INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

An increase in intracellular Ca²⁺ concentration is the signal that triggers contraction in muscle cells. A specialized intracellular store, the sarcoplasmic reticulum (SR), functions to finely control myoplasmic [Ca²⁺] in response to depolarization of the plasmalemma. The series of events that allows the transduction of the sarcolemma depolarization into Ca²⁺ release is called excitation-contraction (e-c) coupling. The two membrane systems, plasmalemma and SR, communicate very efficiently with each other at specific structures called calcium release units (CRUs) or junctions. These structures contain proteins that have been identified as key elements in the process: the ryanodine receptors (RyRs), large intracellular channels (~2260 kDa) that allow Ca²⁺ to exit the SR in response to depolarization of the plasma membrane (for reviews see Coronado et al., 1994; Meissner, 1994; Sutko and Airey, 1996; Franzini-Armstrong and Protasi, 1997) and the dihydropyridine receptors (DHPRs), L-type voltage-gated Ca²⁺ channels that are present in the exterior membranes of muscle cells and that control the opening of RyRs (Fosset et al., 1983; Pincon-Raymond et al., 1985; Rios and Brum, 1987; Tanabe et al., 1987, 1988).

While DHPRs and RyRs are the two proteins primarily responsible for transduction of the electrical signal in both cardiac and skeletal muscle cells, it has become increasingly evident that the way in which they communicate with each other is different in the two muscle types. In both types of muscle, depolarization of the plasma membrane activates the voltage-gated DHPRs, thereby allowing Ca²⁺ to enter the cells from the extracellular space. In cardiac muscle, this Ca²⁺ influx through the _1C(₁ cardiac)DHPR is the principal signal that induces the opening of RyR2 and the consequent massive release of Ca²⁺ into the myoplasm (calcium induced calcium release, CICR; Fabiato, 1983, 1985). In fact, e-c coupling in cardiac muscle fails in the absence of extracellular Ca²⁺. In skeletal muscle, however, Ca²⁺ entry is not required for the communication between _1S(₁ skeletal)DHPR and RyR1, and e-c coupling continues in the absence of extracellular Ca²⁺ (Rios et al., 1991; Schneider, 1994). To elucidate the e-c coupling mechanism in skeletal muscle cells it is essential to understand how RyRs and DHPRs communicate with each other without the need for Ca²⁺ as a messenger. Electron microscopy studies reveal that RyR1 and _1SDHPR have a highly specific association that results in the formation of tetrads, groups of four DHPRs linked to subunits of alternate RyRs (Block et al., 1988; Franzini-Armstrong and Kish, 1995; Protasi et al., 1997). Taken together, the structural and functional observations point to a mechanical coupling of RyR1/_1SDHPR similar to the mechanism first suggested by Schneider and Chandler (1973). The DHPR functions as a voltage sensor that changes its conformation in response to depolarization, and this change in conformation then triggers RyR opening directly without the need of a diffusible messenger (orthograde signaling). RyR1, in turn, affects gating properties of the DHPR (Nakai et al., 1996; Avila and Dirksen, 2000), with the result that the amplitude of the L-type Ca²⁺ current is increased (retrograde signaling).

Two animal models have been extremely valuable tools in dissecting skeletal e-c coupling: the dysgenic mouse (mdg), having a spontaneous mutation in the _1SDHPR gene that eliminates slowly activating Ca²⁺ current and intramembrane charge movement (Beam et al., 1986; Adams et al., 1990; Chaudhari, 1992), and the dyspedic mouse, having a targeted null mutation of RyR1 (Takeshima et al., 1994; Buck et al., 1997). In both cases e-c coupling fails, but can be restored by transfection of the myotubes with cDNA encoding the missing protein (Tanabe et al., 1988; Nakai et al., 1996; Moore et al., 1998), thus directly proving the primary roles of _1SDHPR and RyR1 in e-c coupling. The identification of key domains of _1SDHPR that allow communication with RyR1 has come from the use of skeletal-cardiac and, more recently, skeletal-insect chimeras. In these chimeric DHPRs the II-III loop of the skeletal DHPR, or even a shorter subdomain of it (containing only 46 amino acids), is sufficient to restore skeletal-type e-c coupling and retrograde signaling (Tanabe et al., 1990; Nakai et al., 1998b; Grabner et al., 1999). Even drastic alteration of the sequence surrounding those 46 amino acids does not abolish the ability to support skeletal type e-c coupling or retrograde signaling (Wilkens et al., 2001). For the RyR, an analogous approach has been taken of constructing chimeras in which portions of RyR2 (the cardiac RyR isoform) are replaced with the corresponding segments of RyR1. Expression of such chimeras in dyspedic myotubes shows that both skeletal-type e-c coupling and retrograde signaling are strongly restored by the chimera R10, which contains skeletal residues 1635-2559, whereas only retrograde coupling is strong for the chimera R9, which contains skeletal residues 2659-3720 (Nakai et al., 1998a). Subsequently, it has been shown that the chimera R16, which contains RyR1 residues 1837-2154, mediates weak skeletal-type coupling (Proenza et al., 2002).

Continuing the search for the molecular and structural basis for RyR1/_1SDHPR interactions, we have expressed RyR1, RyR2, and four different RyR1/RyR2 chimeras (R9, R10, R16, and also R4, which contains skeletal residues 1635-3720) in 1B5 cells (a mouse skeletal muscle cell line that carries a null mutation for RyR1) and tested for a correlation between the presence of tetrads and Ca²⁺-entry independent e-c coupling. Differentiated 1B5 cells develop a SR system that forms junctions with the surface membrane and with primitive transverse (T)-tubules despite the lack of RyRs (Protasi et al., 1998). These junctions contain triadin and DHPRs, but of course lack RyRs, or feet (dyspedic CRUs), and thus do not support Ca²⁺ release in response to depolarization, caffeine or 4-m-chloro-cresol (Moore et al., 1998; Fessenden et al., 2000). Dyspedic CRUs (dCRUs) in 1B5 cells resemble junctions in developing myotubes of RyR1-null mice (Takeshima et al., 1994; Takekura et al., 1995; Takekura and Franzini-Armstrong, 1999). In the absence of RyRs, DHPRs do not maintain the normal tetradic arrangement that is found in wild-type skeletal muscle cells (Protasi et al., 1998). Arrays of DHPR tetrads and skeletal-type e-c coupling can be restored in differentiated 1B5 cells by transfection with RyR1 cDNA (Fessenden et al., 2000; Protasi et al., 2000). Our results with chimeric RyRs show that the presence of at least two non-overlapping regions of RyR1 primary sequence are sufficient to restore tetrads and skeletal-type e-c coupling, although with a somewhat variable degree of success, suggesting multiple sites of interaction between RyR1 and the DHPR. Interestingly, the RyR1 sites responsible for functional coupling with the DHPR only partially correlate with those that allow the structural linkage between the two proteins.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culturing

The methods used to create the 1B5 cell line are described in detail elsewhere (Moore et al., 1998). The cells were expanded at 37°C in low-glucose DME medium containing 20% fetal bovine serum, 100 U/ml penicillin, 100 µg/ml streptomycin, and an additional 2 mM L-glutamine (growth medium). After ~48 h the cells were re-plated in either 1) 35-mm dishes containing thermanox coverslips for electron microscopy and immunocytochemistry (Nunc Inc., Naperville, IL); or 2) in 96-well plates with ultra-thin, clear bottoms for Ca²⁺ imaging experiments (Corning Incorporated, Costar, NY) coated with Matrigel (Collaborative Biomedical Products, Bedford, MA). When cells reached ~70% confluence, growth medium was replaced with differentiation medium (growth medium with 5% heat-inactivated horse serum replacing the 20% fetal bovine serum) to induce differentiation. The medium was changed daily.

cDNA packaging in HSV-1 virions and cell transfection

The cDNAs encoding for RyR1, RyR2, and the four RyR1/RyR2 chimeras were packaged into HSV-1 amplicon virions using the helper virus-free packaging system. The methods are described in detail elsewhere (Fraefel et al., 1996; Wang et al., 2000). Four to five days after the beginning of differentiation, the cells were infected with 1 ml differentiation medium containing HSV1 virions at 4 × 10⁵ infectious units/ml (a moiety of infection of ~4). This mixture was removed ~2 h later and replaced with differentiation medium. The cells were either fixed or imaged ~24-36 h later.

Immunohistochemistry

The cells were fixed in methanol for a minimum of 20 min at 20°C, blocked in PBS (phosphate-buffered saline: 26.7 mM Na₂HPO₄, 7.3 mM KH₂PO₄, 136.8 mM NaCl, 2.6 mM KCl) containing 1% BSA and 10% goat serum for 1 h, incubated first with primary antibodies and then with secondary antibodies (cyanine 3 conjugated; Jackson ImmunoResearch Laboratories, Lexington, KY), respectively, for 2 h and 1 h at room temperature. Code, specificity, working dilution, original reference, and the sources of primary antibodies are as follows: 34C, recognizes RyR1, R4, and R9, 1:20, Airey et al., 1990, Developmental Studies Hybridoma Bank, The University of Iowa; C3-33 recognizes RyR2, R10, and R16, 1:20, gift of G. Meissner (Junker et al., 1994). The specimens were viewed on an inverted fluorescence microscope (Olympus IX70).

Electron microscopy

The cells were washed twice in PBS at 37°C, fixed in 3.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, and then kept in fixative for up to one to four weeks at 4°C before further use. For thin-sectioning the cells were post-fixed in 2% OsO₄ for 2 h at room temperature and then contrasted in saturated uranyl acetate either for 4 h at 60°C or overnight at room temperature. The samples were embedded in Epon 812, and the sections stained in uranyl acetate and lead for ~8 min each. For freeze-fractures the glutaraldehyde-fixed cells were infiltrated with 30% glycerol. A small piece of the coverslip was mounted with the cells facing a droplet of 30% glycerol, 20% polyvinyl alcohol on a gold holder, and then frozen in liquid nitrogen-cooled propane (Cohen and Pumplin, 1979; Osame et al., 1981). The coverslip was flipped off to produce a fracture that followed the culture surface originally facing the coverslip. The fractured surfaces were shadowed with platinum unidirectionally at 45^o and then replicated with carbon in a freeze-fracture apparatus (Balzers, model BFA 400; Balzers S.p.A., Milan, Italy). Sections and replicas were photographed in a 410 Electron Microscope (Philips Electron Optics, Mahwah, NJ).

Ca²⁺ imaging

Twenty-four to thirty-six hours after transduction, cells were loaded with the Ca²⁺ fluorescent dye, Fluo-4 AM (Molecular Probes, Eugene, OR) to monitor intracellular changes in [Ca²⁺]. The loading procedure was 1) differentiation media was removed and cells were washed twice with imaging buffer (IB) containing 125 mM NaCl, 5 mM KCl, 1.2 mM MgSO₄, 6 mM glucose, 25 mM HEPES, 0.05% BSA, 2 mM CaCl₂, pH 7.4; 2) the cells were incubated for 30 min in 100 µl IB containing 5 µM Fluo-4 AM; 3) cells were washed again with IB. The 96-well plates, containing a monolayer of 1B5 cells transduced with one of the RyR constructs, were placed on the stage of a Nikon Diaphot 300 inverted microscope equipped with an Olympus Uapo/340 40× oil immersion objective (numerical aperture 1.35) for study at room temperature (~21°C). The microscope was modified to incorporate a pair of separate 3-D micromanipulators on either side of the vertical post holding the condenser. Each well in the 96-well plate could be perfused in a controllable, accurate way using an AutoMate 8-channel air pressure-controlled system incorporating multiple 50-ml reservoirs (Automate Scientific Inc., Berkeley, CA). Each reservoir could be rapidly switched into or out of the perfusion pathway with a ~50 µl dead volume. The 3-D micromanipulators were used to place a pair of capillaries precisely into each well: 1) an inlet to allow delivery of the desired medium, and 2) a vacuum suction outlet to keep the level of medium in the well constant. The perfusion inlet was positioned ~1 mm above the imaged cell/myotube to allow a very efficient and rapid change of solution over the imaged area. Before starting the experiments IB was replaced in all Fluo-4 AM-loaded wells with nominally Ca²⁺-free IB, in which the CaCl₂ was replaced with 0.5 mM CdCl₂ and 0.1 mM LaCl₃ to block Ca²⁺ currents through DHPRs. Four regions with a size of ~1/2 of the cell diameter were selected in four different myotubes. The selected regions were from apparently well-differentiated (large) myotubes and contained no nuclei. The cells were imaged using a PTI delta-RAM as the light source with a Stanford Photonics 12-bit digital intensified CCD and the data displayed and analyzed using QED imaging software (v1.3. QED Software, Pittsburgh, PA). The four areas were simultaneously recorded using the strip chart utility in the QED software to monitor image intensity. In each well, the cells were first perfused for at least 1 min with Ca²⁺-free IB and then stimulated by separate 15-s exposures to Ca²⁺-free 80 mM K⁺ and 40 mM caffeine, with a 15-s recovery period in between. The Ca²⁺-free 80 mM K⁺ consisted of Ca²⁺-free IB, which had equimolar substitution of 80 mM K⁺ for Na⁺. The resulting fluorescence changes were corrected for background by subtraction of the average fluorescence value in the 5 s preceding the test stimulus. Two measures were used to estimate the ability of each of the RyR constructs to restore skeletal-type e-c coupling. The first was to determine the number of cells producing a Ca²⁺ transient in response to K⁺ depolarization relative to the number of cells producing a Ca²⁺ transient in response to caffeine. The second was to determine the ratio of the magnitudes of the responses to K⁺ depolarization and caffeine, respectively.

Preparation of figures

Pictures and negatives were scanned using a Color Flatbed Scanner UMAX Power Look II at 300 dpi. Figures were mounted and labeled using Adobe Photoshop v5.5, Canvas v3.5.4 (Deneba Software), and Microsoft Power Point 98.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Immunocytochemistry

Differentiating 1B5 cells have been previously described in detail (Protasi et al., 1998). Of relevance to this work is the relationship between the formation of large multinucleated myotubes and the presence of dCRUs (dyspedic peripheral couplings, dyads, and triads) containing co-localized DHPRs and triadin foci (Protasi et al., 1998, 2000).

Infection of 1B5 cells with HSV-1 amplicon virions containing cDNA encoding any of the six constructs tested in this work (Fig. 1) results in a large fraction of all cells expressing protein at a substantial level. Fig. 2 illustrates low-magnification images of myotubes immunostained with anti-RyR antibodies (either 34C or C3-33) 24-36 h after infection. The number of cells stained with anti-RyR antibodies agrees with functional results showing that ~80% of the cells respond to caffeine in each sample group (see Ca²⁺ imaging section, Fig. 8, and Table 2 for more details). Infection with HSV-1 amplicon virions does not appear to cause any cell death, obvious changes in cell size and shape, or the level of differentiation. High-magnification immunofluorescence images reveal that all RyRs expressed in our cells were clustered in intense foci (Fig. 3). Focusing up and down clearly shows that the majority of these immunopositive foci are located at, or very near, the surface membrane, on both ventral and dorsal sides of the myotubes. The cortical localization of these positive foci coincides with that previously shown for CRUs in both RyR1-containing and RyR1-lacking myotubes (Protasi et al., 1998, 2000). Control 1B5 myotubes (non-transduced) do not react with either one of the antibodies (34C against RyR1 and RyR3 and C3-33 against RyR2) used to detect RyR expression in transduced myotubes (Fig. 2, A and B, Fig. 3, A and B), confirming that these cells do not express detectable levels of any RyR.

View larger version (27K): [in this window] [in a new window]   FIGURE 1   Schematic representation of RyR chimeras. Skeletal sequence (from RyR1) is represented in dark gray; cardiac sequence (from RyR2) is represented in lighter gray. Amino acid composition of the chimeras R4, R9, and R10 are as in Nakai et al., 1998a. R4: Ca(1-1625), Sk(1635-3720), Ca(3687-4968); R9: Ca(1-2624), Sk(2659-3720), Ca(3687-4968); R10: Ca(1-1625), Sk(1635-2636), Ca(2603-4968). Amino acid composition of R16 is as in Proenza et al., 2002: Ca(1-1817), Sk(1837-2154), Ca(2118-4968).

View larger version (163K): [in this window] [in a new window]   FIGURE 2   HSV1 virions give transfection of high efficiency. (A and B) Lack of labeling of control cells with 34C and C3-33, the two antibodies used to detect RyR1, RyR2 and the four chimeric RyRs. (C-H) Immunostaining of transduced cells shows that a high percentage of myotubes react positively to anti-RyR antibodies. Functional studies confirm that ~80% of the cells respond functionally to caffeine (Fig. 8, Table 2). Bar: 200 µm.

View larger version (147K): [in this window] [in a new window]   FIGURE 3   RyRs are clustered in foci. (A and B) Untreated cultures show no labeling with either 34C or C3-33 antibodies. (C-H) All RyRs expressed in 1B5 cells cluster within intensely labeled foci located at the periphery of large and differentiated myotubes. The images are not confocal, but were taken with a high numerical aperture objective. The focal plane in all cases is at or close to the lower, flat surface of the cell. There are no obvious differences between the various constructs. Clustering in foci is consistent with the correct targeting of all proteins to CRUs, as confirmed by thin sectioning (see Fig. 4). Bar: 25 µm.

Electron microscopy

Thin sections

View larger version (87K): [in this window] [in a new window]   FIGURE 4   RyRs are targeted to junctions between SR terminal cisternae and the exterior membrane. (A) Control 1B5 cells have dyspedic junctions, i.e., no feet are visible in the junctional gap between the two membranes. (B) RyR1 is clustered in fairly extensive and ordered arrays (arrows). (C) RyR2 is usually present within small junctions and makes smaller clusters (arrow, compare with RyR1). (D and E) R4 and R9 make fairly extensive and ordered arrays, similarly to RyR1 (arrows). (F and G) R10 and R16 are present within the junctional gap, but the individual feet are not as clearly visible (arrows). This may be due to variable orientation and/or disordered array (see Fig. 7 B), resulting in poor overlap of the two layers of feet that are included in a thin section of standard thickness (50-60 nm). Bar: 0.1 µm.

Freeze-fracture

View larger version (175K): [in this window] [in a new window]   FIGURE 5   RyR1 and R4 restore extensive DHPR junctional tetrads in surface membranes of 1B5 myotubes. (A) In control 1B5 cells, DHPRs are clustered within junctional domains but they do not form tetrads (see Protasi et al., 1998). (B) RyR1 expression restores DHPR tetrads and tetrad arrays, even though some tetrads are incomplete (see Protasi et al., 2000 for further details). The boxed tetrad is shown at higher magnification in the inset. (C) RyR2 expression does not restore complete DHPR-tetrads, but occasional groups of three tetrad-like particles (boxed) are seen. (D) Expression of R4, the chimera containing the skeletal segment 1635-3720, restores DHPR tetrads even though some of them are incomplete. The boxed tetrad is shown at higher magnification in the inset, and well-defined tetrads from other R4-expressing cells are shown in the inset at the lower left and far right of panel E. (E) Dotting the center of each tetrad in the same array as shown in panel D makes it evident that the tetrads are arrayed in an ordered orthogonal arrangement. The spacing between the dots marking the center of tetrads is consistent with tetrads being superimposed on alternate RyRs, indicating that the normal skeletal 2:1 ratio between feet and tetrads is restored. Bars: 0.1 µm.

View larger version (170K): [in this window] [in a new window]   FIGURE 6   R9, R10, and R16 restore junctional tetrads and, to some extent, tetrad arrays. (A and B) An array of DHPR tetrads in an R9-expressing cell: dots marking the center of complete and incomplete tetrads form an orthogonal arrangement much like tetrad arrays in normal skeletal muscle cells and in 1B5 myotubes expressing R4 (see Fig. 5, D and E) or RyR1 (Protasi et al., 2000). Higher-magnification images are shown for the boxed tetrad in A (inset) and for four tetrads from other R9-expressing cells are below panel B. (C, D and F, G) Clusters of tetrads in cells expressing R10 and R16, respectively. As shown by the dots, the tetrads form orthogonal arrays in which the order extends for only a limited distance. The examples shown here are among the most ordered clusters found for R10 and R16. (E) A cluster of DHPRs in another R10-expressing cell. Higher magnification views of the tetrads outlined by squares are shown in the insets in E and F, and additional higher-magnification views of selected tetrads from R10- and R16-expressing cells are shown below E and F, respectively. The presence of only short-range order in the tetrad arrays for R10 and R16 is probably due to the fact that these two chimeras fail to assemble into ordered arrays of feet in the SR membrane (see Figs. 4 and 7 for comparison). Bars: 0.1 µm.

View larger version (37K): [in this window] [in a new window]   FIGURE 7   Models depicting possible reciprocal arrangements of RyRs and DHPRs within calcium release units of 1B5 cells expressing wild-type and chimeric RyRs. See text for details. At the top of each panel both RyRs (gray) and DHPRs (black) are represented. At the bottom only DHPRs are shown to mimic what is seen in freeze-fracture replicas. (A) Tetrad arrays are quite well-ordered in 1B5 cells expressing RyR1, R4, and R9, suggesting that both chimeras contain in their sequence the RyR region necessary for the physical link with 1SDHPRs. (B) R10 and R16 restore tetrads, but the tetrad arrays seem to be less ordered than in cells expressing R4 or R9, because these two chimeras do not form well-organized arrays of feet in the SR membrane. (C) RyR2 expression does not restore DHPR tetrads in 1B5 myotubes, but an occasional group of three-tetrad-like particles is seen.

Ca²⁺ imaging

Fig. 8 A illustrates Ca²⁺ transients recorded from control and RyR-transduced 1B5 cells in response to two sequential stimuli: Ca²⁺-free 80 mM KCl and 40 mM caffeine (for each construct, the responses of two separate cells are shown). All experiments were performed in nominally Ca²⁺-free IB, supplemented with 0.5 mM CdCl₂ and 0.1 mM LaCl₃ to completely block Ca²⁺ currents through DHPRs. A response to caffeine indicates the presence of RyRs in the cell being examined, whereas a response to the Ca²⁺-free 80 mM K⁺ suggests the occurrence of depolarization-induced intracellular Ca²⁺ release that does not rely upon entry of extracellular Ca²⁺ (i.e., skeletal-type e-c coupling). Control cells do not respond to either Ca²⁺-free 80 mM KCl or 40 mM caffeine, which is expected for cells that do not express RyRs. In all the transduced cultures tested, ~80% of cells (of at least 140 for each group) respond to caffeine (see Table 2), suggesting a very high efficiency of transfection/expression. Of those cells that produce caffeine transients, a variable fraction also produces a transient in response to Ca²⁺-free 80 mM KCl. The fraction of cells that respond to depolarization, and the normalized peak amplitude of this depolarization-response, were taken as the two criteria of the ability of each construct to mediate skeletal-type e-c coupling (Fig. 8 B). The depolarization-response is normalized by the caffeine-response as a way of correcting for possible variations in level of RyR expression, relative volume of junctional SR, extent of SR Ca²⁺ loading, and loading of the indicator dye (Fluo-4). In many cases the magnitude of K⁺-induced transients is larger than that of the subsequent caffeine response, especially in RyR1- and R4-expressing myotubes. Because K⁺ depolarization elicits large transients in RyR1- and R4-expressing cells, and because the caffeine challenge occurs only 15 s after the K⁺ depolarization, it seemed possible that the SR may have been partially depleted before the application of caffeine. To test this possibility, we performed identical experiments on primary myotubes from wild-type mice (results not shown) that express more constant amounts of both _1SDHPR and RyR1. In these experiments we observed that regardless of the order in which the two pulses were applied (K⁺ first or caffeine first), the response to K⁺ was always larger than the response to caffeine. This result suggests that 80 mM K⁺ depolarization is more effective than 40 mM caffeine as a Ca²⁺-releasing stimulus.

View larger version (29K): [in this window] [in a new window]   FIGURE 8   (A) Response of control and RyR-expressing cells to K+-induced depolarization and caffeine. All the experiments were performed in nominal Ca2+-free solution containing 0.1 mM La3+ and 0.5 mM Cd2+ to block Ca2+ entry through the DHPR. Data are plotted as F/Fmax, where Fmax is the peak amplitude of the transient induced by caffeine (40 mM), and the pre-stimulus fluorescence background has been subtracted. In each panel traces from two simultaneously imaged cells are shown superimposed. Control cells did not respond either to K+ or to caffeine. In transduced cells, responses to Ca2+-free 80 mM KCl and to caffeine were restored, but the relative magnitude of the responses varied with the construct. RyR1-expressing cells responded to Ca2+-free 80 mM KCl with a Ca2+ transient often larger than the caffeine one. RyR2 expressing cells gave very small Ca2+ transients in response to Ca2+-free 80 mM KCl in very few cells (<10%), but a robust caffeine response. Various chimeras showed an intermediate behavior: R4 gave the largest responses, followed by R10 and R16, with R9 giving the smallest responses. Note that in the experiments illustrated for RyR2 and R9, only one of the two cells responded to Ca2+-free 80 mM KCl and that one of each of the cells expressing RyR2 and R16 showed spontaneous activity. Quantitative analysis of functional experiments is reported in Table 2. (B) Comparison of intracellular Ca2+ responses of wild-type and chimeric RyRs. Light gray bars: percentage of RyR-expressing cells (i.e., cells responding to caffeine) that respond to Ca2+-free 80 mM KCl (data from Table 2). Dark gray bars: magnitude of the Ca2+-free 80 mM KCl transient as a percentage of the caffeine transient in the same cell (see Methods). RyR1 gave both the highest fraction of cells responding to Ca2+-free 80 mM KCl and the largest amount of Ca2+ released, while RyR2 had the lowest values for both parameters. The chimeric RyRs gave intermediate results in the order: R4 > R10 > R16 > R9.

RyR1 is the most effective construct at restoring skeletal-type e-c coupling: a large fraction of RyR1-expressing cells respond to depolarization, giving a normalized response of large amplitude (Fig. 8 B and Table 2). By contrast, a small fraction of RyR2-expressing cells respond to Ca²⁺-free 80 mM KCl, and the normalized amplitude of the response is much smaller. Among the chimeras, R4 is the most effective in restoring Ca²⁺ release induced by Ca²⁺-free 80 mM KCl, both in number of cells responding and in the magnitude of the transient. R10, R16, and R9 follow this in order of decreasing effectiveness. However, even R9 gives much more frequent and larger responses to Ca²⁺-free 80 mM KCl than did RyR2 (Fig. 8 B and Table 2). Overall, Fig. 8 B shows a striking correspondence between the frequency of cells that were capable of a response to the Ca²⁺-free K⁺-challenge and the normalized magnitude of the Ca²⁺ transient. Based on these data, the constructs can be ordered based on their efficiency in restoring skeletal-type e-c coupling: RyR1 > R4 > R10 R16 > R9 RyR2. Note that in addition to releasing Ca²⁺ in response to stimulation, the transduced cells also produce spontaneous Ca²⁺ transients. This occurs very rarely in RyR1-expressing cells and very frequently in RyR2-expressing cells, with the chimeras showing intermediate behaviors. In Fig. 8 A, spontaneous activity is clearly visible in one of the traces in the RyR2 and R16 panels.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In the present study we have analyzed freeze-fracture replicas and intracellular Ca²⁺ transients of 1B5 myotubes expressing various RyR constructs, with the goal of comparing the ability of these constructs to restore skeletal-type e-c coupling and their ability to organize _1SDHPRs into junctional tetrads. The constructs examined were the skeletal-RyR1, the cardiac-RyR2, and the chimeric RyR1/RyR2 constructs R4 (Sk: 1635-3720), R9 (Sk: 2659-3720), R10 (Sk: 1635-2559), and R16 (Sk: 1837-2154). Based on immunostaining, thin-section electron microscopy, and response to caffeine all of these proteins were expressed at similar levels, targeted to junctions between the SR and plasma membrane and functional as Ca²⁺ release channels. With respect to their ability to organize _1SDHPRs into tetrads, they ranked in decreasing effectiveness as RyR1 > R4 = R9 > R10 = R16 RyR2. As judged by Ca²⁺ transients elicited by K⁺ depolarization in a Ca²⁺-free external medium, the rank order for restoration of skeletal-type e-c coupling was RyR1 > R4 > R10 > R16 > R9 RyR2.

The method we used to test for skeletal e-c coupling (15 s in a Ca²⁺-free medium containing Cd²⁺ and La³⁺) made it possible to measure internal Ca²⁺ releases of varying magnitudes in large numbers of cells and thus to obtain a rank-order of the ability of the various RyR constructs to restore skeletal-type e-c coupling (Fig. 8 B and Table 2). Several of these RyR constructs have also been examined in previous physiological studies. In one study, Nakai et al. (1998a) applied 10 ms electrical stimuli to Fluo-3 AM-loaded myotubes in a Ca²⁺-free medium and found that RyR1-, R4-, and R10-expressing cells produced similar Ca²⁺ transients (indicative of skeletal e-c coupling), whereas R9 and RyR2 failed to produce transients. More recently, Proenza et al. (2002) found that 10-ms stimuli caused 10/13 intact myotubes expressing RyR1 to contract in a Ca²⁺-free medium, whereas 3/30 expressing R16 contracted, and 0/30 expressing RyR2 contracted. Thus, these previous results are consistent with a rank-order of RyR1/R4/R10 > R9/RyR2 (Nakai et al., 1998a) and RyR1 > R16 > RyR2 (Proenza et al., 2002), which is generally in agreement with the rank-order found for skeletal e-c coupling in our present work (Fig. 8 B). However, unlike the present work, the earlier studies failed to produce evidence for skeletal-type coupling for R9 or RyR2. Most likely, the much longer depolarizations used in the present work (15 s) made it possible to detect releases too small to be detected in response to the briefer stimuli (10 ms). In addition, in the present work we used HSV1 virions to transduce 1B5 myotubes. This system delivers cDNA into muscle cells much more efficiently and gives a much higher expression level than the mono-nuclear injection of primary myotubes (Wang et al., 2000; Fessenden et al., 2000; Protasi et al., 2000). It is also important to note that spatially averaged measurements of myoplasmic free Ca²⁺ are only an indirect assay of RyR activation, which means that different assays can give different quantitative estimates of the relative ability of different RyRs to mediate e-c coupling. For example, the data of Fig. 8 B indicate that R16 is about half as effective as RyR1 in mediating skeletal e-c coupling. By contrast, the average value of peak Ca²⁺ transients in R16-expressing myotubes was found to be only ~14% of that in RyR1-expressing cells when measured with 200-ms depolarizations applied via whole patch clamping (Proenza et al., 2002). Whether or not this is actually a better estimate, it emphasizes that while the data of Fig. 8 B are important for providing a rank ordering, they cannot be taken as a linear comparison between constructs.

Of the chimeras examined, R4 (which contains the most RyR1 sequence) is closest to RyR1 in ability to restore skeletal e-c coupling (Fig. 8 B). Moreover, the function of R4 can be partially carried out by subregions within its amino-terminal half (R10 and the shorter segment, R16). The carboxyl-terminal half of R4 (contained in R9 and not in R10 or R16) also rescues e-c coupling, although less effectively than either the R10 or R16 segments. Thus, we conclude that skeletal e-c coupling depends on a functional interaction between _1SDHPR and RyR1, which involves at least two separate regions of RyR1, and that the full functional interaction results from additive effects of two or possibly more regions. A similar conclusion was also reached in an earlier study of R16 and its complementary construct, "R16-reverse" (Proenza et al., 2002).

As mentioned earlier, signaling between the _1SDHPR and RyR1 is bidirectional: in addition to e-c coupling (orthograde signaling), there is also retrograde signaling that increases the magnitude of Ca²⁺ current carried via the _1SDHPR (Nakai et al., 1996). Based on measurements of Ca²⁺ current density in dyspedic myotubes expressing RyR constructs (Nakai et al., 1998a; Proenza et al., 2002), the rank-order for retrograde signaling is RyR1 > R4 = R9 > R10 R16, that is, very similar to their ability to organize _1SDHPRs into tetrads. Thus, the rank-orders of R9 and R10 in orthograde signaling and retrograde signaling/tetrad restoration are opposite. The bidirectional signaling between _1SDHPR and RyR1 strongly suggests a mechanical interaction between the two proteins, and it is difficult to imagine how such a mechanical interaction could occur without physical linkages. In this regard, the occurrence of tetrads is very important because it is indicative of a stereospecific DHPR/RyR association that, in turn, implies physical linkages. We observed tetrads for all those RyR constructs that were more effective in restoring skeletal e-c coupling. The different constructs showed a slightly different rank-order for the restoration of tetrads (Table 1) than for the restoration of skeletal e-c coupling (Fig. 8 B and Table 2). In particular, the subdomain of R4 that is most effective in forming tetrads (the one contained in R9) is not the segment that is most effective in restoring skeletal type e-c coupling (the one contained in R10 or R16). Although this lack of correlation might appear to be at odds with the notion that e-c coupling requires physical linkages, it should be recalled that the signaling between RyR1 and _1SDHPR is bidirectional, and that for retrograde signaling R9 appears to be more effective than R10. Furthermore, one would suppose that the interaction between the _1SDHPR and a chimeric RyR responsible for e-c coupling would depend on the presence of two features: 1) one or more sites of the chimeric RyR that participates in physical linkage to the _1SDHPR, and 2) structures within the RyR that couple mechanical alteration of the linkage site to opening of the Ca²⁺ release channel. A chimera could contain regions that participate in physically linking to the DHPR but lack all or part of the appropriate structures for the intermolecular conformational changes (within the RyR) which convert _1SDHPR conformational changes to RyR channel opening. According to this reasoning, it is not necessarily surprising that a chimera unable to mediate effective e-c coupling could nonetheless be effective at organizing DHPRs into tetrads. However, the hypothesis that e-c coupling depends on mechanical interactions would be called seriously into question if one were to find an RyR construct that was functionally equivalent to, but that lacked the ability to, organize DHPRs into tetrads. That was not the case in our results.

As for functional rescue by the RyR chimeras, the structural rescue of tetrads appears to depend upon two separate regions of RyR1. Both the R4 region and its carboxyl-terminal half, R9, are sufficient to produced organized arrays of DHPR tetrads. However, the amino-terminal half of R4 also appears to be involved to some extent in the physical interaction, because both R10 and its R16 subdomain are able to restore tetrads to a lesser degree (Table 1). Although the entire R4 region represents a substantial amount of primary sequence, folding of this region might produce a single pocket that physically links to the DHPR. Fig. 9 illustrates a hypothetical model of how the folded RyR might interact with the _1SDHPR II-III loop. Of course, neither the functional nor the freeze-fracture data provide any evidence as to whether the physical linkage between _1SDHPR and RyR1 is direct or involves other proteins. A recent report by O'Reilly et al. (2002) suggests a role of FKBP12 in the interaction between the II-III loop and RyR1. However, yeast two-hybrid analysis demonstrates a direct in vitro interaction between the R16 region of RyR1 and a 46-residue segment of the _1SDHPR II-III loop (Proenza et al., 2002), consistent with the idea that these two regions contact one another in vivo. Interestingly, a somewhat weaker interaction was also observed between the II-III loop segment and the portion of RyR2 that corresponds to the R16 region of RyR1 (Proenza et al., 2002). If this interaction can also occur in vivo, it might help to explain the very weak functional and structural coupling that we observed between RyR2 and _1SDHPR.

View larger version (78K): [in this window] [in a new window]   FIGURE 9   Model suggesting how noncontiguous regions of RyR1 could fold to interact with the 1SDHPR. The chimeras examined in this study contain RyR1 regions that are all part of the foot region of the protein according to the folding models of both Takeshima et al. (1989) and Zorzato et al. (1990). The R4 region, shown in three gradations of gray, is illustrated as interacting with the II-III loop of the 1SDHPR, which has been shown to be critical for skeletal-type e-c coupling (Tanabe et al., 1990; Grabner et al., 1999; Wilkens et al., 2001). Within R4, R10 (1635-2636, medium gray) and its R16 subdomain (1837-2154, dark gray) are much more effective in carrying out skeletal e-c coupling than is R9 (2659-3720, lighter gray); the latter region is shown as having a greater region of contact with the II-III loop because R9 is more effective at organizing DHPRs into tetrad arrays. However, our data cannot exclude the possibility that domains of the 1SDHPR other than the II-III loop, or of RyR1 outside of the R4 region, are involved in "mechanical" coupling.