Conservation of cardiac L-type Ca²⁺ channels and their regulation in Drosophila: a novel genetically-pliable channelopathic model

Abstract

Dysregulation of L-type Ca²⁺ channels (LTCCs) underlies numerous cardiac pathologies. Understanding their modulation with high fidelity relies on investigating LTCCs in their native environment with intact interacting proteins. Such studies benefit from genetic manipulation of endogenous channels in cardiomyocytes, which often proves cumbersome in mammalian models. Drosophila melanogaster, however, offers a potentially efficient alternative as it possesses a relatively simple heart, is genetically pliable, and expresses well-conserved genes. Fluorescence in situ hybridization confirmed an abundance of Ca-α1D and Ca-α1T mRNA in fly myocardium, which encode subunits that specify hetero-oligomeric channels homologous to mammalian LTCCs and T-type Ca²⁺ channels, respectively. Cardiac-specific knockdown of Ca-α1D via interfering RNA abolished cardiac contraction, suggesting Ca-α1D represents the primary functioning Ca²⁺ channel in Drosophila hearts. Moreover, we successfully isolated viable single cardiomyocytes and recorded Ca²⁺ currents via patch clamping, a feat never before accomplished with the fly model. The profile of Ca²⁺ currents recorded in individual cells when Ca²⁺ channels were hypomorphic, absent, or under selective LTCC blockage by nifedipine, additionally confirmed the predominance of Ca-α1D current across all activation voltages. T-type current, activated at more negative voltages, was also detected. Lastly, Ca-α1D channels displayed Ca²⁺-dependent inactivation, a critical negative feedback mechanism of LTCCs, and the current through them was augmented by forskolin, an activator of the protein kinase A pathway. In sum, the Drosophila heart possesses a conserved compendium of Ca²⁺ channels, suggesting that the fly may serve as a robust and effective platform for studying cardiac channelopathies.

Introduction

Cardiac action potentials, which drive rhythmic contractions of the heart, are the result of well-choreographed opening and closing of multiple ion channels. Among these are the L-type Ca²⁺ channels (LTCCs). LTCCs not only permit Ca²⁺ entry to initiate myocyte shortening, but also set the length of the plateau phase of the cardiac action potential and, thereby, determine its duration [1]. Disruption of these critical channels or their precise regulation underlies numerous pathologies. For example, mutations in Ca_V1.2, a major LTCC in heart muscle, lead to Timothy syndrome, a multi-system disorder featuring autism, polydactyly, and long-QT syndrome [2-5], while mutations in calmodulin, a regulator of Ca_V1.2 Ca²⁺-dependent inactivation (CDI), result in a malignant form of long-QT syndrome [6-9]. Mutations in Ca_V1.3, an LTCC in nodal cells, can lead to bradycardia and congenital deafness [10, 11]. Additionally, changes in channel density and function are associated with atrial fibrillation [12, 13] and heart failure [14, 15].

Mechanistic dissection of alterations in LTCC regulation is typically conducted using recombinant channels expressed in a heterologous system. Unfortunately, these systems often lack key auxiliary elements that are readily available in the context of cardiac myocytes. Studying Ca²⁺channel regulation, including CDI and protein kinase A (PKA)-mediated current augmentation in native mammalian myocytes is, however, also challenging due to cellular complexity, including redundancy among genes, and the difficulty of endogenously manipulating the channels. Thus, an intermediate system, which is both genetically pliable and reflective of a cardiac muscle cell, is desirable.

Cardiomyocytes from Drosophila melanogaster, the fruit fly, offer an attractive, yet incompletely explored alternative. Drosophila benefit from a completely sequenced genome, conservation of disease orthologs [16-19], and a host of genetic tools [16, 19-21]. For example, the bipartite Gal4-UAS expression system [22] permits straightforward control of gene expression across space (by means of tissue-specific promoters) and time (via drug- or temperature-inducible expression) [23]. This system readily supports investigation of gain- or loss-of-function of genes of interest, owing to large resources and libraries including thousands of independent GAL4 and UAS Drosophila lines that permit selective transgene overexpression or RNAi (interfering RNA)-mediated gene silencing [24, 25]. Importantly, mutant genes can easily be studied in the same genetic environment as their wild-type counterparts, thus minimizing confounding effects, such as those that may result from insertion of transgenes in different locations throughout the genome [26-28]. Many Drosophila proteins are encoded by a single gene that generates multiple distinct isoforms through alternative splicing of the primary transcript, simplifying knockout and/or gene suppression experiments [21]. Severe manipulation of heart components can also be tolerated because oxygen transport occurs through trachea that invaginate from the cuticle into the interior of the fly [29]. The effects of potentially lethal cardiac mutations can therefore be studied in vivo without necessarily initiating death. Furthermore, the recent development of techniques to image [30-35] and record electrical field potentials [34, 36] of both larval [32-35] and adult [30-33, 36] hearts allows functional assessment of different developmental stages of Drosophila myocardium after manipulation of particular genes. Finally, flies are economical and easy to breed, generate numerous offspring, and feature a short life cycle, making large-scale genetic and small-compound screens, in wild-type and disease models, eminently feasible.

Drosophila have an open circulatory system with a dorsal vessel or “cardiac tube” (Figure 1A), which in many ways functionally and developmentally resembles the embryonic vertebrate heart [21, 37, 38]. The tube is composed of a single layer of two opposing rows of cardiac myocytes [35, 39] whose action potential, as found in higher organisms, is myogenic in origin, i.e. action potentials originate from muscle itself as opposed to in response to neuronal impulses (neurogenic origin) [40-42]. The activity of these myocytes is modulated by a neurohormonal system that features the epinephrine-like compound, octopamine [43]. Action potentials of Drosophila heart tubes have been coarsely measured [34, 36, 44, 45], and the voltage waveforms appear comparable in duration and amplitude to those recorded in vertebrate myocytes. As LTCCs play a major role in shaping vertebrate cardiac action potentials, it is plausible that a Drosophila analog is playing a similar role in the heart tube. Like vertebrates, flies express hetero-oligomeric voltage gated Ca²⁺ channels (Ca_V) composed of α₁, β, α₂δ, and, in some tissues, ϒ subunits [17, 46, 47]. Three distinct genes, Ca-α1D (CG4894), cacophony (CG43368), and Ca-α1T (CG15899) encode the α₁-subunits A1D, cac, and T-type, which specify three Ca_V that correspond to Ca_V1, Ca_V2, and Ca_V3, the main Ca²⁺ channel families in vertebrates (Supplementary Table 1). Drosophila A1D is similar to the dihydropyridine-sensitive (L-type) channels of vertebrates; cac to N-, P/Q-, R-type channels; and T-type to Ca_V3 channels [17]. RNA microarray data demonstrate enrichment of Ca-α1D in the hearts of adult fruit flies compared to the whole body, hinting at the potential presence of A1D in Drosophila cardiac tubes [16, 48]. Still, the complete molecular biosignature of heart-tube Ca²⁺ channels is unknown.

Figure 1. Ca-α1D and Ca-α1T transcripts are abundant in Drosophila heart tubes.

A) Confocal micrograph of a semi-intact wild-type w¹¹¹⁸ Drosophila heart tube extending along the dorsal side of the abdomen. The anterior conical chamber is outlined and displayed in (Figure 1B). Note the non-cardiac alary muscles (AM) and the retractors of tergite muscles (RT). Scale bar = 100 μm. B) Micrograph of the conical chamber after Ca-α1D (red) and GAPDH (white) mRNA molecules were labeled with FISH probes. A representative small region of a single cardiomyocyte used for mRNA quantitation is outlined in white. PC, pericardial cell. Scale bar = 25 μm. C) Examples of Ca_v α₁-subunit mRNA particle densities in cardiomyocyte areas of interest (e.g. white box in Figure 1B) and the quantitative determination of the number of subunit messages normalized to the number of GAPDH messages within the same regions of interest. There were significantly more Ca-α1D and Ca-α1T transcripts compared to cac transcripts. *** p < 0.001. (n = 11, 10, 11 animals for Ca-α1D, cac, and Ca-α1T).

Here, we have elucidated the identity of the Ca²⁺ channels in the adult Drosophila heart. Direct visualization and quantitation of mRNA revealed an abundance of Ca-α1D and Ca-α1T and a paucity of cacophony Ca²⁺ channel messages in the cardiac tube. Suppression of Ca-α1D effectively eliminated contraction and Ca²⁺ cycling activity, while knocking down other channels minimally disrupted contraction and Ca²⁺ activity. Although these two lines of evidence suggest the presence of A1D and T-type channels, definitive validation required direct measurement of Ca²⁺ currents across the sarcolemma of Drosophila cardiomyocytes. To this end, we devised a method for isolating viable single cardiomyocytes from heart tubes. These myocytes enabled whole-cell patch clamp recordings of Ca²⁺ currents. Utilizing pharmacological agents and existing Drosophila Ca²⁺ channel mutants, we verified the presence of both A1D and T-type Ca²⁺ currents in fly cardiomyocytes with A1D being the major contributor of the Ca²⁺ influx for contraction. Similar to mammalian LTCCs, Drosophila A1D also exhibits CDI, a critical negative feedback process that helps with channel regulation, and Ca²⁺ current amplification through PKA signaling. Overall, we resolved the ensemble of Ca²⁺ channels functioning in the adult Drosophila heart and devised a novel technique for isolating single cardiomyocytes, which highlight the fly as a feasible alternative for studying diseases involving misregulation of cardiac LTCCs.

Materials and Methods

Drosophila strains and maintenance

TinCΔ4-Gal4 [49], TinCΔ4-Gal4;UAS-GCaMP3 (a generous gift from Dr. Rolf Bodmer), and Hand^4.2-GS-Gal4 [50] Drosophila were employed to drive UAS-transgene expression in a cardiac-restricted fashion. UAS-RNAi stocks, obtained from the Vienna Drosophila Resource Center (VDRC), included UAS-Ca-α1D^RNAi: 51491 (RNAi Ca-α1D #1), 52644 (RNAi Ca-α1D #2), UAS-cacophony^RNAi: 104168, and UAS-Ca-α1T^RNAi: 108827. Background (control) RNAi strains included w¹¹¹⁸ wild-type and KK injection lines. GFP-Zasp52 (BDSC Stock no. 6838, which expresses a fluorescently labeled scaffold protein that binds α-actinin and localizes to muscle Z-discs), hypomorphic cacophony cac^S (a generous gift from Dr. Chun-Fang Wu) [51], and T-type channel knock out (a generous gift from Dr. Carsten Duch) [52] lines were used for electrophysiological studies. All flies were maintained at room temperature (25 °C) on a standard cornmeal-yeast-sucrose-agar medium.

Heart tube morphology

Confocal microscopy was performed as detailed by Alayari et al. (2009) [53]. Briefly, w¹¹¹⁸ Drosophila hearts were surgically exposed according to Vogler and Ocorr (2009) [32] and arrested using 10mM EGTA in artificial hemolymph. The relaxed, semi-intact hearts were fixed (4% formaldehyde in 1X PBS) and washed three times with 1X PBST (PBS with 0.1% Triton X-100). Fixed hearts were then stained with Alexa594 TRITC-phalloidin (1:1000 in PBST), rinsed three times in 1X PBST, mounted and imaged with a Leica TCS SPE RGBV confocal microscope.

RNA fluorescence in situ hybridization (FISH)

TinCΔ4-Gal4 virgin females were crossed with w¹¹¹⁸ males. Five days post eclosion, the heart tubes of the progeny were surgically exposed under artificial hemolymph [32] and the semi-intact preparations fixed as described above. Cardiac in situ hybridization was performed as reported previously [54] using the QuantiGene® ViewRNA Cell Assay kit from Panomics (Affymetrix Inc.), following the manufacturer’s recommendation. Fluorophore-tagged probes were custom designed to target Drosophila mRNA coding for Ca-α1D (VF1-18656, VF4-18190), cacophony (VF4-18657), and Ca-α1T (VF4-18658), and a house keeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH, VF6-18191). The sequences chosen for probe design are present in all splice variants of each channel subunit to ensure complete coverage of mRNA molecules encoding each Ca²⁺ channel. A pre-designed probe targeting human alpha feto-protein (h-AFP, VA1-10125) was used as a negative control (Supplementary Figure 1).

Following RNA hybridization, heart tubes were mounted on coverslips in Prolong Gold Antifade mounting media with DAPI (Thermo Fisher Scientific, Catalog No. P36941) and imaged using a Leica TCS SPE confocal microscope (Leica Microsystems) at 40X magnification. Micrographs of heart tubes were analyzed with ImageJ software (National Institute of Health). To quantitate transcripts from the confocal micrographs, individual channels were separated following binary conversion of the images using the same intensity threshold for each channel across all samples. Regions of interest, which included only the cardiomyocytes, were outlined using the freehand selection tool. An automated ImageJ particle counting plugin was used to determine the number of mRNA particles. Ca_v α₁-subunit particle numbers were normalized to total area and to GAPDH.

Cardiac functional analysis and Ca²⁺ transient imaging

Homozygous TinCΔ4-Gal4;UAS-GCaMP3 or Hand^4.2-GS-Gal4 virgin females were crossed with males carrying a UAS-RNAi cassette that targeted a specific Ca²⁺ channel subunit. Two weeks post eclosion, the female progenies’ hearts were surgically exposed under artificial hemolymph and the beating, semi-intact heart tubes were filmed using a Hamamatsu Orca Flash 2.8 CMOS camera at ~120 frames per second on a Leica DM5000B TL microscope with a 10× (NA, 0.30) immersion lens. Cardiac performance was assessed from the videos using the Semi-automated Optical Heartbeat Analysis (SOHA) program [30, 55]. M-modes, which provide an edge trace documenting heart wall movement over time, were generated via the program. Cardiac function metrics used in this study are similar to those described in Fink et al (2009) [55]. Heart rate variability, akin to the previously reported arrhythmicity index [55], was calculated as heart rate standard deviation divided by median heart rate.

After filming, heart tubes were incubated with 5 μM CellTracker Orange CMTR dye (Thermo Fisher Scientific, Catalog No. C2927), a live cell-permeant fluorescence dye, at room temperature for 30 minutes and rinsed with hemolymph thrice at room temperature for 10-15 minutes. Fly hearts were then imaged using a dual-camera Andor Revolution X1 spinning disc confocal on an inverted microscope (Olympus, Inc) in both green (GCaMP3; ex/em. 488/525 nm) and red (CellTracker Orange CMTR dye; ex/em 561/617 nm) channels simultaneously.

Ca²⁺ transient analysis was performed using in-house Matlab (MathWorks, Inc.)-based algorithms. The green signal was normalized to the red signal and the fractional change of this ratio was used to gauge cardiomyocyte Ca²⁺ handling activity. A single exponential fit was performed on the decay phase of the Ca²⁺ transient to estimate the decay kinetics.

Isolation of live cardiomyocytes from Drosophila heart tubes

Heart tubes from 30-40 two-week-old adult flies were surgically removed and placed into modified Ca²⁺-free hemolymph. The solution was supplemented with collagenase type I to achieve a final concentration of 0.2% w/v and incubated at room temperature on a shaker for 20 minutes. Trypsin was added to the suspension to achieve a final concentration of 0.1% w/v and incubated for 10 minutes with gentle shaking at room temperature. The digesting tissue was gently triturated every 3 minutes after addition of trypsin. The reaction was quenched by adding fetal bovine serum at a 2:1 ratio. The solution was centrifuged at 550 × g for 5 minutes. The supernatant was aspirated and the cell pellet resuspended in Ca²⁺-free hemolymph. Cell suspensions were plated on glass coverslips 1 hour prior to electrophysiological studies to allow the cardiomyocytes to attach to the substrate.

Assessment of Drosophila cardiomyocyte and sarcomeric dimensions

Heart tubes of GFP-Zasp52 Drosophila were surgically exposed and maintained in artificial hemolymph under low Ca²⁺ as previously described [32]. The semi-intact, relaxed hearts were either 1) fixed in 4% formaldehyde in 1X PBS, washed, and mounted, 2) completely removed from the abdominal segment, fixed, washed, and mounted, or 3) removed and dissociated into individual cardiomyocytes as outlined above, prior to fixation and mounting. These three sample groups were also prepared under conditions that prevented active myosin crossbridge cycling. Thus, preceding and during fixation, the samples were exposed to 100 μM blebbistatin in 0.1% DMSO v/v or 0.1% DMSO v/v in artificial hemolymph. Individual cardiomyocytes were imaged at 63X with a Leica TCS SPE RGBV confocal microscope and cellular dimensions (length, width, area) determined using ImageJ software. Resting sarcomere lengths were measured from micrographs of semi-intact and detached hearts, also imaged at 63X, and isolated cardiomyocytes, under low Ca²⁺, blebbistatin-treated, and DMSO-incubated conditions.

Electrophysiology

Whole-cell recordings of individual cardiomyocytes were acquired at room temperature using an Axopatch 200B amplifier (Axon Instruments). Internal solutions contained (in mM): CsMeSO₃, 114; CsCl, 5; MgCl₂, 1; MgATP, 4; HEPES (pH 7.3), 10; and BAPTA, 10; at 295 mOsm adjusted with CsMeSO₃. Seals were formed in artificial hemolymph and following patch rupture, the bath solution was switched to Ca²⁺- or Ba²⁺-external solution containing (in mM): TEA-MeSO₃, 140; HEPES (pH 7.4), 10; and CaCl₂ or BaCl₂, 5; at 300 mOsm, adjusted with TEA-MeSO₃. Traces were lowpass filtered at 2 kHz, and digitally sampled at 10 kHz. A P/8 leak subtraction protocol, where a leak pulse is 1/8 of the test pulse, was used with series resistances of 1-2 MΩ. 10 μM nifedipine (Sigma-Aldrich, N7634) or 10 μM forskolin (Sigma-Aldrich, F6886) was added to the bath solution in certain experiments to assess the pharmacological specificity of the observed Ca²⁺ currents.

Ca²⁺ channel sequence alignment

Amino acid sequences of CACNA1C from Oryctolagus (Gene ID: 100101555; mRNA ID: NM_001136522) and A1D from Drosophila (Gene ID: 34950 or CG4894; mRNA ID: NM_165147) were aligned using Clustal Omega [56] to assess sequence homology.

Statistical Methods

Statistical analysis was performed using GraphPad Prism 7. Data are presented as mean ± SEM. One-way ANOVAs followed by Tukey’s multiple comparisons tests were used to determine whether statistically significant differences existed between the means of 3 or more groups. When data were not normally distributed, Kruskal-Wallis one-way ANOVAs followed by Dunn’s post hoc tests were employed. For comparisons between two unmatched groups, unpaired Student’s t-tests were used to determine if the data sets significantly differed from each other. Significance was assessed at p < 0.05.

Results

Drosophila cardiomyocytes express genes that encode specific voltage-gated Ca²⁺ channels

The fly genome encodes three α₁-subunits (Ca-α1D, cacophony, and Ca-α1T) of Ca_V. These subunits define three distinct Drosophila hetero-oligomeric channels: A1D, cac, and T-type [17, 57-59] that are homologous to the major mammalian classes Ca_V1, Ca_V2, and Ca_V3, respectively. FISH, which allows direct visualization and relative quantitation of individual mRNA molecules, was utilized to decipher the cardiomyocyte Ca_V biosignature in control TinCΔ4-Gal4 x w¹¹¹⁸ Drosophila. Semi-intact fly hearts, which remain suspended within the dissected abdominal segment (Figure 1A), were fixed, permeabilized, and incubated with fluorescent probes that possess base sequence complementarity to specific α₁-subunit mRNAs (Figure 1B). On average, when normalized to GAPDH particles, Drosophila cardiomyocytes showed an abundance of Ca-α1D (0.51±0.03) and Ca-α1T (0.45±0.03) mRNA molecules relative to a limited number of cacophony (0.06±0.01) messages (Figure 1C). Thus, the fly heart expresses significantly higher amounts of Ca-α1D and Ca-α1T vs. cacophony Ca_V α₁-subunit mRNA.

RNA interference reveals the predominant types of functioning voltage-gated Ca²⁺ channels in Drosophila hearts

The prevalence of Ca-α1D and Ca-α1T mRNA implies that A1D and T-type are the major Ca_V that orchestrate cardiac contraction in Drosophila. To verify a direct role of A1D, T-type, and potentially of cac Ca_V, in cardiomyocyte Ca²⁺ signaling, the functional consequences of heart-specific RNAi-mediated silencing of each gene were evaluated. Unlike what is found with the main classes of mammalian Ca_V, each Drosophila Ca_V α₁-subunit is encoded by a single gene, thus limiting the number of targets to be tested. Previous studies have shown that transmembrane Ca²⁺ and not Na⁺ current substantially contributes to Drosophila cardiac action potentials [40, 60]. Therefore, alteration in heart tube contraction after Ca_V α₁-subunit knockdown could hint at the predominant type(s) of Ca²⁺ channels operating in fly cardiomyocytes. A cardiac-specific driver line, TinCΔ4-Gal4; UAS-GCaMP3, was crossed with multiple UAS-RNAi lines, yielding progenies with selective reduction of one of the three Ca²⁺ channels (Figure 2A). A highly significant reduction of Ca-α1D, cacophony, or Ca-α1T transcripts was verified using FISH (Supplementary Figure 3). Spontaneous, myogenic contractions and Ca²⁺ cycling properties in semi-intact heart tubes were assessed in two-week-old adult offspring.

Figure 2. Cardiac contraction and Ca²⁺ transients are suppressed by Ca-α1D knockdown.

A) The Drosophila UAS-GAL4 bipartite expression system. Cardiac-specific TinCΔ4-Gal4 drives expression of UAS-transgenes including the simultaneous expression of a Ca²⁺ biosensor (UAS-GCaMP3) and UAS-RNAi. B) Exemplar M-mode tracings of heart tubes following Ca-α1D, cacophony, or Ca-α1T knockdown (Ca-α1D(−), cac(−), and Ca-α1T(−)). The progenies of w¹¹¹⁸ or the injection lines (KK) crossed with the driver line served as controls. Ca-α1D knockdown completely suppressed contraction. C) Quantitative measurements of cardiac physiological parameters. Population data confirmed substantially altered contraction following Ca-α1D knockdown. *** p < 0.0001 compared to w¹¹¹⁸; ## p < 0.01 compared to KK. There were no significant differences in cardiac variables between the controls. (n=25-31 animals) D) Representative Ca²⁺ transient recordings from individuals of the same population of heart tubes examined in Figure 2B–C. Ca-α1D silencing effectively abolished Ca²⁺ transients in the heart tubes. E) Measurements of Ca²⁺ transients confirmed a significant reduction in Ca²⁺ transient frequency and magnitude of the peak Ca²⁺ transient upon Ca-α1D RNAi expression. The time required to reach the peak Ca²⁺ transient magnitude (time to peak) and the time constant for the Ca²⁺ transient decay (tau decay) are not shown for Ca-α1D(−)#1 and Ca-α1D(−)#2 because of inaccurate measurements due to minimal Ca²⁺ activity in these hearts. Knockdown of cacophony or Ca-α1T produced no significant change in Ca²⁺ transient frequency, Ca²⁺ transient magnitude, time to peak, or tau decay. *** p < 0.001 compared to w¹¹¹⁸. (n=21-30 animals).

TinCΔ4-Gal4; UAS-GCaMP3 > RNAi Ca-α1D #1 Drosophila exhibited minimal heart tube motion compared to the rhythmic contractions in TinCΔ4-Gal4; UAS-GCaMP3 x w¹¹¹⁸ controls as demonstrated by M-mode recordings (Figure 2B). To account for potential confounding positional effects that may result from insertion of the RNAi cassette in discrete locations throughout the Drosophila genome, two different RNAi lines with the same genetic background as the control were evaluated, and the effects of gene silencing on several cardiac parameters were quantified. Compared to control, both TinCΔ4-Gal4; UAS-GCaMP3 > RNAi Ca-α1D #1 and > RNAi Ca-α1D #2 showed significantly decreased heart rates (0.35±0.16 and 0.36±0.10 vs. 1.24±0.10 beats/s in control) and increased heart rate variabilities (0.45±0.12 and 0.22±0.06 vs. 0.10±0.02 in control) (Figure 2C). Moreover, the extent of contraction was significantly diminished in the RNAi-expressing heart tubes as demonstrated by ~4 fold reduction in fractional shortening (0.08±0.01 and 0.05±0.01 vs. 0.42±0.01 in control) and ~4.5 fold reduction in shortening velocities (52.9±17.7 and 69.3±18.7 vs. 872.7±21.3 μm/s in control).

We next ascertained if Ca²⁺ signaling was altered following Ca-α1D knock down. GCaMP3-based green fluorescence, emitted from actively beating heart cells, was recorded simultaneously with orange fluorescence that originated from CellTracker, a dye that was passively loaded into the cardiomyocytes to monitor cell movement. The relative change in ratio between the two signals was used to determine the Ca²⁺ cycling properties. The ratiometric approach helped correct for heart contraction motion artifacts and for different amounts of GCaMP3 expression or CellTracker loading within and among the samples. Following Ca-α1D knockdown, fluorescent signal analysis suggested a completely abolished Ca²⁺ transient (Figure 2D). Moreover, population data of cyclical fluorescent fluctuations confirmed significant reductions of Ca²⁺ transient rates (RNAi Ca-α1D #1, #2 vs. control: 0.04±0.02, 0.03±0.02 vs. 1.32±0.08 Hz) and peak Ca²⁺ transient magnitudes (0.005±0.003, 0.001±0.001 vs. 0.182±0.017) upon suppression of Ca-α1D expression (Figure 2E) compared to control. This observation was consistent with the nearly complete cessation of contraction upon Ca-α1D knockdown described above.

These experimental results demonstrate the effects of constitutive Ca-α1D suppression, i.e. both during and after heart tube development. To assess functional changes in hearts with Ca-α1D knockdown post cardiogenesis, the same UAS-RNAi lines were crossed with the inducible, cardiac-specific Hand^4.2-GS-Gal4 driver line. Expression of Ca-α1D RNAi in the offspring was activated from two days after eclosion by supplementing the food with RU486 and the heart tubes of two-week-old Drosophila were imaged. Reduced fractional shortening and shortening velocity were observed when Ca-α1D was knocked down post-developmentally, although to a lesser extent compared to flies with cardiac-restricted Ca-α1D knockdown throughout development (Supplementary Figure 2). Overall, these results suggest a key role of A1D Ca_V in cardiac function in flies, both during and post-development.

In addition to A1D, potential contributions from cac and, given its high expression levels, T-type Ca²⁺ channels in defining Drosophila cardiac contraction and Ca²⁺-handling properties, were also explored. Individually suppressing expression of these channels in TinCΔ4-Gal4; UAS-GCaMP3 > cacophony^RNAi and > Ca-α1T^RNAi did not significantly alter heart rates (cacophony^RNAi, Ca-α1T^RNAi vs. control: 2.50±0.17, 1.70±0.14 vs. 1.95±0.19 beats/s), heart rate variabilities (0.16±0.02, 0.15±0.03 vs. 0.12±0.03), or shortening velocities (1056±79.3, 960±58.4 vs. 982.4±67.1 μm/s) (Figures 2B-C) compared to the TinCΔ4-Gal4; UAS-GCaMP3 x KK control line. Although there was a statistically significant reduction of fractional shortening subsequent to cacophony or Ca-α1T knockdown (0.40±0.01 or 0.40±0.01 vs. 0.44±0.01 in control), the extent of reduction was minimal compared to that following Ca-α1D knockdown. Similarly, cacophony or Ca-α1T knockdown did not yield statistically significant changes in Ca²⁺ transient rates (cacophony^RNAi, Ca-α1T^RNAi vs. control: 2.04±0.15, 1.32±0.11 vs. 1.66±0.10 Hz), Ca²⁺ transient magnitudes (0.16±0.02, 0.15±0.02 vs. 0.17±0.02), time to peak (165.0±10.8, 214.0±8.5 vs. 184.6±8.5 ms), or decay time constants (261.2±16.9, 332.6±18.6 vs. 303.3±16.6 ms) compared to control (Figures 2E). Collectively, these data illustrate that A1D plays a major role in defining the Drosophila cardiac Ca²⁺ transient and myocardial contraction with potentially minor contributions from cac and T-type Ca²⁺ channels.

Isolation and morphological characterization of Drosophila cardiomyocytes

So far, we have examined the sources of plasmalemmal Ca²⁺ flux in Drosophila cardiomyocytes using multiple indirect approaches. However, voltage clamping individual cells and directly measuring Ca²⁺ current across the membrane could ultimately confirm the major types of active Ca²⁺ channels functioning in Drosophila myocardium. Although isolated cardiomyocytes have been a mainstay for cellular electrophysiology in mammalian systems for decades, no published reports of analogous protocols for flies exist. Therefore, we devised a method for isolating viable single cells from Drosophila heart tubes, which consist of a single layer of bilateral rows of opposing cardiomyocytes. Following enzymatic dissociation of GFP-Zasp52 hearts, individual cells, which maintain their curved shape (Figure 3A, Supplementary Figure 4), were successfully obtained. Consistent with a gradual tapering of the heart’s diameter along its length, the cardiomyocytes exhibited moderate variability in dimensions, which on average were 51.3±2.7 μm in length, 32.2±1.0 μm in width, and 1561±115 μm² in maximally projected area. Sarcomere lengths were determined by measuring the distance between peak fluorescent signals emanating from the Z-disc-associated, Zasp52-GFP. Interestingly, the average resting sarcomere length along myofibrils of myocytes isolated and maintained under low Ca²⁺ (1.38±0.04 μm) was significantly less than that determined for semi-intact (2.50±0.03 μm) or detached whole hearts (2.54±0.02 μm) maintained under similar conditions (Figure 3B). Despite variability in cellular dimensions, the sarcomere lengths were consistent among isolated cardiomyocytes (Supplementary Figure 5). To determine if the shortened sarcomeres resulted from excessive actomyosin associations activated upon myocyte dissociation, we compared the distance between the middle of consecutive Z-discs along myofibrils of semi-intact hearts, detached whole hearts, and single cells maintained in DMSO or in the presence of blebbistatin, a small-molecule inhibitor of several striated muscle myosins. The sarcomere lengths in semi-intact and detached cardiac tubes in the presence of blebbistatin (2.53±0.02 and 2.53±0.02 μm, respectively) were not significantly different from those determined in DMSO (2.52±0.02 and 2.55±0.02 μm, respectively) or under conditions of low Ca²⁺ (Figure 3B). Sarcomeres of isolated cells incubated in blebbistatin or in DMSO did not differ in length from each other (1.46±0.04 vs. 1.36±0.04 μm, respectively) or from those incubated under low Ca²⁺, but were significantly shorter than the sarcomeres of semi-intact or detached whole hearts (Figures 3B and 3C). These data suggest the shortened sarcomeres result from passive processes that accompany cellular separation. Despite reduced sarcomere lengths, the individual myocytes nonetheless remained viable, as demonstrated by rhythmic contractions (Supplementary Movie 3) and the presence of Ca²⁺ transients (Supplementary Figure 4) for up to 2-3 hours in Ca²⁺-containing artificial hemolymph at 25°C.

Figure 3. Morphological characterization of isolated Drosophila cardiomyocytes.

A) GFP-ZASP52 cardiomyocytes after dissociation from detached heart tubes maintained their curved morphology. Note, the left-most cardiomyocyte originated from the conical chamber of the heart. The remaining cells most likely came from the middle one third of the cardiac tube and are representative of those commonly isolated. Scale bar = 20 μm. B) Sarcomere lengths along myofibrils within cardiomyocytes of semi-intact and detached whole hearts were not significantly different under low Ca²⁺ conditions or following exposure to DMSO or 100 μM blebbistatin. After cardiomyocyte dissociation, the average sarcomere length of isolated cells under all three conditions did not significantly differ; however, the sarcomeres were significantly shorter than those of semi-intact and detached whole hearts. *** p < 0.0001. (n=22-31 cells) C) Enlarged views of myofibrils demonstrating the effect of cardiomyocyte isolation on resting sarcomere length.

Voltage clamp experiments confirm A1D as a predominant mediator of Ca²⁺ current in Drosophila cardiomyocytes

Isolated Drosophila cardiomyocytes were voltage clamped at −80 mV (resting potential) and the Ca²⁺ current, induced by various depolarizing voltage steps, was measured. For example, the Ca²⁺ current evoked by a 0-mV step potential in control (GFP-Zasp52) cells was characterized by a rapid influx of Ca²⁺ ions, which slowly decayed over time (Figure 4A, black) and resembled the characteristic inactivating mammalian ventricular cardiomyocyte Ca_V1 Ca²⁺ current. Peak Ca²⁺ current density at various test potentials is shown in Figure 4B (black). To explore potential contributions of Ca²⁺ current from cac and T-type channels, peak Ca²⁺ current density was measured in cardiomyocytes isolated from hypomorphic cac and null T-type Drosophila lines. No significant differences in the peak current density between cardiomyocytes from the control (Figure 4B, black) and hypomorphic cac hearts (Figure 4B, blue) across all test voltages were observed. For example, the current densities of cac hypomorphic vs. control cardiomyocytes at −30 mV and 0 mV were −9.3±1.7 vs. −11.0±2.6 pA/pF and −13.9±0.9 vs. −12.2±1.4 pA/pF, respectively. These results imply a negligible contribution of cac channels in establishing Drosophila myocardial Ca²⁺ currents. However, peak Ca²⁺ current density measured from T-type null cardiomyocytes was significantly reduced relative to control at the lower voltage range (defined here as V < 0 mV with the employed external and internal solutions), e.g. −3.2±0.5 pA/pF at −30 mV, while there was no significant difference in peak current density between T-type null and control cells at the higher voltage range (defined as V > 0 mV), e.g. −13.0±2.3 pA/pF at 0 mV. As both mammalian Ca_V3 and Drosophila T-type Ca²⁺ channels are activated mainly at low voltages [61-64], the current density profile from T-type null cells compared to control suggests the presence of functional T-type Ca²⁺ channels in Drosophila heart tubes.

Figure 4. Voltage clamp recordings confirm A1D channels conduct the predominant transsarcolemmal Ca²⁺ current in Drosophila cardiomyocytes.

A) Ca²⁺ current recording of a dissociated fly myocyte. The current in control cardiomyocytes was suppressed by a dihydropyridine, nifedipine, a hallmark of Ca_V1 channels. The average capacitance of the Drosophila cardiomyocyte membrane was 125±11 pF. B) Mean peak current density in control, hypomorphic cacophony (cac^s), and T-type (Ca-α1T) null cardiomyocytes across test potentials. At high voltages, comparable to the plateau phase of mammalian cardiac action potentials, peak currents of cac^s hypomorphic and T-type null cardiomyocytes were similar to that observed in control, indicating that A1D is the major high-voltage-activated Ca²⁺ channel isoform in the Drosophila heart. At low voltages, T-type null cardiomyocytes showed reduced current densities, suggesting a contribution of T-type channels at low activation voltages. * p < 0.5 compared to control. (n=4, 6, and 4 cells for control, cac^s, and T-type null). C) Population data of the Ca²⁺ current response to 10 μM nifedipine as compared to untreated control myocytes. (n=5 cells).

Finally, to unequivocally validate a requirement for A1D Ca_V in Ca²⁺ signaling in Drosophila cardiomyocytes, dihydropyridine nifedipine, a specific mammalian Ca_V1 and Drosophila A1D channel blocker, was added to the bathing solution of control cells. Application of 10 μM nifedipine significantly decreased the magnitude of Ca²⁺ current (Figure 4A, blue) as confirmed by population data of current reduction after application of nifedipine across multiple test potentials (Figure 4C). In sum, these results corroborate the FISH data (Figure 1), and indicate that both A1D and T-type Ca²⁺ channels are abundantly present and functionally active in Drosophila cardiomyocytes.

Properties of cardiac A1D channels in Drosophila

Multiple lines of evidence illustrate that A1D is a major contributor to transmembrane Ca²⁺ currents in fly cardiomyocytes. Because A1D is homologous to mammalian Ca_V1, it may possess similar properties to its mammalian counterparts, including CDI. CDI is a negative feedback mechanism and crucial feature of Ca_V1 channels (Ca_V1.2, Ca_V1.3, and Ca_V1.4) that helps regulate the level of intracellular Ca²⁺. Key components for orchestrating CDI include calmodulin (CaM), a resident Ca²⁺ sensor molecule that is pre-bound to the C-terminus of the channel [65-67], the N-terminal spatial Ca²⁺ transforming element (NSCaTE) [68-70] (on the N-terminus of the channel), two EF hands [65, 71, 72], an IQ domain [65, 71, 73], and Ca²⁺-free CaM (apoCaM) binding sites (on the C-terminus of the channels) [65, 66] (Figure 5A). Since Drosophila A1D possesses domains with homology to all aforementioned CDI components, it also likely exhibits CDI. Ca²⁺ current recordings of fly cardiomyocytes (Figure 5B, red) showed Ca²⁺ influx due to channel opening followed by a gradual decrease in the current size or channel inactivation (Figure 5B, red). CDI can be distinguished from voltage-dependent inactivation (VDI), another form of channel feedback regulation, by comparing Ca²⁺ vs. Ba²⁺ current (Figure 5B, black) passing through the same channel. Because Ba²⁺ is unable to efficiently bind to CaM, Ba²⁺ current inactivation solely represents the degree of VDI. Thus, the true extent of CDI (Figure 5B, shaded pink) can be calculated as f₃₀₀ = (r_300/Ba − r_300/Ca)/r_300/Ba, where r_300/x is the fraction of Ca²⁺ and Ba²⁺ currents remaining after 300 ms of channel opening. For example, at a 0-mV test potential the f₃₀₀ = 0.35±0.14, which confirms that Drosophila A1D demonstrates robust CDI (Figure 5C).

Figure 5. Ca²⁺-dependent inactivation is a conserved feature of Drosophila A1D.

A) Illustration of the alpha subunit of the vertebrate LTCC with CDI interface regions NSCaTE (yellow), two EF hands (rose, green), and the IQ domain (blue). Sequence comparison of Oryctolagus’s CACNA1C and Drosophila’s A1D with CDI components highlighted in the same color as in the diagram on the left. The binding sites of Ca²⁺-free CaM (apoCaM) are highlighted in grey with key amino acid residues interacting with N-, C-, and both lobes of apoCaM bolded in blue, red, and black. B) Ca²⁺ current in fly cardiomyocytes decayed more rapidly than Ba²⁺ current in the same cell, demonstrating CDI (shaded rose) as Ba²⁺ cannot effectively bind calmodulin. C) Population data showing the fraction of current that remained after 300 ms of activation (r₃₀₀). The different degree of decay between Ca²⁺ and Ba²⁺ currents represents the extent of CDI (shaded rose). (n=7 cells).

Another key property of mammalian cardiac Ca_V1.2 is current augmentation in response to PKA-mediated phosphorylation. Adrenergic-like octopamine receptors (OctαRs, and OctβRs) [74], adenyl cyclase (rutabaga) [75], phosphodiesterase (dunce) [75], and both regulatory [76] and catalytic [77] subunits of PKA are also expressed in Drosophila (Figure 6A), suggesting that A1D may also exhibit PKA-mediated current enhancement. Application of 10 μM forskolin, an activator of adenyl cyclase, increased the amplitude of Ca²⁺ current through the A1D channels (Figure 6B). For example, at 0-mV test potential, the peak Ca²⁺ current was greatly enhanced by 1.58±0.13 fold after forskolin application (Figure 6C). Similar current amplification by forskolin was observed across multiple voltages, confirming the presence of PKA-mediated current augmentation of A1D channels in Drosophila cardiomyocytes.

Figure 6. Ca²⁺ current augmentation by protein kinase A is conserved in Drosophila cardiomyocytes.

A) Depiction of the β-adrenergic pathway. Epinephrine binds to a G-protein coupled receptor and activates the enzyme adenyl cyclase receptor, which converts ATP to cyclic AMP (cAMP). This small signaling molecule activates protein kinase A (PKA), which phosphorylates LTCCs, augmenting their current size. Phosphodiesterase (PDE) deactivates cAMP. Forskolin bypasses this signaling cascade by directly activating adenyl cyclase. Drosophila homologs of the β-adrenergic pathway elements are in parentheses. Oct, octopamine. B) Ca²⁺ current in Drosophila cardiomyocytes increased in amplitude after application of 10 μM forskolin. C) Population data confirm augmentation of Drosophila A1D current by PKA. (n=4 cells).

Discussion

Drosophila represents a potentially ideal platform for studying regulation of, and diseases involving, cardiac Ca²⁺ channel function and dysfunction. In addition to their genetic pliability, fly cardiomyocytes, as shown here, harbor A1D and T-type Ca²⁺ channels (orthologs of mammalian Ca_V1 and Ca_V3 channels, respectively [17]), which is analogous to the Ca²⁺ channel ensemble of mammalian cardiomyocytes [78-80]. Moreover, Drosophila A1D also possesses striking conservation of two key properties of Ca_V1 channels, which are CDI and PKA-dependent current augmentation.

In the current study, we show for the first time, that Drosophila cardiomyocytes express significantly higher amounts of Ca-α1D and Ca-α1T relative to cacophony mRNA. Functional analyses revealed drastically impaired heart tube contraction and decreased Ca²⁺ transient amplitude when Ca-α1D expression was suppressed by RNAi as opposed to a minimal change in contractile properties resulting from Ca-α1T or cacophony silencing. These results imply that not only is A1D highly enriched in Drosophila cardiomyocytes, but that A1D channels are the major contributor to Ca²⁺ flux responsible for orchestrating contraction. However, despite an abundance of Ca-α1T mRNA, T-type channels were shown to contribute minimally to heart tube contraction. These results suggest that there is a potential discordance between Ca-α1T mRNA versus protein load, a high transcriptional reserve to ensure an adequate T-type channel stoichiometry, and/or poorly effective RNA interference and gene suppression. Direct assessment of plasmalemmal Ca²⁺ currents in cardiomyocytes that are genetically devoid of T-type channels could reveal insight into this discrepancy.

Although multiple reports previously described changes in contractions of intact Drosophila hearts [34, 40, 42, 55, 60, 81], at baseline or in response to pharmacological manipulation, approaches to evaluate plasmalemmal currents in adult cardiomyocytes have remained somewhat coarse. Field potentials across the surface of heart tubes were recorded [34, 36] while a rudimentary current clamp recording, performed by inserting electrodes into whole hearts, successfully measured transmembrane voltages [44, 82, 83]. However, the gold standard to directly probe plasmalemmal current is the voltage clamp technique, which yields highly resolved data when performed on single cells as opposed to cellular networks. Voltage clamp recordings have been acquired from neurons of Drosophila embryos [84], pupae [57], and adult [57] brains, motoneurons of adult indirect flight muscle [52] and of larval body-wall muscle [85], larval retinal cells [86], and larval body wall muscles [87-89]. Nonetheless, despite well-established methods for the isolation of cardiomyocytes from various mammalian species and subsequent transsarcolemmal ion current measurements, voltage clamp experiments using single Drosophila cardiomyocytes have not been conducted. Therefore, based on approaches employed for mammalian tissue, we developed a method to enzymatically dissociate viable cardiomyocytes from adult Drosophila heart tubes.

Brief enzymatic digestion of ~35 fly hearts resulted in a population of individual cells that appeared well suited for electrophysiological studies. As reported for vertebrate cardiomyocytes, which show a significant 17% reduction in sarcomere length upon isolation from the intact heart [90], Drosophila sarcomeres likewise displayed reduced lengths following cardiomyocyte dissociation. Comparing the sarcomere lengths from cells of semi-intact relative to those of detached heart tubes revealed that changes in sarcomere length did not transpire in response to heart extraction. Moreover, incubating and digesting the hearts in artificial hemolymph containing either EGTA and cell-permeant EGTA-AM, which chelates and minimizes intracellular Ca²⁺ (data not shown), or blebbistatin did not prevent sarcomere shortening upon cell-to-cell separation. These findings suggest that the change in sarcomere length is not due to excessive or unregulated Ca²⁺-mediated tension or active actomyosin-based contraction in general.

The heightened sarcomeric shortening event that occurs upon fly vs. rat cardiomyocyte isolation (i.e. ~40 vs. 17% reduction in sarcomere length) may be related to several factors. Firstly, the extreme geometric constraints that are imposed upon fly cardiomyocytes, due to two opposing cells establishing the tubular nature of the organ, are partially released following proteolytic digestion. Consequently, coinciding with changes in cellular strain/stress upon separation, relatively large changes in Drosophila myofibril and therefore sarcomere deformation are expected to occur. The elastic properties of the fly connecting filaments also likely differ from those of vertebrate cardiac titin isoforms, which are the predominant determinants of cardiomyocyte passive tension [91]. Such differences may uniquely influence cellular and molecular recoil post digestion.

Despite the exaggerated changes in resting sarcomere length that transpire upon dissociation, we show that Drosophila cardiomyocytes are exceptionally amenable to voltage clamp experiments. We confirmed the presence of Ca²⁺ currents through both A1D and T-type Ca²⁺ channels, which are remarkably similar to the those observed in mammalian cardiomyocytes [79, 80]. In mammals, T-type channels or Ca_V3 are activated at lower potentials and are characterized by a small conductance [79, 80]. Cardiac LTCCs or Ca_V1.2, however, are activated at relatively higher potentials compared to Ca_V3 with higher pore conductance [79, 80] and, thus, are better-suited to drive the shape and duration of the cardiac action potential [1]. As heart tube contraction is the end-point of a multi-step process, from pacemaker cells firing, action potential transmission to myocytes, and Ca²⁺-centric conversion of electrical to mechanical activities, alteration in contractile properties could stem from deviations in any of these critical steps. The T-type Ca²⁺ current was proposed to play a modulatory role in pacemaking, which is predominantly driven by a Na⁺/Ca²⁺ exchanger-dependent Ca²⁺ clock [92, 93]. Therefore, consistent with our findings, suppression of Ca-α1T expression by RNAi minimally affects the contractile properties of the heart tubes. LTCCs, on the other hand, are responsible for the transsarcolemmal Ca²⁺ flux that triggers Ca²⁺-induced Ca²⁺ release from the sarcoplasmic reticulum leading to cardiomyocyte contraction. Hence, it is not surprising that Ca-α1D knock down nearly abolishes heart tube contraction (Figure 2).

Most interestingly, Drosophila A1D also shares key regulatory features with those of mammalian Ca_V1.2, which include CDI and PKA-mediated current augmentation. After entering through Ca_V, Ca²⁺ ions bind to CaM, and Ca²⁺/CaM induces a conformational change that decreases open probability of individual channels. This leads to a reduction of current amplitude at the whole cell level [69, 94] and, thereby, accounts for CDI. During the fight-or-flight response, epinephrine binds to the β₁-adrenergic receptor, a type of G-protein-coupled receptor, and initiates a signaling cascade that results in PKA-mediated Ca_V1.2 phosphorylation. The phosphorylated channels display higher open probability, which leads to an increased whole cell Ca²⁺ current size. The larger Ca²⁺ influx enhances the force of cardiac contraction that is necessary for the flight response. The conservation of these vital features in fly cardiomyocytes hints at the evolutionary importance of LTCCs and, thus, their tightly controlled functions.

Despite the advantages afforded by Drosophila myocytes, there are considerations prior to their widespread use for mechanistic studies of cardiac Ca²⁺ channels. For example, while we confirmed the presence of CDI and PKA-mediated current enhancement, several additional up- and downstream modifiers of Ca²⁺ flux, which exist in cardiomyocytes from higher organisms, have yet to be characterized in insects and may be absent. Additionally, the cells that comprise the complex, four-chambered mammalian heart display regional differences in their Ca²⁺ current and electrophysiological properties⁷⁸. Myocytes originating from different locations along the Drosophila cardiac tube may also display regional variation in Ca²⁺ currents; however, resolving such differences would require careful cellular purification and handling procedures. Moreover, limited tissue yields could make the fly model cumbersome for certain applications. Multiple surgeries are required to obtain enough cells for well-powered studies. Nonetheless, our data suggest Drosophila proves to be an effective and novel platform to investigate regulation of and diseases involving cardiac Ca²⁺ channels. Due to the fly’s genetic versatility, efficient tools, and multiple modalities for functional assessment, screens to dissect physiological or unravel pathological mechanisms are likely eminently feasible using this animal model.

Highlights.

• Drosophila cardiomyocytes primarily express mRNA encoding L- and T-type Ca channels

• A1D L-type Ca channels are required for contraction of the fly heart

• Isolated Drosophila cardiomyocytes are amenable to patch clamp experiments

• A1D is the main conduit for sarcolemmal Ca flux in fly cardiomyocytes

• Drosophila can serve as an efficient model to study cardiac Ca channel regulation