Invited Mini Review
Ca2+-regulated ion channels Daniel H. Cox* Department of Neuroscience, Tufts University School of Medicine, Boston MA, 02420, USA
Due to its high external and low internal concentration the Ca2+ ion is used ubiquitously as an intracellular signaling mol2+ ecule, and a great many Ca -sensing proteins have evolved to 2+ receive and propagate Ca signals. Among them are ion channel proteins, whose Ca2+ sensitivity allows internal Ca2+ to influence the electrical activity of cell membranes and to 2+ feedback-inhibit further Ca entry into the cytoplasm. In this review I will describe what is understood about the Ca2+ sensing mechanisms of the three best studied classes of Ca2+-sensi2+ + tive ion channels: Large-conductance Ca -activated K chan2+ + nels, small-conductance Ca -activated K channels, and voltage-gated Ca2+ channels. Great strides in mechanistic understanding have be made for each of these channel types in just the past few years. [BMB reports 2011; 44(10): 635-646]
2+ two, and I will review what is understood about the Ca -sens2+ + ing mechanisms of the Ca -activated K channels and the 2+ channels here. For recent reviews of voltage-gated Ca 2+ Ca -activated Cl channels and TRP channels see (4-6).
Ca2+-ACTIVATED POTASSIUM CHANNELS Ca2+-activated K+ channels are classified into two types, the large conductance, or BKCa channels, and the smaller conductance or SKCa channels. Both are homotetramers and both contain an integral membrane domain much like that found in purely voltage-gated K+ channels. The BKCa and SKCa channels, however, show very little sequence homology, and they have 2+ evolved unique intracellular sequences to subserve Ca sensing.
LARGE CONDUCTANCE Ca2+-ACTIVATED K+ CHANNELS
The calcium ion might reasonably be considered the preeminent intracellular signaling molecule, as it is involved in a myriad of physiological processes muscle contraction, neurotransmission, the regulation of gene expression, mitosis, and 2+ fertilization, to name a few (1, 2) and hundreds of Ca sensing proteins have evolved to act as effectors in calciumdependent signaling pathways (3). Unique among these are 2+ the Ca - sensitive ion channels, whose gating is modulated 2+ by Ca binding. These channels allow for crosstalk between chemical and electrical systems and in particular for the feed2+ entry through voltage-gated Ca2+ back control of Ca channels. They are best known for enabling changes in intra2+ cellular Ca concentration to influence neuronal firing patterns and the strength of muscle contraction. They fall into four 2+ + main categories, Ca -activated K channels, voltage-gated 2+ 2+ 2+ Ca channels, Ca -activated Cl channels, and Ca -regulated TRP channels. Of these, the mechanisms governing Ca2+-dependent regulation are best understood for the first
BKCa channels are composed of four identical pore-forming α subunits, and in some tissues auxiliary beta subunits (7). The α subunits alone are sufficient to form a functional channel (8), while the beta subunits play a modulatory role (9-11). A single gene encodes for the BKCa α subunit (KCNMA1), and its gene product is referred to as Slo1 (8, 12, 13) (Fig. 1A). Slo1 has an in+ tegral-membrane domain much like a voltage-gated K channel (Kv) although the homology between Kv channels and BKCa channels is low and it has an extra transmembrane domain (S0) that places its N terminus outside the cell (14). It also has a very large intracellular C-terminus (-800 amino acids) that is unique to this channel and appears to have evolved to confer 2+ + Ca sensitivity on an otherwise voltage-gated K channel 2+ structure. Indeed, BKCa channels are both Ca and voltage activated, and there is a synergy between these stimuli that is seen as a leftward shifting of the channel’s conductance-voltage relation 2+ (G-V), as the internal Ca concentration is raised (15, 16) (Fig. 1B). This leftward shifting starts at 100 nm Ca2+ and continues at concentrations over 1 mM, traveling over 200 mV. Thus, the BKCa channel has a very large dynamic range in terms of its ability to sense Ca2+. Surprisingly, however, Slo1’s amino-acid se2+ quence contains no canonical Ca binding motifs (EF hand or C2 domains), so upon its cloning, the nature of the BKCa chan2+ nel’s Ca binding sites was immediately a central issue. Over the last decade site-directed-mutagenesis experiments 2+ have lead to the following picture of the BKCa channel’s Ca binding sites. The Slo1 subunit appears to contain three types
*Corresponding author. Tel: +1-617-636-0305; Fax: +1-617-6362413; E-mail: [email protected]
http://dx.doi.org/10.5483/BMBRep.2011.44.10.635 Received 20 September 2011 Keywords: BK channel, Calcium, Calcium channel, Calmodulin, SK channel http://bmbreports.org
Ca2+-regulated ion channels Daniel H. Cox
Fig. 1. (A) Membrane topology of the α subunit of the BKCa channel. (B) Conductance-voltage relations for the BKCa channel determined at a series of Ca2+ concentrations as indicated. (C) Overall topology of the BKCa channel showing its integral membrane portion and its gating ring. (D) Crystal structure of the BKCa channel’s gating ring from Wu et al. 2010 (26).
of Ca2+ binding sites, one of low affinity (17-19) (mM) that lies in a region of the channel’s large C-terminus proximal to the membrane known as the RCK1 domain, and two of higher affinity, one that also lies within the RCK1 domain and another that lies further downstream in a region of the channel known 2+ as the Ca bowl (20). The low affinity site is disabled by mutations at positions D374 or E399 (17, 19). It is responsible for the leftward shifting of the channel’s G-V relation at very high 2+ Ca concentrations, over 100 µM, and it is also the site of the modulatory action of Mg2+ (17-19). Indeed, in vivo this site it 2+ 2+ is likely to be occupied by Mg rather than Ca . The high-affinity RCK1 site can be disabled by mutations at any of three positions M513 (21), D367 (19), or E535 (22), and these muta2+ tions reduce the ability of Ca to shift the BKCa channel’s G-V relation leftward by about half. Similarly, the Ca2+-bowl site can be disabled by mutations at either of two positions, D898
636 BMB reports
2+ or D900, and disabling the Ca bowl also reduces the left2+ ward-shifting effect of Ca by about half (20, 21, 23, 24). When mutations are made at all three sites together, the channel’s Ca2+ sensitivity is completely eliminated (19). 2+ Recently the study of the Ca sensing mechanism of the BKCa channel was greatly advanced by the publication of a cryo-electron-microscopy structure of the full channel (25) and two crystal structures of the C-terminal domain of the channel (26, 27). All three structures indicated that each Slo1 subunit contains two structurally homologous domains RCK1 and RCK2 that form an interlocking dimer-like structure (Fig. 2A), and four of these dimer-like structures come together to form a ring known as the channel’s gating ring (Fig. 1D). In the intact channel this gating ring hangs below the pore (25) (Fig. 2+ 1C), and Ca -induced changes in its structure are thought to induce channel opening. Indeed, the BKCa channel’s “gating
Ca2+-regulated ion channels Daniel H. Cox
Fig. 2. (A) The BKCa channel’s RCK1-RCK2 pseudo dimer in the presence of 50 mM Ca . The crystal structure is from Yaun et al. 2010 (27). Four such pseudo dimers form the channel’s gating ring. The Ca2+ bowl is indicated in green, the Ca2+ ion in yellow. (B) The BKCa channel’s 2+ RCK1-RCK2 pseudo dimer in the absence of Ca from the crystal structure of Wu et al. 2010 (26). The residues thought to be essential for 2+ 2+ the high-affinity RCK1 Ca binding site are highlighted in the inset. (C) Structure of the Ca bowl from the crystal of Yaun et al. 2010 (27). (D) Results of the site-directed mutagenesis studies of Bao et al. 2004 (23). Bars indicate the leftward shift of the channel’s G-V curve in response 2+ 2+ to 10 µM Ca . Circled are at positions at which mutations reduced the channel’s Ca response. These residues are highlighted with the same color in (C).
ring” is structurally similar to that of the prokaryotic Ca2+-acti+ vated K channel MthK, whose structure was determined in 2002 (25, 28, 29). When Ca2+ binds to the MthK channel’s gating ring, it expands (28, 30), and this expansion is thought to pull open the pore via linkers to the pore helices. Perhaps this is what occurs in the BKCa channel as well. The MthK channel, however, has a thousand-fold lower affinity for Ca2+ 2+ than does the BKCa channel (31, 32), and its Ca binding sites are also very different (33). So although the two channels have similar gating rings, the mechanisms by which Ca2+ binding alters the structures of these rings must be different. Fortunately, one of the BKCa crystal structures, that of Yuan et al. (27), came from a crystal grown in the presence of 50
mM Ca2+, and a bound Ca2+ ion was evident in the crystal (Fig. 2A). Indeed, consistent with the electrophysiological data, 2+ 2+ the Ca was found to lie in a loop formed by the Ca bowl, and as predicted by Bao et al. (23), it was coordinated by the side-chain oxygens of D898 and D900, and as well by two backbone carbonyl oxygens. Thus the structural and functional 2+ data related to the Ca bowl agree quite well (Fig. 2C, D). 2+ Surprisingly, however, the Ca -saturated crystal did not 2+ contain a second bound Ca , nor a third. The high-affinity RCK1 site and the low-affinity site were not evident. In fact, in the case of the high-affinity site the residues implicated as being essential to this site, M513, D367, E535, while in the same general vicinity, did not appear to be close enough together to
Ca2+-regulated ion channels Daniel H. Cox
form the second site. Thus, it appeared that perhaps there was not a second high-affinity binding site. But if this were the case, it would be very hard to explain the mutagenesis data. Shedding light on this issue, however, was the second crystal structure, that of Wu et al. (26), which came from a crystal 2+ grown in the absence of Ca . In this structure the essential high-affinity, RCK1-site residues as determined from site-directed mutagenesis were in sufficiently close proximity to 2+ 2+ perhaps be forming a Ca binding site, although no Ca was present, because none was included in the crystal (Fig. 2B). This result, although not conclusive, taken together with the mutagenesis work, supports the hypothesis that M513, D367, 2+ E535 are indeed involved in forming a high-affinity Ca bind2+ ing site in the native channel, and that the Ca -saturated crystal did not have a bound Ca2+ in this area, nor in fact in the low affinity site, because it was to some degree distorted. Perhaps the loss of the channel’s integral membrane domain 2+ alters the structure. What residues actually coordinate Ca at this second site, however, remains to be established. In particular it would be very unusual for the side chain of M513, which contains a sulfur rather that an oxygen atom, to be in2+ volved in Ca binding, so perhaps M513 makes contacts critical for the structural integrity of the protein in this area but does not directly coordinate Ca2+ (22). 2+ In order for Ca binding to activate the BKCa channel, or 2+ any channel, Ca must bind more tightly to the open channel than to the closed. Thus, each Ca2+ binding site must have two dissociation constants, one for the open state (KO) and one for the closed (KC). Indeed, it is the ratio of these dissociation constants (KC/KO) that determines the effect of Ca2+ binding on the channel’s open probability. Thus, in order to quantitatively 2+ understand the influence of Ca binding on channel gating, these dissociation constants must be determined. To address this issue Sweet and Cox (2008) (34) did and extensive analy2+ sis of the Ca dose-response curves of wildtype and mutant 2+ BKCa channels under conditions where their Ca -binding con2+ stants could be well determined. They found that the Ca 2+ bowl has an affinity for Ca of 3.1 µM, when it is closed, and 0.9 µM, when it is open. And the high-affinity RCK1-site has an affinity of 23 µM, when it is closed, and 4.9 µM, when it is 2+ 2+ open. Thus, the Ca bowl binds Ca more tightly than does the RCK1 site in both the open and closed conformations. 2+ However, when a Ca ion binds, the RCK1 site has more influence on channel opening than does the Ca2+-bowl site. This is because its ratio KC/KO is larger. Indeed KC/KO for the RCK1 2+ site is 4.7, which indicates that a Ca ion binding at this site decreases the energy difference between open and closed by 3.8 KJ/mol, while for the Ca2+ bowl site KC/KO is 3.5, which corresponds to an energy difference of 3.1 KJ/mol. These values were determined at a membrane voltage of -80 mV; however, when Sweet and Cox repeated their experiments at 0 mV, they found that the depolarization increased the affinity of the RCK1 site in both the closed and open conformations, but 2+ it had no effect on the Ca -bowl site (34). They proposed that 638 BMB reports
this arises because the RCK1 site interacts with the channel’s voltage sensor and is influenced by its movement, while the 2+ Ca -bowl site does not. The affinities of the channel’s low af2+ finity Ca binding site have been less well determined, however, they have been estimated to be 2-3 mM, when the channel is closed, and 0.6-0.9 mM when the channel is open (35). 2+ In general, then the current view of the BKCa channel’s Ca sensing mechanism is as follows. Under physiological conditions Ca2+ influences channel opening by binding to two 2+ types of micromolar-affinity Ca binding sites on the channel’s gating ring, eight sites in all. This likely induces an expansion of the channel’s gating ring, which then pulls open 2+ the pore. In the absence of bound Ca (and voltage-sensor activation) the energy difference between open and closed is 32 KJ/mol which corresponds to an open probability of 2 6 2+ × 10 (almost always closed). When Ca binds to all sites, it reduces the energy difference between open and closed by 60 KJ/mol to -28 KJ/mol. This corresponds to an open probability of 0.99 (almost always open). Depending on the membrane potential, active voltage sensors may also help open the channel, such that as voltage sensors become active, fewer bound 2+ Ca are needed to achieve the same open probability. This leads to the synergy observed in Fig. 1B. 2+
SMALL CONDUCTANCE Ca -ACTIVATED K CHANNELS Based on sequence homology with other K+ channels in their pore regions, the Adelman group cloned three SK channel subunits (SK1, SK2, and SK3) from rat brain cDNAs in 1996 (36). All three are very similar in overall topology (Fig. 3A) and all form homotetrameric channels. They are between 553 and + 580 amino acids long, and they each contain a canonical K channel pore region and six transmembrane segments (S1-S6), + which places them in the voltage-gated-K -channel gene family (Fig. 3A). SKCa channels, however, are not voltage-sensitive, apparently because they have lost most of the positively charged residues normally associated with voltage-dependent gating (37, 38). The three SK channels display 80 to 90% sequence identity across their transmembrane regions and diverge somewhat in their N and C termini (36). A fourth member of this family was also subsequently identified (39). It is 2+ termed IK or SK4. All SKCa channel’s are activated by Ca 2+ and, as shown in Fig. 3B, they display remarkably similar Ca sensing properties (39, 40). Each responds rapidly to changes in Ca2+ with a time constant between 5 and 15 ms and each 2+ has an apparent affinity for Ca of 0.3 µM and a hill co2+ efficient for Ca -dependent activation of between 4 and 5 (40). Thus, Ca2+ binds rapidly and tightly to SKCa channels, 2+ and multiple Ca ions bind in a cooperative manner. Despite this impressive Ca2+ sensitivity, as with the BKCa channel, it was noticed early on that SKCa channels contain no 2+ canonical Ca binding motifs (36). The answer to the question: where does their Ca2+ sensitivity come from, however, turned out to be quite different. In 1998 Xia et al. showed by http://bmbreports.org
Ca2+-regulated ion channels Daniel H. Cox
Fig. 3. (A) Membrane topology of the pore-forming subunit of the SKCa channel, from Faber 2009 (80). (B) Ca dose-response curves of the three SKCa channels, from Xia et al. 1998 (40). (C) Crystal structure of two CaM (green) in complex with two CaM binding domains (blue and yellow) of the SK2 channel. The structure is from Schumacher et al. 2001 (44). Ca2+ ions bound to the N lobes of CaM are indicated in red. (D) Model of SKCa-channel gating of Schumacher et al. 2001 (44).
yeast-two hybrid, and biochemical assays that the ubiquitous 2+ Ca -binding protein calmodulin (CaM) constitutively associates with all three SKCa channels (40). And, more telling, 2+ when mutant CaMs that contain two or more disabled Ca binding sites were co-expressed with the SKCa channels, the channels’ apparent Ca2+ affinities where greatly reduced or eliminated (40). Thus, constitutively-associated CaM appears 2+ to serve as the SKCa channel’s Ca sensor, presumably one per subunit, four per channel. CaM is a 148 amino-acid cytoplasmic protein found in virtually all cell types. It serves in a great variety of contexts as a 2+ Ca sensor. Indeed over 300 CaM binding proteins have been identified (2, 41). As shown in Fig. 4, CaM has a bilobe struc2+ ture with each lobe contain two Ca binding sites of the EF-hand type. Upon binding Ca2+, each EF hand changes conformation and this causes a bending of the helix separating the two lobes. This change in the central helix’s structure often greatly increases the affinity of CaM for its target proteins (42). In some instances, however, CaM has been found to associate http://bmbreports.org
2+ with target proteins in a Ca -independent manner, as is the case with the SKCa channel. Most such interactions occur through a well characterized 11 amino-acid motif termed an IQ motif (IQXXXRGXXXR) (43). Surprisingly, however, none of the SK subunits contain such a motif. Thus, the nature of their constitutive association with CaM is considered novel and perhaps unique. 2+ That CaM is in fact the SKCa channel’s Ca sensor was further confirmed in 2001, when the Adelman laboratory solved the crystal structure of CaM bound to a 95-amino-acid peptide that corresponded to the proximal half of SK2’s intracellular C-terminal domain (44). In the structure this CaM-binding domain (CaMBD) forms a long α helix, a turn, and then a shorter helix (Fig. 3C). And surprisingly, although there is only one CaMBD in each channel subunit, there were two CaMBDs and two CaMs in the asymmetric unit of the crystal. That is, two CaMBDs and two CaMs where forming a dimer, with each CaM lying perpendicular to the CaMBDs and making contacts with both (Fig. 3C). This was surprising because it suggests that
Ca2+-regulated ion channels Daniel H. Cox
the energetic aspects of this channel’s Ca2+-dependent gating have yet to be addressed. When it is free in solution, each lobe of CaM is known to bind Ca2+ cooperatively with apparent affinities in the low micromolar range, such that all sites are sa2+ turated above 10 µM Ca (46, 47). CaM’s N-lobe affinities, however, may be higher when it is associated with an SKCa 2+ channel, as the Ca dose-response curves of these channels saturates by 1 µM (Fig. 3B). 2+
Fig. 4. Schematic diagram of CaM with four bound Ca , from Keen et al. 1999 (81).
in the native channel the CaM constitutively associated with one SK subunit, upon binding Ca2+, reaches over and grabs the C-terminal CaMBD of one of its neighboring subunits, and this new interaction is then reciprocated by the CaM associated with the second subunit (Fig. 3D). If one considers that there are four SK subunits per channel, if each grabs a partner 2+ upon Ca binding, the intracellular potion of the channel must necessarily change from four-fold to two-fold symmetry 2+ as Ca binds, a tetramer in the membrane and a dimer of dimers beneath. Exactly how this linking of CaMBDs leads to channel opening has yet to be elucidated, but it appears likely that this interaction creates a torque on the CaMBDs that is translated through the S6 helices to the channel’s gate (45). This notion has lead to the model of SKCa channel gating depicted in Fig. 3D (44). Another interesting aspect of this mech2+ anism is that, in the crystal structure, Ca was found to bind only to the N-lobe of CaM, while the C-lobe formed a constitutive association with the CaMBD (44). Indeed it appears that in forming a tight association with the CaMBD with an affinity on the order of 50-100 nM (45) the structure of the 2+ C-lobe becomes altered such that the Ca -coordinating residues of its EF hands are no longer properly oriented to coordinate Ca2+. As of yet there are no good estimates of the affinities of the 2+ Ca binding sites of the N-lobe of CaM when it is attached to the SKCa channel, or how they change as the channel gates. So 640 BMB reports
Voltage-gated Ca2+ channels (CaV) occupy a unique position 2+ in cellular Ca -signaling, as they are often both the source of 2+ Ca entry and the object of Ca2+-dependent feedback. Indeed, many experiments with both single CaV channels and 2+ macroscopic CaV currents have demonstrated that Ca ions, either entering through a given channel, or provided by other sources, can bind to intracellular regions of a CaV channel to regulate its voltage-dependent gating, and they can do so in either a faciliatory or an inhibitory manner (48-54). This feedback regulation allows for subtle control, in both space and 2+ time, of the intracellular Ca profile, and it has been shown to play an important role in such processes as the regulation of the duration of the cardiac action potential (55) and short-term synaptic plasticity (56, 57). CaV channels are formed by pore-forming α1 subunits (Fig. 5B) and auxiliary β, α2-δ and sometimes γ subunits (Fig. 5A) (58). It is the α1 subunit that determines a CaV channel’s essential phenotype (48). There are three α1 gene families that give rise to channels of type: CaV1, CaV2 or CaV3 (59). CaV3 2+ channels lack Ca -dependent regulation, while CaV1 and CaV2 channels display either of two types. First, in response to a sustained depolarization sufficient to activate CaV current, 2+ the subsequent intracellular Ca rise causes a slow decline in 2+ the observed Ca current known as Ca2+-dependent inactivation or CDI (60). CDI proceeds over the course of hundreds of milliseconds and often leads to near complete in2+ hibition of the current (Fig. 5C) (61, 62). When Ba rather 2+ than Ca is used as the charge carrier, CDI is no longer ob2+ served, as Ba does not bind well to the relevant intracellular 2+ Ca binding sites. Apart from CDI, some CaV channels also display CDF, Ca2+-dependent facilitation, which is observed as a rapid increase in open-probability, and thus macroscopic current, just after channel activation (63, 64). It can be elicited with a brief depolarizing prepulse and then observed in a subsequent test pulse (50, 53, 65, 66) (Fig. 5D). CDF requires 2+ 2+ Ca influx and like CDI is eliminated when Ba is substituted for Ca2+. CDF develops more rapidly than does CDI, with a time constant of 10 ms (50). All CaV1 and CaV2 channels display CDI and some display both CDI and CDF (CaV1.2 and CaV2.1) (51, 67) Over the course of the last decade a good deal has been learned about the molecular underpinnings of CDI and CDF, and what has become clear is that both processes depend on a http://bmbreports.org
Ca2+-regulated ion channels Daniel H. Cox
Fig. 5. (A) Schematic diagram of the voltage-gated Ca2+ channel, from Lacinová, 2005 (58). (B) Membrane topology of the voltage-gated Ca2+ channel, adapted from Lacinová, 2005 (58). Depicted is the PreIQ-IQ motif (red) and a bound CaM molecule (blue and green). (C) Illustration 2+ 2+ of Ca -dependent inactivation, from Dick et al. 2008 (61). (D) Illustration of Ca -dependent facilitation, from Chaudhuri et al. 2004 (65).
single CaM being constitutively bound to the channel’s intracellular C-terminal domain in a region fairly close (within 150 amino acids) to the channel’s last transmembrane segment (53, 61, 66, 68-73). Indeed, both CaV1 and CaV2 channels contain an IQ motif in this region that has been shown in crystallography experiments to bind CaM (74-77), and a CaM mutant 2+ with disabled Ca binding sites has been shown to act in a dominant negative manner to inhibit both CDI and CDF (50). 2+ Thus, CaM does not need bound Ca for it to effectively compete with native CaM in binding to the channel. 2+ More interesting, however, unlike the apoCaM (no Ca bound) constitutively bound to the SKCa channel, the CaV channel’s apoCaM has both lobes free to bind Ca2+, and experiments with mutant CaMs have shown that in CaV2.1 2+ channels CDI requires Ca binding to the N-lobe of CaM, 2+ while CDF requires Ca binding to the C-lobe (50). Thus, remarkably, two lobes of the same molecule subserve two differ2+ ent and partially opposing Ca -dependent regulatory processes, and these processes proceed at very different rates. How this occurs is not yet fully understood, however, we 2+ may assume that, upon binding Ca , each lobe is either rehttp://bmbreports.org
positioned on the channel or else induced to make further contacts with other regions of the channel, and that these new contacts allosterically influence channel gating either positively (CDF) or negatively (CDI). Furthermore, it must be that both CaM lobes can form new contacts at separate sites but at the same time, and presumably with differing affinities and kinetics. To gain a clearer understand of how this works, however, will likely require biochemical and/or structural experiments that delineate the changes that occur in the interactions of each lobe of CaM with the CaV channel as each lobe be2+ comes occupied by Ca . There are however two intriguing characteristics of CDI and CDF that are worth mentioning here and have been largely explained. The first is that, as stated above, in CaV2 channels 2+ CDI requires Ca binding to the N-lobe of CaM and CDF re2+ quires Ca binding to the C-lobe, but in CaV1 channels the 2+ situation is reversed. CDI requires Ca binding to the C-lobe and CDF requires Ca2+ binding to the N-lobe (51). What is the origin of this role reversal? The answer has come largely from x-ray crystallography. Crystal structures of IQ domains from 2+ CaV1 channels in complex with Ca -saturated CaM have reBMB reports
Ca2+-regulated ion channels Daniel H. Cox
Fig. 6. Crystal Structure of CaM in complex with the CaM binding domains of CaV1.2 (A) and CaV2.1 (B). Note the orientation of the CaM molecule is reversed between the two structures. The structure in A is from Van Petegem et al. (2005) (77). The structure in B is from Kim et al. (2008) (75). The figure is adapted from Minor and Findeisen, 2010 (72) (C) Table showing the relationship between channel type, CaM lobe, and functional effect.
vealed that Ca2+/CaM binds to these channels’ IQ motifs in a parallel configuration (Fig. 6A) (74, 77). That is, the N-lobe of CaM makes contacts with the IQ peptide just upstream of where the C-lobe makes contacts. But when similar experiments were done with IQ peptides from CaV2 channels, the 2+ orientation of the Ca /CaM was found to be reversed (75). 2+ That is, Ca /CaM adopted an anti-parallel configuration (Fig. 6B). Thus, it appears that it is the position of the CaM lobe along the channel’s IQ domain that determines which process, 2+ CDI or CDF, will be initiated upon Ca binding to that lobe, not the exact nature of the lobe itself. The reason for the in2+ version in Ca /CaM-binding orientation between CaV1 and CaV2 is not apparent from the crystal structures, but it is not a simple inversion, but rather in the CaV2/Ca2+/CaM structure CaM is not only in an anti-parallel orientation, but it also sits 3 residues further up the IQ helix than it does in the CaV1/ 2+ Ca /CaM structure (72). Thus, to a CaV channel it appears that a lobe is a lobe, and any lobe of CaM will do for either CDI or CDF, so long as it is oriented on the channel correctly. But there is another striking feature of these two processes that does depend on the nature of the lobe itself, rather than on its position, and that is the lo2+ cal or global nature of its Ca sensing. In CaV2 channels CDI mediated by the N-lobe of CaM is blocked by millimolar concentrations of internal EGTA (67, 78). This is unexpected, because millimolar concentrations of EGTA do not effectively
642 BMB reports
2+ buffer the very high Ca concentration (tens of micromolar) that develops around (within 10 nm) the inner mouth of a 2+ Ca channel within microseconds of its opening (79). CaM is attached to the channel close to its inner mouth, so one would expect that both of its lobes would be exposed to this high 2+ Ca concentration every time the channel opens. Indeed, supporting this notion CDF in these same channels which is mediated by the C-lobe unlike CDI is not blocked by EGTA (54), which indicates that the C-lobe is situated within the 2+ Ca nanodomain of the open channel. How can the C-lobe be in the Ca2+ nanodomain and the N-lobe not be? Could it be that, upon channel opening, the 2+ C-lobe is somehow exposed to an appreciably higher Ca concentration because is lies slightly upstream of the N lobe along the IQ motif? No apparently not, as in the CaV1.2 channel the downstream lobe is the C-lobe and the upstream lobe is the N-lobe, yet here still the process mediated by the C-lobe (with this channel CDI) is insensitive to EGTA, while the process mediated by the N-lobe (with this channel CDF) is blocked by EGTA (67). Thus, there appears to be something about the nature of the N and C-lobes of CaM that allows the N-lobe to 2+ be responsive to changes in global Ca (which can be buffered by EGTA) and the C-lobe to be responsive to changes in 2+ local Ca . But exactly how a single molecule could be simul2+ taneously sensitive to both global and local Ca seems a mystery.
Ca2+-regulated ion channels Daniel H. Cox
A solution to this mystery, however, has been proposed by Tadross et al. (2008) (54). They demonstrated with computer simulations that, if one supposes that the a CaM lobe binds 2+ Ca quickly and lets it go slowly, then during the rapid openings and closing of a Ca2+ channel, Ca2+ would bind and accumulate on the lobe, which could then bind to its target site and bring about either CDI or CDF. That is, it would sense the 2+ 2+ local Ca signal coming from the Ca spikes that occur every time the channel opens, and it would integrate them until 2+ were bound to initiate lobe-dependent sufficient Ca signaling. Furthermore, these transients would be insensitive to EGTA. This is what Tadross et al. proposed is occurring with 2+ the C-lobe. Conversely If a CaM lobe were to unbind Ca rapidly, rather than slowly, then, even though it may have a high affinity for Ca2+, the rapid unbinding would render the lobe in2+ 2+ sensitive to local Ca transients. Although Ca would per2+ haps bind during a Ca transient, it would also likely unbind during the subsequent channel closure. Thus, such a lobe 2+ would very seldom spend enough time in a Ca -bound state to effectively initiate Ca2+-dependent processes. This would render the lobe insensitive to local Ca2+ transients, and this is what Tadross et al. have proposed is occurring with the N-lobe. Furthermore, under this hypothesis, the N-lobe would 2+ still be quite responsive to slow changes in Ca concentration, that is to say global changes, even if they were small in 2+ amplitude. This is because with a sustained Ca level, when 2+ one Ca unbinds, another quickly binds, and the N-lobe 2+ would therefore spend enough time Ca saturated to initiate either CDI or CDF. Thus, what at first glance seemed like an irreconcilable conundrum can be nicely explained by differ2+ ences in Ca binding kinetics between the N and C-lobes. It should also be mentioned, however, that Tadross et al. (54) al2+ so demonstrated that a fairly weak affinity of the Ca -bound N lobe for its target site is also required for it to be insensitive 2+ to local Ca transients, and definitive evidence in favor of this 2+ theory in the form of kinetic measurements of Ca binding to channel-resident CaM, have yet to be acquired. Tadross et al. did, however, make a compelling case for their model based on the observation that it quite accurately predicts the relationship between channel open probability and the extent of CDI under a variety of conditions (54). As is the case for the SKCa channel, the affinities of CaM’s 2+ Ca binding sites when attached to a CaV channel, or how they change as the channel gates, have yet to be determined, so the energetics of CDI and CDF have yet to be addressed. But perhaps most pressing, so far little is known about how the 2+ interactions that occur between CaM/Ca /CaMBD and the rest of the channel lead to either inactivation or facilitation. This issue appears to be the next frontier in the study of this 2+ channel’s Ca dependent regulation.
CONCLUSION Over the course of the last decade a great deal has been http://bmbreports.org
2+ 2+ learned about the Ca sensing mechanisms of Ca -activated + 2+ K channels and voltage-gated Ca channels. Each channel type uses unique molecular machinery to regulate its gating, however, the use of constitutively-associated calmodulin is emerging as an important common theme. X-ray crystallography and electrophysiology together have proven to make a 2+ powerful duo in the investigation of Ca -dependent gating. Indeed, in the long run the exceptionally quantitative nature of electrophysiology, combined with the high resolution of x-ray crystallography, promises perhaps a better understanding of 2+ the Ca -dependent regulation of ion-channel gating than the Ca2+-dependent regulation of most other processes.
REFERENCES 1. Clapham, D. E. (2007) Calcium signaling. Cell 131, 10471058. 2. Williams, R. J. P. (1999) Calcium as a Cellular Regulator. Oxford University press, Oxford. 3. Celio, M. R., Pauls, T. L. and Schwaller, B. (1996) Guidebook to the Calcium-binding Proteins. Oxford University Press, Oxford. 4. Kunzelmann, K., Kongsuphol, P., Chootip, K., Toledo, C., Martins, J. R., Almaça, J., Tian, Y., Witzgall, R., Ousingsa2+ wat, J. and Schreiber, R. (2011) Role of the Ca -activated Cl- channels bestrophin and anoctamin in epithelial cells. Biol. Chem. 392, 125-134. 5. Ferrera, L., Caputo, A. and Galietta, L. J. V. (2010) TMEM 2+ 16A protein: a new identity for Ca -dependent Cl channels. Physiology (Bethesda, Md.) 25, 357-363. 6. Ramsey, I. S., Delling, M. and Clapham, D. E. (2006) An introduction to TRP channels. Annu. Rev. Physiol. 68, 619647. 7. Lee, U. S. and Cui, J. (2010) BK channel activation: structural and functional insights. Trends Neurosci. 33, 415-423. 8. Butler, A., Tsunoda, S., McCobb, D. P., Wei, A. and Salkoff, L. (1993) mSlo, a complex mouse gene encoding "maxi" calcium-activated potassium channels. Science 261, 221224. 9. Brenner, R., Jegla, T. J., Wickenden, A., Liu, Y. and Aldrich, R. W. (2000) Cloning and functional characterization of novel large conductance calcium-activated potassium channel beta subunits, hKCNMB3 and hKCNMB4. J. Biol. Chem. 275, 6453-6461. 10. McManus, O. B., Helms, L. M., Pallanck, L., Ganetzky, B., Swanson, R. and Leonard, R. J. (1995) Functional role of the beta subunit of high conductance calcium-activated potassium channels. Neuron 14, 645-650. 11. Lu, R., Alioua, A., Kumar, Y., Eghbali, M., Stefani, E. and Toro, L. (2006) MaxiK channel partners: physiological impact. J. Physiol. 570, 65-72. 12. Adelman, J. P., Shen, K. Z., Kavanaugh, M. P., Warren, R. A., Wu, Y. N., Lagrutta, A., Bond, C. T. and North, R. A. (1992) Calcium-activated potassium channels expressed from cloned complementary DNAs. Neuron 9, 209-216. 13. Atkinson, N. S., Robertson, G. A. and Ganetzky, B. (1991) A component of calcium-activated potassium channels encoded by the Drosophila slo locus. Science 253, 551-555. BMB reports
Ca2+-regulated ion channels Daniel H. Cox
14. Meera, P., Wallner, M., Song, M. and Toro, L. (1997) Large conductance voltage- and calcium-dependent K+ channel, a distinct member of voltage-dependent ion channels with seven N-terminal transmembrane segments (S0-S6), an extracellular N terminus, and an intracellular (S9-S10) C terminus. Proc. Natl. Acad. Sci. U. S. A. 94, 14066-14071. 15. Barrett, J. N., Magleby, K. L. and Pallotta, B. S. (1982) Properties of single calcium-activated potassium channels in cultured rat muscle. J. Physiol. (Lond) 331, 211-230. 16. Cui, J., Cox, D. H. and Aldrich, R. W. (1997) Intrinsic volt2+ age dependence and Ca regulation of mslo large conductance Ca-activated K+ channels. J. Gen. Physiol. 109, 647-673. 2+ 17. Shi, J. and Cui, J. (2001) Intracellular Mg enhances the 2+ + function of BK-type Ca -activated K channels. J. Gen. Physiol. 118, 589-606. 18. Shi, J., Krishnamoorthy, G., Yang, Y., Hu, L., Chaturvedi, N., Harilal, D., Qin, J. and Cui, J. (2002) Mechanism of magnesium activation of calcium-activated potassium channels. Nature 418, 876-880. 19. Xia, X. M., Zeng, X. and Lingle, C. J. (2002) Multiple regulatory sites in large-conductance calcium-activated potassium channels. Nature 418, 880-884. 20. Schreiber, M. and Salkoff, L. (1997) A novel calcium-sensing domain in the BK channel. Biophys. J. 73, 1355-1363. 21. Bao, L., Rapin, A. M., Holmstrand, E. C. and Cox, D. H. (2002) Elimination of the BK(Ca) channel's high-affinity 2+ Ca sensitivity. J. Gen. Physiol. 120, 173-189. 22. Zhang, G., Huang, S.-Y., Yang, J., Shi, J., Yang, X., Moller, A., Zou, X. and Cui, J. (2010) Ion sensing in the RCK1 domain of BK channels. Proc. Natl. Acad. Sci. U. S. A. 107, 18700- 18705. 23. Bao, L., Kaldany, C., Holmstrand, E. C. and Cox, D. H. 2+ (2004) Mapping the BKCa channel's "Ca bowl": sidechains essential for Ca2+ sensing. J. Gen. Physiol. 123, 475-489. 2+ 24. Bian, S., Favre, I. and Moczydlowski, E. (2001) Ca -binding activity of a COOH-terminal fragment of the Drosophila BK channel involved in Ca2+-dependent activation. Proc. Natl. Acad. Sci. U. S. A. 98, 4776-4781. 25. Wang, L. and Sigworth, F. J. (2009) Structure of the BK potassium channel in a lipid membrane from electron cryomicroscopy. Nature 461, 292-295. 26. Wu, Y., Yang, Y., Ye, S. and Jiang, Y. (2010) Structure of the 2+ gating ring from the human large-conductance Ca -gated + K channel. Nature 466, 393-397. 27. Yuan, P., Leonetti, M. D., Pico, A. R., Hsiung, Y. and MacKinnon, R. (2010) Structure of the human BK channel 2+ Ca -activation apparatus at 3.0 A resolution. Science (New York, NY) 329, 182-186. 28. Jiang, Y., Lee, A., Chen, J., Cadene, M., Chait, B. T. and MacKinnon, R. (2002) Crystal structure and mechanism of a calcium-gated potassium channel. Nature 417, 515-522. 29. Kim, H.-J., Lim, H.-H., Rho, S.-H., Eom, S. H. and Park, C.-S. + (2006) Hydrophobic interface between two regulators of K conductance domains critical for calcium-dependent activation of large conductance Ca2+-activated K+ channels. J. Biol. Chem. 281, 38573- 38581. 30. Ye, S., Li, Y., Chen, L. and Jiang, Y. (2006) Crystal structures of a ligand-free MthK gating ring: insights into the ligand gat644 BMB reports
32. 33. 34. 35.
41. 42. 43. 44.
ing mechanism of K+ channels. Cell 126, 1161-1173. Pau, V. P. T., Abarca-Heidemann, K. and Rothberg, B. S. (2010) Allosteric mechanism of Ca2+ activation and H+-in+ hibited gating of the MthK K channel. J. Gen. Physiol. 135, 509-526. Zadek, B. and Nimigean, C. M. (2006) Calcium-dependent gating of MthK, a prokaryotic potassium channel. J. Gen. Physiol. 127, 673-685. Cox, D. H. (2006) BKCa-channel structure and function; in Biological Membrane Ion Channels, pp. 171-219, Springer Science+Business Media LLC. Sweet, T.-B. and Cox, D. H. (2008) Measurements of the 2+ BKCa channel‘s high-affinity Ca binding constants: effects of membrane voltage. J. Gen. Physiol. 132, 491-505. Zhang, X., Solaro, C. R. and Lingle, C. J. (2001) Allosteric 2+ 2+ regulation of BK channel gating by Ca and Mg through a nonselective, low affinity divalent cation site. J. Gen. Physiol. 118, 607-636. Köhler, M., Hirschberg, B., Bond, C. T., Kinzie, J. M., Marrion, N. V., Maylie, J. and Adelman, J. P. (1996) Smallconductance, calcium-activated potassium channels from mammalian brain. Science (New York, NY) 273, 17091714. Aggarwal, S. K. and MacKinnon, R. (1996) Contribution of + the S4 segment to gating charge in the Shaker K channel. Neuron 16, 1169-1177. Sigg, D. and Bezanilla, F. (1997) Total charge movement per channel. The relation between gating charge displacement and the voltage sensitivity of activation. J. Gen. Physiol. 109, 27-39. Ishii, T. M., Silvia, C., Hirschberg, B., Bond, C. T., Adelman, J. P. and Maylie, J. (1997) A human intermediate conductance calcium-activated potassium channel. Proc. Natl. Acad. Sci. U.S.A. 94, 11651-11656. Xia, X. M., Fakler, B., Rivard, A., Wayman, G., JohnsonPais, T., Keen, J. E., Ishii, T., Hirschberg, B., Bond, C. T., Lutsenko, S., Maylie, J. and Adelman, J. P. (1998) Mechanism of calcium gating in small-conductance calcium-activated potassium channels. Nature 395, 503-507. Yap, K. L., Kim, J., Truong, K., Sherman, M., Yuan, T. and Ikura, M. (2000) Calmodulin target database. J. Struct. Funct. Genomics 1, 8-14. Yap, K. L., Ames, J. B., Swindells, M. B. and Ikura, M. (1999) Diversity of conformational states and changes within the EF-hand protein superfamily. Proteins 37, 499-507. Bahler, M. and Rhoads, A. (2002) Calmodulin signaling via the IQ motif. FEBS Lett. 513, 107-113. Schumacher, M. A., Rivard, A. F., Bachinger, H. P. and Adelman, J. P. (2001) Structure of the gating domain of a 2+ + 2+ Ca -activated K channel complexed with Ca /calmodulin. Nature 410, 1120-1124. Schumacher, M. A., Crum, M. and Miller, M. C. (2004) Crystal structures of apocalmodulin and an apocalmodulin/ SK potassium channel gating domain complex. Structure (London, England : 1993) 12, 849-860. Linse, S., Helmersson, A. and Forsén, S. (1991) Calcium binding to calmodulin and its globular domains. J. Biol. Chem. 266, 8050-8054. Stefan, M. I., Edelstein, S. J. and Le Novère, N. (2008) An allosteric model of calmodulin explains differential activation http://bmbreports.org
Ca2+-regulated ion channels Daniel H. Cox
48. 49. 50.
51. 52. 53. 54. 55.
of PP2B and CaMKII. Proc. Natl. Acad. Sci. U. S. A. 105, 10768-10773. Catterall, W. A. (2011) Voltage-gated calcium channels. Cold Spring Harbor Perspectives in Biology 3, 1-23. Chaudhuri, D., Issa, J. B. and Yue, D. T. (2007) Elementary mechanisms producing facilitation of Cav2.1 (P/Q-type) channels. J. Gen. Physiol. 129, 385-401. DeMaria, C. D., Soong, T. W., Alseikhan, B. A., Alvania, R. S. and Yue, D. T. (2001) Calmodulin bifurcates the local 2+ 2+ Ca signal that modulates P/Q-type Ca channels. Nature 411, 484-489. Dunlap, K. (2007) Calcium Channels Are Models of SelfControl. J. Gen. Physiol. 129, 379-383. 2+ Imredy, J. P. and Yue, D. T. (1994) Mechanism of Ca - sen2+ sitive inactivation of L-type Ca channels. Neuron 12, 1301-1318. 2+ Lee, A., Scheuer, T. and Catterall, W. A. (2000) Ca /calmodulin-dependent facilitation and inactivation of P/Q-type 2+ Ca channels. J. Neurosci. 20, 6830-6838. Tadross, M. R., Dick, I. E. and Yue, D. T. (2008) Mechanism of local and global Ca2+ sensing by calmodulin in complex 2+ with a Ca channel. Cell 133, 1228-1240. Alseikhan, B. A., DeMaria, C. D., Colecraft, H. M. and Yue, D. T. (2002) Engineered calmodulins reveal the unexpected 2+ eminence of Ca channel inactivation in controlling heart excitation. Proc. Natl. Acad. Sci. U. S. A. 99, 17185-17190. Catterall, W. A. and Few, A. P. (2008) Calcium channel regulation and presynaptic plasticity. Neuron 59, 882-901. Chaudhuri, D., Alseikhan, B. A., Chang, S. Y., Soong, T. W. and Yue, D. T. (2005) Developmental activation of calm2+ odulin-dependent facilitation of cerebellar P-type Ca current. J. Neurosci. 25, 8282-8294. Lacinová, L. (2005) Voltage-dependent calcium channels. Gen. Physiol. Biophys. 24 (Suppl 1), 1-78. Ertel, E. A., Campbell, K. P., Harpold, M. M., Hofmann, F., Mori, Y., Perez-Reyes, E., Schwartz, A., Snutch, T. P., Tanabe, T., Birnbaumer, L., Tsien, R. W. and Catterall, W. A. (2000) Nomenclature of voltage-gated calcium channels. Neuron 25, 533-535. Brehm, P., Eckert, R. and Tillotson, D. (1980) Calciummediated inactivation of calcium current in Paramecium. J. Physiol. 306, 193-203. Dick, I. E., Tadross, M. R., Liang, H., Tay, L. H., Yang, W. 2+ and Yue, D. T. (2008) A modular switch for spatial Ca selectivity in the calmodulin regulation of CaV channels. Nature 451, 830-834. Lee, A., Wong, S. T., Gallagher, D., Li, B., Storm, D. R., 2+ Scheuer, T. and Catterall, W. A. (1999) Ca /calmodulin binds to and modulates P/Q-type calcium channels. Nature 399, 155-159. Borst, J. G. and Sakmann, B. (1998) Facilitation of presynaptic calcium currents in the rat brainstem. J. Physiol. 513 (Pt 1), 149-155. Cuttle, M. F., Tsujimoto, T., Forsythe, I. D. and Takahashi, T. (1998) Facilitation of the presynaptic calcium current at an auditory synapse in rat brainstem. J. Physiol. 512 (Pt 3), 723-729. Chaudhuri, D., Chang, S. Y., DeMaria, C. D., Alvania, R. S., Soong, T. W. and Yue, D. T. (2004) Alternative splicing as a 2+ molecular switch for Ca /calmodulin-dependent facilita-
78. 79. 80. 81.
tion of P/Q-type Ca2+ channels. J. Neurosci. 24, 6334-6342. Zuhlke, R. D., Pitt, G. S., Tsien, R. W. and Reuter, H. (2000) Ca2+-sensitive inactivation and facilitation of L-type Ca2+ channels both depend on specific amino acid residues in a consensus calmodulin-binding motif in the(alpha)1C subunit. J. Biol. Chem. 275, 21121-21129. Liang, H., DeMaria, C. D., Erickson, M. G., Mori, M. X., Alseikhan, B. A. and Yue, D. T. (2003) Unified mechanisms 2+ 2+ of Ca regulation across the Ca channel family. Neuron 39, 951-960. Erickson, M. G., Alseikhan, B. A., Peterson, B. Z. and Yue, D. T. (2001) Preassociation of calmodulin with voltage-gated 2+ Ca channels revealed by FRET in single living cells. Neuron 31, 973-985. Mori, M. X., Erickson, M. G. and Yue, D. T. (2004) Functional stoichiometry and local enrichment of calmodulin inter2+ acting with Ca channels. Science (New York, NY) 304, 432-435. Zühlke, R. D., Pitt, G. S., Deisseroth, K., Tsien, R. W. and Reuter, H. (1999) Calmodulin supports both inactivation and facilitation of L-type calcium channels. Nature 399, 159-162. Lee, A., Zhou, H., Scheuer, T. and Catterall, W. A. (2003) 2+ Molecular determinants of Ca /calmodulin-dependent regulation of Ca(v)2.1 channels. Proc. Natl. Acad. Sci. U. S. A. 100, 16059-16064. Minor, D. L. and Findeisen, F. (2010) Progress in the structural understanding of voltage-gated calcium channel (CaV) function and modulation. Channels (Austin, Tex) 4, 459474. Peterson, B. Z., DeMaria, C. D., Adelman, J. P. and Yue, D. 2+ 2+ T. (1999) Calmodulin is the Ca sensor for Ca -dependent inactivation of L-type calcium channels. Neuron 22, 549-558. Fallon, J. L., Halling, D. B., Hamilton, S. L. and Quiocho, F. A. (2005) Structure of calmodulin bound to the hydrophobic IQ domain of the cardiac Ca(v)1.2 calcium channel. Structure (London, England : 1993) 13, 1881-1886. Kim, E. Y., Rumpf, C. H., Fujiwara, Y., Cooley, E. S., Van Pete2+ gem, F. and Minor, D. L. (2008) Structures of CaV2 Ca / CaM-IQ domain complexes reveal binding modes that underlie calcium-dependent inactivation and facilitation. Structure (London, England : 1993) 16, 1455-1467. Mori, M. X., Vander Kooi, C. W., Leahy, D. J. and Yue, D. T. (2008) Crystal structure of the CaV2 IQ domain in complex 2+ with Ca /calmodulin: high-resolution mechanistic implications for channel regulation by Ca2+. Structure (London, England : 1993) 16, 607-620. Van Petegem, F., Chatelain, F. C. and Minor, D. L. (2005) Insights into voltage-gated calcium channel regulation from 2+ the structure of the CaV1.2 IQ domain-Ca /calmodulin complex. Nat. Struct. Mol. Biol. 12, 1108-1115. Cox, D. H. and Dunlap, K. (1994) Inactivation of N-type calcium current in chick sensory neurons: calcium and voltage dependence. J. Gen. Physiol. 104, 311- 336. Fakler, B. and Adelman, J. P. (2008) Control of K(Ca) channels by calcium nano/microdomains. Neuron 59, 873-881. Faber, E. S. L. (2009) Functions and modulation of neuronal SK channels. Cell Biochem. Biophys. 55, 127- 139. Keen, J. E., Khawaled, R., Farrens, D. L., Neelands, T., BMB reports
Ca2+-regulated ion channels Daniel H. Cox
Rivard, A., Bond, C. T., Janowsky, A., Fakler, B., Adelman, J. P. and Maylie, J. (1999) Domains responsible for constitutive and Ca2+-dependent interactions between calm-
646 BMB reports
odulin and small conductance Ca2+-activated potassium channels. J. Neurosci. 19, 8830-8838.