Chitin-binding proteins in invertebrates and plants comprise a common chitin-binding structural motif.

Tachycitin, a 73-residue polypeptide having antimicrobial activity is present in the hemocyte of horseshoe crab (Tachypleus tridentatus). The first three-dimensional structure of invertebrate chitin-binding protein was determined for tachycitin using two-dimensional nuclear magnetic resonance spectroscopy. The measurements indicate that the structure of tachycitin is largely divided into N- and C-terminal domains; the former comprises a three-stranded beta-sheet and the latter a two-stranded beta-sheet following a short helical turn. The latter structural motif shares a significant tertiary structural similarity with the chitin-binding domain of plant chitin-binding protein. This result is thought to provide faithful experimental evidence to the recent hypothesis that chitin-binding proteins of invertebrates and plants are correlated by a convergent evolution process.

dues, and glycines and is frequently referred to as a hevein domain (8). It has been well demonstrated that this domain is indispensable for the antimicrobial activity and exhibits a significant conservation in primary sequence (Ͼ40%) and in three-dimensional (3D) 1 structure (9 -12). Although this advanced knowledge has been provided for the plant chitin-binding proteins, less is known for the invertebrate chitin-binding proteins including tachycitin (1,(13)(14)(15)(16). Kawabata et al. (1) identified that tachycitin is a 73-residues chitin-binding protein having antimicrobial activity. They also revealed that tachycitin consists of five intramolecular disulfide bridges; the connected Cys pairs are 6 -33, 12-30, 24 -61, 25-68, and 40 -53. For invertebrates, the chitin-binding domain was assumed to comprise about 65 residues (17) involving a high percentage of cysteine and aromatic residues in a similar manner to the plant chitin-binding domain. On the basis of such similarity between plant and invertebrate chitin-binding proteins, Shen and Jacobs-Lorena (17) proposed a hypothesis that they are correlated by a rare evolutional process, convergent evolution, i.e. proteins from different origins develop to construct the same active site structure to acquire the same function. However, complete lack of 3D-structural information of invertebrate chitin-binding protein obscures the evolutional relationship between invertebrate and plant chitin-binding proteins. The present study determines the solution structure of tachycitin using NMR spectroscopy, which provides the first 3D structural information of invertebrate chitin-binding protein.

EXPERIMENTAL PROCEDURES
An invertebrate chitin-binding protein, tachycitin, was isolated from hemocyte debris of horseshoe crab (Tachypleus tridentatus) as described previously (1) and used without further purification. The NMR samples were prepared by dissolving tachycitin in either 0.3 ml of D 2 O or H 2 O containing 10% D 2 O to give a final concentration of 1-2 mM, whose pH values were adjusted to be 4.0 -6.5 by addition of DCl and/or NaOD. The NMR experiments were performed on JEOL JNM-Alpha 500 and 600 spectrometers operating at temperatures of 15, 20, 30, and 40°C. The two-dimensional experiments, DQF-COSY, TOCSY (mixing time ϭ 75, 85 ms), and NOESY (mixing time ϭ 75, 250 ms), were acquired with low-power (20 Hz) presaturation on the water. The temperature coefficient (Ϫ⌬␦/⌬T, ppb K Ϫ1 ) was estimated from the temperature dependence (15-40°C) of the chemical shift of the H N resonance. The chemical shifts were referenced to the internal standard, TSP (0.00 ppm). Interproton distance restraints were derived from NOE crosspeaks in the NOESY spectra (mixing time ϭ 75 ms), calibrated the peak intensities with known distances (2.2 Å for H ␣ (i)-H N (i ϩ 1) of ␤-sheet and 1.75 Å for H ␤ -H ␤ Ј), and were used as inputs for 3D structural calculations of tachycitin. The NOEs were classified into strong, medium, and weak, corresponding to three distance restraints with an upper limit of 2.7, 3.5, and 5.0 Å, respectively. The upper distance limit was corrected for methyl and methylene protons that were not assigned stereospecifically. The 35 dihedral angle restraints were obtained by measuring 3 J NH-H␣ coupling constants; the angle restraint of Ϫ60 Ϯ 30°was used for the residues having 3 J HN-H␣ coupling constants less than 6 Hz, and that of Ϫ120 Ϯ 30°was used for the residues having 3 J HN-H␣ constants larger than 8 Hz. Hydrogen bond distance restraints were applied between nitrogen and oxygen atoms (2.4 -3.5 Å) and Hn and oxygen atoms (1.5-2.5 Å) for regular secondary structures. The hydrogen bonding was assumed for the residues 18, 27-29, 31, 34, 36, 38, 45, 47, 52, 54, and 59, which show low temperature coefficients * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The atomic coordinates and structure restraints ( (Ͻ4.5 ppb K Ϫ1 ). Our previously determined disulfide bond formations were also used as the distance restraints (1). All NMR data were processed using NMRPipe (18) and PIPP (19), and the structure calculations were performed using simulated annealing (SA) protocol in X-PLOR 3.851 (20). An extended structure of tachycitin was used as the starting structure for calculation, whose C-terminal end is patched (1) with an amide group, CONH 2 . The initial set of restraints included only NOE restraints, no dihedral restraints, and no hydrogen bond restraints. From this initial set of restraints, 100 structures were generated using the SA protocol with heating for 60 ps and cooling for 30 ps. Of those 100 structures, 50 structures with lower total energy were selected as starting structures for the next refinement cycles. The refinement cycles were performed using the SA protocol with heating for 30 ps and cooling for 20 ps. The 25 lowest-energy structures presented in this paper were selected from the 50 structures calculated from the final round of refinement. The program PROCHECK (21) reveals that for all residues, 97.3% of the backbone and dihedral angles fell into core and allowed regions of the Ramachandran map.

RESULTS AND DISCUSSION
The structural determination was performed (Table I) using NMR-derived 1,070 experimental restraints, on the basis of the whole assignments of the 1 H resonances of tachycitin at pH 4.2 and at 30°C, which were deposited to BioMagResBank with the accession number of 4290. Structure of tachycitin ( Fig. 1) appears to comprise a three-stranded ␤-sheet (␤1, ␤2, and ␤3; residues 17-19, 26 -31, and 34 -39) in the N-terminal region and a two-stranded ␤-sheet (␤4 and ␤5; residues 45-47 and 52-54) following a short helical turn (␣1; residues 56 -59) in the C-terminal region. Such arrangements of the secondary structures of tachycitin are not similar to those of any other known antimicrobial peptides in invertebrates; for example, the insect defensin family consists of one long loop, one ␣-helix, and one ␤-sheet from the N terminus (22). As shown in Fig. 1B, a distorted ␤-sandwich structure is constructed by the threestranded and two-stranded ␤-sheets connected through a bending loop (Cys-40 -Leu-44). It appears that this bending loop involves a type III' ␤-turn contributed by the residues Pro-41-Leu-44, for which the formation of a hydrogen bond between Leu-44 H N and Pro-41 OЈ is evidenced by low temperature coefficient of Leu-44 (3.0 ppb K Ϫ1 ). A short segment (residues His-31-Leu-34) flanked between the strands ␤2 and ␤3 constructs a ␤-turn conformation. For this ␤-turn, molecular motional restraint contributed by a disulfide bond (Cys-6 -Cys-33) is suggested by observations of significant line-broadening of the H N resonances for Lys-32 and Cys-33. Another segment comprising six residues (Asn-47-Val-52) flanked by strands ␤4 and ␤5 adapt a ␤-hairpin structure. In this ␤-hairpin, Asn-47 OЈ presumably forms a hydrogen bond with the Lys-51 H N , which is supported by the extremely low temperature coefficient (2.0 ppb K Ϫ1 ) obtained for Lys-51. Overall, the structure of tachycitin is characterized by ␤-sheets flanking short loops and turns, which is typical for most of the small disulfide-rich polypeptides (23).
It was revealed that tachycitin shares a remarkable local structural similarity with a plant chitin-binding protein named hevein. Comparison between our determined structure of tachycitin ( Fig. 2A) and a previously reported structure of hevein (9) (Fig. 2B) clearly shows that an antiparallel ␤-sheet (colored in blue) and a helical turn (colored in red) are constructed in both proteins in highly similar manners. In addition, formation of a disulfide bridge (between Cys-40 and Cys-53) connecting the middle of ␤5 and the C terminus of ␤3 for tachycitin (colored in green, Fig. 2A) is similarly identified in hevein (Fig. 2B). The structural similarity further includes the loop regions, e.g. a hairpin loop structure involved in the antiparallel ␤-sheet (colored in orange). It should be noted that the hairpin loop of tachycitin (Asn-47-Val-52) comprises six residues with ␤␣␣␥␣ L ␤ conformation whereas the corresponding loop of hevein comprises five residues with ␤␣␥␣ L ␤ conformation.
Kawabata et al. (1) reported that the N-terminal 5-28 region of tachycitin shows sequence similarity with the N-terminal 2-21 region of hevein. However, such similarity is not identified by the present study; the secondary structural arrangement, as well as the disulfide-bond patterns, appears to be quite different for the suggested regions.
The structural similarity between segment Cys-40 -Gly-60 of tachycitin and segment Cys-12-Ser-32 of hevein, both comprising the antiparallel ␤-sheet (␤4 and ␤5), was examined by looking at the superimpositions of the segments (Fig. 3). The structural motif shown in Fig. 3 has been found in several plant chitin-binding proteins (10 -12) (Fig. 4). For hevein, segment Cys-12-Ser-32 was identified as an essential chitin-binding domain (24). It appears that arrangements of the two structural motifs shown in Fig. 3 are significantly consistent with each other (backbone RMSD ϭ ϳ1.5 Å). The aromatic sidechain groups of Trp-21 and Trp-23 of hevein (Fig. 3) are known to bind specifically to chitin-derived oligosaccharides through hydrophobic interactions (24,25). This binding is further strengthened by a hydrogen bonding with Ser-19 of hevein (25). As shown in Fig. 3, the residues of Asn-47, Tyr-49, and Val-52 of tachycitin are located at perfectly corresponding positions to the residues of Ser-19, Trp-21, and Trp-23 of hevein. Therefore, one could assume that the region shown in Fig. 3 comprising an antiparallel ␤-sheet and a helical turn (␤4, ␤5, and ␣1; Fig. 2A) in the C-domain of tachycitin serves as an essential chitinbinding site, which protrudes the side-chains of the putative functional residues, Asn-47, Tyr-49, and Val-52. Overall, it could be assumed that the N-terminal region comprising ␤1-␤3  Fig. 2A) behaves as a stable domain so as to locate the C-terminal domain chitin-binding site proper for its function.
Conservation of the chitin-binding structural motif among the chitin-binding proteins in invertebrates and plants was further examined by alignment tests of the proteins with regard to their amino acid sequences corresponding to Cys-40 -Gly-60 of tachycitin (Fig. 4). The 3D structural information has been available for the plant chitin-binding proteins (9 -11). The information is now available for only tachycitin among the invertebrate chitin-binding proteins. It appears that the residues of Cys, Pro, and Gly, all of which have significant influence on the structural constructions, are well conserved in the chitin-binding proteins listed in Fig. 4. Conservation of polar and hydrophobic residues is further identified for the putative chitin-binding residues (e.g. Asn-47, Tyr-49, and Val-52 for tachycitin). For all plant chitin-binding proteins, the 21-residues segments listed in Fig. 4 appear to construct a closely similar 3D structure (9 -11) to the putative chitin-binding site of tachycitin. Further similarity in primary sequence identified between tachycitin, Ag-chit, Pj-chit1, Ch-chit, Peritrophin-44, and Tn-IM (nomenclatures described in the figure legend) assumes that these segments of the invertebrate chitin-binding proteins commonly comprise the chitin-binding structural motif as identified in tachycitin. In 1999, Shen and Jacobs-Lorena (17) proposed a hypothesis that chitin-binding proteins in invertebrates and plants are correlated by a rare evolutional process, convergent evolution. Our present structural determination of tachycitin and the 3D structure-based sequence alignment are thought to provide faithful evidences for the proposed idea of the convergent evolution relationship between invertebrate and plant chitin-binding proteins.  (29). Plants are as follows: hevein from rubber tree (Hevein) (9), Amaranthus caudatus antimicrobial protein 2 (Ac-AMP2) (10), and four homologous domains of wheat germ agglutinin (WGA A, -B, -C, and -D) (11). Residue numbers for each segment are indicated in parentheses. Amino acids conserved between invertebrate and plant proteins are indicated with bold letters. The chitin-binding residues in plants and the corresponding residues in invertebrates are found to be aligned (indicated by asterisks at the bottom), for which polar and hydrophobic residues are colored in red and blue, respectively.
NMR Structure of the Invertebrate Chitin-binding Protein 17931