- Email: [email protected]
Crystal Structure of Echicetin from Echis carinatus (Indian Saw-scaled Viper) at 2.4 Å Resolution
doi:10.1016/j.jmb.2003.10.048
J. Mol. Biol. (2004) 335, 167–176
Crystal Structure of Echicetin from Echis carinatus ˚ Resolution (Indian Saw-scaled Viper) at 2.4 A Jayasankar Jasti, M. Paramasivam, A. Srinivasan and T. P. Singh* Department of Biophysics, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110 029, India
Echicetin is a heterodimeric protein from the venom of the Indian sawscaled viper, Echis carinatus. It binds to platelet glycoprotein Ib (GPIb) and thus inhibits platelet aggregation. It has two subunits, a and b, consisting of 131 and 123 amino acid residues, respectively. The two chains are linked with a disulphide bond. The level of amino acid sequence homology between two subunits is 50%. The protein was purified from the venom of E. carinatus and crystallized using ammonium sulphate as a ˚ resolution precipitant. The crystal structure has been determined at 2.4 A and refined to an R-factor of 0.187. Overall dimensions of the heterodimer ˚ £ 35 A ˚ £ 35 A ˚ . The backbone folds of the two subunits are are , 80 A similar. The central portions of the polypeptide chains of a and b-subunits move into each other to form a tight dimeric association. The remaining portions of the chains of both subunits fold in a manner similar to those observed in the carbohydrate-binding domains of C-type lectins. In echicetin, the Ca2þ-binding sites are not present, despite being topologically equivalent to other similar Ca2þ-binding proteins of the superfamily. The residues Ser41, Glu43 and Glu47 in the calcium-binding proteins of the related family are conserved but the residues Glu126/120 are replaced by lysine at the corresponding sites in the a and b-subunits. q 2003 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: echicetin; crystal structure; molecular association; domain swapping; platelet aggregation
Introduction Echicetin is a heterodimeric protein present in the venom of the Indian saw-scaled viper, Echis carinatus.1 It is composed of two subunits designated as a and b, with molecular mass of 16 kDa and 14 kDa, respectively. It binds to platelet glycoprotein Ib (GPIb) and hence inhibits platelet aggregation. GPIb is involved in platelet adhesion and plays a significant role in the primary hemostasis.2 The process of adhesion of platelets at the sites of vascular injury is induced by specific platelet Abbreviations used: GPIb, glycoprotein Ib; vWf, von Willebrand factor; MBP, mannose-binding protein; IX-BP, coagulation factor IX-binding protein; IX/X-BP, coagulation factor IX/X-binding protein; X-BP, coagulation factor X-binding protein; CRD, carbohydrate-recognition domain; IgMk, immunoglobulin Mk; GITC, guanidium isothiocyanate; MMLV, Moloney murine leukemia virus; RT-PCR, reverse transcriptase polymerase chain reaction. E-mail address of the corresponding author: [email protected]
agonists.1 In addition to it, the low concentrations of thrombin were found to initiate platelet aggregation. This has been shown to be inhibited by echicetin.3 Actually, echicetin interferes with the interactions of von Willebrand factor (vWf) with GPIb. The other venom proteins that induce aggregation are botrocetin from Bothrops jararaca,4 bitiscetin from Bitis arientans5 and flavocetin-A from the habu snake Trimeresurus flavoviridis.6 All these proteins are heterodimers with homologous a and b-subunits, and bind to vWf to form an active complex. However, they choose different sites in vWf for binding, indicating unique specificities. Some other similar proteins such as blood coagulation factor IX-binding proteins (IX-BP),7 X-binding proteins (X-BP)8 and IX/X-binding proteins (IX/X-BP)9 have been characterized. The a-subunit of the present echicetin shows a level of sequence homology of 60% with the echicetin from E. carinatus sochureki venom,10 while its b-subunit exhibits 90% sequence homology.10 Although the a and b-subunits of these proteins show a low level of sequence identity with the carbohydraterecognition domains (CRDs) of C-type lectins,
0022-2836/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
168
Crystal Structure of Echicetin from Saw-scaled Viper
Figure 1. (a) Nucleotide and deduced amino acid sequence for the a-subunit of echicetin. The arrow indicates the amino terminus of the matured protein. The stop codon is indicated by p p p. (b) Nucleotide and deduced amino acid sequence for the b-subunit of echicetin. The arrow indicates the amino terminus of the matured protein. The stop codon is indicated by p p p.
169
Crystal Structure of Echicetin from Saw-scaled Viper
they adopt similar conformations.11 Hence, these proteins are considered to be members of the C-type lectin superfamily although they have no lectin activity. The loss of lectin activity might have arisen due to a unique form of dimerization, which might have given rise to new functions. In order to understand the functional specificity and the molecular recognition of echicetin, and its precise role in the inhibition of platelet aggregation under one situation and the promotion of platelet agglutination in another, structural studies of echicetin and other related proteins of the superfamily have been initiated in our laboratory. The crystal structures of only a few members of the family of C-type lectin-like proteins have been determined.4 – 9 However, so far, no crystal structure of echicetin from any source is known. Here, we report the three-dimensional structure of echicetin from the venom of E. carinatus determined by X-ray diffraction. The structure is unique in showing a domain-swapped a,b-heterodimer lacking a calcium-binding site. The dimerization is mediated by an interchain disulphide bridge, newly identified between the a and the b-heterodimers. This is the first reported crystal structure of an echicetin protein from any source.
Results and Discussion Sequence analysis of echicetin The cDNA of the a (b)-subunit of echicetin reported here is of 615(610) bp in length (the values in parentheses are for the b-subunit). It comprises of 12(103) nucleotides from the 50 -untranslated region (UTR) and an open reading frame (ORF) of 393(438) nucleotides encoding for 131(146) amino acid residues. The signal peptide is of 0(23) amino acid residues long. There are 210(69) nucleotides in the 30 -UTR region, including a stop codon. The nucleotide and derived amino acid sequences for the a and b-subunits of echicetin are given in Figure 1(a) and (b), respectively. The two subunits show 50% sequence homology between them. The a-subunit of echicetin shows 55 – 60% sequence homology with the a-subunits of flavocetin-A and echicetin from E. c. sochureki, C-type lectin-like proteins of snake venom. In contrast, the b-subunit shows a high level of sequence homology, 90%, with the b-subunit of echicetin from the venoms of other sources and 65– 70% with other members of the superfamily of C-type lectin-like proteins. It is noteworthy that the b-subunit resembles the corresponding subunits of other proteins of the family more closely than the a-subunit does. Quality of the model Echicetin contains two chains, the a and b-subunits. The a-subunit consists of 131 amino acid residues and the b-subunit is comprised of 123
Table 1. Crystallographic data and refinement statistics PDB code Space group ˚) Unit-cell dimensions (A ˚) a (A ˚) b (A ˚) c (A b (deg.) ˚) Resolution range (A Number of unique reflections ˚ ) (%) Completeness (overall: 20.0–2.4 A ˚ ) (%) Completeness (last shell: 2.5–2.4 A Rsym (overall) (%) Rsym (last shell) (%) I/s(I) (overall) I/s(I) (last shell) R-factor/Rfree (%) Protein atoms Water molecules ˚ )a r.m.s. deviations in bond lengths (A r.m.s. deviations in bond angles (deg.)a Residues in most allowed regions (%) Overall average G-factorb Backbone quality indexc a b c
1oz7 P21 38.0 95.3 42.6 112.6 20.0– 2.4 9776 95 85 9.8 21.0 9.0 2.5 18.7/24.8 ˚ 2) 2083 (Bavg ¼ 37.4 A ˚ 2) 142 (Bavg ¼ 39.6 A 0.012 2.0 85 0.15 0.06
Target stereochemistry from Engh & Huber.35 G-factor as reported by PROCHECK.12 Backbone Z score reported by WHAT CHECK.36
amino acid residues. The a and b-subunits form a tight heterodimer through domain swapping that is further stabilized by a disulphide linkage. Structural evaluations of the final model of the protein using PROCHECK12 indicated that 85% of the residues were in the most allowed regions of the Ramachandran plot.13 The model has a good geometry and all the residues fit well in the electron density, except for Leu16, Leu40, Glu48, Asp52, Met62, Tyr97, Val102, Met124 and Arg131 in the a-subunit and Phe56 and Tyr71 in the b-subunit. These residues showed poor electron density for the side-chains, possibly because of conformational flexibility, and were mutated to Ala. The refined model included all the 254 residues from two chains and 154 water molecules, yielding an R-factor of 0.187 and Rfree of 0.248 for 9776 unique ˚. reflections in the resolution range of 20.0 – 2.4 A The root mean square deviations (r.m.s.) in bond ˚ and 2.08, lengths and angles were 0.012 A respectively (Table 1). Overall structure The overall folding of the protein chain is shown in Figure 2. Each subunit comprises a globular unit with an extending loop contributing to a number of interactions between the two subunits that resulted in the formation of a stable dimer. The heterodimer is an elongated molecule with overall dimensions ˚ £ 35 A ˚ £ 35 A ˚ . The precise molecular of , 80 A mass as determined by mass spectrometry (MALDI-TOF) was 30,647 Da. The globular domain of each subunit consists of two a helices and five major b strands. Cys78 from the a-subunit and Cys75 from the b-subunit form an interchain
170
Crystal Structure of Echicetin from Saw-scaled Viper
Table 2. Hydrogen bonds between a and b-subunits a-Subunit Glu27 O11 Glu27 O12 Ser41 Og Ser41 Og Gly68 O Ser70 O Ser70 N Glu71 O12 Arg72 N Arg72 N1 Arg72 NH1 Ser73 N Lys74 N Glu75 N Ser80 Og Ser83 N Asp84 Od2 Phe91 O Glu92 O11 Thr111 Og1 Trp112 N Trp112 O
˚) Distance (A
b-Subunit ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···
Ser80 Og Ser80 Og Asp81 Od1 Asp81 Od2 Ser80 N Glu78 N Glu78 O Asn76 Nd2 Asn76 O Trp77 O Asp74 O Asn76 Od1 Asn76 Od1 Asn76 Od1 Asp70 Od1 Gly67 O Ser41 Og Trp106 N Arg108 NH Ala89 O Ala89 O Arg94 NH1
··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···
2.5 3.3 2.5 3.4 2.8 2.5 2.6 2.8 2.6 2.9 2.8 3.4 2.7 3.5 3.3 3.2 3.4 2.8 3.5 3.1 2.7 3.2
echicetin it extends to the adjoining domain to form a loop-swapping dimer. Calcium-binding sites Figure 2. Overall structure of echicetin. The a-subunit is represented in green and the b-subunit in red. The two chains are interconnected by a disulphide-bond formed between Cys78(a) and Cys75(b). The Figure was produced using the programs RIBBONS31 and POVRAY (http://www.povray.org).
disulphide linkage (Figure 2). The superimposition of Ca atoms for the common residues (96 residues ˚. used) of two chains shows an r.m.s. shift of 1.2 A The overall folding of the two subunits is similar except for the flexible swapped loops, which are somewhat differently oriented in order to maximize the interactions between the two subunits (Table 2 and Figure 3). The conformation of echicetin shows significant structural similarities with those of homologous proteins such as coagulation factors IX/X-BP9 (r.m.s. shift of equivalent ˚ and 0.9 A ˚, Ca atoms for the a and b-subunits 0.8 A 7 respectively), IX-BP (r.m.s. shift of equivalent Ca ˚ and 1.0 A ˚ , respectatoms of a and b-subunits 1.0 A ively), X-BP8 (r.m.s. shift of equivalent Ca atoms of ˚ and 1.0 A ˚ , respectively), a and b-subunits 0.9 A 6 flavocetin-A (r.m.s. shift of equivalent Ca atoms of ˚ and 1.0 A ˚ , respectively), a and b-subunits 0.9 A botrocetin4 (r.m.s. shift of equivalent Ca atoms of a ˚ and 0.9 A ˚ , respectively) and and b-subunits 0.8 A 5 bitiscetin (r.m.s. shift of equivalent Ca atoms of a ˚ and 1.0 A ˚ , respectively). Both and b-subunits 0.9 A chains of echicetin are homologous to CRDs of mannose-binding protein (MBP)11 (r.m.s. shift of ˚ ) and other Ca atoms for common residues 1.3 A 14,15 In the C-type lectins, the long C-type lectins. loop folds back to the globular domain, while in
Echicetin does not contain bound calcium ions, although it was crystallized in the presence of 10 mM CaCl2. In the a-subunit of echicetin, Lys126 has replaced the glutamic acid residue that is conserved in the conventional Ca2þ-binding sites in other members of the superfamily, such as, IX-BP, IX/X-BP and X-BP. Similarly, in the b-subunit of echicetin, Glu120 and Glu43 have been replaced by Lys120 and Arg43, respectively. The Nz atom of
Figure 3. Superposition of the Ca trace of the a-subunit (blue) over that of the b-subunit (red). The Figure was drawn with MOLSCRIPT.32
171
Crystal Structure of Echicetin from Saw-scaled Viper
Lys126 in the a-subunit is situated close to the position occupied by Ca2þ in IX/X-BP,9 IX-BP7 and X-BP,8 and contributed to the neutralization of the negative charges of the calcium-binding site. It further stabilized the structure by forming hydro˚ ), O12 (3.2 A ˚ ) of gen bonds with atoms O11 (2.7 A g ˚ ˚ Glu47 and atoms O (3.0 A) and O (3.3 A) of Ser43. The corresponding Lys120 atom in the b-subunit is involved in hydrogen bonds with atoms O11 ˚ ), O12 (3.2 A ˚ ) of Glu47 and atoms O (3.1 A ˚) (3.0 A g ˚ and O (2.6 A) of Ser41. In addition to these hydrogen bonds in the b-subunit, Arg43 N1 forms a ˚ ). This feature hydrogen bond with Glu47 O12 (3.3 A of echicetin is unique and possibly contributes to the differences in the observed activity of the heterodimer when compared with the activities of other members of the family. Comparison with other members of the superfamily So far, the crystal structure of only two other C-type lectin-like proteins, flavocetin-A6 and bitiscetin,5 that lack a Ca2þ-binding site have been studied. The overall backbone conformations of these proteins are similar to that of echicetin. However, the relative orientation of the a and b-subunits of echicetin is markedly different from that found in flavocetin-A6 and in bitiscetin.5 If the b-subunit of echicetin is superimposed on the b-subunits of flavocetin-A and bitiscetin, the additional rotations required for the best fit of the a-subunit of echicetin with the a-subunits of flavocetin-A and bitiscetin were 31.68 and 30.48, respectively. The superpositions of Ca traces of the a and b-subunits of echicetin on the corresponding ˚ and 1.0 A ˚, units show r.m.s. shifts of 0.9 A ˚ ˚, respectively, in flavocetin-A, and 0.9 A and 1.0 A respectively, in bitiscetin, while the r.m.s. displacements for the central loops of the a and b-subunits ˚ and 2.6 A ˚ , respectively, in were found to be 4.2 A ˚ ˚ flavocetin-A, and 4.4 A and 2.0 A, respectively, in bitiscetin. The other similar proteins of the superfamily, such as coagulation factors IX/X-BP,9 IX-BP7 and X-BP,8 are involved in the binding of calcium ions. The calcium atoms in the two subunits form sevenfold coordination spheres. The orientations of coordinating residues in these structures differ significantly with the corresponding residues in the proteins of the superfamily that lack Ca2þ-binding capability. The presence of Ca2þ in the structures of these proteins seems to provide a tightening effect between the N-terminal region of a2 and the C-terminal tail of b5 (Figure 2). The formation of a well defined structure may be essential for ligand recognition in these proteins.7 The relative orientation of a and b-subunits in echicetin is considerably different from that observed in IX/X-BP, IX-BP and X-BP. If the b-subunit of echicetin is superimposed on the b-subunits of the IX/X-BP, IX-BP and X-BP, the additional rotations required for the best fit of the a-subunit
of echicetin with the a-subunits of these proteins are found to be of the order of 308. However, the differences in the relative orientations of the two subunits among IX/X-BP, IX-BP and X-BP are of the order of 68. This shows that the overall shaping of echicetin is distinct and represents a unique structural class among the proteins of the superfamily. Domain swapping Echicetin forms a dimer through domain swapping.16 This is a fascinating arrangement, in which the N and C-terminal parts of each subunit are interchanged by an identical domain from each other. This forms a common interface between the two monomer chains. The region between the loop and the body of the other subunit in echicetin is called a C-interface. The prominent residues of the C-interface of the b(a)-subunit in echicetin are Trp77, Trp79(Trp82), (Phe87), Leu85(Phe88), Tyr87(Tyr91) and Trp90 in the swapped loop and Trp23(Trp23), Leu68(Leu69), Phe97(Phe101) and Trp106(Trp112) in the body part of the b(a)-subunit. The two subunits are held together firmly by an intersubunit disulphide bridge involving Cys78(a)– Cys75(b). The two subunits are further stabilized by several hydrogen bonds (Table 2). The domain swapping is a strong feature of the proteins of this family and appears to be an essential aspect for the dimerization of this class of proteins. The function of the loop that works as a hinge between the domains is dependent on its length. A long enough loop will be able to turn back on the main body of monomer, whereas deletion of one or more residues will make it difficult to fold back and may result in the exposure of hydrophobic residues. This will make the monomer structure unfavourable. The heterodimeric proteins such as echicetin and other members of the family that form domain-swapped dimers are shorter by six or more residues than the monomeric proteins such as MBP,11 E-selectin,17 lithostathine18 and tetranectin.19 This suggests that the individual heterodimeric proteins might have evolved using the mechanism of loop swapping from their own C-type CRD monomers.7,9 There is another class of C-type lectins found in snake venoms that do not form monomers. They form disulphide-linked homodimers and were present in the venoms of rattlesnake (Crotalus atrox),20 B. arietans21 and Lachesis muta stenophyrs.22 These proteins contain a unique sequence of Gln-ProAsp without any hinge loop deletion. This facilitates the folding of the central loop back into the body of the C-type lectin in a similar way, as observed in MBP.11 In this case, the homodimers are arranged as closed monomers linked with a disulphide bond. It indicates that the heterodimers belonging to the proteins of the echicetin superfamily and the homodimers of venom C-type lectins might have originated from a common ancestor and acquired different structures
172
subsequently. It is noteworthy here that deletions at the hinge loop regions of echicetin-like proteins contain residues that are required for carbohydrate binding resulting in the loss of lectin activity in the heterodimers but acquired new functions through unique folding and specific sequence of residues. Platelet glycoprotein Ib-binding site Echicetin inhibits platelet aggregation by interfering with the association of GPIb and the human vWf.1 Structurally, a concave site is formed in echicetin as a result of unique dimerization through domain swapping. The groove-like binding sites in echicetin and other related proteins are further characterized by the presence of specific amino acid sequences. For example, botrocetin and bitiscetin contain a negatively charged stretch at the central concave surface for binding to the positively charged surface of the vWf A1 domain,4,5 while flavocetin-A has two hydrophilic patches that form a putative binding site on the concave surface for GPIb.6 Other coagulation factor-binding proteins (IX-BP, IX/X-BP and X-BP) have a positively charged patch on the corresponding region that is involved in the interaction with the g-carboxyl glutamic acid (Gla) residues of the coagulation factors.7 – 9 In the present case, it is presumed that the concave groove, which has an overall positive charge potential, is involved in the binding with GPIba. The residues, Arg20, Lys60, Asp62, Lys100, Thr102, Asp103, Gln105, Thr107, Arg109, Asp110, Trp113 and Thr114 from the b-subunit, form the lower side of the site; residues Trp72, Asp91,
Crystal Structure of Echicetin from Saw-scaled Viper
Arg94, His95 and Arg108 from the b-subunit form the middle portion of the site (Figure 4); and Tyr63, Lys64, Arg96, Ser98, Glu104, Lys105, Gln106, Thr111, Thr115 and Asn119 from the a-subunit form the upper side of the site (Figure 4). Thus, the predominantly positively charged surface at the putative GPIba-binding site on echicetin might be interacting with the complementary acidic surface on the GPIba (Figure 5). Although the target concave binding site in echicetin displays a basic surface similar to that observed in the blood coagulation factor-binding proteins such as IX-BP, IX/X-BP and X-BP, it does not bind to coagulation factors, primarily because of different positioning of the charged residues in a unique structural environment. It may be further noted that the residues Arg107 and Arg112 from the b chain and Lys100 from the a chain are conserved in IX-BP, IX/X-BP and X-BP, and seem to be critical in the binding of g-carboxyl glutamic acid residues of the Gla-domain of the respective coagulation factors. The corresponding residues in echicetin are substituted by Thr107(b), Thr112(b) and Ser98(a). Similarly, Glu98 from the A-chain, which is conserved in IX-BP, IX/X-BP and X-BP (which binds to Arg428 of the Gla-domain), was replaced by Arg96 in echicetin. Thus, the fine distributions of charges at the concave grooves seem to clearly differentiate in identifying the specific target molecule for binding. IgMk-binding site Recently, echicetin was shown to aggregate platelets upon binding to a plasma component,
Figure 4. Residues of echicetin that are assumed to be involved in binding with glycoprotein Iba (GPIba). The a-subunit is represented in green and the b-subunit in yellow. The Figure was produced using the programs MOLSCRIPT32 and Raster3D.33
Crystal Structure of Echicetin from Saw-scaled Viper
173
Figure 5. Surface potential representations of echicetin as viewed along the interchain disulphide bond (a, b and c). The blue colour corresponds to the positive and the red to negative potentials. The upper and lower halves of the molecule correspond to the a and the b-subunits, respectively. The Figure was produced using the program GRASP.34
IgMk, and was found to induce signal transduction.23,24 This seems to be contrary to the earlier reports about its role in the inhibition of platelet aggregation.1,3 The platelet agglutination was observed when echicetin came in contact with IgMk.24 The clustering of platelets decreases the free platelets in circulation and leads to thrombocytopenia.24 Thus, it was assumed that in vivo, echicetin seems to activate platelet aggregation rather than inhibiting platelet association. However, the structural details regarding the binding site on echicetin for IgMk are not known. The IgMk-binding site on echicetin can be understood by comparing the structure of echicetin with other known structures of C-type lectins for unique structural features that enable echicetin to bind to IgMk. Only echicetin displays an additional concave surface groove rich in positively charged residues and is contributed mostly by the a-subunit. The site is contributed mostly by the globular unit and the extended loop of a-subunit. The extended loop of the b-subunit, where it interacts with the globular unit of the a-subunit contributes the remaining part of the site. The extended loop of the a-subunit expands to form a groove. Glu71 and Glu99 are located at the base of the extended loop of the a-subunit, which could make the loop open. It may be emphasized that this type of extension has not been observed in other snake C-type lectin structures and could enable echicetin to bind to IgMk. As seen from Figure 6, the residues Tyr12(a), Trp23(a), Asp24(a), Glu27(a), Lys28(a), Lys35(a), Asp36(a), His38(a), Glu43(a), Ser70(a), Glu71(a), Arg72(a), Glu99(a), Glu118(a), Glu78(b), Asp81(b), Gln84(b) and Asp86(b) from the globular side of the a-subunit, and the residues
Ser73(a), Lys74(a), Glu75(a), Gln76(a), His77(a), Ser90(a), Glu92(a), Asp70(b) and Arg108(b) from the extended loop side of the a-subunit could
Figure 6. Residues of echicetin involved in the proposed IgMk-binding site. The a-subunit is represented in green and the b-subunit in yellow. The Figure was produced using the programs MOLSCRIPT32 and Raster3D.33
174
probably interact with IgMk by electrostatic interactions. As seen from Figure 5(b) and (c), the protein exhibits a positively charged patch near the concave surface that is thought to be for interaction with the GPIb. The acidic residues on the opposite side of the concave surface (Figure 5(a) and (b)) are involved in the binding to IgMk. Echicetin binds to one molecule of GPIb but one molecule of IgMk can bind up to five molecules of echicetin. The binding of echicetin to GPIb in the absence of IgMk prevents the binding of vWf and thus prevents the platelet aggregation. However, in the presence of IgMk, the clustering of GPIb-bound echicetin from different platelets causes agglutination. Thus, binding of echicetin to GPIba receptor on platelets alone results in inhibition of platelet agglutination, while binding to both GPIba receptor and IgMk promotes platelet aggregation and signal transduction.
Conclusions The most prominent feature of the structure of echicetin is the projection of a central portion of the polypeptide chain towards the adjoining subunit, forming a tight dimeric association. Excluding this feature, each subunit has a fold similar to the C-type CRD fold found in MBP. Echicetin blocks the interaction of vWf with GPIba and thus inhibits platelet aggregation of washed platelets in vitro. However, in addition to binding to the GPIba receptor, in vivo, it binds to IgMk in plasma and initiates platelet agglutination and signal transduction. The structure of echicetin suggests that it might bind to the acidic platelet glycoprotein Iba receptor through its basic concave groove surface formed by dimerization of the two subunits. Echicetin possesses a unique acidic patch located on the a-subunit, which is presumed to be involved in the binding to IgMk through electrostatic interactions. The binding of GPIb-bound echicetin from different platelets to IgMk could lead to platelet aggregation. This is the first reported structure of echicetin and the results provided by these investigations could be used as a model for the development of drugs against thrombosis.
Materials and Methods Sequence determination The tissues from the venom glands of Indian sawscaled viper were obtained from the Irula snake farm with permission from the Government of Tamil Nadu, India. The venom glands were homogenized in 4 M GITC (pH 4.0) with 0.1% b-mercaptoethanol in ice-cold conditions. The total RNA was extracted by the phenol/ chloroform method. The poly(A)þ mRNA was isolated from total RNA using an oligo(dT)-cellulose column
Crystal Structure of Echicetin from Saw-scaled Viper
(Amersham-Pharmacia). The small syringe column packed with oligo(dT)pcellulose was washed with 10 ml of high-salt buffer (1 M NaCl, 1 mM Na2EDTA, 40 mM Tris – HCl, pH 7.4). The total RNA was mixed with an equal volume of high-salt buffer and warmed to 65 8C and cooled immediately by placing it in ice. The chilled RNA was passed through the column. The column was washed with 3 ml of low-salt buffer (0.1 M NaCl, 1 mM Na2EDTA, 20 mM Tris – HCl, pH 7.4). The RNA was eluted using elution buffer (1 mM Na2EDTA, 10 mM Tris – HCl, pH 7.4) which was prewarmed to 65 8C. The poly(A)þ RNA was used for cDNA synthesis, which was carried out with Moloney murine leukemia virus (MMLV)-reverse transcriptase using oligo(dT) primers. A portion (2 ml) of the reverse transcriptase polymerase chain reaction (RT-PCR) reaction was used for the PCR amplification of the echicetin a and b-subunits. The conserved nucleotide sequences from snake venom C-type lectins were used for the design of primers. The sequences 50 CTACCTGTGGAGGCCAAAGGA30 and 50 AATTTATT GGACCTTCTGGCC30 were used as forward and reverse primers, respectively. The PCR was performed with Taq polymerase (Promega, USA) using MJ Research thermal cycler model PTC-100. The clones of a and b-subunits of echicetin were distinguished by Sty1 digestion. The nucleotide sequencing was carried out using the cloned double-stranded DNA (pGEM-T) on an automatic sequencer model ABI-377. Both strands were used for sequencing.
Purification Lyophilized venom of E. carinatus was obtained from Irula cooperative snake farm, Tamil Nadu, India and echicetin was purified by a two-step affinity and anionexchange chromatography. The crude venom (250 mg) was dissolved in 50 mM ammonium acetate buffer (pH 6.0) to a concentration of 25 mg/ml. In order to remove insoluble material, the venom solution was centrifuged at 10,000g for 20 minutes. The supernatant was loaded onto a Cibacron blue F3GA (Pharmacia) column (15 cm £ 2.5 cm), which was pre-equilibrated with 50 mM ammonium acetate (pH 6.0) at 293 K. The column was washed with the same buffer to remove unbound proteins. It was eluted step-wise with 50 mM ammonium bicarbonate (pH 8.5) and then with 50 mM ammonium carbonate (pH 10.5) to obtain echicetin. The fractions were pooled and dialyzed against 20 mM phosphate buffer (pH 6.5) and loaded onto a UNO-Q6 (Pharmacia) anion-exchange column (6.0 cm £ 2.0 cm), which was pre-equilibrated with 20 mM phosphate buffer (pH 6.5). The column was then eluted with a linear gradient of 0.0– 0.3 M NaCl in 20 mM phosphate buffer (pH 6.5). The fourth peak corresponding to echicetin was confirmed by SDS-PAGE, by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, and by N-terminal sequencing.
MALDI-TOF analysis Echicetin (9 ml of 5 mg/ml) in de-ionized water was mixed with 1 ml of 1% (v/v) trimethyl amine and the data were obtained on a Kratos PCKompact SEQ mass spectrometer using software version V1.2.2 (Kratos Analytical, Japan).
175
Crystal Structure of Echicetin from Saw-scaled Viper
Crystallization Lyophilized echicetin was dissolved in acetate buffer (0.1 M sodium acetate trihydrate, pH 4.6) to a final concentration of 15 mg/ml. The crystals of echicetin were obtained using the hanging-drop, vapor-diffusion method with ammonium sulphate as the precipitant: 5 ml of protein solution was mixed with an equal amount of the reservoir solution and equilibrated against 1.0 ml of reservoir solution (2.0 M ammonium sulphate in 0.1 M sodium acetate, pH 4.6 (Hampton Crystal Screen condition # 47), 10 mM CaCl2). Rod-shaped crystals of size 0.3 mm £ 0.2 mm £ 0.2 mm appeared within seven days. X-ray intensity data collection and processing The X-ray intensity data were collected at 283 K using an MAR300 imaging plate scanner mounted on a Rigaku RU-200 rotating-anode X-ray generator. The crystal to detector distance was kept at 200 mm. A total of 130 images were collected with 18 rotation for each image. ˚ resolution. The data The crystals diffracted to 2.4 A were processed and scaled with DENZO and SCALEPACK,25 respectively. The crystals belong to monoclinic space group P21 with one molecule in the asymmetric unit. The results of the data collection and processing statistics are given in Table 1. Structure determination and refinement The structure was determined by molecular replacement using Auto-AMoRe26 with the a chain from coagulation factor IX-binding protein from the venom of T. flavoviridis (PDB code 1bj3) as the search model to locate the two subunits. The residues from 71 to 100, which were suspected to adopt an extended loop conformation, were deleted from the search model as it exhibits conformational flexibility. The rotation and translation ˚ and 4.0 A ˚ functions calculated with data between 12.0 A yielded a clear solution with correlation coefficient of 36.8 (28.6) and an R-factor of 42.1% (45.0) (values in parentheses were of the first noise peak). The stacking arrangement of the molecules in the unit cell for this solution yielded no unfavourable intermolecular contact. The transformed coordinates from AMoRe were subjected to 20 cycles of rigid-body refinement and then to restrained refinement with REFMAC527 from the CCP4i V4.2 software suite.28 This reduced the R-factor to 36.4% with an Rfree of 44.7% (5% of the reflections were used to calculate Rfree and were not included in the refinement). The manual model building of the protein into ð2Fo 2 Fc Þ Fourier and ðFo 2 Fc Þ difference Fourier maps was carried out with the graphics program O29 on a Silicon Graphics O2 workstation. Water molecules were added using ARPP/REFMAC30 until there was no further decrease in free R-factor and were checked and adjusted manually for good hydrogen-bonding geometry with the model using O.29 The final refined coordinates consist of 131 residues in the a-subunit, 123 residues in the b-subunit and 142 water molecules with good geometry. The final R-factor and Rfree were 18.7% and 24.8%, respectively, for all the data in the resolution range ˚ (Table 1). 20.0– 2.4 A Data Bank accession numbers The cDNA sequences for the a and b-subunits of
echicetin have been deposited in GenBank with the accession codes AY268947 and AY268948, respectively.
Acknowledgements J.J. thanks the Council of Scientific and Industrial Research, New Delhi, India, for the award of a fellowship.
References 1. Peng, M., Lu, W., Beviglia, L., Niewiarowski, S. & Kirby, E. P. (1993). Echicetin: a snake venom protein that inhibits binding of von Willebrand factor and alboaggregins to platelet glycoprotein Ib. Blood, 81, 2321 –2328. 2. Kroll, M. H., Hellums, J. D., McIntyre, L. V., Schafer, A. I. & Moake, J. L. (1996). Platelets and shear stress. Blood, 88, 1525– 1541. 3. Peng, M., Emig, F. A., Mao, A., Lu, W., Kirby, E. P., Niewiarowski, S. & Kowalska, M. A. (1995). Interaction of echicetin with a high affinity thrombin binding site on platelet glycoprotein GPIb. Thromb. Haemost. 74, 954– 957. 4. Sen, U., Vasudevan, S., Subbarao, G., McClintock, R. A., Celikel, R., Ruggeri, Z. M. & Varughese, K. I. (2001). Crystal structure of the von Willebrand factor modulator botrocetin. Biochemistry, 40, 345– 352. 5. Hirotsu, S., Mizuno, H., Fukuda, K., Qi, M. C., Matsui, T., Hamako, J. et al. (2001). Crystal structure of bitiscetin, a von Willebrand factor-dependent platelet aggregation inducer. Biochemistry, 40, 13592 – 13597. 6. Fukuda, K., Mizuno, H., Atoda, H. & Morita, T. (2000). Crystal structure of flavocetin-A, a platelet glycoprotein Ib-binding protein, reveals a novel cyclic tetramer of C-type lectin-like heterodimers. Biochemistry, 39, 1915–1923. 7. Mizuno, H., Fujimoto, Z., Koizumi, M., Kano, H., Atoda, H. & Morita, T. (1999). Crystal structure of coagulation factor IX-binding protein from habu ˚ : implication of central loop snake venom at 2.6 A swapping based on deletion in the linker region. J. Mol. Biol. 289, 103– 112. 8. Mizuno, H., Fujimoto, Z., Atoda, H. & Morita, T. (2001). Crystal structure of an anticoagulant protein in complex with the Gla domain of factor X. Proc. Natl Acad. Sci. USA, 98, 7230– 7234. 9. Mizuno, H., Fujimoto, Z., Koizumi, M., Kano, H., Atoda, H. & Morita, T. (1997). Structure of coagulation factors IX/X-binding protein, a heterodimer of C-type lectin domains. Nature Struct. Biol. 4, 438 –441. 10. Polgar, J., Magnenat, E. M., Peitsch, M. C., Wells, T. N., Saqi, M. S. & Clemetson, K. J. (1997). Amino acid sequence of the alpha subunit and computer modelling of the alpha and beta subunits of echicetin from the venom of Echis carinatus (saw-scaled viper). Biochem. J. 323, 533–537. 11. Weis, W. I., Kahn, R., Fourme, R., Drickamer, K. & Hendrickson, W. A. (1991). Structure of the calciumdependent lectin domain from a rat mannosebinding protein determined by MAD phasing. Science, 254, 1608– 1615. 12. Laskowski, R. A., MacArthur, M. W., Moss, D. S. &
176
13. 14.
15. 16. 17.
18.
19.
20.
21.
22.
23.
Crystal Structure of Echicetin from Saw-scaled Viper
Thornton, J. M. (1993). PROCHECK: a program to check stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283–291. Ramachandran, G. N. & Sasisekharan, V. (1968). Conformation of polypeptides and proteins. Advan. Protein Chem. 23, 283– 438. Morita, T., Atoda, H. & Sekiya, F. (1996). Structure and functions of coagulation factor IX/factor X-binding protein isolated from the venom of Trimeresurus flavoviridis. In Natural Toxins (Singh, B. R. & Tu, A. T., eds), vol. 2, pp. 187– 196, Plenum Press, New York. Drickamer, K. (1993). Evolution of Ca2þ-dependent animal lectins. Prog. Nucl. Acid Res. Mol. Biol. 45, 207– 232. Bennett, M. J., Schlunegger, M. P. & Eisenberg, D. (1995). 3D domain swapping: a mechanism for oligomer assembly. Protein Sci. 4, 2455– 2468. Graves, B. J., Crowther, R. L., Chandran, C., Rumberger, J. M., Li, S., Huang, K. S. et al. (1994). Insight into E-selectin/ligand interaction from the crystal structure and mutagenesis of the lec/EGF domains. Nature, 367, 532– 538. Bertrand, J. A., Pignol, D., Bernard, J.-P., Verdier, J.-M., Dagorn, J. C. & Fontecilla-Camps, J. C. (1996). Crystal structure of human lithostathine, the pancreatic inhibitor of stone formation. EMBO J. 15, 2678– 2684. Nielsen, B. B., Kastrup, J. S., Rasmussen, H., Holtet, T. L., Graversen, J. H., Etzerodt, M. et al. (1997). Crystal structure of tetranectin, a trimeric plasminogen-binding protein with an a-helical coiled coil. FEBS Letters, 412, 388– 396. Hirabayashi, J., Kusunoki, T. & Kasai, K. (1991). Complete primary structure of a galactose-specific lectin from the venom of the rattlesnake Crotalus atrox: homologies with Ca2þ-dependent-type lectins. J. Biol. Chem. 266, 2320–2326. Nikai, T., Suzuki, J., Komori, Y., Ohkura, M., Ohizumi, Y. & Sugihara, H. (1995). Primary structure of the lectin from the venom of Bitis arietans (puffadder). Biol. Pharm. Bull. 18, 1620– 1622. Aragon-Ortiz, F., Mentele, R. & Auerswald, E. A. (1996). Amino acid sequence of a lectin-like protein from Lachesis muta stenophyrs venom. Toxicon, 34, 763– 769. Navdaev, A. & Clemetson, K. J. (2002). Glycoprotein Ib cross-linking/ligation on echicetin-coated surfaces
24.
25. 26. 27.
28. 29.
30.
31. 32. 33. 34.
35. 36.
or echicetin-IgMkappa in stirred suspension activates platelets by cytoskeleton modulated calcium release. J. Biol. Chem. 277, 45928– 45934. Navdaev, A., Dormann, D., Clemetson, J. M. & Clemetson, K. J. (2001). Echicetin, a GPIb-binding snake C-type lectin from Echis carinatus, also contains a binding site for IgMkappa responsible for platelet agglutination in plasma and inducing signal transduction. Blood, 97, 2333– 2341. Otwinowski, Z. & Minor, W. (1997). Processing of Xray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307– 326. Navaza, J. (1994). AMoRe: an automated package for molecular replacement. Acta Crystallog. 50, 157– 163. Murshudov, G. N., Vagin, A. A., Lebedev, A., Wilson, K. S. & Dodson, E. J. (1999). Efficient anisotropic refinement of macromolecular structures using FFT. Acta Crystallog. sect. D, 55, 247– 255. Collaborative Computational Project, Number 4 (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760– 763. Jones, T. A., Zou, J., Cowan, S. & Kjeldgaard, W. J. M. (1991). Improved methods for building models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110 – 119. Perrakis, A., Morris, R. J. & Lamzin, V. S. (1999). Automated protein model building combined with iterative structure refinement. Nature Struct. Biol. 6, 458– 463. Carson, M. (1997). Ribbons. Methods Enzymol. 277, 493– 505. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallog. 24, 946– 950. Merritt, E. A. & Murphy, M. E. P. (1994). Raster3D version 2.0: a program for photorealistic molecular graphics. Acta Crystallog. sect. D, 50, 869– 873. Nicholls, A., Sharp, K. A. & Honig, B. (1991). Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins: Struct. Funct. Genet. 11, 281– 296. Engh, R. A. & Huber, R. (1991). Accurate bond and angle parameters for X-ray protein structure refinement. Acta Crystallog. sect. A, 47, 392– 400. Hooft, R. W. W., Vriend, G., Sander, C. & Abola, E. E. (1996). Errors in protein structures. Nature, 381, 272.
Edited by R. Huber (Received 9 September 2003; received in revised form 15 October 2003; accepted 21 October 2003)
J. Mol. Biol. (2004) 335, 167–176
Crystal Structure of Echicetin from Echis carinatus ˚ Resolution (Indian Saw-scaled Viper) at 2.4 A Jayasankar Jasti, M. Paramasivam, A. Srinivasan and T. P. Singh* Department of Biophysics, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110 029, India
Echicetin is a heterodimeric protein from the venom of the Indian sawscaled viper, Echis carinatus. It binds to platelet glycoprotein Ib (GPIb) and thus inhibits platelet aggregation. It has two subunits, a and b, consisting of 131 and 123 amino acid residues, respectively. The two chains are linked with a disulphide bond. The level of amino acid sequence homology between two subunits is 50%. The protein was purified from the venom of E. carinatus and crystallized using ammonium sulphate as a ˚ resolution precipitant. The crystal structure has been determined at 2.4 A and refined to an R-factor of 0.187. Overall dimensions of the heterodimer ˚ £ 35 A ˚ £ 35 A ˚ . The backbone folds of the two subunits are are , 80 A similar. The central portions of the polypeptide chains of a and b-subunits move into each other to form a tight dimeric association. The remaining portions of the chains of both subunits fold in a manner similar to those observed in the carbohydrate-binding domains of C-type lectins. In echicetin, the Ca2þ-binding sites are not present, despite being topologically equivalent to other similar Ca2þ-binding proteins of the superfamily. The residues Ser41, Glu43 and Glu47 in the calcium-binding proteins of the related family are conserved but the residues Glu126/120 are replaced by lysine at the corresponding sites in the a and b-subunits. q 2003 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: echicetin; crystal structure; molecular association; domain swapping; platelet aggregation
Introduction Echicetin is a heterodimeric protein present in the venom of the Indian saw-scaled viper, Echis carinatus.1 It is composed of two subunits designated as a and b, with molecular mass of 16 kDa and 14 kDa, respectively. It binds to platelet glycoprotein Ib (GPIb) and hence inhibits platelet aggregation. GPIb is involved in platelet adhesion and plays a significant role in the primary hemostasis.2 The process of adhesion of platelets at the sites of vascular injury is induced by specific platelet Abbreviations used: GPIb, glycoprotein Ib; vWf, von Willebrand factor; MBP, mannose-binding protein; IX-BP, coagulation factor IX-binding protein; IX/X-BP, coagulation factor IX/X-binding protein; X-BP, coagulation factor X-binding protein; CRD, carbohydrate-recognition domain; IgMk, immunoglobulin Mk; GITC, guanidium isothiocyanate; MMLV, Moloney murine leukemia virus; RT-PCR, reverse transcriptase polymerase chain reaction. E-mail address of the corresponding author: [email protected]
agonists.1 In addition to it, the low concentrations of thrombin were found to initiate platelet aggregation. This has been shown to be inhibited by echicetin.3 Actually, echicetin interferes with the interactions of von Willebrand factor (vWf) with GPIb. The other venom proteins that induce aggregation are botrocetin from Bothrops jararaca,4 bitiscetin from Bitis arientans5 and flavocetin-A from the habu snake Trimeresurus flavoviridis.6 All these proteins are heterodimers with homologous a and b-subunits, and bind to vWf to form an active complex. However, they choose different sites in vWf for binding, indicating unique specificities. Some other similar proteins such as blood coagulation factor IX-binding proteins (IX-BP),7 X-binding proteins (X-BP)8 and IX/X-binding proteins (IX/X-BP)9 have been characterized. The a-subunit of the present echicetin shows a level of sequence homology of 60% with the echicetin from E. carinatus sochureki venom,10 while its b-subunit exhibits 90% sequence homology.10 Although the a and b-subunits of these proteins show a low level of sequence identity with the carbohydraterecognition domains (CRDs) of C-type lectins,
0022-2836/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
168
Crystal Structure of Echicetin from Saw-scaled Viper
Figure 1. (a) Nucleotide and deduced amino acid sequence for the a-subunit of echicetin. The arrow indicates the amino terminus of the matured protein. The stop codon is indicated by p p p. (b) Nucleotide and deduced amino acid sequence for the b-subunit of echicetin. The arrow indicates the amino terminus of the matured protein. The stop codon is indicated by p p p.
169
Crystal Structure of Echicetin from Saw-scaled Viper
they adopt similar conformations.11 Hence, these proteins are considered to be members of the C-type lectin superfamily although they have no lectin activity. The loss of lectin activity might have arisen due to a unique form of dimerization, which might have given rise to new functions. In order to understand the functional specificity and the molecular recognition of echicetin, and its precise role in the inhibition of platelet aggregation under one situation and the promotion of platelet agglutination in another, structural studies of echicetin and other related proteins of the superfamily have been initiated in our laboratory. The crystal structures of only a few members of the family of C-type lectin-like proteins have been determined.4 – 9 However, so far, no crystal structure of echicetin from any source is known. Here, we report the three-dimensional structure of echicetin from the venom of E. carinatus determined by X-ray diffraction. The structure is unique in showing a domain-swapped a,b-heterodimer lacking a calcium-binding site. The dimerization is mediated by an interchain disulphide bridge, newly identified between the a and the b-heterodimers. This is the first reported crystal structure of an echicetin protein from any source.
Results and Discussion Sequence analysis of echicetin The cDNA of the a (b)-subunit of echicetin reported here is of 615(610) bp in length (the values in parentheses are for the b-subunit). It comprises of 12(103) nucleotides from the 50 -untranslated region (UTR) and an open reading frame (ORF) of 393(438) nucleotides encoding for 131(146) amino acid residues. The signal peptide is of 0(23) amino acid residues long. There are 210(69) nucleotides in the 30 -UTR region, including a stop codon. The nucleotide and derived amino acid sequences for the a and b-subunits of echicetin are given in Figure 1(a) and (b), respectively. The two subunits show 50% sequence homology between them. The a-subunit of echicetin shows 55 – 60% sequence homology with the a-subunits of flavocetin-A and echicetin from E. c. sochureki, C-type lectin-like proteins of snake venom. In contrast, the b-subunit shows a high level of sequence homology, 90%, with the b-subunit of echicetin from the venoms of other sources and 65– 70% with other members of the superfamily of C-type lectin-like proteins. It is noteworthy that the b-subunit resembles the corresponding subunits of other proteins of the family more closely than the a-subunit does. Quality of the model Echicetin contains two chains, the a and b-subunits. The a-subunit consists of 131 amino acid residues and the b-subunit is comprised of 123
Table 1. Crystallographic data and refinement statistics PDB code Space group ˚) Unit-cell dimensions (A ˚) a (A ˚) b (A ˚) c (A b (deg.) ˚) Resolution range (A Number of unique reflections ˚ ) (%) Completeness (overall: 20.0–2.4 A ˚ ) (%) Completeness (last shell: 2.5–2.4 A Rsym (overall) (%) Rsym (last shell) (%) I/s(I) (overall) I/s(I) (last shell) R-factor/Rfree (%) Protein atoms Water molecules ˚ )a r.m.s. deviations in bond lengths (A r.m.s. deviations in bond angles (deg.)a Residues in most allowed regions (%) Overall average G-factorb Backbone quality indexc a b c
1oz7 P21 38.0 95.3 42.6 112.6 20.0– 2.4 9776 95 85 9.8 21.0 9.0 2.5 18.7/24.8 ˚ 2) 2083 (Bavg ¼ 37.4 A ˚ 2) 142 (Bavg ¼ 39.6 A 0.012 2.0 85 0.15 0.06
Target stereochemistry from Engh & Huber.35 G-factor as reported by PROCHECK.12 Backbone Z score reported by WHAT CHECK.36
amino acid residues. The a and b-subunits form a tight heterodimer through domain swapping that is further stabilized by a disulphide linkage. Structural evaluations of the final model of the protein using PROCHECK12 indicated that 85% of the residues were in the most allowed regions of the Ramachandran plot.13 The model has a good geometry and all the residues fit well in the electron density, except for Leu16, Leu40, Glu48, Asp52, Met62, Tyr97, Val102, Met124 and Arg131 in the a-subunit and Phe56 and Tyr71 in the b-subunit. These residues showed poor electron density for the side-chains, possibly because of conformational flexibility, and were mutated to Ala. The refined model included all the 254 residues from two chains and 154 water molecules, yielding an R-factor of 0.187 and Rfree of 0.248 for 9776 unique ˚. reflections in the resolution range of 20.0 – 2.4 A The root mean square deviations (r.m.s.) in bond ˚ and 2.08, lengths and angles were 0.012 A respectively (Table 1). Overall structure The overall folding of the protein chain is shown in Figure 2. Each subunit comprises a globular unit with an extending loop contributing to a number of interactions between the two subunits that resulted in the formation of a stable dimer. The heterodimer is an elongated molecule with overall dimensions ˚ £ 35 A ˚ £ 35 A ˚ . The precise molecular of , 80 A mass as determined by mass spectrometry (MALDI-TOF) was 30,647 Da. The globular domain of each subunit consists of two a helices and five major b strands. Cys78 from the a-subunit and Cys75 from the b-subunit form an interchain
170
Crystal Structure of Echicetin from Saw-scaled Viper
Table 2. Hydrogen bonds between a and b-subunits a-Subunit Glu27 O11 Glu27 O12 Ser41 Og Ser41 Og Gly68 O Ser70 O Ser70 N Glu71 O12 Arg72 N Arg72 N1 Arg72 NH1 Ser73 N Lys74 N Glu75 N Ser80 Og Ser83 N Asp84 Od2 Phe91 O Glu92 O11 Thr111 Og1 Trp112 N Trp112 O
˚) Distance (A
b-Subunit ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···
Ser80 Og Ser80 Og Asp81 Od1 Asp81 Od2 Ser80 N Glu78 N Glu78 O Asn76 Nd2 Asn76 O Trp77 O Asp74 O Asn76 Od1 Asn76 Od1 Asn76 Od1 Asp70 Od1 Gly67 O Ser41 Og Trp106 N Arg108 NH Ala89 O Ala89 O Arg94 NH1
··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ··· ···
2.5 3.3 2.5 3.4 2.8 2.5 2.6 2.8 2.6 2.9 2.8 3.4 2.7 3.5 3.3 3.2 3.4 2.8 3.5 3.1 2.7 3.2
echicetin it extends to the adjoining domain to form a loop-swapping dimer. Calcium-binding sites Figure 2. Overall structure of echicetin. The a-subunit is represented in green and the b-subunit in red. The two chains are interconnected by a disulphide-bond formed between Cys78(a) and Cys75(b). The Figure was produced using the programs RIBBONS31 and POVRAY (http://www.povray.org).
disulphide linkage (Figure 2). The superimposition of Ca atoms for the common residues (96 residues ˚. used) of two chains shows an r.m.s. shift of 1.2 A The overall folding of the two subunits is similar except for the flexible swapped loops, which are somewhat differently oriented in order to maximize the interactions between the two subunits (Table 2 and Figure 3). The conformation of echicetin shows significant structural similarities with those of homologous proteins such as coagulation factors IX/X-BP9 (r.m.s. shift of equivalent ˚ and 0.9 A ˚, Ca atoms for the a and b-subunits 0.8 A 7 respectively), IX-BP (r.m.s. shift of equivalent Ca ˚ and 1.0 A ˚ , respectatoms of a and b-subunits 1.0 A ively), X-BP8 (r.m.s. shift of equivalent Ca atoms of ˚ and 1.0 A ˚ , respectively), a and b-subunits 0.9 A 6 flavocetin-A (r.m.s. shift of equivalent Ca atoms of ˚ and 1.0 A ˚ , respectively), a and b-subunits 0.9 A botrocetin4 (r.m.s. shift of equivalent Ca atoms of a ˚ and 0.9 A ˚ , respectively) and and b-subunits 0.8 A 5 bitiscetin (r.m.s. shift of equivalent Ca atoms of a ˚ and 1.0 A ˚ , respectively). Both and b-subunits 0.9 A chains of echicetin are homologous to CRDs of mannose-binding protein (MBP)11 (r.m.s. shift of ˚ ) and other Ca atoms for common residues 1.3 A 14,15 In the C-type lectins, the long C-type lectins. loop folds back to the globular domain, while in
Echicetin does not contain bound calcium ions, although it was crystallized in the presence of 10 mM CaCl2. In the a-subunit of echicetin, Lys126 has replaced the glutamic acid residue that is conserved in the conventional Ca2þ-binding sites in other members of the superfamily, such as, IX-BP, IX/X-BP and X-BP. Similarly, in the b-subunit of echicetin, Glu120 and Glu43 have been replaced by Lys120 and Arg43, respectively. The Nz atom of
Figure 3. Superposition of the Ca trace of the a-subunit (blue) over that of the b-subunit (red). The Figure was drawn with MOLSCRIPT.32
171
Crystal Structure of Echicetin from Saw-scaled Viper
Lys126 in the a-subunit is situated close to the position occupied by Ca2þ in IX/X-BP,9 IX-BP7 and X-BP,8 and contributed to the neutralization of the negative charges of the calcium-binding site. It further stabilized the structure by forming hydro˚ ), O12 (3.2 A ˚ ) of gen bonds with atoms O11 (2.7 A g ˚ ˚ Glu47 and atoms O (3.0 A) and O (3.3 A) of Ser43. The corresponding Lys120 atom in the b-subunit is involved in hydrogen bonds with atoms O11 ˚ ), O12 (3.2 A ˚ ) of Glu47 and atoms O (3.1 A ˚) (3.0 A g ˚ and O (2.6 A) of Ser41. In addition to these hydrogen bonds in the b-subunit, Arg43 N1 forms a ˚ ). This feature hydrogen bond with Glu47 O12 (3.3 A of echicetin is unique and possibly contributes to the differences in the observed activity of the heterodimer when compared with the activities of other members of the family. Comparison with other members of the superfamily So far, the crystal structure of only two other C-type lectin-like proteins, flavocetin-A6 and bitiscetin,5 that lack a Ca2þ-binding site have been studied. The overall backbone conformations of these proteins are similar to that of echicetin. However, the relative orientation of the a and b-subunits of echicetin is markedly different from that found in flavocetin-A6 and in bitiscetin.5 If the b-subunit of echicetin is superimposed on the b-subunits of flavocetin-A and bitiscetin, the additional rotations required for the best fit of the a-subunit of echicetin with the a-subunits of flavocetin-A and bitiscetin were 31.68 and 30.48, respectively. The superpositions of Ca traces of the a and b-subunits of echicetin on the corresponding ˚ and 1.0 A ˚, units show r.m.s. shifts of 0.9 A ˚ ˚, respectively, in flavocetin-A, and 0.9 A and 1.0 A respectively, in bitiscetin, while the r.m.s. displacements for the central loops of the a and b-subunits ˚ and 2.6 A ˚ , respectively, in were found to be 4.2 A ˚ ˚ flavocetin-A, and 4.4 A and 2.0 A, respectively, in bitiscetin. The other similar proteins of the superfamily, such as coagulation factors IX/X-BP,9 IX-BP7 and X-BP,8 are involved in the binding of calcium ions. The calcium atoms in the two subunits form sevenfold coordination spheres. The orientations of coordinating residues in these structures differ significantly with the corresponding residues in the proteins of the superfamily that lack Ca2þ-binding capability. The presence of Ca2þ in the structures of these proteins seems to provide a tightening effect between the N-terminal region of a2 and the C-terminal tail of b5 (Figure 2). The formation of a well defined structure may be essential for ligand recognition in these proteins.7 The relative orientation of a and b-subunits in echicetin is considerably different from that observed in IX/X-BP, IX-BP and X-BP. If the b-subunit of echicetin is superimposed on the b-subunits of the IX/X-BP, IX-BP and X-BP, the additional rotations required for the best fit of the a-subunit
of echicetin with the a-subunits of these proteins are found to be of the order of 308. However, the differences in the relative orientations of the two subunits among IX/X-BP, IX-BP and X-BP are of the order of 68. This shows that the overall shaping of echicetin is distinct and represents a unique structural class among the proteins of the superfamily. Domain swapping Echicetin forms a dimer through domain swapping.16 This is a fascinating arrangement, in which the N and C-terminal parts of each subunit are interchanged by an identical domain from each other. This forms a common interface between the two monomer chains. The region between the loop and the body of the other subunit in echicetin is called a C-interface. The prominent residues of the C-interface of the b(a)-subunit in echicetin are Trp77, Trp79(Trp82), (Phe87), Leu85(Phe88), Tyr87(Tyr91) and Trp90 in the swapped loop and Trp23(Trp23), Leu68(Leu69), Phe97(Phe101) and Trp106(Trp112) in the body part of the b(a)-subunit. The two subunits are held together firmly by an intersubunit disulphide bridge involving Cys78(a)– Cys75(b). The two subunits are further stabilized by several hydrogen bonds (Table 2). The domain swapping is a strong feature of the proteins of this family and appears to be an essential aspect for the dimerization of this class of proteins. The function of the loop that works as a hinge between the domains is dependent on its length. A long enough loop will be able to turn back on the main body of monomer, whereas deletion of one or more residues will make it difficult to fold back and may result in the exposure of hydrophobic residues. This will make the monomer structure unfavourable. The heterodimeric proteins such as echicetin and other members of the family that form domain-swapped dimers are shorter by six or more residues than the monomeric proteins such as MBP,11 E-selectin,17 lithostathine18 and tetranectin.19 This suggests that the individual heterodimeric proteins might have evolved using the mechanism of loop swapping from their own C-type CRD monomers.7,9 There is another class of C-type lectins found in snake venoms that do not form monomers. They form disulphide-linked homodimers and were present in the venoms of rattlesnake (Crotalus atrox),20 B. arietans21 and Lachesis muta stenophyrs.22 These proteins contain a unique sequence of Gln-ProAsp without any hinge loop deletion. This facilitates the folding of the central loop back into the body of the C-type lectin in a similar way, as observed in MBP.11 In this case, the homodimers are arranged as closed monomers linked with a disulphide bond. It indicates that the heterodimers belonging to the proteins of the echicetin superfamily and the homodimers of venom C-type lectins might have originated from a common ancestor and acquired different structures
172
subsequently. It is noteworthy here that deletions at the hinge loop regions of echicetin-like proteins contain residues that are required for carbohydrate binding resulting in the loss of lectin activity in the heterodimers but acquired new functions through unique folding and specific sequence of residues. Platelet glycoprotein Ib-binding site Echicetin inhibits platelet aggregation by interfering with the association of GPIb and the human vWf.1 Structurally, a concave site is formed in echicetin as a result of unique dimerization through domain swapping. The groove-like binding sites in echicetin and other related proteins are further characterized by the presence of specific amino acid sequences. For example, botrocetin and bitiscetin contain a negatively charged stretch at the central concave surface for binding to the positively charged surface of the vWf A1 domain,4,5 while flavocetin-A has two hydrophilic patches that form a putative binding site on the concave surface for GPIb.6 Other coagulation factor-binding proteins (IX-BP, IX/X-BP and X-BP) have a positively charged patch on the corresponding region that is involved in the interaction with the g-carboxyl glutamic acid (Gla) residues of the coagulation factors.7 – 9 In the present case, it is presumed that the concave groove, which has an overall positive charge potential, is involved in the binding with GPIba. The residues, Arg20, Lys60, Asp62, Lys100, Thr102, Asp103, Gln105, Thr107, Arg109, Asp110, Trp113 and Thr114 from the b-subunit, form the lower side of the site; residues Trp72, Asp91,
Crystal Structure of Echicetin from Saw-scaled Viper
Arg94, His95 and Arg108 from the b-subunit form the middle portion of the site (Figure 4); and Tyr63, Lys64, Arg96, Ser98, Glu104, Lys105, Gln106, Thr111, Thr115 and Asn119 from the a-subunit form the upper side of the site (Figure 4). Thus, the predominantly positively charged surface at the putative GPIba-binding site on echicetin might be interacting with the complementary acidic surface on the GPIba (Figure 5). Although the target concave binding site in echicetin displays a basic surface similar to that observed in the blood coagulation factor-binding proteins such as IX-BP, IX/X-BP and X-BP, it does not bind to coagulation factors, primarily because of different positioning of the charged residues in a unique structural environment. It may be further noted that the residues Arg107 and Arg112 from the b chain and Lys100 from the a chain are conserved in IX-BP, IX/X-BP and X-BP, and seem to be critical in the binding of g-carboxyl glutamic acid residues of the Gla-domain of the respective coagulation factors. The corresponding residues in echicetin are substituted by Thr107(b), Thr112(b) and Ser98(a). Similarly, Glu98 from the A-chain, which is conserved in IX-BP, IX/X-BP and X-BP (which binds to Arg428 of the Gla-domain), was replaced by Arg96 in echicetin. Thus, the fine distributions of charges at the concave grooves seem to clearly differentiate in identifying the specific target molecule for binding. IgMk-binding site Recently, echicetin was shown to aggregate platelets upon binding to a plasma component,
Figure 4. Residues of echicetin that are assumed to be involved in binding with glycoprotein Iba (GPIba). The a-subunit is represented in green and the b-subunit in yellow. The Figure was produced using the programs MOLSCRIPT32 and Raster3D.33
Crystal Structure of Echicetin from Saw-scaled Viper
173
Figure 5. Surface potential representations of echicetin as viewed along the interchain disulphide bond (a, b and c). The blue colour corresponds to the positive and the red to negative potentials. The upper and lower halves of the molecule correspond to the a and the b-subunits, respectively. The Figure was produced using the program GRASP.34
IgMk, and was found to induce signal transduction.23,24 This seems to be contrary to the earlier reports about its role in the inhibition of platelet aggregation.1,3 The platelet agglutination was observed when echicetin came in contact with IgMk.24 The clustering of platelets decreases the free platelets in circulation and leads to thrombocytopenia.24 Thus, it was assumed that in vivo, echicetin seems to activate platelet aggregation rather than inhibiting platelet association. However, the structural details regarding the binding site on echicetin for IgMk are not known. The IgMk-binding site on echicetin can be understood by comparing the structure of echicetin with other known structures of C-type lectins for unique structural features that enable echicetin to bind to IgMk. Only echicetin displays an additional concave surface groove rich in positively charged residues and is contributed mostly by the a-subunit. The site is contributed mostly by the globular unit and the extended loop of a-subunit. The extended loop of the b-subunit, where it interacts with the globular unit of the a-subunit contributes the remaining part of the site. The extended loop of the a-subunit expands to form a groove. Glu71 and Glu99 are located at the base of the extended loop of the a-subunit, which could make the loop open. It may be emphasized that this type of extension has not been observed in other snake C-type lectin structures and could enable echicetin to bind to IgMk. As seen from Figure 6, the residues Tyr12(a), Trp23(a), Asp24(a), Glu27(a), Lys28(a), Lys35(a), Asp36(a), His38(a), Glu43(a), Ser70(a), Glu71(a), Arg72(a), Glu99(a), Glu118(a), Glu78(b), Asp81(b), Gln84(b) and Asp86(b) from the globular side of the a-subunit, and the residues
Ser73(a), Lys74(a), Glu75(a), Gln76(a), His77(a), Ser90(a), Glu92(a), Asp70(b) and Arg108(b) from the extended loop side of the a-subunit could
Figure 6. Residues of echicetin involved in the proposed IgMk-binding site. The a-subunit is represented in green and the b-subunit in yellow. The Figure was produced using the programs MOLSCRIPT32 and Raster3D.33
174
probably interact with IgMk by electrostatic interactions. As seen from Figure 5(b) and (c), the protein exhibits a positively charged patch near the concave surface that is thought to be for interaction with the GPIb. The acidic residues on the opposite side of the concave surface (Figure 5(a) and (b)) are involved in the binding to IgMk. Echicetin binds to one molecule of GPIb but one molecule of IgMk can bind up to five molecules of echicetin. The binding of echicetin to GPIb in the absence of IgMk prevents the binding of vWf and thus prevents the platelet aggregation. However, in the presence of IgMk, the clustering of GPIb-bound echicetin from different platelets causes agglutination. Thus, binding of echicetin to GPIba receptor on platelets alone results in inhibition of platelet agglutination, while binding to both GPIba receptor and IgMk promotes platelet aggregation and signal transduction.
Conclusions The most prominent feature of the structure of echicetin is the projection of a central portion of the polypeptide chain towards the adjoining subunit, forming a tight dimeric association. Excluding this feature, each subunit has a fold similar to the C-type CRD fold found in MBP. Echicetin blocks the interaction of vWf with GPIba and thus inhibits platelet aggregation of washed platelets in vitro. However, in addition to binding to the GPIba receptor, in vivo, it binds to IgMk in plasma and initiates platelet agglutination and signal transduction. The structure of echicetin suggests that it might bind to the acidic platelet glycoprotein Iba receptor through its basic concave groove surface formed by dimerization of the two subunits. Echicetin possesses a unique acidic patch located on the a-subunit, which is presumed to be involved in the binding to IgMk through electrostatic interactions. The binding of GPIb-bound echicetin from different platelets to IgMk could lead to platelet aggregation. This is the first reported structure of echicetin and the results provided by these investigations could be used as a model for the development of drugs against thrombosis.
Materials and Methods Sequence determination The tissues from the venom glands of Indian sawscaled viper were obtained from the Irula snake farm with permission from the Government of Tamil Nadu, India. The venom glands were homogenized in 4 M GITC (pH 4.0) with 0.1% b-mercaptoethanol in ice-cold conditions. The total RNA was extracted by the phenol/ chloroform method. The poly(A)þ mRNA was isolated from total RNA using an oligo(dT)-cellulose column
Crystal Structure of Echicetin from Saw-scaled Viper
(Amersham-Pharmacia). The small syringe column packed with oligo(dT)pcellulose was washed with 10 ml of high-salt buffer (1 M NaCl, 1 mM Na2EDTA, 40 mM Tris – HCl, pH 7.4). The total RNA was mixed with an equal volume of high-salt buffer and warmed to 65 8C and cooled immediately by placing it in ice. The chilled RNA was passed through the column. The column was washed with 3 ml of low-salt buffer (0.1 M NaCl, 1 mM Na2EDTA, 20 mM Tris – HCl, pH 7.4). The RNA was eluted using elution buffer (1 mM Na2EDTA, 10 mM Tris – HCl, pH 7.4) which was prewarmed to 65 8C. The poly(A)þ RNA was used for cDNA synthesis, which was carried out with Moloney murine leukemia virus (MMLV)-reverse transcriptase using oligo(dT) primers. A portion (2 ml) of the reverse transcriptase polymerase chain reaction (RT-PCR) reaction was used for the PCR amplification of the echicetin a and b-subunits. The conserved nucleotide sequences from snake venom C-type lectins were used for the design of primers. The sequences 50 CTACCTGTGGAGGCCAAAGGA30 and 50 AATTTATT GGACCTTCTGGCC30 were used as forward and reverse primers, respectively. The PCR was performed with Taq polymerase (Promega, USA) using MJ Research thermal cycler model PTC-100. The clones of a and b-subunits of echicetin were distinguished by Sty1 digestion. The nucleotide sequencing was carried out using the cloned double-stranded DNA (pGEM-T) on an automatic sequencer model ABI-377. Both strands were used for sequencing.
Purification Lyophilized venom of E. carinatus was obtained from Irula cooperative snake farm, Tamil Nadu, India and echicetin was purified by a two-step affinity and anionexchange chromatography. The crude venom (250 mg) was dissolved in 50 mM ammonium acetate buffer (pH 6.0) to a concentration of 25 mg/ml. In order to remove insoluble material, the venom solution was centrifuged at 10,000g for 20 minutes. The supernatant was loaded onto a Cibacron blue F3GA (Pharmacia) column (15 cm £ 2.5 cm), which was pre-equilibrated with 50 mM ammonium acetate (pH 6.0) at 293 K. The column was washed with the same buffer to remove unbound proteins. It was eluted step-wise with 50 mM ammonium bicarbonate (pH 8.5) and then with 50 mM ammonium carbonate (pH 10.5) to obtain echicetin. The fractions were pooled and dialyzed against 20 mM phosphate buffer (pH 6.5) and loaded onto a UNO-Q6 (Pharmacia) anion-exchange column (6.0 cm £ 2.0 cm), which was pre-equilibrated with 20 mM phosphate buffer (pH 6.5). The column was then eluted with a linear gradient of 0.0– 0.3 M NaCl in 20 mM phosphate buffer (pH 6.5). The fourth peak corresponding to echicetin was confirmed by SDS-PAGE, by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry, and by N-terminal sequencing.
MALDI-TOF analysis Echicetin (9 ml of 5 mg/ml) in de-ionized water was mixed with 1 ml of 1% (v/v) trimethyl amine and the data were obtained on a Kratos PCKompact SEQ mass spectrometer using software version V1.2.2 (Kratos Analytical, Japan).
175
Crystal Structure of Echicetin from Saw-scaled Viper
Crystallization Lyophilized echicetin was dissolved in acetate buffer (0.1 M sodium acetate trihydrate, pH 4.6) to a final concentration of 15 mg/ml. The crystals of echicetin were obtained using the hanging-drop, vapor-diffusion method with ammonium sulphate as the precipitant: 5 ml of protein solution was mixed with an equal amount of the reservoir solution and equilibrated against 1.0 ml of reservoir solution (2.0 M ammonium sulphate in 0.1 M sodium acetate, pH 4.6 (Hampton Crystal Screen condition # 47), 10 mM CaCl2). Rod-shaped crystals of size 0.3 mm £ 0.2 mm £ 0.2 mm appeared within seven days. X-ray intensity data collection and processing The X-ray intensity data were collected at 283 K using an MAR300 imaging plate scanner mounted on a Rigaku RU-200 rotating-anode X-ray generator. The crystal to detector distance was kept at 200 mm. A total of 130 images were collected with 18 rotation for each image. ˚ resolution. The data The crystals diffracted to 2.4 A were processed and scaled with DENZO and SCALEPACK,25 respectively. The crystals belong to monoclinic space group P21 with one molecule in the asymmetric unit. The results of the data collection and processing statistics are given in Table 1. Structure determination and refinement The structure was determined by molecular replacement using Auto-AMoRe26 with the a chain from coagulation factor IX-binding protein from the venom of T. flavoviridis (PDB code 1bj3) as the search model to locate the two subunits. The residues from 71 to 100, which were suspected to adopt an extended loop conformation, were deleted from the search model as it exhibits conformational flexibility. The rotation and translation ˚ and 4.0 A ˚ functions calculated with data between 12.0 A yielded a clear solution with correlation coefficient of 36.8 (28.6) and an R-factor of 42.1% (45.0) (values in parentheses were of the first noise peak). The stacking arrangement of the molecules in the unit cell for this solution yielded no unfavourable intermolecular contact. The transformed coordinates from AMoRe were subjected to 20 cycles of rigid-body refinement and then to restrained refinement with REFMAC527 from the CCP4i V4.2 software suite.28 This reduced the R-factor to 36.4% with an Rfree of 44.7% (5% of the reflections were used to calculate Rfree and were not included in the refinement). The manual model building of the protein into ð2Fo 2 Fc Þ Fourier and ðFo 2 Fc Þ difference Fourier maps was carried out with the graphics program O29 on a Silicon Graphics O2 workstation. Water molecules were added using ARPP/REFMAC30 until there was no further decrease in free R-factor and were checked and adjusted manually for good hydrogen-bonding geometry with the model using O.29 The final refined coordinates consist of 131 residues in the a-subunit, 123 residues in the b-subunit and 142 water molecules with good geometry. The final R-factor and Rfree were 18.7% and 24.8%, respectively, for all the data in the resolution range ˚ (Table 1). 20.0– 2.4 A Data Bank accession numbers The cDNA sequences for the a and b-subunits of
echicetin have been deposited in GenBank with the accession codes AY268947 and AY268948, respectively.
Acknowledgements J.J. thanks the Council of Scientific and Industrial Research, New Delhi, India, for the award of a fellowship.
References 1. Peng, M., Lu, W., Beviglia, L., Niewiarowski, S. & Kirby, E. P. (1993). Echicetin: a snake venom protein that inhibits binding of von Willebrand factor and alboaggregins to platelet glycoprotein Ib. Blood, 81, 2321 –2328. 2. Kroll, M. H., Hellums, J. D., McIntyre, L. V., Schafer, A. I. & Moake, J. L. (1996). Platelets and shear stress. Blood, 88, 1525– 1541. 3. Peng, M., Emig, F. A., Mao, A., Lu, W., Kirby, E. P., Niewiarowski, S. & Kowalska, M. A. (1995). Interaction of echicetin with a high affinity thrombin binding site on platelet glycoprotein GPIb. Thromb. Haemost. 74, 954– 957. 4. Sen, U., Vasudevan, S., Subbarao, G., McClintock, R. A., Celikel, R., Ruggeri, Z. M. & Varughese, K. I. (2001). Crystal structure of the von Willebrand factor modulator botrocetin. Biochemistry, 40, 345– 352. 5. Hirotsu, S., Mizuno, H., Fukuda, K., Qi, M. C., Matsui, T., Hamako, J. et al. (2001). Crystal structure of bitiscetin, a von Willebrand factor-dependent platelet aggregation inducer. Biochemistry, 40, 13592 – 13597. 6. Fukuda, K., Mizuno, H., Atoda, H. & Morita, T. (2000). Crystal structure of flavocetin-A, a platelet glycoprotein Ib-binding protein, reveals a novel cyclic tetramer of C-type lectin-like heterodimers. Biochemistry, 39, 1915–1923. 7. Mizuno, H., Fujimoto, Z., Koizumi, M., Kano, H., Atoda, H. & Morita, T. (1999). Crystal structure of coagulation factor IX-binding protein from habu ˚ : implication of central loop snake venom at 2.6 A swapping based on deletion in the linker region. J. Mol. Biol. 289, 103– 112. 8. Mizuno, H., Fujimoto, Z., Atoda, H. & Morita, T. (2001). Crystal structure of an anticoagulant protein in complex with the Gla domain of factor X. Proc. Natl Acad. Sci. USA, 98, 7230– 7234. 9. Mizuno, H., Fujimoto, Z., Koizumi, M., Kano, H., Atoda, H. & Morita, T. (1997). Structure of coagulation factors IX/X-binding protein, a heterodimer of C-type lectin domains. Nature Struct. Biol. 4, 438 –441. 10. Polgar, J., Magnenat, E. M., Peitsch, M. C., Wells, T. N., Saqi, M. S. & Clemetson, K. J. (1997). Amino acid sequence of the alpha subunit and computer modelling of the alpha and beta subunits of echicetin from the venom of Echis carinatus (saw-scaled viper). Biochem. J. 323, 533–537. 11. Weis, W. I., Kahn, R., Fourme, R., Drickamer, K. & Hendrickson, W. A. (1991). Structure of the calciumdependent lectin domain from a rat mannosebinding protein determined by MAD phasing. Science, 254, 1608– 1615. 12. Laskowski, R. A., MacArthur, M. W., Moss, D. S. &
176
13. 14.
15. 16. 17.
18.
19.
20.
21.
22.
23.
Crystal Structure of Echicetin from Saw-scaled Viper
Thornton, J. M. (1993). PROCHECK: a program to check stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283–291. Ramachandran, G. N. & Sasisekharan, V. (1968). Conformation of polypeptides and proteins. Advan. Protein Chem. 23, 283– 438. Morita, T., Atoda, H. & Sekiya, F. (1996). Structure and functions of coagulation factor IX/factor X-binding protein isolated from the venom of Trimeresurus flavoviridis. In Natural Toxins (Singh, B. R. & Tu, A. T., eds), vol. 2, pp. 187– 196, Plenum Press, New York. Drickamer, K. (1993). Evolution of Ca2þ-dependent animal lectins. Prog. Nucl. Acid Res. Mol. Biol. 45, 207– 232. Bennett, M. J., Schlunegger, M. P. & Eisenberg, D. (1995). 3D domain swapping: a mechanism for oligomer assembly. Protein Sci. 4, 2455– 2468. Graves, B. J., Crowther, R. L., Chandran, C., Rumberger, J. M., Li, S., Huang, K. S. et al. (1994). Insight into E-selectin/ligand interaction from the crystal structure and mutagenesis of the lec/EGF domains. Nature, 367, 532– 538. Bertrand, J. A., Pignol, D., Bernard, J.-P., Verdier, J.-M., Dagorn, J. C. & Fontecilla-Camps, J. C. (1996). Crystal structure of human lithostathine, the pancreatic inhibitor of stone formation. EMBO J. 15, 2678– 2684. Nielsen, B. B., Kastrup, J. S., Rasmussen, H., Holtet, T. L., Graversen, J. H., Etzerodt, M. et al. (1997). Crystal structure of tetranectin, a trimeric plasminogen-binding protein with an a-helical coiled coil. FEBS Letters, 412, 388– 396. Hirabayashi, J., Kusunoki, T. & Kasai, K. (1991). Complete primary structure of a galactose-specific lectin from the venom of the rattlesnake Crotalus atrox: homologies with Ca2þ-dependent-type lectins. J. Biol. Chem. 266, 2320–2326. Nikai, T., Suzuki, J., Komori, Y., Ohkura, M., Ohizumi, Y. & Sugihara, H. (1995). Primary structure of the lectin from the venom of Bitis arietans (puffadder). Biol. Pharm. Bull. 18, 1620– 1622. Aragon-Ortiz, F., Mentele, R. & Auerswald, E. A. (1996). Amino acid sequence of a lectin-like protein from Lachesis muta stenophyrs venom. Toxicon, 34, 763– 769. Navdaev, A. & Clemetson, K. J. (2002). Glycoprotein Ib cross-linking/ligation on echicetin-coated surfaces
24.
25. 26. 27.
28. 29.
30.
31. 32. 33. 34.
35. 36.
or echicetin-IgMkappa in stirred suspension activates platelets by cytoskeleton modulated calcium release. J. Biol. Chem. 277, 45928– 45934. Navdaev, A., Dormann, D., Clemetson, J. M. & Clemetson, K. J. (2001). Echicetin, a GPIb-binding snake C-type lectin from Echis carinatus, also contains a binding site for IgMkappa responsible for platelet agglutination in plasma and inducing signal transduction. Blood, 97, 2333– 2341. Otwinowski, Z. & Minor, W. (1997). Processing of Xray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307– 326. Navaza, J. (1994). AMoRe: an automated package for molecular replacement. Acta Crystallog. 50, 157– 163. Murshudov, G. N., Vagin, A. A., Lebedev, A., Wilson, K. S. & Dodson, E. J. (1999). Efficient anisotropic refinement of macromolecular structures using FFT. Acta Crystallog. sect. D, 55, 247– 255. Collaborative Computational Project, Number 4 (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760– 763. Jones, T. A., Zou, J., Cowan, S. & Kjeldgaard, W. J. M. (1991). Improved methods for building models in electron density maps and the location of errors in these models. Acta Crystallog. sect. A, 47, 110 – 119. Perrakis, A., Morris, R. J. & Lamzin, V. S. (1999). Automated protein model building combined with iterative structure refinement. Nature Struct. Biol. 6, 458– 463. Carson, M. (1997). Ribbons. Methods Enzymol. 277, 493– 505. Kraulis, P. J. (1991). MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallog. 24, 946– 950. Merritt, E. A. & Murphy, M. E. P. (1994). Raster3D version 2.0: a program for photorealistic molecular graphics. Acta Crystallog. sect. D, 50, 869– 873. Nicholls, A., Sharp, K. A. & Honig, B. (1991). Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins: Struct. Funct. Genet. 11, 281– 296. Engh, R. A. & Huber, R. (1991). Accurate bond and angle parameters for X-ray protein structure refinement. Acta Crystallog. sect. A, 47, 392– 400. Hooft, R. W. W., Vriend, G., Sander, C. & Abola, E. E. (1996). Errors in protein structures. Nature, 381, 272.
Edited by R. Huber (Received 9 September 2003; received in revised form 15 October 2003; accepted 21 October 2003)