A novel fucose recognition fold involved in innate immunity

The structure of the AAA−fucose complex described here reveals that the protein has no similarities to any other structurally characterized fucose-recognizing protein and represents the fold of an entire lectin family. In addition, it reveals the architecture of the physiological trimer and allows the identification of the interactions that define the specificity of animal fucose-binding lectins and several related proteins. The presence of the AAA fold can be recognized in the sequence of several other proteins from organisms ranging from bacteria to vertebrates. Some contain the conserved sequence motif that we found characterizes binding to -L-fucose. On the basis of the presence of a fully conserved fucose-binding motif, these proteins can be also expected to behave as fucolectins. The structure also identifies exposed loops around the binding site that contain higher sequence variability among the seven A. japonica fucolectins, suggesting how these loops modulate the fine specificity of lectin for different pathogens.

A. anguilla agglutinin obtained from a commercial source was re-purified by affinity chromatography and crystallized with l-fucose present in the crystallization medium. The crystal structure was determined at 1.9 Å resolution by single isomorphous replacement using the anomalous scatter (SIRAS) signal of a gold derivative crystal (Table 1). In total, 158 residues, 3 ligands — 1 -L-fucose, 1 anion (Cl^-) and 1 cation (either Na⁺ or Ca²⁺) — and a 130 solvent molecules were built into the resulting high quality electronic density map.

Table 1. Crystallographic data collection, phasing and refinement statistics Full Table

AAA folds as a -barrel with jellyroll topology (Fig. 1a). The bulk of the fold consists of eight major antiparallel -strands arranged in two -sheets of five (2, 3, 10, 6 and 7) and three (11, 5 and 8) strands packed against each other. Two short antiparallel strands (4 and 9) close one end of the barrel (bottom in Fig. 1a). The N- and C-terminal strands (1 and 11) protrude 15 Å from this end of the barrel, forming an antiparallel two-stranded -sheet. Five loops connect the two main -sheets at the other end of the barrel (top in Fig 1a); we named these loops complementarity determining regions (CDR) 1−5 by analogy to the immunoglobulin nomenclature. The CDR loops encircle a heavily positively charged hollow that binds the carbohydrate (Fig. 1b). These loops are spanned by the following residue ranges: Gln 22−Ser 31 (CDR1), Arg 41−Cys 50 (CDR2), Ser 53−Asn 58 (CDR3), Arg 79−Ala 90 (CDR4) and Pro 138−Ser 142 (CDR5). Of the five loops, CDR1 protrudes farthest away from the barrel; at its exposed apex, the aliphatic segment of Glu 26 stacks over the aromatic ring of His 27, both reaching toward the central depression. On the other side of the hollow, CDR4 has two solvent-exposed Cys residues (Cys 82−Cys 83) that form a rare disulfide bridge between consecutive residues. At the side of the barrel, a loop-rich substructure, which includes CDR1 and CDR2, tightly binds a cation. A disulfide bridge between Cys 50 and Cys 146 links the N-terminus of 11 with the C-terminus of CDR2, and an Arg 41−Glu 149 salt bridge attaches the loop between 2 and 3 to 11. Both of these interactions hold the cation-binding substructure against the bulk of the fold. A disulfide bridge between Cys 108 and Cys 124 links strands 7 and 8, attaching the two main -sheets. An intrasheet salt bridge between Asp 64 and Arg 131 links the N-terminus of 3 with the C-terminus of 10 (in the front sheet, Fig. 1a). The two disulfide bridges, the two salt bridges and the cation clamp the structure together. The number of structural elements associated with the cation site involved in shaping CDR1 and CDR2 stresses the importance of the conformation of these two loops for the in vivo function of AAA.

Figure 1. The structure of the AAA in complex with -L-fucose. a, Ribbon diagram. The labeled secondary structure elements are represented as arrows (-strands) and helices (310-helices). b, Electrostatic potential energy surface. Electrostatic potential at the molecular surface is colored blue for positive values (>10 kT) and red for negative values (<-10 kT). c, Physiological trimer. This figure, as well as Figs 2, 4 and 5b, were prepared using BobScript41, Raster3D42 and GRASP43. The figure is colored from N- to C-terminus in a progression from blue to red. Bound fucose is shown as stick models. Color coding: nitrogen, blue; chlorine, green; oxygen, red; carbons, gold; and calcium, cyan. Full Figure and legend (120K)

The cation is coordinated with seven oxygen atoms from the main chain or the side chain of six residues: Asn 35 (O), Asp 38 (O1), Asn 40 (O), Ser 49 (O, O1), Cys 146 (O) and Glu 147 (O1). Coordination distances (2.4−2.6 Å) and pentagonal bi-pyramidal geometry are compatible with two types of cations, Na⁺ and Ca²⁺, which are both present in the crystallization medium. We tentatively interpret this feature as a Ca²⁺ ion. Although the cation site is distant from the residues involved in carbohydrate binding, enhancement of carbohydrate affinity by calcium addition has been observed in related proteins³ (see below). This enhancement may result from possible changes induced by cation binding on the conformation of contiguous loops CDR1 and CDR2, which are involved in binding of the ligand.

The crystal structure of the complex shows the bound fucose resting on a highly positively charged depression encircled by the CDRs (Fig. 1b). AAA recognizes the fucose ring O5 and 3-OH and 4-OH hydroxyls using the N of His 52 and the guanidinium groups of Arg 79 and Arg 86 (Fig. 2). A network of hydrogen bonds maintains this triad of residues in optimal positions to provide all the polar interactions to the carbohydrate: His 144 interacts with Arg 86, the main chain oxygen of Ser 53 interacts with His 52 and Asp 81 interacts with Arg 79 to form a salt bridge (Fig. 2). The triad of residues at the center of the binding site forms hydrogen bonds with the fucose axial 4-OH with perfect tetrahedral geometry. Completing the list of polar interactions between AAA and the fucose, Arg 79 hydrogen bonds to O5 and Arg 86 hydrogen bonds to the equatorial 3-OH. The network of hydrogen bonds, together with one of the disulfide bridge (Cys 82−Cys 83), makes CDR4 a rigid and wide loop. This loop provides four of the five polar interactions with the carbohydrate. Its sequence, RGDCCER, contains the RGD motif found in cell adhesion proteins¹⁷. The monosaccharide also makes key van der Waals contacts with the binding site: its ring atoms C1 and C2 rest over the CDR4 disulfide and its C6 docks loosely in a hydrophobic pocket stacking against the rings of His 27 (on top) and Phe 45. These residues, together with Leu 23 and Tyr 46 on the sides, form the binding pocket.

Figure 2. Fucose bound to the AAA-binding site. Stereo view of the final model shows the 2Fo - Fc density around the bound carbohydrate in green. Residues and ligand are represented using the same representation as in Fig. 1. Broken cyan lines represent polar interaction between -L-Fuc and protein residues. Full Figure and legend (51K)

Figure 3. Schematic representation of protein interactions with ligands. a, Interactions observed in the complex. b, Interactions predicted between a D-galactose derivate and AAA. Full Figure and legend (9K)

AAA recognizes blood group oligosaccharides such as the fucosyl terminal in the H type 1 antigen (Fuc1-2 Gal1-3GlcNAc1-3Gal1-4Glc) and in Le^a antigen (Gal1-3[Fuc1-4]GlcNAc1-3Gal1-4Glc) but not the Le^x antigen (Gal1-4[Fuc1-3]GlcNAc1-3Gal1-4Glc)⁶. In oligosaccharides containing these antigens, the -linked fucose protrudes from the linear -linked carbohydrate backbone, making the fucose a suitable target of lectin recognition. The fucosyl terminal trisaccharides of Le^x and Le^a are rigid as a result of close packing between their fucose and galactose moieties^{19, 20}. Crystallographic and NMR studies agree that the conformations of the protein bound carbohydrates are closely related to predicted low energy conformations in solution, although not always that with the lowest energy^{21, 22}. The preferred conformations in solution of Le^a and Le^x terminal trisaccharides are with their GlcNAc (N-acetylglucosamine) bridging the partially juxtaposed Fuc (Fucose) and Gal (Galactose) moieties. In Le^a, the GlcNAc moiety has the 2-N-acetyl group pointing toward the Gal moiety side and the 5-CH₂OH group pointing toward the Fuc side of the trisaccharide. This is reversed in the Le^x, because the Fuc and Gal moieties switch places (3-linkage to 4-linkage and vice versa) in linking to GlcNAc. In the fucosyl terminal trisaccharide of the H type 1 antigen (Fuc1-2Gal1-3GlcNAc), GlcNAc and the smaller Gal moieties switch places with respect to the ones in Le^a and Le^x trisaccharides.

Figure 4. Model of the interactions between AAA and a terminal fucosyl trisaccharide. a, Lea antigen. b, H-type 1. Residues and trisaccharides are shown using the same representation and color scheme of Fig. 1, with the exception of the carbon atoms (pink) of the ligand. Full Figure and legend (56K)

The AAA fold has no similarity to other fucose-recognition proteins, including selectins (C-lectin fold) and plant fucolectins (legume lectin fold), and represents a novel fold within the entire lectin family²³. However, a search for similar structures in the DALI database²⁴ reveals three proteins with a fold similar to AAA, none with significant sequence similarity to AAA: the C1 and C2 repeats of blood coagulation factor V⁸ (FVa-C1 and -C2; FVa-C2 PDB entry 1CZT), the C-terminal domain of sialidase¹⁰ (CSIase and neuramidase; PDB entries 1EUT and 1EUU) and the N-terminal domain of galactose oxidase⁹ (NGOase; PDB entry 1GOF). Recently, other proteins with a similar fold and even lower sequence similarities have been reported: APC10/DOC1 (a ubiquitin ligase)¹¹ and XRCC1 (part of the single-stranded DNA repair complex)¹². An alignment of CSIase, NGOase, FVa-Ca and AAA sequences on the basis of the superposition of structurally similar domains indicates sequence identities with AAA in the 2−14% range (Fig. 5). This alignment shows that only three residues, a Pro, an Asp and an Arg, are strictly conserved; these are equivalent to AAA residues Asp 64, Pro 106 and Arg 131. In the four structures, the Pro residue is in different conformations, but the Asp and the Arg form an equivalent salt bridge. The core and the bottom (as viewed in Fig. 5b) of the -barrel are similar in all these structures. The loops at the top of the barrel, although varied in size and conformation, all encircle a central depression. This depression is shallower and has fewer aromatic residues in AAA. In NGOase and CSIase, the hollow conserves two members of the triad of basic residues that holds the axial hydroxyl of the ligand in AAA, the His and one Arg residue. The third partner, the second Arg residue, is replaced by a Glu in CSIase and an Asn residue in a distant position in NGOase. FVa-C2 has lost all the triad residues related to carbohydrate binding, rendering its pocket the most hydrophobic and the deepest.

Figure 5. Structure-based sequence alignment of AAA with three proteins that share the AAA fold. a, Sequence alignment of AAA on the basis of b, the structural superposition of AAA (cyan) with CSIase (red), NGOase (blue) and FVa-C2 (green). Conserved residues are white in a magenta box; regions of residues with high similarity are boxed in blue and colored red. The positions of residues participating in polar interactions with the fucose are marked with a red star; residues that coordinate the cation are marked with green triangles. This sequence alignment figure and Fig. 6 were prepared with ESPrint44. Full Figure and legend (134K)

Figure 6.  Sequence alignments of AAA with proteins of similar sequence. Sequences are selected from those present in the 12/2001 version of the NCBI nonredundant protein sequence database with >25% sequence identity with A. anguilla agglutinin (AAA). The sequences are from AAA; eFL1-7, A. japonica fucolectins (Entrez accession numbers: BAB0352(3-9); eFL1−4 residues 21−172 and eFL5−7 residues 32−182); S.pneum (1,2,3), S. pneumoniae TIGR4 (NP 346573; 1 is residues 611−741, 2 is residues 762-892 and 3 is residues 904-1034); Tach4, T. tridentatus tachylectin-4 (ref. 3); XPLN, X. laevis (P49263; residues 39-179); fw, D. melanogaster furrowed receptor (NP 511136; residues 65-203); and CG9095 D. melanogaster (AAF48429; residues 15−154). The color scheme is similar to that of Fig. 5, with the exception of the structural disulfide bridge Cys residues (green box). Full Figure and legend (127K)

Except for the Drosophila proteins, most of these sequences conserve a fucose-binding sequence motif: HX₂₄RXDX₄(R/K) (where X stands for any residue). Also highly conserved are the size and hydropathic profile of CDR2, which forms part of the hydrophobic pocket for the fucose methyl group. We expect that these sequences fold as fucolectin-like domains and have similar monosaccharide specificity for l-fucose. Another conserved motif is h₂DGx, where h stands for a small hydrophobic residue (Val, Ala or Ile) and 'x' for a small hydrophilic residue (Asn, Asp or Ser). This motif, located in AAA just after the 3₁₀-helix 3, forms the cation-binding site and provides three of the seven oxygens that bind the cation. Horseshoe crab tachylectin-4, which conserves this motif, shows a four-fold enhancement of its hemagglutination activity for A-type erythrocytes upon the addition of Ca²⁺; this enhancement is abolished by EDTA³. In supporting the function assigned to this motif, two of the other proteins of known structures that conserve this motif also show a bound cation (NGOase and CSIase; Fig. 5).

In the sequence alignment (Fig. 6), all non-eel sequences possess shorter CDR1, lacking five of the residues that form 3₁₀-helix 2. This eliminates the loop protruding-feature observed in AAA, reducing or removing interactions between CDR1 and the putative oligosaccharide antennae and perhaps broadening the specificity spectra of these fucolectins to include Le^x oligosaccharides. Organisms other than bacteria conserve the Cys residues involved in the two interstrand disulfide bridges of AAA (50−146 and 104−128).

Similar to the fucolectin from A. japonica, several isoforms of AAA are present in European eel serum^{16, 29}. The isoform predominant in the crystal shows sequence identities with the seven fucolectins from A. japonica (eFL-1−7 in Fig. 6) in the 68−78% range. All fucolectins from the Anguilla genus (Fig. 6) conserve the residue triad associated with polar interactions with the fucose, showing complete conservation of His 52 and the CDR4 sequence. These fucolectins conserve the size of CDR1 and CDR2, although the loops have interesting sequence variations in residues associated with the fucose C6 pocket. CDR1 shows the greatest sequence variability; most eel fucolectins conserve polar residues at the apex of this loop (such as Glu 26 and His 27 in AAA), probably to interact with the third moiety of their oligosaccharide ligands. Also, most eel sequences conserve two aromatic CDR2 residues in the N-terminus of 4 (such as Phe 45 and Tyr 46 in AAA) that form the C6 pocket. In the eFL-1 and eFL-5 isoforms, the residues in the apex of CDR1 are replaced with smaller residues, making this loop thinner and more flexible, and opening the C6 hydrophobic pocket to the solvent. The eFL-5 isoform maximizes changes in this pocket, replacing all residues involved in the pocket with smaller ones (Leu 23 by Ser, Glu 26 by Gly, His 27 by Gly, Phe 45 by Ser and Tyr 46 by Ser), perhaps conferring a broader specificity on eFL-5. Ser substitutions of Leu 23 and Phe 45 may provide additional polar interactions with hydroxyl groups, such as the 6-OH of a galactose-containing oligosaccharide.

A broader sequence alignment adds mostly adhesion domains of bacterial virulence factors (including sialidases, hyaluronidases, galactosidases and chitinases) and coagulation factors. This broad alignment shows two low sequence variability regions corresponding to 3 (W[W/I/L][R/K/Q]hDh) and the turn between 9 and b10 (RY). Three of these residues play roles in the structural stability of the fold; in AAA, these are Trp 61, which stacks in the core of the barrel, and Asp 64 and Arg 131, which form a salt bridge. However, the most conserved residues in the broad alignment are exposed and clustered in a small area of the protein surface (Trp 60, Asp 64, Arg 131 and Tyr 132 in AAA), suggesting the conservation of an interaction region with another protein or specific receptor.

The crystal structure of the AAA defines the fold of this type of fucose-specific lectin. It provides the basis for understanding its -L-fucose specificity and allows the identification of this fold and prediction of function in many other proteins or protein domains of organisms ranging from bacteria to animals. Interestingly, the CRDs from AAA are one of only a few domains (R-lectins and C-lectins) present in animals that have also been found in bacteria^{30, 31, 32}.

Anguilla anguilla agglutinin was purchased from Sigma. The lyophilized protein was resuspended in 10 mM phosphate, pH 7.2. The active lectin was re-purified by affinity chromatography on an L-fucose-Sepharose 6B matrix and eluted with 200 mM -L-fucose. HPLC size exclusion chromatography was preformed in a Superose-12HR (10 300 mm²) (Amersham Biosciences) that was pre-equilibrated in 10 mM Tris-HCl, 10 mM CaCl₂ and 100 mM NaCl, pH 7.5, buffer. Ovalbumin (Ov; 43 kDa) and bovine serum albumin (BSA; 67 kDa) were used as standards. Filtration was carried out at 4 °C at a flow rate of 0.4 ml min^-1. The eluted protein was detected, by measuring the absorbance at 280 nm, at an elution volume of 12.3 ml, between the two standard elution volumes (12 ml for BSA and 12.8 ml for Ov).

Crystals of AAA were grown in hanging drops of equal amounts of reservoir solution (40−50% (v/v) 2-methyl-2,4-pentanediol in pH 7−9) and protein solution (10 mg ml^-1 protein, 5 mM -L-fucose, 10 mM NaCl, 10 mM Tris-HCl, pH 7.5, and 1 mM CaCl₂) equilibrated against a volume of 1 ml of reservoir solution. Rhombohedrally shaped crystals of eel lectin reached a size of 0.3 mm in 72 h at 18 °C. These crystals belong to the space group R32, with unit cell dimensions a = 65.5, b = 65.5 and c = 245.1 Å (Table 1).

Diffraction data from crystals frozen at 100 K were obtained using CuK radiation from a rotating anode RU-H3R with an image plate detector Raxis IV⁺⁺. All data were processed with CrystalClear (RIGAKU).

A gold heavy atom isomorphous derivative was obtained by overnight soaking of a crystal in mother liquor, with the addition of K₂AuCl₄ at a final concentration of 500 M.

A single gold heavy atom position, found by inspection of the difference Patterson map, was refined using MLPHARE³⁷. The residual map calculated using phases from the single atom partial solution showed the presence of two additional minor sites that were added to the refinement. Heavy atom parameters were initially refined using only centric reflections to a resolution of 3 Å and finally refined including acentric reflections and anomalous differences. The final heavy atom model includes anisotropic temperature factors as parameters. Phases from SIRAS, calculated to 2.5 Å resolution (heavy atom derivative limit), were modified by solvent flattening and histogram matching, and extended to 1.9 Å using DM³⁷. The resulting electron density map shows clear density for secondary structure elements, side chains and ligands. After identification of the correct enantiomer, using the right-handedness of consecutive -strands as an indicator, the initial model was built without difficulty using O³⁸ in the high quality experimental map and refined with CNS³⁹. The sequences for the N- and C-termini were taken from the work of Kelly¹⁶. The last three residues, which did not show interpretable density for their side chains, were modeled as Ala residues. For the rest of the residues, the tentative sequence was obtained by selecting on the basis of the electron density from the seven sequences of the isoforms of A. japonica fucolectin. In cases in which the density suggested a residue not present in those sequences, a new residue based in density shape and protein environment was assigned. It is interesting to note that, toward the end of the refinement, weak indications of alternative sequences could be observed in some cases, suggesting a minor presence in the crystal of other isolectins. The final structure has excellent geometry and average B-values compatible with the resolution of the data and no cut-off in Fs (Table 1).

	Top

1. Hoffmann, J.A., Kafatos, F.C., Janeway, C.A. & Ezekowitz, R.A. Phylogenetic perspectives in innate immunity. Science 284, 1313−1318 (1999). | Article | PubMed | ISI | ChemPort |
2. Vasta, G.R., Quesenberry, M., Ahmed, H. & O'Leary, N. C-type lectins and galectins mediate innate and adaptive immune functions: their roles in the complement activation pathway. Dev. Comp. Immunol. 23, 401−420 (1999). | Article | PubMed | ISI | ChemPort |
3. Saito, T., Hatada, M., Iwanaga, S. & Kawabata, S. A newly identified horseshoe crab lectin with binding specificity to O-antigen of bacterial lipopolysaccharides. J. Biol. Chem. 272, 30703−30708 (1997). | Article | PubMed | ISI | ChemPort |
4. Honda, S., Kashiwagi, M., Miyamoto, K., Takei, Y. & Hirose, S. Multiplicity, structures, and endocrine and exocrine natures of eel fucose-binding lectins. J. Biol. Chem. 275, 33151−33157 (2000). | Article | PubMed | ISI | ChemPort |
5. Springer, G.F. & Desai, P.R. Monosaccharides as specific precipitinogens of eel anti-human blood-group H(O) antibody. Biochemistry 10, 3749−3761 (1971). | Article | PubMed | ISI | ChemPort |
6. Baldus, S.E. et al. Characterization of the binding specificity of Anguilla anguilla agglutinin (AAA) in comparison to Ulex europaeus agglutinin I (UEA-I). Glycoconj. J. 13, 585−590 (1996). | Article | PubMed | ISI | ChemPort |
7. Morgan, W.T. & Watkins, W.M. Unravelling the biochemical basis of blood group ABO and Lewis antigenic specificity. Glycoconj. J. 17, 501−530 (2000). | Article | PubMed | ISI | ChemPort |
8. Macedo-Ribeiro, S. et al. Crystal structures of the membrane-binding C2 domain of human coagulation factor V. Nature 402, 434−439 (1999). | Article | PubMed | ISI | ChemPort |
9. Ito, N. et al. Novel thioether bond revealed by a 1.7 Å crystal structure of galactose oxidase. Nature 350, 87−90 (1991). | Article | PubMed | ISI | ChemPort |
10. Gaskell, A., Crennell, S. & Taylor, G. The three domains of a bacterial sialidase: a -propeller, an immunoglobulin module and a galactose-binding jelly-roll. Structure 3, 1197−1205 (1995). | Article | PubMed | ISI | ChemPort |
11. Wendt, K.S. et al. Crystal structure of the APC10/DOC1 subunit of the human anaphase-promoting complex. Nature Struct. Biol. 8, 784−788 (2001). | Article | PubMed | ISI | ChemPort |
12. Marintchev, A. et al. Solution structure of the single-strand break repair protein XRCC1 N-terminal domain. Nature Struct. Biol. 6, 884−893 (1999). | Article | PubMed | ISI | ChemPort |
13. Kairies, N. et al. The 2.0 Å crystal structure of tachylectin 5A provides evidence for the common origin of the innate immunity and the blood coagulation systems. Proc. Natl. Acad. Sci. USA 98, 13519−13524 (2001). | Article | PubMed | ChemPort |
14. Rudd, P.M., Elliott, T., Cresswell, P., Wilson, I.A. & Dwek, R.A. Glycosylation and the immune system. Science 291, 2370−2376 (2001). | Article | PubMed | ISI | ChemPort |
15. Bezkorovainy, A., Springer, G.F. & Desai, P.R. Physicochemical properties of the eel anti-human blood-group H(O) antibody. Biochemistry 10, 3761−3764 (1971). | PubMed | ISI | ChemPort |
16. Kelly, C. Physicochemical properties and N-terminal sequence of eel lectin. Biochem. J. 220, 221−226 (1984). | PubMed | ISI | ChemPort |
17. Pierschbacher, M.D., Hayman, E.G. & Ruoslahti, E. The cell attachment determinant in fibronectin. J. Cell. Biochem. 28, 115−126 (1985). | Article | PubMed | ISI | ChemPort |
18. Springer, G.F. & Desai, P.R. The immunochemical requirement for specific activity and the physiological properties of eel anti-human blood-group H(O) 7 S Globulin. Vox Sang. 18, 551−554 (1970). | PubMed | ISI | ChemPort |
19. Imberty, A. Oligosaccharide structures: theory versus experiment. Curr. Opin. Struct. Biol. 7, 617−623 (1997). | Article | PubMed | ISI | ChemPort |
20. Imberty, A. et al. Computer simulation of histo-blood group oligosaccharides: energy maps of all constituting disaccharides and potential energy surfaces of 14 ABH and Lewis carbohydrate antigens. Glycoconj. J. 12, 331−349 (1995). | Article | PubMed | ISI | ChemPort |
21. Bush, C.A., Martin-Pastor, M. & Imberty, A. Structure and conformation of complex carbohydrates of glycoproteins, glycolipids, and bacterial polysaccharides. Annu. Rev. Biophys. Biomol. Struct. 28, 269−293 (1999). | Article | PubMed | ISI | ChemPort |
22. Perez, S. et al. Crystal and molecular structure of a histo-blood group antigen involved in cell adhesion: the Lewis x trisaccharide. Glycobiology 6, 537−542 (1996). | PubMed | ISI | ChemPort |
23. Rini, J.M. & Lobsanov, Y.D. New animal lectin structures. Curr. Opin. Struct. Biol. 9, 578−584 (1999). | Article | PubMed | ISI | ChemPort |
24. Holm, L. & Sander, C. Protein structure comparison by alignment of distance matrices. J. Mol. Biol. 233, 123−138 (1993). | Article | PubMed | ISI | ChemPort |
25. Tettelin, H. et al. Complete genome sequence of a virulent isolate of Streptococcus pneumoniae. Science 293, 498−506 (2001). | Article | PubMed | ISI | ChemPort |
26. Adams, M.D. et al. The genome sequence of Drosophila melanogaster. Science 287, 2185−2195 (2000). | Article | PubMed | ISI |
27. Seery, L.T., Schoenberg, D.R., Barbaux, S., Sharp, P.M. & Whitehead, A.S. Identification of a novel member of the pentraxin family in Xenopus laevis. Proc. R. Soc. Lond. B Biol. Sci. 253, 263−270 (1993). | PubMed | ISI | ChemPort |
28. Meindl, A. et al. A gene (SRPX) encoding a sushi-repeat-containing protein is deleted in patients with X-linked retinitis pigmentosa. Hum. Mol. Genet. 4, 2339−2346. (1995). | PubMed | ISI | ChemPort |
29. Horejsi, V. & Kocourek, J. Studies on lectins. XXXVI. Properties of some lectins prepared by affinity chromatography on O-glycosyl polyacrylamide gels. Biochim. Biophys. Acta 538, 299−315 (1978). | PubMed | ISI | ChemPort |
30. Dodd, R.B. & Drickamer, K. Lectin-like proteins in model organisms: implications for evolution of carbohydrate-binding activity. Glycobiology 11, 71R−79R (2001). | Article | PubMed | ISI | ChemPort |
31. Zhang, H. et al. Crystal structure of YbaK protein from Haemophilus influenzae (HI1434) at 1.8 Å resolution: functional implications. Proteins 40, 86−97 (2000). | Article | PubMed | ISI | ChemPort |
32. Saukkonen, K., Burnette, W.N., Mar, V.L., Masure, H.R. & Tuomanen, E.I. Pertussis toxin has eukaryotic-like carbohydrate recognition domains. Proc. Natl. Acad. Sci. USA 89, 118−122 (1992). | PubMed | ChemPort |
33. Baumgartner, S., Hofmann, K., Chiquet-Ehrismann, R. & Bucher, P. The discoidin domain family revisited: new members from prokaryotes and a homology-based fold prediction. Protein Sci. 7, 1626−1631 (1998). | PubMed | ISI | ChemPort |
34. Vogel, W. Discoidin domain receptors: structural relations and functional implications. FASEB J. 13, S77−S82 (1999). | PubMed | ISI | ChemPort |
35. Poole, S., Firtel, R.A., Lamar, E. & Rowekamp, W. Sequence and expression of the discoidin I gene family in Dictyostelium discoideum. J. Mol. Biol. 153, 273−289 (1981). | Article | PubMed | ISI | ChemPort |
36. Weis, W.I., Drickamer, K. & Hendrickson, W.A. Structure of a C-type mannose-binding protein complexed with an oligosaccharide. Nature 360, 127−134 (1992). | Article | PubMed | ISI | ChemPort |
37. Collaborative Computational Project, Number 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50, 760−763 (1994). | Article | ISI |
38. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47, 110−119 (1991). | Article | PubMed | ISI |
39. Brünger, A. et al. Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D 54, 905−921 (1998). | Article | PubMed | ISI |
40. Altschul, S.F. et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389−402 (1997). | Article | PubMed | ISI | ChemPort |
41. Esnouf, R.M. Further additions to MolScript version 1.4, including reading and contouring of electron density maps. Acta Crystallogr. D 55, 938−940 (1999). | Article | PubMed | ISI |
42. Merrit, E.A. & Bacon, D.J. RASTER3D Methods Enzymol. 277, 505−524 (1997). | ChemPort |
43. Nicholls, A., Sharp, K. & Honig, B. Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins Struct. Func. Genet. 11, 281−286 (1991). | ISI | ChemPort |
44. Gouet, P., Courcelle, E., Stuart, D.I. & Metoz, F. ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15, 305−308 (1999). | Article | PubMed | ISI | ChemPort |

	Top

Competing interests statement:  The authors declare that they have no competing financial interests.