Báo cáo Y học: Boar salivary lipocalin Three-dimensional X-ray structure and androstenol/androstenone docking simulations pot

Boar salivary lipocalin Three-dimensional X-ray structure and androstenol/androstenone docking simulations Silvia Spinelli 1, *, Florence Vincent 1, *, Paolo Pelosi 2 , Mariella Tegoni 1 and Christian Cambillau 1 1 Architecture et Fonction des Macromole ´ cules Biologiques, CNRS and Universite ´ s Aix-Marseille I & II, France; 2 Dipartimento di Chimica e Biotecnologie Agrarie, Universite ´ di Pisa, Italy The X-ray structure of variant A of authentic boar salivary lipocalin (SAL), a pheromone-binding protein specifically expressed in the submaxillary glands of the boar, has been solved and refined at 2.1 A ˚ resolution. The structure displays a classical lipocalin fold with a nine-stranded sandwiched b barrel and an a helix. A putative glycosylation site, at position 53, has been found to carry a GlcNAc sugar residue. In contrast with what was expected on the basis of mass spectroscopy reports, the internal cavity was found to be devoid of bound pheromonal compound (androstenone or androstenol). Instead, a small electron density volume could be satisfied by a glycerol molecule, a component of the cry- oprotecting liquor. The internal cavity was revealed to be very small for steroid compound accommodation. There- fore, docking and molecular dynamics experiments were performed with both pheromonal compounds. These simulations clearly demonstrate a volume increase of the cavity upon steroid binding and the adaptation of the amino- acid side chains to the steroid molecules. This explains the higher affinity of SAL for both steroid molecules compared to other smaller molecules, although no specific interaction is established with either compound. Keywords: crystal structure; odorant binding protein; saliv- ary lipocalin; steroid; pheromone. Boar salivary lipocalin (SAL) is a pheromone-binding protein (PBP) specifically expressed in the submaxillary glands of the boar [1]. While it is produced in very high amounts (hundreds of milligrams) in the boar’s gland, it cannot be detected in the sow, even using sensitive methods such as Western blot. When extracted from its natural source, this protein is associated with both boar’s phero- mones, 5a-androst-16-en-3-one and 5a-androst-16-en-3-ol [2]. The affinity of SAL for 5a-androst-16-en-3-one has been determined using fluorescence by competition with 1-aminoanthracene (AMA), and was found to be 0.4 l M [1]. Two isoforms of this protein have been purified from the same individual, differing by three amino-acid substitutions [1]. It was suggested that the two isoforms could specifically bind the two pheromonal components. Both forms of boar’s SAL are glycosilated at a single position, Asn53, and present a disulphide bridge connecting two of the three cysteines present, Cys68 and Cys160 [1]. Proteins very similar to SAL have also been purified from the nasal mucosa of the pig [2]. These proteins occur in both sexes and present the same amino-acid sequence as SAL, but differ in their glycan region, which is much larger and more complex in the nasal SAL, although linked to the same Asn53. Interestingly, the same two isoforms produced by the boar’s glands are also expressed in the nose [1]. Ligand-binding experiments have shown that the gland’s SAL proteins, in native and recombinant forms, as well as their nasal counterparts, bind a number of odorants, by far the strongest ligand being one of the pheromone’s compo- nent 5a-androst-16-en-3-one [1]. The similar binding beha- viour of proteins differing for their glycosilated region, including the recombinant unglycosilated protein, suggests that the glycan part of SAL does not play a significant role in the structural characteristics of the binding pocket. The phenomenon of similar or even identical lipocalins expressed in the nose and in other organs involved in chemical communication is well known and documented in other species [3,4]. The urinary proteins of mouse and rat were the first to be identified as pheromone carriers [3]. They are specifically produced by mature males and released in the urine at concentrations of several milligrams per millilitre. When excreted, they are strongly associated with specific sex pheromones. Proteins similar to the urinary proteins are also produced by the salivary glands and by nasal glands of both sexes of the same species [3]. In the hamster, a lipocalin (aphrodisin) is secreted in large amounts in the vaginal discharge [5,6]. No investigation on the nasal protein has been reported for this species, but two odorant binding proteins (OBPs) of the mouse (OBP-Ia and OBP-Ib) bear high similarities to aphrodisin [7]. The role of all these pheromone-binding proteins could involve a function more complex than that of simple carriers for hydrophobic molecules. Both physiological and bio- chemical evidence strongly suggest that such proteins could directly act on specific receptors of the nasal area, probably of the vomeronasal organ [8]. It has been shown that urinary Correspondence to M. Tegoni or C. Cambillau, Architecture et Fonction des Macromole ´ cules Biologiques, UMR 6098, CNRS and Universite ´ s Aix-Marseille I & II, 31 Chemin Joseph Aiguier, 13402 Marseille Cedex 20, France. Fax: + 33 491 16 45 36, Tel.: + 33 491 16 45 01, E-mail: tegoni@afmb.cnrs-mrs.fr or cambillau@afmb.cnrs-mrs.fr Abbreviations: AMA, 1-aminoanthracene; OBP, odorant binding protein; SAL, boar salivary lipocalin; Glc-Nac, N-acetylglucosamine; androstenol, 5a-androst-16-en-3-ol; androstenone, 5a-androst-16-en-3-one. *Note: these authors contributed equally to the work. (Received 7 November 2001, accepted 22 March 2002) Eur. J. Biochem. 269, 2449–2456 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02901.x proteins can accelerate the onset of puberty in young immature female mice, when applied to the nasal cavity [9]. It has also been reported that the same proteins are able to activate a G protein of the vomeronasal system [10]. Therefore, if a specific interaction occurs between phero- mone-binding lipocalins and receptors of the vomeronasal organ, it is important to fully understand the structure of these soluble proteins and try to identify regions where interaction with membrane-bound receptors could occur. In this paper, we report the structure of the variant A of native boar salivary lipocalin, solved and refined at 2.1 A ˚ resolution. Due to the lack of naturally occurring phero- mone in the structure, we have performed dynamic docking simulations with the two specific pheromones of SAL. The results indicate a conformational change of the side-chains of the residues belonging to the internal cavity, which adapt to the steroid molecules. MATERIALS AND METHODS Crystallization conditions and data collection The protein purification and characterization has been previously described [2]. Crystals of SAL were obtained with hanging drops over a reservoir containing 1.95 M ammo- nium sulphate, 0.1 M sodium citrate pH 5.5 and 0.2 M potassium/sodium tartrate. The drops contained 3 lLof protein solution at 3 mgÆmL )1 ,2.4lL of reservoir solution and 0.6 lL of cadmium chloride 10 m M .Thecrystalsizewas further improved by macro-seeding. The crystals belong to the tetragonal space group P4 1 2 1 2 with cell parameters a ¼ b ¼ 69.87 A ˚ , c ¼ 71.42 A ˚ . With one molecule per asymmetric unit, the V m has a value of 2.18 A ˚ 3 ÆDa )1 . A first data set has been collected in the lab at 2.7 A ˚ resolution from a flash-frozen (100 K) crystal, soaked in 12.5% (v/v) glycerol as cryoprotectant, with a MAR- research 345-image plate placed on a Rigaku RU200 rotating anode. The diffraction data were processed with DENZO [11] and SCALA [12]; and data reduction was performed by TRUNCATE [12] (Table 1). Structure determination and refinement The structure was solved by molecular replacement, using AMoRe [13] with the horse allergen lipocalin, which has 44% homology with SAL, as a search model (PDB 1EW3). A clear solution was obtained with a correlation coefficient of 0.416 and a R value of 47.6% at 3.5 A ˚ . Cartesian refinement cycles were carried out with CNS [14] using bulk solvent correction and simulated annealing. At each stage of the refinement, the model was examined on a display and manually refitted into the sigmaA-weighted electron density maps [15] with the program TURBO - FRODO [16]. A data set was subsequently collected at 2.1 A ˚ resolution on the ID14- EH2 synchrotron station (ESRF, Grenoble), and refine- ment was extended to this resolution. The final R work and R free were, respectively, 25.4 and 28.2% (Table 1). No density could be observed for the first seven N-terminal residues, or for the 10 remaining saccharidic residues of the complex glycan. These two disordered parts (14% of total mass), together with the high B-factors, explain the high R free obtained after refinement converged. The stereochem- istry was analysed with PROCHECK [17] which assigned 87.6% residues in the most favourable region and 12.4% residues in the additional allowed region. The coordinates have been deposited in the Protein Data Bank at RCSB (http://www.rcsb.org/pdb/) under entry 1 G6. Docking experiment with androstenol and androstenone The previously identified natural ligands androstenol and androstenone have been built with the program ACD / CHEM- SKETCH [18]. The parameter and topology files for CNS [14] were obtained using the automated procedure of XDICT [19]. The ligands were manually introduced in the cavity in a position suitable to get a minimum number of clashes using TURBO - FRODO [16] running on a SGI OCTANE . The automated docking performed with CNS started from this position and involved positional minimization of the ligand and the protein. During the initial minimization steps, the Ca atoms were constrained to their initial positions and a harmonic restraint was applied on the side chain atoms. The ligand was left without any restraint. In subsequent steps, the constraints were removed and the restraints decreased. RESULTS AND DISCUSSION The overall structure of boar SAL SAL belongs to the lipocalin family as its sequence has identity ranging between 64 and 57% in pair-wise compar- isons with other members of the lipocalin family of known three-dimensional structure [1] (Fig. 1). SAL is a monomer, as are horse allergen [20], MUP [21], a2micro-globulin [21] and aphrodisin [22], but not bovine OBP [23]. It contains 165 residues on which residues 8–165 could be identified in the electron density maps. SAL is formed of the classical Table 1. Final statistics of SAL structure. Data collection and refine- ment statistics. Data collection Resolution limits (A ˚ ) 28.8–2.1 Completeness (I/r(I) > I) 99.6 (99.6) I/r(I) > I (all/last shell) 7.3 (2.4) Redundancy 3.9 (4.0) R sym (all/last shell) a 5.6 (29.5) Refinement Resolution limits 15–2.13 Total number reflections 10 107 Nb of protein atom (+ odor) 1284 Final R factor /R free c 25.2/28.3 Average B factors molecules A and B Side chain 42.5 Main chain 40.2 Solvent 53.8 Ligands (Glycerol + NAG + Cd 2+ ) 50.6 Nb of water molecules 112 Rms deviation from ideal value Bonds(A ˚ )/angles 0.011/1.4 Dihedral angles 27.0 Improper angles 1.11 a R sym ¼ S|I i )<I>|/S I i , where I i is the intensity of an indi- vidual reflection and <I> is the mean intensity of that reflection. b Number of reflections: 1063. 2450 S. Spinelli et al. (Eur. J. Biochem. 269) Ó FEBS 2002 nine-stranded b barrel (residues 19–125, 151–153), a hinge region (144–149) and a a helical domain (130–143) (Fig. 2). Two conserved cysteines (Cys68 and Cys160) form the conserved disulphide bridge between the C-terminal tail and the b barrel (Fig. 2). A Glc-NAc moiety is observed at one out of the three glycosylation consensus sequences detected in the natural protein (see below; Fig. 1). A strong Fourier difference density map was identified at the interface between SAL and a symmetry related molecule. This density could be satisfied with a cadmium ion (Cd 2+ ), a cation present in the crystallization medium. Two variants of SAL have been identified [1]. The crystallized protein belongs to variant A, as confirmed by the electron density in accord with the presence of Val45, Ile48 and Ala73, instead of Ala45, Val48 and Val73 present in variant B. Glycosylation site Three consensus glycosylation sites are present in the sequence of SAL at residues Asn53, Asn38 and Asn99 (Fig. 1), but post-translational modifications studies have shown that only one oligosaccharide complex is linked to Asn53 [1]. The crystal structure confirms this assignation because an electron density for the first Glc-NAc residue is observed linked to Asn53. One of the two glycosylation sites identified in horse allergen [24] is identical to that of SAL and takes the same orientation (Fig. 3). Two putative glycosylation sites have been identified in the native hamster aphrodisin, at Asn41 and Asn69 [25] (Fig. 1). The structure of the recombinant protein indeed did not comprise N-linked saccharides, but Glc-NAc residues have been modelled, linked to each Asn22. If we superimpose the three-dimensional structures of SAL with aphrodisin, Asn53 of SAL corresponds to Asn41 in aphrodisin. In the case of aphrodisin, Asn41 is located in a loop, which contains a disulphide bridge characteristic of its PBP subclass. However, the direction of Asn and its bound sugar is different relative to SAL (Fig. 3). The internal cavity The main cavity, delineated by the b barrel, is not shielded from the solvent because a small opening formed by the ring Fig. 1. Amino-acid sequence of SAL aligned with those of other mammalian lipocalins. Horse allergen and PBPs such as a2l-globulin, MUP1 and hamster aphrodisin. The potential glycosylation sites of SAL and aphrodisin are identified by blue filled circles, the red circle marks where a sugar is observed in the electron density map. Fig. 2. Stereoview of SAL X-ray structure in ribbon representation. The helix is in red, strands are blue and turns or coil are yellow. The GlcNAc residue bound to Asn53 is in white ball-and-stick, and the serendipitous ligand, glycerol, in green ball-and-stick. Ó FEBS 2002 Boar salivary lipocalin structure (Eur. J. Biochem. 269) 2451 of side-chains of residues Phe60, Ala73, Val85, Tyr87, Met41, is observed at one extremity of the barrel (not shown). As the protein used in the crystallization is the natural one, we expected to find an electron density accounting for the natural ligands as identified by mass spectroscopy, namely 5a-androst-16-ene-3-one or 5a-androst-16-ene-3a-ol, both known as sex pheromones for the pig [2]. Surprisingly, this cavity is devoid of such ligands, but a smaller electron density volume was identified, compatible with the size and the shape of a glycerol molecule, probably introduced during cryoprotectant soak- ing (Fig. 4). However, all experiments designed to obtain data from crystals containing 5a-androst-16-ene-3-one or 5a-androst-16-ene-3a-ol failed. The wall of the cavity is mostly formed by hydrophobic residues: Val45, Phe58, Ala73, Val85, Phe93, Leu106 and Leu108. Three weakly polar residues, Tyr87, Tyr123 and Cys75, complete the cavity wall (Table 2). As in the case of horse allergen, Glu121 protrudes inside the cavity, making a hydrogen bond with an alcohol function of glycerol (Fig. 4). In most lipocalins, the b barrel contains one or several internal cavities. In SAL, besides the main cavity, four smaller ones are also observed (Fig. 5A). The volume of the main cavity, as calculated with the program GRASP [26], is 315 A ˚ 3 . The four other cavities have volumes of 40, 33, 33 and 32 A ˚ 3 . Two of the smaller cavities are close to the main one, separated only by side chains, while two others are at the periphery of the protein (Fig. 5A). Unexpectedly, the volume of the main cavity is one of the smallest among lipocalins of known three-dimensional structures. The cavity of MUP has a volume of  510 A ˚ 3 (Fig. 5C), porcine OBP of  500 A ˚ 3 , bovine OBP of  441 A ˚ 3 ,hamster aphrodisin  350 A ˚ 3 , and horse allergen (Fig. 5B), has the smallest,  273 A ˚ 3 . The small size of the main cavity in SAL may be due to the presence of few other cavities. Indeed, as described below, the size of the cavity of the SAL model complexed with the two steroids reaches a value  500 A ˚ 3 . The increase in volume is due in part to the fusion of the main cavity with the two adjacent minor cavities, resulting of amino-acid side-chains conformational changes (Fig. 6). Residues 45 and 73 are located in the cavity and are involved in conformational changes. Interestingly, these residues are Val and Ala, respectively, in isoform A of SAL, the structure of which is described here, and are replaced in isoform B by Ala and Val, respectively. Replacement of residues 43 and 75 with those of isoform B in the SAL X-ray model did not produce significant modifications of the cavity upon complexation (see below). Docking experiments As mass spectroscopy showed that androstenone and androstenol were present together in the same sample of SAL, we decided to perform docking modelling experiments with either of them. We have chosen a simple and safe procedure using minimal constraints and restraints, with the Fig. 3. Ribbon view of the lipocalin scaffold and of the three glycosylated loops and GlucNAC in SAL (white), horse allergen (blue) and aphrodisin (violet). Fig. 4. Stereo view of the electron density map of bound glycerol in the SAL combining site. 2452 S. Spinelli et al. (Eur. J. Biochem. 269) Ó FEBS 2002 aim of testing the flexibility of the side-chain residues in the closed cavity upon steroid docking. This procedure cannot be compared to a full molecular dynamics study, but should yield the most striking features of the docking experiment. The structures of the ligands were generated and minim- ized with the program ACD / CHEMSKETCH [18]. The resulting conformations are different as expected from their chemical difference, as the A ring bears a hydroxyl in androstenol and a keto function in androstenone. The conformation of rings D and C is almost superimposed in the two ligands, but they diverge more closer to ring A, which bears the different functional group. Either of the two steroid molecules has a volume of 240 A ˚ 3 , slightly smaller than that of the cavity (315 A ˚ 3 )and an elongated shape, limiting the number of starting orientations to two positions. The first molecule has a polar function pointing to the opening of the cavity, towards the solvent (position ÔoutÕ). The second, at 180° to the first, has the polar function directed to the centre of the protein (position ÔinÕ) (Fig. 7A,B). Table 2. List of the residues forming the wall of several lipocalins. SAL (this work), porcine OBP [27], bovine OBP [28] (F. Vincent, S. Spinelli, R. Ramori, S. Grolli, M. Tegoni, C. Cambillau, unpublished results), horse allergen [30], mouse MUP [21] and hamster aphrodisin [22]. The residues interacting with androstenol (#) and androstenone (§) are identified by crosses, the letters i and o identify the positions ÔinÕ and ÔoutÕ, respectively. The residues in bold are strictly homologous in the six lipocalins, those in italics are the residues of SAL conserved in other lipocalins. Boar SAL Boar-SAL with androstenol (#) /androstenone (§) p-OBP b-OBP Horse allergen Mouse MUP Hamster aphrodisin #i #o §i §o Leu27 Ile21 Ile22 Leu42 Leu28 Ile15 Met41 ··Phe35 Phe36 Met56 Phe42 Leu29 Val43 ··Val37 Thr38 Val58 Leu44 Phe31 Val45 ···Met39 Phe40 Val60 Leu46 Phe33 Phe58 ····Leu53 Phe54 Ala73 Leu58 Ile47 Phe60 ·· ·Phe55 Phe56 Tyr75 Phe60 Phe49 Ala73 ··Leu68 Val69 Met88 Met73 Val62 Cys75 ···Gly70 Ala71 Phe90 Ala75 Gly64 Val85 ···Val80 Ala81 Leu101 Val86 Thr74 Tyr87 ····Tyr82 Tyr83 Tyr103 Tyr88 Phe76 Asn91 ··Asn86 Asn87 Asn107 Asn92 Asn80 Phe93 ····Phe88 Phe89 Phe109 Phe94 Phe82 Leu106 ····Ile100 Ala101 Leu122 Ala107 Phe94 Leu108 ····Asn102 Asn103 Leu124 Leu109 Asn96 Leu119 · Met114 Leu115 Leu135 Leu120 Met108 Glu121 ····Gly116 Gly117 Glu137 Gly122 Val110 Tyr123 ···Leu118 Phe119 Tyr139 Tyr124 Ala112 Fig. 5. GRASP representation of the cavities observed in three lipocalins. (A) SAL (B) horse allergen and (C) MUP. Fig. 6. GRASP representation of the cavities observedinSAL.(A) Native SAL in the X-ray structure (variant ÔAÕ) (B) in model complex with androstenol in position ÔinÕ,and(C)in position ÔoutÕ. The cavities for variant ÔBÕ are undistinguishable from those for variant A. The cavities of the complexes with androste- none are undistinguishable from those with androstenol. Ó FEBS 2002 Boar salivary lipocalin structure (Eur. J. Biochem. 269) 2453 In the four starting positions, ÔinÕ and ÔoutÕ for andros- tenone and androstenol, we have observed clashes between the steroid and the protein that could not be eliminated manually. Therefore, we performed a careful and conser- vative energy minimization with CNS in order to make it possible for the steroid and the internal side-chains to reach an optimum position. We performed nine cycles of energy minimization refinement in which the harmonic restrained forces were progressively reduced. The structures obtained when convergence was achieved were taken as the final results of the modelling. Relative to the initial structures, we observed in these models the displacement of several side- chains covering the wall of the cavity. Indeed, the initial ÔinÕ or ÔoutÕ orientation of the ligand was kept at the end of the minimization and no changes were observed in the confor- mation of the ligand. After docking, irrespectively to the orientation ÔinÕ or ÔoutÕ, two of the four small cavities collapsed with the main one to form a larger binding pocket (Figs 6B,C). The size of the cavity accommodating the ligand reached 560 A ˚ 3 , although three peripheral cavities of 40, 60 and 20 A ˚ 3 were still present. In the final models, the molecules of androstenone and androstenol adopt the two opposite orientations relative to the b barrel (Figs 7A,B). The interactions between the residues covering the wall of the cavity and the two positions of the two steroids are hydrophobic and very similar in the ÔinÕ or ÔoutÕ positions (Table 2). The shape complementarity between the cavity and the ligands is good but not optimal, as extra unfilled space remains (Fig. 7C,D). In Fig. 7C, the opening of the cavity is clearly visible within the sliced portion of the surface. Indeed, side-chain rotations aresufficienttomakeitpossibleforthesteroidstogoinor out through this opening. As a polar moiety is present in both steroids, it was tempting to speculate on the possible interaction of this group with a polar residue in the cavity (Glu121 or Tyr123). Unfortunately, and in contrast to the glycerol binding, no interaction appears to occur between the polar function of either ligand oriented ÔinÕ and a polar residue in the cavity. Thus, the presence of the polar residues (Glu121 or Tyr123) inside the otherwise apolar cavity is difficult to justify on the basis of androstenol/one binding. For the ÔoutÕ positions, Fig. 7. Views of the model of the complex of androstenol bound in SAL, with in (A) and (C) the ligand in position ÔinÕ, with the OH pointing inside the cavity and, in (B) and (D) the ligand in position ÔoutÕ, with the OH pointing towards the bulk solvent. Besides the polar group, the positions of androstenone (not shown) are undistinguishable from those of androstenol. In (A) and (B): view of the lipocalin scaffold and the bound ligand. In (A) and (B) the rings of the ligand are named A to D from the polar function. In the insert, the chemical structure of the steroids is represented. In (C) and (D) the molecular surface is represented, slabbed at the level of the ligand containing cavity. The blue arrows in (C) and (D) indicates the cavity opening. 2454 S. Spinelli et al. (Eur. J. Biochem. 269) Ó FEBS 2002 the polar functions point in the direction of the bulk solvent. As mentioned above, the conformation of residues 45 and 73 changes upon docking in isoform A (Val45, Ala73) but not in isoform B (Ala45, Val73). Besides this difference, the increase of the volume of the cavity upon ligand biding is very similar in the two isoforms, as well as the interactions between the side chains of the three differing residues (45, 73 and48)andthetwosteroids. The SAL residues forming the wall of the cavity are listed in Table 2 and compared with those observed in other lipocalins, OBPs and PBPs, including pig and bovine OBP, horse allergen, MUP and hamster aphrodisin. Among the residues homologous in the six structures (in blue and red), four are strictly conserved (Phe60, Tyr87, Asn91 and Phe93) and they all interact with the steroids; three of them are aromatic and contribute at shaping the cavity. As reported previously, the volume cavities of horse allergen and MUP are very similar [20], and they are the most similar to the cavity of SAL; the cavities of OBPs come next in similarity, followed by aphrodisin. No specificity was reported for pOBP towards several classes of odours [27] and the same seems to be true for bOBP (F. Vincent, S. Spinelli, R. Ramori, S. Grolli, M. Tegoni, C. Cambillau, unpublished results), although a naturally bound ligand was reported to be copurified and visible in the structure [28]. In contrast, rat OBPs 1–3 exhibit some preference, if not specificity, for certain classes of chemical compounds [29]. SAL affinity for 1-amino-anthracene, for androstenone, and for a group of odors such as 3,7-dimethyl-1-octanol, 2-phenyl ethanol, 2-isobutyl-3-methoxypyrazine has been assayed by fluorescence [1]. The affinity of SAL was by far higher for the steroids than for the other compounds [1]. Our results clearly explain this observation on a structural basis. The cavity of SAL can fit rather closely the two steroids, whereas the other compounds, being smaller, will not establish as many favourable van der Waals interactions. In previous reports, such as that with pOBP, it was observed no or little side-chain movement upon odor binding. Indeed, the pOBP cavity is larger, and the odors are smaller than those of SAL. This indicates that SAL, although being able to bind a certain variety of small molecules, is able to perform side-chain conformational changes in order to optimize its cavity to accommodate the naturally occurring ligands. CONCLUDING REMARKS Boar salivary lipocalin displays a close structural relation with other lipocalins such as horse allergen, a2l-globulin and MUP. It contains a main cavity open to the solvent and several minor ones. Likely due to this fragmentation of the inner volume, the main cavity is smaller than those of other lipocalins. The fragmentation itself is due to side-chain orientations of the internal residues differing slightly from those of other structures. After docking experiments with the natural ligands, the main cavity and two smaller cavities fuse, giving rise to an internal cavity (560 A ˚ 3 ) as large as that of other OBPs. The docking experiments explored two sets of starting positions for androstenone and androstenol. The results indicate a close similarity of the respective solutions for both steroids, both with good complementarity with the cavity wall surface. ACKNOWLEDGEMENTS This study was supported in part by a ÔFondation pour la Recherche Me ´ dicaleÕ (FRM) grant and by the EU BIOTECH Structural Biology project OPTIM (BIO4-98-0420 OPTIM). REFERENCES 1. Loebel, D., Scaloni, A., Paolini, S., Fini, C., Ferrara, L., Breer, H. & Pelosi, P. (2000) Cloning, post-translational modifications, heterologous expression and ligand-binding of boar salivary lipocalin. Biochem. J. 350 Part 2, 369–379. 2. Marchese, S., Pes, D., Scaloni, A., Carbone, V. & Pelosi, P. (1998) Lipocalins of boar salivary glands binding odours and phero- mones. Eur. J. Biochem. 252, 563–568. 3. Bacchini, A., Gaetani, E. & Cavaggioni, A. (1992) Pheromone binding proteins of the mouse, Mus musculus. Experientia 48, 419– 421. 4. Magert, H.J., Hadrys, T., Cieslak, A., Groger, A., Feller, S. & Forssmann, W.G. (1995) cDNA sequence and expression pattern of the putative pheromone carrier aphrodisin. Proc. Natl Acad. Sci. USA 92, 2091–2095. 5. Singer, A.G. (1991) A chemistry of mammalian pheromones. J Steroid Biochem. Mol. Biol. 39, 627–632. 6. Singer, A.G. (1990) Aphrodisin: pheromone or tranducer. Chem. Senses. 15, 199–203. 7. Pes, D. & Pelosi, P. (1995) Odorant-binding proteins of the mouse. Comp.Biochem.Physiol.BBiochem.MolBiol.112, 471–479. 8. Brennan, P.A. (2001) The vomeronasal system. Cell. Mol. Life Sci. 58, 546–555. 9. Mucignat-Caretta, C., Caretta, A. & Cavaggioni, A. (1995) Acceleration of puberty onset in female mice by male urinary proteins. J. Physiol. 486, 517–522. 10. Krieger, J., Schmitt, A., Lobel, D., Gudermann, T., Schultz, G., Breer, H. & Boekhoff, I. (1999) Selective activation of G protein subtypes in the vomeronasal organ upon stimulation with urine- derived compounds. J. Biol. Chem. 274, 4655–4662. 11. Otwinovski, Z. (1993) DENZO: oscillation data and reducing program. In Data collection and processing (Sawyer, L., Isaacs, N.W. & Bailey, S., eds), DLSCI/R34 Daresbury Laboratory, Warrington, UK. 12. CCP4 & 4, C.C.P.N. (1994) The CCP4 suite: programs for crys- tallography. Acta Crystallogr. D 50, 760–766. 13. Navaza, J. (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr. A 50, 157–163. 14. Brunger, 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. (1997–99) Crystallographic and NMR System (CNS), CNSSOLVE. Yale Universty Press, New Heaven, CT. 15. Read, R.J. (1986) Improving Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallog. A 42, 140–149. 16. Roussel, A. & Cambillau, C. (1991) The TURBO-FRODO graphics package. Silicon graphics Geometry Partners Directory, Mountain View, USA. 17. Laskowski, R., MacArthur, M., Moss, D. & Thornton, J. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 26, 91–97. 18. Spessard, G.O. (1998) ACD labs/logP dB 3.5 and Chemsketch 3.5. J. Chem. Inf. Comput. Sci. 38, 218–230. 19. Kleywegt, G.J. & Jones, T.A. (1998) Databases in protein crys- tallography. Acta Crystallogr. D 54, 1119–1131 (CCP4 Proceedings). 20. Lascombe, M.B., Gregoire, C., Poncet, P., Tavares, G.A., Rosinski-Chupin, I., Rabillon, J., Goubran-Botros, H., Mazie, J.C., David, B. & Alzari, P.M. (2000) Crystal structure of the Ó FEBS 2002 Boar salivary lipocalin structure (Eur. J. Biochem. 269) 2455 allergen Equ c 1. A dimeric lipocalin with restricted IgE-reactive epitopes. J. Biol. Chem. 275, 21572–21577. 21. Bocskei, Z., Groom, C.R., Flower, D.R., Wright, C.E., Phillips, S.E., Cavaggioni, A., Findlay, J.B. & North, A.C. (1992) Phero- mone binding to two rodent urinary proteins revealed by X-ray crystallography. Nature 360, 186–188. 22. Vincent, F., Lobel, D., Brown, K., Spinelli, S., Grote, P., Breer, H., Cambillau, C. & Tegoni, M. (2001) Crystal structure of aph- rodisin, a sex pheromone from female hamster. J. Mol. Biol. 305, 459–469. 23. Tegoni, M., Ramoni, R., Bignetti, E., Spinelli, S. & Cambillau, C. (1996) Domain swapping creates a third putative combining site in bovine odorant binding protein dimer. Nat.Struct. Biol. 3, 863–867. 24. Gregoire, C., Rosinski-Chupin, I., Rabillon, J., Alzari, P.M., David, B. & Dandeu, J.P. (1996) cDNA cloning and sequencing reveal the major horse allergen Equ c 1 to be a glycoprotein member of the lipocalin superfamily. J. Biol. Chem. 271, 32951–32959. 25. Henzel, W.J., Rodriguez, H., Singer, A.G., Stults, J.T., Macrides, F., Agosta, W.C. & Niall, H. (1988) The primary structure of aphrodisin. J. Biol. Chem. 263, 16682–16687. 26. Nicholls, A., Sharp, K. & Honig, B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins: Struct. Funct. Genet. 11, 281–296. 27. Vincent, F., Spinelli, S., Ramoni, R., Grolli, S., Pelosi, P., Cambillau, C. & Tegoni, M. (2000) Complexes of porcine odorant binding protein with odorant molecules belonging to different chemical classes. J. Mol. Biol. 300, 127–139. 28. Ramoni, R., Vincent, F., Grolli, S., Conti, V., Malosse, C., Boyer, F.D., Nagnan-Le Meillour, P., Spinelli, S., Cambillau, C. & Tegoni, M. (2000) The insect attractant, 1-octen-3-ol is the natural ligand of bovine odorant binding protein. JBiolChem. 276, 7150– 7155. 29. Lobel, D., Strotmann, J., Jacob, M. & Breer, H. (2001) Identifi- cation of a third rat odorant-binding protein (obp3). Chem Senses. 26, 673–680. 30. Gregoire, C., Tavares, G.A., Lorenzo, H.K., Dandeu, J.P., David, B. & Alzari, P.M. (1999) Crystallization and preliminary crystal- lographic analysis of the major horse allergen Equ c 1. Acta Crystallogr. D 55, 880–882. 2456 S. Spinelli et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . Boar salivary lipocalin Three-dimensional X-ray structure and androstenol/androstenone docking simulations Silvia Spinelli 1, *,. Italy The X-ray structure of variant A of authentic boar salivary lipocalin (SAL), a pheromone-binding protein specifically expressed in the submaxillary