Localization in the II-III Loop of the Dihydropyridine Receptor of a Sequence Critical for Excitation-Contraction Coupling*

Skeletal and cardiac muscles express distinct isoforms of the dihydropyridine receptor (DHPR), a type of voltage-gated Ca2+ channel that is important for excitation-contraction (EC) coupling. However, entry of Ca2+ through the channel is not required for skeletal muscle-type EC coupling. Previous work (Tanabe, T., Beam, K. G., Adams, B. A., Niidome, T., and Numa, S. (1990) Nature346, 567–569) revealed that the loop between repeats II and III (II-III loop) is an important determinant of skeletal-type EC coupling. In the present study we have further dissected the regions of the II-III loop critical for skeletal-type EC coupling by expression of cDNA constructs in dysgenic myotubes. Because Ser687 of the skeletal II-III loop has been reported to be rapidly phosphorylatedin vitro, we substituted this serine with alanine, the corresponding cardiac residue. This alanine-substituted skeletal DHPR retained the ability to mediate skeletal-type EC coupling. Weak skeletal-type EC coupling was produced by a chimeric DHPR, which was entirely cardiac except for a small amount of skeletal sequence (residues 725–742) in the II-III loop. Skeletal-type coupling was stronger when both residues 725–742 and adjacent residues were skeletal (e.g. a chimera containing skeletal residues 711–765). However, residues 725–742 appeared to be critical because skeletal-type coupling was not produced either by a chimera with skeletal residues 711–732 or by one with skeletal residues 734–765.

Dihydropyridine receptors (DHPRs) 1 in skeletal and cardiac muscle are closely related proteins encoded by two different genes (1,2). In both muscle types, DHPRs serve dual functions (2)(3)(4)(5)(6) as voltage-gated, L-type Ca 2ϩ channels and as a trigger for excitation-contraction (EC) coupling, which controls the release of Ca 2ϩ through ryanodine receptors (RyRs) (7) in the sarcoplasmic reticulum. However, the mechanism of EC coupling is different in cardiac and skeletal muscle. In cardiac muscle, depolarization causes opening of L-type Ca 2ϩ channels, and the resulting entry of extracellular Ca 2ϩ triggers RyRs to release Ca 2ϩ (6). In skeletal muscle, EC coupling does not require Ca 2ϩ entry (8); rather, depolarization causes some other kind of signal to be transmitted from DHPRs to RyRs. To identify regions that are critical for this skeletal-type EC coupling, we have previously used the approach of constructing cDNAs encoding chimeras of the skeletal and cardiac DHPRs (3). Expression of these chimeric DHPRs in dysgenic myotubes, which lack a functional gene for the skeletal DHPR (9), revealed that the putative cytoplasmic region between repeats II and III (II-III loop, amino acids 666 -791) of the skeletal DHPR is an important determinant of skeletal-type EC coupling (3).
Comparison of the skeletal and cardiac DHPRs reveals differences scattered throughout the II-III loops. One important difference is the presence of a site (Ser 687 ) in the skeletal II-III loop, which is phosphorylated by cyclic AMP-dependent protein kinase (PKA) (1,10) and is lacking in the cardiac loop (2). Another important difference is the presence of a 12-amino acid insertion in the N-terminal half of the cardiac II-III loop. To determine the importance of these and other amino acid differences for skeletal-type EC coupling, we have now examined a skeletal DHPR in which Ser 687 was mutated to alanine and also analyzed chimeric DHPRs in which skeletal sequence was substituted for successively smaller portions of the II-III loop of the cardiac DHPR. We found that removal of the serine phosphorylation site did not alter the function of the skeletal DHPR and that a small region slightly toward the C-terminal half from the midpoint of the II-III loop was critical for the ability of chimeric DHPRs to mediate skeletal-type EC coupling. The results are discussed in light of work from other laboratories in which the function of isolated RyRs was assayed during exposure to peptide fragments corresponding to various regions of the skeletal and cardiac DHPRs (11)(12)(13)(14)(15). sites, and new restriction sites introduced by site-directed mutagenesis (16) and PCR; the nucleotide changes used to introduce the new restriction sites were selected so as not to alter the amino acid sequence present in the cDNA inserts of the expression plasmids pCAC6 (4) and pCARD1 (2,17), which were used as the starting materials. All fragments amplified by PCR were confirmed by sequencing. In the following constructions, an asterisk and the letter "p" are used to designate nucleotides within restriction sites introduced by site-directed mutagenesis and PCR, respectively (nucleotide substitutions are given in brackets): pCSk31, XmnI-StyI (Sk 1964(Sk -2127  Functional Analysis of Chimeric DHPRs-Primary cultures of dysgenic myotubes were prepared and injected with plasmid DNA as described previously (4). Cells expressing the injected plasmid were identified by contraction in response to extracellular stimulation (4). The whole-cell patch clamp technique (18) was used in combination with a photomultiplier system to allow the simultaneous measurement of membrane currents and Ca 2ϩ transients 1-4 days after plasmid injection (19). The patch pipette contained (mM): 140 Cs-aspartate, 5 MgCl 2 , 10 Cs 2 EGTA (20 nM free Ca 2ϩ ), 5 Na 2 ATP, 0.2 pentapotassium-Fluo-3 (Molecular Probes, Eugene, OR), and 10 HEPES (pH 7.4 with CsOH). The bath solution contained (mM): 145 tetraethylammonium ϩ , 165 Cl Ϫ , 10 HEPES (pH 7.4 with CsOH), 0.003 tetrodotoxin, and either 10 Ca 2ϩ (to support Ca 2ϩ current) or 2 Ca 2ϩ , 8 Mg 2ϩ , 0.5 Cd 2ϩ , and 0.3 La 3ϩ (to block Ca 2ϩ current). The voltage clamp command sequence was to step from the holding potential (Ϫ80 mV) to Ϫ50 mV for 25-30 ms, to the test potential for 15 ms, to Ϫ50 mV for 25-30 ms, and back to the holding potential. The parameters GЈ max , VЈ1 ⁄2 , and kЈ were determined from the slow calcium current as described previously (19). Because there was large variability in magnitude of the calcium transients, the transients were not analyzed quantitatively, except to determine the "strength" of skeletal-type EC coupling, which was defined for each chimera as the ratio of the number of cells with a measurable Ca 2ϩ transient in Cd 2ϩ /La 3ϩ solution to the number of cells with a transient in 10 mM Ca 2ϩ -containing solution. Additionally, traces selected for illustration are representative in terms of time course but not necessarily in terms of magnitude. Temperature was 20 -22°C. Fig. 1 schematically illustrates the skeletal muscle and cardiac DHPRs, designated CAC6 (4) and CARD1 (2, 17), respectively, and the chimeric DHPR, CSk3 (3), which has a skeletal II-III loop but otherwise has cardiac sequence. The skeletal II-III loop is 138 amino acids in length, whereas the cardiac II-III loop is 147 amino acids in length (1,2). In the skeletal muscle II-III loop, Ser 687 is phosphorylated in vitro by PKA (10); alanine occurs at the corresponding position (residue 809) of the cardiac II-III loop. To test for the importance of this PKA site, we constructed a mutant skeletal DHPR (S687A) in which Ser 687 was replaced by an alanine.

RESULTS
Intracellular Ca 2ϩ transients in response to depolarization were recorded in dysgenic myotubes expressing CAC6, S687A, CARD1, and CSk3 with the cells bathed first in a solution containing 10 mM Ca 2ϩ (Fig. 1, center column) and then in a solution containing Cd 2ϩ plus La 3ϩ to block the entry of extracellular Ca 2ϩ (Fig. 1, right-hand column). Blocking the entry of extracellular Ca 2ϩ abolished the intracellular Ca 2ϩ transient for constructs mediating cardiac-type EC coupling (CARD1 trace). For constructs mediating skeletal-type EC coupling (e.g. CAC6), the Ca 2ϩ transient was still present (although decreased in amplitude) after Ca 2ϩ entry had been blocked by the addition of Cd 2ϩ and La 3ϩ . The decreased amplitude of the Ca 2ϩ transient (i.e. ⌬F/F) has been previously described for normal myotubes exposed to inorganic Ca 2ϩ channel blockers (20). It probably results from deleterious effects of Cd 2ϩ and La 3ϩ , which cause an essentially irreversible increase in resting fluorescence (F).
In agreement with previous results (3), skeletal-type EC coupling was observed in dysgenic myotubes expressing the chimera CSk3, in which the II-III loop of the cardiac DHPR was replaced with the corresponding skeletal sequence (Fig. 1). The presence of a PKA phosphorylation site did not appear to be critical, because S687A was able to mediate skeletal-type EC coupling (all 7 myotubes tested). Moreover, the magnitude and voltage dependence of the slow Ca 2ϩ current for S687A (G' max ϭ 54.4 Ϯ 44.6 nanosiemens/nanofarads; V'1 ⁄2 ϭ 27.6 Ϯ 9.5 mV; k'ϭ 7.9 Ϯ 2.2 mV; n ϭ 6) were similar to those of the slow Ca 2ϩ current in dysgenic myotubes expressing CAC6 (19).
To localize the region of the II-III loop most important for skeletal-type EC coupling, we constructed new chimeric DHPRs, using CSk3 as the parent structure. These chimeras had cardiac sequence except for a small amount of skeletal sequence within the II-III loop. Schematics of the II-III loops of the chimeras and representative Ca 2ϩ transients observed after expression in dysgenic myotubes are illustrated in Fig. 2. The initial chimeras examined were CSk31 and CSk32, in which either the N-or C-terminal half, respectively, of the II-III loop of CSk3 was replaced with the corresponding cardiac sequence. CSk32, but not CSk31, produced a Ca 2ϩ transient after block of the Ca 2ϩ current by Cd 2ϩ and La 3ϩ (Fig. 2) Fig. 2, which replots data from these cells. The test depolarization was ϩ40 mV except for CAC6-S687A (ϩ30 mV). The vertical calibration (⌬F/F) corresponds to 2.0 for CAC6, 0.2 for S687A, and 0.6 for CARD1 and CSk3.
importance of the C-terminal half of the II-III loop. Thus, we next examined CSk33 and CSk34, in which complementary regions of the skeletal portion of CSk32 were converted to cardiac sequence. CSk33 supported skeletal-type EC coupling, whereas CSk34 did not. Thus, the middle portion of the II-III loop, indicated by the vertical dashed lines in Fig. 2, is important for skeletal-type EC coupling. Successively larger segments of the N-terminal half of this region were converted to cardiac sequence in the constructs CSk53, CSk54, and CSk40 and of the C-terminal half in the constructs CSk56 and CSk39. Of the constructs with less skeletal sequence at the N-terminal half, CSk53 and CSk54 produced skeletal-type EC coupling, but CSk40 did not. Of the constructs with less skeletal sequence at the C-terminal half, CSk56 produced skeletal-type coupling but CSk39 did not. Because there appeared to be insufficient skeletal sequence in CSk39 or CSk40 to support skeletal-type coupling, it suggested that the region of overlap between the next two larger constructs mediating skeletal-type EC coupling (CSk54 and CSk56) might be essential. Thus, we made CSk58, which has skeletal muscle sequence corresponding to this overlap region (indicated by vertical dotted lines in Fig. 2 (skeletal residues 725-742)). CSk58 was able to mediate skeletal-type EC coupling, but the coupling was weak in that a detectable Ca 2ϩ transient was present after block of Ca 2ϩ current in only 3 of 12 cells expressing CSk58 compared with 8 of 8 cells expressing CSk33 (see legend to Fig. 2).

DISCUSSION
In this study, we identified a small region (CSk58, skeletal residues 725-742) of the DHPR II-III loop that is critical for skeletal-type EC coupling. This region is located a little toward the C-terminal half from the middle of the II-III loop and includes 5 conservative and 7 non-conservative amino acid changes from the corresponding portion of the cardiac loop (Fig.  3). Chimeras that included only the N-terminal half (CSk39) or C-terminal half (CSk40) of the CSk58 region were unable to mediate skeletal-type EC coupling. Although an intact CSk58 region appeared to be necessary for skeletal-type EC coupling, this coupling was considerably weaker than that for chimeras containing additional skeletal residues on either side of the CSk58 region. Ser 687 in the skeletal II-III loop, which is phosphorylated by PKA, is not included in the CSk58 region, and a purely skeletal DHPR with the mutation S687A retained the ability to mediate skeletal-type EC coupling (Fig. 1). The cardiac II-III loop contains an inserted segment of 12 amino acids (cardiac residues 831-842) corresponding to a site between skeletal residues 710 and 711 (Fig. 3). The presence of this insertion in CSk53, CSk54, and CSk58 did not prevent them from mediating skeletal-type EC coupling (Fig. 2).
A simple way to account for our results would be to postulate that skeletal-type EC coupling involves a direct interaction between the RyR and the II-III loop of the DHPR, with loop residues 725-742 being especially important. For example, these amino acids might constitute an "agonist," which assumes the right configuration to bind to (and activate) the RyR only when the sarcolemma is depolarized. Unfortunately, the experiments with chimeric DHPRs do not provide a direct test of this hypothesis. An approach that has been used by other investigators to probe interactions between DHPRs and RyRs is to create peptides corresponding to various parts of the DHPR sequence and then to test for effects on RyRs by applying the peptides either to triad/SR vesicular preparations (activation monitored as an increase in [ 3 H]ryanodine binding or Ca 2ϩ efflux) or to RyRs reconstituted in artificial planar bilay- In cells not exhibiting skeletal-type coupling, depolarization-evoked Ca 2ϩ transients were present in the 10 Ca 2ϩ bathing solution. Test potential was ϩ40 mV except for CSk33 (ϩ50), CSk53 (ϩ70), and CSk58 (ϩ90). Vertical calibration (⌬F/F) corresponds to 0.3 (CSk56), 0.6 (CSk3, CSk32, CSk39, CSk40, CSk53, CSk58, CARD1), 0.8 (CSk31, CSk33), 1.0 (CSk34), or 1.2 (CSk54) .   FIG. 3. Alignment of the II-III loops from the skeletal muscle (Sk) and cardiac (C) DHPRs. Vertical lines and asterisks indicate conservative and non-conservative amino acid changes, respectively. Brackets indicate the regions of CSk53 and CSk58 having skeletal sequence. Also indicated by brackets are portions of the skeletal II-III loop that correspond to specific isolated peptides which other investigators have analyzed for effects on function of the RyR: peptides F1 (12) and A (14) were found to activate the skeletal RyR, peptide C was found to partially antagonize the activating effects of peptide A (14), and peptide C1 was found to have no effect (14). ers (activation monitored by an increase in open probability). As discussed below, it is difficult to construct a consistent framework that can account for both the results of application of the constructed peptides and our results on expression of DHPR cDNAs in dysgenic myotubes.
Lu et al. (11) found that [ 3 H]ryanodine binding to skeletal SR vesicles is increased by both a skeletal and a cardiac II-III loop peptide (SDCL and CDCL, respectively). This result seems difficult to reconcile with the observation that CSk3 (the chimera with a skeletal II-III loop) restores skeletal-type EC coupling whereas CARD1 (purely cardiac DHPR) does not ( Fig. 1  (3, 4, 17)). Lu et al. (12) subsequently found that phosphorylated SDCL specifically binds to, but does not activate, the skeletal RyR and that mutation of Ser 687 to alanine abolishes the ability of the SDCL to activate the skeletal RyR. This latter observation contrasts with our finding that both wild-type and S687A mutant skeletal DHPRs are able to mediate skeletaltype EC coupling (Fig. 1). More recently, Leong and MacLennan (13) found that a 37-amino acid segment of the skeletal RyR binds to the skeletal II-III loop and that this binding is mostly lost when Lys 677 and Lys 682 of the skeletal II-III loop are mutated to glutamates, the corresponding cardiac residues. This contrasts with our observation that the skeletal versus cardiac identity of this portion of the II-III loop had no effect on the ability to mediate skeletal-type EC coupling.
To localize regions that might be critical for EC coupling, Lu et al. (12) divided the II-III loop into three smaller peptides, termed F1 (Glu 666 -Glu 726 ), F2 (Pro 709 -Leu 766 ), and F3 (Lys 733 -Leu 791 ). The peptide F1, which represents only about the N-terminal half of the SDCL (Fig. 3), was found both to increase [ 3 H]ryanodine binding to the skeletal RyR and to displace binding of the phosphorylated SDCL, whereas peptides F2 and F3 caused little or no increase in [ 3 H]ryanodine binding and did not detectably displace the binding of the phosphorylated SDCL (12). Using somewhat smaller peptides corresponding to four distinct regions, El-Hayek et al. (14) also found that an N-terminal portion of the II-III loop causes activation of the RyR. Specifically, they found that "peptide A" (Thr 671 -Leu 690 ; see Fig. 3) caused Ca 2ϩ efflux from, and enhanced [ 3 H]ryanodine binding to, triad-enriched microsomes of skeletal muscle. Peptides corresponding to portions of the II-III loop closer to the C-terminal half, including a "peptide C1" (Phe 725 -Gly 743 ), which approximately overlaps the skeletal sequence of our CSk58 chimera (see Fig. 3), did not increase [ 3 H]ryanodine binding or Ca 2ϩ efflux. Although not causing activation, "peptide C" (Glu 724 -Pro 760 ), which roughly overlaps the skeletal sequence in our CSk53 chimera, moderately antagonized the actions of peptide A. Thus, it may be that two portions of the DHPR II-III loop (corresponding to peptides A and C) both bind to the skeletal RyR. However, there is also a discrepancy between the results of El-Hayek et al. (14) and those of Lu et al. (12), who found that the peptide F2 (larger than and encompassing the peptide C region) failed to antagonize binding of the phosphorylated SDCL to the skeletal RyR (12).
To account for the results of the peptide experiments, one could hypothesize that EC coupling involves activation of the RyR via the N-terminal half of the II-III loop. However, if this were the only important interaction between the DHPR and RyR it would be difficult to explain why the identity of this region (skeletal versus cardiac) had no influence on whether or not the II-III loop chimeras mediated skeletal-type EC coupling (Fig. 2). As a way of maintaining the hypothesis that the N-terminal half of the II-III loop is the physiological activator of the skeletal RyR, one might postulate that skeletal-type EC coupling depends upon an additional inhibitory interaction in resting muscle between the RyR and the portion of the II-III loop corresponding to peptide C. Because peptide C overlaps the region found in the loop chimeras to be critical for skeletaltype EC coupling, this hypothetical inhibitory interaction might be lost in chimeras having cardiac sequence in the region corresponding to peptide C. However, it has also been reported that a peptide corresponding to amino acids 1487-1506 of the skeletal DHPR inhibits [ 3 H]ryanodine binding to skeletal and cardiac membranes and also activity of the reconstituted skeletal RyR (15). This peptide corresponds to a region within the C terminus, which was shown not to affect whether chimeric DHPRs could mediate skeletal-type EC coupling (3).
Just as there are important limitations on experiments with chimeric DHPRs (i.e. the inability to test for direct interactions between the DHPR and RyR), there are also significant weaknesses in the experiments with isolated peptides. For example, the isolated peptides may assume conformations different from those of the corresponding regions of the DHPR in living cells. Equally or more important, the isolated peptides may act at sites on the RyR that are different from those actually involved in EC coupling. By expressing cDNAs encoding engineered RyRs in dyspedic myotubes in which the RyR-1 gene is disrupted (21,22), it may be possible to identify regions of the RyR critical for skeletal-type EC coupling. In turn, this will make it possible to determine whether one or more of the isolated peptides acts at sites coincident with those identified by the expression of RyR cDNAs. Thus, progress in our understanding of EC coupling is likely to continue benefiting from analysis of DHPR/RyR interactions using both isolated proteins and expression of cDNAs in dysgenic and dyspedic myotubes.