C2-domains, Structure and Function of a Universal Ca2+-binding Domain*

A vast amount of protein sequence data accumulated over recent years has revealed that protein modules are widespread in nature. Many intracellular and extracellular proteins consist, in part or fully, of combinations of protein modules. C2-domains, together with SH2, PTB, PH, SH3, WW, and PDZ domains, are typical examples of intracellular protein modules. These modules form independently folding domains of 80–160 residues with characteristic binding properties; C2-domains bind Ca 21 and phospholipids, SH2 and PTB domains phosphotyrosine-containing sequences, PH domains phosphatidylinositol phosphates, SH3 and WW domains proline-rich sequences, and PDZ domains C-terminal sequences. C2-domains are unique among these modules because phospholipid binding to many C2-domains is regulated by Ca . For this reason, C2-domains are sometimes referred to as Ca -dependent lipid binding domains. However, C2-domains are not obligatory Ca and phospholipid-binding modules. C2-domains have diverged evolutionarily into Ca-dependent and Ca-independent forms that interact with multiple targets. Thus, although most C2-domains are probably Ca-binding domains, they represent a family of versatile protein modules with diverse functions. C2-domains comprise approximately 130 residues and were first identified in protein kinase C (1). Close to 100 C2-domain sequences are listed in the current data banks. Although reviews of several C2-domain proteins have been published (2–11), recent results on the structure and interactions of C2-domains by x-ray crystallography and NMR spectroscopy offer a new opportunity to rationalize the properties of C2-domains in structural terms. In this minireview, we will attempt to use this opportunity and correlate the functional properties of C2-domains with their structures.

A vast amount of protein sequence data accumulated over recent years has revealed that protein modules are widespread in nature. Many intracellular and extracellular proteins consist, in part or fully, of combinations of protein modules. C 2 -domains, together with SH2, PTB, PH, SH3, WW, and PDZ domains, are typical examples of intracellular protein modules. These modules form independently folding domains of 80 -160 residues with characteristic binding properties; C 2 -domains bind Ca 2ϩ and phospholipids, SH2 and PTB domains phosphotyrosine-containing sequences, PH domains phosphatidylinositol phosphates, SH3 and WW domains proline-rich sequences, and PDZ domains C-terminal sequences. C 2 -domains are unique among these modules because phospholipid binding to many C 2 -domains is regulated by Ca 2ϩ . For this reason, C 2 -domains are sometimes referred to as Ca 2ϩ -dependent lipid binding domains. However, C 2 -domains are not obligatory Ca 2ϩand phospholipid-binding modules. C 2 -domains have diverged evolutionarily into Ca 2ϩ -dependent and Ca 2ϩ -independent forms that interact with multiple targets. Thus, although most C 2 -domains are probably Ca 2ϩ -binding domains, they represent a family of versatile protein modules with diverse functions. C 2 -domains comprise approximately 130 residues and were first identified in protein kinase C (1). Close to 100 C 2 -domain sequences are listed in the current data banks. Although reviews of several C 2 -domain proteins have been published (2)(3)(4)(5)(6)(7)(8)(9)(10)(11), recent results on the structure and interactions of C 2 -domains by x-ray crystallography and NMR spectroscopy offer a new opportunity to rationalize the properties of C 2 -domains in structural terms. In this minireview, we will attempt to use this opportunity and correlate the functional properties of C 2 -domains with their structures.

Proteins Containing C 2 -domains
Most proteins with C 2 -domains function in signal transduction or membrane traffic. The first category includes proteins involved in the generation of lipid second messengers (e.g. cPLA 2 1 (12), PLCs (13), and phosphatidylinositol 3-kinases (14)), in protein phosphorylation (e.g. PKC (15,16)), in activation of GTPases (e.g. Ras-GAP (17)), and in ubiquitin ligation (e.g. Nedd4 (18)). The second category contains synaptotagmins (19,20), rabphilin-3 (21), RIM (22), and Munc13 (2). In addition to many well characterized proteins, several open reading frames with C 2 -domains are reported in Gen-Bank TM . For example, open reading frames in Caenorhabditis elegans and yeast encode transmembrane proteins with three or four C 2 -domains whose biological roles have not been identified. This suggests that additional interesting functions for C 2 -domain proteins remain to be discovered.
Much of the current data on the structures and interactions of C 2 -domains were derived from studies of PLC␦1, PKC, cPLA 2 , and, in particular, synaptotagmin I. PLC␦1 hydrolyzes phosphatidylinositol 4,5-bisphosphate to generate the second messengers diacylglycerol and inositol-1,4,5-trisphosphate (5). PKCs represent a family of protein kinases that are regulated by diacylglycerol and other lipids (1,4). In addition, activation of the classical isoforms of PKC (PKC␣, -␤, and -␥) depends on Ca 2ϩ . cPLA 2 hydrolyzes glycerophospholipids to produce arachidonic acid, a precursor of prostaglandins and leukotrienes, which are involved in inflammation (7). In PLC␦1, PKC, and cPLA 2 , the C 2 -domain is believed to play a regulatory role by mediating the Ca 2ϩ -dependent recruitment of these enzymes to phospholipid membranes. In contrast, synaptotagmin I functions in membrane traffic. Synaptotagmin I belongs to a family of at least 12 transmembrane proteins containing two C 2 -domains, the C 2 A-and C 2 B-domains. The C 2 -domains occupy most of the cytoplasmic region of the synaptotagmins and probably act as Ca 2ϩ effector domains. Synaptotagmin I is localized to synaptic vesicles where it is essential for the fast, Ca 2ϩ -dependent component of neurotransmitter release (23,24). Synaptotagmin I is believed to function as the main Ca 2ϩ sensor in synaptic vesicle exocytosis by a mechanism involving Ca 2ϩ binding to both C 2domains (see below). The functions of the other synaptotagmins and membrane trafficking proteins with C 2 -domains are less well characterized but may be similar.

Diverse Ca 2؉ -dependent Properties of C 2 -domains
The notion that C 2 -domains act as Ca 2ϩ -binding motifs arose from the observation that the classical isoforms of PKC, which contain a C 2 -domain (PKC␣, -␤, and -␥), were regulated by Ca 2ϩ . In contrast, isoforms apparently lacking a C 2 -domain (PKC␦, -⑀, -, and -) were Ca 2ϩ -independent (1). Activation of classical PKCs by Ca 2ϩ involves the translocation of PKCs to the membrane by Ca 2ϩdependent phospholipid binding (4,25). These observations led to the hypothesis that C 2 -domains may mediate Ca 2ϩ -dependent phospholipid binding.
This notion was first actually demonstrated for the C 2 A-domain of synaptotagmin I. The isolated C 2 A-domain was shown to represent an autonomously folding module that binds phospholipids in a Ca 2ϩ -dependent manner (26). The C 2 A-domain bound all negatively charged phospholipids independent of headgroup structure. Mg 2ϩ , Ba 2ϩ , and Sr 2ϩ were unable to stimulate phospholipid binding. Ca 2ϩ acted cooperatively with a Hill coefficient of 3 and an apparent affinity in the low micromolar range. Further studies revealed that the C 2 A-domain binds Ca 2ϩ directly without phospholipids but with a lower apparent Ca 2ϩ affinity (Ϸ0.2 mM compared with Ϸ5 M free Ca 2ϩ ) (27). The C 2 A-domain of synaptotagmin I also bound syntaxin 1 as a function of Ca 2ϩ , with a low apparent Ca 2ϩ affinity resembling that of intrinsic Ca 2ϩ binding (28). These data suggested that the functions of C 2 -domains may be diverse and include Ca 2ϩ -dependent interactions with proteins in addition to phospholipids. Analysis of the C 2 A-domains from other synaptotagmins revealed that those from synaptotagmins II, III, V, and VII bind phospholipids and syntaxin 1, but the C 2 A-domains of synaptotagmins IV, VI, X, and XI do not (28). Although the C 2 Adomains of synaptotagmins I, II, III, V, and VII bind to phospholipids with similar Ca 2ϩ affinities, they exhibit distinct cation specificities (29). All of the C 2 A-domains from synaptotagmins that bind to phospholipids also interact with syntaxin 1 as a function of Ca 2ϩ but with different Ca 2ϩ dependences: synaptotagmins I, II, and V require Ca 2ϩ concentrations of Ͼ0.2 mM whereas synaptotagmins III and VII bind at Ͻ1 M Ca 2ϩ (28).
Sequence analyses of the C 2 A-and C 2 B-domain of synaptotagmin I revealed that they contain evolutionarily conserved differences. The C 2 B-domain does not exhibit the same Ca 2ϩ -dependent phospholipid binding properties as the C 2 A-domain. Nevertheless, most C 2 B-domains probably bind Ca 2ϩ because they contain the requisite Ca 2ϩ -binding sequences defined in C 2 A-domains (see below), and because the C 2 B-domains of synaptotagmins I and II mediate the Ca 2ϩ -dependent self-association of synaptotagmins (30). These results led to a model whereby the C 2 A-and C 2 Bdomains of most synaptotagmins are Ca 2ϩ -binding domains that are specialized for different Ca 2ϩ -dependent activities.
Similar to the C 2 A-domains of synaptotagmins, the C 2 -domains from cPLA 2 (31,32), PKC␤ (27), and Nedd4 (18) bind phospholipids at micromolar Ca 2ϩ concentrations. However, the C 2 A-domains of synaptotagmins and the C 2 -domain of PKC␤ preferentially bind to negatively charged phospholipids whereas the C 2 -domain from cPLA 2 interacts with neutral phospholipids (26,33). Furthermore, although the C 2 -domain from PKC␤ is similar to the synaptotagmin C 2 A-domains, it does not bind to syntaxin 1 as a function of Ca 2ϩ . Therefore, even among C 2 -domains that share Ca 2ϩ -dependent phospholipid binding, there are functional distinctions that may be important for their biological roles.
To complicate matters, some C 2 -domains that are Ca 2ϩ -regulated simultaneously bind other molecules in a Ca 2ϩ -independent manner. For example, the C 2 B-domain of synaptotagmin I interacts with AP-2 (34), inositol polyphosphates (35), ␤-SNAP (36), and Ca 2ϩ channels (37). Finally, many C 2 -domains may not bind Ca 2ϩ at all. Several synaptotagmins appear to be unable to bind Ca 2ϩ , as may be the case with the C 2 -domains of RIM. Interestingly, PKC isoforms that initially were not thought to have a C 2 -domain and are Ca 2ϩ -independent (PKC␦, -⑀, -, and -) actually contain a C 2 -domain that is located at the N terminus and probably does not bind Ca 2ϩ (2,11,38). Thus, as a group C 2 -domains perform multiple biological functions.

Three-dimensional Structures of C 2 -domains
X-ray diffraction analysis of the synaptotagmin I C 2 A-domain yielded the first structure of a C 2 -domain (39). The structure consists of a compact ␤-sandwich composed of two four-stranded ␤-sheets (Fig. 1A). Three loops at the top of the domain and four at the bottom connect the eight ␤-strands. Ca 2ϩ binding occurs exclusively at the top three loops (see below). NMR spectroscopy showed that the solution structure of the C 2 A-domain is identical to the crystal structure (27). 2 Determination of the structures of three other C 2 -domains (from PKC␤, cPLA 2 , and PLC␦1) revealed similar designs and interesting differences. PLC␦1 is a modular protein composed of PH-, EF-hand, C 2 -, and catalytic domains. X-ray diffraction studies of crystals from PLC␦1 lacking the N-terminal PH domain provided the threedimensional structure of a C 2 -domain in the context of a nearly full-length protein (40,41). The three-dimensional structure of the PLC␦1 C 2 -domain ( Fig. 1B) is very similar to that of the synaptotagmin I C 2 A-domain, with a root mean square deviation of 1.4 Å for 109 equivalent ␣-carbons. The topology of the ␤-strands, however, is strikingly different (40). The arrangement of ␤-strands in the PLC␦1 C 2 -domain constitutes a circular permutation of the topology observed in the C 2 A-domain of synaptotagmin I (Fig. 1C). As a result, strand 1 of the synaptotagmin I C 2 A-domain occupies the same position as strand 8 of the PLC␦1 C 2 -domain. The N and C termini are at the top of the C 2 -domain in synaptotagmin I but at the bottom in PLC␦1 (Fig. 1). The two types of topology are referred to as topology I (synaptotagmin I C 2 A-domain) or topology II (PLC␦1 C 2 -domain). The crystal structures of the C 2 -domains from PKC␤ 3 and cPLA 2 (42) are also similar to those of synaptotagmin I and PLC␦1 and exhibit topologies I and II, respectively. It is unclear why C 2 -domains occur in two topologies. One reason may be that the topology influences the relative orientation of a C 2 -domain with respect to its neighboring domains.
There is a high degree of structural homology between C 2domains in the core ␤-sandwich and less similarity in the top and bottom loops (Fig. 1). Accordingly, the C 2 -domain sequences involved in the core ␤-sandwich are highly conserved between C 2 -domains whereas the sequences of the loops, particularly loop 1 and the three bottom loops, are not. The high degree of structural identity between the core ␤-sandwiches of C 2 -domains suggests that the ␤-sandwich represents a scaffold. This scaffold allows the emergence of variable loops at the top and bottom of the domain. As discussed below, the loops are involved in Ca 2ϩ binding and may determine the functional specificity of a C 2domain. The C 2 -domain structures provide a framework to interpret the properties of C 2 -domains and at the same time allow us to predict the minimum sequences required for a complete, well folded ␤-sandwich. Thus results from experiments performed with incomplete C 2 -domain fragments or with mutants containing deletions in a ␤-strand should be interpreted with caution since misfolding is likely.

How Do C 2 -domains Bind Ca 2؉ ?
The Ca 2ϩ binding modes of the C 2 -domains from synaptotagmin I, PKC␤, PLC-␦1, and cPLA 2 were analyzed by x-ray diffraction and NMR spectroscopy (27, 39 -44). 2,3 In all C 2 -domains, multiple Ca 2ϩ ions bind in a cluster exclusively at the top loops (Fig. 1). These loops are widely separated in the primary sequences (Fig. 2). The Ca 2ϩ -binding sites are formed primarily by aspartate side chains that serve as bidentate ligands for two or three Ca 2ϩ ions.
In the C 2 A-domain of synaptotagmin I, loops 1 and 3 contain three Ca 2ϩ -binding sites (Ca1, Ca2, and Ca3 in Fig. 3). The Ca 2ϩbinding sites are formed by five aspartate side chains, one serine side chain, and three carbonyl groups (27,39,44) 2 (Figs. 2 and 3). The presence of three Ca 2ϩ -binding sites in the C 2 A-domain correlates well with the Hill coefficient of 3 observed in Ca 2ϩ -dependent phospholipid binding experiments (26). Ca 2ϩ binding to all three sites is necessary for syntaxin 1 and phospholipid binding (44). 4 The coordination spheres of the bound Ca 2ϩ ions in the C 2 A-domain are incomplete, especially for Ca3. This results in the low apparent intrinsic Ca 2ϩ affinity of this site (Ͼ1.0 mM). When phospholipids A and B, the locations of the N and C termini and of the Ca 2ϩ -binding loops are indicated. Each C 2domain is shown complexed to three Ca 2ϩ ions (orange) (27,40,41,44). 2,3 The diagrams were prepared with the program MOLSCRIPT (51). In C, ␤-strands in the C 2 -domains from synaptotagmin I and PKC␤ (left) and from PLC␦1 and cPLA 2 (right) are numbered in the order of the primary sequences. The three Ca 2ϩ -binding loops at the top of the C 2 -domains are indicated. bind, they probably fill unsatisfied coordination sites on the bound Ca 2ϩ ions, resulting in a Ϸ1000-fold increase in the apparent affinity of the C 2 A-domain for Ca 2ϩ . The Ca 2ϩ binding mode of the C 2 -domain of PKC␤, as determined by x-ray crystallography, is very similar to that of the C 2 A-domain. 3 The C 2 -domain of PLC␦1 shares two of the Ca 2ϩ -binding sites of the synaptotagmin I C 2 A-domain (Ca1 and Ca2) but contains a distinct third Ca 2ϩ -binding site (Ca4) (41, 43) (Fig. 3). Site Ca4 involves one aspartate, one asparagine, and one serine side chain in addition to one backbone carbonyl group. Sites Ca1 and Ca4 were occupied in all complexes of PLC␦1 with Ca 2ϩ and Ca 2ϩ analogs (La 3ϩ , Sm 3ϩ , and Ba 2ϩ ). Site Ca2 was only filled in the La 3ϩ complex, but it seems likely that Ca 2ϩ also binds to this site at Ca 2ϩ concentrations above 1 mM or at lower Ca 2ϩ concentrations in the presence of phospholipids (43). In addition, all side chains from site Ca3 in the synaptotagmin I C 2 A-domain are conserved in the PLC␦1 C 2 -domain and have similar orientations in the structures of both C 2 -domains (44). 2 This strongly suggests that Ca 2ϩ may also occupy this site and that the PLC␦1 C 2 -domain may contain a total of four Ca 2ϩ -binding sites. Bound Ca 2ϩ ions in the C 2 -domain of PLC␦1 have unsatisfied coordination sites, suggest-ing that in the absence of phospholipids they may also exhibit low apparent affinities. The C 2 -domain of cPLA 2 has a Ca 2ϩ binding mode similar to that of the PLC␦1 C 2 -domain but apparently with only sites Ca1 and Ca4 occupied (42).

FIG. 1. Ribbon diagrams of the structures of the C 2 A-domain of synaptotagmin I (A) and the C 2 -domain of PLC␦1 (B) and schematic drawing of their ␤-strand topologies (C). In
The Ca 2ϩ binding modes summarized above can be used to anticipate the Ca 2ϩ binding properties of other C 2 -domains. The aspartate residues involved in Ca 2ϩ binding in the synaptotagmin I C 2 A-domain are conserved in many C 2 -domains. Based on this observation, we proposed that the motif formed by these aspartate residues is widespread and named it the C 2 -motif (27). Sites Ca1 and Ca2 are probably the most common Ca 2ϩ -binding sites in C 2 -domains, and additional Ca 2ϩ -binding sites are likely to exist in many C 2 -domains depending on the side chains present in loops 1-3. Ca 2ϩ -dependent C 2 -domains thus appear to have been designed to concentrate multiple Ca 2ϩ ions in a small region. The Ca 2ϩ ions contain unsatisfied coordination sites that remain available for interaction with target molecules.

Mechanisms of C 2 -domain Function
The three-dimensional structures of C 2 -domains determined so far show no evidence that Ca 2ϩ induces a substantial change from one well defined conformation to another well defined conformation. Comparison of the NMR solution structure of the Ca 2ϩ -bound form of the synaptotagmin I C 2 A-domain with the crystal structure of the Ca 2ϩ -free form demonstrated that Ca 2ϩ binding involves rotations of some side chains but causes no substantial backbone rearrangements (27). 2 The NMR data indicate that the Ca 2ϩ -binding region is flexible in the absence of Ca 2ϩ and is stabilized after Ca 2ϩ binding. Structural stabilization by Ca 2ϩ binding is consistent with decreased B-factors in the crystal structure of the C 2 A-domain after partial Ca 2ϩ saturation (39) and with the observations that Ca 2ϩ causes a large change in denaturation temperature (27) and increases the resistance of the synaptotagmin I C 2 A-domain against proteolysis (45). With regard to the PLC␦1 C 2 -domain, 10 of 12 x-ray structures obtained in the presence or absence of Ca 2ϩ or Ca 2ϩ analogs are very similar, suggesting that Ca 2ϩ binding does not cause conformational changes (40,41,43).
If Ca 2ϩ does not induce a major conformational change in C 2 -domains, how does Ca 2ϩ regulate their function? The structural stabilization induced by Ca 2ϩ probably does not account for Ca 2ϩ regulation because the conformations compatible with binding to target molecules are also available in the absence of Ca 2ϩ . However, Ca 2ϩ binding causes a major change in the electrostatic potential of the synaptotagmin I C 2 A-domain that may be important for regulating interactions. Analysis by NMR spectroscopy showed that the region around the Ca 2ϩ -binding sites of the C 2 A-domain is responsible for Ca 2ϩ -dependent binding to syntaxin (46). This region contains the cluster of aspartate residues that coordinate Ca 2ϩ and a ring of basic amino acids surrounding it. Binding to syntaxin, a negatively charged protein, could therefore be driven by the change in electrostatic potential caused by Ca 2ϩ binding and could be mediated by the basic side chains surrounding the Ca 2ϩ -binding site. Two exposed hydrophobic side chains in the region and coordination of the unsatisfied Ca 2ϩ valences by acidic residues of syntaxin may contribute to the interaction (46).
The mode of interaction between synaptotagmin I and syntaxin 1 suggested that synaptotagmin I acts as an electrostatic switch in neurotransmitter release. The binding of phospholipids by the C 2 Adomain is also best explained by this model. Mutations in basic and FIG. 2. Sequences of the C 2 -domains from synaptotagmin I (S), PLC␦1 (C), and cPLA 2 (A). Identical residues are shown on a yellow background. Aspartate, asparagine, and serine residues that coordinate the Ca 2ϩ ions in the different C 2 -domains are shown on a pink background. Residues whose backbone carbonyl groups coordinate Ca 2ϩ ions are shown on a blue background. Sequences that are structurally almost identical in the three-dimensional structures are indicated by a green bar below the alignment, and locations of ␤-strands by an orange bar. The positions of the three loops involved in Ca 2ϩ binding are shown above the alignment. acidic amino acids that disrupt syntaxin binding (44,46) also inhibit phospholipid binding. 4 Phospholipid binding correlates with the density of negative charges on the surface of the phospholipid bilayer rather than with a specific chemical structure. Furthermore, binding is inhibited by high salt. These results support the importance of electrostatic interactions for the Ca 2ϩ -dependent binding of the synaptotagmin I C 2 A-domain to phospholipids (26). Other C 2 -domains that bind to negatively charged phospholipids in a Ca 2ϩ -dependent manner such as those of classical PKCs may share this mechanism of binding. Substitutions in two of the aspartate residues that bind Ca 2ϩ in PKC␤ have shown, however, that lipid binding is probably not purely electrostatic (47). It is likely that coordination of the Ca 2ϩ ions bound to C 2 -domains by the phosphate groups of the lipids may provide a major contribution to the binding energy, which is supported by the observation that the apparent Ca 2ϩ affinities are much higher in the presence of phospholipids than in their absence. Insertion of highly exposed hydrophobic side chains into the lipid bilayer may contribute to binding as proposed for the Ca 2ϩ -dependent binding of phosphatidylcholine to the C 2 -domain of cPLA 2 (33). The side chains in the Ca 2ϩ -binding loops are likely to influence the preference for types of lipids. Thus, the preference of cPLA 2 for neutral rather than negatively charged phospholipids may be because of the presence of two acidic residues in loops 1 and 3, in addition to the Ca 2ϩ -binding residues, and to the absence of basic residues.

Evolution of C 2 -domains: Example of Synaptotagmins
The differences between C 2 -domains in synaptotagmins may give clues about how this domain adapted to diverse functions. In evolution, the C 2 -domains of the more than 12 different synaptotagmins developed distinct Ca 2ϩ affinities, or in some cases, Ca 2ϩ independence (28,29). Interestingly synaptotagmins IV and XI have a single, evolutionarily conserved amino acid change in the Ca 2ϩ -binding residues of the C 2 A-motif. These C 2 A-domains are unable to bind phospholipids as a function of Ca 2ϩ (48). Reversal of this amino acid change restored the ability of synaptotagmins IV and XI to bind phospholipids as a function of Ca 2ϩ . Thus all other structural requirements for Ca 2ϩ -dependent phospholipid binding were evolutionarily retained in these synaptotagmins, and a single amino acid substitution was selected to abolish Ca 2ϩ -dependent phospholipid binding. This finding supports the notion that at least the C 2 A-domain in synaptotagmins performs other functions in addition to Ca 2ϩ -dependent phospholipid binding.

Conclusion
C 2 -domains are remarkable modules present in a wide variety of proteins that can participate in different types of interactions. Two widespread Ca 2ϩ -binding motifs defined by structural characteristics are known: EF-hands as the most widely distributed motifs with little structural autonomy; and C 2 -domains representing autonomous modules present in probably more than 100 proteins. Although more EF-hands than C 2 -domains have been described, the growing number of C 2 -domains in the data banks suggests that C 2 -domains are universal Ca 2ϩ -binding domains. The Ca 2ϩ -binding sites formed by C 2 -motifs and EF-hands have different architectures and function by distinct mechanisms. The EF-hand is formed by a contiguous helix-turn-helix sequence that binds a single Ca 2ϩ ion and usually is a substructure in an ␣-helical protein domain (49,50). Multiple EF-hands may be present in a protein. Ca 2ϩ binding to EF-hands in contiguous domains often occurs in a concerted manner, causing conformational changes that expose hydrophobic surfaces. In contrast, C 2 -domains are autonomously folding modules with a stable ␤-sheet scaffold. Multiple Ca 2ϩ ions bind in a cluster at the tip of the domain in a region formed by loops that are distant in the sequence. The Ca 2ϩ binding properties of C 2 -domains confer onto them the ability to act as electrostatic switches without requiring large conformational changes. The Ca 2ϩ binding mode of C 2 -domains may be particularly useful for fast Ca 2ϩ -triggered reactions, such as neurotransmitter release. We expect that the number of C 2 -domains and the variety of their interactions will continue to grow, with new developments and surprises for years to come.