Non–Ca²⁺-conducting Ca²⁺ channels in fish skeletal muscle excitation-contraction coupling

Abstract

During skeletal muscle excitation-contraction (EC) coupling, membrane depolarizations activate the sarcolemmal voltage-gated L-type Ca²⁺ channel (Ca_V1.1). Ca_V1.1 in turn triggers opening of the sarcoplasmic Ca²⁺ release channel (RyR1) via interchannel protein–protein interaction to release Ca²⁺ for myofibril contraction. Simultaneously to this EC coupling process, a small and slowly activating Ca²⁺ inward current through Ca_V1.1 is found in mammalian skeletal myotubes. The role of this Ca²⁺ influx, which is not immediately required for EC coupling, is still enigmatic. Interestingly, whole-cell patch clamp experiments on freshly dissociated skeletal muscle myotubes from zebrafish larvae revealed the lack of such Ca²⁺ currents. We identified two distinct isoforms of the pore-forming Ca_V1.1α_1S subunit in zebrafish that are differentially expressed in superficial slow and deep fast musculature. Both do not conduct Ca²⁺ but merely act as voltage sensors to trigger opening of two likewise tissue-specific isoforms of RyR1. We further show that non-Ca²⁺ conductivity of both Ca_V1.1α_1S isoforms is a common trait of all higher teleosts. This non-Ca²⁺ conductivity of Ca_V1.1 positions teleosts at the most-derived position of an evolutionary trajectory. Though EC coupling in early chordate muscles is activated by the influx of extracellular Ca²⁺, it evolved toward Ca_V1.1-RyR1 protein–protein interaction with a relatively small and slow influx of external Ca²⁺ in tetrapods. Finally, the Ca_V1.1 Ca²⁺ influx was completely eliminated in higher teleost fishes.

Excitation-contraction (EC) coupling is understood as the mechanism linking membrane depolarization to the contraction of a muscle cell. Depolarization-induced activation of the sarcolemmal voltage-gated Ca²⁺ channel (Ca_V) triggers opening of the sarcoplasmic Ca²⁺ release channel (ryanodine receptor, RyR) to release from SR stores Ca²⁺ that is needed for myofibril contraction. Though in cardiac muscle the homologous isoform RyR2 is activated by the Ca²⁺ influx through Ca_V1.2, in skeletal muscle, conformational changes of Ca_V1.1 directly activate RyR1 via protein–protein interaction (1, 2). However, a simultaneously occurring Ca²⁺ current through Ca_V1.1 is observed in mammalian skeletal myotubes. In contrast to the large and rapidly activating currents that are needed for efficient RyR2 activation in cardiac myocytes (3, 4), the skeletal muscle Ca_V1.1 Ca²⁺ current is relatively small and slowly activating (5). Phylogenetically interesting is that EC coupling in skeletal muscle of very early chordates, such as amphioxus or tunicates, is activated by Ca²⁺ influx, and in lamprey muscles, maximum-force generation is dependent on Ca²⁺ currents, although direct Ca_V1.1-RyR coupling is already established (6). However, in cultured mammalian skeletal muscle cells the absence of extracellular Ca²⁺ or pharmacological blocking of Ca_V1.1 does not immediately affect intracellular Ca²⁺ release and contractions (7, 8). Despite the outcome of these short-term in vitro experiments, it remains unclear if the Ca_V1.1 Ca²⁺ influx into the intact muscle plays an essential role on the long-term.

Here we show that in skeletal muscle of teleost fishes the voltage-gated Ca²⁺ channel Ca_V1.1 is unable to conduct Ca²⁺ and, thus, turned from an ion channel to a pure voltage sensor to trigger RyR1 opening. Therefore, our findings suggest that on the level of the whole organism, Ca²⁺ influx via Ca_V1.1 is not required in teleost fishes. Furthermore, we discuss that this Ca_V1.1 nonconductivity stands at the topmost position of an evolutionary trajectory from Ca²⁺ influx-dependent, cardiac muscle-like EC coupling in skeletal muscles of early chordates to Ca²⁺ influx-independent EC coupling with small Ca²⁺ currents in mammals, and finally to the complete elimination of Ca_V1.1 Ca²⁺ influx in teleost fishes.

Results

Non–Ca²⁺-Conducting Ca_V1.1 Channels Perform EC Coupling in Zebrafish Skeletal Muscle.

Whole-cell patch clamp analysis of myotubes isolated from wild-type zebrafish larvae revealed the lack of voltage-dependent inward Ca²⁺ currents at physiologically relevant potentials between −40 and +70 mV (Fig. 1A). Also, replacement of the charge carrier Ca²⁺ by Ba²⁺ did not evoke inward L-type currents (Fig. 1D). However, correct voltage-dependent activation of the Ca_V1.1 channels was indicated by robust Ca_V1.1-specific gating currents (Fig. 1B) (9, 10) that reached the levels described for the Ca²⁺-conducting Ca_V1.1 in mouse myotubes (11). Gating currents result from depolarization-induced displacements of charges in the channel's voltage sensing elements, resulting in a conformational change of the channel complex (12). This steric effect is not only inducing channel opening but is also responsible for the direct coupling of Ca_V1.1 with the Ca²⁺ release channel RyR1 in the SR (2). To test for functional Ca_V1.1-RyR1 interaction, depolarization-induced intracellular Ca²⁺ release from the SR through RyR1 was recorded using the intracellular Ca²⁺ indicator dye Fluo-4. Robust Ca²⁺ transients indicated intact EC coupling in these myotubes (Fig. 1C). Thus, zebrafish Ca_V1.1 is able to undergo the voltage-dependent conformational change required for channel activation and to trigger EC coupling, but completely lost its function to conduct calcium ions.

Fig. 1.

Non–Ca²⁺-conducting Ca²⁺ channels in larval zebrafish skeletal myotubes. (A) Depolarizing test pulses to potentials unto +70 mV were unable to elicit Ca²⁺ currents at any potential (n = 14). (B) Robust voltage-dependent α_1S gating is indicated by charge movements with a Q_max value of 10.45 ± 0.58 nC/μF and a half-maximal activation of −4.21 ± 1.36 mV (n = 20), assessed by integrating the “ON” component of α_1S gating currents (Q_on). (C) Depolarization-induced intracellular Ca²⁺ transients had a (ΔF/F₀) _max value of 4.44 ± 0.42 with half-maximal activation at −10.39 ± 0.75 mV (n = 17). (D) Also, 10 mM Ba²⁺ instead of 10 mM Ca²⁺ as a charge carrier did not awake any current at any potential (n = 6).

Two Ca_V1.1α_1S Isoforms Are Differentially Expressed in Zebrafish.

The central α_1S subunit forms the pore, the Ca²⁺ selectivity filter, and the voltage sensor of the skeletal muscle Ca²⁺ channel complex (12, 13), and thus primarily determines the channel’s conductivity properties. Therefore, we performed a BLAST search within the ENSEMBL zebrafish genomic assembly Zv7 with the coding sequence of the rabbit α_1S subunit (rb-α_1S) as a template (12). Surprisingly, two distinct genes coding for Ca_V1.1α_1S subunits were identified in the zebrafish genome and were located on chromosomes 8 and 22. Both genes code for intact α_1S subunits that contain all molecular characteristics, such as the interaction domains with the accessory β_1a subunit (14) and with RyR1 (15). One gene, identified on chromosome 22, consists of 43 exons and codes for an α_1S protein of 1,777 amino acids, hereafter named zf-α_1S-a, with 66% sequence identity to rb-α_1S. The second gene, identified on chromosome 8, is composed of 45 exons that code for the 1,847 amino acids of zf-α_1S-b with 64% sequence identity to rb-α_1S. The two zf-α_1S isoforms share an amino acid homology of 73%. In a phylogenetic tree based on parsimony analysis of DNA sequences (Fig. 2A), both zf-α_1S subunits cluster with rb-α_1S, clearly separated from the clade comprising the cardiac α_1C isoforms of rabbit and also zebrafish, consequently identifying them as α_1S subunits. Ancestral α₁ isoforms from the housefly Musca domestica (16) and the tunicate Halocynthia roretzi (17) were used as outgroups.

Fig. 2.

Two distinct α_1S subunit isoforms are found in the zebrafish skeletal musculature. (A) Singlemost parsimonious tree of aligned DNA sequences groups both zebrafish α_1S subunit isoforms (zf-α_1S-a and zf-α_1S-b) jointly with α_1S of the rabbit (rb-α_1S) and clearly separated from the clade comprising cardiac α_1C subunits of both species. Bootstrap scores of 100% (1,000 replications) for all branches support evolutionary divergence. (B) Quantitative RT-PCR reveals different expression patterns for zf-α_1S-a and zf-α_1S-b. Data represent mean ± SEM. In deep fast muscle we almost exclusively detect zf-α_1S-b, whereas in superficial slow muscle, zf-α_1S-a is the predominant isoform (***P < 0.001 by Student's paired t test). (C) Data obtained for zf-α_1S (B) closely match with expression levels of RyR1. (D–F) Whereas control experiments with rb-α_1S (D) restored Ca²⁺ inward currents (Top and Middle) in murine α_1S-null myotubes with I_max of −2.42 ± 0.18 pA/pF (n = 25), both, zf-α_1S-a (n = 14) and zf-α_1S-b (n = 12) were unable to restore any Ca²⁺ current (E and F). Ca²⁺ transients (Bottom) obtained with zf-α_1S-a had (ΔF/F₀) _max values of 0.84 ± 0.12 (n = 15), with a half-maximal activation at 9.91 ± 1.54 mV, and thus were smaller (P < 0.001) and shifted toward more positive potentials (P < 0.01) when compared with zf-α_1S-b [(ΔF/F₀) _max = 1.76 ± 0.2, with a half-maximal activation at 4.13 ± 1.4 mV, n = 17], which was comparable (P > 0.6) to rb-α_1S [(ΔF/F₀) _max = 1.95 ± 0.24, with a half-maximal activation at 3.52 ± 1 mV, n = 30].

The question now arises, why—in contrast to e.g., mammals—two α_1S subunit isoforms are present in zebrafish, and whether they might follow specific expression patterns. Recently, two isoforms of the skeletal muscle RyR1 were identified in zebrafish (18), with RyR1a being expressed in superficial slow muscle (also called red muscle), which is aerobic and used for persistent swimming, and RyR1b as the predominant isoform in deep fast muscle (white muscle), the glycolytic musculature used for burst activity. To test if RyR1a and RyR1b might be explicit molecular interaction partners for zf-α_1S-a and zf-α_1S-b, respectively, we performed quantitative RT-PCR on each of these muscle layers, isolated from adult zebrafish. In deep fast muscle, zf-α_1S-b was by far the predominant isoform, whereas in superficial slow muscle the expression level of zf-α_1S-a was significantly higher compared with zf-α_1S-b (Fig. 2B). The considerable signal obtained for zf-α_1S-b in superficial slow muscle most likely results from a preparation artifact, because this superficial slow muscle layer is very thin in the relatively small zebrafish and thus hard to separate cleanly. Therefore, we performed the same experiments on the musculature of another species of the same superorder (Ostarioclupeomorpha), the sardine (Sardina pilchardus), in which the tissue separation is much clearer due to a robust superficial slow muscle layer of this very active swimmer. Here, a perfect separation of the two α_1S and RyR1 isoforms in both muscle types was revealed (Fig. S1). As shown in Fig. 2 B and C, the expression patterns of zf-α_1S-a and zf-α_1S-b perfectly match with those of the two RyR1 isoforms, RyR1a and RyR1b.

Both zf-α_1S Isoforms Do Not Conduct Ca²⁺.

To test the two distinct isoforms for their ability to conduct Ca²⁺, both zf-α_1S cDNAs were generated by RT-PCR from zebrafish skeletal muscle and cloned into expression vector pGFP37, allowing N-terminal GFP tagging (7). Constructs were heterologously expressed in murine α_1S-null myotubes (cell line GLT) (19) and patch-clamp analyzed. In stark contrast to rb-α_1S controls (Fig. 2D), Ca²⁺ inward currents could not be detected either through zf-α_1S-a (Fig. 2E) or zf-α_1S-b (Fig. 2F). However, like the rb-α_1S controls, both zf-α_1S isoforms restored EC coupling with robust intracellular Ca²⁺ transients (Fig. 2 D–F). Notably, the Ca²⁺ transients recorded with zf-α_1S-a were smaller and shifted toward more positive potentials compared with those obtained with zf-α_1S-b (Table 1).

Table 1.

Intracellular Ca²⁺ transients recorded in expression experiments in GLT myotubes

Construct	(ΔF/F0)max	1/2V (ΔF/F0)max (mV)	n
rb-α1S	1.73 ± 0.13	2.28 ± 0.75	67
zf-α1S-a	0.84 ± 0.12 *	9.91 ± 1.54 **	15
zf-α1S-b	1.76 ± 0.2 *	4.13 ± 1.4 **	17
rb-α1S ME/TQ	1.93 ± 0.22	1.15 ± 1.4	20
rb-α1S N617D	1.46 ± 0.23	1.92 ± 1.71	12
rb-α1S D296K	1.75 ± 0.3	1.38 ± 1.64	15
RmRR	1.96 ± 0.37	2.15 ± 2.3	5

Robust intracellular Ca²⁺ transients indicate intact EC coupling for all constructs *P < 0.001; **P < 0.01.

Non–Ca²⁺-Conducting Ca_V1.1 Channels Exist Also in Other Teleost Fishes.

The question arises whether non–Ca²⁺-conducting Ca_V1.1 channels are a common attribute of teleost fishes or are an exclusive trait of zebrafish or its order Cypriniformes. Therefore, we patch-clamp analyzed larval myotubes from a phylogenetically closely related fish, the pike characin (Ctenolucius hujeta, Characiformes) (Fig. 3A), and from ricefish medaka (Oryzias latipes, Beloniformes) (Fig. 3B), which is much higher in the phylogenetic scale. Results from both teleost species matched those from zebrafish. Again, no Ca²⁺ inward currents, but pronounced intra-cellular Ca²⁺ transients, were recorded.

Fig. 3.

Non–Ca²⁺-conducting Ca²⁺ channels are found also in other teleost orders. (A and B) In skeletal myotubes from the pike characin Ctenolucius hujeta (A) and the ricefish medaka Oryzias latipes (B), no Ca²⁺ inward currents could be recorded upon test pulses unto +70 mV (Upper), despite the presence of pronounced intracellular Ca²⁺ transients with (ΔF/F₀) _max values of 6.05 ± 0.37 (n = 10) for the pike characin and 3.7 ± 0.54 (n = 6) for medaka, respectively (Lower).

Charged Pore Residues Are Correlated with Ca²⁺ Conductance in Both α_1S Isoforms.

Surprisingly, zf-α_1S / rb-α_1S sequence alignments showed that at position 303 of zf-α_1S-a, glutamate is consistently exchanged by glutamine when compared with rb-α_1S (Fig. 4 A and C). In a properly folded conducting Ca²⁺ channel, the glutamate at this position is the first of four glutamates located in the pore-forming loops (p-loops) of the four homologous repeats (Fig. 4A), thus forming the selectivity filter for Ca²⁺ ions (20). Consequently, Gln303 in zf-α_1S-a makes the formation of an intact Ca²⁺ selectivity filter doubtful. We tested the consequences of the hypothesized loss of an intact selectivity filter by exchanging the corresponding Glu292 together with the preceding Met291 to glutamine and threonine in the conducting rb-α_1S, according to zf-α_1S-a sequence. Construct rb-α_1S ME/TQ was unable to recover L-type Ca²⁺ currents in heterologous expression experiments (Fig. 4D). In contrast to the prominent exchange at position 303 of zf-α_1S-a, zf-α_1S-b shows a perfect conservation of the essential Ca²⁺ selectivity filter glutamates (Glu312/Glu633/Glu1034/Glu1348) (Fig. 4E). Instead, in homologous repeat II of zf-α_1S-b an additional negative charge (aspartate) in close proximity of the selectivity filter glutamate was identified, and this was able to completely block Ca²⁺ conductance when introduced into rb-α_1S in point mutant rb-α_1S N617D (Fig. 4F). Interestingly, this asparagine to aspartate mutation was also found in α_1S-b of the pike characin, but not in α_1S-b of the more distantly related medaka (Fig. 4G). Furthermore, a chimera of rb-α_1S, containing the repeat II S5/p-loop/S6 region of medaka α_1S-b, continued to show Ca²⁺ inward currents (Fig. S2). Unlike the addition of a negative charge (aspartate) in the repeat II p-loop of zf-α_1S-b, in medaka-α_1S-b, a conversion of a negative to a positive charge (aspartate to lysine compared with rb-α_1S) was found in the repeat I p-loop, four residues C-terminal to the selectivity filter glutamate (Fig. 4G). Again, the corresponding point mutant rb-α_1S D296K was unable to conduct Ca²⁺ inward currents (Fig. 4H). Intact EC coupling of all chimeras and point mutants mentioned previously was indicated by robust intracellular Ca²⁺ transients with (ΔF/F₀) _max values >1.4 (Table 1).

Fig. 4.

Amino acids incompatible with Ca²⁺ conductance. (A and B) In murine α_1S-null myotubes, expression of the wild-type α_1S from rabbit (rb-α_1S), with a typical amino acid arrangement of a conducting channel in the pore region of homologous repeats I–IV (RI-RIV), could restore Ca²⁺ inward currents of 1.79 ± 0.16 pA/pF (n = 37). (C) In zf-α_1S-a, a threonine and glutamine at position −1 and 0 relative to the selectivity filter (selectivity filter glutamates are indicated by gray bars) of homologous repeat I (RI) were identified and equally found in α_1S-a of pike characin and medaka. (D) Corresponding replacement of residues Met291 and Glu292 to threonine and glutamine in rb-α_1S (rb-α_1S ME/TQ) completely abolished Ca²⁺ inward currents (n = 9). Similar results were obtained with (F) rb-α_1S N617D (n = 8) in which Asn617 of rb-α_1S was exchanged to aspartate as found in α_1S-b of zebrafish (E) and pike characin, as well as with (H) rb-α_1S D296K (n = 15) in which Asp296 was exchanged to lysine as found in α_1S-b of medaka (G).

Non–Ca²⁺-Conducting Ca_V1.1 Channels Are a Common Trait of Higher Teleosts.

We tested for the presence of one of three marker amino acids incompatible with Ca_V1.1 Ca²⁺ conductance, in a large selection of teleosts using sets of degenerated α_1S-specific primers (Table S1) that amplify short fragments from genomic DNA or cDNA coding for the p-loop regions of homologous repeats I and II of α_1S-a and α_1S-b. Sequence analysis of these fragments from 32 species of 17 different orders of teleost fishes (Figs. S3 and S4) revealed the presence of two α_1S isoforms in all teleosts from the ancestral branch Osteoglossomorpha (Notopterus notopterus) to the most recently branching Beloniformes (Oryzias latipes; Fig. 5). The marker amino acids threonine and glutamine in the selectivity filter of zf-α_1S-a were identified in α_1S-a of all tested higher-teleost fishes starting with the clade comprising the Cypriniformes. Interestingly, sequence analysis of α_1S-b from numerous teleosts revealed either the lysine at position +4 relative to the selectivity filter in repeat I (Fig. S3) or the aspartate at position +3 relative to the selectivity filter in repeat II (Fig. S4), strictly correlating with their position on the evolutionary tree (Fig. 5). Consequently, higher teleosts exclusively express non–Ca²⁺-conducting Ca_V1.1 channels. However, none of the three marker amino acid positions were found in the ancestrally branching superorders Osteoglossomorpha and Elopomorpha, and thus it is feasible that basal teleosts, like nonteleost fishes and tetrapods, still have Ca²⁺-conducting Ca_V1.1 channels.

Fig. 5.

Non–Ca²⁺-conducting Ca_V1.1 channels are a common trait of all higher teleost fishes. Sequencing of a phylogenetic wide range of teleost fishes revealed that two distinct α_1S isoforms (α_1S-a and α_1S-b) are present in all teleosts. Marker amino acids indicating non-Ca²⁺ conductivity (boxed in black) were identified in α_1S-a and α_1S-b of all higher teleosts starting with the clade comprising the order Cypriniformes. The positions of the selectivity filter glutamates are indicated by gray bars. Phylogenetic tree is drawn after the Tree of Life web project (21).

Discussion

Ca²⁺ Influx–Independent EC Coupling Without Simultaneous Ca_V1.1 Ca²⁺-Conductance Sets Teleosts at a Most-Derived Position Concerning the Evolution of Chordate Skeletal Muscle EC Coupling.

Early chordates such as tunicates or cephalochordates (amphioxus) were shown to activate RyR via influx of extracellular Ca²⁺ (Ca²⁺ influx-dependent EC coupling; see Fig. 6) (22, 23) with rapidly activating Ca²⁺ currents (24), rather resembling the kinetics of cardiac Ca_V1.2 currents. At this phylogenetic stage, Ca_V channels are clustered opposite to RyR but are ultrastructurally arbitrarily arranged (25). This cardiac-like Ca_V arrangement, together with the lack of an intact α_1S II-III loop critical domain (15) (Fig. S5), explains the inability for direct Ca_V-RyR coupling. The functional arrangement of Ca_V1.1 in groups of four (tetrads), a prerequisite for direct Ca_V1.1–RyR1 interaction, can first be observed in hagfish and lamprey (25). In these species, blocking of extracellular Ca²⁺ influx could no more fully obstruct EC coupling (24, 26, 27), even though Ca_V1.1 Ca²⁺ influx was found to still contribute to maximum force generation in EC coupling in lampreys (6). In tetrapods, from frog to human, the tetrads changed their arrangement inside the orthogonal arrays, and skeletal-muscle EC coupling became independent of Ca²⁺ influx (29), although a small and slowly activating Ca_V1.1 Ca²⁺ current can be observed, whose role is largely unknown. In higher-teleost fishes, the Ca_V1.1 Ca²⁺ current completely vanished, in part due to substantial sequence exchanges in or close to the selectivity filter of both Ca_V1.1α_1S isoforms, and consequently, RyR1 activation is solely achieved by direct Ca_V1.1–RyR1 interaction. This clear evolutionary trajectory points to an evolutionary pressure on the reduction of Ca_V1.1 Ca²⁺ currents.

Fig. 6.

Non–Ca²⁺-conducting Ca_V1.1 channels position higher teleost fishes as most derived group concerning the evolution of skeletal muscle EC coupling. In Ca²⁺ influx-dependent EC coupling (22, 23) of early prochordates, such as tunicates or amphioxus, RyR opening is induced via rapidly activating Ca²⁺ currents (Ca²⁺ current) (24), and Ca_V channels are arbitrarily arranged in clusters (Ca_V clusters) opposite RyR (25). Later, hagfish and lampreys developed direct Ca_V1.1-RyR1 coupling (24, 26, 27) with Ca_V1.1 arranged in groups of four (tetrads) (25). In tetrapods, the tetrads changed their disposition within the array, and a slowly activating Ca²⁺ current is still observed, despite that EC coupling is independent of the Ca²⁺ influx. The Ca²⁺ inward current completely vanished in teleost fishes, and EC coupling is solely induced via Ca_V1.1-RyR1 protein–protein interaction. This is true for both distinct Ca_V1.1α_1S isoforms (α_1S-a and α_1S-b) that are present in higher teleosts, concordant with the hypothesis of a whole-genome duplication (WGD) at the basis of the teleost linage (28). Selectivity filter glutamates are boxed in gray, marker amino acids indicating non-Ca²⁺ conductivity in black. Phylogenetic tree is drawn after the Tree of Life web project (21).

Organ Differentiation in Teleost Skeletal Muscle.

We identified two different non–Ca²⁺-conducting isoforms of the pore-forming Ca_V1.1α_1S subunit in teleost fishes. Ca_V1.1 α_1S-a and α_1S-b are differentially expressed in superficial slow and deep fast musculature, respectively, which have repeatedly been shown to be very diverse musculature types regarding their transcriptome and physiology (30–32). We further showed that the expression pattern of the two different α_1S subunits nicely matches the expression pattern of the two RyR1 isoforms of teleost skeletal muscle, suggesting the presence of two distinct musculature-type specific Ca_V1.1-RyR1 couplons. This two-couplon adaptation is an intriguing example of how the whole-genome duplication at the basis of the teleost linage (WGD; see Fig. 6) continues into isoform formation during phylogenetic organ differentiation. Future studies will investigate the putative physiological and biophysical differences of EC coupling in these two different skeletal musculatures.

Ca²⁺ Influx in Mammalian Skeletal Muscle—A Tolerated Remnant or a Physiological Necessity?

Apparently, all aspects of teleost fish skeletal muscle work appropriately without Ca²⁺ influx through Ca_V1.1—a fact that also sheds light on the question about the role of the Ca_V1.1 Ca²⁺ influx into mammalian skeletal muscle. Together with the findings that pharmacological blocking of Ca_V1.1 or elimination of extracellular Ca²⁺ does not inhibit EC coupling (7, 8, 29), it is manifested that the Ca²⁺ influx in tetrapods is largely dispensable and could be seen as a tolerated remnant of the ancestral Ca²⁺influx-dependent skeletal muscle EC coupling. However, Ca²⁺ influx in mammalian skeletal muscle seems to be variably regulated, e.g., by the accessory Ca_V1.1 subunits α₂δ-1 (33) and γ (34) or by retrograde current amplification, induced by RyR1-Ca_V1.1 interaction (35). Ca²⁺ current density is also massively up-regulated in developing skeletal muscle by expression of an alternatively spliced α_1S subunit (36). A possible role of this regulated Ca²⁺ influx might be in Ca²⁺ homeostasis by, e.g., SR store filling. However, experimental attempts to evaluate the importance of this Ca²⁺ influx for muscle function in vivo have not yet been realizable due to the lack of an appropriate system. The presence of non–Ca²⁺-conducting Ca²⁺ channels in teleost fishes now provides evidence that on the organism level, skeletal muscle is completely functional without any Ca_V1.1 Ca²⁺ influx even on the long term. However, it is still feasible that principal differences in the physiology of Ca²⁺ handling in mammalian and teleost skeletal muscle might exist, and that in mammalian muscle the Ca²⁺ influx is indeed indispensable. To address this question, a knock-in mouse model carrying a nonconducting Ca_V1.1 and a transgenic zebrafish strain that expresses a conducting Ca_V1.1 in skeletal muscle will be valuable tools in further studies.

Taken together, this report might serve as the starting point for a diverse range of further studies in the fields of skeletal muscle physiology, biophysics of ion channel pores, and evolutionary biology of skeletal muscle function.

Materials and Methods

Cell Culture.

Myotubes of zebrafish, pike characin, and medaka were isolated and cultured exactly as described in detail (9). The murine α_1S-null line GLT was cultured and transfected as described (19, 37).

Biophysical Analysis.

Myotubes were analyzed for L-type Ca²⁺ currents and gating currents by using the whole-cell patch-clamp technique in combination with intracellular Ca²⁺ recordings by loading the cells with the Ca²⁺-sensitive dye Fluo-4 as described for GLT (38) and zebrafish myotubes (39). Charge carrier was 10 mM Ca²⁺, or in one control experiment, 10 mM Ba²⁺ (Fig. 1D).

Quantitative RT-PCR.

RNA samples of deep fast and superficial slow muscle tissue from adult zebrafish or from sardine were isolated using the RNeasy Kit (Qiagen) and reverse-transcribed using random primers and M-MLV reverse transcriptase (Promega). All qPCR assays were designed with at least one primer spanning an exon–exon boundary. Primer sequences are found in Table S2. To generate absolute standard curves for all targets DNA fragments of 400–500 bp containing the qPCR primer annealing regions were cloned into the pGEM-T Easy vector (Promega) and sequence fidelity checked (Eurofins MWG Operon). PCR fragments generated from the pGEM constructs were quantified using the Quant-iT PicoGreen dsDNA Quantitation Kit (Invitrogen) and used at concentrations of 1 × 10² – 1 × 10⁵ molecules/PCR as templates for the site-specific standard curves. All qPCRs were performed on a Stratagene MX3000P using MESA GREEN qPCR Mastermix Plus for SYBR Assay (Eurogentec). Absolute copy numbers of samples were calculated using MxPro 4.0 software (Stratagene). All expression values are given as copy number of target gene divided by the copy number of the housekeeping gene succinate dehydrogenase complex, subunit A (SDHA) (40).

Molecular Cloning.

zf-α_1S-a and zf-α_1S-b cDNA fragments were PCR generated from reverse-transcribed RNA isolated from zebrafish skeletal muscle. Generation of zf-α_1S constructs and of point mutations into rb-α_1S cDNA, all contained in expression plasmid pGFP37 (7), is described in SI Materials and Methods.

Genomic and cDNA Sequencing.

Tissue samples of various species were collected from alive (tail fin clip) or freshly killed fish. gDNA was isolated immediately or after transportation in 100% alcohol. RNA was isolated and reverse transcribed as described previously for quantitative RT-PCR. α_1S-specific but species-unspecific degenerate primers were used to PCR-generate fragments coding for the α_1S subunit pore regions of homologous repeats I and II carrying marker amino acids for nonconductivity (Table S1).