|Home | About | Journals | Submit | Contact Us | Français|
The Popeye domain containing (Popdc) genes encode a family of transmembrane proteins with an evolutionary conserved Popeye domain. These genes are abundantly expressed in striated muscle tissue, however their function is not well understood. In this study we have investigated the role of the popdc2 gene in zebrafish. Popdc2 transcripts were detected in the embryonic myocardium and transiently in the craniofacial and tail musculature. Morpholino oligonucleotide-mediated knockdown of popdc2 resulted in aberrant development of skeletal muscle and heart. Muscle segments in the trunk were irregularly shaped and craniofacial muscles were severely reduced or even missing. In the heart, pericardial edema was prevalent in the morphants and heart chambers were elongated and looping was abnormal. These pathologies in muscle and heart were alleviated after reducing the morpholino concentration. However the heart still was abnormal displaying cardiac arrhythmia at later stages of development. Optical recordings of cardiac contractility revealed irregular ventricular contractions with a 2:1, or 3:1 atrial/ventricular conduction ratio, which caused a significant reduction in heart frequency. Recordings of calcium transients with high spatiotemporal resolution using a transgenic calcium indicator line (Tg(cmlc2:gCaMP)s878) and SPIM microscopy confirmed the presence of a severe arrhythmia phenotype. Our results identify popdc2 as a gene important for striated muscle differentiation and cardiac morphogenesis. In addition it is required for the development of the cardiac conduction system.
Electrical activity is autonomously generated in the vertebrate heart. The sinus node acts as the primary cardiac pacemaker and electrical activity is rapidly propagated through the atria. The electrical impulse is delayed at the atrioventricular (AV) junction due to the presence of slow conducting AV node cells. This arrangement allows the atria to contract before the ventricle. A fast-conducting network of specialized cells in the His-Purkinje system ensures that the ventricle is electrically excited from the apex to the base allowing efficient ejection of blood (Christoffels et al., 2010). Several transcription factors have been recognized that do play an important role in specifying and shaping the transcriptional programs of the cardiac conduction system (CCS) (Blaschke et al., 2007; Hoogaars et al., 2007; Moskowitz et al., 2007). Likewise, signaling molecules involved in CCS development including neuregulin and endothelin have been identified (Mikawa and Hurtado, 2007). However, many aspects of CCS development are currently not well understood.
The zebrafish has recently emerged as a novel system to study cardiac electrophysiology and the development and physiology of the cardiac conduction system (Milan and Macrae, 2008). In zebrafish the cardiac conduction system lacks the anatomical boundaries that are found in the mammalian heart. Nonetheless, a distinct cardiac pacemaker has been functionally defined in the embryonic zebrafish heart (Arrenberg et al., 2010). Impulse delay at the atrioventricular canal (AVC) allowing sequential contraction of atrial and ventricular chambers is first observed 36–40 h post fertilization (hpf) and is correlated with a cell shape change of AVC myocytes (Chi et al., 2008; Milan et al., 2006a). A fast conducting ventricular conduction tissue develops at 100 hpf and its functionality is correlated with the development of ventricular trabeculation (Chi et al., 2008). Despite significant anatomical differences, the electrophysiology and the beating rates of zebrafish and human hearts were found to be surprisingly similar (Kopp et al., 2005; Leong et al., 2010). Many of the ionic currents involved in the generation and shaping of the cardiac action potential are identical, although some differences are also present (Milan et al., 2006b; Nemtsas et al., 2010). Several zebrafish mutants have been identified that faithfully model human conduction diseases such as long or short QT syndromes, which are difficult to model in the mouse due to the high heart rate and differences in potassium currents being active during repolarization (Arnaout et al., 2007; Hassel et al., 2008; Langheinrich et al., 2003; Milan et al., 2003; Salama and London, 2007). Thus, the zebrafish appears to be a powerful system to study cardiac electrophysiology and recent pharmacological and mutagenesis screens in the zebrafish model resulted in the identification of several novel genes involved in the developmental or functional control of the cardiac conduction system (Chi et al., 2008; Milan et al., 2009).
The Popeye domain containing (Popdc) gene family encodes a unique class of membrane proteins harboring three transmembrane domains and an evolutionary conserved intracellular Popeye domain (Andrée et al., 2003, Brand, 2005; Knight et al., 2003; Osler et al., 2006). In vertebrates, the Popdc gene family consists of three members, Popdc1 (also known as Bves), Popdc2, and Popdc3 (Brand, 2005; Osler et al., 2006). Popdc2 is uniquely present in vertebrates, whereas Popdc1 and Popdc3 orthologs are also found in lower chordates and insects (Brand, 2005; Lin et al., 2007). All three genes in vertebrates display an overlapping and evolutionary conserved expression pattern, being highly abundant in cardiac and skeletal muscle, suggesting an important function in these tissues (Andrée et al., 2000; Breher et al., 2004; Froese and Brand, 2008; Hitz et al., 2002; Parnes et al., 2007; Torlopp et al., 2006). Popdc genes are also expressed in other cell types such as smooth muscle cells, neurons, and several types of epithelial cells, which indicate a more wide-spread role for this gene family (Brand, 2005; Osler et al., 2006). In Drosophila, antisense-mediated knockdown of the Popdc1 homolog causes abnormal epithelial movement and failure of pole cell migration (Lin et al., 2007). Similarly, morpholino-mediated knockdown of Popdc1 in Xenopus causes an arrest in gastrulation (Ripley et al., 2006). In contrast, the null mutant of Popdc1 in mice is viable. However, skeletal muscle tissue in this mutant displays impaired regeneration after experimental wounding (Andrée et al., 2002). Popdc1 has been proposed to act in cell adhesion since it is rapidly recruited to sites of cell–cell contact formation (Wada et al., 2001). Moreover, Popdc1 has been reported to control membrane localization of ZO-1 and thereby modulate tight junction integrity in epithelial cells (Osler et al., 2005). Recently, protein partners for Popdc1 have been identified, the guanine nucleotide exchange factor T (GEFT), which acts as a GEF for Rho-family GTPases (Smith et al., 2008), and the SNARE proteins VAMP2/VAMP3, which control vesicular transport and β1-integrin recycling (Hager et al., 2010). A null mutation of Popdc2 has recently been engineered in mice (Froese and Brand, 2008) and interestingly, a cardiac arrhythmia phenotype was present in null mutants (Froese et al., 2012).
Here we demonstrate that during zebrafish development popdc2 is expressed in skeletal muscle and heart. Global loss of popdc2 affected skeletal muscle development resulting in abnormal facial and trunk muscle formation. The alignment of muscle fibers as well as establishment of myotendinous junctions was defective in the morphants. An even more severe defect was seen in craniofacial muscle development. Here many muscles were malformed, or reduced in size. In the heart looping was defective and the chambers were misshaped. After reducing the morpholino concentration, muscle and heart development appeared morphological normal, however the heart became arrhythmic at later stages of development, showing irregular action potential durations, and an AV block. Thus, popdc2 in zebrafish is essential for muscle and heart development and is required for cardiac conduction system development.
Zebrafish were raised under standard conditions at 28.5 °C and staged as previously described (Westerfield, 2003). The following transgenic lines were used: Tg(cmlc2:gCaMP)s878 (Arnaout et al., 2007), Tg(cmlc2:eGFP) (Huang et al., 2003), Tg(flk1:eGFP)s843 (Jin et al., 2005), Tg(cmlc2:eGFP-ras)s883 (D’Amico et al., 2007), Tg(gata1:dsRed)sd2 (Traver et al., 2003), Tg(acta1:GFP) (Higashijima et al., 1997).
Morpholino-modified oligonucleotides (MO) were obtained from Gene-Tools (Philomath, OR). Embryos were injected at the 1–2 cell stage with 0.5 ng to 2 ng of morpholino. Morpholino sequences were directed against the splice-acceptor site of exon 2 of popdc2 (MO1-popdc2: 5′-CTTAATCTGGAATTAAACAGGAGAA-3′) and the splicedonor site of exon 1 of popdc2 (MO2-popdc2: 5′-GTTCAATTGTTTCTCACCTGCCAGA-3′). Embryos were injected at the 1–2 cell stage with 0.5–1 ng MO1-popdc2 for analysis of the cardiac phenotype and 1–2 ng MO1-popdc2 for analyzing the tail and craniofacial muscle defects. In case of the injection of MO1/MO2 morpholinos 1 ng of each morpholino was used. For control purposes a standard control morpholino: 5′-CCTCTTACCTCAGTTACAATTTATA-3′ was used at the respective concentration.
The ability of murine Popdc2 mRNA to rescue the morphant phenotype was tested by subcloning the coding sequence of Popdc2 into the Stu I restriction site of pCS2+. The plasmid was linearized by enzymatic digestion with Sac II and capped mRNA was synthesized using the SP6 mMessage Machine Kit (Ambion). Rescue was achieved using 100 pg cRNA and 1 ng of MO1-popdc2.
RNA of adult tissues and embryos was isolated with Trizol Reagent (Invitrogen) and purified according to manufacturer’s instructions. cDNA was synthesized from DNase-treated total RNA using AMV reverse transcriptase (Promega). The following primer pairs were used for RT-PCR analysis to study the extent of interference with splicing after morpholino injection (Figs. 3B,C): Efw (5′-CTGCTGATGCTGCAGTGTTT-3′) and Erev (5′-GGCACGTCCTCTTTGGTATC-3′); Ifw (5′-TGCAGCTGAAATCTGAATGG-3′) and Irev (5′-GTTTAAAGCAGGCCACCTGA-3′). The following primer pairs were used for RT-PCR analysis of gene expression (Figs. 2U,V): zf-popdc2-fw (5′-GACGGGGAACAGAAGCACAGACA-3′) and zf-popdc2-rev (5′-ACCGCCCATATAGCCCAGAAAAA-3′); β-actin-fw (5′-CTTGCGGTATCCACGAGAC-3′) and β-actin-rev (5′-GCGCCATACAGAGCAGAA-3′.
Zebrafish embryos were fixed overnight at 4 °C with 4% paraformaldehyde in phosphate buffered saline (PBS). After extensive washing in PBS, the embryos were equilibrated for 30 min in 0.2% Triton X-100/PBS. Subsequently the embryos were incubated with rhodamine-phalloidin (Molecular Probes; 1:150 dilution in 0.2% Triton X-100/PBS) for 20 min. Subsequently, the embryos were washed several times in 0.2% Triton X-100/PBS and embedded in 1.5% LMW agarose for confocal microscopy. Confocal stacks were obtained with a Leica TCS-SP2 AOBS.
Whole mount in situ hybridization was carried out as described previously (Jowett and Lettice, 1994). The following digoxygenin-labeled antisense probes were used: cmlc2 (Yelon et al., 1999), vmhc (Yelon et al., 1999), myod (Weinberg et al., 1996), notch1b (Walsh and Stainier, 2001), and bmp4 (Chen et al., 1997). A full-length popdc2 cDNA clone was identified in the IMAGE library (IMAGE clone: 2602337; NCBI accession number: 153906) and was fully sequenced. The nucleotide sequence has been deposited in GenBank (accession no.: AAQ57588). An antisense probe was generated using T3 polymerase after linearization with BamHI.
For whole-mount immunostaining, embryos were fixed in 4% paraformaldehyde or in case of FAK with Dent’s fixative (Dent et al., 1989) at 4 °C overnight. Blocking was performed using PBS with 5% goat serum, 0.2% Triton X-100 and 1% DMSO. Subsequently the embryos were incubated with the primary antibody at 4 °C over night followed by intense washing steps in PBS with 1.5% goat serum, 0.2% Triton X-100 and 1% DMSO. The following primary antibodies were used: monoclonal anti-vinculin (Sigma, 1:400), polyclonal anti-FAK (Santa Cruz Biotechnology, 1:200), monoclonal beta-catenin (BD Transduction Laboratories, 1:200), polyclonal beta-catenin (Cell Signaling, 1:200), monoclonal F59 (DHSB, 1:20), monoclonal F310 (DHSB, 1:20), monoclonal anti-Alcam (Zn8, DHSB, 1:250), monoclonal anti-sarcomeric α-actinin (Sigma, clone EA-53, 1:500), and monoclonal anti-tropomyosin (Sigma, clone TM311, 1:200). The secondary antibody was conjugated with Alexa Fluor 488 or 566 (Invitrogen, 1:400) and incubation was performed for 2–3 h at room temperature.
4 dpf and 5 dpf zebrafish embryos were fixed in 3.5% formaldehyde/0.1 M sodium phosphate buffer for 1 h and stored in 70% methanol. The protocol for simultaneous bone and cartilage staining was previously described in Spoorendonk et al. (2008) and adapted from Walker and Kimmel (2007). Embryos were rinsed in 50% ethanol and cartilage staining was performed using 0.2 mg/ml alcian blue 8 GX (Sigma) in 70% ethanol/80 mM MgCl2. After several wash steps in 0.02% Triton X-100, embryos were bleached in 1% H2O2/1% KOH for 30 min and rinsed in a saturated sodium tetraborate solution. The embryos were digested for 1 h in 1 mg/ml trypsin (Sigma) in 60% saturated sodium tetraborate. The bones were stained with 0.04 mg/ml alizarin red S (Sigma) in 1% KOH, destained with an increasing glycerol series and stored at 4 °C in 70% glycerol.
3 dpf zebrafish embryos were analyzed by electron microscopy using a Zeiss EM-10 following standard procedures according to (Schoft et al., 2003).
Zebrafish embryos (Tg(cmlc2:eGFP-ras)s883 (Jin et al., 2005) and Tg(flk1:eGFP)s843 (D’Amico et al., 2007) were embedded in 2% low melting agarose supplemented with tricaine and imaged by confocal microscopy (Zeiss, LSM 710).
Cardiac contractility was recorded in tricaine anesthetized 5–7 dpf old embryos. The imaging system consisted of an inverted microscope (Zeiss Axiovert 25) equipped with a digital high-speed video camera (Basler Ahrensburg, Germany). Video images were recorded at 28 °C (20,000 images, 50 fps) and heart rates were measured as described previously (Schwerte and Fritsche, 2003). Luminance periodograms of pixels being passed by the ventricle or atrium were filtered by an equi-ripple low pass filter with a pass frequency of 0.1 Hz and a stop frequency of 10 Hz. In this filtered periodograms, peaks were detected and peak-to-peak distances (reflecting beat-to-beat distances) were converted to a beat-to-beat frequency for each interval. These frequencies were plotted against time to a cardiotachogram, which was analyzed by power spectrum analysis to evaluate frequency domains.
For electro-mechanical isolation, 5 and 6 dpf Tg(cmlc2:gCaMP)s878 embryos were transferred to 40 mM butanedione monoxime (BDM) in E3 for 1–2 min and then embedded in 1.5% low melting agarose in E3 supplemented with 20 mM BDM and subjected to SPIM microscopy to detect calcium transients as described (Arnaout et al., 2007). The obtained video sequences were analyzed with Matlab (The Mathworks, Natick, MA). Six areas within the video sequence were selected and fluorescence intensity in these areas was plotted over time in a semi-logarithmic manner.
We identified a popdc2 homolog in zebrafish through database screening (Fig. S1). Expression analysis by RT-PCR revealed that popdc2 mRNA was detectable at 4 hpf (Fig. 1A). Expression level peaked at 24 hpf, and gradually decreased afterwards. Expression analysis using whole mount in situ hybridization revealed that popdc2 was not detectable during early stages of somitogenesis, but when skeletal muscle differentiation commenced at 14–16 somites, popdc2 expression became detectable (Figs. 1B–D). Expression displayed an anteroposterior gradient, since newly formed somites did not express popdc2, however, expression was detected in somites that were more mature. Myocardial expression was first observed at 22 hpf in the forming heart tube (Fig. 1E). At 24 hpf, expression in the heart was confined to differentiated cardiac muscle cells (Figs. 1F,H,J). At the same time popdc2 was expressed in the entire myotome (Figs. 1F,G,I). At 48 and 72 hpf, expression in the trunk muscle cells disappeared, but was present in muscle cells of the pectoral fin bud and facial muscles. At this time in development, prominent expression in both, atrial and ventricular chambers was observed in the heart (Figs. 1K–R). In the adult, popdc2 was present in the heart and faintly expressed in skeletal muscle (Fig. 1S).
In order to study the function of popdc2, we interfered with its expression using two different morpholino oligonucleotides, which were targeted to the splice donor (MO2) and acceptor sequences (MO1) of intron 1, respectively (Fig. S2A). Both morpholinos, which were used singly (Fig. S2B,C), or in combination (Fig. S2D) suppressed expression of the popdc2 transcript with similar efficiency. In each case aberrantly spliced popdc2 transcripts retaining intron 1 sequences were detected by PCR using intron-specific primer (Fig. S2B–D). Injection of MO1, MO2 or a combination of thereof evoked similar phenotypes in heart and skeletal muscle development (Fig. S2E–H). A majority (n=73/92) of MO1-popdc2 morphants displayed aberrant tail morphology at 24 hpf. Identical phenotypes at similar frequency were seen in case of MO2 or MO1/MO2 injected embryos. After onset of blood circulation the morphants developed a pericardial edema (Fig. S2E–H), which was often accompanied by prominent blood accumulation at the venous pole at 48 hpf. At later stages, the edema became more extensive, resulting in stretching of the heart, and ultimately the heart was ruptured due to the massive inflation of the pericardial sac (data not shown).
In order to more closely evaluate the skeletal muscle phenotype, embryos receiving MO1-popdc2, a combination of MO1 and MO2, or control morpholinos, respectively, were stained using rhodamine-labeled phalloidin (Figs. 2A–F). In control embryos, the myotomal segments were chevron-shaped and evenly spaced. The myofibrillar content, which at 24 hpf was initially sparse increased in density by 72 hpf (Figs. 2A,D). In contrast, in the MO1-popdc2 morphants the myotomal segment borders were u-shaped and the segmental borders were ill defined and irregularly spaced (Figs. 2B,E). Myofibrils were often unevenly distributed and areas lacking myofibrils were observed. Somitogenesis was normal as revealed by MyoD staining (Fig. S3), and therefore aberrant segmental borders in trunk muscle tissue were independent of paraxial mesoderm segmentation. A combined injection of MO1- and MO2-morpholinos exaggerated the phenotype (Figs. 2C,F). Knockdown of popdc2 did not affect muscle-fiber type specification as shown by the normal distribution of both fast and slow muscle fibers in the popdc2 morphants (Figs. 3A,D,G,J). However, the overall muscle mass was noticeably reduced in the popdc2 morphant. Both muscle fiber types were morphologically abnormal. While the multinucleated fast muscle fibers in control embryos were relatively even in size and were stacked neatly in parallel rows, morphant fast muscle displayed varying cell shapes and was highly disorganized (Figs. 3B,E). Likewise slow muscle fibers in the popdc2 morphant also displayed malalignment and tissue gaps in the otherwise tight arrangement (Figs. 3H,K). By 3 dpf, both fiber types became increasingly disorganized in the morphant embryos with occurrence of ruptured fast muscle fibers, and many slow muscle fibers with a wavy appearance and the presence of widening gaps (Figs. 3C,F,I,L). The myofibrillar apparatus displayed severe myofiber disarray as revealed by α-actinin staining (Fig. S4). The development of the myotendinous junctions was studied using FAK and vinculin as marker proteins (Henry et al., 2005). Early histogenesis of trunk skeletal muscle at 17 and 18 hpf appeared to be unaffected by the loss of popdc2 (Fig. S5). However by 24 hpf the myotendinous junction was much thinner and discontinuous suggesting a failure to built up or maintain proper myotomal boundaries (Fig. 4). We analyzed whether the severe trunk muscle defect was rescued by mouse Popdc2 cRNA (Fig. 5). Popdc2 cRNA injected zebrafish (n=60/60) displayed a clearly structured tail indistinguishable from control embryos (Fig. 5). In contrast, MO1-popdc2 injected embryos (n=59/62) showed a disorganized tail at 24 hpf and developed pericardial edema, which started to become visible at 48 hpf (Fig. 5). Embryos co-injected with Popdc2 cRNA and MO1-popdc2 displayed a rescue of the trunk muscle phenotype at 24 hpf in 22 out of 82 (27%) embryos and normal heart development at 48 hpf. In addition 8 embryos retained the pericardial edema but a rescue of the tail phenotype was seen (weak phenotype, Fig. 5). We conclude that the observed defects in skeletal muscle development can be partially rescued by the addition of mouse Popdc2 to counteract the knockdown by the popdc2 morpholino.
At 3 dpf, the morphants were often lying motionless on the substrate, or displayed abnormal circular swimming behavior after stimulation. While the observed morphogenetic defects in trunk skeletal muscle development are likely to be causal for the observed muscle dysfunction, we cannot rule out that this phenotype is secondary and caused for example by cardiac insufficiency.
Craniofacial muscle development was also severely affected by the loss of popdc2. Confocal analysis of the heads of 3 dpf old Tg(acta1:GFP) embryos revealed a loss of craniofacial muscles, or a massive reduction in size and those muscles were dysmorphic (Fig. 6). Similar results were obtained after whole mount staining of the head of morphants using MF20 antibody (Fig. S6). The impaired development of craniofacial muscle was not accompanied by aberrant craniofacial bone and cartilage formation, which was unaffected in the popdc2 morphant (Fig. S7). Similarly the formation of pharyngeal arch tissue was indistinguishable between WT and popdc2 morphants (data not shown) suggesting that the cranial phenotype was confined to the skeletal muscle compartment.
At a morpholino concentration of 2 ng/embryo, the heart was dysmorphic and showed aberrant looping and a lack of trabeculation (Figs. 7A,B). Myocyte shape appeared to be unaffected, however the development of the myofibrillar apparatus in the popdc2 morphants was less advanced when compared to the control morphant (Figs. 7C–F). The heart displayed arrhythmia, which became more severe as development progressed and most embryos had a silent ventricle at 3–4 dpf. Due to the development of a severe pericardial edema the heart was stretched and therefore it is unclear whether the arrhythmia phenotype was primary or due to the morphological aberrations. In order to study the conduction phenotype further, the morpholino concentration was reduced to 1 ng/embryo. At this concentration, pericardial edema was not present between 24–72 hpf and skeletal muscle development appeared undisturbed. At 72 hpf some morphants displayed cardiac arrhythmia. At 4 dpf, a mild pericardial edema developed in many morphants (n=38/103) and a minor fraction of them displayed ventricular pauses (n=8/103). At 5 dpf, the total number of animals with cardiac defects increased (n=63/103) as well as the cardiac phenotype became more severe: i.e. pericardial edema (n=16/103), and cardiac arrhythmia (n=41/103). Cardiac arrhythmia was characterized by ventricular pauses with a 2:1 and 3:1 AV block pattern (Fig. 8A; Movies 1 and 2 in the supplementary material). At 5 dpf the 2:1 AV block occurred with high frequency (30±6/min (n=4)), and a few 3:1 ventricular pauses (3±1 (n=4)), were also observed. Due to these ventricular pauses, the resulting heart frequency (182±4) was significantly different from controls (235±18; p≤0.05; Fig. 8B) and similar observations were made at 6 and 7 dpf. At 7 dpf, morphant hearts also displayed extensive pauses of atrial and ventricular contraction. A mean of 1.2±0.4 pauses per min with a duration of 9.8±1.8 s was seen (Movies 3 and 4 in the supplementary material), which resulted in a further drop of the mean heart frequency, which was 165±15 in the popdc2 morphants vs. 201±8 (p≤0.05) in controls.
In order to further study the cardiac conduction phenotype in the popdc2 morphants, cardiac Ca2+ cycling was studied in the Tg(cmlc2:gCaMP)s878 line, which expresses a voltage-sensitive fluorescent Ca2+ sensor in myocytes (Arnaout et al., 2007). In order to image Ca2+-oscillations in these hearts with high spatial and temporal resolution, selective plane illumination microscopy (SPIM) was utilized (Huisken and Stainier, 2009). Cardiac contraction was uncoupled from electrical excitation using butanedione monoxime (BDM) and in vivo images were obtained at 5 and 6 dpf. In wild-type hearts (n=4), repetitive fluorescent waves representing systolic Ca2+-release were visible, spreading from the atrium via the AV-junction to the ventricle (Fig. 9A and Movie 5 in the supplementary material). In MO1-injected embryos (n=6), several different aberrant patterns of Ca2+-waves were visible (Figs. 9B–D and Movies 6, 7, and 8 in the supplementary material). In two of the imaged morpholino-treated embryos (Figs. 9B, C), intervals between individual Ca2+-peaks were irregular in the atrium and this might have its impact on electrical excitation in the ventricle (Fig. 10). Several of the intervals between atrial Ca2+-peaks were very short and the electrical impulses possibly reached the AV-junction at a time when the ventricle was still in its refractory state and therefore unable to respond. In three other cases, the atria displayed normal rhythmicity and no interval variations were seen, however, an AV block with different severity was observed (data not shown). Another phenotype that we observed was a sinoatrial block (Fig. 9D).
Confocal analysis of cardiac myocyte morphology using the Tg(cmlc2:eGFP-ras)s883 line revealed normal atrial and ventricular chamber dimensions and trabeculation (Fig. S8A–D). Moreover, cardiac myocyte morphology in the atrium and in the AV canal showed no obvious alterations (Fig. S8E–H). This was also the case for the endocardium and valve leaflets, which were analyzed in morphants with a Tg(flk1:eGFP)s843 background (Fig. S8I–L). Taken together no evidence was found for a morphological correlate explaining the observed cardiac conduction abnormalities. The expression of vmhc and cmlc2 was also studied and both genes were normally expressed and no difference was seen between popdc2 morphants and control embryos (Fig. S9A–D). Since in the popdc2 morphants an AV-block was observed, the expression of bmp4 and notch1b, two established myocardial and endocardial AV canal marker genes were studied. The expression of both genes was not affected by the loss of popdc2 (Fig. S9E–H). Histology and ultrastructure of 72 hpf popdc2 morphants and control embryos also revealed no difference (Fig. S10). In order to make sure that at the time the hearts of popdc2 morphants became arrhythmic, the morpholinos were biological active, RT-PCR analysis using RNA isolated from embryos at 4 and 5 dpf and exon and intron-specific primers was performed (Fig.S11). At both time points the morpholinos efficiently suppressed popdc2 expression and the transcripts that were present retained intron 1. We therefore conclude that the observed cardiac conduction phenotype was present at a time the popdc2 morpholinos still had their full biological activity.
This study represents the first functional analysis of the popdc2 gene. We report that the zebrafish popdc2 gene is required for skeletal and cardiac muscle development. During trunk muscle development the morphants displayed signs of a myopathy, which was characterized by a loss of individual myofibers, aberrant muscle segmentation, and myofibrillar disarray. A similar phenotype has also been reported in a number of other genes acting in different cellular pathways during muscle development (Dowling et al., 2008; Nixon et al., 2005; Raeker et al., 2010; Zoeller et al., 2008). Interestingly, in case of craniofacial muscle development, we observed a malformation and size reduction of individual muscles. It will be interesting to find out whether the phenotypes in the head and trunk muscles are based on the same molecular mechanism. In order to gain further insight in the underlying pathology, it will be important to study migration, fusion and differentiation of muscle precursor cells by time-lapse analysis in vivo. Interestingly, a muscle phenotype was also observed in case of the related family member Popdc1 in mice (Andrée et al., 2002). The Popdc1 null mutant displayed an impaired ability of skeletal muscle regeneration in the adult. In this regard it is also noteworthy that two recently identified Popdc1-interacting proteins, GEFT and VAMP2/VAMP3 have both been implicated in skeletal muscle regeneration (Bryan et al., 2005; Hager et al., 2010; Smith et al., 2008; Tajika et al., 2007). Thus, apparently members of the Popdc gene family have essential functions in the development and regeneration of skeletal muscle.
Many popdc2 morphants displayed a cardiac conduction defect characterized by 2:1 or 3:1 AV block. In addition, variability in action potential duration in the atrium was observed and in rare cases a sinoatrial block was seen. At later stages of development most of the injected embryos had a silent heart phenotype leading ultimately to the death of the animals. The cardiac electrical phenotype was not observed during early heart development and was first seen in some embryos at 3 dpf. The electrical phenotype became more robustly visible at 4–5 dpf. Recently, it has been proposed that in zebrafish the development of the cardiac conduction system consists of three developmental steps (Chi et al., 2008). In the first step, primary pacemaker development is accomplished at 24 hpf. The atrioventricular delay of electrical conduction develops at 36 to 48 hpf, and at 96 hpf the fast ventricular conduction system becomes fully functional. Cardiac electrical activity is however subject to further steps of maturation such that the QRS interval time is continuously decreasing during larval development (Yu et al., 2010). Maturation of the cardiac electrical system is probably accomplished by anatomical changes and improved electrical coupling between myocytes but may also relate to changes in expression level, membrane localization and complex formation of ion channels and regulatory proteins, which are involved in action potential generation. Whether the defects seen here in the popdc2 morphants are caused by aberrant conduction tissue development or are due to an essential function of popdc2 in action potential generation needs to be further studied. The AV block, which is observed in many popdc2 morphants resembles the phenotype of the breakdance mutant (Kopp et al., 2005), which is defective in the Kcnh2 potassium channel involved in cardiac repolarization (Arnaout et al., 2007; Langheinrich et al., 2003). The variability in the peak intervals in the popdc2 morphant can be explained on the basis of a defect in cardiac repolarization. The mouse Popdc proteins are localized to the plasma membrane in cardiac myocytes and an interaction with TREK-1, a member of the tandem pore family of potassium channels has been observed (Froese et al., 2012). Therefore it will be important to define whether also in case of the zebrafish protein an interaction with ion channels can be demonstrated. Significantly the Popdc1 and Popdc2 null mutants in mice display cardiac arrhythmia (Froese et al., 2012), suggesting an evolutionary conserved function in cardiac electrical activity. Novel molecular insight may also come from studying the electrophysiological properties of cardiac myocytes isolated from popdc2 morphants. It will also be important to study the phenotype of popdc1 and popdc3 morphants in zebrafish. It is likely that cardiac and skeletal muscle phenotypes are also present in these morphants given that in all vertebrates studied so far, the Popdc genes are expressed in muscle and heart (Andrée et al., 2000; Breher et al., 2004; Froese and Brand, 2008; Hitz et al., 2002; Parnes et al., 2007; Torlopp et al., 2006).
Gene defects in ion channel genes causing cardiac conduction defects such as long QT syndrome in humans have been successfully modeled in zebrafish through mutagenesis or morpholino-mediated suppression of the corresponding homologs (Morita et al., 2008). The phenotypes of those mutants resembled to a large extent the conduction defects observed in patients and this is probably due to the fact that man and zebrafish have similar heart rates and utilize similar sets of ion channels for action potential generation (Milan and Macrae, 2008). Therefore the here observed conduction defects in zebrafish could be taken as an indication for a possible association of POPDC2 mutations with cardiac conduction defects in patients.
The expert technical assistance of Reinhild Fischer, Elisabeth Meyer-Natus, and Anneliese Striewe-Conz at the University of Würzburg and Ursula Herbort-Brand at the Heart Science Centre, Imperial College London is gratefully acknowledged. We thank Holger Dill, Utz Fischer and Stefan Schulte-Merker for their support. This research was financially supported by grants from the German Research Foundation (DFG): BR1218/9-4 (T.B.) and GK 1048 (T.B.), the European Community’s Sixth Framework Program contract (Heart Repair) LSHM-CT-2005-018630 (T.B), the NIH (HL54737) (D.Y.R.S.), travel stipends of the Company of Biologists (B.C.K.), and EMBO (ASTF-96-2007; B.C.K.), and the Magdi Yacoub Institute (T.B.). The antibodies F59, F310, Zn8 were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biology, Iowa City, IA 52242.
Supplementary materials related to this article can be found online at doi:10.1016/j.ydbio.2012.01.015.