The II-III Loop of the Skeletal Muscle Dihydropyridine Receptor Is Responsible for the Bi-directional Coupling with the Ryanodine Receptor*

The dihydropyridine receptor (DHPR) in the skeletal muscle plasmalemma functions as both voltage-gated Ca2+ channel and voltage sensor for excitation-contraction (EC) coupling. As voltage sensor, the DHPR regulates intracellular Ca2+ release via the skeletal isoform of the ryanodine receptor (RyR-1). Interaction with RyR-1 also feeds back to increase the Ca2+ current mediated by the DHPR. To identify regions of the DHPR important for receiving this signal from RyR-1, we expressed in dysgenic myotubes a chimera (SkLC) having skeletal (Sk) DHPR sequence except for a cardiac (C) II-III loop (L). Tagging with green fluorescent protein (GFP) enabled identification of expressing myotubes. Dysgenic myotubes expressing GFP-SkLC or SkLC lacked EC coupling and had very small Ca2+currents. Introducing a short skeletal segment (α1Sresidues 720–765) into the cardiac II-III loop (replacing α1C residues 851–896) of GFP-SkLC restored both EC coupling and Ca2+ current densities like those of the wild type skeletal DHPR. This 46-amino acid stretch of skeletal sequence was recently shown to be capable of transferring strong, skeletal-type EC coupling to an otherwise cardiac DHPR (Nakai, J., Tanabe, T., Konno, T., Adams, B., and Beam, K.G. (1998) J. Biol. Chem.273, 24983–24986). Thus, this segment of the skeletal II-III loop contains a motif required for both skeletal-type EC coupling and RyR-1-mediated enhancement of Ca2+ current.

Excitation-contraction (EC) 1 coupling in skeletal muscle depends upon a functional interaction between dihydropyridine receptors (DHPRs) in the plasmalemma and ryanodine receptors (RyRs) in the sarcoplasmic reticulum (SR). In skeletal muscle, the DHPR functions both as an L-type Ca 2ϩ channel and as the voltage sensor, which in response to plasmalemmal depolarization, transmits a signal that causes RyR-1 (the skeletal RyR isoform) to release Ca 2ϩ from the SR (1)(2)(3). The nature of the signal transmitted from the skeletal DHPR to RyR-1 is not yet understood, although there is strong evidence that skeletal-type EC coupling does not rely upon the entry of external Ca 2ϩ (4).
An approach to identifying regions of the skeletal DHPR that are important for EC coupling has been to express cDNAs encoding chimeric DHPRs in dysgenic myotubes, which lack endogenous skeletal DHPRs. This work has shown that a purely cardiac DHPR expressed in dysgenic myotubes results in EC coupling which is cardiac type (dependent on entry of Ca 2ϩ ) (5), whereas skeletal-type EC coupling results from expression of a chimeric DHPR having cardiac sequence except for a skeletal II-III loop (6). More recently, it was shown that strong skeletal-type coupling could be produced by a chimeric DHPR that contained only a 46-amino acid skeletal segment within the II-III loop and weak skeletal-type coupling by a chimera containing only an 18-amino acid skeletal segment (7).
Analysis of myotubes from dyspedic mice, which lack RyR-1, has revealed that in addition to the orthograde (EC coupling) signal transmitted from the skeletal DHPR to RyR-1, there also appears to be a retrograde signal whereby RyR-1 increases the magnitude of the L-type Ca 2ϩ current mediated by the DHPR. In particular, Ca 2ϩ current density is very low in dyspedic myotubes even though the surface density of DHPRs appears to be essentially normal (8,9). Expression of cDNA encoding RyR-1 causes the density of L-type current in dyspedic myotubes to increase toward normal (8). However, these experiments did not reveal whether the region of the skeletal DHPR that is crucial for orthograde coupling is also important for the RyR-1-mediated enhancement of DHPR Ca 2ϩ current.
Here we describe experiments to identify regions of the skeletal DHPR that are critical for the ability of the DHPR to receive the retrograde (current-enhancing) signal from RyR-1. The results demonstrate that the II-III loop is critical for both orthograde and retrograde signaling. Within the II-III loop, the 46-amino acid segment found to be important for skeletal-type EC coupling is also important for transducing the retrograde signal from RyR-1.

Construction of Chimeric DHPRs
Chimeras between the ␣ 1 subunits of the skeletal muscle DHPR (Sk (10)) and the cardiac muscle DHPR (C (11)) had amino acid composition (numbers in parentheses) as follows.
SkLC SkLC-The EcoRI-XmnI fragment of Sk (nt 1007-1964) was coligated with the ligation product from the XmnI-HincII fragment of C (nt 2330 -2782) plus the HincII-XhoI fragment of Sk (nt 2389 -2654) into the corresponding EcoRI/XhoI restriction sites of a SacII-XhoI subclone of Sk (nt 86 -2654) in pBluescript SKϩ (Stratagene). Finally, the SacII-XhoI insert of the modified subclone (now carrying the cardiac II-III loop sequence) was ligated into the corresponding SacII/XhoI restriction sites of plasmid pCAC6, which contains the complete skeletal DHPR coding region in the mammalian expression vector pKCRH2 (6).
GFP-␣ 1S -The coding sequence of the ␣ 1 subunit of the skeletal muscle DHPR (10) was inserted in-frame and downstream of the coding region of a modified green fluorescence protein (GFP), cloned in a proprietary mammalian expression vector (kindly provided by P. Seeburg) as described in detail elsewhere (12).
␤ 1b -The Ca 2ϩ channel ␤ 1b subunit cDNA (kindly provided by K. Campbell) was cloned as a SacI-HindIII fragment (5Ј and 3Ј polylinker, respectively) into the SacI/HindIII polylinker sites of the mammalian expression vector pSV-SPORT1 (LifeTechnologies, Inc.). The integrity of all the chimeric DHPRs was confirmed by sequence analysis using an ABI 377 automatic sequencer.

Expression of cDNA
Primary cultures of myotubes isolated from newborn dysgenic mice were prepared as described previously (13). Approximately 1 week after plating, myotubes were microinjected (2) into a single nucleus with solutions of expression plasmids (300 -600 ng/l) carrying cDNAs for either GFP-␣ 1S , GFP-SkLC, GFP-SkLCS 46 , or GFP-SkLCS 18 . Injected myotubes were subsequently examined for the development of green fluorescence. Expressing cells were evaluated for contraction (2) in response to electrical stimulation (80 V, 10 -30 ms), macroscopic Ca 2ϩ currents, immobilization-resistant intramembrane charge movement (14), and subcellular channel distribution (only for GFP-SkLC). In a separate set of experiments examining the role of the ␤ 1b subunit for Ca 2ϩ channel enhancement, dysgenic myotubes were coinjected with GFP-SkLC cDNA (600 ng/l) and 350 ng/l ␤ 1b -carrying mammalian expression plasmid. Additionally, dyspedic myotubes were grown in primary culture as described for dysgenic myotubes (13) and mononuclearly injected (2) with 350 ng/l ␤ 1b -carrying mammalian expression plasmid together with pure GFP vector (25 ng/l) to enable the identification of expressing cells.

Electrophysiological Characterization
Macroscopic Ca 2ϩ currents were measured using the whole-cell patch clamp technique (15). The patch pipettes (borosilicate glass) had resistances of 1.5-1.9 M⍀ when filled with an internal solution containing 140 mM cesium aspartate, 10 mM Cs 2 -EGTA, 5 mM MgCl 2 , and 10 mM HEPES (pH 7.4 with CsOH). The composition of the external bath solution was 10 mM CaCl 2 , 145 mM tetraethylammonium chloride, 3 M tetrodotoxin, and 10 mM HEPES (pH 7.4 with tetraethylammonium hydroxide). Test pulses were preceded by a 1-s prepulse to Ϫ30 mV to inactivate endogenous T-type Ca 2ϩ currents (14). Test currents were corrected for linear components of leakage and capacitative currents by digitally scaling and subtracting the average of 10 preceding control currents, elicited by hyperpolarizing voltage steps (20 -40 mV amplitude) applied from the holding potential of Ϫ80 mV. Ca 2ϩ currents were normalized by linear cell capacitance (expressed in pA/pF). After the recording of whole-cell Ca 2ϩ currents, 0.5 mM Cd 2ϩ , and 0.1 mM La 3ϩ were added to the external bath solution to enable the recording of immobilization-resistant intramembrane charge movement (gating currents). The procedure for recording and calculating maximum charge movement densities and the prepulse protocol used was described in detail elsewhere (14,16). To examine the effect of Ca 2ϩ release on sarcolemmal Ca 2ϩ current, Ca 2ϩ current and Ca 2ϩ transients were measured (17) in normal myotubes with the external solution described above for Ca 2ϩ currents and patch pipettes containing an internal solution composed either of 145 mM cesium glutamate, 8 mM MgATP, 0.5 mM K 5 -Fluo-3 (Molecular Probes, Eugene, OR), 2 mM CsCl, 10 mM EGTA, 10 mM HEPES, pH 7.4, with CsOH (10 EGTA solution) or 65 mM cesium glutamate, 5 mM MgCl 2 , 0.5 mM K 5 -Fluo-3, 40 mM BAPTA, 10 mM HEPES, pH 7.4 with CsOH (40 BAPTA solution). For the measurement of Ca 2ϩ transients in dysgenic myotubes expressing chimeric DHPRs, the pipette contained 145 mM cesium glutamate, 8 mM MgATP, 0.5 mM K 5 -Fluo-3, 0.1 mM EGTA, 2 mM CsCl, 10 mM HEPES (pH 7.2 with CsOH), and the external solutions was 150 mM tetraethylammonium chloride, 10 mM HEPES, 5 mM CaCl 2 , 1 mM MgCl 2 , 1 M tetrodotoxin (pH 7.2 with tetraethylammonium hydroxide). For the measurements of Ca 2ϩ transients, it was not suitable to use the GFP-tagged constructs that had fluorescence excitation and emission wavelengths close to those of Fluo-3. Thus, cDNAs coding for SkLC, SkLCS 46 , and SkLCS 18 were inserted into the expression plasmid pKCRH2 (18) and were coinjected with cDNA encoding the ␣ subunit of the human surface antigen CD8 (19). Myotubes expressing the mutant channels were identified using polystyrene beads coated with CD8 antibodies as described previously (20). Transient changes in fluorescence (⌬F) were normalized by the resting fluorescence (F). The maximum rate of change of ⌬F/F was determined by fitting a line segment to the steepest portion of the transient. All recordings were made at room temperature (ϳ20°C) and data are reported as mean Ϯ S.D.

Laser-scanning Confocal Microscopy
GFP-SkLC-expressing dysgenic myotubes cultured on 35-mm culture dishes were superfused with a normal rodent Ringer solution (145 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES, pH 7.4 with NaOH) and mounted under a glass coverslip. The culture dish was subsequently fastened upside-down on the stage of a Nikon inverted microscope. Fluorescing cells were analyzed using a Sarastro 2000 confocal laser-scanning microscope (Molecular Dynamics) with a Nikon 60ϫ PlanApo oil immersion objective (numerical aperture 1.40) and the ImageSpace™ software (Silicon Graphics Inc., Mt. View, CA). GFP excitation/emission was achieved with a filter set (488 nm/510 nm) designed for fluorescein detection. Images were 1024 ϫ 1024 pixels with a pixel size of 0.11 m.
Step size between confocal sections was 2 m. Images were processed using the Adobe Photoshop software (ADOBE Systems, Mountain View, CA).

The Skeletal Muscle DHPR II-III Loop Is Essential for Receiving the Retrograde (Current-enhancing) Signal-Previous
work showed that the cardiac DHPR expressed in dysgenic (DHPR-deficient) myotubes was unable to mediate orthograde signaling (i.e. skeletal EC coupling). However, a chimera with cardiac sequence except for a skeletal II-III loop (CSk3 (5)) could mediate orthograde signaling. More recently, it was found that the expression of RyR-1 in dyspedic (RyR-deficient) myotubes increased the amplitude of slow L-type Ca 2ϩ current produced by the skeletal DHPR (8). To determine whether the II-III loop plays an important role in "receiving" this currentenhancing signal from RyR-1, we constructed a chimeric DHPR (SkLC) that was the "inverse" of CSk3: a skeletal DHPR except for a cardiac II-III loop. Electrically evoked contractions were never observed in dysgenic myotubes that had been injected with SkLC, consistent either with the possibility that the chimera was nonfunctional or that a cardiac II-III loop abolished orthograde signaling. Because electrically evoked contraction could not be used to identify myotubes expressing SkLC, we constructed a cDNA that fused GFP to the amino terminus of SkLC ( Fig. 1a) to allow definitive identification by means of in situ fluorescence. Fusion proteins of this sort were shown previously not to affect the function of either muscle or brain Ca 2ϩ channels (12). Myotubes expressing GFP-SkLC displayed slowly activating Ca 2ϩ currents (Fig. 2b), which were much smaller in amplitude than those present in GFP-␣ 1S -expressing myotubes (Fig. 2a). To allow comparisons between cells, peak current-voltage relationships were fitted (14) to yield a value of maximal Ca 2ϩ conductance (G max ). The value of G max for GFP-SkLC was significantly (p Ͻ 0.005) smaller than for GFP-␣ 1S (Table I). This decrease in G max for GFP-SkLC did not appear to be a consequence of a reduced number of DHPRs expressed in the surface membrane because values for maximal charge movement (Q max ) were similar (p Ͼ 0.05) for GFP-SkLC and GFP-␣ 1S (Fig. 2, d and e; Table I). The ratio of G max to Q max Ј (Q max Ј equals Q max minus the average, endogenous charge in dysgenic myotubes; Ref. 14) for GFP-SkLC was less than half that for GFP-␣ 1S (Table I). Thus, it appears that the presence of a cardiac II-III loop prevents GFP-SkLC from receiving the current-enhancing signal from RyR-1. Indeed, the value of G max /Q max Ј for GFP-SkLC was very close to the value found for dyspedic myotubes (8), which have ␣ 1S but lack RyR-1.
Identification of a Skeletal Segment in the II-III Loop Sufficient to Restore Both Skeletal-type EC Coupling and Enhancement of Ca 2ϩ Current-Nakai et al. (7) previously showed that substitution of a 46-amino acid segment of skeletal sequence into the II-III loop of an otherwise cardiac DHPR produced a chimera (CSk53; ␣ 1S residues 720 -765) capable of mediating strong, skeletal-type EC coupling upon expression in dysgenic myotubes. To determine whether this motif (Fig. 1b) is also sufficient to allow reception of the Ca 2ϩ current-enhancing signal from RyR-1, we substituted this 46-residue segment into the cardiac II-III loop of the otherwise skeletal chimera GFP-SkLC. The resulting chimera, GFP-SkLCS 46 (Fig. 1a), not only mediated skeletal-type EC coupling (electrically evoked contraction of more than half of the fluorescent cells tested in Cd 2ϩ /La 3ϩ , n Ͼ 50; data not shown) but also produced large Ca 2ϩ current densities (Fig. 2c) (Table I).
Nakai et al. (7) also showed that an otherwise cardiac chimera containing an even shorter (18-residue) skeletal segment (CSk58, ␣ 1S residues 725-742) was still able to mediate skeletal-type EC coupling; however, this coupling was weak (7). To test if this 18-amino acid segment allows reception of the channel-enhancing signal from RyR-1, we constructed chimera GFP-SkLCS 18 (Fig. 1, a and b). The value of G max for GFP-SkLCS 18 was not significantly different from that of GFP-SkLC (Table I; p Ͼ 0.05). Additionally, the G max /Q max Ј ratio for GFP-SkLCS 18 was similar to that found for GFP-SkLC expressed in dysgenic myotubes or that of endogenous ␣ 1S in dyspedic myotubes that lack RyR-1 (Table I). Therefore, the minimal DHPR sequence that allows strong enhancement of Ca 2ϩ current by RyR-1 is incomplete in, or missing from, the 18-residue skeletal segment in the II-III loop of GFP-SkLCS 18 . However, this minimal sequence is contained within the 46-residue skeletal segment of the GFP-SkLCS 46 II-III loop.
Skeletal-type EC Coupling Is Very Weak for SkLCS 18  . Current amplitudes were normalized by linear cell capacitance and are expressed as pA/pF. d-f, immobilization-resistant intramembrane charge movements were recorded in response to a depolarization to ϩ40 mV following a prepulse protocol (14). Recordings of charge movement were obtained from the same myotubes as shown above under a-c after blocking Ca 2ϩ currents with a test solution containing 0.5 mM Cd 2ϩ and 0.1 mM La 3ϩ . The linear cell capacitance (C) for each cell was as follows: a and d, cell b67, C ϭ 436 pF; b and e, cell b48, C ϭ 520 pF; c and f, cell c08, C ϭ 588 pF. of the strength of retrograde coupling (current enhancement from RyR-1). To obtain a similarly quantitative assessment of the strength of orthograde (EC) coupling to RyR-1, we measured depolarization-induced Ca 2ϩ transients. Depolarizationinduced Ca 2ϩ transients were never observed with SkLC (Fig.  3a, 0 of 13 cells tested) but were routinely observed for SkLCS 46 (Fig. 3b, 16 of 16 cells tested). The transients support the conclusion that SkLCS 46 mediates skeletal-type EC coupling because they were of similar magnitude for test pulses to ϩ30 mV (where Ca 2ϩ current is near maximal) and ϩ80 mV (where Ca 2ϩ current is small as a result of reduced driving force). By the same logic, SkLCS 18 was also able to mediate skeletal-type coupling because the Ca 2ϩ transients were again similar at ϩ30 and ϩ80 mV (Fig. 3c). However, the maximal rate of increase of the ⌬F/F signal (at ϩ80 mV) was only 0.023 Ϯ 0.012 ms Ϫ1 (n ϭ 12) for SkLCS 18 , which is almost 5-fold lower than the value of 0.112 Ϯ 0.025 ms Ϫ1 (n ϭ 16) for SkLCS 46 . Thus, skeletal-type coupling is much weaker for SkLCS 18 than for SkLCS 46 .
It is possible to calculate the enhancement of Ca 2ϩ current that might have been expected for SkLCS 18 relative to that measured for SkLCS 46 under the assumption that there is a linear relationship between the strengths of retrograde and orthograde signaling. The enhancement of current for SkLCS 46 can be defined as (G/Q 46 Ј Ϫ G/QЈ) Ϭ Q 46 Ј , where G/Q 46 Ј and G/QЈ are the values of G max /Q max Ј for GFP-SkLCS 46 and GFP-SkLC, respectively. With the values from Table I, the enhancement of current for SkLCS 46 was ϳ1.5. If the enhancement of current for SkLCS 18 was, like orthograde signaling, 5-fold smaller than for SkLCS 46 (see above), it would yield a predicted enhancement of only ϳ0.3, a value probably too small to have been detectable.
Ca 2ϩ Release Is Not Responsible for Enhancement of Ca 2ϩ Current-As described above, the chimeric DHPR constructs able to "receive" the Ca 2ϩ current-enhancing signal from RyR-1 were exactly the same as those able to "transmit" the EC coupling signal to RyR-1. Thus, it seemed possible that Ca 2ϩ released from RyR-1 (in response to the EC coupling signal) represented the feedback signal, causing the enhancement of Ca 2ϩ current. As a way of testing this possibility, we carried out simultaneous measurements of Ca 2ϩ currents and Ca 2ϩ transients in normal myotubes using two different pipettefilling solutions. One of these solutions (10 EGTA) contained 10 mM EGTA (to mimic the standard solution we used for measuring Ca 2ϩ currents) plus ATP to support Ca 2ϩ re-uptake into the SR. The other solution (40 BAPTA) contained 40 mM BAPTA (to buffer myoplasmic Ca 2ϩ strongly) and lacked ATP so as to hinder Ca 2ϩ re-uptake into the SR. To ensure thorough dialysis, we selected only small myotubes (232 Ϯ 76 pF, n ϭ 9) with compact geometry and used low resistance patch pipettes (0.9 to 1.4 M⍀). For the cells analyzed, the uncompensated access resistance remained low (1.85 Ϯ 0.33 M⍀, n ϭ 9) after entry into whole-cell mode. Fig. 4 illustrates Ca 2ϩ currents and Ca 2ϩ transients evoked by constant amplitude depolarizations applied at the indicated times after breaking into a normal

FIG. 3. Depolarization-induced Ca 2؉ transients in dysgenic myotubes expressing DHPR chimeras SkLC (a), SkLCS 46 (b) and
SkLCS 18 (c). The holding potential was Ϫ90 mV, and cells were depolarized to the indicated test potentials following a prepulse protocol (14). Note that both the rate of change and maximal increase of ⌬F/F are much smaller for SkLCS 18 (14); I, peak current activated at test potential V; V rev , extrapolated reversal potential; V G , potential for activation of half-maximal conductance; k G , slope factor. Values of immobilization-resistant Q ON were determined as described previously (14) and were fitted according to ; Q max , maximum immobilizationresistant charge movement; V, test potential; V Q , potential at which half the charge has moved; k Q , slope factor. QЈ max is the difference between Q max and the average, endogenous charge movement Q dys(max) found in dysgenic myotubes (Q dys(max) ϭ 2.5 nC/F; (14)). For all the data given, the estimated series resistance error was Ͻ10 mV. Brackets indicate two data sets compared statistically by an unpaired two-sample t test. Asterisks indicate statistically significant differences (p Ͻ 0.005), whereas no asterisk indicates p Ͼ 0.05. Values for dyspedic myotubes and for ␣ 1S -expressing dysgenic myotubes were listed for comparison and were published previously (8,14).  myotube with either 10 EGTA (a) or 40 BAPTA (b). Similar results were obtained for a total of 5 cells studied with 40 BAPTA and 4 cells with 10 EGTA. With 10 EGTA in the pipette, depolarization-evoked Ca 2ϩ release was sufficient to cause a transient increase in the fluorescence (⌬F) of the indicator dye Fluo-3. Note that both ⌬F and the base-line fluorescence (F) increased between 2.5 and 7.5 min after break-in with 10 EGTA, suggesting that during this time Fluo-3 was diffusing into the cell. Because both ⌬F and F remained stable at longer times, it appeared that 7.5 min was sufficient for equilibration between the pipette solution and the myoplasm. With 40 BAPTA in the pipette, depolarization failed to elicit a transient increase in fluorescence, and the base-line fluorescence remained very low, presumably because Ca 2ϩ was buffered so strongly that virtually all of the Fluo-3 entering the cell remained in the Ca 2ϩ -free form. The absence of evoked fluorescence increases with 40 BAPTA indicates that there was effective buffering of Ca 2ϩ released from the SR (where Ca 2ϩ stores had likely been depleted). Therefore, the measurements with 40 BAPTA should give an indication of the behavior of Ca 2ϩ currents in myotubes where Ca 2ϩ transients near release sites were substantially suppressed.
As is evident in Figs. 4, a and b, Ca 2ϩ current amplitude ran down as a function of time after breaking into the cell with either 10 EGTA or 40 BAPTA. In fact, the rundown in these experiments (with small cells and low access resistance) was faster than that observed under the conditions we normally used for measurements of Ca 2ϩ currents (larger cells, higher access resistance). Several factors may have contributed to the more rapid rundown with the 40 BAPTA, including the much lower level of resting free Ca 2ϩ and the absence of ATP. However, because large Ca 2ϩ currents were still present at times when junctional Ca 2ϩ transients were effectively suppressed, it seems unlikely that Ca 2ϩ release represents the critical feedback signal whereby RyR-1 enhanced Ca 2ϩ current.
GFP-SkLC Ca 2ϩ Channels Cluster in Punctate Foci-By means of immunostaining, skeletal DHPRs in normal myotubes were shown to cluster in foci that colocalize with RyR clusters (21). Confocal microscopy also reveals focal clusters in living, dysgenic myotubes expressing GFP-tagged DHPRs (12). To determine whether the absence of either orthograde or retrograde signaling by GFP-SkLC (Figs. 2a and 3a, respectively) was a consequence of failure to co-localize with RyRs, we used confocal microscopy to determine whether or not focal clusters were present in dysgenic myotubes expressing this chimeric construct. As shown in Fig. 5, focal clusters were present in GFP-SkLC-expressing dysgenic myotubes. The pattern of distribution of these clusters does not appear qualitatively different from that of GFP-tagged DHPRs (12), which are capable of interacting with the RyRs of the SR. DISCUSSION We have found that replacing the II-III loop of the skeletal DHPR with the corresponding region of the cardiac DHPR causes the loss of two functions. This skeletal DHPR with a cardiac II-III loop (SkLC) can neither transmit the orthograde (EC coupling) signal to the skeletal ryanodine receptor (RyR-1) nor receive the retrograde (current enhancing signal) from RyR-1. Substitution of a 46-amino acid segment of skeletal sequence into the cardiac loop of SkLC restores both orthograde and retrograde signaling.
A Role for the ␤ 1b Subunit?-Because Ca 2ϩ currents are of small amplitude in skeletal muscle cells lacking RyR-1, we suggested in an earlier study that the small Ca 2ϩ currents observed after heterologous expression of skeletal DHPRs in nonmuscle cells might be a consequence of the absence of RyR-1 in these cells (8). Recently, however, it was shown that large Ca 2ϩ currents could be produced with skeletal DHPRs expressed in Xenopus oocytes if the ␤ 1b subunit was used instead of ␤ 1a (22), the predominant ␤ isoform in skeletal muscle (23). This work did not establish whether the ␤ 1b subunit simply increased expression of DHPRs in the oocyte plasmalemma or actually increased the current without changing the number of plasmalemmal DHPRs (as we have shown is likely the case for enhancement of current by RyR-1). Our preliminary experiments suggest that for DHPRs in their normal environment (muscle cells), expression of ␤ 1b does not overcome the loss of interaction with RyR-1. In particular, neither expression of ␤ 1b in dyspedic myotubes (G max /Q max Ј ratio: 14 nS/pC; n ϭ 6) nor co-expression of ␤ 1b together with GFP-SkLC in dysgenic myotubes (G max /Q max Ј ratio: 12 nS/pC; n ϭ 4) yielded values that were much different from the corresponding values obtained without ␤ 1b co-expression (G max /Q max Ј ratios of 12 nS/pC and 15 nS/pC, respectively; see Table I).
What is the Mechanism of Enhancement of Current?-The mechanisms of orthograde and retrograde signaling between the skeletal DHPR and RyR-1 remain to be established. One possible explanation of retrograde signaling is that the Ca 2ϩ released during EC coupling feeds back onto the DHPR to enhance current. This hypothesis is compatible with the observation that precisely those chimeras that did not show enhancement of current were those that also lacked (SkLC) or had only weak (SkLCS 18 ) EC coupling. Furthermore, Feldmeyer et al. (24) have presented evidence that Ca 2ϩ release may modulate the Ca 2ϩ current in cut fibers from frog skeletal muscle, including the demonstration that prolonged , suggesting that there may have been disruption of the t-tubular system or a loss of the ability of the DHPR to undergo the voltage-driven conformational changes producing charge movement. Either of these kinds of changes would not have affected our analysis, which indicates that functional coupling of the DHPR to RyR-1 is associated with large differences in G max /Q max Ј , the ratio of Ca 2ϩ conductance to charge movement. Furthermore, in contrast to the results on frog skeletal muscle (24), our experiments showed that large Ca 2ϩ currents were present in mouse myotubes in which Ca 2ϩ transients near release sites should have been largely suppressed by dialysis with 40 mM BAPTA (Fig. 4). Negligible effects on maximal Ca 2ϩ conductance have also been previously reported for dialysis of mouse myotubes with 1 mM ryanodine, 200 M ruthenium red, or 20 mM BAPTA (25).
Results from work on DHPR chimeras (14) also argue against an essential role of Ca 2ϩ release in enhancement of current. In that study it was found that G max /Q max Ј was 55 nS/pC for CARD1 (the cardiac DHPR) and 157 nS/pC for CSk3 (the cardiac DHPR with a skeletal II-III loop). Thus, it appears that the presence of a skeletal II-III loop enhanced the current via a mechanism not strongly dependent on Ca 2ϩ release, because both CARD1 and CSk3 support depolarization-induced Ca 2ϩ release under the conditions used for measurement of Ca 2ϩ currents (7). Data from RyR-1/RyR-2 chimeras provide another argument that Ca 2ϩ released via skeletal-type EC coupling is not required for enhancement of current. In particular, expression in dyspedic (RyR-1 lacking) myotubes of the chimera R9 produced enhancement of Ca 2ϩ current but not restoration of skeletal-type EC coupling (26). Finally, recent experiments show that Ca 2ϩ currents are enhanced in dyspedic myotubes after expression of a mutated ryanodine receptor, which releases almost no Ca 2ϩ in response to depolarization (27).
An alternative to the idea that the release of Ca 2ϩ from RyR-1 causes enhancement of current is to suppose that protein-protein interactions are responsible. Fig. 6 illustrates a model in which EC coupling involves transmission of a signal from the skeletal DHPR to RyR-1 via the II-III loop, and enhancement of current involves transmission of a retrograde signal from RyR-1 to the DHPR, again via the II-III loop (an intermediary protein coupling between the II-III loop and RyR-1 is another possibility). The nature of both the orthograde and retrograde signals remains unknown (for example, the retrograde signal might correspond to a covalent modification of the DHPR). However, in the illustrated model, interaction of RyR-1 with the II-III loop stabilizes the DHPR in a conformation (Fig. 6a), which increases single channel current and/or Po (channel open probability) compared with the conformation of the DHPR found in the absence of this interaction (Fig. 6b, no RyR-1; Fig. 6c, cardiac II-III loop). Both skeletal-type EC coupling and the enhancement of current are restored by introduction of a small segment of the skeletal II-III loop (Fig. 6d).
The importance of the II-III loop is emphasized by complementary gain-of-function and loss-of-function experiments. A gain of function (skeletal-type EC coupling) was shown with CSk3 in which the skeletal II-III loop was transplanted into the cardiac DHPR (5). As discussed above, these same experiments also suggest a second gain of function (enhancement of current) because G max /Q max Ј was ϳ3-fold larger for CSk3 than for CARD1 (14). The experiments reported here now demonstrate a loss of both functions (skeletal-type EC coupling, enhancement of current) when the cardiac II-III loop is transplanted into the skeletal DHPR (i.e. SkLC) and a restoration of both functions with SkLCS 46 .  (5), or in dyspedic myotubes injected with RyR-1 cDNA (8). The cytoplasmic II-III loop of the skeletal DHPR is critical for transmitting the signal controlling the release of Ca 2ϩ ions via RyR-1 (5) in the SR membrane. This skeletal-type EC coupling (Skeletal ECC) is not dependent on influx of extracellular Ca 2ϩ . The II-III loop is also essential for receiving the current-enhancing signal from RyR-1 (Channel enhancement). Panel b depicts the situation as found in dyspedic muscle. Dyspedic myotubes lack RyR-1 but have intact DHPRs at a density comparable with normal myotubes (8). Dyspedic myotubes display no EC coupling and also show significantly reduced slow Ca 2ϩ current densities through the DHPR (8). In the absence of contact with RyR-1, the skeletal DHPR assumes a conformation (symbolized by the tilted cylinders representing homologous repeats I-IV) that produces reduced Ca 2ϩ current (symbolized by the smaller arrow). Panel c models the behavior of chimera GFP-SkLC expressed in dysgenic myotubes. The cardiac II-III loop (Cardiac loop) in an otherwise skeletal DHPR prevents the DHPR-RyR-1 interaction so that there is neither EC coupling nor appreciable Ca 2ϩ current. Panel d shows that the introduction of a short skeletal segment (␣ 1S residues 720 -765, symbolized by a bold line), which sufficed to transfer strong skeletal-type EC coupling to the cardiac DHPR (as in CSk53 described in Ref. 7), is also sufficient to restore wild-type Ca 2ϩ current densities (as in chimera SkLCS 46 ). Together, these observations suggest that these 46 amino acids of the skeletal II-III loop contain residues that are required for both strong skeletal-type EC coupling and RyR-1-mediated enhancement of skeletal Ca 2ϩ current.
For a model like the one in Fig. 6, how would one interpret the observation that orthograde and retrograde coupling are weak for SkLCS 18 ? One possibility is that the great majority of SkLCS 18 DHPRs and RyRs are simply not in physical contact because SkLCS 18 lacks part of the required sequence. However, it seems very likely that SkLCS 18 clusters into foci at sites where the plasmalemma forms junctions with RyR-containing regions of the SR, because even SkLC clusters into foci (Fig. 5). Of course, a demonstration of co-localization of DHPRs and RyRs at the light microscopic level does not imply direct physical contact. Suggestive evidence for direct physical contact between DHPRs and RyRs in skeletal muscle has been provided by freeze-fracture analysis. This analysis has shown that skeletal DHPRs appear to be organized in characteristic tetrads (thought to be four DHPRs, each of which is in contact with one of the four subunits of a RyR) (28). By contrast, cardiac DHPRs appear to be located close to, but not in contact with, RyRs, because tetrads are not observed in cardiac muscle (29). Thus, it will be important to carry out freeze-fracture analysis to determine whether or not tetrads are formed upon expression of CSk3, SkLC, SkLCS 46 , and SkLCS 18 . If very few tetrads are observed for SkLCS 18 , it would suggest that the weak coupling for this construct was a result of loss of physical contact with RyR-1. If tetrad formation is comparable for SkLCS 18 and SkLCS 46 , it would suggest that tetrads of SkLCS 18 induce a lower channel activity of a RyR-1 tetramer than do tetrads of SkLCS 46 .