Crystal Structures of S100A6 in the Ca2+-Free and Ca2+-Bound States

Ludovic R. Otterbein, Jolanta Kordowska, Carlos Witte-Hoffmann, C.-L.Albert Wang, Roberto Dominguez
2002 Structure  
manner. Most S100 genes appear clustered within human chromosome 1q21 [11], a region that is often altered in transformed cells, consistent with a role in tumor progression [6]. The expression of the S100A6 gene, in particular, is increased in leukemia cells [12] and during the G1 phase of the cell cycle [13], implying a role in cell cycle progression (thus, S100A6 is also known as calcyclin). Experiments at the protein level show that S100A6 is overexpressed in numerous human cancer Summary
more » ... ls [14-16]. It has also been reported that S100A6 may be involved in reversing the inhibition exerted by S100A6 is a member of the S100 family of Ca 2؉ binding caldesmon on smooth muscle contraction [17]. Binding proteins, which have come to play an important role of S100A6 to the regulatory domain of annexin XI, a in the diagnosis of cancer due to their overexpression Ca 2ϩ -dependent phospholipid binding protein, has been in various tumor cells. We have determined the crystal connected with a potential role in cell proliferation and structures of human S100A6 in the Ca 2؉ -free and Ca 2؉differentiation [18]. bound states to resolutions of 1.15 Å and 1.44 Å , re-Three-dimensional structures of S100 proteins in the spectively. Ca 2؉ binding is responsible for a dramatic Ca 2ϩ -bound form have been determined by X-ray cryschange in the global shape and charge distribution tallography [7, 19-22] and NMR spectroscopy [23-25]. of the S100A6 dimer, leading to the exposure of two Recently, three structures of S100 proteins bound to symmetrically positioned target binding sites. The retarget peptides have been also reported: S100A10sults are consistent with S100A6, and most likely other annexin II [8], S100A11-annexin I [9], and S100B-p53 S100 proteins, functioning as Ca 2؉ sensors in a way regulatory peptide [26] . Contrary to what might be exanalogous to the prototypical sensors calmodulin and pected, the binding site and orientation for the two troponin C. The structures have important implications annexin peptides is different from that of the p53 pepfor our understanding of target binding and cooperatide. Thus, interpretation of the current structural data tivity of Ca 2؉ binding in the S100 family. would suggest that target binding by different S100 proteins involves different sites. Introduction Three NMR structures of S100 proteins, corresponding to the Ca 2ϩ -free (or apo) state, are also available In higher eukaryotes, an extensive repertoire of Ca 2ϩ [27] [28] [29] . However, there is disagreement concerning binding proteins translate transient increases in intracelthese Ca 2ϩ -free structures [24, 30, 31]. For example, the lular Ca 2ϩ levels into a multitude of cellular processes structures of Ca 2ϩ -free and Ca 2ϩ -bound rabbit S100A6 [1]. The vast majority of these proteins bind Ca 2ϩ through are very similar, leading the authors to question the role helix-loop-helix structures that usually exist in pairs and of S100 proteins as Ca 2ϩ sensors [23, 29]. In contrast, are referred to as EF-hand motifs [2, 3]. While several the structures of Ca 2ϩ -free and Ca 2ϩ -bound rat S100B, EF-hand proteins, including calmodulin and troponin C, which is similar in sequence to S100A6, reveal a major work as Ca 2ϩ sensors, undergoing structural changes conformational change of helix III upon Ca 2ϩ binding that expose target binding sites upon Ca 2ϩ binding, oth-[24, 28]. A higher resolution update of the structure of ers, such as calbindin D 9k , exhibit little structural change Ca 2ϩ -free rabbit S100A6 appears to confirm the exisand are thought to work as Ca 2ϩ buffers [4]. tence of significant differences between the structures Most S100 proteins form dimers in solution. Each subof S100A6 and S100B [31]. Groves et al. [30] have sugunit is a 10-12 kDa acidic molecule with two EF-hand gested that these discrepancies might arise from model-Ca 2ϩ binding domains. While the C-terminal EF-hand ing errors due to insufficient structural constraints. displays a canonical 12 amino acid, parvalbumin-like We describe here the X-ray crystal structures of E. Ca 2ϩ binding loop, the N-terminal EF-hand, which is coli-expressed human S100A6 in the Ca 2ϩ -free and usually referred to as the S100-hand, encompasses 14 Ca 2ϩ -bound states to resolutions of 1.15 Å and 1.44 Å , amino acids and has a lower affinity for Ca 2ϩ [5, 6]. respectively. In both states, S100A6 forms dimers stabi-Observed deviations from this common description are lized by hydrophobic interactions between helices I and the facts that (1) some S100 proteins have lost their IV of each monomer. Ca 2ϩ binding is responsible for an ability to bind either one or both of the Ca 2ϩ ions while ‫68ف‬Њ reorientation of helix III and marked changes in still adopting a Ca 2ϩ -bound-like conformation [7-9]; (2) the positioning and structure of helix II, the linker loop there is at least one monomeric member in the family, between helices II and III, and the C-terminal end of calbindin D 9k [10]; and (3) some S100 proteins form hethelix IV. The combination of these structural differences erodimers [5]. results in a dramatic change in the global shape and In contrast to calmodulin, which is ubiquitously excharge distribution of the S100A6 dimer, leading to the pressed, S100 proteins are expressed in a cell-specific Key words: S100A6; calcyclin; Ca 2ϩ binding; Ca 2ϩ sensor; X-ray; Ca 2ϩ -free structure; Ca 2ϩ -bound structure 1 Correspondence:
doi:10.1016/s0969-2126(02)00740-2 pmid:11937060 fatcat:obvzqng2ebcdvo3lxad4m5rxhy