Functional Consequences of Mutations of Conserved, Polar Amino Acids in Transmembrane Sequences of the Ca2+ Release Channel (Ryanodine Receptor) of Rabbit Skeletal Muscle Sarcoplasmic Reticulum*

The potential role in Ca2+ release channel function of highly conserved, polar, and small amino acids in predicted transmembrane sequences in the rabbit skeletal muscle ryanodine receptor (RyR1) was investigated through mutagenesis. Acidic amino acids Asp3987, Glu4032, Asp4815, Asp4917, Asp4938, and Asp4969 and amidated residues Asn4034, Asn4037, Asn4574, Asn4805, Asn4806, and Gln4933, and Gly4033 were mutated to Ala, and Ala3988 was mutated to Val. When expressed in HEK-293 cells and challenged with either caffeine or 4-chloro-m-cresol, mutants E4032A, N4806A, D4815A, and D4917A did not respond, indicating that Ca2+ release channel function was impaired. None of these mutants exhibited specific binding of [3H]ryanodine. Mutants N4805A and Q4933A showed a diminished response to both caffeine and 4-chloro-m-cresol, but [3H]ryanodine binding was not altered. Other mutant responses and the responses of mutants E4032D, N4806Q or D, D4815N or E, and D4938N or E were unaltered when compared with RyR1. However, mutants E4032Q, D4917N or E, and Q4933N or E displayed neither caffeine nor 4-chloro-m-cresol response nor [3H]ryanodine binding. Sedimentation assays indicated that the nonfunctional mutants did contain tetrameric complexes, implying that defects in the assembly of a functional channel did not occur with specific mutations in transmembrane sequences. These results support the view that amino acids Glu4032 (M2), Asn4806 (M7), Asp4815 (M7), Asp4917 (M10), and Gln4933 (M10) are involved in channel function and regulation.

In skeletal muscle, excitation-contraction coupling involves the interaction between two important proteins, the ␣1-subunit of the L-type Ca 2ϩ channel (dihydropyridine receptor or DHPR), 1 located in the transverse tubular membrane, and the Ca 2ϩ release channel (ryanodine receptor or RyR), located in the junctional terminal cisternae of the sarcoplasmic reticu-lum. The DHPR acts as a voltage sensor in response to depolarization of the transverse tubule, whereas RyR responds as a release channel for Ca 2ϩ stored in the sarcoplasmic reticulum. Ca 2ϩ channel function of the DHPR is probably irrelevant for excitation-contraction coupling in skeletal muscle (1)(2)(3). Analysis of the deduced amino acid sequence of rabbit skeletal muscle RyR has led to the prediction that transmembrane sequences near the COOH terminus form the Ca 2ϩ -conducting pore and the binding sites for ryanodine and Ca 2ϩ (4 -6), whereas the remainder of the molecule forms a series of cytoplasmic domains.
The number of transmembrane sequences in the COOHterminal region is still undefined. We proposed that there are 12 transmembrane sequences, but MЈ and MЉ were tentative (5). Although we no longer consider M3 and M4 to be transmembrane sequences, since they are not conserved in the RyR/ IP 3 R family, we retain the Zorzato numbering system in this study. Studies using site-directed antibodies have confirmed that the NH 2 and COOH termini of RyR1 are located in the cytosol (7,8). Recent investigations of an expressed, truncated RyR1 protein have demonstrated that the COOH-terminal portion of RyR1, containing M1 to M10, is sufficient to form a functional Ca 2ϩ release channel (9).
Studies of structure/function relationships in transmembrane sequences in the Ca 2ϩ -ATPase (SERCA1) have revealed that acidic or amidated amino acids provide ligand-binding sites for Ca 2ϩ , whereas the juxtaposition of a small residue with an acidic residue (EG in M6, GE in M7, and DG in M8) is a critical motif in SERCA molecules (10 -14). Analysis of the transmembrane sequences M1 to M10 of RyR1 and alignment of the sequences of 6 RyR and 5 IP 3 R members of the Ca 2ϩ release channel family (4,(15)(16)(17)(18)(19)(20)(21)(22)(23) revealed that several negatively charged and amidated residues are highly conserved. Among these, Glu 4032 , Asp 4815 , and Asp 4917 are absolutely conserved ( Fig. 1), suggesting that these amino acids might play an important role in structure/function relationships in Ca 2ϩ release channels.
In this study, we mutated 14 negatively charged and amidated amino acids (plus vicinal Gly and Ala residues), which were conserved among 11 members of the RyR and IP 3 R families. These residues are located in M1, M2, M5, M7, and M10 of RyR1. We used Ca 2ϩ microfluorimetry to measure Ca 2ϩ release in response to caffeine and 4-chloro-m-cresol for wild type and mutant RyR1 proteins transiently expressed in HEK-293 cells. We also measured [ 3 H]ryanodine binding by wild type and mutant RyR1 proteins, and sedimentation was used to assess the oligomeric status of the expressed receptors. Our results indicate that residues in M2, M7, and M10 are involved in channel function.

EXPERIMENTAL PROCEDURES
Materials-Pfu polymerase, restriction endonucleases, and other DNA-modifying enzymes were purchased from Stratagene, Boehringer Mannheim, New England Biolabs, Promega, and Amersham Pharmacia Biotech. Fura-2 acetoxymethyl ester was from Molecular Probes. Caffeine was from Sigma and 4-chloro-m-cresol from Fluka. Unlabeled ryanodine was obtained from Calbiochem, and [ 3 H]ryanodine was purchased from NEN Life Science Products. CHAPS was from Bio-Rad. Expression vector pcDNA 3.1(Ϫ) was from Invitrogen. Monoclonal antibody 34C was a kind gift from Dr. Judith Airey (24). Horseradish peroxidase-and alkaline phosphatase-conjugated goat anti-mouse secondary antibody (IgG) were obtained from Promega, and SuperSignal Ultra chemiluminescent substrate was obtained from Pierce. Nitrocellulose membranes were from Schleicher & Schuell. All other reagents were of reagent grade or highest grade available.
Oligonucleotide-directed Mutagenesis-Site-directed mutagenesis was carried out in cassettes that were excised from the full-length sequence in pBluescript KS(ϩ) (Stratagene) (pBS-RyR1) using unique restriction endonucleases (25). In the current experiments, mutagenesis involved cassettes C9 (NdeI-NheI), C10 (NheI-ClaI), and C11 (ClaI-HindIII) in vectors pBS and pBS2 (25). Shorter fragments were excised from C9 (EcoRI-XhoI) and (XhoI-NheI) and ligated into the multiple cloning sites of pBluescript or pBS2. Site-specific mutagenesis was performed on these fragments by Pfu polymerase-based polymerase chain reaction using the Quickchange kit (Stratagene). The complete sequence of the mutated fragment was confirmed using the dideoxynucleotide chain termination method (26). Fragments with the desired mutations were subcloned back into their original positions in pBS-RyR1, and then the entire sequence was excised with XbaI and HindIII and subcloned into the pcDNA 3.1(Ϫ) vector for expression in HEK293 cells.
Cell Culture and DNA Transfection-Culture of HEK-293 cells and their transfection by the calcium phosphate precipitation method (27) was carried out as described previously (25,28).
Immunoblotting-SDS-polyacrylamide gel electrophoresis (29) and immunoblotting (30) were performed on total proteins from 25 l of cell lysate obtained from one 35-mm plate as described previously (25).
Fluorescence Measurements-A microfluorimetry system (Photon Technologies Inc.) was used to monitor the fluorescence changes before and after addition of caffeine or 4-chloro-m-cresol to the surface of HEK-293 cells that had been transfected with wild type or mutant RyR1 cDNA, as described previously (25). Cells were cultured and loaded with Fura-2AM on a glass coverslip. The coverslip was then inserted into a chamber mounted on the stage of a Zeiss inverted Diaphot microscope. The chamber was used in the open bath configuration so that solution changes could be made at a rate of ϳ2.5 ml/min from a pipette tip placed near the target cells. A vacuum line was placed on the side of the chamber opposite to the pipette tip and adjusted so that the volume was maintained at ϳ0.5 ml. For each mutant, clusters of about 30 cells were monitored in 3-4 fields in each coverslip, and 4 -6 different transfections were carried out. Wild type RYR1 cDNA was included in each transfection. Data were averaged and expressed as mean Ϯ S.E.
Microsome Preparations from HEK-293 Cells-Microsome preparations were carried out as described previously (31). Solubilization of Transfected HEK-293 Cells or Microsomes and Sucrose Density Gradient Centrifugation-Transfected HEK-293 cells grown in 100-mm Petri dishes were washed twice with 5 ml of PBS (147 mM NaCl, 2.7 mM KCl, 8 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , pH 7.4) and harvested in PBS containing 5 mM EDTA. Cells were collected by centrifugation at 4,000 rpm for 10 min in a Sorvall SS-34 rotor. Cell pellets were washed with PBS and centrifuged again. Cell pellets from 100-mm dishes were then solubilized on ice for 45 min in 1 ml of buffer containing 150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 1% CHAPS, 5 mg/ml L-phosphatidylcholine, and the protease inhibitor mix described above. Solubilized proteins were obtained by removing the debris through centrifugation at 4,000 rpm at 4°C for 10 min. Microsome preparations were solubilized as described previously (31). The resulting supernatant from cells or microsomes was placed on the top of a 7.5-25% (w/v) linear sucrose gradient solution containing 50 mM Tris-Hepes, pH 7.4, 0.3 M NaCl, 0.1 mM CaCl 2 , 0.3% CHAPS, 0.15% L-phosphatidylcholine, and protease inhibitor mix and was centrifuged at 28,000 rpm in a Beckman SW 40 rotor for 18 -20 h at 4°C. Fractions of 1.0 ml each were collected from bottom to top and measured for immmunoreactivity.
Protein Assay-Protein concentration was determined by dye binding using bovine serum albumin as a standard (32).

Mutagenesis and Expression of Mutant RyR1 cDNAs-
Alignment of sequences in 6 RyR and 5 IP 3 R members of the Ca 2ϩ release channel family (4, 15-23) that we have predicted to be transmembrane sequences (5,15) showed that conserved acidic and amidated amino acids are present in predicted transmembrane sequences M1, M2, M5, M6, M7, and M10 in RyR1 and that Gly or Ala residues vicinal to an acidic amino acid are present in M1 (DA) and M2 (EG) (Fig. 1). Glu 4032 , Asp 4815 , and Asp 4917 are 100% conserved in the RyR/IP 3 R family, but Asn 4662 is poorly conserved and was not analyzed. Conserved, negatively charged amino acids, Asp 3987 , Glu 4032 , Asp 4815 , and Asp 4815 , amidated residues Asn 4034 , Asn 4937 , Asn 4575 , Asn 4805 , Asn 4806 , Gln 4933 , and Ala 3988 , and small residue Gly 4033 were selected for mutagenesis. Two absolutely conserved residues Asp 4938 in the COOH terminus and Asp 4938 and Asp 4969 , located on the predicted cytoplasmic junction of M10, were also included. In initial studies, most residues were mutated to the small, non-polar residue Ala, whereas Ala 3988 was mutated to Val.
Wild type and mutant RyR1 proteins were expressed in HEK-293 cells. Western blotting of whole cell lysates with monoclonal antibody 34C showed no expression of RyR1 in non-transfected or vector-transfected HEK-293 cells, but wild type and all of the mutants were expressed at readily detectable levels (Fig. 2). Although the extent of staining of Western blots of whole cell lysates varied from lane to lane, the differences averaged out over 5 independent experiments, so that we did not observe any significant differences in the expression level of any of the mutants described in this study. There was also no significant difference in apparent molecular mass between wild type and mutant RyR1 proteins (Fig. 2).
Fluorescence Measurement of Ca 2ϩ Release in Response to Caffeine and 4-Chloro-m-cresol-Caffeine releases Ca 2ϩ from internal stores through RyR, by sensitizing the Ca 2ϩ -induced Ca 2ϩ release process (33), whereas 4-chloro-m-cresol stimulates the Ca 2ϩ release channel through an unknown mechanism (34). The peak amplitude of fura-2 fluorescence following each application of caffeine or 4-chloro-m-cresol was measured to estimate the function of each expressed wild type or mutant RyR1. Fig. 3A shows an example of the fluorescence changes elicited in wild type RyR1-transfected cells in response to caffeine and 4-chloro-m-cresol. Fura-2-loaded cells were first challenged with 20 mM caffeine. This caused a significant upward shift in the fluorescence ratio, indicating that Ca 2ϩ stored internally was released into the cytosol. After peak fluorescence was reached, caffeine was washed away by perfusion, allowing Ca 2ϩ to return to the internal store before 4-chloro-m-cresol was added to obtain a second peak of Ca 2ϩ release. We did not observe any changes in the 340/380 nm ratio after application of either 20 mM caffeine or 0.3 mM 4-chloro-m-cresol to either non-transfected or vector-transfected HEK 293 cells. We did, however, observe that concentrations of 4-chloro-m-cresol higher than 0.5 mM could induce a slow Ca 2ϩ release phase in non-transfected cells. Fig. 3B shows the results of fluorescence measurements for control, wild type, and mutant RyR1 proteins. Mutation to Ala of three absolutely conserved residues, Glu 4032 , Asp 4815 and Asp 4917 , and of the less well conserved Asn 4806 , led to the loss of response to caffeine and 4-chloro-m-cresol. Two mutants, N4805A and Q4933A, showed decreased response to both agents. The response of other mutants, to caffeine and 4-chlorom-cresol, including control mutants Asp 4938 and Asp 4969 , was not different from wild type RyR1. Mutants A3988V and G4033A were also normally responsive to caffeine and 4-chlorom-cresol. The location of Gly vicinal to an acidic amino acid, a critical feature in the formation of Ca 2ϩ -binding sites in Ca 2ϩ -ATPase molecules (12,14), is obviously not critical to RyR1 function.
Since mutation to Ala changed size and charge for acidic residues, and size and polarity for amidated residues, it was of interest to determine the consequences of less drastic mutations. Asp was mutated to Asn and Glu, Glu to Gln and Asp, Asn to Gln and Asp, and Gln to Asn and Glu. These included Glu 4032 in M2; Asn 4805 , Asn 4806 and Asp 4815 in M7; and Asp 4917 and Gln 4933 in M10. In addition, Asp 4938 in the COOH terminus was again included as a control. The response of these mutants to caffeine and 4-chloro-m-cresol is presented in Fig.  3C.
When Glu 4032 in M2 was mutated to Asp, wild type function was retained, but this was not the case for mutation to Gln, indicating that charge is important for RyR1 function. In M5, mutation of Asn 4805 to Gln and Asp and mutation of Asp 4815 to Asn and Glu did not result in any significant defect in response to caffeine or 4-chloro-m-cresol, implying that the size of these residues is more important than charge. Mutation of Asp 4917 in M10 to Ala, Asn, and Glu led to loss of caffeine and 4-chlorom-cresol responses, implying that retention of Asp at position 4917 is essential. When Gln 4933 was mutated to Ala, a decrease in the peak amplitude of the fluorescence ratio was observed, but mutation to Asn and Glu, of similar size and polarity, abolished responses to both caffeine and 4-chloro-m-cresol. When Asp 3938 was mutated to Asn, the mutant response to caffeine and 4-chloro-m-cresol was not different from wild type, and when mutated to Glu, the response was diminished.
[ 3 H]Ryanodine Binding Assay-Binding of ryanodine to the single high affinity binding site on the ryanodine receptor requires channel opening (35)(36)(37). Microsomes from cells expressing each of the mutant RyR1 proteins were incubated with 60 nM [ 3 H]ryanodine under conditions promoting Ca 2ϩ release channel opening. As shown in Fig. 4, those mutants lacking response to caffeine and 4-chloro-m-cresol also had no significant [ 3 H]ryanodine binding, indicating that the channels were either closed, that they did not form a tetrameric complex, or that they did not fold and assemble properly and were vulnerable to proteolytic digestion.
Western blotting of wild type and nonfunctional mutant proteins in cell lysates (Fig. 2) or in microsomal membranes (not shown) revealed bands of comparable content and molecular

FIG. 2. Immunoblot analysis of the expression of wild type and mutant
RyR1 cDNAs in HEK-293 cells. Samples from whole cell lysates containing about 50 g of protein were separated by 6% SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and probed with monoclonal antibody 34C and secondary horseradish peroxidase-conjugated anti-mouse IgG.
mass. Therefore, the defective [ 3 H]ryanodine binding in microsome preparations from those mutants that were not responsive to caffeine was not due to their poor expression in HEK-293 cells. The fact that the proteins were located in the microsomal fraction indicated that the mutations did not affect membrane insertion. Mutants Q4933A and D4938E, which were weakly responsive to caffeine and 4-chloro-m-cresol, retained normal [ 3 H]ryanodine binding capacity, confirming that an active tetrameric channel was formed.
Sedimentation-In order to determine whether ryanodine binding deficient channels were tetrameric, sedimentation assays in linear sucrose gradients were carried out. The location of wild type RyR1 proteins in gradient fractions was determined using monoclonal antibody 34C. Fig. 5 shows the distribution of the expressed RyR proteins in a linear sucrose density gradient. A single peak of immunoreactivity in the position corresponding to an oligomer in the gradient fractions was found in solubilized preparations from wild type RyR1, mutants G4033A and Q4933A, and all the non-functional mutants in M2, M7, and M10 ( Fig. 5 and Table I). These expressed proteins clearly formed an oligomeric complex of similar size to that formed by the native receptor (31), indicating that they have normal structure but abnormal function.  Table I. In M2, mutation of Glu 4032 to Ala or Gln, but not to Asp, led to loss of channel response to caffeine and 4-chloro-mcresol and to loss of [ 3 H]ryanodine binding, but tetramer formation was retained. In M7, mutation of Asn 4806 to Ala, but not to Gln or Asp, and mutation of Asp 4815 to Ala, but not to Asn or Glu, had similar deleterious effects. It is of special interest that neither M2 nor M7 are among the 4 most hydrophobic sequences first identified by Takeshima et al. (4). Clearly these sequences do play a role in functions associated with transmembrane sequences (38).

Structure/Function Relationships in
In M10, mutation of Asp 4917 to Ala, Asn or Glu, and of Gln 4933 to Asn or Glu led to loss of Ca 2ϩ release channel function, whereas tetramer formation was retained. Asp 4917 seemed to be irreplaceable, since change to Ala, Asn, or Glu abolished channel function. These observations highlight M10 as an interesting sequence for further detailed analysis of channel pore formation. Selected fields containing about 30 cells were challenged first with 20 mM caffeine. Caffeine was washed out after peak amplitude (peak of change in the ratio of fluorescence at 340/380 nm, indicating peak changes in [Ca 2ϩ ] i ) was obtained, thereby restoring the resting Ca 2ϩ level. About 60 s after resting Ca 2ϩ level was achieved, 0.3 mM 4-chloro-mcresol was applied to obtain a second peak amplitude. B and C, averaged peak amplitudes for the 340/380 nm ratio (fluorescence ratio at the highest response to caffeine or 4-chloro-m-cresol minus the ratio at rest) for wild type and mutant RyR1 and for vector alone. Data were averaged and expressed as mean Ϯ S.E. Data for nonresponsive mutants and wild type RyR1 in B also appear in C to facilitate comparison.

FIG. 4. [ 3 H]ryanodine binding to microsome preparations from HEK-293 cells transfected with wild type or mutant RyR1
cDNAs or pcDNA vector alone. [ 3 H]Ryanodine binding was determined by membrane filtration, as described under "Experimental Procedures." Data are expressed as mean Ϯ S.E. in cpm ϫ 10 3 /mg protein.
For mutants nonresponsive to caffeine and 4-chloro-m-cresol, the assay was carried out 3-4 times, but for responsive mutants, the fluorescence measurements were carried out twice.
The reasons for the defects in caffeine response and [ 3 H]ryanodine binding in our mutant channels are not known. The physical structure of the ion-conduction pathway may be altered in these mutants. Since transmembrane sequences of the Ca 2ϩ release channel are likely to form the pore of the channel, substitutions of critical amino acids in the pore region may cause narrowing or distortion of the interior of the pore and thus not allow Ca 2ϩ ions to pass through. It is well known that ryanodine binding to Ca 2ϩ release channels requires opening of the channel. Thus our [ 3 H]ryanodine binding data indicate that those mutant channels that are inactive and do not respond to caffeine are closed. Channel opening is associated with a 4°r otation of the transmembrane region with respect to the cytosolic region (39). Disruption of the channel pore might prevent the rotation of the transmembrane region that leads to channel opening. A second possibility is that transmembrane sequences M2, M7, and M10 form part of the ion selectivity filter (40 -42) in the Ca 2ϩ release channel and that alterations in ion selectivity occur in [ 3 H]ryanodine binding deficient mutants.
A third possibility is that the Ca 2ϩ sensitivity of the mutant channels is decreased. Activation of the Ca 2ϩ release channel by Ca 2ϩ is mediated through high affinity Ca 2ϩ -binding sites that are likely to be located within the channel. Although several Ca 2ϩ -binding sites have been proposed and experimentally demonstrated (3), none of them has been confirmed as the critical site of Ca 2ϩ activation. Strong evidence has been presented to show that negatively charged amino acids in transmembrane sequences are major contributors to the high affinity Ca 2ϩ -binding sites in SERCA molecules (10,14) and in the L-type Ca 2ϩ channel (41). In fact, the high affinity binding site determines the selectivity of L-type Ca 2ϩ channels for Ca 2ϩ over other ions (41). Support for the view that negatively charged amino acids in transmembrane sequences might be major contributors to the high affinity Ca 2ϩ -binding sites in RyR1 comes from the work of Chen et al. (43) who reported that the mutation E3885A in M2 in RyR3 (equivalent to E4032A in RyR1) resulted in the loss of caffeine response in transfected HEK-293 cells, in agreement with our observations. They used single channel recordings in planar bilayers to show that the mutant RYR3 forms a functional channel with normal conductance in which Ca 2ϩ sensitivity for activation is reduced by 10,000-fold. The mutant channel retained normal modulation by ryanodine, caffeine, and ATP. Formation of heterotetramers of wild type and mutant channels by coexpression led to channels with intermediate Ca 2ϩ sensitivity. These observations led Chen et al. (43) to propose that Glu 3885 is a key residue in formation of the Ca 2ϩ sensor of RyR3. Our data for mutant E4032A in RYR1 are consistent with the observations of Chen et al. (43) and extend them to mutants N4806A, D4815A, and D4917A, which also did not respond to either caffeine or 4-chloro-m-cresol when expressed in HEK-293 cells.
A potential inconsistency lies in the fact that Chen et al. (43) demonstrated a response of the E3885A mutant to ryanodine, suggesting that the ryanodine-binding site is intact in activated channels, whereas we could not detect [ 3 H]ryanodine binding to our E4032A or E4032Q mutants in the presence of 10 mM ATP and 100 M free Ca 2ϩ , recreating conditions that activate the E3885A mutant channel. We also tested the E4032A and E4032Q mutants for [ 3 H]ryanodine binding in a novel assay 2 using a series of Ca 2ϩ concentrations, ranging from pCa 8 to 1.5 in the presence of 5 mM caffeine and 1 mM ATP, but we did not find any specific [ 3 H]ryanodine binding in either of these two mutants (data not shown). It is possible that the E4032A and E4032Q mutations are even more insensitive to Ca 2ϩ than the E3885A mutant so that they do not open under the conditions defined. Alternatively, it is possible that these mutations in RyR1 also affect ryanodine binding directly.   Single channel recordings across a planar lipid bilayer provide an excellent assay for mutant RyR1 function. However, some mutants may not form functional channels in the bilayer system (44) and, since only a tiny fraction of molecules is sampled, errors can arise from inadvertent assay of endogenous channels. In unpublished studies using Ca 2ϩ imaging, 3 we have noted that caffeine or halothane-stimulated Ca 2ϩ release occurs in less than 3% of untransfected cells, in line with the observation that RyR1 is not detectable through Western blotting of whole cell lysates (45). Ca 2ϩ photometry and ryanodine binding, in which fluorescence or binding data are gathered from a large number of intact cells, are very useful assays for the screening and partial characterization of RyR1 proteins carrying mutations. In earlier studies, we and others (25,46,47) used Ca 2ϩ photometry to show enhanced caffeine, halothane, and 4-chloro-m-cresol sensitivity in RyR1 molecules carrying mutations associated with malignant hyperthermia. Thus Ca 2ϩ photometry and [ 3 H]ryanodine binding provide a relatively rapid and efficient first screen for the detection of interesting mutants that can then be subjected to more timeconsuming, detailed analyses.
RyR1 Transmembrane Sequences-Although there is no consensus regarding the number of transmembrane sequences in the COOH-terminal region of RyR1, hydrophobicity of four predicted transmembrane sequences is obvious. They are referred to as M1-M4 in the sequence of Takeshima et al. (4), and as M5, M6, M8, and M10 in Zorzato et al. (5). Transmembrane sequences are believed to be key determinants of both tetramer formation and channel pore formation (3,38).
Attempts have been made to analyze the function of the COOH terminus of RyR1, containing transmembrane sequences. A truncated RyR1 transcript is expressed in brain (48). It is probably initiated at Met 4382 , resulting in the expression of the COOH-terminal 656 amino acid residues of RyR1, and would be missing M1 and M2, as defined in Fig. 1. This protein has been expressed in Chinese hamster ovary cells, where it was detectable in the endoplasmic reticulum membrane. However, neither caffeine response nor [ 3 H]ryanodine binding activity could be detected in the expressed, truncated protein. A longer truncated COOH-terminal fragment of RyR1, containing M1 and M2 (see Fig. 1), has also been expressed in Chinese hamster ovary cells (9). This expressed protein formed a functional channel that was shown in single channel recordings to be responsive to ryanodine, but to be otherwise unregulated. This fragment must form a functional tetramer and a channel pore, but regulation may involve other upstream sequences. The NH 2 -terminal portion of RyR1 clearly contributes additional regulatory sites, exemplified by the two MH domains (49). The data presented in this paper lend support to the notion that M2, M7, and M10 are transmembrane sequences that participate in tetramer assembly and channel pore formation.