The mammalian skeletal muscle DHPR has larger Ca2+ conductance and is phylogenetically ancient to the early ray-finned fish sterlet (Acipenser ruthenus).

The L-type Ca2+ channel or dihydropyridine receptor (DHPR) in vertebrate skeletal muscle is responsible for sensing sarcolemmal depolarizations and transducing this signal to the sarcoplasmic Ca2+ release channel RyR1 via conformational coupling to initiate muscle contraction. During this excitation-contraction (EC) coupling process there is a slow Ca2+ current through the mammalian DHPR which is fully missing in euteleost fishes. In contrast to ancestral evolutionary stages where skeletal muscle EC coupling is still depended on Ca2+-induced Ca2+-release (CICR), it is possible that the DHPR Ca2+ conductivity during mammalian (conformational) EC coupling was retained as an evolutionary remnant (vestigiality). Here, we wanted to test the hypothesis that due to the lack of evolutionary pressure in post-CICR species skeletal muscle DHPR Ca2+ conductivity gradually reduced as evolution progressed. Interestingly, we identified that the DHPR of the early ray-finned fish sterlet (Acipenser ruthenus) is phylogenetically positioned above the mammalian rabbit DHPR which retained robust Ca2+ conductivity, but below the euteleost zebrafish DHPR which completely lost Ca2+ conductivity. Remarkably, our results revealed that sterlet DHPR still retained the Ca2+ conductivity but currents are significantly reduced compared to rabbit. This decrease is due to lower DHPR membrane expression similar to zebrafish, as well as due to reduced channel open probability (Po). In both these fish species the lower DHPR expression density is partially compensated by higher efficacy of DHPR-RyR1 coupling. The complete loss of Po in zebrafish and other euteleost species was presumably based on the teleost specific 3rd round of genome duplication (Ts3R). Ts3R headed into the appearance of two skeletal muscle DHPR isoforms which finally, together with the radiation of the euteleost clade, fully lost the Po.

Introduction

Excitation-contraction (EC) coupling links the electrical excitation of the sarcolemma to the contractile activation of the myofilaments. In vertebrate skeletal muscle, this mechanism depends on the close interplay of the sarcolemmal voltage-gated L-type Ca2+ channel or dihydropyridine receptor (DHPR) and the SR Ca2+ release channel or ryanodine receptor type-1 (RyR1). Depolarization of the sarcolemma generates conformational changes in the voltage-sensing DHPRα1S subunit [1] which via inter-channel protein – protein interaction opens the RyR1 to release Ca2+ from SR stores, required for muscle contraction [2,3]. Simultaneous to this conformational DHPR-RyR1 interaction, there is a small and slowly activating inward Ca2+ current through the DHPR, found from lower vertebrates like sharks to higher vertebrates like mammals. The physiological significance of this L-type Ca2+ current which is not immediately required for EC coupling [4], is still a matter of uncertainty even after more than 40 years of continuous research [5-7]. In contrast to the skeletal DHPR, activation of the cardiac DHPR isoform leads to a large and rapidly activating inward Ca2+ current. Subsequently, this DHPR Ca2+ influx triggers opening of the Ca2+-sensitive cardiac ryanodine receptor type-2 (RyR2) to release Ca2+ ions from the intracellular SR Ca2+ stores. This Ca2+-induced Ca2+ release (CICR) mechanism consequently leads to cardiac muscle contraction [8-10].

Very interesting from the phylogenetic view is that the EC coupling mechanism in skeletal muscle fibers of invertebrates and also of early chordates like tunicates or cephalochordates (amphioxus) resembles the cardiac-type EC coupling of vertebrates, since the activation fully depends on external Ca2+ influx through the DHPR [11-18]. In skeletal muscles of hagfish and lamprey which are higher up on the phylogenetic tree, conformational (vertebrate skeletal-muscle-type) DHPR-RyR1 coupling is already established [13,19] but maximum force generation still depends on external Ca2+ influx [20]. On the evolutionary trajectory to the vertebrates EC coupling became fully Ca2+-influx independent (see Fig. 6 in Ref. [18]). Despite loss of this direct role of Ca2+ influx in skeletal muscle EC coupling, robust DHPR currents were recorded from shark muscles [21], which belong to the ancient class “cartilaginous fish” (Chondrichthyes) (Fig. 1). Ca2+ influx through the skeletal muscle DHPR continued to exist at least until the phylogenetic branch of the “four limbed vertebrates” (Tetrapods) which is the most successful and species-rich clade within the ancient group of Sarcopterygii (Fig. 1). Finally, evolution led to a complete abolition of DHPR Ca2+ influx in the class of “ray-finned fishes” (Actinopterygii), more precisely in their highest developed infraclass of “modern bony fishes” (Euteleostei) containing the model organism zebrafish (Danio rerio). Hence, euteleost fish DHPRs fully lost their conductance and solely work as voltage sensors to detect membrane depolarization and as transmitters of this signal to the RyR1 [18]. Therefore, it is very likely that evolutionary replacement of the CICR mechanism in skeletal muscles of early chordates by conformational EC coupling in vertebrates eliminated the evolutionary pressure on the DHPR to conduct Ca2+. Thus, it is feasible that the remaining DHPR Ca2+ influx in tetrapods is just an evolutionary remnant (vestigiality) of the ancient CICR stage.

Fig. 6

The phylogenetic position of the sterlet α1S subunit. (A) Alignment of amino acid sequences forming and framing the DHPR Ca2+ selectivity filter of rb-α1S, st-α1S, zf-α1S-a, and zf-α1S-b. Pore loops I–IV (I, II, III, IV) are aligned and the critical EEEE residues of the Ca2+ selectivity filter are shown in yellow. The blue rods on top indicate the position of the two pore helices, P1 and P2 and the blue line indicates the selectivity filter vestibule (SF) [65]. The critical exchange of the selectivity filter glutamate (E303Q) in pore loop I of zf-α1S-a, as well as the amino acid substitution N636D in pore loop II of zf-α1S-b are indicated in red. Both these substitutions led to DHPR Ca2+ non-conductivity [18]. Amino acid exchanges are indicated as unboxed letters. (B) Phylogenetic tree based on parsimony analysis of full-length amino acid sequence alignments of rb-α1S, st-α1S, zf-α1S-a and zf-α1S-b, as well as cardiac DHPR counterparts including the garfish isoform (gar-α1C; see Results). Ancestral DHPR isoforms from housefly (Musca-α1) and tunicate (Tunicate-α1) were used as outgroups. In both, the DHPRα1S as well as DHPRα1C cluster, rabbit isoforms are the most ancestral. Scale bar indicates evolutionary distance in substitutions per site (s/s). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 1

Vertebrate phylogeny and positions of the three species examined in the study. Phylogenetic relationship between the major lineages of actinopterygian (ray-finned) fishes and sarcopterygian (lobe-finned) fishes, including tetrapods as direct descendants of this clade. A single sister group to the Osteichthyes (bony fishes) are the Chondrichthyes (cartilaginous fishes; including sharks, rays, and chimaeras) [61]. Whole genome duplications (WGDs) are indicated by red asterisks, with the first two rounds (1R and 2R) at the vertebrate stem and the teleost-specific third round (Ts3R) at the basis of the teleost radiation [23,62,63]. Red arrows indicate the phylogenic status of the three species – rabbit, sterlet, and zebrafish examined in the study. Phylogenetic tree is drawn after the Tree of Life web project [64]. Part of animal outlines was modified from [18]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

One possibility to reach the state of DHPR non-conductivity in euteleost fishes could be a gradual diminution in conductance during the course of vertebrate phylogeny, due to random mutations channel regions previously essential for Ca2+ conductance. Alternatively, a sharp transition from fully conducting skeletal muscle DHPRs all the way throughout vertebrate phylogeny until a stage of sudden non-conductivity only on the trajectory to the euteleost branch is likewise possible. However, such an abrupt switch would rather suggest novel physiological requirements for DHPR Ca2+-non-conductivity in euteleost fishes, for so far unknown reasons.

To shed light on the mechanism how the complete loss of DHPR Ca2+ conductivity in skeletal muscle of euteleost fishes arose – by “fading out” or “switching off” – we investigated DHPR current properties in a selected species, which is evolutionary quite away from the Ca2+-conducting tetrapods but somewhere on the trajectory towards the non-Ca2+-conducting euteleost fishes (Fig. 1). Therefore, we cloned and sequenced the channel-pore-forming DHPRα1S subunit of sterlet (Acipenser ruthenus) which is one of the 27 species from the family of sturgeons, close to the basis of Actinopterygii (Fig. 1). Sturgeons, together with the other basal groups, namely bichirs, gars, and bowfins consist of less than 50 extant species and are often considered as “living fossils” since their phylogenetic development has been very slow and their morphology and physiology has remained unchanged since earliest times [22,23]. However, it was unknown if DHPR Ca2+ conductivity already disappeared in those early ray-finned fishes.

In the present study, we focused on the biophysical characterization of the skeletal muscle DHPR Ca2+ conductivity and EC coupling parameters of sterlet in comparison to the mammalian rabbit (Oryctolagus cuniculus) and the euteleost zebrafish (Danio rerio). Sterlet DHPRα1S was cloned and expressed in the DHPRα1S-null murine muscle cell line GLT [24]. Interestingly, sterlet DHPRα1S was found to conduct Ca2+ currents but at lower expression density and open probability (Po) compared to the mammalian control DHPR. However, reduced DHPR expression density combined with enhanced DHPR-RyR1 coupling efficacy in sterlet was identical to zebrafish. Overall, our results suggest that DHPR Ca2+ conductivity “fades out” on the phylogenetic track from the mammals to euteleost fishes. But the final abrupt “turn off” in euteleost fish was presumably based on an additional extraordinary evolutionary event – the 3rd round of teleost specific genome duplication (Ts3R) at the basis of the teleost branch (Fig. 1). Ts3R led to the emergence of two skeletal muscle DHPRα1S isoforms in teleosts which finally turned non-Ca2+ conducting in euteleost fish.

Materials and Methods

Molecular cloning of the sterlet DHPRα1S subunit

Sterlet DHPRα1S subunit (st-α1S) cDNA was created as follows, with nucleotide numbers (nt) given in parentheses and asterisks indicating restriction enzyme (RE) sites introduced by PCR using proofreading Phusion DNA polymerase (Finnzymes reagents). Total RNA from adult sterlet skeletal muscle was isolated using the RNeasy Mini Kit (Qiagen) and reverse transcribed using the Ready-To-Go T-Primed First-Strand Kit (Amersham Pharmacia Biotech). Three cDNA fragments covering the entire open reading frame (ORF) of st-α1S were PCR generated from first-strand cDNA with primers designed according to the zebrafish DHPRα1S-a sequence (GenBank accession no. NM 001146150): SalI*-KpnI (nt −11 to 1119), KpnI-EcoRI (nt 1119–3004), EcoRI-XbaI* (nt 3004–5643). Two subclones were created by co-ligating fragments SalI*-KpnI (nt −11 to 1119) plus KpnI-EcoRI (nt 1119–3004), as well as ligating fragment EcoRI-XbaI* (nt 3004–5643) into the corresponding polylinker RE sites of pBluescript SK+ (pBS) (Stratagene). For the final construct and N-terminal GFP tagging, excised fragments SalI*-EcoRI (nt −11 to 3004) and EcoRI-XbaI* (nt 3004–5643) were in-frame co-ligated into the SalI/XbaI polylinker RE sites of pGFP37 [25]. For an exemplar scheme of the GFP-α1S construct and its expression observed as GFP fluorescence in α1S-null myotubes, see Figs. 1 and 2 of Ref. [25]. After confirming the integrity of the cDNA by sequence analysis (Eurofins MWG Operon, Martinsried, Germany), sterlet DHPRα1S subunit cDNA was deposited in Gen-Bank database (GenBank accession no. HQ636615).

Fig. 2

DHPR Ca2+ currents are smaller in sterlet than in rabbit skeletal muscle. (A) GLT myotubes transfected with st-α1S (red circles; n = 20) expressed substantial DHPR currents, but significantly lower (p < 0.001) than rb-α1S expressing myotubes (blue squares; n = 28). Control recordings with zf-α1S (green diamonds; n = 19) displayed the expected lack of inward Ca2+ currents [18]. (B) Overexpression of st-α1S (n = 6) or rb-α1S (n = 12) in zebrafish (zf) myotubes showed in general, a higher current density but again a significant (p < 0.01) difference of similar dimension like observed in GLT myotubes. Representative current traces are shown on the right and current-voltage (I–V) graphs on the left. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Expression plasmids

Identical to st-α1S, rabbit and zebrafish DHPRα1S subunits (rb-α1S, zf-α1S-a, and zf-α1S-b, respectively) were N-terminally GFP-tagged by cloning the respective cDNAs into expression plasmid pGFP37 as described previously [18,25].

Zebrafish

Wild-type (wt) zebrafish were bred and maintained under standard laboratory conditions according to the established protocols [31], approved by the Tierethik-Beirat of the Medical University Innsbruck and the Bundesministerium für Wissenschaft und Forschung.

Culture and transfection of GLT and primary zebrafish myotubes

Murine DHPRα1S-null (GLT) myotubes were cultured on 35-mm culture dishes at 37 °C and 10% CO2 as described previously [24]. Approximately 4 days after plating, at the onset of myoblast fusion, myotubes were transfected with 1 μg of N-terminally GFP-tagged DHPRα1S cDNA constructs using FuGENE®HD transfection reagent (Promega, Germany). Three to five days after transfection, myotubes were used for whole-cell patch clamp recordings.

Isolation, transfection, and culturing of zebrafish (zf) myotubes were done as described previously [27]. In brief, myoblasts were isolated from 25 to 28 hpf wild-type zebrafish by digesting the tail tissue with collagenase type I (Sigma) and transfected with 2 μg of N-terminally GFP-tagged DHPRα1S cDNA constructs using the AMAXA™ rat neonatal cardiomyocyte nucleofector kit (Lonza, Germany) and plated on 35-mm culture dishes. For whole cell patch clamp recordings myocytes were cultured for 4–6 days at 28 °C.

Whole cell patch clamp recordings

L-type Ca2+ currents in combination with intracellular SR Ca2+ release were recorded from transfected (GFP-positive) GLT and primary zf myotubes by loading the cells with the Ca2+-sensitive dye Fluo-4 as described previously in detail [26,27]. Concisely, all recordings were performed at room temperature with an Axopatch 200 B amplifier controlled by the pClamp software. Patch pipettes were fabricated from borosilicate glass (Harvard Instruments) and had resistances of 1.5–2 MΩ for GLT and 3.5–5 MΩ for zf myotubes when filled with the internal solution. Spontaneous contractions of zebrafish myotubes were blocked by adding 100 μM of the myosin-II blocker N-benzyl-p-toluene sulfonamide (Sigma) to the bath solution. 10 mM Ca2+ in the bath solution served as the charge carrier and DHPR Ca2+ currents were elicited by 200-ms depolarizing test pulses from a holding potential of −80 mV to test potentials from −50 and +80 mV in 10-mV increments and filtered at 1 kHz and digitized at 5 kHz. To correct for leak currents a P/4 prepulse protocol was employed and to inactivate Na+ and T-type Ca2+ channels a 1-s prepulse to −30 mV was administered before every test pulse. Linear cell capacitance was used to normalize current amplitudes (pA/pF). Following Boltzmann expression was used to fit the current-voltage (I–V) curves: where I is the current; V is the test potential; V is the reversal potential; V is the half maximal activation potential; G is the maximum channel conductance and k is the slope factor.

In parallel to current recordings, the average fluorescence intensity via Fluo-4 fluorescence was measured, normalized to the resting fluorescence and expressed as ΔF/F0. Following Boltzmann distribution was used to fit the voltage dependence of ΔF/F0: where V1/2 is the potential at which ΔF/F = (ΔF/F)max/2 and k is the slope factor.

Recordings were analyzed using ClampFit 9.0 and 10.0 (Axon Instruments) and SigmaPlot 10.0 (SPSS Science, Chicago, IL) software.

Statistics

Data are reported as mean ± SE. Statistical significance was calculated using unpaired Student’s t test and p values considered statistically significant are as follows: *, p < 0.05; **, p < 0.01; and ***, p < 0.001.

Results

Myotubes of the immortalized murine muscle cell line GLT, lacking DHPRα1S [24], were transfected with expression plasmids encoding either GFP-tagged sterlet DHPRα1S (st-α1S), or GFP-tagged rabbit DHPRα1S (rb-α1S) [25], or a stoichiometric mix of GFP-tagged DHPRα1S-a (from superficial, slow, red muscle) and DHPRα1S-b (from deep, fast, white muscle) of zebrafish (zf-α1S) as controls [18].

DHPR Ca2+ influx into sterlet skeletal muscle is considerably lower compared to mammals

As shown in Fig. 2A, whole cell patch-clamp analysis of GLT myotubes transfected with st-α1S revealed significantly smaller (p < 0.001) L-type Ca2+ currents (0.6 ± 0.1 pA/pF) compared to rb-α1S transfected cells (2.1 ± 0.2 pA/pF). As expected, zf-α1S-expressing GLT myotubes showed no inward Ca2+ currents [18]. Despite the difference in I between st-α1S and rb-α1S, no significant difference (p > 0.05) in half-maximal current activation potential (V1/2) was observed (34.4 ± 2.2 mV and 37.0 ± 1.1 mV for st-α1S and rb-α1S, respectively). Lower membrane expression of st-α1S in the mammalian expression system (GLT) accounting for reduced current density might be due to partial DHPR subunit incompatibilities. An example in this context is the crucial role of the DHPRβ1a subunit in membrane trafficking of DHPRα1S and in the conformational interaction of DHPR and RyR1via tetrad formation [28-31]. Consequently, a putatively weaker interaction between the heterologously expressed fish α1S subunits and the endogenous murine β1a of the GLT myotubes might be responsible for the lower membrane expression of the fish DHPR. To test this, we overexpressed st-α1S as well as the control rb-α1S in primary zebrafish myotubes which completely lack DHPR Ca2+ influx (18) and thus serve well as a heterologous expression system for studying Ca2+-conducting DHPR α1S subunits.

However, our results exclude fish/mammalian species-related DHPR subunit incompatibilities, because I of rb-α1S (8.6 ± 1.3 pA/pF) was again ∼3-fold larger compared (p < 0.01) to st-α1S (2.9 ± 0.7 pA/pF) (Fig. 2B). Interestingly, however the current size in freshly dissociated zebrafish muscle cells is more than 4-fold higher for both st-α1S and rb-α1S compared (p < 0.001) to that obtained from the GLT cell line.

To test whether ∼3.5-fold smaller Ca2+ currents through st-α1S compared to rb-α1S are due to differences in membrane expression densities rather than biophysical channel properties, we analyzed intramembrane charge movement (Q) at reversal potentials (V) upon st-α1S and rb-α1S expression in GLT myotubes (Fig. 3A). At V (68.7 ± 0.2 mV for st-α1S and 79.4 ± 1.5 for rb-α1S), maximum of DHPRs are activated with the key advantage that the Q signal is not contaminated by inward Ca2+ currents. Thus, Q recordings provide a convenient measure for the number of functional channels in the membrane [32,33]. As depicted in Fig. 3A, GLT myotubes expressing st-α1S displayed significantly lower (p < 0.001) Q (2.8 ± 0.5 nC/μF) compared to rb-α1S-expressing myotubes (8.5 ± 0.8 nC/μF), indicating a substantial difference in membrane expression of these two DHPRs. Interestingly, reduction in membrane expression of st-α1S is identical (p > 0.05) to that of zf-α1S (2.5 ± 0.2 nC/μF) – the second ray-finned-fish DHPR used in this study, suggesting a fish-specific reduction in DHPR expression density (Fig. 3A).

Fig. 3

Reduced DHPR membrane expression and channel open probability in sterlet skeletal muscle. (A) Charge movement recordings at reversal potential (Q) from GLT myotubes expressing rb-α1S (blue bar; n = 24), st-α1S (red bar; n = 17), or zf-α1S (green bar; n = 21) indicated significantly lower (p < 0.001) membrane expression of both fish DHPRs. Representative charge movement traces are depicted on the right. Q (indicated by the left arrow) is evaluated by integrating the area under the peak, induced by applying a 200-ms test potential at the reversal potential of +70 or +80 mV. (B) Relative DHPR open probability (P) was calculated by plotting Q versus [Q− Q] and comparing the slopes of the regression lines. Q (indicated by the right arrow in A, right panel) is calculated by integrating the area under the tail current occurring at the end of the 200-ms test potential. P of st-α1S (red circles) is significantly smaller (p < 0.01) than of rb-α1S (blue squares). As expected, expression of zf-α1S in GLT myotubes (green diamonds) displayed no conductance and thus the slope of the regression line was around zero, significantly different (p < 0.001) to sterlet and rabbit DHPRs. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

In addition to the observed reduction in membrane expression, distinct biophysical channel properties might also contribute to smaller Ca2+ currents through sterlet DHPR. To test for differences in DHPR channel open probability (P), we analyzed the size of the integrated tail currents (Q) that emerge after the end of the 200-ms test pulses at V upon expression of st-α1S, rb-α1S, or zf-α1S in GLT cells (Fig. 3A; right panel). On plotting Q against [Q − Q] the slope of the linear fit provides a convenient index of relative P [32,33]. Interestingly, we found a ∼3.3-fold smaller slope (y = 0.77 x) and thus lower P for st-α1S compared to (p < 0.01) rb-α1S (y = 2.5 x; Fig. 3B). As expected, the slope factor of the non-Ca2+-conducting zf-α1S control was around zero. Thus both parameters, reduced membrane expression as well as a lower channel open probability of st-α1S contribute to the reduced DHPR inward Ca2+ current in sterlet skeletal muscle.

Accelerated kinetics of the DHPR Ca2+ current in sterlet skeletal muscle

A hallmark of the mammalian skeletal muscle DHPR L-type Ca2+ current is its slow activation and inactivation kinetics [34,35]. Interestingly, GLT myotubes expressing st-α1S displayed DHPR currents with significantly (p < 0.001) accelerated activation kinetics compared to rb-α1S. As shown in Fig. 4A, the average time to peak measured from I traces at +40 or +50 mV was 35.1 ± 7.1 ms for st-α1S and 87.2 ± 7.3 ms for rb-α1S. Likewise, DHPR current inactivation, analyzed as fractional inactivation 80 ms after the current peak (R80), was slightly but significantly (p < 0.01) enhanced in st-α1S (76.5 ± 3.9%) compared to rb-α1S (88.2 ± 1.7%) (Fig. 4A, right panel). Previously, DHPR transmembrane segment IS3 together with the extracellular linker between transmembrane segments IS3 and IS4 have been identified to be responsible for slow or fast current activation [36,37]. Sequence alignments of this defined region revealed 19 amino acid exchanges between the slow activating rb-α1S and the fast activating st-α1S (Fig. 4B). But, most interestingly, in this region 8 amino acids were identical between st-α1S and the fast cardiac DHPRα1C isoforms, regardless if it was from rabbit (rb-α1C) or garfish (Lepisosteus oculatus; gar-α1C). Gars are another ancient order of Actinopterygii, phylogenetically close to the sturgeons (see Fig. 1). We used gar-α1C as a cardiac reference sequence because sterlet α1C sequence is not available so far. However, sterlet cardiac Ca2+ currents have been characterized as fast activating and inactivating [38] similar to mammalian cardiac L-type currents [10,36]. Altogether, it appears that the identified amino acid exchanges in the IS3 and IS3-IS4 linker region of st-α1S are responsible for its α1C-type accelerated current activation kinetics.

Fig. 4

DHPR Ca2+ currents activate and inactivate faster in sterlet than in rabbit skeletal muscle. (A) (Left panel): DHPR Ca2+ currents peak significantly faster (p < 0.001) in st-α1S (red bar; n = 20) compared to rb-α1S (blue bar; n = 30) expressing GLT myotubes. (Middle panel): Representative traces of maximum DHPR current (I) at +40 mV test potential. The red arrowhead indicates the peak of sterlet I and the blue arrowhead of rabbit I. (Right panel): Analyses of fractional inactivation, 80 ms after the current peak (R80) indicates acceleration of inactivation of sterlet DHPR currents compared to (p < 0.01) rabbit. (B) Amino acid sequence alignment of transmembrane segment IS3 and the adjacent IS3-IS4 linker (upper panel) known to be responsible for determining slow or fast activation of the DHPR current [36,37] revealed 19 amino acid exchanges (unboxed letters) between the slow activating rb-α1S and the fast activating st-α1S. Amino acid sequence alignment of fast activating cardiac DHPRs of rabbit and garfish (lower panel), revealed 8 amino acids to be identical to corresponding st-α1S positions (red arrows). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Reduced SR Ca2+ release but enhanced DHPR – RyR1 coupling efficacy in sterlet skeletal muscle

The conformational change during DHPR activation, measured as charge movement, Q not only opens the DHPR to allow Ca2+ influx (except for the teleost zf-α1S), but also leads to the activation of the SR Ca2+ release channel, RyR1 via conformational DHPR – RyR1 coupling [3]. Robust intracellular SR Ca2+ release in response to 200-ms test pulses from −50 to +70 mV was observed with all the three constructs (Fig. 5A, right panel). However, maximal SR Ca2+ release (ΔF/F0) values were significantly lower in GLT myotubes expressing either st-α1S (1.1 ± 0.1; p < 0.01) or zf-α1S (1.1 ± 0.1; p < 0.001) when compared to rb-α1S (1.8 ± 0.2) (Fig. 5A and B, left graphs). The half-maximum activation potential of Ca2+ release (V1/2) was identical (p > 0.05) for all the three DHPRs (st-α1S, −0.70 ± 1.6 mV; zf-α1S, 0.90 ± 0.8 mV; and rb-α1S, −0.34 ± 1.2 mV).

Fig. 5

Reduced SR Ca2+ release but higher DHPR-RyR1 coupling efficacy in sterlet and zebrafish DHPRs. (A) GLT myotubes transfected with st-α1S (red bar; n = 23) or zf-α1S (green bar; n = 52) displayed significantly smaller maximal SR Ca2+ release ((ΔF/F0)) compared to (p < 0.01 and p < 0.001, respectively) rb-α1S-expressing myotubes (blue bar; n = 39). Representative SR Ca2+ release traces are depicted on the right. (B) Voltage dependence of SR Ca2+ release shows identical (p > 0.05) half-maximum activation for all the three DHPRs. Comparison of the absolute voltage-dependence of Ca2+ release (left graphs) to the Ca2+ release normalized to DHPR expression levels (measured as Q; right graphs) indicated significantly higher (p < 0.001) DHPR-RyR1 coupling efficacy in sterlet and zebrafish skeletal muscle compared to rabbit. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Next we were interested, if the significant reduction in DHPR charge movement of st-α1S and zf-α1S compared to rb-α1S (Fig. 3) is quantitatively reflected in the observed drop in intracellular SR Ca2+ release. To test this, we normalized the ΔF/F0 values (Fig. 5B, left graphs) to the individual Q values (see Fig. 3A). Conversely, normalized graphs (Fig. 5B, right graphs) indicated that the resulting maximal (ΔF/F0/Q) values were significantly higher (p < 0.001) in GLT myotubes expressing st-α1S or zf-α1S (0.42 ± 0.04 and 0.38 ± 0.02, respectively) compared to rb-α1S (0.17 ± 0.02). Taken together, these results indicate that DHPR – RyR1 couplons of sterlet and zebrafish skeletal muscles have more than 2-fold enhanced efficacy of SR Ca2+ release compared to rabbit.

Does the “living fossil” sterlet retain a phylogenetically ancient DHPR in skeletal muscle?

Despite surprisingly close similarities between the skeletal muscle DHPRs of the ancestral sterlet and modern teleost zebrafish regarding functional sarcolemmal expression levels and efficient coupling to the SR release channel, there is one fundamental difference between the DHPRs of these two ray-finned fish: Sturgeon DHPR conducts inward Ca2+ currents whereas zebrafish DHPR is non-conducting (see Figs. 2 and 3). Previously, we identified crucial single amino acid substitutions (in reference to rb-α1S) responsible for the loss of this Ca2+ influx in both zebrafish skeletal muscle DHPR isoforms. In zf-α1S-a the selectivity filter glutamate in repeat I was substituted by glutamine (E303Q) and in zf-α1S-b the neutral asparagine in repeat II, three residues C-terminal to the selectivity filter glutamate, was exchanged to a negatively charged aspartate (N636D) [18]. As displayed in Fig. 6A, in st-α1S, sequences adjacent to the critical selectivity filter glutamates of repeats I–IV are fully homologous to rb-α1S despite the large evolutionary distance between these two different classes – Actinopterygii and Sarcopterygii (see Fig. 1). Hence, any sequence exchange, including the essential substitutions in repeats I and II responsible for Ca2+ non-conductivity in teleost fishes, must have evolved within the Actinopterygii, but after the separation of the sturgeon branch.

Even though the central st-α1S selectivity filter region embracing repeats I–IV is 100% homologous to rb-α1S, other areas might be quite heterologous viz., the IS3, IS3-IS4 loop region (see Fig. 4B) and rather share higher homology to zf-α1S, if considering the surprising similarities in DHPR surface expression level and EC coupling efficiency between st-α1S and zf-α1S (see above). Hence, we were interested in the phylogenetic position of st-α1S in comparison to our other two candidate DHPRs, rb-α1S and zf-α1S. So the question arose if the DHPR of the “living fossil” sterlet [22,23] is far the oldest Ca2+ channel in this ensemble? If so, are then the similarities in channel properties with the evolutionary most derived euteleost zebrafish rather an analogy than the manifestation of evolutionary closeness? To this aim, we calculated a phylogenetic tree based on parsimony analysis of full-length amino acid sequence alignments of rb-α1S, st-α1S, and zf-α1S together with the cardiac DHPR counterparts (Fig. 6B). Ancestral DHPR isoforms from housefly Musca domestica [39] and the tunicate Halocynthia roretzi [40] were used as outgroups. Strikingly, the mammalian rb-α1S emerges as the ancestral skeletal muscle DHPR, with st-α1S in-between and the two zf-α1S paralogs as the phylogenetically most derived. We see a similar picture in the clade comprising the cardiac DHPRs. The mammalian isoform, rb-α1C is basic to the cardiac DHPRs of both actinopterygian species.

Discussion

To shed light on a fascinating phenomenon of muscle physiology – the loss of skeletal muscle DHPR Ca2+ influx during vertebrate evolution, we performed sequence comparisons, phylogenetic tree analysis, and patch-clamp electrophysiology on three phylogenetically distant skeletal muscle DHPRs in this study.

Phylogenetically most derived DHPRs in skeletal muscles of the euteleost zebrafish

Previously, we detected that zebrafish has not only two isoforms of the DHPRα1S subunit, but most interestingly that both paralogs implemented Ca2+-non-conductivity by different point mutations leading to distinct molecular strategies of Ca2+ blockade [18]. Additionally, we showed that all tested euteleost DHPRα1S isoforms carried amino acid substitutions that prevented Ca2+ influx in skeletal muscle. Remarkably, on the basis of the teleost branch in the order Osteoglossiformes (see Fig. 1), DHPRα1S already exist in two paralogs, but the amino acid exchanges essential for blocking the Ca2+ influx did not take place yet [18].

Teleost fishes are by far the highest derived clade in the class of Actinopterygii and the most diversified and dominant group among vertebrates with over 30,000 species, which is approximately equal to all other vertebrate groups combined [41]. A vast majority of them (>95%) belong to the highest developed Euteleostei, which are certainly the most successful vertebrate group, probably due to their rich gene supply as a result of the 3rd teleost specific genome duplication (Ts3R, see Fig. 1) in this lineage [23]. Interestingly, alone this fact means that half of the vertebrate species possess non-Ca2+-conducting skeletal muscle DHPRs which exemplifies that “silencing” of skeletal muscle DHPRs is in no case a marginal phenomenon. Hence, in this study we choose as the first of three model organisms, the euteleost zebrafish (Danio rerio) as the canonical vertebrate species with non-Ca2+ conducting skeletal muscle DHPR (see Fig. 1).

The skeletal muscle DHPR of sterlet is Ca2+ conducting

Sterlet (Acipenser ruthenus) from the family of sturgeons, with unknown DHPR Ca2+ -conductivity status and with its position at the basis of the Actinopterygii phylogenetically sufficiently distant to the euteleost zebrafish, was selected as the second model organism. Sturgeons belong to an ancient clade which separated from the other ray-finned fishes around 360 million years ago and thus are phylogenetically located far below the teleosts [18,42-45]. As mentioned above, sturgeons together with the other basal actinopterygian (bichirs, gars, and bowfins) are often considered as “living fossils” with very slow phylogenetic development and apparently stagnating morphological and physiological advances [22,23]. However, due to the lack of biophysical studies on skeletal muscle DHPRs of these early ray-finned fishes, it was unknown if DHPR Ca2+ conductivity was existing or rather lost due to the lack of evolutionary pressure since the stages of earliest vertebrates (see Introduction). Finally, our present study provides clear evidence that sturgeons (and consequently also the clade phylogenetically beneath – the bichirs) possess Ca2+-conducting DHPRs in skeletal muscle.

The mammalian DHPR is evolutionary ancient to sterlet and zebrafish

The third model organism in this study is the mammalian species rabbit (Oryctolagus cuniculus), which is archetypical for a Ca2+-conducting skeletal muscle DHPR [46]. Rabbit DHPRα1S shares similar amino acid sequence identity to the representatives (n = 11) of different mammalian families, ranging from 89% (African elephant), 93% (human, mouse, rat, cat, and horse) to 94% (walrus). Mammals are a clade within the tetrapods which evolved around 390 million years ago [47] as the most advanced branch from the evolutionary less successful (because most species are extinct) second class of “bony fishes” (Osteichtyes), the so called “lobe-finned fishes” (Sarcopterygii) (Fig. 1). Since skeletal muscle DHPRs of amphibians [48,49], birds [50], and mammals [46,51] were characterized to conduct Ca2+, we can unequivocally conclude that also their ancestors, and thus the entire clade of Sarcopterygii, embraces species with Ca2+-conducting DHPRs.

Surprisingly, our phylogenetic and biophysical analyses revealed that the mammalian rabbit not only possesses DHPRs (skeletal muscle as well as cardiac) that are evolutionary ancient to both selected actinopterygian DHPRs (Fig. 6), but also that only the rabbit skeletal muscle DHPR produces a robust inward Ca2+ current. Thus, skeletal and cardiac muscle DHPRs of the so called “living fossil” sterlet are phylogenetically at a more derived position compared to corresponding rabbit isoforms. DHPR Ca2+ influx in sterlet is significantly reduced compared to rabbit (see Fig. 2), which in addition underlines the more advanced evolutionary stage and accordingly supports the hypothesis of a gradual disappearance of skeletal muscle DHPR Ca2+ conductance during vertebrate evolution. Interestingly, the reduction of Ca2+ conductance of sterlet DHPR is based on two mechanisms: (i) lower membrane expression and (ii) lower open probability (Po) of the channel. Lower DHPR surface expression in sterlet already anticipates the reduced DHPR expression found in the euteleost zebrafish. But in contrast to sterlet, Po of the zebrafish DHPR already reached the zero level (see Fig. 3). However, it is likewise possible that single channel conductance (γ) or both single channel parameters, Po and γ, are reduced in sterlet. Remarkably, this drop in DHPR surface expression is not 1:1 reflected in lower intracellular Ca2+ release, as might be expected from a lower number of actively interacting DHPR-RyR1 couplons. Our results reveal that both, sterlet DHPR as well as the most derived zebrafish DHPR possess higher DHPR-RyR coupling efficacy which partially compensates for the lower DHPR membrane expression.

Another interesting aspect of this study is that the current through the sterlet DHPR does not follow the archetypical skeletal muscle L-type current kinetics. Instead of activating and inactivating in a slow mode, it is significantly faster in both aspects to all the so far described skeletal muscle DHPRs [46,48-51]. Consequently, whereas the reduced size of the sterlet DHPR current points to a more advanced evolutionary state (supported by the pedigree in Fig. 6), the fast current kinetics resembles rather the phylogenetically ancient cardiac isoform. Additionally, the 8 amino acids of cardiac sequence retained in transmembrane segment IS3 and the adjacent IS3-IS4 linker, considered to be responsible for the fast activation kinetics, seems to present an interesting atavism.

The loss of DHPR Ca2+ conductance in Euteleostei – fading out or switching off?

The answer to our primary question, if the complete loss of DHPR Ca2+ conductance in skeletal muscle of euteleost fishes emerged by slowly “fading out” due to the lack of evolutionary pressure or rather was “switched off” due to e.g., new muscle-physiological concepts (see above), appears to be bilateral. Still existent but considerably lower Ca2+ influx in sterlet, a species that is phylogenetically (see Fig. 1) and also in terms of DHPR evolution (see Fig. 6) located between rabbit and zebrafish, strongly suggests a steady loss of Ca2+ conductance during vertebrate evolution. However, the final abrupt “switch off” needed an extraordinarily evolutionary event, the Ts3R, which occurred at the base of the teleost radiation [23,52] (Fig. 1). The significance of whole-genome duplications (WGDs) for shaping the architecture of vertebrate genomes, yielding vast biological diversity and hence evolutionary success, are widely accepted and increasing evidences for two rounds of WGDs (1R and 2R) at the basis of vertebrate radiation as well as a 3rd WGD (Ts3R) at the basis of the teleost clade (see Fig. 1) emerged in recent years [23,52-57]. Numerous crucial progresses in muscle evolution are known to originate from 1R- and 2R-duplications. One was the differentiation of muscle-specific myosin II into cardiac and skeletal muscle isoforms [13,55,58]. Similarly, the diversification of the Ca2+ release channel into RyR1 and RyR2 [59] and of DHPRα1 into DHPRα1S and DHPRα1C occurred, allowing the acquisition of direct DHPRα1S-RyR1 interaction in skeletal muscle in parallel to the ancient DHPRα1C-RyR2 CICR mechanism still persisting in cardiac muscle [16]. DHPRα1S isoforms evolved to a considerably higher degree compared to DHPRα1C (Fig. 6), pointing to a tighter selection pressure on DHPRα1S due to the requirement to develop efficient conformational EC coupling. Finally, Ts3R led to splitting of DHPRα1S into DHPRα1S-a and DHPRα1S-b isoforms in teleosts, specific for slow and fast skeletal muscles, respectively [18]. Amino acid substitutions induced by this new round of DHPRα1S isoform formation, together with the silencing of both DHPRs led to by far the most derived DHPRα1S isoforms on the phylogenic tree, with the fast muscle DHPRα1S-b at the topmost position (Fig. 6).

One possible physiological conception for this obvious final “enforcement” of DHPR non-Ca2+-conductivity following DHPRα1S duplication during Ts3R might involve adaptations in the cellular energy management. Thus, the block of Ca2+ influx during EC coupling abolishes the need for Ca2+ extrusion from the myoplasm, allowing a pure Ca2+ recycling pathway between the myoplasm and internal SR Ca2+ stores. This ATP-saving concept might be of special relevance in the poikilothermic teleosts but would play only a minor role in mammals where a constant basal metabolic activity of skeletal muscles is nonetheless required for heat production for maintenance of a stable body temperature. To test this hypothesis swim tunnel respirometry experiments [60] on a transgenic zebrafish strain expressing a Ca2+ conducting DHPR in skeletal muscle are on the way.

Conclusion

Biophysical and phylogenetic analyses of the skeletal muscle DHPR of the early ray-finned fish species sterlet (Acipenser ruthenus) support the hypothesis that DHPR Ca2+ conductivity “fades out” during vertebrate evolution due to the lack of evolutionary pressure on this presumably vestigial current. This conclusion is drawn from the significantly lower open probability (Po) of the phylogenetically derived sterlet DHPR when compared to the relatively ancestral mammalian rabbit DHPR. However, the final “turn off” of DHPR Ca2+-conductivity did not take place before the stage of euteleost fishes. Radiation of this clade was preceded by the teleost-specific 3rd round of whole genome duplication, Ts3R. Consequently, different amino acid substitutions in the duplicated DHPRα1S isoforms enforced the complete block of DHPR Ca2+-conductivity in euteleost fishes.