Different Regions in Skeletal and Cardiac Muscle Ryanodine Receptors Are Involved in Transducing the Functional Effects of Calmodulin*

Calmodulin (CaM) inhibits the skeletal muscle ryanodine receptor-1 (RyR1) and cardiac muscle RyR2 at micromolar Ca 2 (cid:1) but activates RyR1 and inhibits RyR2 at submicromolar Ca 2 (cid:1) by binding to a single, highly conserved CaM-binding site. To identify regions responsible for the differential regulation of RyR1 and RyR2 by CaM, we generated chimeras encompassing and flanking the CaM-binding domain. We found that the exchange of the N- and C-terminal flanking regions differentially affected RyR1 and RyR2. A RyR1/RyR2 chimera with an N-terminal flanking RyR2 substitution (RyR2 amino acid (aa) 3537–3579) was activated by CaM in single channel measurements at both submicromolar and micromolar Ca 2 (cid:1) . A RyR2/RyR1 chimera with a C-terminal flanking the 86-amino acid RyR1 substitution (RyR1 aa 3640–3725) bound 35 S-CaM but was not inhibited by CaM at submicromolar Ca 2 (cid:1) . In this region, five non-conserved amino acid residues (RyR1 aa 3680 and 3682–3685 and RyR2 aa 3647 and 3649–3652) differentially affect RyR helical probability. Substitution of the five amino acid residues in RyR1 with those of RyR2 showed responses to CaM comparable with wild type RyR1. In contrast, substitution of the five amino acid residues in RyR2 with those of RyR1 showed loss of CaM inhibition, whereas substitution of the five RyR2 sequence residues in the RyR2 chimera containing the RyR1 calmodulin-binding domain and C-flanking sequence restored wild type RyR2 inhibition by CaM at submicromolar Ca 2 (cid:1) . The results suggest that different regions are involved in CaM modulation of RyR1 and RyR2. They further suggest that five non-conserved amino acids in the C-terminal region flanking the CaM-binding domain have a key role in CaM inhibition of RyR2.

Ca 2ϩ -release channels, known as ryanodine receptors (RyR), 1 play a key role in a variety of cells by increasing the cytoplasmic Ca 2ϩ concentration in response to cellular signals. Three different RyR genes have been identified in mammalian cells: skeletal type (RyR1), cardiac type (RyR2), and brain type (RyR3) (1)(2)(3). The RyRs form large protein complexes composed of four 565-kDa subunits and four small FK506-binding proteins. Multiple small physiological molecules such as Ca 2ϩ , Mg 2ϩ , and ATP and proteins such as protein kinases and calmodulin (CaM) modulate RyR ion channel function (1)(2)(3).
CaM is a small cytosolic Ca 2ϩ -binding protein that regulates various cellular functions (4,5). Direct CaM binding inhibits all three RyR isoforms at [Ca 2ϩ ] Ͼ 1 M, whereas at [Ca 2ϩ ] Ͻ 1 M RyR1 and RyR3 are activated, but RyR2 is inhibited by CaM (6,7). 35 S-CaM binding to sarcoplasmic reticulum vesicles (8,9) and purified RyR1 and RyR2 (10) showed that the two receptor isoforms bind one molecule of CaM per RyR subunit in the absence and presence of Ca 2ϩ . CaM protection of trypsin digestion of RyR1 (11) and site-directed mutagenesis of RyR1 (12) and RyR2 (13) demonstrated that the two RyRs have a single CaM regulatory binding domain (CaMBD) (aa 3614 -3643 in RyR1) for the Ca 2ϩ -free and Ca 2ϩ -bound (Ca 2ϩ -CaM) forms of CaM. The CaM-binding domain is highly conserved among the mammalian isoforms, yet corresponding mutations in the RyR1 and RyR2 CaM-binding domain resulted in a different response to CaM in 35 S-CaM binding and functional measurements (13). This result suggests that other isoform-specific regions are responsible for the differential modulation of RyR1 and RyR2 by CaM.
To test the hypothesis of an involvement of isoform-specific regions in regulating RyR1 and RyR2 by CaM, we constructed a series of RyR1/RyR2 chimeras and determined their CaMbinding properties and modulation by CaM. The results suggest that multiple regions are involved in transducing the functional effects of CaM in RyR1 and RyR2. The present study focused on the role of one of these regions. We found that substitution of five amino acid residues in RyR2 with those of RyR1 eliminated CaM inhibition but not CaM binding at [Ca 2ϩ ] Ͻ 1 M.

Materials-[ 3 H]Ryanodine was obtained from PerkinElmer
Life Sciences, Tran 35 S-label from ICN Radiochemicals (Costa Mesa, CA), unlabeled ryanodine from Calbiochem, Complete protease inhibitors from Roche Applied Science, and HEK293 cells from the American Type Culture Collection. RyR2 cDNA was kindly provided by Dr. Junichi Nakai of RIKEN Brain Science Institute (Wako, Japan). Unlabeled CaM and 35 S-CaM were prepared as described previously (10).
Construction of RyR cDNAs-The full-length rabbit RyR1 (14) and RyR2 (15) cDNAs were subcloned into expression vectors pCMV5 and pCIneo, respectively. Single and multiple base changes were introduced by Pfu-turbo polymerase-based chain reaction, using mutagenic oligonucleotides and the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The complete mutated DNA fragments amplified by PCR were confirmed by DNA sequencing. A schematic summary and the amino acid numbers of RyR chimeras are shown in Fig. 1.
The chimeric ClaI/SacII fragment was subcloned back to the original position of a vector containing the BbrPI/SacII-(5038 -11203) fragment of RyR2 and then to full-length RyR2 in the pCIneo expression vector to construct the full-length RyR2-based chimera expression plasmids. The chimera EclXI/NdeI or EclXI/BamHI fragment was subcloned back to a vector containing a PvuI/NdeI-(8469 -11173) fragment of RyR1 and then to a vector containing PvuI/XbaI-(8469-polylinker) fragment. Fulllength expression plasmids for RyR1-based chimeras were prepared by ligation of two fragments (ClaI/PvuI and PvuI/XbaI) and expression vector pCMV5 (ClaI/XbaI).
To construct RyR1-10M, a vector containing the PinAI/XhoI-(4935-6467) fragment of RyR1 was used as a template for mutagenesis. The mutated fragment was subcloned back to a vector containing ClaI/PvuI-(polylinker-8469), and then a full-length expression plasmid was prepared by 3Ј fragment ligation as described above. Nucleotide numbering is as described previously (15,16).
Expression of Full-length RyRs in HEK293 Cells-RyR cDNAs were transiently expressed in HEK293 cells with FuGENE 6 (Roche Applied Science) according to the manufacturer's instruction. Cells were maintained at 37°C and 5% CO 2 in high glucose Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and were plated the day before transfection. For each 10-cm tissue culture dish, 3.5 g of cDNA was used. Cells were harvested 48 h after transfection, and crude membrane fractions were prepared as described previously (12). Single Channel Analysis-Single channel measurements of WT and mutant RyRs were performed in planar lipid bilayers containing phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine in the ratio of 5:3:2 (25 mg of total phospholipid/ml of n-decane) as described previously (13). Crude membrane fractions expressing WT and mutant RyRs were pretreated for 30 min at room temperature with 1 M myosin light chain kinase-derived CaM-binding peptide to dissociate endogenous CaM (10). Final peptide concentration was Ͻ10 nM following the addition of membranes to the cis (cytosolic) chamber of the bilayer apparatus. A strong dependence of single channel activities on cis-Ca 2ϩ concentration indicated that the large cytosolic "foot" region faced the cis chamber of the bilayers. The trans (lumenal) side of the bilayer was defined as ground. Measurements were made with symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4, with 0.4 M or 2 M free Ca 2ϩ in the cis chamber. Single channel recordings with WT-RyR1 and chimeras were made in the presence of 0.4 M Ca 2ϩ and 1 mM ATP to increase the otherwise very low channel activities. Exogenous CaM was added to the cis solution. Electrical signals were filtered at 2 kHz, digitized at 10 kHz, and analyzed as described previously (17). P o values in multichannel recordings were calculated using the equation P o ϭ ⌺ iP o,i /N, where N is the total number of channels, and P o,i is channel open probability of the ith channel.
Biochemical Assays and Data Analysis-Free Ca 2ϩ concentrations were obtained by including in the solutions the appropriate amounts of Ca 2ϩ and EGTA as determined using the stability constants and computer program published by Shoenmakers et al. (18). Free Ca 2ϩ con- centrations of Ն1 M were verified with the use of a Ca 2ϩ -selective electrode. Results are given as mean Ϯ S.E. Significance of differences in the data (p Ͻ 0.05) was determined using Student's t test.

RESULTS
To identify the isoform-specific regions involved in CaM modulation, we prepared 10 RyR1-and RyR2-based chimeras and multisite mutants (Fig. 1B) and transiently expressed them in HEK293 cells. All chimeras and mutants showed a Ca 2ϩ dependence of [ 3 H]ryanodine binding corresponding to that of WT-RyR1 or WT-RyR2 (19).  (Fig.  2), which were in good agreement with values reported previously (13). Fig. 2 also shows that all RyR2-based chimeras and mutants (Fig. 1B, left panel) examined in this study bound 35 S-CaM. Some differences in the binding levels were observed; however, none of these were significantly different from WT-RyR2. The expression levels of RyR1-based chimeras and mutants (Fig. 1B, right panel) were too low to determine their CaM binding levels. The To identify the region that was responsible for loss of a significant effect of calmodulin on 2-1-2 at 0.4 M Ca 2ϩ , the chimeric region was subdivided into an N-terminal flanking (2-1-2(N)), CaM-binding, and C-terminal flanking (2-1-2(C)) region. As described in a previous study (13), substitution of the four non-identical amino acids in the RyR2 CaM-binding domain (Fig. 1A) with the corresponding amino acids in RyR1 yielded results essentially identical to WT-RyR2. Replacement of the N-terminal flanking region (2-1-2(N)) also did not alter CaM inhibition at 0.4 M and 2 M Ca 2ϩ (Fig. 1B). In contrast, exchange of the C-terminal flanking region (2-1-2(C)) eliminated CaM inhibition at 0.4 M Ca 2ϩ and significant inhibition at 2 M Ca 2ϩ by 50 nM (but not by 1 M) CaM, notwithstanding that 2-1-2(C) bound 35 S-CaM with an affinity comparable with WT-RyR2.
The N-and C-terminal flanking regions of RyR1 and RyR2 have a sequence identity of 41 and 63%, respectively, compared with 86% for the CaMBD. Secondary structure predictions, using a protein sequence analysis program (bmerc-www. bu.edu/psa/) (20) indicated a nearly indistinguishable helical probability for the RyR1 and RyR2 CaMBDs (Fig. 4A). In contrast, major differences were predicted for both the N-flanking and C-flanking regions. The differences in helical probability of N-terminal flanking regions are caused to a large extent by an RyR2-specific insert (RyR2 aa 3564 -3575). We previously showed that an RyR2 mutant lacking this insert bound 35 S-CaM and was inhibited by CaM in a manner essentially identical to WT-RyR2 (13). Therefore, this region does not appear to have a major role in CaM modulation of RyR2. In the Cterminal flanking region, protein sequence analysis predicted several differences in helical probability. In the present study, we address the role of five non-conserved RyR amino acid residues (RyR2 aa 3647 and 3649 -3652) that have a profound effect on helical probability. Substitution of these five amino acid residues with those of RyR1 in RyR2 (RyR2-5M) increased helical probability close to that in RyR1 (Fig. 4B). Single chan-  1B and 5, B and C). The difference in the ability of the five RyR2 amino acids to restore CaM inhibition was likely the result of the presence of the N-flanking RyR1 region in 2-1-2, because this region contributes to CaM modulation of RyR1 as described below. At 2 M Ca 2ϩ , RyR2-5M, 2-1-2rev5M, and 2-1-2(C)rev5M were inhibited by CaM (Figs. 1B and 5); however, inhibition was only significant at 1 M CaM and not 50 nM CaM (Fig. 1B), which indicated a decreased efficacy of CaM in modulating the mutant channels.
CaM Modulation of RyR1-based Chimeras and Mutants-We also constructed and expressed four RyR1-based reverse chimeras and mutants, 1-2-1, 1-2-1(N), 1-2-1(C), and RyR1-5M. Characterization of RyR1-based chimeras relied on single channel recordings because their expression levels were too low to determine 35 S-CaM binding levels (see above). CaM failed to either activate or inhibit the 1-2-1 chimera (Fig. 1B). Replacement of only the N-terminal flanking region decreased the efficacy of CaM activation at 0.4 M Ca 2ϩ and, surprisingly, at 2 M Ca 2ϩ resulted in significant activation of 1-2-1(N) following the addition of 1 M CaM. In contrast, replacement of the C-terminal region failed to disrupt CaM modulation of 1-2-1(C) at either Ca 2ϩ concentration. Substitution of the five amino acid residues in RyR1 with those of RyR2 decreased helical probability (Fig. 4B) without affecting CaM activation and inhibition of RyR1-5M (Fig. 1B), which was consistent with the results of 1-2-1(C). Interestingly, at 0. are 10 non-identical RyR1 and RyR2 amino acid residues in this region. We substituted the 10 aa residues in RyR1 with the corresponding RyR2 residues. The RyR1-10M mutant was activated and inhibited by CaM in single channel (Fig. 6) and [ 3 H]ryanodine binding (not shown) measurements that were essentially identical to WT-RyR1, which suggests that this region is not responsible for the differential regulation of RyR1 and RyR2 by CaM at [Ca 2ϩ ] Ͻ 1 M.

DISCUSSION
Cryomicroscopy studies revealed that the CaM-binding site in the skeletal muscle RyR isoform is at least 10 nm away from the effector site (pore) and that Ca 2ϩ binding causes an ϳ33-Å shift of the binding site (22). Therefore, long range conformational changes are likely involved in the modulation of RyR activity by CaM. RyR1 and RyR2 each have a single highly conserved CaM-binding region; however, the two RyR isoforms are differentially regulated by CaM, which suggests that isoform-specific regions are involved in transducing the functional effects of CaM. In the present study, we identified one of these regions. Secondary structure predictions indicated a large difference between the helical probability of five non-conserved amino acid residues in the C-terminal region flanking the CaMbinding domain of RyR1 and RyR2. Substitution of the five amino acid residues in RyR2 with those of RyR1 eliminated CaM inhibition at submicromolar Ca 2ϩ and decreased the effi-cacy of CaM inhibition at micromolar Ca 2ϩ . Conversely, substitution of the corresponding RyR2 amino acid residues in RyR1 was without effect on CaM activation and CaM inhibition at Ca 2ϩ concentrations below and above 1 M, respectively. Thus, five amino acid residues in the C-terminal region flanking the CaM-binding domain appear to have a specific role in CaM modulation of RyR2.
The RyR2-based chimeras were characterized in 35 S-CaM binding and single channel measurements. Characterization of RyR1-based chimeras relied on single channel measurements because expression levels were too low to obtain reliable 35 S-CaM binding data. Although all RyR2-based constructs bound 35 S-CaM, regardless of whether the constructs contained the RyR1 or RyR2 CaMBD, 2-1-2, 2-1-2(C), and RyR2-5M were unable to convert CaM binding into a functional signal. Interestingly, loss of function was confined to [Ca 2ϩ ] Ͻ 1 M because CaM inhibition was observed at [Ca 2ϩ ] Ͼ 1 M. We also note that the "uncoupled" constructs had a K ϩ conductance and Ca 2ϩ dependence of [ 3 H]ryanodine binding comparable with WT-RyR2. The results suggest disruption of the interactions of the CaMBD with protein domains that transduce the functional effects of CaM in RyR2 at [Ca 2ϩ ] Ͻ 1 M.
Ikemoto and Yamamoto (23) have shown that exogenously added RyR1 synthetic peptides alter channel activity by "unzipping" interdomain interactions in the large RyR1 channel complex. RyR1 mutations giving rise to malignant hyperthermia were suggested to result in elevated channel activities by disrupting interdomain interactions. Zhu et al. (24) found that synthetic peptides corresponding to the CaMBD affect RyR1 channel function, which suggested that the exogenously added peptides bound to a region that directly interacts with the CaMBD. Previously, we presented evidence for the importance of three amino acid residues in this interaction in RyR2. Three mutations in the RyR2 CaMBD (W3587A, L3591D, and F3603A) eliminated CaM inhibition at 0.4 M Ca 2ϩ , without an effect on CaM binding (13). The present work extends the previous studies by providing the first clues regarding the identity of regions that transduce the functional effects of CaM in RyR1 and RyR2 Analysis revealed that substitution of five non-conserved RyR1 amino acid residues in 2-1-2(C) restored CaM inhibition at [Ca 2ϩ ] Ͻ 1 M. On the other hand, substitution of the five RyR1 residues in 2-1-2 did not restore CaM inhibition. Thus, the additional presence of the N-terminal region likely introduced conformational changes that could not be reversed by replacing the five amino acid residues in 2-1-2rev5M. We also tested the role of the two regions flanking the CaMBD in RyR1. Chimera 1-2-1 was not modulated by CaM, in contrast to chimera 2-1-2, which was inhibited at [Ca 2ϩ ] Ͼ 1 M. Further analysis showed that exchange of only the C-terminal flanking region did not eliminate CaM activation and CaM inhibition in the RyR1-based chimera 1-2-1(C), whereas CaM inhibition of 2-1-2(C) was disrupted at [Ca 2ϩ ] Ͻ 1 M. Opposite effects were observed when the N-terminal flanking region was exchanged in RyR1 and RyR2. Replacement of this region did not interfere with CaM inhibition in 2-1-2(N). The efficacy of CaM activation of 1-2-1(N) was reduced at 0.4 M Ca 2ϩ and unexpectedly resulted in CaM activation of 1-2-1(N) at 2 M Ca 2ϩ . Further studies will be required to determine whether the structural elements responsible for activation of 1-2-1(N) lie within the N-terminal flanking region.
Zhang et al. (21) suggested that the carboxyl-terminal half of CaM binds to the RyR1 CaM-binding domain, whereas the N-terminal half interacts with RyR1 amino acid residues 1975-1999. We substituted the 10 non-identical amino acid residues in RyR1 with the corresponding RyR2 residues and found that the mutant was regulated by CaM essentially identical to WT-RyR1, which suggests that this region is not responsible for the differential regulation of RyR1 and RyR2 by CaM at [Ca 2ϩ ] Ͻ 1 M.
Our data show that the 1-2-1 chimera was not inhibited by CaM, notwithstanding that the chimera contained the RyR2 CaMBD and N-and C-terminal flanking regions. Chimera 2-1-2 was not significantly activated at [Ca 2ϩ ] Ͻ 1 M Ca 2ϩ , despite the presence of the RyR1 CaMBD and flanking regions. These results suggest that additional regions are involved in the CaM modulation of RyR1 and RyR2, as expected considering the large size of the RyRs.
In summary, our structure-function analysis shows that it is possible to use chimeras and multiple site mutants to define the Ca 2ϩ -dependent and isoform-dependent regulation of RyR1 and RyR2 by CaM. The chimeras described here formed channels with a Ca 2ϩ dependence and K ϩ conductance comparable with WT-RyRs, which suggests that our structural manipulations did not introduce major global conformational changes, and thus can reveal regions that have a role in the Ca 2ϩ -dependent and isoform-dependent CaM modulation of RyR1 and RyR2. Our work points to one of these regions consisting of five amino acid residues in the C-terminal flanking region of the CaMBD of the RyR2 ion channel.