Crystal Structure of Human Cystatin D, a Cysteine Peptidase Inhibitor with Restricted Inhibition Profile
2005; Elsevier BV; Volume: 280; Issue: 18 Linguagem: Inglês
10.1074/jbc.m411914200
ISSN1083-351X
AutoresMarcia Alvarez-Fernandez, Yu‐He Liang, Magnus Abrahamson, Xiao‐Dong Su,
Tópico(s)Folate and B Vitamins Research
ResumoCystatins are natural inhibitors of papain-like (family C1) and legumain-related (family C13) cysteine peptidases. Cystatin D is a type 2 cystatin, a secreted inhibitor found in human saliva and tear fluid. Compared with its homologues, cystatin D presents an unusual inhibition profile with a preferential inhibition cathepsin S > cathepsin H > cathepsin L and no inhibition of cathepsin B or pig legumain. To elucidate the structural reasons for this specificity, we have crystallized recombinant human Arg26-cystatin D and solved its structures at room temperature and at cryo conditions to 2.5- and 1.8-Å resolution, respectively. Human cystatin D presents the typical cystatin fold, with a five-stranded anti-parallel β-sheet wrapped around a five-turn α-helix. The structures reveal differences in the peptidase-interacting regions when compared with other cystatins, providing plausible explanations for the restricted inhibitory specificity of cystatin D for some papain-like peptidases and its lack of reactivity toward legumain-related enzymes. Cystatins are natural inhibitors of papain-like (family C1) and legumain-related (family C13) cysteine peptidases. Cystatin D is a type 2 cystatin, a secreted inhibitor found in human saliva and tear fluid. Compared with its homologues, cystatin D presents an unusual inhibition profile with a preferential inhibition cathepsin S > cathepsin H > cathepsin L and no inhibition of cathepsin B or pig legumain. To elucidate the structural reasons for this specificity, we have crystallized recombinant human Arg26-cystatin D and solved its structures at room temperature and at cryo conditions to 2.5- and 1.8-Å resolution, respectively. Human cystatin D presents the typical cystatin fold, with a five-stranded anti-parallel β-sheet wrapped around a five-turn α-helix. The structures reveal differences in the peptidase-interacting regions when compared with other cystatins, providing plausible explanations for the restricted inhibitory specificity of cystatin D for some papain-like peptidases and its lack of reactivity toward legumain-related enzymes. Cystatins are natural inhibitors of family C1 (papain-like) cysteine peptidases. In mammals, cystatins inhibit peptidases such as cathepsins B, H, K, L, and S both intra- and extracellularly following a reversible, tight binding mechanism (1Abrahamson M. Alvarez-Fernandez M. Nathanson C.M. Biochem. Soc. Symp. 2003; 70: 179-199Crossref PubMed Scopus (350) Google Scholar). The family C1 enzymes are involved in the normal lysosomal turnover of proteins, but they are also implicated in many disease processes, such as tumor invasion and connective tissue destruction on inflammation (2Sloane B.F. Moin K. Krepela E. Rozhin J. Cancer Metastasis Rev. 1990; 9: 333-352Crossref PubMed Scopus (188) Google Scholar, 3Mort J.S. Recklies A.D. Poole A.R. Arthritis Rheum. 1984; 27: 509-515Crossref PubMed Scopus (123) Google Scholar, 4Buttle D.J. Burnett D. Abrahamson M. Scand. J. Clin. Lab. Investig. 1990; 50: 509-516Crossref PubMed Google Scholar). The cystatins constitute a superfamily of related proteins. The mammalian superfamily members are of three major types (1Abrahamson M. Alvarez-Fernandez M. Nathanson C.M. Biochem. Soc. Symp. 2003; 70: 179-199Crossref PubMed Scopus (350) Google Scholar, 5Barrett A.J. Rawlings N.D. Davies M.E. Machleidt W. Salvesen G. Turk V. Barrett A.J. Salvesen G. Proteinase Inhibitors. 12. Elsevier Science Publishers BV, New York1986: 515-569Google Scholar, 6Rawlings N.D. Barrett A.J. J. Mol. Evol. 1990; 30: 60-71Crossref PubMed Scopus (270) Google Scholar). Type 1 cystatins (also called stefins) are primarily cytoplasmatic, single-domain proteins composed of ∼100 amino acid residues, with no disulfide bridges and no signal peptide. Type 2 cystatins are secreted inhibitors and are also single-domain proteins (but are about 120 residues long) and present two well-conserved disulfide bridges and typical signal peptides. Type 3 cystatins, or kininogens, are multidomain proteins presenting three tandemly repeated type 2 cystatin-like domains. Chicken egg-white (CEW) 1The abbreviations used are: CEW, chicken egg-white; BSL, backside loop; r.m.s.d., root mean square deviation; PDB, Protein Data Bank. cystatin, an avian type 2 cystatin, was the first cysteine peptidase inhibitor for which the three-dimensional structure was determined by x-ray crystallography (7Bode W. Engh R. Musil D. Thiele U. Huber R. Karshikov A. Brzin J. Kos J. Turk V. EMBO J. 1988; 7: 2593-2599Crossref PubMed Scopus (547) Google Scholar). The structures of two type 1 cystatins have also been determined: human cystatin A (or stefin A), determined by NMR spectroscopy (8Martin J.R. Craven C.J. Jerala R. Kroon-Zitko L. Zerovnik E. Turk V. Waltho J.P. J. Mol. Biol. 1995; 246: 331-343Crossref PubMed Scopus (102) Google Scholar) and recently by x-ray crystallography of a complex with cathepsin H (9Jenko S. Dolenc I. Guncar G. Dobersek A. Podobnik M. Turk D. J. Mol. Biol. 2003; 326: 875-885Crossref PubMed Scopus (97) Google Scholar); and human cystatin B (or stefin B) in complex with papain, determined by x-ray crystallography (10Stubbs M.T. Laber B. Bode W. Huber R. Jerala R. Lenarcic B. Turk V. EMBO J. 1990; 9: 1939-1947Crossref PubMed Scopus (470) Google Scholar). All three cystatins show the same overall structure, with a five-stranded anti-parallel β-sheet wrapped around a five-turn α-helix. In these cystatin structures the papain-binding site is a tripartite, wedge-shaped edge, formed by the N-terminal segment and the first and second hairpin loops, which are called L1 and L2. There is also a NMR model for the plant inhibitor oryzacystatin, which shows the same "cystatin fold" as the animal cystatins (11Nagata K. Kudo N. Abe K. Arai S. Tanokura M. Biochemistry. 2000; 39: 14753-14760Crossref PubMed Scopus (113) Google Scholar). In addition, the structure of a dimeric form of human cystatin C has been published (12Janowski R. Kozak M. Jankowska E. Grzonka Z. Grubb A. Abrahamson M. Jaskolski M. Nat. Struct. Biol. 2001; 8: 316-320Crossref PubMed Scopus (349) Google Scholar). Although this dimeric form of cystatin C is inactive as a papain inhibitor due to shedding of the binding site, each of the two domains formed by three-dimensional subdomain swapping adopt the monomeric cystatin fold. Despite these quite extensive structural data, detailed knowledge of what determines the specificity profiles of different cystatins is lacking. Cystatin D is a type 2 cystatin thus far found only in human saliva and tear fluid (13Freije J.P. Balbin M. Abrahamson M. Velasco G. Dalboge H. Grubb A. Lopez-Otin C. J. Biol. Chem. 1993; 268: 15737-15744Abstract Full Text PDF PubMed Google Scholar). It is produced as a preprotein of 142 amino acid residues, of which the first 20 residues constitute a typical signal peptide (14Freije J.P. Abrahamson M. Olafsson I. Velasco G. Grubb A. Lopez-Otin C. J. Biol. Chem. 1991; 266: 20538-20543Abstract Full Text PDF PubMed Google Scholar). Cystatin D was originally found as the product of a gene segment displaying a high degree of homology to the human cystatin C gene (15Abrahamson M. Olafsson I. Palsdottir A. Ulvsbäck M. Lundwall Å. Jensson O. Grubb A. Biochem. J. 1990; 268: 287-294Crossref PubMed Scopus (522) Google Scholar). Its complete amino acid sequence displays 55% identical residues compared with the cystatin C sequence, with all sequence motifs known to be essential for cysteine peptidase inhibition well conserved (14Freije J.P. Abrahamson M. Olafsson I. Velasco G. Grubb A. Lopez-Otin C. J. Biol. Chem. 1991; 266: 20538-20543Abstract Full Text PDF PubMed Google Scholar). However, the inhibition profile of cystatin D for human family C1 peptidases is clearly different from that of cystatin C (1Abrahamson M. Alvarez-Fernandez M. Nathanson C.M. Biochem. Soc. Symp. 2003; 70: 179-199Crossref PubMed Scopus (350) Google Scholar, 16Balbin M. Hall A. Grubb A. Mason R.W. Lopez-Otin C. Abrahamson M. J. Biol. Chem. 1994; 269: 23156-23162Abstract Full Text PDF PubMed Google Scholar). Unlike the latter, cystatin D is unable to inhibit cathepsin B. Besides, it shows a preferential inhibition of cathepsin S over cathepsins H and L (16Balbin M. Hall A. Grubb A. Mason R.W. Lopez-Otin C. Abrahamson M. J. Biol. Chem. 1994; 269: 23156-23162Abstract Full Text PDF PubMed Google Scholar). By a site-directed mutagenesis approach to alter residues in the N-terminal segments of cystatin D and C, it has been shown that these residues can interact with the non-primed substrate pockets of the enzymes in a substrate-like manner (17Hall A. Ekiel I. Mason R.W. Kasprzykowski F. Grubb A. Abrahamson M. Biochemistry. 1998; 37: 4071-4079Crossref PubMed Google Scholar). Moreover, evidence was presented that N-terminal sequence differences partly explain the specificity differences between cystatins D and C. However, by analysis of engineered hybrid cystatins, it was apparent that structural differences in the framework of the cystatin molecule also must have a large effect on the inhibitory specificity of cystatin D (17Hall A. Ekiel I. Mason R.W. Kasprzykowski F. Grubb A. Abrahamson M. Biochemistry. 1998; 37: 4071-4079Crossref PubMed Google Scholar). It was recently reported that some type 2 cystatins can inhibit mammalian legumain, a cysteine peptidase of family C13, through a novel reactive site located on the side opposite to the papain-binding site (18Alvarez-Fernandez M. Barrett A.J. Gerhartz B. Dando P.M. Ni J. Abrahamson M. J. Biol. Chem. 1999; 274: 19195-19203Abstract Full Text Full Text PDF PubMed Scopus (244) Google Scholar), in a loop referred to as the back-side loop (BSL). This site results in tight reversible inhibition of pig legumain and is active on human cystatins C, E/M, and F, but not on cystatins A, B, and D. Thus, also with respect to inhibition of family C13 enzymes, cystatin D displays a more restricted and specific inhibition profile than other type 2 cystatins. In the present study, we have crystallized and determined the three-dimensional structure of recombinant human cystatin D, with the aim of clarifying the structural reasons for its selectivity at target enzyme inhibition. Expression and Purification of Recombinant Cystatin D—Human Arg26-cystatin D (one of the two allelic variants present in approximately equal proportions in the population) (19Balbin M. Freije J.P. Abrahamson M. Velasco G. Grubb A. Lopez-Otin C. Hum. Genet. 1993; 90: 668-669Crossref PubMed Scopus (16) Google Scholar) was overexpressed in an Escherichia coli expression system as described previously (13Freije J.P. Balbin M. Abrahamson M. Velasco G. Dalboge H. Grubb A. Lopez-Otin C. J. Biol. Chem. 1993; 268: 15737-15744Abstract Full Text PDF PubMed Google Scholar). After expression, the protein was purified by anion exchange chromatography on a Q-Sepharose column (30 × 300 mm2; Amersham Biosciences), followed by size exclusion chromatography on a Superdex 75 10/30 column (Amersham Biosciences) connected to a fast protein liquid chromatography system. The anion exchange chromatography was performed using 20 mm ethanolamine, pH 9.0, containing 1 mm benzamidinium chloride as elution buffer, and the size exclusion chromatography was performed using 50 mm Tris buffer, pH 7.5, with 150 mm NaCl. The fractions of highest purity were pooled and dialyzed against 100 mm Tris buffer, pH 7.5. The protein solution was then concentrated using a Vivaspin column with a cut-off limit of 5000 Da (Vivascience, Lincoln, UK) to a final concentration of ∼8 mg/ml. Protein concentrations were determined by UV absorption spectroscopy at 280 nm using ϵ = 18,200 m-1 cm-1 as extinction coefficient (A280, 0.1% = 1.29) (17Hall A. Ekiel I. Mason R.W. Kasprzykowski F. Grubb A. Abrahamson M. Biochemistry. 1998; 37: 4071-4079Crossref PubMed Google Scholar). The purity of the protein in size exclusion chromatography fractions was determined by size- and charge-separating electrophoreses in 16.5% SDS-PAGE gels (20Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207227) Google Scholar) and 1% agarose gels (21Jeppson J.O. Laurell C.B. Franzen B. Clin. Chem. 1979; 25: 629-638Crossref PubMed Scopus (218) Google Scholar), respectively. Crystallization—Crystallization plates were prepared using the hanging-drop vapor diffusion method in 24-well VDX plates (Hampton Research, Laguna Nigel, CA). Initial screening of crystallization conditions, at 18 °C, was done using the Crystal Screen kits 1 and 2 (22Jancarik J. Kim S-H. J. Appl. Crystallogr. 1991; 24: 409-411Crossref Scopus (2079) Google Scholar, 23Cudney R. Patel S. Weisgraber K. Newhouse Y. McPherson A. Acta Crystallogr. 1994; D50: 414-423Google Scholar) (Hampton Research). Five-μl droplets were used in the initial screens (2.5 μl of protein solution and 2.5 μl of precipitant solution), and 6–10-μl droplets were used in optimization trials. Reservoirs contained 750 μl in initial screens and 1000 μl at optimization. X-ray Data Collection and Processing—Room temperature data were collected on a Mar image plate system (Marresearch GmbH, Hamburg, Germany) mounted on a Rigaku RU-200 rotating anode generator operating at 50 kV, 90 mA. Crystals were mounted in a quartz capillary for data collection. A full data set was collected from a single crystal. Cryo conditions for data collection were worked out using sucrose as cryo protectant. The crystal was equilibrated with the mother liquor in the presence of 15% sucrose for a few minutes. The crystal was then mounted in a nylon CryoLoop (Hampton Research) and flash-cooled directly in a cold nitrogen stream at about 100 K. Diffraction data were collected at crystallographic beamline BL711 at the MAX-II synchrotron laboratory (Lund, Sweden) using a Mar345 image plate detector (X-ray Research GmbH, Norderstedt, Germany). A typical exposure time was 60 s/frame with 1° oscillation. All data sets were processed using the DENZO and SCALEPACK packages (24Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref PubMed Scopus (38572) Google Scholar). Structure Determination—The structure of cystatin D was solved by molecular replacement methods. For the room temperature data, this was done by using CEW cystatin (PDB code 1CEW) (7Bode W. Engh R. Musil D. Thiele U. Huber R. Karshikov A. Brzin J. Kos J. Turk V. EMBO J. 1988; 7: 2593-2599Crossref PubMed Scopus (547) Google Scholar) as a search model. Different modifications, such as poly(A) and poly(S), were tried. The programs AMoRe (25Navaza J. Acta Crystallogr. 1994; 50: 157-163Crossref Scopus (5029) Google Scholar) and Crystallography & NMR System (CNS) (26Brünger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J-S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. 1998; D54: 905-921Crossref Scopus (16967) Google Scholar) were used for the replacement search of the data between 15.0 and 4.0 Å. The molecular replacement solution was refined using the program CNS on the complete room temperature data (30.0 to 2.5 Å). The refined room temperature structure was then used as model for the rigid-body refinement on the cryo data. Structural Alignment and Graphical Illustrations—Multiple sequence alignment of cystatins with known structures was initially done by the Genetics Computer Group (GCG) Wisconsin Package software. The alignment was modified using the multiple structure alignment obtained with the program Multiple Alignment of Protein Structures (MAPS). 2G. Lu, manuscript in preparation. Web server: bioinfo1.mbfys.lu.se/TOP/maps.html. The structures used in the alignment were obtained from the Protein Data Bank (27Berman H.M. Westbrook J. Feng Z. Gilliland G. Bhat T.N. Weissig H. Shindyalov I.N. Bourne P.E. Nucleic Acids Res. 2000; 28: 235-242Crossref PubMed Scopus (27551) Google Scholar): CEW cystatin, PDB code 1CEW (7Bode W. Engh R. Musil D. Thiele U. Huber R. Karshikov A. Brzin J. Kos J. Turk V. EMBO J. 1988; 7: 2593-2599Crossref PubMed Scopus (547) Google Scholar); dimeric human cystatin C, PDB code 1G96 (12Janowski R. Kozak M. Jankowska E. Grzonka Z. Grubb A. Abrahamson M. Jaskolski M. Nat. Struct. Biol. 2001; 8: 316-320Crossref PubMed Scopus (349) Google Scholar); cystatin A, PDB code 1DVD (8Martin J.R. Craven C.J. Jerala R. Kroon-Zitko L. Zerovnik E. Turk V. Waltho J.P. J. Mol. Biol. 1995; 246: 331-343Crossref PubMed Scopus (102) Google Scholar); cystatin B, PDB code 1STF (10Stubbs M.T. Laber B. Bode W. Huber R. Jerala R. Lenarcic B. Turk V. EMBO J. 1990; 9: 1939-1947Crossref PubMed Scopus (470) Google Scholar); and oryzacystatin, PDB code 1EQK (11Nagata K. Kudo N. Abe K. Arai S. Tanokura M. Biochemistry. 2000; 39: 14753-14760Crossref PubMed Scopus (113) Google Scholar). If not otherwise indicated, the amino acid numbering used is that of human cystatin C, 3Human cystatin C numbering (51Grubb A. Löfberg H. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 3024-3027Crossref PubMed Scopus (301) Google Scholar) is used for cystatin D and other cystatins in this study. as previously used for cystatin D and other human type 2 cystatins (13Freije J.P. Balbin M. Abrahamson M. Velasco G. Dalboge H. Grubb A. Lopez-Otin C. J. Biol. Chem. 1993; 268: 15737-15744Abstract Full Text PDF PubMed Google Scholar, 28Ni J. Abrahamson M. Zhang M. Alvarez-Fernandez M.A. Grubb A. Su J. Yu G.L. Li Y. Parmelee D. Xing L. Coleman T.A. Gentz S. Thotakura R. Nguyen N. Hesselberg M. Gentz R. J. Biol. Chem. 1997; 272: 10853-10858Abstract Full Text Full Text PDF PubMed Scopus (145) Google Scholar, 29Ni J. Fernandez M.A. Danielsson L. Chillakuru R.A. Zhang J. Grubb A. Su J. Gentz R. Abrahamson M. J. Biol. Chem. 1998; 273: 24797-24804Abstract Full Text Full Text PDF PubMed Scopus (138) Google Scholar). Graphical representations were prepared with the programs MOLMOL (30Koradi R. Billeter M. Wuthrich K. J. Mol. Graph. 1996; 14 (29-32): 51-55Crossref PubMed Scopus (6490) Google Scholar) and GRASP (31Nicholls A. Sharp K.A. Honig B. Proteins. 1991; 11: 281-296Crossref PubMed Scopus (5316) Google Scholar). Crystallization of Cystatin D and Crystal Data Collection— Recombinant human cystatin D crystals appeared during the first week in Crystal Screen kit 1 condition 39 (100 mm sodium-HEPES buffer, pH 7.5, with 2% (w/v) polyethylene glycol 400 and 2 m (NH4)2SO4) at 18 °C. Finer grids based on this condition were settled at the same temperature by using either Tris or HEPES as buffer at a pH interval between 6.5 and 8.0 and by varying the ammonium sulfate (0.4–2.4 m) and polyethylene glycol (1–2%, w/v) concentrations. Crystals were obtained under several conditions. They were stable and presented typical shapes as long rods or plates. Two of the well-diffracting crystals were used for structure determination. These crystals were grown at 18 °C in 100 mm Tris, pH 7.5, with 2.4 m (NH4)2SO4 and 2.5% polyethylene glycol 400. A room temperature data set was collected from a plate-shaped crystal with dimensions of about 0.5 × 0.3 × 0.1 mm3. The crystal diffracted beyond 2.5 Å, and a full data set could be collected from a single crystal (Table I). A second similar crystal soaked in 15% sucrose was used to collect a data set at about 100 K, giving diffraction beyond 1.8 Å (Table I).Table IData collection and refinement statisticsParameterCrystal ICrystal IIDetection typeRoom temperatureCryo (about 100 K)λ = 1.5418λ = 0.9979Space groupP21212P21212Unit cell parametersa = 34.90a = 34.05b = 84.37b = 81.72c = 47.65c = 46.74α = β = γ = 90°α = β = γ = 90°No. of waters3285No. of non-H protein atoms912907Diffraction limit (Å)30-2.530-1.8Mosaicity (from Denzo)0.4330.366Solvent content (%)55.451.9R-merge (%)7.8 (25.2)5.8 (34.2)I/σ(I)25.1 (8.0)17.5 (3.0)No. of unique reflections495512,036Completeness (%)94.5 (96.8)94.9 (98.1)Rfree0.2290.278Rconv0.1910.250Averaged B-factor (Å2)All atoms44.2733.47Main chain38.9227.77Side chains49.0637.33Solvent48.4442.34Ramachandran plot statistics (%)Most favored region8790Additional allowed region119Generously allowed region10Disallowed region11 Open table in a new tab Molecular replacement was used as method to solve the structure of cystatin D from the room temperature data set. This was accomplished using the crystal structure of CEW cystatin as search model. Using AMoRe (25Navaza J. Acta Crystallogr. 1994; 50: 157-163Crossref Scopus (5029) Google Scholar) and CNS with data collected between 15.0 and 4.0 Å, we obtained the same rotational solutions, which were well above background, regardless of which model was used. From the extinction list, two axes were clearly shown as screw axes. Thus, the space groups P212121 and P21212 were both tested for translational search. The space group P21212 gave the correct solution. The rigid body refinement using CNS (32Brünger A.T. XPLOR Version 3.1 A System for X-ray Crystallography and NMR. Yale University Press, New Haven, CT1992Google Scholar) with the room temperature data (30.0 to 2.5 Å) lowered the Rcryst/Rfree from 0.438/0.438 to 0.349/0.342, respectively. The simulated annealing method was then applied for further refinement. The maps were calculated and inspected, and the residues of the search model were changed to the correct ones. Composite-omit maps were then calculated to remove model bias. A total of 112 residues, from position Ala10 to Val120 (human cystatin C numbering, Figs. 1A and 2), are included in the final room temperature structure model. No electron density was detected for the residues in the N-terminal segment before Ala10 (Fig. 1A). This region must thus be disordered in structure. The flexible region around residues 80–84 was difficult to build in before the composite-omit maps were made. Thirty-two water molecules were added to the model where strong difference densities (>3σ) were shown and the hydrogen-bond geometry was good. The individual B-factor refinement was applied to the final room temperature model.Fig. 2Alignment of cystatins with determined structures. The sequence alignment shown is based on a structural alignment performed by MAPS of the human cystatin D structure and those known for type 1 and 2 cystatins (human cystatin A, human cystatin B, CEW cystatin, human cystatin C from the dimer structure, and the plant cystatin, oryzacystatin). α-Helices and β-sheets are indicated in yellow and blue, respectively. The conserved papain-binding site is marked by boxes in magenta. The red asterisk indicates the position of the Asn residue, which is necessary for legumain inhibition. Arrows indicate the two conserved disulfide bridges in type 2 cystatins. Human cystatin C amino acid numbering was used, with the letter "a" indicating residues inserted in the cystatin D structure compared with the other cystatins structures. The residues closest in space at structural alignment are aligned in this figure, but it should be noted that structural positions of cystatin D residues 37–37a-38 all deviate very much from those for residues 37–38 in CEW cystatin and human cystatin C. Which of the three cystatin D residues that should be seen as the inserted one could therefore not be predicted with certainty (see Fig. 4). Similarly, in the larger L1 loop of cystatin D, residues 57a-58 correspond to residue 58 of the other two type 2 cystatins, with the latter located in a position intermediate to those of cystatin D residues 57a and 58 (see Fig. 3).View Large Image Figure ViewerDownload Hi-res image Download (PPT) The structure solution was straightforward for the cryo data after the room temperature structure was refined. The cryo structure had slight but significant changes in the cell dimensions (Table I). The refined room temperature model was then used for the rigid-body refinement on the cryo data in the resolution range from 30.0 to 1.8 Å. It was trivial to perform the subsequent refinement steps by CNS and to add a total of 85 water molecules to the cryo model (Table I). Despite the significant unit cell changes, particularly on the length of the b axis, the room temperature and cryo structures turned out to be very similar, with root mean square deviation (r.m.s.d.) on the Cα trace of 0.36 Å and an overall r.m.s.d. with side chains of 0.91 Å. At the C1 peptidase binding region, the r.m.s.d. values when comparing the cryo and room temperature structures are 0.23, 0.13 and 0.63 Å for the main chain atoms of the N-terminal part (amino acid residues Gly11-Ala15), the L1 loop (Gln55-Gly59), and the L2 loop (Val104-Asp108), respectively. The relatively large r.m.s.d. value for L2 is attributable to the contribution from Pro105. At the putative legumain (C13 peptidase) binding site, the r.m.s.d. value for the main chain atoms of the BSL (Val37-Glu41) is 0.15 Å. The strong similarity between the room temperature and cryo structures can also be seen from a B-factor plot (supplemental Fig. 1). Overall Structure—Human cystatin D adopts the so-called cystatin fold (Fig. 1B), as do its five homologues with known structures (7Bode W. Engh R. Musil D. Thiele U. Huber R. Karshikov A. Brzin J. Kos J. Turk V. EMBO J. 1988; 7: 2593-2599Crossref PubMed Scopus (547) Google Scholar, 8Martin J.R. Craven C.J. Jerala R. Kroon-Zitko L. Zerovnik E. Turk V. Waltho J.P. J. Mol. Biol. 1995; 246: 331-343Crossref PubMed Scopus (102) Google Scholar, 10Stubbs M.T. Laber B. Bode W. Huber R. Jerala R. Lenarcic B. Turk V. EMBO J. 1990; 9: 1939-1947Crossref PubMed Scopus (470) Google Scholar, 11Nagata K. Kudo N. Abe K. Arai S. Tanokura M. Biochemistry. 2000; 39: 14753-14760Crossref PubMed Scopus (113) Google Scholar). The core structure is built from a five-stranded anti-parallel β-sheet consisting of β1 (Ile13-Thr16), β2 (Ser44-Ile57), β3 (Val60-Thr71), β4 (Glu95-Val104), and β5 (Lys109-Lys119) that is wrapped around a five-turn α-helix (Lys21-Lys36) (Figs. 1B and 2). Comparison with CEW cystatin (Fig. 1, C and D) revealed some notable differences in the overall structure of cystatin D: 1) the loop at the C-terminal end of the α-helix is larger than that in CEW cystatin, which deforms the last turn of the helix. 2) Cystatin D does not present a bulge in the middle of the second strand of the β-sheet around position 49. 3) There is no helix in the appendix loop of cystatin D. Instead, it presents a disordered conformation. This is also the case for the monomeric domains in the crystal structure of dimeric human cystatin C (12Janowski R. Kozak M. Jankowska E. Grzonka Z. Grubb A. Abrahamson M. Jaskolski M. Nat. Struct. Biol. 2001; 8: 316-320Crossref PubMed Scopus (349) Google Scholar). 4) Marked differences in the putative peptidase-interacting regions of cystatin D are observed (see below). Comparison of the electrostatic potential surfaces of the two proteins (Fig. 1E) revealed further differences. Cystatin D has a narrower and more elongated shape than CEW cystatin. This might be a result of the missing bulge in the β2-strand, straightening up the β-sheet in cystatin D. Also, the two cystatins differ quite significantly with respect to the charge distribution on their surfaces. In CEW cystatin, positive and negative charges are evenly distributed on the protein surface. In cystatin D, however, the surface presents some strongly (mainly negatively) charged areas, whereas other areas are pronounced hydrophobic. Human cystatin D is present in two natural forms due to a gene polymorphism (19Balbin M. Freije J.P. Abrahamson M. Velasco G. Grubb A. Lopez-Otin C. Hum. Genet. 1993; 90: 668-669Crossref PubMed Scopus (16) Google Scholar). It has been shown that this variation neither significantly affects the enzyme binding properties of the inhibitor nor has drastic effects on protein stability (16Balbin M. Hall A. Grubb A. Mason R.W. Lopez-Otin C. Abrahamson M. J. Biol. Chem. 1994; 269: 23156-23162Abstract Full Text PDF PubMed Google Scholar), but the structural consequences of the variation have not been elucidated. The form crystallized here, Arg26-cystatin D, in which the 26th residue of the 122-residue predicted mature cystatin D sequence (13Freije J.P. Balbin M. Abrahamson M. Velasco G. Dalboge H. Grubb A. Lopez-Otin C. J. Biol. Chem. 1993; 268: 15737-15744Abstract Full Text PDF PubMed Google Scholar) is Arg, has a population frequency of 0.45 (19Balbin M. Freije J.P. Abrahamson M. Velasco G. Grubb A. Lopez-Otin C. Hum. Genet. 1993; 90: 668-669Crossref PubMed Scopus (16) Google Scholar). Because the predicted amino acid sequence of mature cystatin D is 1 residue longer in the N terminus than the reference sequence of cystatin C,3 the polymorphic residue is number 25 in an alignment of cystatin sequences (Fig. 2). The other natural form of human cystatin D (population frequency of 0.55) has Cys in this position, unlike other type 2 cystatins, for which Arg25 is well conserved (Fig. 2). The present structure shows that the Arg25 residue in cystatin D is situated in the second turn of the α-helix. Although it seems to be exposed at the surface of the protein, its side chain is undoubtedly oriented toward the cavity formed by the bent L2 loop, as revealed by the density map. The side chain is likely trapped by a salt bridge with the side chain of either Glu103 or Asp108 in the proximity of the L2 loop. The electron density for the amine groups further out in the side chain of Arg25 is weak or almost absent in both the cryo and the room temperature structures. This indicates a large flexibility in the conformation of these amine groups, reflected by high B-factor values, suggesting that they can alternate between the two anchoring sites formed by Glu103 and Asp108. Similarly, the conserved Arg residue in CEW cystatin seems to form a hydrogen bond with th
Referência(s)