine exopeptidases [14]. Cathepsin H is [36] are topologically equivalent, 173 with papain [39] an aminopeptidase [26,27], cleaving a single N-terminal and 156 with human cathepsin B [34], yielding root mean residue from a polypeptide chain, while is an square (rms) deviations of 0.85 Å, 0.85 Å and 0.83 Å, aminodipeptidase [28] and cathepsin B is a carboxydipep- respectively. Structure-based alignment of those enzymes tidase [29]. It has been reported that cathepsin H also is presented in Figure 2. The numbers of equivalently exhibits endopeptidase activity [30]. The active-site cleft positioned residues indicate that cathepsin H is struc- of papain and the related endopeptidases, cathepsins K turally most similar to actinidin. The differences between [31,32] and L [33], is empty, whereas the occluding loop the two structures are mostly insertions and deletions in of cathepsin B [34] occupies the back of the active-site the loop regions. cleft, thus making cathepsin B a carboxydipeptidase. Using chemical and enzymatic cleavage methods, it has Mature porcine cathepsin H has two N-glycosylation sites, previously been shown that the mini-chain is bound via a Asn112 on the body of the enzyme and Asn79P on the disulfide bond to Cys205 of the body of cathepsin H [35]. mini-chain [40]. Carbohydrate residues are frequently not It was proposed [35] that the mini-chain is involved in the visible in a crystal structure. In cathepsin H, only the first steric regulation of the accessibility of the active-site cleft. three carbohydrate rings attached to Asn112 and located in The way in which the mini-chain specifies the aminopep- the front region of the are clearly visible in the tidase activity is shown in detail in the structure of cathep- electron-density maps. This is the result of these carbohy- sin H presented here. drate rings and the mini-chain being closely packed in the active-site cleft (Figure 3). The Asn79P sidechain is glyco- Results sylated, but the observed electron-density map extending Papain numbering of residues is used throughout this from its sidechain is not clear enough to allow us to recog- paper. The insertions of cathepsin H residues are denoted nize sugar rings. by appending a letter to the sequence number preceding each insertion, starting with ‘A’. The mini-chain residues The active-site cleft are numbered as they occur in pro-cathepsin H, as deduced The active-site cleft of a papain-like cysteine protease from the cDNA sequence, with the letter ‘P’ added. runs across the top of the molecule. The cleft has broad ends but narrows in the middle, where the catalytic Overall structure residues Cys25, His159 and Gln19 reside. In the cathepsin The cathepsin H molecule is well defined by the electron- H structure, the front part of the cleft is filled with carbo- density map, the only exceptions being the N-terminal hydrate residues on the left and mini-chain in the middle, residue, Tyr1, the N-terminal residue of the mini-chain whereas the back part is unoccupied and thus free for sub- Glu76P and a few sidechains. The structure confirms the strate binding (Figure 3). pattern of the disulfide bridges in bovine cathepsin H [35], including that formed between Cys205 and the mini- The backbone and sidechains of the catalytic residues are chain Cys80P. The three disulfide bridges Cys22–Cys63, found at the positions usual for a papain-like enzyme. The Cys56–Cys95 and Cys154–Cys200 are topologically equiva- only exception is the His159 imidazole ring which in, con- lent to the disulfide bridges in actinidin [36]. Cathepsin H trast to all other known structures of cysteine proteases, is an ellipsoidal molecule with dimensions 32 × 26.5 × 24 Å does not form a thiolate–imidazolium ion pair with Cys25. (calculated using GRASP [37]). Its fold is typical of a papain The His159 is instead rotated 80° about the Cα–Cβ bond [38] superfamily member. The structure consists of right and its Nε proton is involved in a salt bridge with the (R-) and left (L-) domains, forming the body of the enzyme, C-terminal carboxyl group of Val212A from a neighboring Research Paper Crystal structure of porcine cathepsin H Guncar et al. 53

Figure 1

The overall structure of cathepsin H. (a) (a) Ribbon diagram of the body of cathepsin H (blue ribbons) in the standard view with stick representations of the catalytic residue Cys25 (white) at the top of the central helix, the mini-chain in the active-site cleft (orange) and the carbohydrate moiety connected to Asn112 (red). The nitrogen, oxygen and sulfur atom sticks in the mini-chain and Cys25 are colored blue, red and yellow, respectively. The figure was prepared using the program RIBBONS [66]. (b) Stereo picture of the Cα trace of cathepsin H (thick lines) in standard orientation, superimposed with actinidin (thin lines). Marked residues are those of cathepsin H. Only the disulfide bridge between the mini- chain and the body of cathepsin H is shown. The figure was produced using the program MAIN [61]. (b) 20 180 20 180

87 138 87 138 25 25 50 159 148 50 159 148 C C 97 97 30 155D 190 30 155D 190 68 164 170 68 164 170 58B 80P 168C 58B 80P 168C 78 20510 78 20510 107 40 118 107 40 118 C 127 C 127 N N N N Structure

molecule in the crystal (Figure 4). Its Nδ hydrogen atom occupy the cavity behind the His159 imidazole ring, where stabilizes the orientation of the ring on the other side by this residue would normally have been expected to be forming a hydrogen bond with Oδ1 from the preceding found. We account for this unusual positioning of the histi- residue, Asn158. Two well-defined solvent molecules dine imidazole ring by crystal packing. It is evident,

Figure 2

cath YPPSMDWRKKG--NF--VS-PVKNQGSCGSCWTFSTTGALESAVAIATG-KMLSLAEQQLVDCAQNFN--NH 9pap IPEYVDWRQ-K--GA--VT-PVKNQGSCGSCWAFSAVVTIEGIIKIRTG-NLNQYSEQELLDCDR--R--SY 2act LPSYVDWRS-A--GA--VV-DIKSQGECGGCWAFSAIATVEGINKITSG-SLISLSEQELIDCGRT-QN-TR 1huc LPASFDAREQWPQC-PTI-KEIRDQGSCGSCWAFGAVEAISDRICIHT-NVSVEVSAEDLLTCCGSM--CGD cath GCQGGLPSQAFEYIRYNKGIMGE------DTYPYKGQ------D-D-HCK-F---QP- 9pap GCNGGYPWSALQLVAQY-GIHYR------NTYPYEGV------Q-R-YCR-SREK--- 2act GCDGGYITDGFQFIINDGGINTE------ENYPYTAQ------D-G-DCDV--ALQD- 1huc GCNGGYPAEAWNFWTRK-GLVSGGLYESHVGCRPYSI-PPCEHHVNGSRPPCTGEGDTP-K-CSK---I--- cath ------DK--AIAFVKDVANITMNDEEAMVEAVALYNPVSFAFEV-TNDFLMYRKGIYSST-SCHKTPD 9pap ------GP---YAAKTDGVRQVQPYNQGALLYSIA-NQPVSVVLQAAGKDFQLYRGGIFVG-PCGN---- 2act ------Q--KYVTIDTYENVPYNNEWALQTAVT-YQPVSVALDAAGDAFKQYASGIFTG-PCGT---- 1huc CEPGYSPTYK-QDK-HYGYNSYSVSN-SEKDIMAEIYKNGPVEGAFSV-YSDFLLYKSGVYQ--H------cath ------KVNHAVLAVGYGEENGIPYWIVKNSWGPQWGMNGYFLIERG--K--NMCGLAACASYPIPL-V-- 9pap ------KVDHAVAAVGYGP----NYILIKNSWGTGWGENGYIRIKRGTGNSYGVCGLYTSSFYPVK-N--- 2act ------AVDHAIVIVGYGTEGGVDYWIVKNSWDTTWGEEGYMRILRNV-GGAGTCGIATMPSYPVK-Y--- 1huc VTGEMMG--GHAIRILGWGVENGTPYWLVANSWNTDWGDNGFFKILRG--Q--DHCGIESEVVAGIPR--TD Structure

Structure-based alignment of the sequences of cathepsin H (cath) with papain (9pap [39]), actinidin (2act [36]) and cathepsin B (1huc [34]). Identical residues are shaded in black, similar residues in gray. 54 Structure 1998, Vol 6 No 1

Figure 3

Stereo view of the mini-chain binding in the active-site cleft. residues mentioned in text are marked. The structure is color coded as in Figure 4. The N-terminal Glu76P residue, which is undefined by electron density, is drawn in green. Connolly surface was calculated without the mini-chain. The figure was prepared with the program MAIN [61].

however, that the catalytic histidine sidechain is capable of consisting of residues Gln64–Gln70 comes closer (0.3 Å) adopting various conformations within the active-site cleft. to the R-domain. In addition, the latter residues are pre- ceded by a short insertion of two residues (Asn58A and The mini-chain and the Asn112-bound carbohydrate region Phe58B) which extend the loop between Gln58 and The mini-chain runs in an extended conformation in the Asn59 into the active-site cleft. The narrowest place in the substrate-binding direction along the front side of the cleft is between Leu67 and Pro155C, in the middle of active-site cleft (Figures 3 and 5). The residues Pro77P which is Cys205, which is involved in the disulfide bridge to Cys80P are in an extended β-strand conformation. with the mini-chain, completely closing the active-site Ser81P and Ala82P continue the β-strand-like orientation cleft. These three residues are involved in noncovalent of peptide bonds, but the chain forms a twist, which interactions and actually split the active-site cleft into two orients the C-terminal Thr83P down against the bottom of parts (Figures 3 and 6). the L- and R-domains. The structure confirms that the mini-chain residue Cys80P is covalently linked to the R- In addition to the mini-chain, part of a carbohydrate chain domain through a disulfide bridge to Cys205. runs across the R- and L-domain interface. Three sugar rings starting from Asn112 are unambiguously defined in The active-site cleft in the region of the bound mini-chain the electron-density map. Their structure is in agreement is narrower than those of other related structures. The with the nuclear magnetic resonance (NMR) observations major narrowing feature on the right side is an insertion [40] that these are two N-acetylglucosamine residues loop of four residues (Lys155A, Thr155B, Pro155C and (NAG) and a mannose residue, linked by β(1–4) glycosidic Asp155D), while on the left side a part of the L-domain bonds. The rings shield a substantial part of surface of the

Figure 4

Stereo view of the final |2|Fobs|–|Fcalc|| electron-density map of the cathepsin H active site, contoured at 1.5σ (blue), superimposed on a stick model of cathepsin H. The main body of cathepsin H molecule is shown in blue, the mini-chain in orange and the symmetry-related molecule in pink. The oxygen, nitrogen and sulfur atoms are represented as small spheres in red, blue and yellow, respectively. Water molecules are drawn as pink spheres. The figure was prepared with the program MAIN [61]. Research Paper Crystal structure of porcine cathepsin H Guncar et al. 55

Figure 5

Stereo view of the ||Fobs|–|Fcalc|| kicked omit electron-density map of the mini-chain contoured at 1.3σ (blue). Structure factors for the kicked omit map were calculated from randomly displaced atoms, up to 0.3 Å along each coordinate. The mini-chain is shown in orange sticks, the oligosaccharide moiety in red, catalytic Cys25 and His159 in yellow and all other residues, which are represented only as a mainchain trace, in blue. The figure was prepared with the program MAIN [61].

residues Val110, Glu73, Phe58B and Tyr77 from solvent. anchored to the bottom of the front pocket, where it spans The ring attached to Asn112 is involved in interactions between the Oγ atom of Ser69 and carbonyl group of with the vicinal Pro77P of the mini-chain. Ala206, thus contributing two additional hydrogen bonds bridging the L- and R-domains. The rest of the pocket is The mini-chain binds into the active-site cleft through the filled with three well-defined solvent molecules. The sidechains of Gln78P, Cys80P and Thr83P (Figure 6). The strong binding of the Gln78P residue in the pocket is other sidechains point away from the bottom of the cleft demonstrated by its atomic temperature factors (18.0 Å2), into the solvent. Their conformation is not as clearly which are the lowest in the mini-chain. defined as the rest of the mini-chain structure and the posi- tion and conformation of the N-terminal Glu76P residue is The pocket behind Leu67 and the mini-chain disulfide is not defined. No mainchain interactions with the underlying the structural equivalent of the S2- in other enzyme surface, other than the C-terminal Thr83P and cysteine proteases. In cathepsin H, this space is filled by Ser81P, are observed. However, a hydrogen bond is formed Thr83P. The sidechains of Ala133 and Val157 from the R- between the peptide bond amide hydrogen atom of Gln78P domain provide the hydrophobic part of the surface, and the O6 atom of sugar residue NAG112A. against which the methyl group of the Thr83P sidechain is packed. The carbonyl and amide groups of Gly66, The hydrophobic wall formed by the sidechain contacts of together with the aromatic ring of Trp26, provide a favor- Leu67, disulfide bond Cys205–Cys80P and Pro155C splits able environment for binding the amide and carboxyl the active-site cleft underlying the mini-chain into two groups of Thr83P. parts. Each part contains a pocket. The surface of the front pocket has primarily hydrophilic character, whereas the Discussion surface of the back pocket is more hydrophobic (Figure 3). The mini-chain defines the cathepsin H aminopeptidase The backside of the front pocket is formed by the side- function chains of Ser9 and Ser207 and a solvent molecule placed All papain-like cysteine proteases have the same basic beneath Leu67 and Cys205. On the right, there is the mechanism of action, but their specificity and point of glycosylated Asn112 residue, spanning with its carbohy- cleavage in the substrate differs from one member of the drate moiety to the left, and on the right, there are several family to another. The active-site cleft of cysteine endo- carbonyl oxygen atoms and amide protons pointing into peptidases like papain [38], actinidin [36], and cathepsins the pocket. From the top right, Met115 extends its hydro- L [33] and K [31,32] is unoccupied and so able to bind phobic surface into the hydrophobic region of Pro155C substrates along its full length, whereas the active-site clefts and Cys205. The amide group of the Gln78P sidechain is of cathepsins H and B, both exopeptidases, are partially 56 Structure 1998, Vol 6 No 1

Figure 6

Schematic representation of the mini-chain binding. The mini-chain (bold lines) is TRP177 covalently attached by the Cys80P–Cys205 GLN19 disulfide bridge to the R-domain. Three other N residues of the mini-chain form hydrogen bonds (dashed lines) and stabilize the mini- chain position. CENTRAL HELIX N CYS25 N

HIS159 N ASN158 GLY66 THR83P N N O ALA82P ASP155D LEU67 SER81P LYS155A PRO155C THR155B

CYS80P ALA206

SER69 CYS205 HO N 2 ALA204

ASN79P HO2 HO2 GLN78P MET115 NAG112A N NAG112B MAN12C1 ASN112 PRO77P

GLU76P

Structure

filled, thereby limiting the free substrate-binding sites. In The interatomic distance (peptide bond) between the P2 cathepsin B, a carboxydipeptidase, the active-site cleft carbon atom and P1 nitrogen atom is at least 1.5 Å shorter behind the S2′-binding site is occupied [34,41,42] and in than the interatomic distance between the C-terminal cathepsin H, an aminopeptidase, the active-site cleft in carbon atom of the mini-chain and the N-terminal nitrogen front of the S1-binding site is blocked. atom of a P1 residue, which are involved in a non-covalent interaction. Thus, it would appear that there could not be Cathepsin H utilizes a part of its propeptide, the mini- sufficient space for a substrate to bind productively. This chain, to partially fill the cleft and to provide a negatively difference is accommodated by the cathepsin H structure charged C-terminal carboxyl group of the mini-chain, which as follows. The relative positioning of the central catalytic presumably interacts with the positively charged N-terminal residues (Cys25, His159 and Gln19) is preserved through- amino group of a substrate P1 residue (Figure 7). The out the structure of cathepsin H as compared with those of mini-chain C-terminal residue Thr83P is located in, what actinidin [36], papain [39] and cathepsin B [34]. However, in the papain structure, is the S2-binding site. in cathepsin H, the distance between the Cα atoms of the Research Paper Crystal structure of porcine cathepsin H Guncar et al. 57

Figure 7

Stereo view of the modeled substrate in the active site. The structure of the body of cathepsin H (blue sticks) is shown in the Connolly surface (blue dots) with the predicted substrate (green sticks) in the active site. The mini-chain is represented as orange sticks and was not included in Connolly surface calculation. The catalytic residue His159 was rotated about its Cα–Cβ bond to form an active thiolate–imidazolium ion pair with Cys25. The N terminus of the substrate is anchored to the C-terminal carboxyl of Thr83P. Non-carbon atoms are represented as small spheres, color coded as in Figure 4. Possible hydrogen bonds are drawn as dashed pink lines. The figure was prepared with the program MAIN [61].

catalytic residue Cys25 and Trp177 is shorter, whereas and the positioning of the corresponding P1, P1′ and P2′ there is more space between the Cα atom of Cys25 and the substrate residues can be predicted on the basis of the Cα atoms of residues equivalent to those forming the S2 related structures. Four substrate-binding subsites — S2, subsite in papain, actinidin and cathepsin B (Figure 8). The S1, S1′ and S2′ — of the cysteine proteases were unam- shift in position of the scissile bond in the substrate may not biguously identified with the help of the crystal structures be sufficient to accommodate all the required increase of of substrate analogue chloromethyl-ketone inhibitors in interatomic distance. An additional source of accommoda- complex with papain [43] and the structure of the epoxy- tion to substrate might come from the mini-chain. The succinyl-based inhibitor (CA030) in complex with cathep- mini-chain Thr83P residue is placed in a pocket much sin B [44]. The fifth site, the S3 substrate-binding site, is larger then the residue itself, suggesting that when sub- somewhat ambiguous, since the complexes between the strate binds to cathepsin H, the Thr83P responds with a substrate analogue chloromethyl-ketone inhibitors and shift towards the hydrophobic wall of Leu67 and the disul- papain and the structure of the stefin–papain complex [45] fide. This would allow substrate to bind further towards its indicate different binding possibilities. On this basis, the N-terminal direction, thus bringing the Cys25 thiol into line substrate model for cathepsin B was generated [44] and with the scissile bond. then transferred to the cathepsin H environment by super- position of the protein structures. The P2 residue was Interactions with substrate removed from the substrate model. In cathepsin H, there are no S3 or S2 substrate-binding subsites, because the region equivalent to the related The substrate P1 residue can form an ionic interaction endopeptidase subsites is occupied by the mini-chain with the mini-chain C-terminal carboxyl group. Its car- residues. The S1, S1′ and S2′ subsites are, however, free bonyl occupies the and the carbonyl of the

Figure 8

Stereo view of the superimposed structures of TRP177 TRP177 cathepsin H (thick lines), cathepsin B (medium lines) and papain (thin lines) showing GLN19 GLN19 the displacement of the residues in the active- site cleft of cathepsin H. Disulfide bridges are shown as dashed lines. Cathepsin B (1huc) CYS25 CYS25 and papain (9pap) structures were HIS159 HIS159 superimposed on cathepsin H, minimizing rms deviations of topologically equivalent Cα atoms. GLY65 GLY65 GLY66 GLY66

CYS205 CYS205

Structure 58 Structure 1998, Vol 6 No 1

P1′ residue forms a hydrogen bond with the hydrogen of shown that E-64 binds into the S1, S2 and S3 substrate- the Trp177 Nε atom (Figure 7). The sidechain interac- binding sites of the proteases. If it were to bind in the tions can, however, not be predicted as exactly as the same way to cathepsin H, it is evident that it would mainchain interactions because of lack of structural data, compete with the mini-chain residues for the same binding but their binding sites on the cathepsin H surface can be sites. This could be the reason for the slow binding of clearly located, allowing us to discuss their specificity. E-64 to cathepsin H compared to other related enzymes [21]. For the same reason, leupeptin [24] and diazo- Cathepsin H shows a preference for residues with large methane dipeptides [25] do not bind to the enzyme. hydrophobic (Phe, Trp, Leu and Tyr) or basic (Arg and Lys) sidechains at the P1 position over residues containing Modeling of complexes between stefins and cathepsin H small or polar sidechains [26]. The predicted substrate (models not shown), based on the stefin B–papain complex model shows that the P1 residue sidechain points along [45], raises questions about the structure of these com- the wall of the active-site cleft towards the solvent region. plexes. According to the models, both hairpin loops of There are no special groups on the surface of the S1 human stefin B fit well into the active-site cleft of cathep- subsite that favor a particular type of residue, so the larger sin H, but the N-terminal trunk of stefin B collides with amino-acid sidechains at the P1 position presumably bind the mini-chain residues, in particular with the Thr83P. more strongly, owing to their more extensive contact surface with the wall of the enzyme. Processing The sequence of preprocathepsin H was obtained from The S1′ subsite is, like that of the other enzymes in the the cDNA sequence. In vivo, the prepeptide is cleaved co- family, expected to be more specific, because the sidechain translationally, and the proenzyme is processed post-trans- of a P1′ substrate residue can form a relatively large contact lationally during lysosome maturation [49]. surface within a pocket. No experimental data for the speci- ficity of this site, however, are known to us. The residues As seen from the amino acid sequence and the structure, Val136, Leu142 and Phe141 form the walls of the pocket. maturation of cathepsin H includes at least three cleav- These residues are identical in composition and conforma- ages of the proenzyme chain: at the enzyme N terminus tion to those of cathepsin B, suggesting that binding of and at both termini of the mini-chain. During processing, isoleucine and leucine may be preferred, as in the case of a ten-residue peptide, connecting the mini-chain with the cathepsin B. There is, however, no equivalent residue to N-terminal residue of the enzyme Tyr1 and the propep- Met196 in cathepsin B, which encloses the hydrophobic tide chain preceding the mini-chain, is removed. surface of that S1′ subsite from the top. However, Asn158 and Val157, preceding His159, are placed similarly to the Structures of related zymogens of procathepsin B [42,50, equivalent residues in the endopeptidase structures. The 51], procathepsin L [52] and procaricain [53] have revealed chain preceding these residues, including a short four- that the propeptide of a cysteine protease binds along the residue insertion, follows a unique route before it merges active-site cleft in the direction opposite to that of the at the position of Tyr149 into the path of the chain of the bound substrate. The mini-chain, however, as part of the other related structures. This part actually narrows the cathepsin H propeptide, binds in the mature enzyme active-site cleft and thereby helps to stabilize the position- along the active-site cleft in the substrate-binding direc- ing of the mini-chain residues within the cleft. The pre- tion. The question here is whether the positioning of the dicted S2′ subsite of cathepsin H is quite broad and exposed mini-chain observed in the mature enzyme is present also to solvent. The structure reveals no particular feature, and in the zymogen form. Assuming that the zymogen is a we may assume that hydrophilic and aromatic residues, due form designed in order to allow controlled activation of the to exposure of the site to solvent, would be preferred. enzyme, the cathepsin H propeptide should not bind along the active-site cleft in the substrate-binding direc- In the light of recent work [46], the unusual positioning of tion, otherwise it would be cleaved. This means that the the His159 imidazole ring may not be as unusual as it mini-chain would be positioned differently in the zymogen appears. Negative charges from the top or bottom of the from the mature enzyme. enzyme may provide an electrostatic switch between the active form of a cysteine protease, in which the thiolate– On this basis, we suggest that the processing of procathep- imidazole pair is formed, and an inactive form in which sin H occurs in at least three steps, in which the molecule the positively charged histidine imidazole ring is exposed is transformed from its zymogen form, via an enzymati- to solvent. cally active intermediate exhibiting endopeptidase activ- ity, into an aminopeptidase. In the first step, a cleavage Interactions with cysteine protease inhibitors liberates the mini-chain N terminus from the main body The crystal structures of complexes of the inhibitor E-64 of the propeptide chain. The cleavage site is probably at with actinidin [47], papain [48] and cathepsin L [33] have the mini-chain N-terminal residue Glu76P. In the second, Research Paper Crystal structure of porcine cathepsin H Guncar et al. 59

step a peptide bond close to the mature enzyme N termi- of proteolytic activity, rather than resorting to a basi- nus is cleaved. The mini-chain remains attached to the cally different structure. The modification also involves body of the molecule only by means of the disulfide bond. the use of the pre-existing pro-peptide which evolved In the third step, the mini-chain reorients and positions originally for the different purpose of controlled switch- along the active-site cleft in the direction of a bound sub- ing of the proteolytic activity. strate and gets processed to its eventual C terminus (Thr83P) by the catalytic site of the same molecule. After Materials and methods this cleavage, the mini-chain pulls back and adopts the Isolation observed conformation and thus turns the enzyme into an Cathepsin H was isolated from porcine spleen. The purification proce- dure was based on the method described by Popovic et al. [9]. The aminopeptidase. The N terminus of the enzyme is addi- purity and identity of cathepsin H was checked by SDS–PAGE and N- tionally processed to its final sequence. terminal amino-acid sequence analysis. The purified protein was con- centrated in a spin concentrator (Centricon (Amicon)) to a concentration Biological implications of 11 mg/ml. N-terminal analysis of the reduced enzyme showed three sequences that correspond to heavy, light and mini-chain. Cathepsin H Papain-like cysteine proteases are important enzymes was isolated in two- and three-chain forms. Both forms contain the involved in intracellular protein turnover. They can also eight-residue long mini-chain. The three-chain form has N-terminal degrade extracellular matrix components, for which heavy and C-terminal light chains, consisting of 177 and 43 amino acid reason they are associated with numerous pathological residues respectively. conditions. All of these enzymes have the same basic Primary structure determination (cDNA cloning) mechanism of action, but their specificity and point of Total RNA was isolated from the buffy coat obtained from fresh heparin- cleavage in the substrate differs from one member of the treated porcine peripheral blood by homogenization in guanidinium thio- family to another. Endopeptidases cleave to cyanate [54] and centrifugation in a solution of cesium trifluoroacetate + large fragments and exopeptidases, like cathepsin H, [55]. Poly(A) RNA was purified on oligo(dT)-cellulose and cDNA syn- thesized using an Amersham cDNA synthesis system. EcoRI–XhoI digest those fragments to single amino acids. adapters (Stratagene) were ligated to the cDNA and a cDNA library was constructed in λgt11 using a kit (Amersham). A 939 bp fragment of Cathepsin H is a lysosomal proteolytic enzyme, special- human procathepsin H cDNA (prepared by polymerase chain reaction ized to remove the N-terminal residue from proteins and (PCR) from human placental cDNA using primers corresponding to the 35 α peptides. Its sequence shows it to belong to the papain sequence reported by Fuchs et al. [56]) was labeled by [ S]dATP S (Amersham) using Random Primed DNA Labeling Kit (Boehringer). A family of enzymes, with which it shares a common three- portion of the cDNA library of about 105 plaque-forming units was spread dimensional structure, with a thiol group as one of the on LBA plates and blotted to Hybond-N membranes (Amersham). The catalytic residues. It is synthesized as an inactive proen- membranes were hybridized with a 35S-labeled human procathepsin H zyme from which it is activated by controlled proteolysis. fragment according to Sambrook et al. [57]. A single positive plaque was picked and subjected to two rounds of purification. From the posi- The X-ray structure described here confirms that the tive clone, λ DNA was isolated, digested with EcoRI and both the result- overall structure is similar to that of other cysteine pro- ing fragments, of about 1.15 kb and 1 kb, inserted into pUC19. The teases, but it shows how the basic endopeptidase structure 1.15 kb fragment, which comprised the entire coding sequence of is modified to achieve the aminopeptidase activity. porcine preprocathepsin H and 115 bp of the 3’ untranslated region, was completely sequenced on both strands using a T7 sequencing kit (Pharmacia). The cDNA sequence of porcine preprocathepsin H cDNA It has previously been shown that, on activation, a small reported here has been submitted to the DDBJ/EMBL/GenBank data- portion of the pro-region — the mini-chain — remains bases under accession number AF001169. bound to the enzyme through a disulfide bond. The structure of the enzyme described here shows that this Crystallization, X-ray data collection and structure determination mini-chain is bound in one end of the active-site cleft, Crystals were grown by the sitting drop vapor diffusion method. The thus limiting the size of substrate that can be bound and reservoir contained 1 ml 0.05 M sodium acetate and 5% mme-PEG 5K cleaved. There is room for only one residue to be bound buffer pH 4.8. The drop was composed of 2 µl reservoir solution and on the N-terminal side of the catalytic site. Other minor 2 µl cathepsin H in 20 mM sodium acetate and 1 mM EDTA buffer structural changes are seen that fine tune the location of pH 5.0. Concentration of applied protein was 11 mg/ml. the substrate, so that the peptide bond joining the N-ter- Diffraction data were collected from a single crystal using Cu Kα radia- minal residue to the rest of the protein is located pre- tion from a Rigaku rotating anode X-ray generator and recorded on cisely over the thiol group. The mini-chain also performs 18 cm MAR Research image plate detector. Auto indexing and scaling another role in substrate binding. Its position and orien- was done using HKL programs [58]. The crystal diffracted beyond 1.98 Å resolution and belonged to the orthorhombic space group tation are such that its negatively charged terminal car- P212121 with cell dimensions a = 38.4 Å, b = 68.7 Å, c = 86.8 Å. It boxyl group points into the cleft, where it can interact contained one molecule in the asymmetric unit. Orientation and transla- with the N-terminal amino group of the substrate. tion of the cathepsin H molecule were determined using a molecular replacement method implemented in AMoRe program [59]. The crystal This structure provides an informative example of structures of papain [39], actinidin [36], papaya proteinase omega [60] and cathepsin B [34] together and separately were used as a search economy in evolution. It shows how one type of 3D model. A set of consistent solutions was found, the actinidin structure structure has been modified to produce a different kind giving the best result, with a correlation factor of 0.29 and R value of 60 Structure 1998, Vol 6 No 1

Table 1 Accession numbers The coordinates have been deposited in the Brookhaven Protein Data Crystallographic data and refinement statistics. Bank with accession code 8pch. Data collection Acknowledgements Spacegroup P212121 We thank B Kobe for help with initial attempts of processing the data with Cell parameters (Å) a = 38.4 HKL suite of programs, Anka Ritonja for performing the N-terminal b = 68.7, c = 86.8 sequence analysis and Bojana Mozetic-Francky for a generous gift of human α = β = γ = 90° procathepsin H cDNA. RH Pain and B Turk are kindly acknowledged for Molecules in asymmetric unit 1 their critical reading of the manuscript and discussions. This work was sup- Limiting resolution (Å) 1.98 ported by the Slovenian Ministry of Science and Technology. Measured reflections 102,384 Independent reflections 15,946 References R * 0.11 (99–2.1 Å) 1. Bond, J.S. & Butler, P.E. (1987). Intracellular proteases. Annu. Rev. sym 56 Completeness 0.98 (99–2.1Å) Biochem., , 333–364. 2. Tsushima, H., Ueki, A., Matsuoka, Y., Mihara, H. & Hopsu-Havu, V.K. Final refinement parameters (1991). Characterization of a cathepsin-H-like enzyme from a human No. scattering protein atoms 1803 melanoma cell line. Int. J. Cancer 48, 726–732. No. solvent molecules 179 3. Gabrijelcic, D., et al., & Turk, V. (1992). Cathepsins B, H and L in Resolution range in refinement (Å) 8.0–2.1 human breast carcinoma. Eur. J. Clin. Chem. Clin. Biochem. 30, Reflections used in refinement (1σ cutoff) 12,932 69–74. R factor (1σ cutoff) (%) 19.8 4. Schweiger, A., Stabuc, B., Popovic, T., Turk, V. & Kos, J. (1997). σ † Enzyme-linked immunosorbent assay for the detection of total Rfree (1 cutoff) (%) 24.5 Geometry of the final model cathepsin H in human tissue cytosols and sera. J. Immunol. Methods 201, 165–232. Rms deviation of bond distances (Å) 0.012 5. Kopitz, J., Arnold, A., Meissner, T. & Cantz, M. (1993). Protein Rms deviation of bond angles (°) 1.53 catabolism in fibroblasts cultured from patients with mucolipidosis II B-factor values (rms by bonds) and other lysosomal disorders. Biochem. J. 295, 577–650. Overall (rms deviation) (Å2) 13.7 (4.4) 6. Chapman, H.A., Riese, R.J. & Shi, G.-P. (1997). Emerging roles for Mini-chain (rms deviation) (Å2) 27.2 (4.8) cysteine proteases in human biology. Annu. Rev. Physiol. 59, 63–88. Cathepsin H without mini-chain and water 11.7 (4.4) 7. Machleidt, W., et al., & Mueller-Esterl, W. (1986). Human cathepsins (rms deviation) (Å2) B, H and L: characterization by amino acid sequences and some kinetics of inhibition by the kininogens. In Cysteine Proteinases and Σ Σ 〈 〉 /ΣΣ 〈 〉, 〈 〉 their Inhibitors. (Turk, V., ed.), pp. 3–18. Walter de Gruyter and Co., *Rsym = h i | Iih– Ih | h i Ih where Ih is the mean intensity of symmetry-equivalent reflections. †R was calculated using 10% of Berlin-New York. free 8. Takio, K., Towatari, T., Katunuma, N., Teller, D.C. & Titani, K. (1983). reflections. Homology of amino acid sequences of rat liver cathepsins B and H with that of papain. Proc. Natl. Acad. Sci. USA 80, 3666–3670. 9. Popovic, T., et al., & Turk, V. (1988). A new purification procedure of 0.51 using data in the 15 Å to 4.0 Å resolution range. The crystal pack- human kidney cathepsin H, its properties and kinetic data. Biol. Chem. aging was inspected on a graphical display using MAIN [61]. The first Hoppe Seyler 369, 175–253. density maps were calculated using the actinidin structure and reflec- 10. Ritonja, A., Popovic, T., Kotnik, M., Machleidt, W. & Turk, V. (1988). tion data between 15 Å and 3.5 Å resolution. The subsequent interac- Amino acid sequences of the human kidney cathepsins H and L. FEBS Lett. 228, 341–305. tive sessions using MAIN included model building, crystallographic 11. Lafuse, W.P., Brown, D., Castle, L. & Zwilling, B.S. (1995). IFN-gamma refinement and structure analysis. The data resolution was gradually increases cathepsin H mRNA levels in mouse macrophages. expanded to the final range (8.0 Å–2.1 Å). Refinement also included J. Leukocyte Biol. 57, 663–609. simulated annealing and conjugate gradient minimization procedures in 12. Ishidoh, K., Kominami, E., Katunuma, N. & Suzuki, K. (1989). X-PLOR [62]. Standard parameters for protein [63] and carbohydrate structure of rat cathepsin H. FEBS Lett. 253, 103–107. [64] parameters, as provided by X-PLOR, were used for geometry reg- 13. Fuchs, R., Machleidt, W. & Gassen, H.G. (1988). Molecular cloning ularization. Kicked omit maps were used throughout the structure and sequencing of a cDNA coding for mature human kidney cathepsin determination process to reveal ambiguous parts of the structure. H. Biol. Chem. Hoppe Seyler 369, 469–535. 14. Turk, B., Turk, V. & Turk, D. (1997). Structural and functional aspects Structure factors for the kicked omit maps were calculated from ran- of papain-like cysteine proteinases and their protein inhibitors. Biol. domly displaced atoms, up to 0.3 Å along each coordinate. Water mol- Chem. 378, 141–150. ecules were generated using an automatic procedure in MAIN and 15. Rogers, J.C., Dean, D. & Heck, G.R. (1985). Aleurain: a barley thiol then corrected manually in several steps during refinement. After the protease closely related to mammalian cathepsin H. Proc. Natl. Acad. crystallographic R value dropped to 0.25 at 2.1 Å resolution, the tem- Sci. USA 82, 6512–6516. perature factor refinement was applied. The mini-chain was built into a 16. Watanabe, H., Abe, K., Emori, Y., Hosoyama, H. & Arai, S. (1991). kicked omit map. The mini-chain direction was additionally verified Molecular cloning and gibberellin-induced expression of multiple using an R criterion: the R value for the structure with the mini-chain cysteine proteinases of rice seeds (oryzains). J. Biol. Chem. 266, free 16897–16902. oriented in the correct direction was 0.236 and the R value 0.267, free 17. Jerala, R., Kroon-Zitko, L., Popovic, T. & Turk, V. (1994). Elongation on while the values for the structure with the mini-chain running in the the amino-terminal part of stefin B decreases inhibition of cathepsin H. opposite direction were 0.251 and 0.316, respectively. The final refine- Eur. J. Biochem. 224, 797–802. ment included all reflections and yielded the final model at 2.1 Å resolu- 18. Turk, B., et al., & Turk, V. (1995). Regulation of the activity of tion, with the crystallographic R value being 0.198 and the Rfree value lysosomal cysteine proteinases by pH-induced inactivation and/or for 10% of the reflections, 0.245. The geometry of the final model was endogenous protein inhibitors, cystatins. Biol. Chem. Hoppe-Seyler inspected with MAIN and PROCHECK [65]. All residues lie in the 376, 225–230. allowed regions of the Ramachandran plot, 84.3% of residues in most 19. Lenarcic, B., Krizaj, I., Zunec, P. & Turk, V. (1996). Differences in specificity for the interactions of stefins A, B and D with cysteine favored regions and 15.7% in additional allowed regions. Other struc- proteinases. FEBS Lett. 395, 113–108. tural parameters for the structure refined at this resolution show no sig- 20. Abrahamson, M., Mason, R.W., Hansson, H., Buttle, D.J., Grubb, A. & nificant deviations from the expected values (overall G factor is 0.2 Ohlsson, K. (1991). Human C. Role of the N-terminal which is better than expected for 2.1 Å resolution). The crystallographic segment in the inhibition of human cysteine proteinases and in its data and refinement statistics are summarized in Table 1. inactivation by leucocyte elastase. Biochem. J. 273, 621–606. Research Paper Crystal structure of porcine cathepsin H Guncar et al. 61

21. Barrett, A.J., et al., & Hanada, K. (1982). L-trans-Epoxysuccinyl- 46. Pinitglang, S., et al. & Brocklehurst, K. (1997). A classical enzyme leucylamido(4-guanidino)butane (E-64) and its analogues as inhibitors active center motif lacks catalytic competence until modulated of cysteine proteinases including cathepsins B, H and L. Biochem. J. electrostatically. Biochemistry 36, 9968–9982. 201, 189–278. 47. Varughese, K.I., Su, Y., Cromwell, D., Hasnain, S. & Xuong, N.H. 22. Shi, G.P., Munger, J.S., Meara, J.P., Rich, D.H. & Chapman, H.A. (1992). Crystal structure of an actinidin-E-64 complex. Biochemistry (1992). Molecular cloning and expression of human alveolar 31, 5172–5176. macrophage , an elastinolytic cysteine protease. J. Biol. 48. Varughese, K.I., Ahmed, F.R., Carey, P.R., Hasnain, S., Huber, C.P. & Chem. 267, 7258–7262. Storer, A.C. (1989). Crystal structure of a papain-E-64 complex. 23. Bossard, M.J., et al., & Levy, M.A. (1996). Proteolytic activity of human Biochemistry 28, 1330–1332. . Expression, purification, activation, and 49. Nishimura, Y., Kawabata, T., Yano, S. & Kato, K. (1990). Inhibition of substrate identification. J. Biol. Chem. 271, 12517–12524. intracellular sorting and processing of lysosomal cathepsins H and L 24. Schwartz, W.N. & Barrett, A.J. (1980). Human cathepsin H. Biochem. at reduced temperature in primary cultures of rat hepatocytes. Arch. J. 191, 487–577. Biochem. Biophys. 283, 458–513. 25. Angliker, H., Wikstrom, P., Kirschke, H. & Shaw, E. (1989). The 50. Cygler, M., Sivaraman, J., Grochulski, P., Coulombe, R., Storer, A.C. & inactivation of the cysteinyl exopeptidases cathepsin H and C by Mort, J.S. (1996). Structure of rat procathepsin B: model for inhibition affinity-labelling reagents. Biochem. J. 262, 63–68. of cysteine protease activity by the proregion. Structure 4, 405–416. 26. Takahashi, T., Dehdarani, A.H. & Tang, J. (1988). Porcine spleen 51. Turk, D., Podobnik, M., Kuhelj, R., Dolinar, M. & Turk, V. (1996). cathepsin H hydrolyzes oligopeptides solely by aminopeptidase Crystal structures of human procathepsin B at 3.2 and 3.3 Ångstroms activity. J. Biol. Chem. 263, 10952–10957. resolution reveal an interaction motif between a papain-like cysteine 27. Rothe, M. & Dodt, J. (1992). Studies on the aminopeptidase activity of protease and its propeptide. FEBS Lett. 384, 211–204. rat cathepsin H. Eur. J. Biochem. 210, 759–764. 52. Coulombe, R., Grochulski, P., Sivaraman, J., Menard, R., Mort, J.S. & 28. Dolenc, I., Turk, B., Pungercic, G., Ritonja, A. & Turk, V. (1995). Cygler, M. (1996). Structure of human procathepsin L reveals the Oligomeric structure and substrate induced inhibition of human molecular basis of inhibition by the prosegment. EMBO J. 15, cathepsin C. J. Biol. Chem. 270, 21626–21631. 5492–5503. 29. Aronson N.N., Jr. & Barrett, A.J. (1978). The specificity of cathepsin B. 53. Groves, M.R., Taylor, M.A., Scott, M., Cummings, N.J., Pickersgill, Hydrolysis of glucagon at the C-terminus by a peptidyldipeptidase R.W. & Jenkins, J.A. (1996). The prosequence of procaricain forms an mechanism. Biochem J. 171, 759–765. alpha-helical domain that prevents access to the substrate-binding 30. Koga, H., Mori, N., Yamada, H., Nishimura, Y., Tokuda, K., Kato, K. & cleft. Structure 4, 1193–1203. Imoto, T. (1992). Endo- and aminopeptidase activities of rat cathepsin 54. Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. & Rutter, W.J. (1979). H. Chem. Pharm. Bull. (Tokyo) 40, 965–970. Isolation of biologically active ribonucleic acid from sources enriched 31. McGrath, M.E., Klaus, J.L., Barnes, M.G. & Bromme, D. (1997). Crystal in ribonuclease. Biochemistry 18, 5294–5299. structure of human cathepsin K complexed with a potent inhibitor. Nat. 55. Mirkes, P.E. (1985). Simultaneous banding of rat embryo DNA, RNA, Struct. Biol. 4, 105–109. and protein in cesium trifluoroacetate gradients. Anal. Biochem. 148, 32. Zhao, B., et al., & Abdel-Meguid, S.S. (1997). Crystal structure of 376–383. human osteoclast cathepsin K complex with E-64. Nat. Struct. Biol. 4, 56. Fuchs, R., & Gassen, H.G. (1989). Nucleotide sequence of human 109–111. preprocathepsin H, a lysosomal cysteine proteinase. Nucleic Acids 33. Fujishima, A., Imai, Y., Nomura, T., Fujisawa, Y., Yamamoto, Y. & Res. 17, 9471. Sugawara, T. (1997). The crystal structure of human cathepsin L 57. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989). Molecular Cloning. A complexed with E-64. FEBS Lett. 407, 47–50. Laboratory Manual, 2nd edn. Cold Spring Harbor Laboratory Press, 34. Musil, D., et al., & Bode, W. (1991). The refined 2.15 A X-ray crystal Cold Spring Harbor, NY. structure of human liver cathepsin B: the structural basis for its 58. Otwinowski, Z. & Minor, W. (1996). Processing of X-ray diffraction specificity. EMBO J. 10, 2321–2330. data collected in oscillation mode. Methods Enzymol. 276, 307–326. 35. Baudyš, M., et al., & Blake, C.C. (1991). S–S bridges of cathepsin B 59. Navaza, J. (1994). AMoRe: an automated package for molecular and H from bovine spleen: a basis for cathepsin B model building and replacement. Acta Cryst. A 50, 157–163. possible functional implications for discrimination between exo- and 60. Pickersgill, R.W., Rizkallah, P., Harris, G.W. & Goodenough, P.W. endopeptidase activities among cathepsins B, H and L. Biomed. (1991). Determination of the structure of papaya protease omega. Biochim. Acta 50, 569–637. Acta Cryst. B 47, 766–771. 36. Baker, E.N. (1980). Structure of actinidin, after refinement at 1.7 A 61. Turk, D. (1992). Weiterentwicklung eines programmes für resolution. J. Mol. Biol. 141, 441–524. molekülgraphik und electronendichte-manipulation und seine 37. Nicholls, A., Sharp, K.A. & Honig, B. (1991). Protein folding and anwendungen auf verschiedene protein-strukturaufklärungen. PhD association: insights from the interfacial and thermodynamic thesis, Technische Universität München, Germany. properties of hydrocarbons. Proteins 11, 281–376. 62. Brünger, A.T. (1992). X-PLOR, Version 3.1. A System for X-ray 38. Drenth, J., Jansonius, J.N., Koekoek, R., Swen, H.M. & Wolthers, B.G. Crystallography and NMR. Yale University Press, New Haven, CT. (1968). Structure of papain. Nature 218, 929–932. 63. Engh, R.A. & Huber, R. (1991). Accurate bond and angle parameters 39. Kamphuis, I.G., Kalk, K.H., Swarte, M.B. & Drenth, J. (1984). Structure for X ray protein structure refinement. Acta Cryst. A 47, 392–400. of papain refined at 1.65 A resolution. J. Mol. Biol. 179, 233–286. 64. Jeffrey, G.A. (1990). Crystallographic studies of carbohydrates. Acta 40. Takahashi, T., Dehdarani, A.H., Schmidt, P.G. & Tang, J. (1984). Cryst. B 46, 89–103. Cathepsins B and H from porcine spleen. Purification, polypeptide 65. Laskowski, R.A., MacArthur, M.W, Moss, D.S. & Thornton, J.M. (1993). chain arrangements, and carbohydrate content. J. Biol. Chem. 259, PROCHECK: a program to check the stereochemical quality of 9874–9882. protein structures. J. Appl. Cryst. 26, 283–291. 41. Illy, C., Quraishi, O., Wang, J., Purisima, E., Vernet, T. & Mort. J. S. 66. Carson, M. (1991). Ribbons 2.0. J. Appl. Cryst. 24, 958–961. (1997). Role of the occluding loop in cathepsin B activity. J. Biol. Chem. 272, 1197–1202. 42. Podobnik, M., Kuhelj, R., Turk, V. & Turk, D. (1997). Crystal structure of the wild-type human procathepsin B at 2.5 Å resolution reveals the native active site of a papain-like cysteine protease zymogen. J. Mol. Biol. 271, 774–788. 43. Drenth, J., Kalk, K.H. & Swen, H.M. (1976). Binding of chloromethyl ketone substrate analogues to crystalline papain. Biochemistry 15, 3731–3738. 44. Turk, D., et al., & Turk, V. (1995). Crystal structure of cathepsin B inhibited with CA030 at 2.0 Å resolution: A basis for the design of specific epoxysuccinyl inhibitors. Biochemistry 34, 4791–4797. 45. Stubbs, M.T., et al., & Turk, V. (1990). The refined 2.4 Å X-ray crystal structure of recombinant human stefin B in complex with the cysteine proteinase papain: a novel type of proteinase inhibitor interaction. EMBO J. 9, 1939–1947.