Portal de Periódicos da CAPES

Structural Mechanism Governing the Quaternary Organization of Monocot Mannose-binding Lectin Revealed by the Novel Monomeric Structure of an Orchid Lectin

2005; Elsevier BV; Volume: 280; Issue: 15 Linguagem: Inglês

10.1074/jbc.m411634200

1083-351X

Wei Liu, Na Yang, Jingjin Ding, Ren‐Huai Huang, Zhong Hu, Dacheng Wang,

Complement system in diseases

Two isoforms of an antifungal protein, gastrodianin, were isolated from two subspecies of the orchid Gastrodia elata, belonging to the protein superfamily of monocot mannose-specific lectins. In the context that all available structures in this superfamily are oligomers so far, the crystal structures of the orchid lectins, both at 2.0 Å, revealed a novel monomeric structure. It resulted from the rearrangement of the C-terminal peptide inclusive of the 12th β-strand, which changes from the "C-terminal exchange" into a "C-terminal self-assembly" mode. Thus, the overall tertiary scaffold is stabilized with an intramolecular β-sheet instead of the hybrid observed on subunit/subunit interface in all known homologous dimeric or tetrameric lectins. In contrast to the constrained extended conformation with a cis peptide bond between residues 98 and 99 commonly occurring in oligomers, a β-hairpin forms from position 97 to 101 with a normal trans peptide bond at the corresponding site in gastrodianin, which determines the topology of the C-terminal peptide and thereby its unique fold pattern. Sequence and structure comparison shows that residue replacement and insertion at the position where the β-hairpin occurs in association with cis-trans inter-conversion of the specific peptide bond (97–98) are possibly responsible for such a radical structure switch between monomers and oligomers. Moreover, this seems to be a common melody controlling the quaternary states among bulb lectins through studies on sequence alignment. The observations revealed a structural mechanism by which the quaternary organization of monocot mannose binding lectins could be governed. The mutation experiment performed on maltose-binding protein-gastrodianin fusion protein followed by a few biochemical detections provides direct evidence to support this conclusion. Potential carbohydrate recognition sites and biological implications of the orchid lectin based on its monomeric state are also discussed in this paper. Two isoforms of an antifungal protein, gastrodianin, were isolated from two subspecies of the orchid Gastrodia elata, belonging to the protein superfamily of monocot mannose-specific lectins. In the context that all available structures in this superfamily are oligomers so far, the crystal structures of the orchid lectins, both at 2.0 Å, revealed a novel monomeric structure. It resulted from the rearrangement of the C-terminal peptide inclusive of the 12th β-strand, which changes from the "C-terminal exchange" into a "C-terminal self-assembly" mode. Thus, the overall tertiary scaffold is stabilized with an intramolecular β-sheet instead of the hybrid observed on subunit/subunit interface in all known homologous dimeric or tetrameric lectins. In contrast to the constrained extended conformation with a cis peptide bond between residues 98 and 99 commonly occurring in oligomers, a β-hairpin forms from position 97 to 101 with a normal trans peptide bond at the corresponding site in gastrodianin, which determines the topology of the C-terminal peptide and thereby its unique fold pattern. Sequence and structure comparison shows that residue replacement and insertion at the position where the β-hairpin occurs in association with cis-trans inter-conversion of the specific peptide bond (97–98) are possibly responsible for such a radical structure switch between monomers and oligomers. Moreover, this seems to be a common melody controlling the quaternary states among bulb lectins through studies on sequence alignment. The observations revealed a structural mechanism by which the quaternary organization of monocot mannose binding lectins could be governed. The mutation experiment performed on maltose-binding protein-gastrodianin fusion protein followed by a few biochemical detections provides direct evidence to support this conclusion. Potential carbohydrate recognition sites and biological implications of the orchid lectin based on its monomeric state are also discussed in this paper. Plant lectins are defined as proteins possessing at least one non-catalytic domain that binds reversibly to a specific mono- or oligosaccharide (1Peumans W.J. Van Damme E.J. Plant Physiol. 1995; 109: 347-352Crossref PubMed Scopus (847) Google Scholar). Because of the increasing availability of sequence and structure information, the majority of all known plant lectins have been recently subdivided into seven structural and evolutional related groups (2Van Damme E.J.M. Peumans W.J. Barre A. Rouge P. Crit. Rev. Plant Sci. 1998; 17: 575-692Crossref Scopus (566) Google Scholar). Among these classes are monocot mannose-binding lectins that preferably bind to 1–3- or 1–6-linked d-mannoses with the highest affinity (3Van Damme E.J.M. Allen A.K. Peumans W.J. Physiol. Plant. 1988; 73: 52-57Crossref Scopus (126) Google Scholar). Lectins from this protein superfamily having been isolated and characterized so far come from Amaryllidaceae, Alliaceae, Araceae, Liliaceae, Orchidaceae, and recently reported Iridaceae families (2Van Damme E.J.M. Peumans W.J. Barre A. Rouge P. Crit. Rev. Plant Sci. 1998; 17: 575-692Crossref Scopus (566) Google Scholar, 4Van Damme E.J. Astoul C.H. Barre A. Rouge P. Peumans W.J. Eur. J. Biochem. 2000; 267: 5067-5077Crossref PubMed Scopus (42) Google Scholar). All these proteins are found to present three potential carbohydrate binding motifs per subunit, each of which contains a consensus sequence signature QXDXNXVXY, essential for mannose binding (5Ramachandraiah G. Chandra N.R. Proteins. 2000; 39: 358-364Crossref PubMed Scopus (46) Google Scholar).All currently known monocot mannose binding lectins with available three-dimensional structures in the Protein Data Bank are hololectins built from two or four identical one-domain promoters. It has been widely believed that oligomerization and multivalency thereby play an important role in variability of carbohydrate recognition among plant lectins. For instance, tetrameric proteins such as snowdrop and daffodil lectins bind with mannan epitopes on GP120, the major glycoprotein of human immunodeficiency virus, with a high degree of affinity (6Balzarini J. Schols D. Neyts J. Van Damme E. Peumans W. De Clercq E. Antimicrob. Agents Chemother. 1991; 35: 410-416Crossref PubMed Scopus (223) Google Scholar, 7Marchetti M. Mastromarino P. Rieti S. Seganti L. Orsi N. Res. Virol. 1995; 146: 211-215Crossref PubMed Scopus (23) Google Scholar), whereas garlic agglutinin, a dimer, does not (8Barre A. Van Damme E.J. Peumans W.J. Rouge P. Plant Physiol. 1996; 112: 1531-1540Crossref PubMed Scopus (106) Google Scholar, 9Dam T.K. Bachhawat K. Rani P.G. Surolia A. J. Biol. Chem. 1998; 273: 5528-5535Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar). Galanthus nivalis agglutinin (GNA) 1The abbreviations used are: GNA, G. nivalis agglutinin; MBP, maltose-binding protein; r.m.s.d., root mean square (r.m.s.) deviation.1The abbreviations used are: GNA, G. nivalis agglutinin; MBP, maltose-binding protein; r.m.s.d., root mean square (r.m.s.) deviation. from snowdrop bulbs was the first member of this protein superfamily to have its crystal structure determined (10Hester G. Kaku H. Goldstein I.J. Wright C.S. Nat. Struct. Biol. 1995; 2: 472-479Crossref PubMed Scopus (180) Google Scholar), and since then, a series tetrameric or dimeric structures of monocot lectins from amaryllis (Hippeatrum hybrid) (11Chantalat L. Wood S.D. Rizkallah P.J. Peynolds C.D. Acta Crystallogr. D Biol. Crystallogr. 1996; 52: 1146-1152Crossref PubMed Scopus (28) Google Scholar), bluebell (Scilla campanulata) (12Wood S.D. Wright L.M. Reynolds C.D. Rizkallah P.J. Allen A.K. Peumans W.J. Van Damme E.J. Acta Crystallogr. D Biol. Crystallogr. 1999; 55: 1264-1272Crossref PubMed Scopus (41) Google Scholar), daffodil (Narcissus pseudonarcissus) (13Sauerborn M.K. Wright L.M. Reynolds C.D. Grossmann J.G. Rizkallah P.J. J. Mol. Biol. 1999; 290: 185-199Crossref PubMed Scopus (57) Google Scholar), and garlic (Allium sativum) (14Chandra N.R. Ramachandraiah G. Bachhawat K. Dam T.K. Surolia A. Vijayan M. J. Mol. Biol. 1999; 285: 1157-1168Crossref PubMed Scopus (96) Google Scholar, 15Ramachandraiah G. Chandra N.R. Surolia A. Vijayan M. Acta Crystallogr. D Biol. Crystallogr. 2002; 58: 414-420Crossref PubMed Scopus (25) Google Scholar) have been reported as well as a two-domain lectin from S. campanulata (16Wright L.M. Reynolds C.D. Rizkallah P.J. Allen A.K. Van Damme E.J. Donovan M.J. Peumans W.J. FEBS Lett. 2000; 468: 19-22Crossref PubMed Scopus (21) Google Scholar).The basic structural unit for bulb lectins is a dimer, as revealed in the cited literature. In all the studied structures two subunits are related by a pseudo-2-fold symmetry axis and assemble into a tightly bound dimer by exchanging their C-terminal β-strands (residues 101–105 in GNA) to form a hybrid β-sheet (10Hester G. Kaku H. Goldstein I.J. Wright C.S. Nat. Struct. Biol. 1995; 2: 472-479Crossref PubMed Scopus (180) Google Scholar). This mode often referred to as the C-terminal exchange confers a large buried area on the subunit/subunit interface through which a stable dimer is established. An unusual cis peptide bond between Gly-98 and Thr-99 is commonly involved in those oligomers and believed to play a vital role for strand exchange (10Hester G. Kaku H. Goldstein I.J. Wright C.S. Nat. Struct. Biol. 1995; 2: 472-479Crossref PubMed Scopus (180) Google Scholar, 12Wood S.D. Wright L.M. Reynolds C.D. Rizkallah P.J. Allen A.K. Peumans W.J. Van Damme E.J. Acta Crystallogr. D Biol. Crystallogr. 1999; 55: 1264-1272Crossref PubMed Scopus (41) Google Scholar, 13Sauerborn M.K. Wright L.M. Reynolds C.D. Grossmann J.G. Rizkallah P.J. J. Mol. Biol. 1999; 290: 185-199Crossref PubMed Scopus (57) Google Scholar). From an energetic and structural point of view, monomers cannot exist stably when a dimer dissociates, as expected and verified calorimetrically (17Bachhawat K. Kapoor M. Dam T.K. Surolia A. Biochemistry. 2001; 40: 7291-7300Crossref PubMed Scopus (22) Google Scholar). However, three merolectins containing a single protomer have already been isolated and sequenced from Orchidaceae, i.e. Lister ovata and Epipactis helleborine (18Van Damme E.J. Balzarini J. Smeets K. Van Leuven F. Peumans W.J. Glycoconj. J. 1994; 11: 321-332Crossref PubMed Scopus (34) Google Scholar, 19Van Damme J.M. Smeets K. Torrekens S. Van Leuven F. Peumans W.J. Eur. J. Biochem. 1994; 221: 769-777Crossref PubMed Scopus (69) Google Scholar) and Gastrodia elata (20Hu Z. Huang Q.Z. Acta Bot. Yunnanica. 1994; 16: 169-177Google Scholar). In this context it would be of great interest to uncover unique structural features of these monomers different from those presented in oligomers for the purpose of gaining deeper insight into structural mechanisms governing the quaternary association for monocot mannose-binding lectins.Gastrodianin, also known as Gastrodia antifungal protein (GAFP) previously, is purified from the nutritive corms of the orchid G. elata, a traditional Chinese medicinal herb cultured for thousands of years. The plant lacks chlorophyll and leads a parasitic life on the fungus Armillaria mellea. The fungal hyphae invade the nutritive and the primary corm of G. elata during its development. The cortical cells in corms, however, capture and digest the infecting hyphae and transport the released nutrients into a terminal corm for sustaining its growth (21Yang Z. Hu Z. Acta Bot. Yunnanica. 1990; 12: 421-426Google Scholar). Histochemical localization studies in vivo showed that a mannose-binding lectin, later named gastrodianin, accumulates in nutritive corms where fungal infection takes place (20Hu Z. Huang Q.Z. Acta Bot. Yunnanica. 1994; 16: 169-177Google Scholar, 22Jiang L. Xu J.T. Wang H. Liu H.X. Sun Y.R. Acta Bot. Sin. 1993; 35: 593-599Google Scholar). Antifungal assays in vitro also confirmed a strong inhibitive activity of gastrodianin toward a wide range of phytopathogenic fungi, such as Trichoderma viride, Valsa ambiens, Rhizoctonia solani, Gibberella zeae, Ganoderma lucidum, and Botrytis cinerea (23Xu Q. Liu Y. Wang X.C. Gu H. Chen Z.L. Plant Physiol. Biochem. 1998; 36: 899-905Crossref Scopus (53) Google Scholar). As a common phenomenon a number of isolectins simultaneously occur in one plant, and several isoforms have been identified from G. elata, differing subtly from each other in their sequence. At least five of them have been distinguished on the level of cDNA sequences (24Hu Z. Huang Q.Z. Liu X.Z. Yang J.B. Acta Bot. Yunnan. 1999; 21: 131-138Google Scholar, 25Wang X. Bauw G. Van Damme E.J. Peumans W.J. Chen Z.L. Van Montagu M. Angenon G. Dillen W. Plant J. 2001; 25: 651-661Crossref PubMed Google Scholar, 26Wang X.C. Diaz W.A. Bauw G. Xu Q. Montagu M.V. Chen Z.L. Dillen W. Acta Bot. Sin. 1999; 41: 1041-1045Google Scholar). The mature protein is estimated to be 112 amino acids long and behaves as a monomer in solution (25Wang X. Bauw G. Van Damme E.J. Peumans W.J. Chen Z.L. Van Montagu M. Angenon G. Dillen W. Plant J. 2001; 25: 651-661Crossref PubMed Google Scholar). Sequence alignment (Fig. 1) indicates that gastrodianins belong to the superfamily of monocot lectins and share the highest degree of identity with two other orchid merolectins, L. ovata and E. helleborine mannose-binding proteins (18Van Damme E.J. Balzarini J. Smeets K. Van Leuven F. Peumans W.J. Glycoconj. J. 1994; 11: 321-332Crossref PubMed Scopus (34) Google Scholar).We report here the native crystal structures of two gastrodianin isoforms purified from two subspecies of G. elata named gastrodianin-1 and gastrodianin-4 according to their cDNA numbers. With a background of a plethora of attention given to oligomeric plant lectins, the refined structures of gastrodianin provide us with a scientific glimpse of a monomeric mannose-binding lectin for the first time. It also reveals a potential structural mechanisms governing protein assembly in this superfamily, with possible links to their biological roles in the plant. Furthermore, the conclusion from the structural analysis was generally confirmed by mutagenesis experiments on recombinant maltose-binding protein (MBP)-gastrodianin fusion protein.MATERIALS AND METHODSPurification and Crystallization—Gastrodianin-1 and gastrodianin-4 were isolated from newborn terminal corms of two subspecies of the orchids, namely G. elata B1 f. elata and G. elata B1 f. glauca, respectively. The purification and crystallization of gastrodianin-1 have been described previously (27Liu W. Hu Y.L. Wang M. Xiang Y. Hu Z. Wang D.C. Acta Crystallogr. D Biol. Crystallogr. 2002; 58: 1833-1835Crossref PubMed Scopus (13) Google Scholar), and an identical scheme was applied on gastrodianin-4. The only difference in crystallization trials between the two proteins was the different additives used; that is, 2.5% of dioxane for gastrodianin-1 and 3∼5% m-phenylenediamene for gastrodianin-4. Both crystals underwent a long period of growth before shooting.Data Collection—Data collection, processing, and preliminary analysis for gastrodianin-1 were carried out as described previously (27Liu W. Hu Y.L. Wang M. Xiang Y. Hu Z. Wang D.C. Acta Crystallogr. D Biol. Crystallogr. 2002; 58: 1833-1835Crossref PubMed Scopus (13) Google Scholar). A 2.0-Å data set for gastrodianin-4 was collected at room temperature on the beamline of BL-6B at the Photon Factory in KEK, Tsukaba, Japan, with a synchrotron radiation source. X-ray diffraction patterns were record on a Weissenberg camera. Evaluation and scaling of intensity data were processed using DENZO and SCALEPACK in HKL program package (28Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38361) Google Scholar). The completeness of the whole data sets reached 95.3%.Structure Determination—The structure solution for gastrodianin-4 was first achieved by a molecular replacement technique using the program AMoRe (29Navaza J. Acta Crystallogr. Sect. A. 1994; 42: 140-149Google Scholar), with a search probe constructed on a truncated model of subunit C of GNA (PDB code 1MSA (10Hester G. Kaku H. Goldstein I.J. Wright C.S. Nat. Struct. Biol. 1995; 2: 472-479Crossref PubMed Scopus (180) Google Scholar)) without the 11 residues at the C terminus. A reasonable R-factor of 40.8% and a correlation coefficient of 60.7% were given from the top solution after rigid body refinement. In the case of gastrodianin-1, four molecules were estimated in an asymmetric unit of the crystal (27Liu W. Hu Y.L. Wang M. Xiang Y. Hu Z. Wang D.C. Acta Crystallogr. D Biol. Crystallogr. 2002; 58: 1833-1835Crossref PubMed Scopus (13) Google Scholar), and the program MOLREP (30Vagin A. Teplyakov A. Acta Crystallogr. D Biol. Crystallogr. 2000; 12: 1622-1624Crossref Scopus (689) Google Scholar) in the CCP4 program suite (31Collaborative Computational Project N. Acta Crystallogr. D Biol. Crystallogr. 1994; 50: 760-763Crossref PubMed Scopus (19704) Google Scholar) was used instead of AMoRe with the refined model of gastrodianin-4 as the search probe. The correlation factor rose from 25.7 to 58.5% in four cycles, whereas the R-factor went down from 55.1 to 41.3%. Very few close contacts indicated that the final combined coordinates from MOLREP were acceptable.Model Building and Refinement—Although a truncated search model from a subunit of snowdrop lectin was used in molecular replacement, extra electron densities were clearly present at the C terminus, accommodating additional residues beyond residue 98. Proper residues in accordance with the sequence were manually fitted one by one into the density map with the program O (32Jones T.A. Kieldgaard M. O. The manual, Uppsala University, Uppsala, Sweden1993Google Scholar), and several rounds of positional refinement were needed at intervals to improve map quality.After the completion of model building at the C terminus and the replacement of all various residues that differ between gastrodianin and GNA, the final models from molecular replacement for both orchid lectins were refined by simulated annealing at the initial stages followed by energy minimization with the CNS program suite (33Brunger 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. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16929) Google Scholar) and with REFMAC (34Murshudov G.N. Vagin A.A. Dodson E.J. Acta Crystallogr. D Biol. Crystallogr. 1997; 53: 240-255Crossref PubMed Scopus (13776) Google Scholar) at the final round. Anisotropic scaling for initial B factor and solvent contribution was included in the process of refinement with both CNS and REFMAC. The thermal parameters were refined on individual atoms.Non-crystallographic symmetry restraints were applied between four monomers in the gastrodianin-1 structure during the first simulated annealing runs to prevent over-refinement but with decreasing Wa (Non-crystallographic symmetry restraint weight) from cycle to cycle until the last rounds of refinement when neither positional nor B-factor restraints were applied. This allowed for conformational variability among monomers. In both cases solvent molecules as well as sulfate ions were incorporated when the polypeptides became stable until no significant feature remained in the difference map. The occupancy of sulfate ions was refined within a reasonable range of thermal factor. The agreement between the atomic models and x-ray data was checked by SFCHECK (35Vagin A.A. Richelle J. Wodak S.J. Acta Crystallogr. D Biol. Crystallogr. 1999; 55: 191-205Crossref PubMed Scopus (857) Google Scholar), and proper geometries were verified using PROCHECK (36Laskowski R. MacArthur M. Moss D. Thornton J. J. Appl. Crystallogr. 1993; 26: 283-291Crossref Google Scholar).Expression of MBP-Gastrodianin Fusion Protein—Because of the unavailability of soluble recombinant gastrodianin from various expression systems including Escherichia coli and yeast, oligomeric studies had to be performed with resorting to fusion protein expression. The cDNA encoding gastrodianin-1 was obtained from the topical corm of G. elata B1 elata using 3′-rapid amplification of cDNA ends as reported previously (24Hu Z. Huang Q.Z. Liu X.Z. Yang J.B. Acta Bot. Yunnan. 1999; 21: 131-138Google Scholar). A nucleotide sequence corresponding to the mature lectin from residue 1 to 112 was inserted into pMAL-p2 (New England Biolabs), downstream of the gene encoding the MBP, following the protocol described by Wang et al. (25Wang X. Bauw G. Van Damme E.J. Peumans W.J. Chen Z.L. Van Montagu M. Angenon G. Dillen W. Plant J. 2001; 25: 651-661Crossref PubMed Google Scholar). MBP-gastrodianin fusion protein was produced in E. coli host strain TB1 and purified through affinity chromatography on a column packed with amylose resin (New England Biolabs) in the first step. The fraction of interest was subsequently concentrated with ultrafiltration (Millipore) and then loaded onto a Superdex 75 10/300 GL column (Amersham Biosciences) preequilibrated with 50 mm phosphate buffer, pH 7.0, and 0.15 m NaCl at room temperature.Mutation Experiments—To confirm the conclusion from the structure analysis, the mutation from Asp-97—Asn-98 —Ser-99 —Asn-100 —Asn-101—Ala-102—Ile-103 into Gly-97—Thr-98 —Asp-99 —Arg-100 on the loop sequence of gastrodianin gene was fulfilled by multiple-round PCR. Four mutagenic primers were synthesized for the purpose of stepwise nested amplification. A pair of primers, 5′-primer (5′-TCAGATCGGTTGAATTCGGGC-3′) and one of 3′-primers (primer 1, 5′-ACGGTCAGTACCGTATATGACGACGTTACGATC-3′; primer 2, 5′-ATTTCCAACGTTGGTGTGGGTTGCCCAACGGGTCAGTACCGTATATGAC-3′; primer 3, 5′-GTGGGATCCATTAATTTCCAACGTTGGTGTGGG-3′) was applied for each run in sequence, and the intermediate amplified product was used as the PCR template for the following round after purification from agarose gel. DNA sequencing was preformed to guarantee the correctness of the final mutant gene. The expression and purification of the mutant fusion protein were carried out with the same protocol as that for the wild type.Detection of Oligomerization—Oligomerization of the mutant MBP-gastrodianin fusion protein was detected by the following experiments with the wild type fusion protein used as the control. First, the wild and mutant proteins were comparatively analyzed by gel filtration chromatography on Hiload 16/60 Superdex 75 column. The molecular size of wild and mutant fusion proteins were then measured by using a Superdex 75 column (separation range from 3,000 to 70,000) and a Superdex 200 10/300 HR column (separation range from 10,000 to 600,000) (Amersham Biosciences). The two columns were calibrated with low and high Mr protein standard kits, respectively, the former comprising ribonuclease A (Mr 13.7), chymotrypsinogen A (Mr 25.0), ovalbumin(Mr 43.0) and bovine serum albumin (Mr 67.0), and the latter including aldolase (Mr 158), catalase (Mr 232), ferritin (Mr 440), and thyroglobulin (Mr 690). The elution solution used in the experiments contained 0.15 m NaCl in 50 mm phosphate buffer, pH 7.0, which was eluted at the flow rate of 0.5 ml/min.RESULTS AND DISCUSSIONQuality of Model—Both structures determined by molecular replacement and refined at 2.0 Å resolution showed acceptable values of either crystallographic or free R-factors. Reasonable r.m.s. statistics concerning bond length and angles indicated the stereochemical correctness of the final coordinates. All the relative data are listed in Table I.Table IData processing and refinement statisticsGastrodianin-1Gastrodianin-4CrystalSpace groupP21212 (No. 18)C2221 (No. 20)Unit cell dimensions (Å)a = 61.1, b = 91.5, c = 81.1a = 78.1, b = 97.5, c = 36.1Contents of the asymmetric unit4 monomers (A and B, 112 residues; C, 111 residues; D, 110 residues)1 monomer (110 residues)Data collectionNo. of unique reflections31,3779,216Redundancy6.23Resolution range (Å) (last shell)24.69-2.0 (2.11-2.0)18.93-2.0 (2.07-2.0)Completeness (%) (last shell)99.8 (99.8)95.3 (91.2)I/σ (last shell)6.7 (2.5)6.3 (2.9)Rmerge (%) (last shell)9.7 (29.6)6.7RefinementReflections used in refinement31,3399,092Size of test set (%)5.07.5Rcryst/Rfree (%)16.6/20.619.9/23.8Ramachandran plotCore region (%)84.988.5Allowed region (%)15.110.5ModelNo. of protein atoms3,442847No. of hetero atoms105No. of solvent molecules26543r.m.s.d. of bond lengths (Å)0.0080.010r.m.s.d. of bond angles (degree)1.291.45r.m.s.d. of dihedral angles (degree)27.1127.66r.m.s.d. of improper angles (degree)1.181.31Estimated coordinate error (Å)aEstimated according to a Luzzati plot (54) and the Cruickshank method (55).0.189/0.1830.244/0.194Averaged B-factors (Å2)Protein atomsMonomer A, 13.624.7Monomer B, 14.1Monomer C, 22.5Monomer D, 27.0Hetero atoms17.133.0Solvent atoms31.138.2a Estimated according to a Luzzati plot (54Luzzati P.V. Acta Crystallogr. D Biol. Crystallogr. 1952; 5: 802-810Crossref Google Scholar) and the Cruickshank method (55Cruickshank D.W. Acta Crystallogr. D Biol. Crystallogr. 1999; 55: 583-601Crossref PubMed Scopus (500) Google Scholar). Open table in a new tab Four molecules of gastrodianin-1, named A, B, C, and D, resided in the asymmetric unit and behaved somewhat dissimilarly during refinement. High quality electron density accommodated all 112 amino acids for monomer A and B, whereas one or two residues of C and D were missing due to the lack of density at their C termini. Molecules A and B both have far lower averaged B-factors than those of C or D as a result of more stabilizing contacts from neighboring molecules. Two disordered loop regions 45–52 and 77–82 of molecule D, which protrude from molecular surface into solvent, are poorly defined in the 2Fo – Fc map.Given the different packing mode, the crystal structure of gastrodianin-4 contained only one molecule in the asymmetric unit belonging to a different space group (Table I). The model of gastrodianin-4 was refined to a little higher R-factor and has a lower degree of precision than gastrodianin-1, ascribed from higher solvent content and fewer stabilizing contacts among protein chains. Still, the loop region 45–52 is stabilized through a lattice contact, whereas the loop 77–82 keeps disorder, and the density for two residues at the C terminus is missing.C Terminus—Bulb lectins are believed to undergo a post-translational modification during maturation. As for gastrodianin, even if the cDNA sequences corresponds to a 171-amino acid pre-polypeptide, the precise cleavage site at the C terminus is still in debate (24Hu Z. Huang Q.Z. Liu X.Z. Yang J.B. Acta Bot. Yunnan. 1999; 21: 131-138Google Scholar, 25Wang X. Bauw G. Van Damme E.J. Peumans W.J. Chen Z.L. Van Montagu M. Angenon G. Dillen W. Plant J. 2001; 25: 651-661Crossref PubMed Google Scholar). The same pending problems were left on other orchid lectins inclusive of L. ovata mannose-binding protein and E. helleborine mannose-binding protein as well, both of which were thought to exist as monomers (19Van Damme J.M. Smeets K. Torrekens S. Van Leuven F. Peumans W.J. Eur. J. Biochem. 1994; 221: 769-777Crossref PubMed Scopus (69) Google Scholar). In the final model of gastrodianin-1, the C termini of two monomers (A and B) can be clearly traced in the density map because they are strongly fixed by neighboring monomers (D and C respectively) via salt bridges from their N-terminal cationic groups. In this case definite electron densities reasonably account for the presence of the carboxyl group at Asn-112 as the last residue of the chain (Fig. 2). We thereby identified that the mature gastrodianin is comprised of 112 residues as the total length and could even presume that the mature peptides of L. ovata mannose-binding protein and E. helleborine mannose-binding protein are also composed of 112 amino acids since the sequence identity among them reaches as high as 83% (Fig. 1).Fig. 2Stereo view of the C terminus of molecule A in crystal structure of gastrodianin-1, with the corresponding electron density (2Fo – Fc) map contoured at 1.0 σ level. The carboxyl of Asn-112 at the carboxyl C end of the main chain can be clearly defined.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Overall Structure of Gastrodianin—The four crystallographic independent molecules of gastrodianin-1 form two pseudo-dimers (AD and BC) in the asymmetric unit, which related by a non-crystallographic 2-fold axis almost parallel to the y axis (179.36°) (Fig. 3a). Seven hydrogen bonds between A and D or B and C and two salt bridges between A and B are involved in stabilizing intermolecular contacts (Fig. 3b). In the crystallographic lattice of gastrodianin-4, a number of hydrogen bonds are involved in contacts between symmetric related molecules, among which Asn-36 is supposed to be most notable. Both main chain and the side chain of the residue contribute to interactions with neighboring molecules, and very probably for this reason, its dihedral angles fall in the disallowed region of the Ramachandran plot, as is often observed in the oligomeric structures of bulb lectins (12Wood S.D. Wright L.M. Reynolds C.D. Rizkallah P.J. Allen A.K. Peumans W.J. Van Damme E.J. Acta Crystallogr. D Biol. Crystallogr. 1999; 55: 1264-1272Crossref PubMed Scopus (41) Google Scholar, 13Sauerborn M.K. Wright L.M. Reynolds C.D. Grossmann J.G. Rizkallah P.J. J. Mol. Biol. 1999; 290: 185-199Crossref PubMed Scopus (57) Google Scholar). Least-square superpositions of main chain atoms among all independent molecules in the gatrodianin-1 structure give r.m.s differences ranging from 0.27 Å (A-B) to 0.43 Å (A-D) and r.m.s. differences of 0.51 Å between the two isolectins, gas