Ryanodine receptor Ca2+ release channels: does diversity in form equal diversity in function?

I. INTRODUCTION

It is well established that changes in the level and distribution of intracellular calcium are used to regulate many different cellular processes. During the past decade, the use of calcium-sensitive dyes and photoproteins in combination with increasingly sophisticated imaging hardware and software has led to a more detailed picture of how calcium levels change inside of cells. A common finding that has emerged from these studies is that calcium signals are complex. The changes seldom involve a simple shift from one steady-state level of cytoplasmic calcium to another, but frequently they are regulated in the three spatial dimensions as well as time. This results in calcium oscillations and calcium waves that assume different formats and that can be either global or localized to specific regions of the cell (25, 26, 179, 272, 350, 352). In addition, these signals may involve not only changes in cytoplasmic calcium but also less well understood alterations in the level of calcium within cellular organelles, including the nucleus and mitochondria, and in the content of calcium in the storage pools themselves.

As might be anticipated, the generation of such complex signals appears to require an equally diverse array of proteins specialized for releasing, binding, sequestering, and storing calcium within cells. Functionally specialized forms of the proteins involved in each of these activities have been identified. Moreover, in an increasing number of cases, more than one form of these proteins is coexpressed within the same cell, creating the potential for multiple calcium release systems having different operational properties. In some instances, these systems may be arranged in parallel to generate independent calcium signals, while in others, they may work in series and provide the basis for complex regulation and amplification of a single calcium release event.

In this article, we focus on one of two families of proteins that release calcium from intracellular stores, the ryanodine receptors (RyR). During recent years, there have been a significant number of reviews concerning the RyRs and closely related topics (48, 62, 72, 85, 94, 148, 196, 224, 228, 229, 255, 276, 283, 284, 295, 307, 315, 340, 357, 379). Rather than recapitulate the fundamental information covered in these articles, we have taken advantage of this situation to explore more speculative topics. In particular, we examine the diversity existing within this family of proteins and explore how this diversity may contribute to the properties of the calcium signals mediated by the RyRs. As specific examples, we consider the complexities associated with coexpression of RyRs in mature vertebrate skeletal muscles, whose functional properties do not lead to an obvious need for multiple calcium release channels, and embryonic muscles, where diversity in many cellular processes may provide the flexibility essential for meeting the changing needs of a developing tissue. Although we focus on the RyRs in this article, a similar story could be constructed for the other family of intracellular calcium release channels, the inositol trisphosphate receptors (I[P.sub.3]R), which also consists of multiple isoforms and alternatively spliced variants and involves the coexpression of different family members within the same cell (32, 65, 233, 246, 251, 293, 327).

II. MOLECULAR DIVERSITY

A. Vertebrate Ryanodine Receptor Isoforms and Splice Variants

The vertebrate RyRs identified to date are homotetramers composed of polypeptide subunits with molecular masses of 500-600 kDa encoded by mRNAs in excess of 15 kb. These proteins are similar in their overall topology, and the ion channel-forming membrane-spanning regions are highly conserved and localized to the COOH-terminal of the protein. The RyR channel domain has not been completely defined and has been suggested to consist of either 4 (343) or 10 (394) membrane-spanning regions. In either case, the remaining N[H.sub.2]-terminal part of the RyR, which comprises [approximately]80% of the protein mass, forms a large cytoplasmic foot domain that is quatrefoil in shape (194, 300, 367). Cryoelectron microscopic images reveal a four-fold symmetry with a channel extending as a single pore from the SR luminal end through the central region of the receptor. Further details about the molecular features of the RyRs can be found in several recent reviews (62, 255, 315, 340).

The RyRs comprise a growing family of proteins. Messenger RNAs for three RyRs have been cloned and sequenced from mammalian tissues and found to be encoded by separate genes: ryr 1 from skeletal muscle (217, 220, 259, 260, 343, 394), ryr2 from cardiac muscle (220, 247, 259, 260), and ryr3 from brain (119, 139, 220, 314). Both RyR1 and RyR2 have been isolated and purified based on their ability to bind [3H]ryanodine (45, 160, 161, 162, 194), and RyR3 has not yet been purified and studied in vitro, primarily due to a failure to identify a tissue expressing this isoform in suitable abundance.

Three RyR isoforms have also been identified in avian, amphibian, and piscine muscles. These include two isoforms, termed [Alpha] and [Beta], found initially in skeletal muscle and a third isoform found in heart. The avian RyR isoforms have been isolated, purified, and shown by several criteria to represent three unique proteins (6, 8, 76). Similarly, the amphibian and piscine [Alpha]- and [Beta]-RyRs have been purified (243, 252, 258). The mRNAs for the amphibian and the avian [Alpha]- and [Beta]-RyRs have been cloned and sequenced (170, 255, 260a, 264; Airey, unpublished data). On the basis of sequence comparisons with the mammalian isoforms, the amphibian and avian [Alpha]- and [Beta]-RyRs are most similar to the mammalian RyR1 and RyR3 isoforms, respectively (255, 260a, 264; Airey, unpublished data); while on the basis of immunological similarities, the amphibian [Alpha]- and [Beta]-RyRs have been suggested to be homologues of RyR1 and RyR2 (195). As discussed below, the extent that these isoforms are functional homologues remains an important question. Although ultimately a common nomenclature should be adapted for the mammalian and nonmammalian RyRs, at present it may be best to retain a distinction between these isoforms until their functional similarities are established more completely. Therefore, in this review, the mammalian isoforms are termed RyR1, RyR2, and RyR3, while the nonmammalian vertebrate isoforms are referred to as [Alpha], [Beta], and cardiac. Other, as yet uncharacterized, RyR isoforms are designated by the species and/or tissue in which they have been identified.

Alternative splice variants of both RyR1 and RyR2 mRNAs have been identified. In the case of RyR1, two separate splice events result in the presence or absence of a 15- or a 18-bp exon. Because the corresponding five([Ala.sup.3481]-[Gln.sup.3485]) (116, 395) and six-amino acid ([Val.sup.3865]-[Ash.3870]) (116) sequences occur in the region of RyR1 containing putative binding sites for channel modifiers, such as calcium, calmodulin, and ATP and a Ser residue ([Ser.sup.2843]) that is phosphorylated by both the calcium/calmodulin protein kinase (CaMK) and the adenosine 3[prime],5[prime]-cyclic monophosphate (cAMP)-dependent protein kinase (PKA) (328), it was suggested that these splice events may alter RyR1 channel function (116). The possibility that the resulting RyR1 proteins have unique functional roles is supported by the findings that 1) the RyR1 splice variant mRNAs are expressed in both a tissue- and a developmental stage-specific manner; 2) they are coexpressed in some tissues, e.g., skeletal muscle and brain; and 3) the same RyR1 splice variants are found in mouse, rabbit, and human muscles, indicating that their expression has been conserved in different species (116).

An alternative splice variant of RyR2 involving a 24-bp segment, encoding amino acids 3716-3723, has also been identified (247). In addition, mRNA encoding one of the RyR1 splice variants has been found in the heart (116), indicating that this RyR1 isoform may be coexpressed with RyR2.

The alternative splicing events identified to date for RyR1 and RyR2 have the potential of resulting in three additional RyR1 and one additional RyR2 isoforms. When one considers these results in terms of the primary topic of this review, RyR diversity, it is important to note that the protein products of the splice variants have yet to be identified. Thus their relative abundance and whether they differ functionally is unknown. Antibodies capable of distinguishing the alternative splice variants have not been described; consequently, their cellular expression patterns and the possibility that more than one is coexpressed in the same cell also remain to be determined. The latter results will be of interest, because as considered in greater detail in section VI, an unexpected difference observed for vertebrate fast-twitch muscle fibers is that some fibers coexpress equivalent levels of two RyR isoforms, while others express one predominant RyR. One way this apparent paradox could be resolved is if a RyR1 splice variant that is functionally equivalent to the low abundance isoform is also expressed in the latter fibers.

The results of pharmacological studies also suggest that RyR isoforms other than RyR1 and RyR2 exist in mammalian tissues. At present, it is unclear whether these additional RyRs are the RyR3 isoform, one of the splice variants described above, or as yet unidentified isoforms because the pharmacological and ion channel properties of the RyR3 and alternatively spliced RyRs have not been established. For example, unlike RyR1 and RyR2 channels that are each stabilized in subconductance states and low (less than micromolar) and blocked by high (greater than micromolar) concentrations of ryanodine (see below), a RyR in canine and porcine aortic smooth muscle has been found to be blocked by ryanodine, but not modified to a conductance substate by this ligand (146). A RyR in cultured human uterine artery smooth muscle cells is unlike RyR1 and RyR2 in that it is not affected by caffeine (213). These isoforms may be RyR3. The RyR3 mRNA has been found in a number of smooth muscles (119, 120, 139, 201), and this isoform has been suggested to either be insensitive (119) or to have a reduced sensitivity (345) to caffeine. In addition, recent data indicate that a RyR(s) other than RyR1 and RyR2 may be expressed in liver (22, 207, 304, 305).

B. Invertebrate Ryanodine Receptors

The RyRs expressed in a number of invertebrates appear to differ from the isoforms found in vertebrate tissues. A RyR in lobster muscle differs from RyR1 and RyR2 in that it requires millimolar, rather than micromolar, concentrations of calcium for channel activation in vitro (257, 299). From measurements of calcium release transients in permeabilized skeletal muscle fibers from the dyspedic (a RyR1 knockout) mouse, RyR3 appears to require greater levels of calcium for activation than either RyR1 or RyR2 (345), but RyR3 can be activated by calcium levels that are less than those required by the lobster muscle RyR. A preliminary report of the partial sequence of the lobster RyR did not indicate a greater similarity with one of the vertebrate RyRs (383). A RyR has been cloned and sequenced from Drosophila (142, 342). On the basis of the derived amino acid sequence, this RyR exhibits the most extensive differences with the other RyRs identified to date, having [less than]50% similarity with the mammalian isoforms. The Drosophila RyR protein has not yet been isolated, characterized, or shown to be an ion channel. In addition, as shown in Table 1, RyRs have been identified in several other invertebrates. These RyRs have yet to be characterized at a molecular level.

 
TABLE 1. Ryanodine receptor: species distribution 
 
Animal                Assay                      References Cited 
 
Mammals 
 
Human                 RS                         394 
Dog                   RE                         278 
Rabbit                RE, RS                     278, 343 
Rat                   RE                         278 
Mouse                 RE                         278 
Guinea pig            RE                         278 
Pig                   RE                         278 
Mink                  RS                         119 
Sheep                 BL                         379 
Cow                   WA, RE                     61, 331 
Cat                   RE                         278 
 
Birds 
 
Chicken               WA, RB, IH, BL             6, 270 
Finch                 WA, RB                     177 
Pigeon                WA, RB                     177 
Duck                  RB                         75 
Quail                 RB                         271 
 
Amphibians 
 
Frog                  WA, RB, BL, RE             40, 243, 258, 277 
Xenopus               WA, RE                     243, 263 
 
Reptiles 
 
Rattlesnake           WA                         252 
Garter snake          WA                         252 
Lizard                WA                         252 
Turtle                WA                         252 
 
Fish 
 
Toadfish              US, RB, WA, BL, IH         30, 252, 258 
Trout                 RE, WA                     154, 252 
Amphioxus             US                         135 
Lancelet              IP, WA, RB                 24 
Knifefish             IH                         396 
Marlin                WA, RB, IH                 31, 253 
Swordfish             WA, IH                     31 
Bass                  WA                         31 
 
Chelicerates 
 
Scorpion              US                         210 
Spider                US                         103, 303 
Horseshoe crab        US                         100 
 
Insects 
 
Moth                  RE                         221 
Cockroach             RE                         77 
Drosophila            RS                         142, 342 
Honeybee              RE                         370 
Grasshopper           US                         210 
Damselfly             US                         210 
Katydid               US                         78 
 
Crustaceans 
 
Crayfish              US, RE                     132, 210 
Lobster               RB, WA, BL, CaF, RS        257, 299, 383 
Barnacle              US, RE                     28, 155 
 
Nematodes 
 
C. elegans            RB, BL                     186 
 
Echinoderms 
 
Sea urchin            RE                         39, 112 
 
Mollusks 
 
Scallop               US                         210, 292 
Snail                 RE                         4 
 
US, ultrastructure; RS, ryanodine receptor sequence; RE, ryanodine 
effects on function; BL, bilayer; WA, Western analysis; IH, 
immunohistochemistry; RB, [3H]ryanodine binding; CaF, SR 
membranes-45Ca fluxes. 

C. Ryanodine Receptor-Related Proteins

An additional form of RyR1 may result from the utilization of a second transcription initiation site in the 3[prime]-third of the ryr1 gene. This site is used in rabbit brain to produce a truncated form of this isoform, as RyR1-specific cDNA probes identified a 2.4-kb mRNA in Northern analyses (344). Expression of a cDNA encoding the COOH-terminal 656 amino acids of RyR1 contained in the truncated form of this protein in Chinese hamster ovary (CHO) cells resulted in the appearance of a 75-kDa polypeptide that was recognized by anti-RyR1 polyclonal serum in Western analysis. Cells expressing the truncated protein exhibited neither [3H]ryanodine binding nor caffeine-stimulated calcium release, raising a question about the extent to which this protein retains the properties of the complete RyR1 isoform (344). This is of interest, because using different approaches Callaway et al. (43) and Witcher et al. (381) have independently localized the high-affinity ryanodine binding site to the COOH-terminal 76 kDa of the protein. In both studies, ryanodine was bound and stabilized to the intact protein, before fragmentation of the isoform by proteolysis; therefore, part of the protein on the N[H.sub.2]-terminal side of this region may be required for initially binding ryanodine at its high-affinity site under normal (intracellular) conditions. The truncated 75-kDa protein is also of interest because it contains all 4 of the membrane-spanning regions proposed by Takeshima et al. (343) to form the RyR1 calcium channel domain, but only 6 of the 10 proposed by Zorzato et al. (394). Therefore, the truncated protein has the potential to be a new type of calcium channel and to provide a test of the two possible channel domain structures. The function of this protein remains to be established.

A 106-kDa protein capable of binding [3H]ryanodine and forming calcium channels in bilayers has been isolated from skeletal muscle sarcoplasmic reticulum (SR) (392). The relationship between this protein and the larger RyRs described above has yet to be established.

On the basis of only the preceding data, there may be up to nine RyR isoforms expressed in vertebrate tissues. Moreover, it is likely that with further analysis this complexity will be increased by the finding of additional alternative splice variants and perhaps of new isoforms. An important issue is the extent to which these RyRs differ functionally, because it is ultimately their functional differences that will determine the complexity of the calcium release events they mediate. Another question concerns whether sequence similarities can be taken to indicate that RyRs expressed in different species or tissues, such as RyR3 and the [Beta]-RyR, are functional homologues. This point is considered in section VI in terms of the functional properties of these isoforms.

[TABULAR DATA FOR TABLE 2 OMITTED]

III. SPECIES AND TISSUE DISTRIBUTIONS

Ryanodine receptors were first identified pharmacologically because of the pronounced actions of the plant alkaloid ryanodine on insects and on vertebrate striated muscles (see Ref. 169 for references). Indeed, these proteins were purified initially from striated muscles based on the ability to bind [3H]ryanodine. Ryanodine receptors have subsequently been shown to be widely distributed both across species (Table 1) and in different tissues within the same species (Table 2) by pharmacological and/or molecular data. It is likely that more species and tissues will be added to these lists as antibody and nucleic acid probes capable of identifying the various RyR isoforms described above become available.

Findings relevant to this review are that calcium release events frequently involve combinations of calcium release channels. For example, the results of functional studies indicated that a RyR and an I[P.sub.3]R are present within the same cells (25, 26), an observation that has been confirmed by immunolocalization studies (e.g., Refs. 302, 369). More recently, it has been shown that two RyR isoforms are coexpressed in cells. This was observed first in chicken fast-twitch skeletal muscles (6) and subsequently in frog (195, 243, 244, 258) and fish (252, 258) fast-twitch skeletal muscles and avian neurons (81, 261, 262, 369). The latter findings demonstrate that RyR coexpression is not a species- or a tissue-specific event. The coexpressed RyRs have different mobilities in sodium dodecyl sulfate gels, raising the possibility that one was derived from proteolysis of the other. This was ruled out for the two RyRs, [Alpha] and [Beta], found in chicken skeletal muscle by showing that the peptide maps generated for these proteins using three separate proteases differed (6). Similarly, comparisons of a number of biochemical and immunological properties, in particular, of peptide maps generated for the chicken [Alpha]-, [Beta]-, and cardiac RyR isoforms, demonstrated that neither the [Alpha]- nor the [Beta]-isoform was the result of expression of the cardiac RyR gene (8). Ryanodine receptor-isoform specific antibodies were used to provide definitive proof that both RyRs coexisted within the same cell (6). The initial studies of this phenomenon were facilitated by the fact that the two RyR isoforms in nonmammalian vertebrate muscles and nerves are expressed at comparable levels.

More recently, the results of several studies indicate that two RyRs are also coexpressed in mammalian tissues. Both RyR1 and RyR3 mRNA have been identified in mink skeletal muscle using a ribonuclease (RNase) protection assay (119). Consistent with RyR3 existing as a second calcium release channel in skeletal muscle, Takeshima and co-workers (341, 345) found both RyR3 mRNA and a ryanodine-sensitive calcium transient in skeletal muscle fibers from RyR1 "knock-out" (dyspedic) mice in which both alleles of the RyR1 gene had been inactivated by homologous recombination. Studies from several laboratories using immunolocalization, in situ hybridization, reverse transcription (RT)-polymerase chain reaction (PCR), and RNase protection assays (120, 201, 345) have shown that the distributions of RyR1, RyR2, and RyR3 overlap in regions and in specific neuron types in the mammalian brain (114, 120, 191, 193, 225). Whether these isoforms are coexpressed in the same neurons remains to be determined. The latter three techniques have also been used to show that the mammalian RyRs have overlapping distributions in a number of peripheral tissues (see Table 2).

Although the data in the preceding paragraph demonstrate that more than one RyR isoform is expressed in mammalian tissues, coexpression within the same cell has yet to be demonstrated directly. This will require the use of antibody and nucleic acid probes that distinguish the alternatively spliced variants of each isoform, as well as the individual isoforms to colocalize the RyRs in individual cells. The overlapping distributions of the RyR isoforms in different tissues noted above could reflect the presence of a multiple RyR calcium release system within the same cell, of a heterologous multirelease channel system composed of RyR and IP3R isoforms, or the independent use of the individual calcium release channels by different cell types within a tissue. In any case, the selective expression of the RyR isoforms either alone or in combination indicates that these proteins have unique functional properties that are used to fit the needs of different cells and tissues.

In addition to demonstrating directly whether two RyRs coexist in different cells, it will be important to establish the relative abundance and the cellular loci (see sect. IV) of each isoform to evaluate the contributions made by each protein to cellular calcium release events. The properties of RyR-mediated calcium transients may depend not only on the types of RyRs present but also on the relative abundance of each isoform. A release system containing similar quantities of two RyRs may operate differently from one composed of a single RyR, or of disparate levels of two release channels. Such differences in RyR abundance have been shown to exist in nonmammalian vertebrate muscles and neurons. For example, during embryonic and posthatch development, subsets of Purkinje neurons in the chicken cerebellum express either the [Alpha]-RyR alone, the [Alpha]- and [Beta]-RyR in combination, or only the [Beta]-RyR (81, 261, 262). In addition, in contrast to fast-twitch muscles in frogs, fish, and birds, which express two RyR isoforms at similar levels, only a single RyR protein (identified as the [Alpha]-RyR) was detected in muscles capable of very fast contractions, including the toadfish swimbladder muscle, extraocular muscles from toadfish, chickens, and cats, and the rattler muscle from rattlesnakes (252). A single [Alpha]-RyR was also found in locomotory muscles from lizards and snakes, but two RyRs, designated as [Alpha] and cardiac or [Beta], were identified in muscles from turtles and alligators (252).

Western analysis was used to identify the RyRs in the latter studies (252), and a potential caveat concerns whether this approach as applied in these studies had sufficient sensitivity to detect low levels of a second iso-form. For example, only one laboratory has reported detection of RyR3 protein in mammalian skeletal muscle by Western analysis (61). Whether a second low-abundance RyR is expressed in nonmammalian vertebrate muscles is of interest. If the second isoform is absent in some muscles, then at least three different types of calcium release systems are likely to exist in vertebrate skeletal muscles: one containing only a (or RyR1), one containing similar levels of [Alpha]- and [Beta]-RyRs, and one containing predominately RyR1 (or [Alpha]-RyR) and lower levels of RyR3 (or [Beta]-RyR). Unfortunately, the in situ hybridization, RT-PCR, and RNase protection analyses used in the studies of mammalian tissues described above yield only estimates of the relative levels of mRNAs encoding each RyR. If the isoforms have different rates of synthesis and/or degradation, these estimates may not reflect the abundance of the

RyR proteins in the different cell types. In any case, it is clear that RyRs are expressed at different relative levels in different vertebrate muscles. Possible reasons for this molecular diversity are discussed further in section VI.

IV. CELLULAR DISTRIBUTIONS

Correlations between cellular ultrastructure and the distribution of the proteins involved in calcium metabolism have contributed important information about both the anatomy of calcium release systems and the association of these systems with specific cellular domains and structures.

A. Nonmuscle Cells

Examples of complexities in the distribution of calcium release channels in nonmuscle cells are found first in the studies by Burgoyne et al. (41), which show two SERCA1 [Ca.sup.2+]-ATPase isoforms are differentially localized in adrenal chromaffin cells. Interestingly, the loci of these isoforms corresponded to sites at which calcium transients initiated by I[P.sub.3]-generating agonists and caffeine originated. These data indicate the existence of distinct calcium release systems in these cells that 1) involve different protein components; 2) can be activated independently, perhaps as well as interactively; and 3) are likely to generate signals that impart unique information to the cell.

A second example involves RyR distribution in cerebellar Purkinje neurons. Stimulated by the suggestions by Henkart and colleagues (143-145) of structural similarities between subsurface membrane cisternae in neurons and triad junctions in skeletal muscle, we investigated whether RyRs are expressed in either the central and/or peripheral nervous systems. These studies have provided another example of differential localization of calcium release systems that may have physiological relevance, namely, that of RyRs and I[P.sub.3]Rs in avian cerebellar Purkinje neurons. With the use of antibodies specific for each receptor type, the I[P.sub.3]Rs and RyRs were found to have overlapping distributions in these cells and in some cases to exist in close proximity in the same membranes (369). A caveat associated with the latter observation is that unless antibodies that bind to a functionally active conformation of each protein are used, this approach identifies immunoreactive protein and not necessarily functional calcium release channels. For example, both receptor types could be translated at the same site in the cell, but not become functional as release channels until they are trafficked to different calcium stores. Because ryanodine binds with high affinity to a conformation of the RyR associated with an open state of the channel, the development of a fluorescently labeled ryanodine (374) may provide a probe that is suitable for detecting RyRs that are active as release channels.

One difference that was noted in distributions of the RyRs and I[P.sub.3]Rs in Purkinje neurons was that while both RyRs and I[P.sub.3]R are present in the main shafts of the dendrites (in particular at branch points), only the I[P.sub.3]Rs were found in the spines that project laterally from the shafts. This created regions containing either both I[P.sub.3]R and RyR release channels or regions containing only I[P.sub.3]R channels. Interestingly, three-dimensional reconstruction of the endomembrane systems present in the dendrites by Martone et al. (218) demonstrate that the membranes present within the shafts are continuous with those that extend into the spines. Thus, while the RyRs and I[P.sub.3]Rs contain signals that lead to their localization to the same regions in membranes in dendritic shafts, the I[P.sub.3]R must also contain an additional signal that targets it to another site in the same membrane system present in the dendritic spines. The functional significance of this difference in localization of the two receptor types is discussed below and may be illustrated by recent findings, such as those of Seymour-Laurent and Barrish (301).

The preceding data indicate that multiple calcium release channels are used in the same cell and provide important insights into their physical relationships and possible functions. As mentioned previously, proteins responsible for calcium uptake [the [Ca.sup.2+]-ATPase (178)], binding [calsequestrin (95) and calreticulin (230)], and release (RyR and I[P.sub.3]R; see above) all exist as multiple forms. As reagents capable of detecting the different isoforms of these proteins become more widely available, it will be interesting to learn if assemblies of these proteins are used in a predictable manner to produce calcium transients having similar properties in different cell types. Recent studies of RyRs in osteoclasts (390, 391) suggest that RyRs in these cells may function to sense extracellular calcium, in a heretofore unappreciated type of activity.

B. Vertebrate Skeletal Muscles

The calcium release system whose morphological features have been described most extensively is that associated with the SR of striated muscles. This work has been facilitated by the association of the release system with distinctive junctional structures consisting of terminal cisternae of the SR and either the surface membrane of the cell (peripheral couplings) or invaginations of the surface membrane known as transverse (t) tubules (triad and dyad junctions) (101, 102, 105; [ILLUSTRATION FOR FIGURE 1 OMITTED]). The insights provided by morphological analyses have been essential for the progress made in this field. They have provided the basis for verifying the identity of isolated membrane preparations, such as heavy SR (terminal cisternae) (89, 226, 290) and intact triads (51, 236, 237). These preparations have been used in vitro to establish the biochemical and pharmacological properties of SR calcium release, as well as the identity of proteins involved in coupling muscle cell excitation to SR calcium release (e.g., Refs. 42, 49, 50). Not the least of the latter is the RyR (e.g., Refs. 180, 290), which had been termed initially the foot protein (101) based on its appearance in electron micrographs.

Morphological analyses of the SR calcium release system have also had important influences on models currently conceived for excitation-contraction (E-C) coupling in vertebrate skeletal muscle. The results of freeze-fracture studies of toadfish swim bladder muscle provided structural evidence for the existence of two classes of RyRs (30). Regular arrays of particles were observed in fractured faces of both the t-tubular and the SR terminal cisternae membranes. The t-tubular particles, which occurred in clusters of four and were termed tetrads, were proposed to represent four dihydropyridine receptors (DHPR). Groupings of four particles that represented the tetrameric RyR were observed in the SR membranes. Composites of the images obtained for both membranes yielded a relative positioning of the two particle types, such that only every other RyR was placed sufficiently close to a tetrad to permit the two proteins to interact. This result indicated the existence of two types of RyRs, that differed in their positioning relative to the DHPRs. Moreover, this physical arrangement has been interpreted to indicate that different mechanisms are used to activate calcium release by each type of RyR (see sect. VI).

The results from three additional lines of experimentation conducted by Franzini-Armstrong and colleagues (106, 107, 336, 337) support the interpretation of the above results and indicate that the association of alternate RyRs with the tetrad cluster of DHPRs observed in the toadfish swim bladder muscle may be representative of the placement of these proteins in skeletal muscles from other vertebrates. First, the identity of the tetrad of particles in t-tubular membranes as the DHPR has been substantiated further. While tetrads are observed in freeze-fracture replicas of surface membranes from cultured wild-type mouse skeletal muscle cells, they are not found in membranes from muscle cells from mice homozygous for the muscular dysgenesis (mdg) mutation (107). This mutation resides in the gene encoding the skeletal muscle isoform of the [[Alpha].sub.1]-DHPR subunit (52, 347) and precludes expression of a normal functional channel (347). Importantly, tetrads are found following expression of a wild-type [[Alpha].sub.1]-subunit in cultured mdg/mdg muscle cells (336). Second, the identity of the SR membrane particles identified as RyR1 is similarly supported by the absence of these particles in SR membranes obtained using skeletal muscles from dyspedic (RyR1 knock-out) mice (337). Third, comparison of the organization of tetrads observed in freeze-fracture replicas of surface membranes of skeletal muscles from mouse embryonic myotubes, frog tonic fibers, and cultured human myotubes (each of which contains peripheral couplings), with the organization observed for RyRs in separate studies, indicates that only alternate RyRs are positioned to interact with a tetrad (DHPRs) in these species (106). Tetrads were observed to be larger than the SR membrane (RyR) particles, and an explanation offered for the alternate RyR association pattern is that the size of the tetrads precludes them from being packed close enough to interact with every RyR (106).

As just described, evidence exists for two morphologically distinct classes of RyRs, for the expression of two biochemically distinct RyRs, and as described below, for the utilization of two mechanisms for activating SR calcium release. Thus things would appear to have fallen into place, as far as the RyRs and their roles in skeletal muscle SR calcium release are concerned. Unfortunately, biochemical data described here and functional data discussed in section v are not consistent with such a simple picture. Two of the muscles shown to exhibit alternating RyR/DHPR associations, toadfish swim bladder muscle and mouse skeletal muscle, appear to express solely (or predominantly) a single RyR isoform (252, 345). Thus, at least in these muscles, it does not appear necessary to express more than one RyR isoform to have two morphologically distinct classes of this protein. The possibility exists, of course, that these muscles also express either an as yet unidentified RyR isoform or an alternatively spliced variant of RyR1 that serves as the second RyR. At present, due to the lack of probes for such additional RyRs, there are no data either supporting or negating either possibility.

 
TABLE 3. Ratio of dihydropyridine and ryanodine receptors 
 
                                             Ratio of 
                                               DHPR     References 
Muscle Studied                                to RyR      Cited 
 
Skeletal muscle 
 
One RyR isoform expressed 
 
Toadfish super fast                             2.0         30 
Rabbit fast twitch                              1.3         27 
Rabbit adductor (fast)                          1.8        216 
Rat tibialis anterior (fast)                    1.2        216 
Rabbit masseter (mixed, mainly fast)            1.6        216 
Rabbit leg (mixed) and psoas (fast)             1.0         12 
Pig longissimus dorsi (mixed, mainly fast) 
(normal, heterozygous, homozygous MH)         0.6-0.9      231 
Rabbit soleus (slow)                            2.1        216 
Rat soleus (slow)                               1.2        216 
 
Two RyR isoforms expressed 
 
Frog fast twitch                                0.6         12 
Frog sartorius (fast)                           1.1        216 
 
Cardiac muscle 
 
Rabbit                                          0.3         27 
Guinea pig                                      0.2         27 
Rat                                             0.1         27 
Ferret                                          0.1         27 
 
RyR, ryanodine receptor; DHPR, dihydropyridine receptor; MH, 
malignant hyperthermia. In all studies except Ref. 30, ratio was 
determined from radioligand binding studies. In Ref. 30, ratio was 
determined from morphometric analysis. 

The existence of an alternating RyR/DHPR association pattern in muscles with a single (or predominant) RyR suggests that the ability of a RyR to associate with DHPRs is not determined by RyR sequence-specific features. Thus conditions exogenous to the RyR may dictate the organization of these proteins. There are several possibilities for such conditions. First, it could be simply that the numbers of the RyRs and DHPRs differ. As noted by Franzini-Armstrong and Irish (106), because a tetrad is composed of four DHPRs and there are two RyRs for every tetrad, the alternate RyR association pattern would predict a 2:1 ratio of DHPR to RyR. Interestingly, estimates of the abundance of these proteins from three radioligand binding studies (based on the assumptions that each tetrameric RyR binds 1 molecule of ryanodine with high affinity and that the 4 DHPRs present in a tetrad each bind 1 dihydropyridine molecule) yield ratios for different skeletal muscles from different species that range from 2.0 to 0.5 (Table 3 and references therein). These studies included muscles that express a single …