C-terminal residues of skeletal muscle calsequestrin are essential for calcium binding and for skeletal ryanodine receptor inhibition

Skeletal muscle function depends on calcium signaling proteins in the sarcoplasmic reticulum (SR), including the calcium-binding protein calsequestrin (CSQ), the ryanodine receptor (RyR) calcium release channel, and skeletal triadin 95 kDa (trisk95) and junctin, proteins that bind to calsequestrin type 1 (CSQ1) and ryanodine receptor type 1 (RyR1). CSQ1 inhibits RyR1 and communicates store calcium load to RyR1 channels via trisk95 and/or junctin. In this manuscript, we test predictions that CSQ1’s acidic C-terminus contains binding sites for trisk95 and junctin, the major calcium binding domain, and that it determines CSQ1’s ability to regulate RyR1 activity. Progressive alanine substitution of C-terminal acidic residues of CSQ1 caused a parallel reduction in the calcium binding capacity but did not significantly alter CSQ1’s association with trisk95/junctin or influence its inhibition of RyR1 activity. Deletion of the final seven residues in the C-terminus significantly hampered calcium binding, significantly reduced CSQ’s association with trisk95/junctin and decreased its inhibition of RyR1. Deletion of the full C-terminus further reduced calcium binding to CSQ1 altered its association with trisk95 and junctin and abolished its inhibition of RyR1. The correlation between the number of residues mutated/deleted and binding of calcium, trisk95, and junctin suggests that binding of each depends on diffuse ionic interactions with several C-terminal residues and that these interactions may be required for CSQ1 to maintain normal muscle function.


Background
In skeletal muscle, the rapid and co-ordinated release of calcium ions from the internal sarcoplasmic reticulum (SR) store is essential in triggering muscle contraction. Skeletal muscle activation, in response to an action potential on the surface membrane, activates voltagegated L-type Ca 2+ channels which, in turn, initiate SR Ca 2+ release. The SR Ca 2+ release channel is the ligandgated ryanodine receptor (RyR), a large (>2 mDa) homotetrameric ion channel, which releases between 10% to 17% of the total SR Ca 2+ with each action potential (reviewed in [1]). Ryanodine receptor type 1 (RyR1) is central to excitation-contraction coupling, and its activity and ability to release Ca 2+ is refined by the level of Ca 2+ load inside the SR and by a luminal SR protein complex, including calsequestrin (CSQ), skeletal triadin 95 kDa (trisk95), and junctin.
CSQ is the most abundant Ca 2+ buffering protein found within the SR, with its concentration reported to be between 11 and 36 μmol (l fiber volume) −1 , dependent on muscle type [2]. Calsequestrin type 1 (CSQ1) is the only isoform expressed in fast twitch muscle fibers, while equal amounts of CSQ1 and the so-called cardiac CSQ2 isoform are expressed in slow twitch fibers [2]. Both isoforms display a high degree of homology, with the C-terminal tail extended in CSQ2. CSQ1 is a low affinity, moderate to high-capacity Ca 2+ -binding protein, binding between 50 and 80 mol Ca 2+ /mole CSQ1 with a K D of approximately 1 to 2 mM [3,4]. CSQ1 buffers free SR Ca 2+ to approximately 1 mM during the contraction/relaxation cycle.
Research over the past 2 decades has shown that CSQ1 plays additional roles in global regulation of Ca 2+ signaling in skeletal and cardiac muscles. A retrograde signal from CSQ1 is thought to be important for store-operated Ca 2+ entry in skeletal muscle [5]. CSQ1 inhibits RyR1, while in the heart, CSQ2 activates the cardiac RyR [6]. CSQ is considered to be a luminal Ca 2+ sensor for the RyR, communicating Ca 2+ store load to the channel [7][8][9]. In skeletal muscle, CSQ1 acts as a break on the channel, to curtail SR Ca 2+ release under conditions of lower store load and is essential for normal muscle function. Knockout of CSQ1 leads to a malignant hyperthermia phenotype with excess Ca 2+ release under stress [10]. Despite the functional interaction between CSQ1 and RyR1, there is no evidence that they bind directly in vivo. Instead, both proteins bind to junctin and trisk95 [11][12][13][14][15][16], forming a luminal protein complex, which allows CSQ1-RyR1 communication. The importance of the luminal protein complex in modulating RyR activity and excitation-contraction coupling is highlighted by the fact that mutation or alteration in expression of these proteins results in severe contractile dysfunction [17][18][19][20][21][22][23].
At in vivo luminal [Ca 2+ ]s, CSQ1 monomers selfassociate to form polymers, which are observed as long linear strand-like structures closely associated with the junctional face membrane in electron micrographs [24]. The dynamic polymerization of CSQ1 is highly dependent on Ca 2+ . At low [Ca 2+ ] of 100 nM, CSQ1 exists in an unfolded randomly coiled structure [25]. As [Ca 2+ ] is increased toward 100 μM, CSQ1 undergoes a conformational change [25][26][27] and increases in helicity [26,28]. The conformational changes include the folding of the three thioredoxin domains within CSQ1 and the subsequent formation of front-to-front dimers between two folded monomers [27]. In this process, the Nterminus from one monomer inserts into a groove between two β-strands of domain II of the second monomer [27]. A secondary intermolecular interaction among front-to-front dimers allows them to associate in a backto-back configuration, bringing together two electronegative surfaces to form linear polymers [27].
Several factors influence CSQ1's moderate to high Ca 2+ binding capacity. Ca 2+ binding capacity is partially determined by overall net charge [29]; however, the calculated net charge of skeletal CSQ1 is insufficient for the amount of Ca 2+ that it can bind in its polymer form [30]. This suggests that additional sites for Ca 2+ binding must be formed as CSQ1 polymerizes. The roles of the acidic-rich C-terminal tail (residues 354 to 367 in rabbit) has been postulated to play an important role in inferring the Ca 2+ binding capacity of CSQ1 [15,29,31], although the mechanisms have not been well understood. Recent molecular dynamic modeling studies postulate that Ca 2+ binding sites within the C-terminus of CSQ1 serve to neutralize the interface and enable CSQ polymerization and subsequent high-capacity Ca 2+ binding [32]. Furthermore, the existence of low-affinity Ca 2+ binding sites on CSQ1 may serve, at least in part, to induce additional Ca 2+ binding sites [32], allowing for CSQ's high-capacity Ca 2+ binding. That the C-terminal tail is essential for binding capacity is supported experimentally, as its deletion in CSQ2 severely reduces Ca 2+ binding capacity [29]. The Cterminal tail contains the highest surface negative charge density [30] and was initially thought to be responsible for at least 26% of the Ca 2+ binding, although simulations suggest this to be as low as 10% [32].
CSQ1's acidic-rich C-terminal tail also constitutes a "hot spot" for CSQ1 protein associations. CSQ1 residues 354 to 367 contain the binding site for trisk95 and have been suggested to contain critical residues that support its interaction with junctin [15]. We are yet to understand the true nature of this region, because the anatomical details of the acidic C-terminal tail are not resolved in published crystal structures of CSQ1 (reviewed in [33]) and because there is little information on the role of specific residues within the Cterminal tail in the various functions of CSQ1.
In this paper, we have dissected the functional importance of residues within CSQ1's C-terminal tail. We have examined the effects of substitutions or deletions within the C-terminal tail on CSQ1's Ca 2+ binding capacity, its ability to associate with trisk95 and junctin and its ability to regulate RyR1 channel activity to ensure normal muscle function.

Ethics approval
All animal work was approved by the Australian National University and University of Canberra Animal Ethics Committees.

CSQ1 mutation, expression, and purification
To probe the functional importance of the acidic-rich C-terminal tail in CSQ1, wild-type (WT) CSQ1 and several alanine substitution and deletion mutants (see Figure 1) were generated by PCR of rabbit CSQ1 cDNA using the Stratagene multi-mutagenesis kit. All CSQ1 constructs were subcloned into a pGEX5x1 vector, containing a N-terminal glutathione S-transferase (GST) tag and expressed as previously described [34]. The expressed CSQ1 proteins were dialyzed against 20 mM MOPS, 150 mM NaCl, and 1 mM CaCl 2 (pH 7.4). Where necessary, the protein underwent secondary purification using anti-CSQ1 immunoprecipitation (Pierce crosslinking IP kit), following manufacturer's instruction.

SR vesicle preparation
SR vesicles from back and leg muscles from New Zealand White Rabbits were prepared as described in [7,9,35].
RyR1 purification from skeletal SR vesicles was performed as described in [36]. The purified RyR1 was concentrated, snap-frozen, and stored at −70°C. The protein was run on SDS polyacrylamide gels, and immunoblots probed with anti-RyR1, anti-CSQ1, anti-triadin and antijunctin antibodies to detect contamination by these proteins.
SDS-PAGE and Western blot were performed according to [37,38]. Briefly, proteins were separated on 4% to 15% or 4% to 20% SDS polyacrylamide gels and transferred to polyvinylidene difluoride (PVDF) membrane for Western blot. PVDF membranes were exposed to primary antibodies to CSQ1, junctin, trisk95, and RyR1 (as appropriate) and secondary HRP-conjugated antibody, prior to chemiluminsence detection. Images were developed using a Kodak OXmat M20 film processor onto Kodak Hyper ECL film.

Purification of trisk95 and junctin from skeletal SR vesicles
Trisk95 and junctin were isolated by SDS preparative gel electrophoresis, as previously described [39].

Ca 2+ binding capacity
The Ca 2+ binding capacities of CSQ1 constructs were determined using a modified 45 Ca 2+ spin dialysis binding assay [14]. In brief, 200 μg of CSQ1 constructs were conjugated with 45 Ca 2+ by incubation in 150 mM NaCl, 20 mM MOPS, 100 μM to 5 mM CaCl 2 (pH 7.4), and 2.6 kBq 45 Ca 2+ for 15 min at room temperature. Unconjugated 45 Ca 2+ was removed by centrifugation of the samples in a Microcon centrifugal filter concentrator at 12,000 × g for 10 min. 45 Ca 2+ radioactivity of the retentate (containing CSQ1-45 Ca 2+ ) and of an unfiltered sample aliquot were counted using a Packard 1500 Tri-Carb liquid scintillation analyzer (Packard Instrument Co., Downers Grove, IL, USA). Protein concentration of the retentate (containing CSQ1-45 Ca 2+ ) was determined using a Bio-Rad DC protein determination assay, according to manufacturer's instruction. Data is presented as nmol 45 Ca 2+ /mg CSQ1.

Co-immunoprecipitation
In vitro binding of CSQ1 to purified trisk95 and junctin was performed using a Pierce co-immunoprecipitation kit as per manufacturer's instruction, with the following changes. Co-immunoprecipitation was performed in a buffer containing 20 mM MOPS, 150 mM NaCl, and 1 mM CaCl 2 (pH 7.4), using a 2:1 protein ratio (weight) of CSQ1 to either junctin, trisk95, or RyR1. Affinity chromotography CSQ1-RyR1 interactions were investigated using affinity chromatography as previously described [34].

Turbidity measurements
Solution turbidity [40,41] was monitored spectrophotometrically in a 1-cm path length quartz cuvette. Three micromolar protein was suspended in a buffer containing 20 mM Tris and 100 mM KCl pH 7.4. Small aliquots of CaCl 2 were added to the cuvette to final concentrations of 0.1 to 3 mM. After each addition, the cuvette was stirred and allowed to equilibrate at room temperature for 7.5 min, after which the absorbance was recorded at 350 nm. Absorbance values were corrected for any change in absorbance due to buffer alone. Turbidity is a measure of CSQ1 transition from soluble to insoluble forms as a function of [Ca 2+ ], and increased absorbance at 350 nm reflects the shift toward insoluble particles.

Single channels
Artificial planar bilayers separating two baths (cis and trans) were formed as described previously [7,34]. Native SR vesicles (50 μg) or purified RyR1 (10 μg) were added to the cis solution so that the cytoplasmic surface of the SR and RyR1 faced this solution after incorporation. Solution constituents were as follows: , and open frequency (F o ) in single channel recordings using threshold discrimination or indirectly from the fractional mean current (I' F ) when more than one channel was opening [9]. For simplicity, P o and I' F are combined in calculations of average P o or relative P o [9].

Statistics
Average data are presented as mean ± SE. The significance of differences between control and test values was tested using a Student's t-test for paired data. In some cases, to reduce the effects of variability in control parameters (P oCon , T oCon , T cCon , and F oCon ) and to evaluate parameters after CSQ1 construct addition (P oCSQ , T oCSQ , T cCSQ , and F oCSQ ), data were expressed as the difference between log 10 X CSQ and log 10 X Con for each channel (for example, logP oCSQ − logP oCon ). The difference from control was assessed with a paired t-test applied to log10X Con and log10X CSQ . A P value of ≤0.05 was considered to be significant.

Expression of CSQ1
CSQ1 contains a highly acidic-rich C-terminal domain, 354 EDDDDEDDDDDDDD 367 , purported to be a Ca 2+ binding motif, and contains residues essential for binding both junctin and trisk95 [15]. To investigate the function of the CSQ1 C-terminal domain, we generated two deletion mutants of GST-tagged rabbit CSQ1 and three alanine substitution mutants. The two CSQ1 deletion mutants were generated by deleting either the whole C-terminal tail (CSQ1 ΔC14 ) or 50% of the tail (CSQ1 ΔC7 ) ( Figure 1A). As CSQ Ca 2+ binding motifs consist of a pair of acidic residues located close in space (with each pair binding one calcium ion), alternate acidic residues were substituted to disrupt one (CSQ1 DA1 ), two (CSQ1 DA2 ), or three (CSQ1 DA3 ) Ca 2+ binding sites within the sequence ( Figure 1A). For single channel and 45 Ca 2+ binding studies, CSQ1 constructs were cleaved from the GST by incubation with the serine endopeptidase Factor Xa and purified to homogeneity ( Figure 1B).

Effects of C-terminal tail modification on the Ca 2+ binding capacity of CSQ
The Ca 2+ binding capacity of the CSQ1 constructs was determined using 45 Ca 2+ binding. Ca 2+ binding capacity is dependent on CSQ1 structure and increases sharply as CSQ1 polymerizes. Given that CSQ1's C-terminus is believed to form a large acidic Ca 2+ binding pocket upon polymerization [27], it follows that disruption of this pocket would reduce Ca 2+ binding capacity. Mutation of one or two acidic residues (CSQ1 DA1 and CSQ1 DA2 ) within the C-terminus had no significant effect on Ca 2+ binding capacity from 0.1 to 1 mM Ca 2+ , and the binding curves were similar to that we have previously reported [42], but CSQ1 DA2 bound significantly less Ca 2+ than WT CSQ1 at 2 mM Ca 2+ (Figure 2 (Figure 2). This provides further strong evidence that many of the C-terminal acidic residues of CSQ1 are key to the formation of a Ca 2+ binding pocket, which enhances CSQ's Ca 2+ binding capacity.

CSQ1 aggregation
To determine whether mutation of the acidic C-terminus disrupted CSQ1 aggregation, turbidity measurements were performed. CSQ1 and CSQ2 undergo Ca 2+ -induced compaction or a shift from the proteins soluble to insoluble form [41]. CSQ1 aggregation is measured as turbidity (absorbance at 350 nm), which is proportional to the levels of insoluble CSQ1 [40,41]. The turbidity of WT CSQ1 increases sigmoidily across the [Ca 2+ ] range tested, reaching a constant concentration of insoluble protein at between 1 and 1.5 mM Ca 2+ (Figure 3). Mutation of one or two alternate residues yielded near identical results, while mutation of three residues resulted in a small approximately 20% decrease in absorbance across [Ca 2+ ]'s of 0.7 to 2 mM ( Figure 3; compare CSQ1 DA3 with WT, CSQ1 DA1 and CSQ1 DA2 ). These data indicate that these residues do not play a major role in Ca 2+ -induced CSQ1 compaction. CSQ1 ΔC7 and CSQ1 ΔC14 displayed marked differences in the Ca 2+ -dependent aggregation of CSQ1 ( Figure 3). CSQ1 ΔC7 showed some levels of Ca 2+ -induced aggregation; these were much less than WT, CSQ1 DA1 , CSQ1 DA2 , or CSQ1 DA3 , indicating that residues 361 DDDDDDD 367 are important in the formation of the higher molecular weight entities. Remarkably, deleting CSQ1 C-terminus severely disrupted CSQ1 ability to aggregate, presumably by disrupting Ca 2+ -induced polymerization.

Effects of CSQ modification on in vitro interactions with trisk95 and junctin
We tested the ability of CSQ1 to associate with trisk95 and junctin via its C-terminal residues, using co-immunoprecipitation. Purified CSQ constructs were coupled to anti-CSQ1/agarose protein A/G, prior to exposure to trisk95 or junctin. As has been previously reported [42], there are significant interactions between WT CSQ1/trisk95 and WT CSQ1/junctin at the resting luminal [Ca 2+ ] of 1 mM Ca 2+ and ionic strength (150 mM) ( Figure 4A,C first lane). A substantial interaction between the proteins remained after mutation of one or two alternate residues. There was a small but significant decrease in trisk95 association with CSQ1 DA3 (Figure 4A,D), indicating that 363 D may play a minor role in the trisk95-CSQ1 association. This result suggests that these three alternate C-terminal residues are not critical for the interaction with either trisk95 or junctin.
The deletion constructs CSQ1 ΔC7 and CSQ1 ΔC14 reveal that the overall C-terminal tail domain heavily influences trisk95's interaction with CSQ1. Deletion of the entire C-terminal tail abolished CSQ1 association with trisk95 ( Figure 4A,D). Interestingly, there is a significantly reduced band of trisk95 indicating minimal association with the CSQ1 ΔC7 construct ( Figure 4A,D), indicating that residue 361 DDDDDDD 367 contributes significantly to the trisk95-CSQ1 interaction, with a minor contribution from 363 D. The reduced CSQ1-trisk95 interaction may be due to altered Ca 2+ sensitivity of the trisk95-CSQ1 interaction, which has been shown to be highly Ca 2+ -dependent [15]. There is only modest trisk95-CSQ1 binding at resting luminal [Ca 2+ ] (1 mM) and much higher trisk95-CSQ1 association at lower Ca 2+ concentrations (approximately 100 μM) [15]. However, lowering [Ca 2+ ] to 100 μM did not significantly alter the degree of  trisk95 association with CSQ1 ΔC7 or CSQ1 ΔC14 (compare CSQ1 ΔC7 and CSQ1 ΔC14 in Figure 4A (1 mM Ca 2+ ) with CSQ1 ΔC7 and CSQ1 ΔC14 in Figure 4B (100 μM)). Therefore, the impaired interaction is not due to a shift in Ca 2+ dependence of the association but most likely due to deletion of the interaction site on CSQ1 for trisk95. In contrast to the effects on CSQ/trisk95 binding, deletion of the C-terminal tail of CSQ1 severely impeded, but did not abolish, junctin association with CSQ1 ( Figure 4C, D). Both the CSQ1 ΔC7 and CSQ1 ΔC14 constructs bound approximately 50% to 80% less junctin than did WT CSQ1 ( Figure 4C; compare WT with CSQ1 ΔC7 and CSQ1 ΔC14 ). This implies that the C-terminal tail forms a major part of the junctin binding motif but does not constitute the entire binding domain.

Effects of CSQ1 modification on the regulation of RyR1 channels by CSQ1
The functional consequence of the CSQ1 C-terminal mutations on RyR1 regulation by CSQ1 was studied in lipid bilayers. At a physiological resting [Ca 2+ ] of 1 mM, CSQ1 polymers are tethered to the RyR1 through its interactions with trisk95 and junctin, allowing CSQ1 to inhibit native rabbit skeletal RyR1, specifically through its association with junctin [39]. Thus, it is likely that mutation of residues that reduce Ca 2+ binding capacity and/or inhibit interactions with junctin would alter CSQ1's functional effect on RyR1 gating.
Native skeletal rabbit SR vesicles (which contain the RyR1 and its full complement of associated co-proteins) were incorporated into lipid bilayers, and baseline activity was recorded after addition of 2 mM cis ATP and 4.5 mM trans BAPTA (to lower trans Ca 2+ to 100 nM, that is, a sub Ca 2+ -activated level; Figure 5A, top panel). After approximately 3 to 5 min of stable baseline recording, the Cs + concentration in the trans chamber was increased to 500 mM for 5 min to strip endogenous CSQ1 from the native RyR1 [7]. The trans chamber was then perfused with trans solution to remove dissociated CSQ1 and to reduce trans [Cs + ] to 250 mM. Dissociation of endogenous inhibitor CSQ1 from RyR1 (to yield CSQ-dissociated RyR1) caused an increase in channel open probability (P o ), which is sustained after the removal of high trans Cs + ( Figure 5A, middle trace). As previously reported [7], re-association of WT CSQ1 with the luminal face of the CSQ-dissociated RyR1, caused significant inhibition of RyR1 activity, restoring channel activity to the baseline level ( Figures 5A and 6A).
The decrease in P o in the presence of WT CSQ1 and alanine substituted mutants CSQ1 DA1 -CSQ1 DA3 is due to a significant decrease in mean open time (T o ) ( Figure 6B; Table 1). Similar decreases in (T o ) were observed upon reassociating the CSQ1 alanine mutants ( Figure 6B; Table 1). There were also significant increases in mean closed times (T c ) of the channels, in line with those recorded in the presence of WT CSQ1 ( Figure 6C; Table 1). As a result of the opposing changes in T o and T c , there was no change in the overall frequency of opening ( Figure 6D).
In contrast to substitution within the final five Cterminal residues (CSQ1 DA3 ), deletion of the final seven C-terminal acidic residues reduced CSQ1's ability to inhibit RyR1 channels ( Figures 5C and 6A,B,C,D; Table 1). While the CSQ1 ΔC7 mutant inhibited channels, the inhibition was significantly less than the potent inhibition caused by WT CSQ1 and CSQ1 DA1 -CSQ1 DA3 . The decrease in P o was accompanied by significant changes in T o and T c (Figure 6B,C; Table 1). This data correlates with the reduced ability of this construct to induce Ca 2+ compaction ( Figure 3) and to bind to junctin (Figure 4), which communicates CSQ1 signaling to RyR1 [39]. Removal of the entire C-terminal tail abolished CSQ1's ability to inhibit the RyR1 channels ( Figures 5D and 6A,B,C,D; Table 1). Indeed, CSQ1 ΔC14 evoked a small but significant activation of RyR1. This activation is similar to the activation seen when CSQ1 is added to purified RyR1, that is, only seen when trisk95 and junctin are absent [7,43] and indicative of a direct, but probably non-physiological, interaction between CSQ1 and RyR1.
To investigate the nature of the small RyR1 activation by CSQ1 ΔC14 , we examined its effect on purified RyR1 (Figure 7). There was no contaminant trisk95, junctin, or CSQ1 in this sample ( Figure 7D), so that any effects of CSQ1 ΔC14 would be due to its association directly with RyR1, rather than via the anchoring proteins. Purified RyR1 activity rose significantly approximately 1.9- fold after the trans addition of 16 mg/ml CSQ1 ΔC14 ( Figure 7A,B), although the degree of purified RyR1 activation was not as high as the approximately 2.8fold activation induced by WT CSQ1 (Figure 7 and [7]). In addition, affinity chromatography shows that both the WT CSQ1 and CSQ1 ΔC14 bound to purified RyR1 ( Figure 7C). This indicates that although there was no difference in the ability of CSQ1 to bind directly to RyR1 in the presence or absence of the C-terminal tail, full activation of RyR1 by CSQ1 depends on binding to residues in both the C-terminal domain of CSQ1 and in the N-terminal domains. We have previously suggested that despite CSQ1 causing a modest activation of purified RyR1 in the bilayer, it is the overwhelming inhibition of native RyR1 (through the CSQ1-junctin-RyR1 interaction; [39]) which drives the overall effect of CSQ1 in the cell, which would acts as a brake on SR Ca 2+ release.

Overview
Here, we present evidence that the C-terminal tail of CSQ1 forms the primary binding pocket for Ca 2+ ions and that it contains the CSQ1 binding sites for trisk95 and junctin and for functional interactions with RyR1. This is also the first evaluation of the influence of specific acidic residues on the functional characteristics of CSQ1. Our novel data indicates the extent to which specific acidic residues in CSQ1's C-terminal tail influence the Ca 2+ binding capacity and Ca 2+ -induced aggregation of CSQ1, its ability to bind to trisk95 and junctin and its ability to inhibit RyR1. In summary, Ca 2+ binding declines with progressive removal of acidic residues, and all 14 acidic residues are required for the full Ca 2+ binding capacity of the wild-type protein. Deletion of one to three of the acidic residues has little effect on Ca 2+ -induced compaction or CSQ1 association with trisk95 and junctin and does not alter the inhibitory action on RyR1 of the CSQ1 binding to junctin. CSQ1 aggregation, binding of trisk95 and junctin to CSQ1, and the inhibitory effect of the binding on RyR1 activity are severely disrupted by deletion of seven of the acidic residues and are all but abolished by deletion of all 14 C-terminal residues.

CSQ Ca 2+ binding capacity
The Ca 2+ binding capacity of CSQ1 decreased when two or more residues within the acidic C-terminal tail were modified or deleted. CSQ1's moderate Ca 2+ binding capacity has been thought to depend on CSQ1's polymerization which occurs at relatively high [Ca 2+ ] (>0.5 mM). This model is based on the published crystal structure of rabbit skeletal CSQ1 but is not supported by the crystal structure as residues 352 to 367 are not observed in the electron density mapping (reviewed in [33]). CSQ1 dimers and polymers condense at [Ca 2+ ]s between 100 μM and 1 mM with a cellular ionic strength of 150 mM [6]. The Ca 2+ dependence of CSQ1's structure suggests that Ca 2+ may act as a "glue", interacting with key acidic residues when a polymer is formed. It is further hypothesized that residues within the C-terminus form a Ca 2+ binding sink within the structured protein which is stabilized by the formation of three salt bridges ( 215 E-86 K, 216 E-24 K, and 169 E-85 K) [27]. The linear sequence of full-length CSQ1 cannot account for its reported Ca 2+ binding capacity, of 34 Ca 2+ ions per monomer at resting (1 mM) [Ca 2+ ] [4]. There are insufficient acidic residues on CSQ1 to form the 34 acidic doublets, assuming that one acidic double is required to bind each Ca 2+ ion [44]. The Ca 2+ sink hypothesis was recently refined, with simulation model data showing that Ca 2+ preferentially saturates the C-terminal tail and induces protein polymerization which stabilizes the formation of many Ca 2+ binding sites on CSQ1's surface [32] which and could rationally account for the reported Ca 2+ binding capacity. Our data provides the first experimental evidence supporting this refined hypothesis. Furthermore, the mutation of three acidic residues in the C-terminal tail could only account for the loss of 3 of the 34 Ca 2+ ions bound per monomer of CSQ1, yet Ca 2+ binding data illustrates an approximately 45% loss of Ca 2+ binding capacity. One interpretation of our data is that these residues are important in stabilization of the Ca 2+ binding pocket formed upon CSQ1 polymerization in which the ratio of Ca 2+ ions to acidic doublets residues is ≫ 1. This is supported by data showing CSQ1 structure and Ca 2+ binding are drastically altered when the whole C-terminal tail is deleted (Figure 2 and [15,29]) and the inability of CSQ1 ΔC14 to form polymers (Figure 3) under our experimental conditions. CSQ1 is reported to be the key luminal Ca 2+ sensor for the RyR1 and to guide RyR1's response to the fluctuations in luminal [Ca 2+ ] that occur during contraction and relaxation [6,8,34]. CSQ1 binds and releases Ca 2+ in response to store load and to maintain free Ca 2+ at approximately 1 mM and allows CSQ1 to sense the level of Ca 2+ inside the SR. Given that mutation and deletion of the C-terminus has a profound effect on Ca 2+ binding capacity, it is likely that the C-terminal tail is important in communicating luminal [Ca 2+ ] to RyR1.

Loss of CSQ1 binding to trisk95 and junctin
It was not surprising that the absence of the C-terminal tail impeded CSQ1-trisk95 association as the CSQ1 binding site for trisk95 has previously been localized to residues within the C-terminal tail [16]. Further dissection illustrates that residues 363, 365, and 367 do not play a pivotal role in this association, although comparison of binding of CSQ1 DA2 and CSQ1 DA3 do suggest a minor role for 363 D. It appears that residues 354 to 362 are essential in forming CSQ1's trisk95 binding site.
It is significant that a modest amount of junctin was able to bind to CSQ1 ΔC14 , given that others have reported that the C-terminal domain of CSQ1 is the sole binding site for junctin in skeletal muscle [15]. This is consistent with the fact that more than one binding site on CSQ1 may be involved in its association with junctin in skeletal muscle, as with cardiac CSQ2 [13]. It is likely that a binding motif, in N-terminal domains of CSQ1, contributes to junctin association and would account the fraction of junctin binding remaining with CSQ1 ΔC7 and CSQ1 ΔC14 .

Loss of CSQ1 regulatory function of RyR1 channel activity
The significant decrease in CSQ1's ability to inhibit RyR1 activity following deletion of 361 to 367, but that substitution of residues 363, 365, and 367 maintain normal RyR1 regulation implies that residues 361, 362, 364, and 366 contribute to maintaining the efficiency of CSQ1's regulation of RyR1 [9]. The abolition of RyR1 inhibition by CSQ1 when the C-terminal tail was removed confirms both that this tail is essential in CSQ1's role in inhibition of Ca 2+ release through RyR1 and an important role for residues 354 to 360. CSQ1 inhibition of RyR1 activity is mediated via a RyR1/junctin/CSQ1 interaction, and not through RyR1/trisk95/CSQ1 [39]. Therefore, it is curious that CSQ1 inhibition of the channel is lost despite the remaining modest association of junctin with CSQ1 ΔC14 . It is possible that CSQ1's function as a channel inhibitor is directly related to its Ca 2+ binding capacity and/or structure, which is severely reduced when the C-terminal tail is removed from CSQ1 (Figures 2 and 3).
CSQ1 activation of RyR1 -not a physiological regulatory mechanism?
CSQ1 ΔC14 increases the activity of native RyR1 (Figure 6), and both WT CSQ1 [7,43] and CSQ1 ΔC14 activate purified RyR1 (Figure 7). It is likely that this activation is relatively unimportant physiologically, as full-length CSQ1 addition to RyR1 in the presence of anchors trisk95 and junctin results in a strong channel inhibition. In light of this, native RyR1 activation by CSQ1 ΔC14 can be explained in either of two ways. Firstly, WT CSQ1 associated with junctin exerts a dominant inhibitory effect on RyR1 gating, so that channel inhibition overshadows activation arising from the direct RyR1-CSQ1 association. When the binding site on CSQ1 for junctin is compromised in CSQ1 ΔC14 , channel activation by the direct CSQ1-RyR1 coupling is unmasked. On the other hand, the binding site for WT CSQ1 on RyR1 may be occluded by CSQ1's interaction with trisk95 and/or junctin. Thus, the reduced CSQ1 ΔC14 association with trisk95 and/or junctin may expose the binding site on RyR1 for CSQ1, allowing channel activation when CSQ1 ΔC14 is added to the trans chamber.

Conclusions
CSQ1 is unarguably the major the Ca 2+ buffer inside the SR. In addition, CSQ1 is both an inhibitor of native RyR1 under resting conditions and also a luminal Ca 2+ sensor for the channel, so that it may act as a brake on RyR1 Ca 2+ release in times of low store load [7,9]. As mentioned above, CSQ1 has been thought to regulate RyR1 via interactions with trisk95 and junctin. Our previous single channel data suggests that junctin, but not trisk95, is the key intermediate protein in RyR1 inhibition [39], although evidence from one trisk95 and junctin knockout study suggests that trisk95 is the preferential anchor for CSQ1 [45]. On the other hand, studies using myotubes deficient of trisk95 and junctin confirm our observations that junctin's primary role is in communication of luminal Ca 2+ to RyR1 [46]. It remains possible that anchoring and functional regulation are in fact separate processes. Regardless, CSQ1 undisputedly plays an important role in regulating RyR1 and in excitation-contraction coupling. The essential nature of CSQ1 is illustrated in CSQ1-null mice, which exhibit increased susceptibility to stress in a similar manner to exertional/environmental heat stroke and to human malignant hyperthermia (MH) a life-threatening hypermetabolic disorder induced by treatment with volatile anesthetics and the muscle relaxant succinylcholine [10,47,48]. This complex phenotype is likely initiated by abnormal Ca 2+ handling by the SR, due to lack of CSQ1 inhibition on RyR1 activity [47].