Portal de Periódicos da CAPES

Identification of a Dantrolene-binding Sequence on the Skeletal Muscle Ryanodine Receptor

2002; Elsevier BV; Volume: 277; Issue: 38 Linguagem: Inglês

10.1074/jbc.m205487200

1083-351X

Kalanethee Paul-Pletzer, Takeshi Yamamoto, Manjunatha B. Bhat, Jianjie Ma, Noriaki Ikemoto, Leslie S. Jimenez, Hiromi Morimoto, Philip G. Williams, Jerome Parness,

Neurobiology and Insect Physiology Research

Dantrolene is a drug that suppresses intracellular Ca2+ release from sarcoplasmic reticulum (SR) in skeletal muscle and is used as a therapeutic agent in individuals susceptible to malignant hyperthermia. Although its precise mechanism of action has not been elucidated, we have identified the N-terminal region (amino acids 1–1400) of the skeletal muscle isoform of the ryanodine receptor (RyR1), the primary Ca2+ release channel in SR, as a molecular target for dantrolene using the photoaffinity analog [3H]azidodantrolene. Here, we demonstrate that heterologously expressed RyR1 retains its capacity to be specifically labeled with [3H]azidodantrolene, indicating that muscle specific factors are not required for this ligand-receptor interaction. Synthetic domain peptides of RyR1 previously shown to affect RyR1 function in vitro andin vivo were exploited as potential drug binding site mimics and used in photoaffinity labeling experiments. Only DP1 and DP1–2s, peptides containing the amino acid sequence corresponding to RyR1 residues 590–609, were specifically labeled by [3H]azidodantrolene. A monoclonal anti-RyR1 antibody that recognizes RyR1 and its 1400-amino acid N-terminal fragment recognizes DP1 and DP1–2s in both Western blots and immunoprecipitation assays and specifically inhibits [3H]azidodantrolene photolabeling of RyR1 and its N-terminal fragment in SR. Our results indicate that synthetic domain peptides can mimic a native, ligand-binding conformation in vitro and that the dantrolene-binding site and the epitope for the monoclonal antibody on RyR1 are equivalent and composed of amino acids 590–609. Dantrolene is a drug that suppresses intracellular Ca2+ release from sarcoplasmic reticulum (SR) in skeletal muscle and is used as a therapeutic agent in individuals susceptible to malignant hyperthermia. Although its precise mechanism of action has not been elucidated, we have identified the N-terminal region (amino acids 1–1400) of the skeletal muscle isoform of the ryanodine receptor (RyR1), the primary Ca2+ release channel in SR, as a molecular target for dantrolene using the photoaffinity analog [3H]azidodantrolene. Here, we demonstrate that heterologously expressed RyR1 retains its capacity to be specifically labeled with [3H]azidodantrolene, indicating that muscle specific factors are not required for this ligand-receptor interaction. Synthetic domain peptides of RyR1 previously shown to affect RyR1 function in vitro andin vivo were exploited as potential drug binding site mimics and used in photoaffinity labeling experiments. Only DP1 and DP1–2s, peptides containing the amino acid sequence corresponding to RyR1 residues 590–609, were specifically labeled by [3H]azidodantrolene. A monoclonal anti-RyR1 antibody that recognizes RyR1 and its 1400-amino acid N-terminal fragment recognizes DP1 and DP1–2s in both Western blots and immunoprecipitation assays and specifically inhibits [3H]azidodantrolene photolabeling of RyR1 and its N-terminal fragment in SR. Our results indicate that synthetic domain peptides can mimic a native, ligand-binding conformation in vitro and that the dantrolene-binding site and the epitope for the monoclonal antibody on RyR1 are equivalent and composed of amino acids 590–609. Dantrolene is a hydantoin derivative used to treat malignant hyperthermia (MH), 1The abbreviations used are: MH, malignant hyperthermia; AMP-PCP, β,γ-methyleneadenosine 5′-triphosphate; CHO, Chinese hamster ovary; DHPR, dihydropyridine receptor; DP, domain peptide; mAb, monoclonal antibody; PIPES, piperazine-N,N′-bis(2-ethanesulfonic acid); PVDF, polyvinyl difluoride; RyR, ryanodine receptor; SCR, scrambled peptide; SR, sarcoplasmic reticulum. a rare, pharmacogenetic disorder of skeletal muscle characterized by uncontrolled Ca2+ release from sarcoplasmic reticulum (SR) stores in response to volatile anesthetics. Triggering of MH results in hypercontracture, hyperthermia, and eventually death. Therapeusis with dantrolene results from its effective suppression of skeletal muscle SR Ca2+ release, presumably by modulating the activity of the ryanodine receptor (RyR1), the primary Ca2+ release channel in skeletal muscle via that Ca2+ stored in the SR is released into the myoplasm to initiate muscle contraction in response to membrane depolarization (1Fruen B.R. Mickelson J.R. Louis C.F. J. Biol. Chem. 1997; 272: 26965-26971Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar). The genetic defect that causes MH in humans has been linked to point mutations in RyR1 in about 50% of patients. To date, some 26 MH-linked mutations in RyR1 have been identified: 9 in the extreme N-terminal region, 16 in the central region, and 1 in the extreme C-terminal region (2McCarthy T.V. Quane K.A. Lynch P.J. Hum. Mutat. 2000; 15: 410-417Crossref PubMed Scopus (307) Google Scholar, 3Sambuughin N. Nelson T.E. Jankovic J. Xin C. Meissner G. Mullakandov M., Ji, J. Rosenberg H. Sivakumar K. Goldfarb L.G. Neuromuscul. Disord. 2001; 11: 530-537Abstract Full Text Full Text PDF PubMed Scopus (43) Google Scholar, 4Froemming G.R. Ohlendieck K. Front. Biosci. 2001; 6: D65-D74Crossref PubMed Scopus (26) Google Scholar, 5MacLennan D.H. Eur. J. Biochem. 2000; 267: 5291-5297Crossref PubMed Scopus (188) Google Scholar, 6Jurkat-Rott K. McCarthy T. Lehmann-Horn F. Muscle Nerve. 2000; 23: 4-17Crossref PubMed Scopus (280) Google Scholar, 7Denborough M. Lancet. 1998; 352: 1131-1136Abstract Full Text Full Text PDF PubMed Scopus (234) Google Scholar). Interestingly, a single amino acid deletion, rather than mutation, in the central region of RyR1 (Glu2347) has also been found to be associated with MH susceptibility (8Sambuughin N. McWilliams S. de Bantel A. Sivakumar K. Nelson T.E. Am. J. Hum. Genet. 2001; 69: 204-208Abstract Full Text Full Text PDF PubMed Scopus (42) Google Scholar). Whether dantrolene suppression of Ca2+ release occurs via direct interaction with RyR1 is not entirely clear. Some have found evidence suggesting that RyR1 is not the target (9Palnitkar S.S. Mickelson J.R. Louis C.F. Parness J. Biochem. J. 1997; 326: 847-852Crossref PubMed Scopus (41) Google Scholar, 10Szentesi P. Collet C. Sarkozi S. Szegedi C. Jona I. Jacquemond V. Kovacs L. Csernoch L. J. Gen. Physiol. 2001; 118: 355-375Crossref PubMed Scopus (81) Google Scholar), and others have found evidence that it is (1Fruen B.R. Mickelson J.R. Louis C.F. J. Biol. Chem. 1997; 272: 26965-26971Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 11Nelson T.E. Lin M. Zapata-Sudo G. Sudo R.T. Anesthesiology. 1996; 84: 1368-1379Crossref PubMed Scopus (76) Google Scholar, 12Zhao F., Li, P. Chen S.R. Louis C.F. Fruen B.R. J. Biol. Chem. 2001; 276: 13810-13816Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar). Given the controversy as to the mechanism and, hence, the targets of dantrolene action, we have embarked on a project to directly identify the molecular target(s) of dantrolene by photoaffinity labeling. Knowing the target(s) of dantrolene binding would allow genetic and physiological manipulation of these molecular entities to not only elucidate the mechanism of action of this drug but also provide insights into the in vivo mechanisms controlling the RYR1 Ca2+ release channel in skeletal muscle excitation-contraction coupling and the pathophysiology of MH. Recently, we have demonstrated that [3H]azidodantrolene, a pharmacologically active, photoaffinity analog of dantrolene, specifically labels the N-terminal, 1400-amino acid residue fragment of RyR1 cleaved by n-calpain, a tissue-specific isoform of this Ca2+- and thiol-activated protease (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). Several studies have demonstrated that this portion of the RyR plays a significant role in the regulation of channel function (14Wu Y. Aghdasi B. Dou S.J. Zhang J.Z. Liu S.Q. Hamilton S.L. J. Biol. Chem. 1997; 272: 25051-25061Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar, 15Aghdasi B. Zhang J.Z., Wu, Y. Reid M.B. Hamilton S.L. J. Biol. Chem. 1997; 272: 3739-3748Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar, 16Yamazawa T. Takeshima H. Shimuta M. Iino M. J. Biol. Chem. 1997; 272: 8161-8164Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar, 17Leong P. MacLennan D.H. J. Biol. Chem. 1998; 273: 29958-29964Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar, 18Leong P. MacLennan D.H. J. Biol. Chem. 1998; 273: 7791-7794Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar). Additionally, the nine, N-terminal mutations in RyR1 linked to MH alluded to above are localized within this region of the channel. The DHPR and RyR1 are intimate physiological partners during skeletal muscle excitation-contraction coupling. The evidence indicates that the DHPR is the voltage sensor in the T-tubule membrane that upon sensing depolarization undergoes a conformational change associated with intramolecular charge movement. This is believed to result in the movement of the intracellular loop between transmembrane domains II and III (II–III loop) of the DHPR α-1 subunit, which physically contacts the RyR1, inducing its opening and resultant Ca2+ release from the SR (for reviews see Refs. 19Protasi F. Front. Biosci. 2002; 7: D650-D658Crossref PubMed Google Scholar and 20Shoshan-Barmatz V. Ashley R.H. Int. Rev. Cytol. 1998; 183: 185-270Crossref PubMed Google Scholar). Previous studies have demonstrated the experimental utility of using synthetic domain peptides derived from the DHPR and RyR1 to define physiologically significant domains within the parent protein. In particular, peptides A and C of the DHPR II–III loop and domain peptides DP1 and DP4 of the RyR1 have been shown to be active in in vitro studies of excitation-contraction coupling (21El Hayek R. Saiki Y. Yamamoto T. Ikemoto N. J. Biol. Chem. 1999; 274: 33341-33347Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar, 22Yamamoto T. Ikemoto N. J. Biol. Chem. 2002; 277: 984-992Abstract Full Text Full Text PDF PubMed Scopus (6) Google Scholar, 23Yamamoto T. Rodriguez J. Ikemoto N. J. Biol. Chem. 2002; 277: 993-1001Abstract Full Text Full Text PDF PubMed Scopus (13) Google Scholar, 24Ikemoto N. Yamamoto T. Front. Biosci. 2002; 7: D671-D683Crossref PubMed Google Scholar). Interestingly, DP1 and DP4 are from the N-terminal and central regions of the RyR1 that are "hot spots" for mutations conferring sensitivity to MH and/or the rare myopathy, central core disease (24Ikemoto N. Yamamoto T. Front. Biosci. 2002; 7: D671-D683Crossref PubMed Google Scholar). We therefore used these peptides as potential in vitro targets for [3H]azidodantrolene photoaffinity labeling, as well as others from RyR1 not demonstrated to have physiological activity as negative controls, in our attempt to define the dantrolene-binding sequence(s) in these proteins. In the present study, we demonstrate specific [3H]azidodantrolene photolabeling of two domain peptides containing the core sequence corresponding to amino acid residues 590–609 on RyR1-DP1 (amino acids 590–609) and DP1–2s, an elongated version of the skeletal sequence present in DP1 (590–628). A monoclonal anti-RyR1 antibody raised against rabbit terminal cisternae (25Campbell K.P. Knudson C.M. Imagawa T. Leung A.T. Sutko J.L. Kahl S.D. Raab C.R. Madson L. J. Biol. Chem. 1987; 262: 6460-6463Abstract Full Text PDF PubMed Google Scholar) that recognizes both the intact rabbit RyR1 and the 172 kDa, n-calpain-cleaved, N-terminal fragment of this channel (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar), also recognizes these synthetic RyR1 peptides. This antibody specifically inhibited [3H]azidodantrolene photolabeling of RyR1 in SR in a concentration-dependent manner. These results indicate, therefore, that the dantrolene-binding site on RyR1 is comprised of amino acids 590–509. Dantrolene sodium·3.5H2O and azumolene sodium·2H2O were generous gifts of Proctor & Gamble (Norwich, NY). Polyclonal sheep anti-rabbit RyR1, monoclonal mouse (IgM) anti-rabbit RyR1 XA7 (mAb), and polyclonal rabbit anti-rabbit RyR1 C-terminal (raised against a synthetic C-terminal RyR1 peptide corresponding to amino acids 5023–5037) antibodies were generous gifts of Dr. K. P. Campbell (University of Iowa, Iowa City, IA). Rabbit fast twitch skeletal muscle was supplied by Dr. H. Weiss (UMDNJ-Robert Wood Johnson Medical School, Piscataway, NJ). [3H]Azidodantrolene was synthesized, purified, and characterized exactly as described (26Palnitkar S.S. Bin B. Jimenez L.S. Morimoto H. Williams P.G. Paul-Pletzer K. Parness J. J. Med. Chem. 1999; 42: 1872-1880Crossref PubMed Scopus (38) Google Scholar), and the specific activity was determined to be 28 Ci/mmol. Rabbit RyR1 domain peptides (DP1, DP1–2s, DP4, DP3, and DP7), DP1 scrambled peptide SCR1, and rabbit skeletal muscle DHPR α1 subunit peptides (PepA and PepC) were synthesized on an Applied Biosystems model 431 A synthesizer and purified by reversed-phase high performance liquid chromatography, as described (21El Hayek R. Saiki Y. Yamamoto T. Ikemoto N. J. Biol. Chem. 1999; 274: 33341-33347Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar, 27Yamamoto T., El Hayek R. Ikemoto N. J. Biol. Chem. 2000; 275: 11618-11625Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar). DP1 scrambled peptide SCR4 was synthesized by Biosynthesis Inc. (Lewisville, TX). Scrambled DP1 sequences were generated by computer program, and SCR1 and SCR4 sequences were picked for subsequent synthesis. The entire cDNA of RyR1 was cloned into the eukaryotic expression vector pRRS11 under the control of the SV40 promoter. Plasmid pRRS11 was introduced into CHO cells using the LipofectAMINE-mediated gene transfection method, as described previously (28Towbin H. Staehelin T. Gordon J. Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 4350-4354Crossref PubMed Scopus (44938) Google Scholar, 29Wabl M. Steinberg C. Curr. Opin. Immunol. 1996; 8: 89-92Crossref PubMed Scopus (46) Google Scholar). Stable clones of CHO cells permanently transfected with RyR1 were selected using G418 (30Bhat M.B. Hayek S.M. Zhao J. Zang W. Takeshima H. Wier W.G. Ma J. Biophys. J. 1999; 77: 808-816Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar, 31Hayek S.M. Zhu X. Bhat M.B. Zhao J. Takeshima H. Valdivia H.H. Ma J. Biochem. J. 2000; 351: 57-65Crossref PubMed Scopus (25) Google Scholar). Crude SR vesicles were prepared from rabbit fast twitch skeletal muscle in the presence of protease inhibitors (200 μm phenylmethylsulfonyl fluoride, 0.3 μm aprotinin, 10 μg/ml soybean trypsin inhibitor, and 2.8 μm pepstatin A) as described in Ref. 32Hawkes M.J. Diaz Munoz M. Hamilton S.L. Membr. Biochem. 1989; 8: 133-145Crossref PubMed Scopus (32) Google Scholar. Microsomal membranes from CHO cells were obtained by ultracentrifugation of cell lysates prepared after sonication of cultured cells, as described (30Bhat M.B. Hayek S.M. Zhao J. Zang W. Takeshima H. Wier W.G. Ma J. Biophys. J. 1999; 77: 808-816Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). Leupeptin was specifically omitted from buffers used in membrane preparation to specifically allow for n-calpain cleavage of RyR1 and subsequent identification of its 172-kDa, N-terminal fragment, if present (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar, 33Shevchenko S. Feng W. Varsanyi M. Shoshan-Barmatz V. J. Membr. Biol. 1998; 161: 33-43Crossref PubMed Scopus (33) Google Scholar). SR vesicles (100 μg) and CHO microsomal membranes (200 μg) were photolabeled with [3H]azidodantrolene (100–200 nm) in binding buffer (20 mm PIPES, pH 7, containing 0.5 mm AMP-PCP) in the absence (T, total binding) or presence (N, nonspecific binding) of azumolene (150–300 μm), as described previously (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). For inhibition of photolabeling by mAb anti-RyR1, mouse ascites were added to the binding buffer at dilutions of 1:50, 1:25, and 1:10 in the presence of protease inhibitors (see "Membrane Preparation" above) with the addition of 1 mm leupeptin. Synthetic peptides (12.5 μM) were photolabeled with 50–100 nm[3H]azidodantrolene in binding buffer containing 10 μg of bovine serum albumin in the absence of AMP-PCP. Following photolabeling, the samples were resolved by SDS-PAGE (5% acrylamide for SR samples and 20% for synthetic peptides) (34Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (207522) Google Scholar). For autoradiography and Western blot analyses, the SDS-PAGE resolved proteins were electroblotted onto PVDF membranes (Sequi-Blot; Bio-Rad) (28Towbin H. Staehelin T. Gordon J. Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 4350-4354Crossref PubMed Scopus (44938) Google Scholar). Autoradiography was performed as described (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). For Western blots, the membranes were probed with primary antibody at appropriate dilution/concentration (polyclonal anti-RyR1, 1:5000; mAb, 1:1000), followed by alkaline phosphatase-conjugated secondary antibody. Immunoreactive bands were visualized with 5-bromo-4-chloro-3-indolyl phosphate/p- nitro blue tetrazolium using standard procedures. The autoradiograph was scanned using Photoshop 6.0 (Adobe) software and the image was saved as a tif file. The saved image was analyzed using ImageQuant for Windows NTTM software (Molecular Dynamics and AmershamBiosciences), and the radiolabeled bands were quantified. The arbitrary values were normalized after subtracting for background and nonspecific binding and expressed as percentages of control values. Of the 5037 amino acids that constitute the RyR1 sequence, we have shown that a dantrolene-binding site resides between amino acids 1 and 1400 of the N-terminal region of the channel (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). In an attempt to devise a molecular biological approach to define the amino acids that comprise the dantrolene-binding site, we first asked whether RyR1 is capable of specifically interacting with dantrolene in the absence of a muscle cell background. We reasoned that if it did, (a) this would corroborate our finding that the RyR1 is a pharmacological target of dantrolene and (b) we might reasonably attempt to express a region of the N-terminal portion of the channel that would retain its ability to bind dantrolene and assist us in localization of the drug-binding site. We used recombinant RyR1, stably expressed in CHO cells, as a target for our photoaffinity labeling experiments. Microsomal membranes prepared from these cells, along with rabbit skeletal muscle SR vesicles as positive control, were photolabeled with [3H]azidodantrolene in the presence or absence of AMP-PCP and excess azumolene, as described under "Experimental Procedures". As shown in Fig. 1 A, specific photolabeling of a 565-kDa protein corresponding to the RyR1 monomer in SR was observed in CHO microsomal membranes expressing RyR1. This photolabeling was dependent on the presence of AMP-PCP. This is identical to the requirements for photolabeling of RyR1 and its 172-kDa fragment in SR reported previously (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). The Western blot (Fig1 B) of the same membrane used for autoradiography, probed with a polyclonal anti-RyR1 antibody, demonstrates equivalent protein loading in all the experimental lanes and confirms that the radiolabeled band in CHO cell extracts is indeed RyR1. No specific photolabeling or anti-RyR1 immunoreactive bands were observed in membranes prepared from untransfected CHO cells. These results demonstrate that (a) dantrolene can bind to heterologously expressed RyR1 in the absence of other muscle-specific proteins and (b) this binding is pharmacologically specific and dependent on adenine nucleotide triphosphate, a known regulator of the channel protein. To determine the amino acids on RyR1 that constitute the binding site for dantrolene, we attempted to specifically photolabel a heterologously expressed, green fluorescent protein-tagged, RyR1-N-terminal fragment containing amino acids 182–1608. This fragment did not label, presumably because of incorrect folding or processing (data not shown). We then turned to synthetic domain peptides of RyR1 as possible surrogate targets for the dantrolene-binding site. We synthesized four peptides with amino acid residues from within the N-terminal, 1400-amino acid region of RyR1 (DP1, DP1–2s, DP3, and DP7), as well as peptides from the central region of RyR1 (DP4) and from the II–III loop of the DHPR-α1 subunit (PepA and PepC) (Fig. 2). These peptides were photoreacted with [3H]azidodantrolene in the presence or the absence of excess azumolene, and the peptides were analyzed for specific photolabeling by autoradiography. The autoradiograph in Fig. 3 Ademonstrates specific photolabeling of DP1 (amino acid residues 590–609) and an elongated version of DP1, DP1–2s (amino acid residues 590–628). DP7 (amino acid residues 543–576) was nonspecifically labeled, whereas the other peptides were not labeled at all. The pharmacological specificity of [3H]azidodantrolene labeling of DP1 was demonstrated by the ability of dantrolene and azumolene, but not the unrelated drug, atropine, to inhibit this interaction (Fig. 3 B). In addition, scrambling of the DP1 sequence (Fig. 3 D, SCR1 and SCR4) led only to nonspecific labeling of varying intensity (Fig. 3 C). These results suggest that DP1 might represent an in vitromodel of the dantrolene-binding site on RyR1.Figure 3[3H]Azidodantrolene photolabeling of synthetic peptides. A, synthetic peptides were photolabeled with [3H]azidodantrolene in the absence (lanes T) or the presence of azumolene (lanes N), resolved on 20% SDS-PAGE, electroblotted onto PVDF membrane, and subjected to autoradiography. B, pharmacological specificity of photolabeling of DP1 was determined in the absence (lane T) or presence of 150 μmazumolene (lane Az), 25 μm dantrolene (Dn), 150 μm atropine (lane At), an unrelated drug as negative control. C, photolabeling of DP1 and scrambled peptides of DP1, SCR1, and SCR4. D, sequence comparison of DP1 and scrambled peptides.View Large Image Figure ViewerDownload (PPT) Our previous work has shown that an anti-RyR1 mAb immunoprecipitates [3H]azidodantrolene-labeled RyR1 and its 172-kDa fragment from solubilized SR (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). Given the epitope specificity of monoclonal antibodies and that the epitope of this monoclonal seemed to be on the 172-kDa protein, we reasoned that it was possible that one of our domain peptides corresponding to sequences from the N-terminal region of RyR1 might interact with this antibody. If it did, it would demonstrate the feasibility of using synthetic peptides to mimic a protein epitope in this channel. First, to confirm our earlier results described above, we demonstrated that only intact RyR1 and its 172-kDa N-terminal fragment were reactive with the monoclonal antibody by photolabeling duplicate SR samples with [3H]azidodantrolene, electroblotting these onto PVDF membranes after SDS-PAGE, and probing one sample with mAb anti-RyR1 and the other with polyclonal anti-RyR1 C-terminal antibody. Fig.4 A clearly demonstrates that the mAb recognizes [3H]azidodantrolene-labeled RyR1 monomer and its 172-kDa fragment but not the C-terminal, 410-kDa portion of the channel. The anti-RyR1 C-terminal antibody, on the other hand, recognizes the [3H]azidodantrolene-labeled RyR1 monomer and its 410-kDa, C-terminal fragment but not its 172-kDa N-terminal fragment (Fig. 4 B). This indicated to us that the epitope of mAb, hitherto unknown, must lie within the first 1400 residues of the N-terminal region of RyR1. To determine whether any of our RyR1 domain peptides could be recognized by this antibody, we resolved the peptides on a 20% SDS-PAGE gel, transferred them onto PVDF membrane, and probed the latter with mAb anti-RyR1. Only DP1 and DP1–2s, both of which share the same corresponding core of amino acid residues (590–609) of RyR1, were recognized by the mAb (Fig. 5 A). The scrambled DP1 peptides SCR1 and SCR4 were not recognized by the antibody (Fig.5 B). These results indicate that the epitope for the mAb is contained within the DP1 sequence.Figure 5Immunorecognition of DP1 sequence-containing peptides by mAb anti-RyR1. RyR1 domain peptides, DP1, DP1–2s, DP7, and DP4; DHPR II–III loop peptide; and PepA (A) and scrambled DP1 peptides SCR1 and SCR4 (B) were resolved on a 20% SDS-PAGE gel and transferred onto PVDF membrane and probed with mAb anti-RyR1. Only the DP1 sequence-containing peptides are recognized.View Large Image Figure ViewerDownload (PPT) Significantly, the results presented above demonstrate a parallelism between mAb recognition and specific [3H]azidodantrolene photolabeling of a sequence in the N-terminal of RyR1. The intersection of these two sets of data predicts that mAb anti-RyR1 should inhibit [3H]azidodantrolene labeling of RyR1 in SR. To test this hypothesis, we photolabeled SR in the absence or presence of increasing concentrations of mAb-containing mouse ascites. The autoradiograph (Fig.6 A) reveals a concentration-dependent inhibition of specific photolabeling of both the RyR1 monomer, as well as its 172-kDa fragment; this was despite equivalent anti-RyR1 immunoreactive protein loading in each lane (Fig. 6 B). Quantitation of the inhibition of photolabeling revealed near linear inhibition over the mAb concentration range tested, particularly for the 172-kDa fragment (Fig. 6 C). The linearity of inhibition of photolabeling of intact RyR1 by mAb anti-RyR1 is less evident in this autoradiograph because of its high concentration and the degree of photolabeling relative to the N-terminal fragment. Optimizing exposure for the inhibition of photolabeling of the 172-kDa fragment results in near saturation of the sensitivity of the x-ray film for the labeling of the 565-kDa RyR1 monomer (data not shown). As a result, the two inhibition curves are not entirely parallel. This does not negate the phenomenon of specific inhibition of [3H]azidodantrolene photolabeling of both forms of the channel protein, n-calpain-cleaved and uncleaved. Moreover, inhibition of labeling is not due to nonspecific effects of IgM itself, because purified mouse IgM at the highest concentration used in mouse ascites did not affect photolabeling of either the intact channel or its 172-kDa fragment (Fig. 6 A). It is our hypothesis that understanding the molecular mechanism of action of dantrolene will lead to a better understanding of the physiology of MH, the mechanism of excitation-contraction coupling, and the regulation of intracellular Ca2+ release from SR. A prerequisite for elucidating molecular mechanism requires the identification of molecular target(s). The biochemical identification of amino acids involved in drug binding allows for the design of mutational experiments to test their functional significance in drug action. Here, we have amassed evidence that the amino acid sequence of the dantrolene-binding site on RyR1 corresponds to amino acids 590–609, corresponding to the synthetic domain peptide DP1. The most compelling evidence for the DP1 sequence being the dantrolene-binding sequence in RyR1 comes from our experiments with the monoclonal antibody, mAb anti-RyR1. In 1987, Campbell et al.(25Campbell K.P. Knudson C.M. Imagawa T. Leung A.T. Sutko J.L. Kahl S.D. Raab C.R. Madson L. J. Biol. Chem. 1987; 262: 6460-6463Abstract Full Text PDF PubMed Google Scholar) produced an IgM anti-RyR1 mAb, clone XA7, of uncharacterized epitope, that specifically immunoprecipitates [3H]ryanodine-bound RyR1. Our present studies demonstrate that this antibody recognizes its epitope not only on the RyR1 monomer and its 172-kDa, N-terminal, calpain cleavage fragment but also on the synthetic peptides, DP1 and DP1–2s, containing the core RyR1 sequence 590–609. Most significantly, this antibody specifically inhibits [3H]azidodantrolene photolabeling of RyR1 and its 172-kDa fragment in SR in a concentration-dependent manner. This result is consistent with our earlier experiments demonstrating the absence of a dantrolene-binding site on RyR1 C-terminal to the n-calpain cleavage site between amino acids 1400–1401 (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). The fact that we were able to demonstrate virtually complete inhibition of [3H]azidodantrolene photolabeling of the 172-kDa fragment with a monoclonal antibody is a strong argument that the epitope defined by this antibody, amino acids 590–609, represents the only dantrolene-binding site on RyR1. The relatively high concentrations (<1:50 dilution) of mouse ascites required to bring about inhibition of specific photolabeling is likely due to the low antigen affinities of IgM antibodies, which are part of the early immune response of the body, relative to mature IgG antibodies (29Wabl M. Steinberg C. Curr. Opin. Immunol. 1996; 8: 89-92Crossref PubMed Scopus (46) Google Scholar). These data also strengthen the experimental evidence that, in vitro, synthetic domain peptides are capable of assuming the native conformation(s) of domains present in the parent protein, RyR1 in this case (35Villen J. Borras E. Schaaper W.M. Meloen R.H. Davila M. Domingo E. Giralt E. Andreu D. Chembiochem. 2002; 3: 175-182Crossref PubMed Scopus (24) Google Scholar, 36De Ferrari G.V. Canales M.A. Shin I. Weiner L.M. Silman I. Inestrosa N.C. Biochemistry. 2001; 40: 10447-10457Crossref PubMed Scopus (373) Google Scholar, 37Langosch D. Crane J.M. Brosig B. Hellwig A. Tamm L.K. Reed J. J. Mol. Biol. 2001; 311: 709-721Crossref PubMed Scopus (116) Google Scholar). Previous studies have already shown that heterologously expressed RyR1 in CHO cells forms functional channels (38Bhat M.B. Zhao J. Takeshima H. Ma J. Biophys. J. 1997; 73: 1329-1336Abstract Full Text PDF PubMed Scopus (116) Google Scholar) and allowed us to address the question of whether dantrolene binding to RyR1 requires the presence of muscle-specific factors in heterologously expressed channels. Here, we have demonstrated that RyR1 can interact directly with dantrolene in the absence of a muscle-specific milieu. That this interaction is modulated by AMP-PCP, as it is in SR (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar), substantiates our assertion that this drug-channel interaction is pharmacologically relevant. It further indicates that this nucleotide analog affects RyR1 directly in enhancing the ability of this channel to bind dantrolene. Specific photolabeling of the synthetic peptide, DP1, on the other hand, should be independent of all specific regulators of RyR1 activity, because none of the putative nucleotide or protein regulatory sites on RyR1 would be expected to be associated with this 20-amino acid peptide. Indeed, our data demonstrate that [3H]azidodantrolene photolabeling of the two DP1 sequence-containing peptides is independent of AMP-PCP (data not shown). Although we have presented strong evidence that the dantrolene-binding site on RyR1 is modulated by this nucleotide analog, it is not yet clear how ATP might modulate the activity of this drugin vivo. A recent, detailed examination of the possible physiological site(s) of dantrolene action in skeletal muscle has shown that dantrolene suppresses but never completely eliminates intracellular Ca2+ release in whole and skinned muscle fibers or from SR vesicles but has no effect on purified RyR1 incorporated into lipid bilayers (10Szentesi P. Collet C. Sarkozi S. Szegedi C. Jona I. Jacquemond V. Kovacs L. Csernoch L. J. Gen. Physiol. 2001; 118: 355-375Crossref PubMed Scopus (81) Google Scholar). These results led those authors to conclude that RyR1 is not likely the site of action of dantrolene. This result is in stark contrast to our results above and those of Nelson and colleagues (11Nelson T.E. Lin M. Zapata-Sudo G. Sudo R.T. Anesthesiology. 1996; 84: 1368-1379Crossref PubMed Scopus (76) Google Scholar). Possible unifying explanations of these disparate results are likely to be methodologic and could include the following: (a) the dantrolene-binding site on RyR1 is uniquely sensitive to modification during the purification process rendering it dantrolene insensitive under certain as yet undefined conditions or (b) the physiological effects of dantrolene binding to RyR1 requires the interaction of other RyR1 interacting factors that might be removed during purification or whose interactions are not modeled well in the artificial lipid bilayer. Comparing the DP1 sequence in RyR1 with the other rabbit isoforms reveals that there is an identical sequence in RyR2 (601–619) and a nearly identical one in RyR3 (577–597), with a conservative amino acid substitution (Val to Leu) at position 596 (39Takeshima H. Ann. N. Y. Acad. Sci. 1993; 707: 165-177Crossref PubMed Scopus (37) Google Scholar). If this sequence is the dantrolene-binding site, all three should be sensitive to this drug. Yet, existing evidence indicates that only the RyR1 and RyR3 isoforms are sensitive to dantrolene. We have shown by [3H]azidodantrolene photolabeling of SR vesicles prepared from skeletal or cardiac muscle that only RyR1 is a good target for dantrolene, not RyR2 (13Paul-Pletzer K. Palnitkar S.S. Jimenez L.S. Morimoto H. Parness J. Biochemistry. 2001; 40: 531-542Crossref PubMed Scopus (41) Google Scholar). Similarly, other laboratories have shown that dantrolene can inhibit [3H]ryanodine binding to RyR1 and RyR3 but not RyR2 (1Fruen B.R. Mickelson J.R. Louis C.F. J. Biol. Chem. 1997; 272: 26965-26971Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 12Zhao F., Li, P. Chen S.R. Louis C.F. Fruen B.R. J. Biol. Chem. 2001; 276: 13810-13816Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar). If the DP1 sequence is the dantrolene-binding site, why is the RyR2 isoform insensitive to this drug when it contains the identical sequence? The most likely explanation at our present state of knowledge is that the domain-domain interactions within RyR2 constrain the conformation around this sequence so that dantrolene is no longer able to bind with high affinity to this site. Other possible explanations include RyR2-associated proteins or post-translational modifications of the channel that block access to the site. The underlying reasons for the differences in dantrolene sensitivity of RyR2 relative to RyR1 are under active investigation in our laboratory. As noted above, the DP1 sequence of RyR1 is nearly identical in RyR3. Whether this isoform is responsive to dantrolene in the organs in which this isoform is expressed remains to be clarified. Early studies indicated that dantrolene has no effect on smooth muscle contractility (41Sengupta C. Meyer U.A. Carafoli E. FEBS Lett. 1980; 117: 37-38Crossref PubMed Scopus (24) Google Scholar), although more recent studies contradict that (42Nasu T. Oosako H. Shibata H. Gen. Pharmacol. 1996; 27: 513-517Crossref PubMed Scopus (17) Google Scholar). Indeed, the latter studies are strongly supported by the demonstration that heterologously expressed RyR3 has been shown to be responsive to dantrolene (12Zhao F., Li, P. Chen S.R. Louis C.F. Fruen B.R. J. Biol. Chem. 2001; 276: 13810-13816Abstract Full Text Full Text PDF PubMed Scopus (250) Google Scholar). If RyR3 is indeed a target of dantrolene, then this last result might explain the various effects of dantrolene seen in nonmuscle tissues that express RyR3, i.e. neuronal and immune cells (43Lokuta A.J. Komai H. McDowell T.S. Valdivia H.H. FEBS Lett. 2002; 511: 90-96Crossref PubMed Scopus (27) Google Scholar, 44Hosoi E. Nishizaki C. Gallagher K.L. Wyre H.W. Matsuo Y. Sei Y. J. Immunol. 2001; 167: 4887-4894Crossref PubMed Scopus (76) Google Scholar). Recently, Ikemoto and co-workers (21El Hayek R. Saiki Y. Yamamoto T. Ikemoto N. J. Biol. Chem. 1999; 274: 33341-33347Abstract Full Text Full Text PDF PubMed Scopus (32) Google Scholar, 27Yamamoto T., El Hayek R. Ikemoto N. J. Biol. Chem. 2000; 275: 11618-11625Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar) demonstrated that DP1 added to SR vesicles activates RyR1, as measured by enhanced [3H]ryanodine binding. These authors have postulated that DP1 may be assuming the native conformation of the identical sequence in the channel (domain x), thereby competing with this domain for its interacting or "mating" domain (domain y) on RyR1, resulting in "unzipping" of the native RyR1 domain x-domain y interactions (24Ikemoto N. Yamamoto T. Front. Biosci. 2002; 7: D671-D683Crossref PubMed Google Scholar,40Ikemoto N. Yamamoto T. Trends Cardiovasc. Med. 2000; 10: 310-316Crossref PubMed Scopus (59) Google Scholar, 45Yamamoto T. Ikemoto N. Biochemistry. 2002; 41: 1492-1501Crossref PubMed Scopus (49) Google Scholar). If, under normal resting conditions, domain x-domain y interactions would keep the channel in a closed state, unzipping these interactions should result in activation of the channel. It follows then that the insinuation of DP1 into the normal domain x position, mating or zipping to domain y of RyR1, would result in loss of the conformational constraints imparted by these zipped domains and lead to channel opening. If the above is true, then the most parsimonious explanation of the results to date is that dantrolene inhibits RyR1-mediated Ca2+ release by stabilizing the interdomain x-y interactions. Moreover, dantrolene stabilization of these interdomain interactions may even result in stabilization of other protein domain-domain interactions, either within RyR1 or with another interacting protein(s) or both, thus reducing the likelihood of channel openings and inhibiting Ca2+ release. Elucidation of the identity of the putative interacting domain(s) and modeling of the dantrolene-binding site in the three channel isoforms would greatly contribute to our understanding of the regulation of RyR channel function. We regret the untimely closing of the National Tritium Labelling Facility as a national resource and acknowledge the formidable contributions of this resource to our research program in particular and national biomedical small molecule research programs in general.