A 37-amino acid sequence in the skeletal muscle ryanodine receptor interacts with the cytoplasmic loop between domains II and III in the skeletal muscle dihydropyridine receptor.

Interactions between the Ca2+ release channel of skeletal muscle sarcoplasmic reticulum (ryanodine receptor or RyR1) and the loop linking domains II and III (II-III loop) of the skeletal muscle L-type Ca2+ channel (dihydropyridine receptor or DHPR) are critical for excitation-contraction coupling in skeletal muscle. The DHPR II-III loop was fused to glutathione S-transferase- or His-peptide and used as a protein affinity column for 35S-labeled in vitro translated fragments from the N-terminal three-fourths of RyR1. RyR1 residues Leu922-Asp1112 bound specifically to the DHPR II-III loop column, but the corresponding fragment from the cardiac ryanodine receptor (RyR2) did not. The use of chimeras between RyR1 and RyR2 localized the interaction to 37 amino acids, Arg1076-Asp1112, in RyR1. The RyR1 922-1112 fragment did not bind to the cardiac DHPR II-III loop but did bind to the skeletal muscle Na+ channel II-III loop. The skeletal DHPR II-III loop double mutant K677E/K682E lost most of its capacity to interact with RyR1, suggesting that two positively charged residues are important in the interaction between RyR and DHPR.

E-C coupling 1 linking electrical stimulation of skeletal muscle and the release of Ca 2ϩ from the sarcoplasmic reticulum to muscle contraction almost certainly involves a direct interaction between the Ca 2ϩ release channel of skeletal muscle sarcoplasmic reticulum (the ryanodine receptor or RyR1) and the L-type Ca 2ϩ channel of skeletal muscle (the dihydropyridine receptor or DHPR) located in the transverse tubule (1)(2)(3)(4). Extracellular Ca 2ϩ is not necessary for E-C coupling in skeletal muscle but is required in cardiac muscle (5). Skeletal muscletype E-C coupling was rescued in RYR1 knockout mice by injection with RYR1 cDNA (6) and in myotubes cultured from dysgenic mouse skeletal muscle, in which functional DHPR and E-C coupling are lacking, by injection with cDNA encoding skeletal DHPR (7,8). Tanabe et al. (8) used chimeras between skeletal and cardiac DHPRs to define a region linking DHPR domains II and III as the determinant of skeletal versus cardiac type E-C coupling. The DHPR II-III loop activates RyR1 in single channel recordings (9,10). Ca 2ϩ release from triads has also been elicited by the addition of the DHPR II-III loop peptide (11).
Marty et al. (12) co-immunoprecipitated RyR1 and DHPR from muscle triads with antibodies against one or the other of the two proteins. RyR1 and the ␣1-subunit of the DHPR in triads were cross-linked with dithiobis(succinimidyl propionate) (13). Physical interactions between RyR1 and DHPR have been inferred from functional interactions between these proteins (9 -11, 14).
In this study, we purified the DHPR II-III loop following its expression in Escherichia coli. RyR1 fragments were prepared from a cell-free in vitro translation system. The purified DHPR II-III loop was bound to a gel matrix to permit affinity chromatography, and 35 S-labeled RyR1 fragments were then passed through the column to demonstrate direct interaction between specific, expressed fragments of the two proteins. This strategy allowed us to identify a 37-amino acid sequence of RyR1 that interacts specifically with the DHPR II-III loop.

MATERIALS AND METHODS
Chemicals and Reagents-NiNTA resin was from Qiagen and glutathione-Sepharose 4B from Pharmacia Biotech Inc. Translational grade [ 35 S]Met was obtained from Amersham Life Science. The coupled in vitro transcription and translation kit (TNT Quick) was from Promega. Rabbit skeletal and cardiac muscle cDNA was purified from phage cDNA libraries (15,16) using mediprep columns from Qiagen. The rSkM1 rat skeletal muscle Na ϩ channel cDNA (17) was a generous gift from Dr. P. Backx (University of Toronto). The full-length ryanodine receptor cDNA clone, pBS SRR10, was described previously (18).
Preparation of Fusion Proteins-The loops linking domains II and III were amplified and cloned using the polymerase chain reaction (PCR): nucleotides 1990 -2381 of the cDNA fragment encoding the rabbit skeletal muscle DHPR II-III loop with the skeletal muscle cDNA library as template; nucleotides 2374 -2772 of the cardiac DHPR II-III loop with the cardiac muscle cDNA library as template; nucleotides 2389 -3059 of the rat skeletal muscle Na ϩ channel with rSkM1, the rat skeletal muscle Na ϩ channel cDNA (17) as template. Oligonucleotide primers flanking the cDNA sequence for the respective loops were designed with exterior BamHI and EcoRI restriction endonuclease sites for in-frame cloning into the ptrcHisC vector (Invitrogen) or the pGEX 3X vector (Pharmacia). All cloned fragments were then verified by DNA sequence analysis. The oligonucleotide attgaattccaccaccaccaccaccaccaccaccaccacaagcttgaattcata and its complementary oligonucleotide were used to add 10 His residues to the C-terminal end of glutathione S-transferase (GST) by self-annealing of the two oligonucleotides, endonuclease restriction digestion at the EcoRI sites flanking the His 10 sequence, and ligation into the EcoRI site of pGEX3X.
E. coli strain DH5␣ (Life Technologies, Inc.) was used for expression. Proteins were purified following standard procedures (19) in the pres-ence of 20 mM imidazole, pH 7.0, and protease inhibitors. His-peptide (Invitrogen) fusion proteins were purified with NiNTA resin, and GST fusion proteins were purified with glutathione-Sepharose 4B. After washing, the His-peptide fusion proteins were eluted with 0.5 M imidazole, pH 7.0, in phosphate-buffered saline (PBS). GST fusion proteins were eluted with 10 mM reduced glutathione and dialyzed against PBS overnight. Eluted proteins were analyzed by SDS-PAGE and Coomassie Blue staining, and yield was determined by the Bradford assay (Bio-Rad).
Preparation of Ryanodine Receptor Fragments-Fragments for in vitro transcription and translation were prepared from pBS SRR10 (18). pBS SRR10 was digested with EcoRI (2394) and XbaI (MCS), blunt ended with Klenow, and self-ligated to make pBS RYR1 F1. PstI restriction endonuclease fragments were subcloned into pBSKSϩ ATG-PstI. pBSSKϩ ATG-PstI was prepared from pBSSKϩ (Stratagene) by ligation of the linker ataccaccatgggcctgcaggcccatggtggtat into the SmaI site, digestion with PstI, and religation to circularize the vector. The PstI site is downstream of a Kozak consensus start sequence (20) and an ATG for translation initiation. The PstI fragments were subcloned into the PstI site of pBSKϩ ATG-PstI, in-frame with the initiator Met, following restriction endonuclease digestion of pBS RYR1 with PstI (see Fig. 1A). The PstI fragments that were subcloned and expressed were:  (20) and in-frame with an initiator Met to form pBS RYR1 F9 and F10b. The AatII (9468)-AatII (10523) fragment from RYRI was cloned into the AatII site of pBS RYR1 F10b to form pBS RYR F10. pBS RYR1 F3a was made by digestion of pBS RYR1 F3 with AatII(3335) and EcoRI (MCS), blunt end formation with Klenow, and religation. The pBS RYR2 F3a fragment, from nucleotides 2797 to 3378, was amplified by PCR from RYR2 cDNA (16) using the oligonucleotide primers ctgcagatgtcacttgagaccttgaag and atagaattcgatccggctgacaacctg, which include PstI and EcoRI restriction sites, respectively, for cloning into pBS-ATG-PstI.
Chimeras between RyR1 and RyR2 were made using pBS RYR1 F3a and pBS RYR2 F3a through the introduction of restriction endonuclease sites by PCR-based mutagenesis (QuikChange kit, Stratagene) (see Fig. 2A Binding Assay-Affinity matrices were prepared by eluting Hispeptide fusion protein or GST-His 10 from NiNTA resin and binding them to 40 l of fresh NiNTA at a concentration of 0.5 mg of fusion protein/ml. Purified, dialyzed GST fusion proteins were bound to fresh glutathione-Sepharose to achieve the desired concentration of fusion protein (mg/ml of Sepharose). The amount of protein bound to affinity matrices was confirmed by eluting a sample of the protein-bound affinity matrix with SDS, separating the eluted protein by SDS-PAGE, and staining with Coomassie Blue. Each stained fusion protein band was quantitated by densitometry (Molecular Analyst, Bio-Rad). Resins to which fusion proteins were bound were washed with 1 ml of PBS, Proteins retained on the affinity columns were eluted with 100 l of SDS sample buffer. SDS-PAGE (15% gel) was used to analyze 1 l of the in vitro translation product (5% of the total column input) and 20 l of the eluate (20% of the total column eluate). The gels were fixed three times in 6 volumes of 10% methanol-12% acetic acid for 20 min each time. The radioactive signal was enhanced with Entensify (NEN Life Science Products), and the gels were dried and exposed to either autoradiogram film (BioMax AR, Kodak) or quantitated by a Molecular Imager (Bio-rad). Specific binding was defined as total binding less nonspecific binding to GST-His 10 columns or GST columns.

Identification of an RyR1
Sequence Interacting with the DHPR II-III Loop-We immobilized the DHPR II-III loop on a Ni ϩ column as a His-peptide fusion protein and passed 35 Slabeled in vitro translated fragments of RyR1 through it (Fig.  1). We found that 20 Ϯ 2.7% of in vitro translated fragment 922-1220 was retained on the DHPR II-III loop affinity column compared with less than 5% retention for any of the other fragments of RyR1 lying between amino acids 1 and 3724. (Fig.  1, B and C). Reduction of the RyR1 fragment to 190 amino acids, spanning residues 922-1112, increased the proportion of the in vitro translated fragment retained on DHPR II-III loop columns to 32 Ϯ 1.6% (Fig. 2). To test isoform specificity, we passed 35 S-labeled in vitro translated fragments of RYR2 over the affinity column. We did not detect any specific binding of In vitro translated fragments of RyR1 were passed through 0.5 mg/ml DHPR II-III His-peptide fusion protein columns as described under "Materials and Methods." A, schematic of RyR1 fragments cloned in-frame with a Kozak consensus start sequence (20) and an initiator ATG codon for in vitro translation. B, autoradiogram of in vitro translated RyR1 fragments (F1-F10) representing 5% of input and 20% of fragments eluted from GST-His 10 (lanes G) and skeletal DHPR (lanes D) affinity columns. C, percentage of specific binding of in vitro translated RyR1 fragments to the skeletal DHPR affinity columns, quantitated by densitometry and expressed as the means Ϯ S.E. from at least four separate experiments. Specific binding was defined as total binding to DHPR II-III His-peptide fusion protein columns, less nonspecific binding to GST-His 10 .
Because further reduction of the size of the RyR1 fragment led to loss of synthesis of the in vitro translated product, we reduced the size of the RyR1 sequence that binds to the DHPR II-III loop by making chimeras between RyR1 and RyR2, as illustrated in Fig. 2A. The RyR1(922-1075)-RyR2(1076 -1112) (C1) chimera did not bind to the II-III loop, but 24 Ϯ 2.7% of the RyR2(922-953)-RyR1(954 -1112) (C3) chimera and 28 Ϯ 2.6% of the RyR2(922-1075)-RyR1(1076 -1112) (C2) chimera were retained on a 0.5 mg/ml DHPR II-III loop affinity column (Fig.  2, B and C). We deduced that the 37-amino acid sequence between Arg 1076 and Asp 1112 of RyR1 is important for binding the DHPR II-III loop.
DHPR Specificity in the RyR1-DHPR Interaction-The chimeric construct (C3), which binds to the skeletal muscle DHPR II-III loop His-peptide fusion protein, did not bind above background to either GST columns or to a GST fusion protein constructed from the II-III loop of the cardiac DHPR receptor (cdDHPR in Fig. 3, A and B). A GST fusion protein affinity column with the II-III loop of the skeletal muscle DHPR II-III (skDHPR in Fig. 3, A and B) retained 12 Ϯ 0.4% of the input of the in vitro translated C3 chimera (Fig. 3, A and B). Surprisingly, 14 Ϯ 0.4% of the input of in vitro translated C3 chimera bound to a GST fusion protein affinity column containing the loop between domains II and III of the homologous skeletal muscle voltage-gated Na ϩ channel (skNa in Fig. 3, A and B).
The region of the skeletal muscle DHPR II-III loop important for activation of RyR1 has been localized to the amino acid sequence between positions 671 and 690 (11). A 14-amino acid sequence within the skeletal muscle DHPR activation region is KAKAEERKRRKMSR (Fig. 3C). The corresponding amino acid sequence in the skeletal muscle Na ϩ channel is RGKILSPKEI-ILSE and in the cardiac DHPR is KEEEEEKERKKLAR. The skeletal muscle DHPR and the Na ϩ channel II-III loop sequences contain positive charges at positions 1, 3, and 8 of the shared sequences, a small amino acid (Ala or Gly) at position 2, and Ser at position 13. By contrast, the cardiac DHPR sequence has negative charges at positions 2, 3, and 8 and an Ala at position 13. We made the double mutation K677E/K682E in the skeletal muscle DHPR II-III loop so that the mutated skeletal muscle DHPR sequence, KAEAEERERRKMSR, would more closely resemble the cardiac sequence in this region. This double mutation resulted in a 66 Ϯ 7.1% decrease in binding of the C3 chimera, from 12 Ϯ 0.4% binding to 4 Ϯ 0.7% binding. DISCUSSION In this study, we localized the binding site for the skeletal DHPR II-III loop to 190 amino acids between 922 and 1112 in in vitro transcribed and translated products of RyR1. Decreasing the size of the fragment resulted in loss of in vitro translation, and GST fusion proteins of fragments shorter than 190 amino acids were unstable, suggesting that this sequence might form a stable structural domain. RyR1, RyR2, and RyR3 contain four repeat sequences (15,16,21). The 922-1112 fragment of RyR1 contains part of repeat 1 and all of repeat 2. We made use of these repeat domain boundaries to create three RyR1-RyR2 chimeras (Fig. 2A). The 37-amino acid sequence between Arg 1076 and Asp 1112 in RyR1, which was retained in chimera C2 (and in C3 but not in C1) was necessary for the binding of about 25% of the input in vitro translation product to the DHPR II-III loop column. This sequence begins virtually at the end of repeat 2. Because the binding site lies outside of the repeat sequences, the repeats are not likely to play a role in the

FIG. 2. Affinity of chimeric RyR1-RyR2 constructs for the skeletal DHPR II-III loop.
A, chimeric RyR1-RyR2 constructs prepared as described under "Materials and Methods." The first line represents a region in RyR1 containing two repeat sequences and defines their respective boundaries. Amino acid numbering refers to RyR1 residues in the chimeric constructs. B, autoradiogram of in vitro translated RyR1-RyR2 chimeric fragments representing 5% of input and 20% of fragments eluted from 0.5 mg/ml GST-His 10 (lanes G) and skeletal DHPR II-III (lanes D) affinity columns. C, percentage of specific binding of in vitro translated RyR1-RyR2 chimeric fragments to the skeletal DHPR affinity columns, quantitated by densitometry and expressed as the means Ϯ S.E. from at least four separate experiments. Specific binding was defined as total binding to DHPR II-III His-peptide fusion protein columns, less nonspecific binding to GST-His 10 . cardiac DHPR II-III loop (cdDHPR); skeletal Na ϩ channel II-III loop (skNa). B, percentage of specific binding of in vitro translated chimeric fragment (C3) to the II-III loop affinity columns, quantitated by densitometry and expressed as the means Ϯ S.E. from at least four separate experiments. Specific binding was defined as total binding to GST fusion protein columns, less nonspecific binding to GST. C, sequence alignment of segments of skeletal and cardiac muscle Ca 2ϩ and Na ϩ channel II-III loops.

RyR1-DHPR Interaction Site 7793
RyR1-DHPR interaction but may form part of a structural domain that includes the DHPR binding site. Only 10 residues between Arg 1076 and Asp 1112 differ in RyR1 and RyR2, and several of these residues are likely to form part of the interaction site between the two proteins. The newly identified RyR1-DHPR interaction site is distinct from the D2 region (amino acids 1303-1406), which was shown to be important for E-C coupling in skeletal muscle (22) by reconstituting E-C coupling in myotubes from RYR1 knockout mice. Takekura et al. (23) have reported that a functional ryanodine receptor must be provided for the formation of normal junctions between the transverse tubule and the sarcoplasmic reticulum in RYR1 knock-out mice. Thus, deletions in RyR1 may lead to the loss of E-C coupling because the interaction site between DHPR and RyR1 is missing or because RyR1 structure is so altered that proper junctions can no longer form, even though interaction sites are intact. The second possibility could explain the apparent discrepancy between the findings of Yamazawa et al. (22) and our own results.
In vivo studies have shown that RyR1, but not RyR2, can restore skeletal muscle-type E-C coupling (24). Peptides from II-III loops of the skeletal muscle DHPR that activate Ca 2ϩ currents in RyR1 reconstituted into lipid bilayers cannot activate RYR2 (9). In our studies, RyR2 sequences did not bind to the DHPR II-III loop (Fig. 2). In contrast to previous in vitro studies in which peptides from either the skeletal or cardiac II-III loops were equally capable of activating RyR1 Ca 2ϩ release channel function (9, 10), we have found that binding of the RyR1 954 -1112 fragment is specific to the skeletal DHPR II-III loop (Fig. 3). Our results are also in agreement with in vivo studies showing that skeletal-type E-C coupling in dysgenic mice lacking DHPR could only be rescued with the skeletal muscle isoform of DHPR (7,8).
An unexpected observation was that chimera C3 was bound to a GST fusion protein affinity column made up from the loop between domains II and III of the homologous skeletal muscle voltage-gated Na ϩ channel (skNa in Fig. 3, A and B). We extended this observation by aligning sequences that include 14 highly charged amino acids in the II-III loop of skeletal and cardiac DHPRs, shown to be important for activation of RyR1 (11), with the corresponding sequence in the skeletal Na ϩ channel sequences (Fig. 3C). A comparison shows that net charge ranges from 0 to ϩ6, that lysine occurs at positions 3 and 8 of the 14-amino acid sequence in the skeletal DHPR and the Na ϩ channel (both of which bind to RyR1), and that glutamate occurs at positions 3 and 8 in the cardiac DHPR (which does not bind to RyR1). The double mutation K677E/K682E in the skeletal DHPR II-III loop decreased chimera C3 binding by twothirds (from 13% to 4%). Thus, the presence of positive charges at positions 3 and 8 seems to be an important feature for binding of RyR1, but net charge does not. Because phosphorylation of Ser 687 (found at position 13 in both the skeletal muscle DHPR and the Na ϩ channel) has been shown to inhibit the interaction of the skeletal DHPR II-III loop with RyR1 (10), charged residues at the C-terminal end of this 14-amino acid sequence in the skeletal DHPR II-III loop may also be important.
Although this study does not deal directly with the functional interaction between the RyR1 and DHPR, numerous studies have implicated the II-III loop of the DHPR in E-C coupling (25,26). The loss of interaction that we observed when two residues were mutated in a region of the DHPR II-III loop, previously implicated in the triggering of Ca 2ϩ release, makes it likely that this interaction site is functional.
Monnier et al. (27) have demonstrated that an Arg to His mutation in the loop linking DHPR domains III and IV in the ␣-subunit of the skeletal muscle DHPR can cause susceptibility to malignant hyperthermia (MH). The physiological basis of MH is an abnormality in the regulation of sarcoplasmic Ca 2ϩ concentration, and the RYR1 gene has been linked to MH in human and pigs (28). The involvement of both RyR1 and DHPR in MH suggests that the pathology of MH is due to aberrant E-C coupling. These observations, together with the observations that a C-terminal peptide of the DHPR can inhibit the Ca 2ϩ release function of RyR1 (14) and that there is a retrograde signal by which RyR1 enhances slow Ca 2ϩ channel function (6,29), suggest that interactions between the RyR1 and DHPR involve many sites in the two proteins. Protein affinity chromatography should be applicable to the study of interactions of other DHPR loops with RyR1.