Tài liệu Báo cáo khoa học: S100–annexin complexes – structural insights pptx

REVIEW ARTICLE S100–annexin complexes – structural insights Anne C. Rintala-Dempsey, Atoosa Rezvanpour and Gary S. Shaw Department of Biochemistry, University of Western Ontario, London, Canada Introduction The fusion of cellular phospholipid membranes is required for processes such as membrane reorganiza- tion, exocytosis and vesicular trafficking. In this manner, annexin A1 has been shown to be involved in the vesiculation and sorting of epidermal growth factor receptors [1]. Annexins perform their function by reversibly binding to membranes in a calcium-depen- dent manner through calcium-binding loops on the convex sides of their highly conserved core domains [2–4]. The unique N-terminal sequences of some annexins (A1, A2) are closely associated with the core domain in the absence of calcium, and are sub- sequently released on binding of the ions. S100 proteins, which are dimeric EF-hand calcium-binding proteins, also coordinate calcium ions, but undergo a significant conformational change to expose hydrophobic residues on their surface [5,6]. Identical hydrophobic surfaces on either side of the S100 molecule are able to bind two separate target molecules, such as the N-terminal sequences of annexin proteins. This heterotetrameric interaction allows two membrane-bound annexin pro- teins to be brought into close proximity via an S100 protein. To more clearly understand the interactions between annexin and S100 proteins, efforts have been made to determine the interactions and structures of these two protein families. Numerous structures of Keywords calcium-binding protein; didalcin; EF-hand; membrane interaction; NMR spectroscopy; protein interaction; S100A11; S100B; structure; X-ray crystallography Correspondence G. S. Shaw, Department of Biochemistry, University of Western Ontario, London, ON N6A 5C1, Canada Fax: 1 519 661 3175 Tel: 1 519 661 4021 E-mail: gshaw1@uwo.ca (Received 16 June 2008, revised 29 July 2008, accepted 5 August 2008) doi:10.1111/j.1742-4658.2008.06654.x Annexins and S100 proteins represent two large, but distinct, calcium- binding protein families. Annexins are made up of a highly a-helical core domain that binds calcium ions, allowing them to interact with phospho- lipid membranes. Furthermore, some annexins, such as annexins A1 and A2, contain an N-terminal region that is expelled from the core domain on calcium binding. These events allow for the interaction of the annexin N-terminus with target proteins, such as S100. In addition, when an S100 protein binds calcium ions, it undergoes a structural reorientation of its helices, exposing a hydrophobic patch capable of interacting with its targets, including the N-terminal sequences of annexins. Structural studies of the complexes between members of these two families have revealed valuable details regarding the mechanisms of the interactions, including the binding surfaces and conformation of the annexin N-termi- nus. However, other S100–annexin interactions, such as those between S100A11 and annexin A6, or between dicalcin and annexins A1, A2 and A5, appear to be more complicated, involving the annexin core region, perhaps in concert with the N-terminus. The diversity of these interac- tions indicates that multiple forms of recognition exist between S100 pro- teins and annexins. S100–annexin interactions have been suggested to play a role in membrane fusion events by the bridging together of two annexin proteins, bound to phospholipid membranes, by an S100 protein. The structures and differential interactions of S100–annexin complexes may indicate that this process has several possible modes of protein–pro- tein recognition. 4956 FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS individual annexin and S100 proteins have been determined and, in addition, two structures of the complexes between the two protein families have been completed. In particular, the structures of the com- plexes between S100A10 and annexin A2 and S100A11 and annexin A1 have been solved [7,8], and reveal a common mode of interaction between these two proteins. However, other types of interaction between annexins and S100 proteins have been observed that utilize other portions of the annexin protein. Included amongst this group are interactions of S100A1, S100A11 and S100B with annexin A6 [9,10], and dicalcin, an S100-like protein, that inter- acts with annexins A1, A2 and A5 in a calcium- dependent manner [11]. These complexes indicate that multiple modes of protein–protein recognition may be present. In this review, the structures of annexins, S100 proteins and the complexes between the two protein families are used to provide insights into their complex biology highlighted in the accompanying review [12]. Structures of the annexin proteins In humans, there are 12 different annexin proteins, annexins A1–A11 and A13, that have orthologues in most vertebrates [13]. As of May 2008, there were 63 three-dimensional structures of annexin proteins, most from X-ray crystallographic methods, deposited in the Protein Data Bank (http://www.rcsb.org). These struc- tures include full-length, truncated and mutant forms of the annexins (particularly annexin A5), as well as annexin–protein complexes. In particular, vertebrate structures of human annexins A1, A2, A3, A5, A8 and bovine annexins A4 and A6 have been determined, some in both the calcium-free and calcium-bound forms. Consistent with the first annexin structure dete- rmined, annexin A5 [14–17], all annexins, except annexin A6, form a core domain consisting of four conserved structural repeat sequences (I–IV), each about 70–75 residues in length. Annexin A6 is a unique member of the annexin family possessing two four-repeat core domains connected by a linker region [18], a result of a gene duplication event. As shown in Fig. 1 for annexin A1 [19], each repeat unit is formed from five a-helices (A–E), arranged such that heli- ces A, B, D and E are roughly antiparallel to each other, with helix C nearly perpendicular to these heli- ces. The repeats pack into two distinct arrangements within the core domain. The repeat pairs I ⁄ IV and II ⁄ III pack together, mostly as a result of hydrophobic interactions between helices B and E in each repeat, arranged in a near-antiparallel fashion [19,20]. For example, in annexin A2 [21], residues in helices B and E from repeat I (V54, V57 and V98, L102) and repeat IV (I289, V293 and A330, Y333, L334) form a tight nonpolar network between these two repeat units. In general, the hydrophobic nature of these residues in the annexin sequences is highly conserved. Calcium binding to the annexins promotes their binding to phospholipid-containing membranes. Most structures of annexins show that the coordination of calcium ions by annexins occurs via three residues in the A ⁄ B loop that ligate the calcium ion using their backbone carbonyl atoms and the bidentate side-chain of either an Asp or Glu 38 residues downstream in the D ⁄ E loop (Fig. 1) [22]. Water molecules satisfy the remaining two coordination sites for each calcium ion. In this manner, each annexin protein coordinates one AB Fig. 1. Extrusion of the N-terminal helix in annexin A1 on calcium binding. Ribbon representations of apo-annexin A1 (1HM6) (A) [19] and Ca 2+ -annexin A1 (1MCX) (B) [24]. The core domain repeats are coloured red for repeat I, blue for repeat II, yellow for repeat III and green for repeat IV. The helices of repeat III are labelled A–E. The N-terminus of apo-annexin A1 was resolved in the calcium-free crystal structure and is shown in magenta. The extreme N-terminal helix of annexin A1 is associated with repeat III in the absence of calcium, and essentially takes the place of helix D. In the presence of calcium, the N-terminal helix is not visible in the structure and is presumed to be expelled from the core domain. The calcium ions are shown as orange spheres. A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS 4957 calcium ion per repeat, giving rise to the trademark annexin sequence pattern GXGT-(38)-D ⁄ E [23]. In addition, secondary coordination of calcium ions with lower affinity has also been noted within the D ⁄ E loops of repeats I and III, as in annexins A1 [24] and A2 [21], or in the A ⁄ B loop near the primary calcium site in this region. In both cases, a larger number of water molecules are used to satisfy the calcium coordi- nation. Furthermore, these sites appear to show greater variability within the annexin structures, proba- bly as a result of differences in crystallization condi- tions. Remarkably, the calcium ions all lie on the same curved side of the annexin structures [25], forming a convex surface which has been proposed to interact with phospholipids (Fig. 1). Unlike the structures of EF-hand signalling pro- teins, such as troponin-C [26] or members of the S100 protein family [27,28], the annexins do not appear to undergo a significant structural change within their core domains on calcium binding. For example, a comparison of the calcium-free and cal- cium-bound forms of annexins A1 and A2 shows only 1.56 and 0.72 A ˚ differences between the back- bone arrangements of these structures. The most significant difference between these structures is a dis- ruption of the packing of the helices in repeat III of calcium-free annexin A1 as a result of the presence of an N-terminal helix. The N-terminal extension ranges between 11 (annexin A6) and more than 50 (annexins A7 and A11) residues, and is only found in some annexins (A1, A2, A6, A7, A9, A11). However, annexin A1 is the only member of the group in which the intact N-terminal sequence is visible in the X-ray structure [19]. In the calcium-free state, this structure shows that the N-terminus of annexin A1 forms a kinked a-helix in which residues A2–A11 from this helix are buried against helix E (D259– A271) of repeat III and back on to helix C. The unique feature of this N-terminal helix is that it essentially replaces helix D from the helical packing arrangement in repeat III found in the calcium-bound form of annexin A2 or in other annexin structures. In the presence of calcium, the N-terminal helix is absent from the annexin A1 structure, suggesting that it has been extruded from the core structure [24]. This calcium-sensitive extrusion is reminiscent of that exhibited by the EF-hand protein recoverin, which releases an N-terminal myristoyl group on calcium binding [29,30]. Studies of peptides derived from the N-terminus of annexin A1 reveal that an extruded N-terminus probably has little regular secondary structure, but undergoes a coil–helix transition on protein or membrane binding [31,32]. Structures of S100 proteins The S100 protein family is a group of 25 members, found solely in vertebrates. These proteins undergo a calcium-induced structural change during signalling events. As a result, the calcium-bound forms of S100 proteins are able to interact with target molecules, giv- ing rise to a variety of biological responses, including protein phosphorylation, cell growth and motility, and gene transcription [5]. The structures of several S100 proteins have been determined using NMR spectro- scopy and X-ray crystallography, and show the details of the calcium-binding sites, dimerization motif and structural changes on calcium binding. Unlike the dumbbell shapes of well-studied EF-hand calcium- binding proteins, such as calmodulin [33] and tropo- nin-C [26], the S100 proteins have a more compact, globular structure. As shown for S100A11 (Fig. 2), each S100 monomer comprises two helix–loop–helix motifs, or EF-hands, connected by a flexible linker. The N-terminal calcium-binding site (site I) is termed a ‘pseudo’ EF-hand, because of the presence of two extra residues in the loop and the coordination of calcium mainly through backbone carbonyls, whereas the tighter binding C-terminal site (site II) is a canon- ical EF-hand, binding calcium through acidic side- chains. The majority of S100 proteins form symmetric noncovalent homodimers, a feature that is unique to these proteins within the EF-hand family of calcium- binding proteins. Heterodimers, such as that formed between S100A8 and S100A9 [34], are also possible. The dimer interface is composed of the antiparallel arrangement of helices I and IV of each monomer. The two calcium-binding loops are held in close prox- imity via a short antiparallel b-sheet, and are on the opposite side of the molecule relative to the N- and C-termini. In the calcium-free (apo) structures of several S100 proteins, including apo-S100A11, helices III and IV are nearly parallel to one another, resulting in a num- ber of residues at their interface being inaccessible to the solvent and giving the protein a more ‘closed’ structure [35]. On calcium binding, the N-terminus of helix III moves by almost 40° relative to helix IV and becomes nearly perpendicular to helix IV, exposing hydrophobic residues on both helices (V57, M61, L85, A88, F93 in S100A11), as well as on helix I (I12, I16) and the linker region (L45, A47, F48), which were pre- viously buried in the apo state (Fig. 2). In S100A11, it has also been noted that helix IV becomes elongated on calcium binding. The large conformational change of helix III and the exposure of the hydrophobic resi- dues, first shown for S100B [36–38], have become a S100–annexin complexes – structural insights A. C. Rintala-Dempsey et al. 4958 FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS trademark of the S100 calcium-binding event and are responsible for the interactions of these proteins with a diverse array of target proteins [5]. One member of the S100 family, S100A10, differs from the others as it is unable to bind calcium ions because of a three-residue deletion in site I (N28, N29, T30 of S100A11 are absent in S100A10) and mutations of acidic calcium- coordinating residues in site II (D68 and E77 of S100A11 are substituted with C and S in S100A10) (see sequences in Fig. 3A). Remarkably, the structure of calcium-free S100A10 [7] is nearly identical (rmsd 0.85 A ˚ ) to that of Ca 2+ -S100A11 [8] despite the presence (or absence) of calcium ions in the calcium- binding loops (Fig. 2). I II III IV I′ II′ III′ IV′ annexin A2 annexin A2 Ca 2+ I II III IV I′ II′ III′ IV′ I II III IV I′ II′ III′ IV ′ annexin A1 annexin A1 helix II I movement annexin A1 apo-S100A1 1C a 2+ -S100A1 1 Ca 2+ -S100A11-annexin A1 complex apo-S100A10-annexin A2 complex I II III IV I′ II′ III′ IV′ A B CD Fig. 2. Calcium-induced conformational change of S100 proteins. Ribbon representations of apo-S100A11 (1NSH) (A) [35] and Ca 2+ -S100A11 (1QLS) (B) [8] are shown in similar orientations to demonstrate the conformational changes that occur on calcium binding. The helices are numbered I–IV for one monomer and I¢–IV¢ for the other. Helix I is shown in red (residues E5–Y20 in apo-S100A11 and E7–A23 in Ca 2+ - S100A11), helix II in yellow (K32–E42 and K34–M41), helix III in green (V55–K62 and G56–D66) and helix IV in blue (Q74–V85 and F75–K99). The b-sheets in the calcium-binding loops are shown in cyan and, in Ca 2+ -S100A11, the short a-helix of the linker is shown in grey. Calcium ions are shown as orange spheres. When calcium binds to S100A11, the largest conformational changes occur in the C-terminal EF-hand, whereas the N-terminal EF-hand remains relatively unchanged. Helix III moves  40° with respect to helix IV (green arrow shows the direc- tion of movement), exposing a hydrophobic cleft between helix IV and the linker of one monomer and helix I¢ of the other monomer. (C) Binding of the N-terminal region of annexin A1 (magenta) is mediated by hydrophobic residues of the binding cleft on either side of the S100A11 dimer, making contacts with helices III and IV from one monomer and helix I¢ of the partner monomer simultaneously. (D) The structure of S100A10, an S100 protein that does not bind calcium, bound to the N-terminal region of annexin A2, is shown to illustrate the similarity to the Ca 2+ -S100A11–annexin A1 structure. When the two S100–annexin structures are superimposed, the rmsd for the polypep- tide backbones is 0.87 A ˚ . A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS 4959 In S100A10 and Ca 2+ -S100A11, helices I and IV comprise the dimer interface as in the other S100 structures; however, helix IV is markedly longer than in apo-S100A11, extending nearly to the C-terminus. Helix III has a similar orientation in both S100A10 and Ca 2+ -S100A11, thus exposing very similar hydro- phobic regions and residue composition (Fig. 2). On the basis of these structural observations, it is clear that S100A10 and Ca 2+ -S100A11 should interact with target proteins in a similar manner. Dicalcin is a unique S100 protein Dicalcin is an S100-like protein (originally named p26olf) isolated from the olfactory epithelium of frog (Rana catesbeiana) [39], which has been implicated in the calcium-dependent regulation of olfactory neurones through interaction with a b-adrenergic receptor kinase-like protein [40]. The protein consists of 217 residues arranged in two homologous halves: an N-ter- minus (1–105) and a C-terminus (119–217) connected though a Pro-rich linker region (residues 106–118). Based on the sequence alignment of dicalcin with dimeric S100B or S100A11, dicalcin is predicted to be composed of a pair of approximately 100 residue halves arranged in tandem, each comprising N-termi- nal pseudo (EF-A and EF-C) and C-terminal canoni- cal (EF-B, EF-D) EF-hand calcium-binding sites (Fig. 3). Multiple sequence alignment of the two halves of dicalcin with the EF-hand motifs of 18 different S100 proteins shows a four-residue insertion in each C-terminal EF-hand and a 13-residue insertion in the linker region connecting the N- and C-domains (Fig. 3). Despite the four-residue insertions in sites EF-B and EF-D, dicalcin is still able to bind four cal- cium ions [40,41]. As a result of the sequence similarity B C A1 M2 V3 S4 E5 F6 L7 K8 Q9 A10 W1 1 F12 I13 D14 annexin A1 S1 T2 V3 H4 E5 I6 L7 S8 K9 L1 0 S1 1 L1 2 E1 3 G1 4 annexin A2 C89, E9', I12', E13' ,I 16' E5', M8', E9', M12' L45, A47, F48, L85 , C89, E9', I12' F38, F41, L78, C82, E5', M8 ' F41, A81, C82, Y8 5 A88, S92 L85, A88, C89, S92 C82, Y85, F86, M90, M12 ′ S100A1 1 contacts S100A10 contact s Fig. 3. Similarity of sequences and protein–protein contacts in S100–annexin structures. (A) A sequence alignment of S100A11 (pig), S100A10 (human) and dicalcin is shown to allow a comparison of the residues involved in interactions with annexin peptides. The helices are shaded in similar colours to those used in Fig. 2 and the calcium-binding loop residues are underlined. The N-terminal and C-terminal halves of dicalcin were aligned with S100A11 and S100A10 as described by Tanaka et al. [42]. The shading of the helices for dicalcin corre- sponds to the observed a-helices in a dicalcin model based on the three-dimensional structure of bovine apo-S100B. Schematic representa- tions of the contacts between S100A11 and the annexin A1 peptide (B) and between S100A10 and annexin A2 (C) are shown to illustrate the similarities between the two complexes. The annexin peptides are shown as helical wheels to illustrate the relative positions of the amino acids in the helices, with the key hydrophobic residues shaded in light purple forming an XOOXXOOX motif. The residues of the respective S100 binding partners that make contact (< 6 A ˚ ) with each of the hydrophobic residues are labelled. For example, the side-chain of L7 of annexin A1 is within 6 A ˚ of the side-chains of L45, A47, F48 within the linker of S100A11, L85 and C89 of helix IV and E9¢ and I12¢ of helix I¢ of the other monomer. L7 of annexin A2 makes nearly identical contacts with F38, F41, L78, C82, E5¢ and M8¢ of S100A10, as can be seen when the residues are compared in the sequence alignment of the two proteins. S100–annexin complexes – structural insights A. C. Rintala-Dempsey et al. 4960 FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS of dicalcin with other S100 proteins, Tanaka et al. [42] proposed a model for apo-dicalcin, in which each half of the dicalcin protein consists of two tightly packed EF-hands similar to the fold of an S100 monomer (Fig. 4). The interface for the two halves of dicalcin is arranged in a four-helix bundle, in which helix I in the N-terminal domain and helix V in the C-terminal domain are nearly antiparallel to each other and roughly perpendicular to helices IV (N-domain) and VIII (C-domain). The X-type arrangement of these four helices contains an extensive hydrophobic inter- face similar to the homodimeric dimer interface of S100B [43–45] or S100A11 [8,35]. However, the non- identity of the N- and C-terminal portions of dicalcin might point to fine tuning of the dicalcin structure which is more reminiscent of a heterodimeric complex, such as that observed for S100A8⁄ S100A9 [34]. Structures of S100 proteins complexed with annexins The first structure of an S100 protein complexed with an annexin protein was solved by Rety et al. in 1999 [7] and comprised S100A10 bound to the first 13 resi- dues of the N-terminus of annexin A2. One year later, the structure of Ca 2+ -S100A11 bound to the 14 N-ter- minal residues of annexin A1 was determined [8]. These two structures (Fig. 2) provide valuable infor- mation on how these two protein families physically interact with one another, and how these interactions give rise to the biological functions that have been observed in the cell. S100 proteins have long been known to interact with members of the annexin family, and these interactions play a role in membrane fusion events [46–48]. In particular, the structures reveal a common mode of interaction between these two protein families, as well as key elements for target specificity. Early studies have shown that the binding of annex- ins A1 and A2 to Ca 2+ -S100A11 and S100A10, respectively, is strongly dependent on the unique N- terminal regions of the annexin proteins [49–52]. This was confirmed by the crystal structures of Ca 2+ - S100A11 [8] and S100A10 [7] in the presence of N-terminal annexin peptides. Despite the number of differences between the sequences of the S100 proteins (Fig. 3) and the calcium-bound states of the proteins, and the fact that the two annexin peptides are from different protein sources, both structures contain two annexin peptides per S100 dimer, located in near-iden- tical binding sites on either side of the S100 molecule (Fig. 2). Each peptide makes contact with both S100 monomers, resulting in the bridging together of the two monomers by the annexin protein. In each case, the peptides form a-helical structures when bound to their S100 binding partners, as predicted previously on the basis of their sequences (acetyl-AMVSEFLKQAW- FID and acetyl-STVHEILSKLSLEG for annexins A1 and A2, respectively) [50,51,53] and the structure of this region in the calcium-free form of annexin A1 [19]. Furthermore, N-acetylation of the peptides has been found to be a requirement for S100–annexin interactions [8,50,53], as removal of the acetyl group in annexin A2 results in a 2700-fold decrease in binding affinity to S100A10 [53]. Although no direct contacts are made between the acetyl groups and the S100 pro- teins, it has been suggested that the acetyl group aids in the stabilization of the helix dipole of the annexin N-terminus, and therefore the required helical confor- mation of the peptide. The rmsd for the backbones for the entire Ca 2+ -S100A11–annexin A1 and S100A10– annexin A2 complexes is 0.87 A ˚ . This is a clear indica- tion of a common mode of interaction between these members of the S100 and annexin families. Hydrophobic interactions between the annexin N-termini and the S100 proteins play a major role in their interactions. The amphipathic nature of the ann- exin peptides presents a series of hydrophobic residues on one face that interact with S100A10 or Ca 2+ - S100A11. In annexin A1, the hydrophobic surface is made up of residues V3, F6, L7 and A10 and, in ann- exin A2, it is made up of residues V3, I6, L7 and L10 (Fig. 3). This representation clearly indicates a strong conservation of hydrophobic residues at these positions (XOOXXOOX; X = hydrophobic residue, O = polar residue), which make the largest number of contacts I IV II III V VI VIII VII Fig. 4. Model of apo-dicalcin based on the three-dimensional struc- ture of bovine apo-S100B [42]. The ribbon diagram of apo-dicalcin is shown to illustrate the first ‘half’ of the dicalcin protein, composed of helices I–IV (residues 1–105) and the second portion formed from helices V–VIII (residues 119–217). The ribbon diagram shows helices I and V (red), II and VI (yellow), III and VII (green), and IV and VIII (blue) for the two homologous halves of the protein. An extended linker region (residues 106–118, shown in grey) joins the C-terminus of helix IV with the N-terminus of helix V. A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS 4961 with the S100 proteins. In the S100A10–annexin A2 structure, V3 of the annexin A2 peptide interacts with a large number of residues in helix I¢ (E5¢,M8¢,E9¢, M12¢). Similarly, E9¢, I12¢, E13¢ and I16¢ of S100A11 are in close proximity to V3 of annexin A1 (Fig. 3B). The importance of this residue is further illustrated by its substitution with a polar amino acid, which leads to a complete loss of binding between annexin A2 and S100A10 [53]. However, Ca 2+ -S100A11–annexin A1 complex formation seems to be less sensitive to substi- tution, as replacement of V3 with Ala results in little change in binding affinity [54]. The residue at posi- tion 6 (I or F) makes numerous contacts with helix IV (C82, Y85, F86, M90 in S100A10 and L85, A88, C89, S92 in S100A11). The side-chain of L7 makes the larg- est number of contacts with the S100 proteins by inter- acting with residues in the linker (F38, F41 in S100A10 and L45, A47, F48 in S100A11), helix IV (L78, C82 in S100A10 and L85, C89 in S100A11) and helix I¢ (E5¢,M8¢ in S100A10 and E9¢, I12¢ in S100A11). A decrease in the size of the hydrophobic side-chain at positions 6 and 7 by substitution with either Ala or Val in both annexins A1 and A2 dramat- ically reduces binding, indicating the close packing of the S100–annexin interaction [53,54]. The residue at position 10 (A or L) is near helix IV (A81, C82, Y85 in S100A10 and A88, S92 in S100A11). Several hydro- gen bonds between the peptides and the S100 proteins stabilize the interaction. The structures of the S100A10–annexin A2 and Ca 2+ -S100A11–annexin A1 heterotetramers show how a single S100 protein may interact with two annexin proteins [7,8,55]. In both cases, the interaction utilizes the N-terminus of the annexin protein, a region of the protein that is expelled from the annexin core structure on calcium binding to the annexin protein. As pro- posed by Gerke and Moss [56], this would provide an elegant mechanism, whereby calcium binding by an annexin protein not only promotes its association with a phospholipid membrane, but also facilitates interac- tion with either S100A10 or Ca 2+ -S100A11, allowing two membrane surfaces to be brought within close proximity for a fusion or vesiculation event. Insights into other S100–annexin interactions Other interactions between the S100 and annexin fami- lies have been reported. Consistent with the calcium- induced conformational change observed in S100A11, most of these complexes require the calcium form of the S100 protein, although there are a few calcium- independent interactions, including S100A4 and annexin A2 [57]. Some annexins, such as annexin A6, have many possible S100 binding partners, e.g. S100A1 [9], S100A6 [58], S100A11 [10] and S100B [9], whereas other S100 proteins can interact with several different annexins. For example, S100A6 has been shown to bind annexins A2 [59], A6 [58] and A11 [60]. The S100A6–annexin A11 interaction appears to involve a similar pattern of hydrophobic residues (XOOXXOOX) from the N-terminal extension of ann- exin A11 (L52, M55, A56 and M59) [61] as observed for annexins A1 and A2, and this region has been pre- dicted to adopt an amphipathic helix. Similar to the S100A10–annexin A2 complex, S100A4 also interacts with the N-terminus of annexin A2 in a calcium-inde- pendent manner [57]. Alternatively, the interaction between S100A11 and annexin A6 has been found to involve residues within each of the two core domains of annexin A6 [10]. A similar conclusion has been reached for the interaction of S100A1 and S100B with annexin A6 [9]. In the latter case, and for the S100A12–annexin A5 interaction [62], it has also been observed that the extreme C-terminus of the S100 pro- tein is not involved in the annexin interaction. This is in contrast with observations for the S100A10–annexin A2 and Ca 2+ -S100A11–annexin A1 complexes (Figs 2 and 3), where the C-termini of the S100 proteins are indispensable, probably as a result of the elongated nature of helix IV which extends nearly to the last resi- due in each protein. Together, these results indicate that multiple modes of binding between S100 proteins and annexins are possible. On the basis of the similarity of the amino acid sequences of the S100 proteins with those of dicalcin (Fig. 3), it is perhaps not surprising that in vivo and in vitro data show that dicalcin interacts with ann- exins A1, A2 and A5 in a calcium-dependent manner [11]. Furthermore, as annexins A1 and A2 utilize the N-terminal helix region to interact with S100A10 and Ca 2+ -S100A11, respectively, it may be suggested that a similar mode of interaction is used for these annexins with dicalcin. In this regard, N-terminally truncated forms of annexins A1 and A2 exhibit a calcium-sensi- tive interaction with dicalcin, albeit approximately four- to five-fold weaker than that for the full-length protein, indicating that the annexin N-terminal helix is not the sole binding site [11]. This finding is consistent with the sequence of annexin A5 which lacks a corre- sponding N-terminal region to annexins A1 and A2, and yet is able to interact with calcium-bound dicalcin. The results may point to two separate regions utilized by annexins A1 and A2 for their interactions with dicalcin. A similar conclusion has been reached for the interaction of Ca 2+ -S100A11 with annexin A6, where S100–annexin complexes – structural insights A. C. Rintala-Dempsey et al. 4962 FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS the N-terminal sequence of annexin A6 is not neces- sary for the interaction [10]. Uncovering the potential binding sites in the core domains of annexins A1, A2 and A5 for dicalcin or annexin A6 for S100A11 has not yet been attempted. Other than the calcium-sensitive extrusion of the N-terminal helix from annexins A1 and A2, the most significant structural change that occurs is the expo- sure of Trp187 in annexin A5 at high calcium ion concentrations [14,17,63]. The exposure of this residue facilitates binding to the phospholipid membrane, and so is unlikely to be used also for interactions with an S100 protein. In addition, this residue is not con- served in annexins A1 and A2, where Lys residues exist. Rather, the most obvious choice for S100 pro- tein binding with the annexins is the opposite ‘side’ of the core domain from the membrane-binding region. An attractive site may be helix C from domain IV, as this helix sits near the bottom of the structure (Fig. 1). In the absence of calcium, helix C is protected by the N-terminal helix in annexins A1 and A2. On calcium binding to the annexin protein, several residues near helix C become mostly exposed (E305, N309, D310, A313, K317 in annexin A1). Sev- eral of the analogous residues are also exposed in annexin A5. However, analysis of this helix does not reveal the XOOXXOOX motif used in the N-terminal helix, suggesting that a different mode of interaction may occur. It is also interesting that residues V287– V298 in annexin A2 match the TRTK-12 consensus motif observed for peptide binding to S100B [64]. However, most of these residues appear to be buried in annexin A2. Further experiments are needed to confirm whether this or some alternative site on the annexin proteins is used to interact with dicalcin and other S100 proteins. Although the structures of Ca 2+ -S100A11 and S100A10 clearly show the surfaces used to interact with annexins A1 and A2, respectively, models of other S100–annexin complexes, such as Ca 2+ -S100A11 with annexin A6 or calcium-bound dicalcin with annexins A1, A2 or A5, are not available. However, some information about potential binding sites can be gleaned by an examination of the S100 protein struc- tures in the apo- and calcium-bound states. For exam- ple, S100A11 utilizes several residues in helix I (E9, I12, I13, I16), the linker (L45, A47, F48) and the extreme C-terminus of helix IV (L85, A88, C89, S92) to interact with annexin A1 (Fig. 3). Many of these residues are inaccessible in the calcium-free state and would be expected to provide an interactive surface not only for annexin A1, but also for other annexin proteins. It will be important to complete site-directed mutagenesis experiments on S100 proteins, such as S100A11 and dicalcin, to understand the roles of these residues in the affinities and interactions with different annexin proteins. Future perspectives Of the 25 members of the S100 protein family, seven (S100A1, S100A4, S100A6, S100A10, S100A11, S100A12 and S100B) have been shown to interact with at least one of the 12 annexin proteins. In addi- tion, some S100 proteins, such as S100A6, appear to form complexes with several annexin proteins (A2, A5, A6 and A11). More recently, the unique S100 protein dicalcin has been shown to bind to annex- ins A1, A2 and A5 in a calcium-sensitive manner. On the basis of the association of annexins with anionic lipid membranes, it is probable that most of these S100 proteins coordinate with annexins to facilitate the association of two membrane surfaces important for cellular events, such as vesicle formation. Struc- tural studies have established that calcium binding to the S100 protein (except S100A10) is required in order to facilitate most S100–annexin interactions. However, only two three-dimensional structures (S100A10–annexin A2, Ca 2+ -S100A11–annexin A1) are available that show how this interaction might occur. Both structures show that the annexin mole- cule utilizes an XOOXXOOX motif in its extreme N-terminal helix to bridge helices III and IV of one subunit with helix I¢ of the other in the S100 protein. Alternatively, several S100–annexin complexes, includ- ing those of S100A1, S100A11 and S100B with ann- exin A6, and dicalcin with annexins A1, A2 and A5, appear to require the annexin core domain for optimal binding. Future experiments are needed to narrow down the unique regions on the annexins most important for their calcium-sensitive interactions with different S100 proteins. Furthermore, the three- dimensional structures of calcium-bound S100 pro- teins complexed with different annexin proteins will be required in order for details of the recognition modes between these important proteins to be identi- fied. Together with advances in S100–annexin biology, this information will provide a detailed description of their roles in calcium signalling. Acknowledgements This work was supported by a grant from the Cana- dian Institutes of Health Research (GSS) and an award from the Canada Research Chairs program (GSS). A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS 4963 References 1 Futter CE, Felder S, Schlessinger J, Ullrich A & Hopkins CR (1993) Annexin I is phosphorylated in the multivesicular body during the processing of the epider- mal growth factor receptor. J Cell Biol 120, 77–83. 2 Gerke V, Creutz CE & Moss SE (2005) Annexins: link- ing Ca 2+ signalling to membrane dynamics. Nat Rev Mol Cell Biol 6, 449–461. 3 Rescher U & Gerke V (2004) Annexins – unique mem- brane binding proteins with diverse functions. J Cell Sci 117, 2631–2639. 4 Lemmon MA (2008) Membrane recognition by phos- pholipid-binding domains. Nat Rev Mol Cell Biol 9 , 99–111. 5 Santamaria-Kisiel L, Rintala-Dempsey AC & Shaw GS (2006) Calcium-dependent and -independent interactions of the S100 protein family. Biochem J 396, 201–214. 6 Smith SP & Shaw GS (1998) A change-in-hand mecha- nism for S100 signalling. Biochem Cell Biol 76, 324–333. 7 Rety S, Sopkova J, Renouard M, Osterloh D, Gerke V, Tabaries S, Russo-Marie R & Lewit-Bentley A (1999) The crystal structure of a complex of p11 with the ann- exin II N-terminal peptide. Nat Struct Biol 6, 89–95. 8 Rety S, Arie J-P, Tabaries S, Seeman J, Russo-Marie F, Gerke V & Lewit-Bentley A (2000) Structural basis of the Ca 2+ -dependent association between S100C (S100A11) and its target, the N-terminal part of ann- exin I. Structure 8, 175–184. 9 Garbuglia M, Verzini M, Hofmann A, Huber R & Donato R (2000) S100A1 and S100B interactions with annexins. Biochim Biophys Acta 1498, 192–206. 10 Chang N, Sutherland C, Hesse E, Winkfein R, Wiehler WB, Pho M, Veillette C, Li S, Wilson DP, Kiss E et al. (2007) Identification of a novel interaction between the Ca(2+)-binding protein S100A11 and the Ca(2+)- and phospholipid-binding protein annexin A6. Am J Physiol Cell Physiol 292, C1417–C1430. 11 Uebi T, Miwa N & Kawamura S (2007) Comprehensive interaction of dicalcin with annexins in frog olfactory and respiratory cilia. FEBS J 274, 4863–4876. 12 Miwa N, Uebi T & Kawamura S (2008) S100–annexin complexes – biology of conditional association. FEBS J 275, 4956–4966. 13 Fernandez MP & Morgan RO (2003) Structure, func- tion and evolution of the annexin gene superfamily. In: Annexins: Biological Importance and Annexin-Related Pathologies (Bandorowicz-Pikula J, ed.), pp. 21–37. Landes Bioscience ⁄ Eurekah.com: Georgetown, TX. 14 Concha NO, Head JF, Kaetzel MA, Dedman JR & Seaton BA (1993) Rat annexin V crystal structure: Ca(2+)-induced conformational changes. Science 261, 1321–1324. 15 Huber R, Romisch J & Paques EP (1990) The crystal and molecular structure of human annexin V, an anti- coagulant protein that binds to calcium and mem- branes. EMBO J 9, 3867–3874. 16 Huber R, Schneider M, Mayr I, Romisch J & Paques EP (1990) The calcium binding sites in human annexin V by crystal structure analysis at 2.0 A ˚ resolution. Implications for membrane binding and calcium chan- nel activity. FEBS Lett 275, 15–21. 17 Sopkova J, Renouard M & Lewit-Bentley A (1993) The crystal structure of a new high-calcium form of annexin V. J Mol Biol 234, 816–825. 18 Avila-Sakar AJ, Creutz CE & Kretsinger RH (1998) Crystal structure of bovine annexin VI in a calcium- bound state. Biochim Biophys Acta 1387, 103–116. 19 Rosengarth A, Gerke V & Luecke H (2001) X-ray structure of full-length annexin I and implications for membrane aggregation. J Mol Biol 306, 489–498. 20 Weng X, Luecke H, Song IS, Kang DS, Kim SH & Huber R (1993) Crystal structure of human annexin I at 2.5 A ˚ resolution. Protein Sci 2, 448–458. 21 Shao C, Zhang F, Kemp MM, Linhardt RJ, Waisman DM, Head JF & Seaton BA (2006) Crystallographic analysis of calcium-dependent heparin binding to annexin A2. J Biol Chem 281, 31689–31695. 22 Seaton BA (1996) Annexin V molecular structure, ligand binding and biological function. In: Annexins: Molecular Structure to Cellular Function (Seaton BA, ed.), pp. 15–29. R.G. Landes: Georgetown, TX. 23 Geisow MJ, Fritsche U, Hexham JM, Dash B & John- son T (1986) A consensus amino-acid sequence repeat in Torpedo and mammalian Ca 2+ -dependent mem- brane-binding proteins. Nature 320, 636–638. 24 Rosengarth A & Luecke H (2003) A calcium-driven conformational switch of the N-terminal and core domains of annexin A1. J Mol Biol 326, 1317–1325. 25 Swairjo MA, Concha NO, Kaetzel MA, Dedman JR & Seaton BA (1995) Ca(2+)-bridging mechanism and phospholipid head group recognition in the membrane- binding protein annexin V. Nat Struct Biol 2, 968–974. 26 Herzberg O & James MNG (1985) Structure of the calcium regulatory muscle protein troponin C at 2.8 A ˚ resolution. Nature 313, 653–659. 27 Nelson MR & Chazin WJ (1998) Structures of EF-hand Ca 2+ -binding proteins: diversity in the organization, packing and response to Ca 2+ binding. Biometals 11, 297–318. 28 Maler L, Sastry M & Chazin WJ (2002) A structural basis for S100 protein specificity derived from compara- tive analysis of apo and Ca 2+ -calcyclin. J Mol Biol 317, 279–290. 29 Ames JB, Porumb T, Tanaka T, Ikura M & Stryer L (1995) Amino-terminal myristoylation induces coopera- tive calcium binding to recoverin. J Biol Chem 270, 4526–4533. 30 Ames JB, Tanaka T, Ikura M & Stryer L (1995) Nuclear magnetic resonance evidence for Ca 2+ -induced S100–annexin complexes – structural insights A. C. Rintala-Dempsey et al. 4964 FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS extrusion of the myristoyl group of recoverin. J Biol Chem 270, 30909–30913. 31 Hu NJ, Bradshaw J, Lauter H, Buckingham J, Solito E & Hofmann A (2008) Membrane-induced folding and structure of membrane-bound annexin A1 N-terminal peptides: implications for annexin-induced membrane aggregation. Biophys J 94, 1773–1781. 32 Yoon MK, Park SH, Won HS, Na DS & Lee BJ (2000) Solution structure and membrane-binding property of the N-terminal tail domain of human annexin I. FEBS Lett 484, 241–245. 33 Babu YS, Bugg CE & Cook WJ (1988) Structure of calmodulin refined at 2.2 A ˚ resolution. J Mol Biol 203, 191–204. 34 Korndorfer IP, Brueckner F & Skerra A (2007) The crystal structure of the human (S100A8 ⁄ S100A9) 2 het- erotetramer, calprotectin, illustrates how conforma- tional changes of interacting alpha-helices can determine specific association of two EF-hand proteins. J Mol Biol 370, 887–898. 35 Dempsey AC, Walsh MP & Shaw GS (2003) Unmask- ing the annexin I interaction from the structure of Apo-S100A11. Structure 11, 887–897. 36 Drohat AC, Baldisseri DM, Rustandi RR & Weber DJ (1998) Solution structure of calcium-bound rat S100B (bb) as determined by nuclear magnetic resonance spec- troscopy. Biochemistry 37, 2729–2740. 37 Matsumura H, Shiba T, Inoue T, Harada S & Kai Y (1998) A novel mode of target recognition suggested by the 2.0 A ˚ structure of holo S100B from bovine brain. Structure 6, 233–241. 38 Smith SP & Shaw GS (1998) A novel calcium-sensitive switch revealed by the structure of human S100B in the calcium-bound form. Structure 6, 211–222. 39 Miwa N, Kobayashi M, Takamatsu K & Kawamura S (1998) Purification and molecular cloning of a novel cal- cium-binding protein, p260lf, in the frog olfactory epi- thelium. Biochem Biophys Res Commun 251, 860–867. 40 Miwa N, Uebi T & Kawamura S (2000) Characterization of p26olf, a novel calcium-binding protein in the frog olfactory epithelium. J Biol Chem 275, 27245–27249. 41 Miwa N, Shinmoyo Y & Kawamura S (2001) Calcium- binding by p26olf, an S100-like protein in the frog olfactory epithelium. Eur J Biochem 268, 6029–6036. 42 Tanaka T, Miwa N, Kawamura S, Sohma H, Nitta K & Matsushima N (1999) Molecular modeling of single polypeptide chain of calcium-binding protein p26olf from dimeric S100Bbb. Protein Eng 12, 395–405. 43 Drohat AC, Tjandra N, Baldisseri DM & Weber DJ (1999) The use of dipolar couplings for determining the solution structure of rat apo-S100B(bb). Protein Sci 8, 800–809. 44 Kilby PM, Van Eldik LJ & Roberts GCK (1996) The solution structure of the bovine S100b protein dimer in the calcium-free state. Structure 4 , 1041–1052. 45 Malik S, Revington M, Smith SP & Shaw GS (2008) Analysis of the structure of human apo-S100B at low temperature indicates a unimodal conformational distribution is adopted by calcium-free S100 proteins. Proteins 73, 28–42. 46 Harder T & Gerke V (1993) The subcellular distribution of early endosomes is affected by the annexin II 2 p11 2 complex. J Cell Biol 123, 1119–1132. 47 Mayorga LS, Beron W, Sarrouf MN, Colombo MI, Creutz C & Stahl PD (1994) Calcium-dependent fusion among endosomes. J Biol Chem 269, 30927–30934. 48 Seemann J, Weber K & Gerke V (1997) Annexin I targets S100C to early endosomes. FEBS Lett 413, 185–190. 49 Johnsson N, Vandekerckhove J, Van Damme J & Weber K (1986) Binding sites for calcium, lipid and p11 on p36, the substrate of retroviral tyrosine-specific protein kinases. FEBS Lett 198, 361–364. 50 Johnsson N, Marriott G & Weber K (1988) p36, the major cytoplasmic substrate of src tyrosine protein kinase, binds to its p11 regulatory subunit via a short amino-terminal amphipathic helix. EMBO J 7, 2435– 2442. 51 Mailliard WS, Haigler HT & Schlaepfer DD (1996) Calcium-dependent binding of S100C to the N-terminal domain of annexin I. J Biol Chem 271, 719–725. 52 Seemann J, Weber K & Gerke V (1996) Structural requirements for annexin I–S100C complex-formation. Biochem J 319, 123–129. 53 Becker T, Weber K & Johnsson N (1990) Protein–pro- tein recognition via short amphiphilic helices; a muta- tional analysis of the binding site of annexin II for p11. EMBO J 9, 4207–4213. 54 Rintala-Dempsey AC, Santamaria-Kisiel L, Liao Y, Lajoie G & Shaw GS (2006) Insights into S100 target specificity examined by a new interaction between S100A11 and annexin A2. Biochemistry 45, 14695– 14705. 55 Lewit-Bentley A, Rety S, Sopkova-De OliveiraSantos J & Gerke V (2000) S100–Annexin complexes: some insights from structural studies. Cell Biol Int 24, 799– 802. 56 Gerke V & Moss SE (2002) Annexins: from structure to function. Physiol Rev 82, 331–371. 57 Semov A, Moreno MJ, Onichtchenko A, Abulrob A, Ball M, Ekiel I, Pietrzynski G, Stanimirovic D & Alak- hov V (2005) Metastasis-associated protein S100A4 induces angiogenesis through interaction with Annexin II and accelerated plasmin formation. J Biol Chem 280, 20833–20841. 58 Zeng FY, Gerke V & Gabius HJ (1993) Identification of annexin II, annexin VI and glyceraldehyde-3-phos- phate dehydrogenase as calcyclin-binding proteins in bovine heart. Int J Biochem 25, 1019–1027. 59 Filipek A, Wojda U & Lesniak W (1995) Interaction of calcyclin and its cyanogen bromide fragments with A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 4956–4966 ª 2008 The Authors Journal compilation ª 2008 FEBS 4965 [...].. .S100–annexin complexes – structural insights A C Rintala-Dempsey et al annexin II and glyceraldehyde 3-phosphate dehydrogenase Int J Biochem Cell Biol 27, 112 3–1 131 60 Tokumitsu H, Mizutani A, Minami H, Kobayashi R & Hidaka H (1992) A calcyclin-associated protein is a newly identified member of the Ca2+ ⁄ phospholipid-binding proteins, annexin family J Biol Chem 267, 891 9–8 924 61 Sudo... refinement Implications for structure, membrane binding and ion channel formation of the annexin family of proteins J Mol Biol 223, 68 3–7 04 64 McClintock KA & Shaw GS (2000) A logical sequence search for S100B target proteins Protein Sci 9, 204 3– 2046 FEBS Journal 275 (2008) 495 6–4 966 ª 2008 The Authors Journal compilation ª 2008 FEBS ... Characterization of the calcyclin (S100A6) binding site of annexin XI-A by sitedirected mutagenesis FEBS Lett 444, 1 1–1 4 62 Hatakeyama T, Okada M, Shimamoto S, Kubota Y & Kobayashi R (2004) Identification of intracellular target 4966 proteins of the calcium-signaling protein S100A12 Eur J Biochem 271, 376 5–3 775 63 Huber R, Berendes R, Burger A, Schneider M, Karshikov A, Luecke H, Romisch J & Paques E (1992) Crystal . for S100B [3 6–3 8], have become a S100–annexin complexes – structural insights A. C. Rintala-Dempsey et al. 4958 FEBS Journal 275 (2008) 495 6–4 966 ª 2008. orange spheres. A. C. Rintala-Dempsey et al. S100–annexin complexes – structural insights FEBS Journal 275 (2008) 495 6–4 966 ª 2008 The Authors Journal compilation