Coronary vessel development is a highly coordinated process during heart formation. Abnormal development and dysfunction of the coronary network are contributory factors in the majority of heart disease. Understanding the molecular mechanisms that regulate coronary vessel formation is crucial for preventing and treating the disease. We report a zebrafish gene-trap vinculin b (vclb) mutant that displays abnormal coronary vessel development among multiple cardiac defects. The mutant shows overproliferation of epicardium-derived cells and disorganization of coronary vessels, and they eventually die off at juvenile stages. Mechanistically, Vclb deficiency results in the release of another cytoskeletal protein, paxillin, from the Vclb complex and the upregulation of ERK and FAK phosphorylation in epicardium and endocardium, causing disorganization of endothelial cells and pericytes during coronary vessel development. By contrast, cardiac muscle development is relatively normal, probably owing to redundancy with Vcla, a vinculin paralog that is expressed in the myocardium but not epicardium. Together, our results reveal a previously unappreciated function of vinculin in epicardium and endocardium and reinforce the notion that well-balanced FAK activity is essential for coronary vessel development.

The bulk of the vertebrate heart consists of cardiomyocytes and fibroblasts. In addition, the heart contains endothelial cells, which constitute the endocardium and coronary vessels, and epicardial cells. Heart development is a highly coordinated sequential process that includes heart field induction, cardiac cell migration, heart tube formation, cardiac looping, chamber specification and maturation, and coronary vessel formation (Harrison et al., 2015; Staudt and Stainier, 2012). Towards the end of cardiac development, the newly formed coronary vessels supply oxygen and nutrients to the inner cardiomyocytes, a process that is essential for survival (Harrison et al., 2015; Reese et al., 2002). Coronary diseases are a leading cause of death in industrialized countries, and understanding the development and physiology of coronary vessels is crucial in their prevention and treatment. Coronary vessel development is proposed to involve both vasculogenesis and angiogenesis (Munoz-Chapuli et al., 2002; Olivey et al., 2004). Previous studies in mouse and chicken showed that precursors of coronary vessels derive from the sinus venous, endocardium and epicardium (Dettman et al., 1998; Red-Horse et al., 2010; Vrancken Peeters et al., 1999; Wu et al., 2012). A recent report demonstrated that coronary vessels originate from endocardium-derived arterial cells by angiogenic sprouting, and identified Cxcr4a-Cxcl12b as a key signaling axis for this process in zebrafish (Harrison et al., 2015). However, the molecular mechanisms that regulate coronary vessel formation and patterning are still not well understood.

Formation of the complex coronary network requires highly coordinated cell-to-cell and cell-to-extracellular matrix (ECM) interactions, in which the cytoskeleton is essential for the transduction of cell force and of extracellular signals into cells. The cytoskeletal protein vinculin is localized in the integrin-mediated cell-ECM adhesions and cadherin-mediated cell-cell junctions (Geiger, 1979; Lifschitz-Mercer et al., 1997; Raz and Geiger, 1982). Vinculin mediates force transmission from cell matrix to actin cytoskeleton by interacting with other cytoskeletal proteins, such as talin, actinin and β-catenin (Zamir and Geiger, 2001; Ziegler et al., 2006). In addition, vinculin integrates extracellular signals essential for cellular survival and processes including differentiation, apoptosis and locomotion (Carisey and Ballestrem, 2011). Coordination of these processes is crucial for embryonic development and tissue homeostasis. The human vinculin (VCL) gene contains 22 exons (Moiseyeva et al., 1993). Exon 19 can be alternatively spliced, resulting in two protein isoforms: VCL and metaVCL (Koteliansky et al., 1992). VCL is ubiquitously expressed, whereas metaVCL, the larger isoform containing an additional 68 amino acid residues from exon 19, is expressed exclusively in cardiac and smooth muscle (Belkin et al., 1988). Mutations in VCL have been identified in human dilated cardiomyopathy (DCM) patients (Olson et al., 2002). In mice, vinculin (Vcl) gene deletion is lethal during gestation, with abnormal neural tube formation and cardiac defects (Xu et al., 1998a). However, mice with a cardiomyocyte-specific deletion of Vcl survive to adulthood. In the first 3 months, the mutant mice show 50% mortality and those that do survive longer exhibit dilated cardiomyopathy (Zemljic-Harpf et al., 2007). These results suggest that vinculin exerts important functions in cells other than cardiomyocytes. It remains incompletely understood how vinculin regulates heart development and function.

Another protein involved in cardiovascular development is focal adhesion kinase (FAK). FAK is ubiquitously expressed and is involved in cell survival and motility by mediating integrin, growth factor, and mechanical stress signaling (Mitra et al., 2005; Parsons, 2003). FAK is highly expressed in the developing vasculature (Polte et al., 1994). Consistently, Fak (Ptk2) null mouse embryos die at E8.5 with multiple defects, including a disorganized cardiovascular system (Ilić et al., 1995). FAK controls cell survival and the motility of endothelial cells and smooth muscle cells during angiogenesis (Braren et al., 2006; Hauck et al., 2000; Peng et al., 2004; Shen et al., 2005). In addition, FAK (Ptk2ab) regulates coronary vessel formation during heart regeneration in adult zebrafish (Missinato et al., 2015). Localized to focal adhesions, FAK interacts with integrin-associated proteins, such as paxillin and talin, and can elicit extracellular signal-regulated kinase (ERK) signaling (Fincham et al., 2000; Hauck et al., 2002; Klemke et al., 1997). Interaction between FAK and the adaptor protein paxillin is crucial for activation of signaling cascades required for cell survival and motility (Turner, 2000). In Vcl knockout mice, FAK activity is upregulated (Xu et al., 1998a). In cultured mouse Vcl null cells, paxillin and FAK interaction is enhanced and FAK phosphorylation is upregulated (Subauste et al., 2004), suggesting that vinculin regulates cell survival and motility by modulating paxillin-FAK interaction. Nevertheless, cellular functions of vinculin remain to be fully elucidated during heart development.

Here, we report a zebrafish gene-trap mutant v12, in which the vclb gene is disrupted by a Tol2 transposon insertion. We found that vclb plays crucial roles during coronary vessel development. Vclb is enriched in the epicardium and endocardium and regulates the formation of the coronary vessel network by fine-tuning the phosphorylation of FAK and ERK. Our study in zebrafish uncovered previously unappreciated functions of vinculin in coronary vessel development, and might have implications regarding therapeutic intervention of certain subtypes of coronary heart disease.

The zebrafish V12 gene-trap line displays specific GFP expression in the heart

In a Tol2 transposon-mediated gene-trap screen in zebrafish, we identified the V12 line, in which the EGFP reporter shows distinct expression patterns in various tissues, including the heart. Specifically, EGFP expression initiates in the somite septum at 24 h post fertilization (hpf). At 35 hpf, EGFP is detected in the heart. From 50 hpf on, EGFP is prominently localized in the ventricle (Fig. 1A-H). Colocalization of EGFP with β-catenin but not with α-actinin (a sarcomere Z-disc marker) suggests that EGFP expression is close to the cell membrane (Fig. 1I-J″). Intercross between heterozygous fish (F1) yielded 908 F2 embryos, which were divided into three groups based on EGFP intensity (no, 23.6%; weak, 52.2%; strong, 24.2%) indicating that alleles of a single locus in the V12 line are segregated at Mendelian ratio. Accordingly, the no, weak and strong EGFP F2 families are likely to represent wild type, heterozygote and homozygote (hereafter referred to as the v12 mutant), respectively. Strong EGFP expression in hearts suggests involvement of the trapped gene during heart development.

Fig. 1.

GFP is highly expressed in the ventricle of v12 zebrafish embryos. (A-D′) GFP is not detected in 75% epiboly embryos (A,A′), and is initially expressed in the somite boundaries at ∼24 hpf (B). Cardiac GFP expression starts at ∼35 hpf (C,C′) and GFP is highly expressed in the ventricle at 50 hpf (D,D′). Arrows indicate ventricle. (E-H) Ventricular GFP expression in v12 heterozygotes (E-F) and v12 homozygotes (G-H) at 84 hpf shown at high magnification. Arrows indicate ventricle. (I-I″) Immunostaining shows that β-catenin and GFP colocalize close to the cell membrane in v12 heterozygous hearts at 78 hpf. (J-J″) Immunostaining shows that GFP is absent from Z-disc (α-actinin) in v12 heterozygous hearts at 78 hpf. Scale bars: 100 μm in A-D,E-H; 20 μm in I; 5 μm in J.

Fig. 1.

GFP is highly expressed in the ventricle of v12 zebrafish embryos. (A-D′) GFP is not detected in 75% epiboly embryos (A,A′), and is initially expressed in the somite boundaries at ∼24 hpf (B). Cardiac GFP expression starts at ∼35 hpf (C,C′) and GFP is highly expressed in the ventricle at 50 hpf (D,D′). Arrows indicate ventricle. (E-H) Ventricular GFP expression in v12 heterozygotes (E-F) and v12 homozygotes (G-H) at 84 hpf shown at high magnification. Arrows indicate ventricle. (I-I″) Immunostaining shows that β-catenin and GFP colocalize close to the cell membrane in v12 heterozygous hearts at 78 hpf. (J-J″) Immunostaining shows that GFP is absent from Z-disc (α-actinin) in v12 heterozygous hearts at 78 hpf. Scale bars: 100 μm in A-D,E-H; 20 μm in I; 5 μm in J.

The vinculin gene (vclb) is disrupted in the V12 line

To identify the trapped gene, we carried out thermal asymmetric interlaced polymerase chain reactions (TAIL-PCR) to amplify the flanking genomic sequences of the Tol2 insertion site, as previously reported (Liu and Chen, 2007). The flanking sequences were mapped to an unreported sequence in chromosome 12, which encodes a polypeptide with high homology to human VCL protein (Fig. 2A, Fig. S1). There are two vinculin genes in the zebrafish genome (Pascoal et al., 2013). vcla is located on chromosome 13 and its full-length cDNA has been cloned (Vogel et al., 2009), whereas the sequence of vclb is unknown. We cloned the full-length vclb cDNA, which encodes a polypeptide of 1065 amino acid residues. The vclb RNA expression pattern resembles that of the EGFP reporter in the trap line (Fig. 2B). Unlike vcla, vclb shows no alternative splicing around the region corresponding to exon 19 in vcla (Fig. 2C). The EGFP-vclb fusion transcript from the v12 mutant encodes a chimeric protein with EGFP fused to the N-terminal 261 amino acids of Vclb (NVclb, Fig. 2D). A protein of predicted size (55 kDa) was detected by western blot using a GFP antibody (Fig. 2E). RT-PCR genotyping of embryos of different EGFP intensity confirmed their respective genotypes (Fig. 2F). RT-PCR and qRT-PCR results show that the full-length vclb mRNA is almost undetectable, whereas vcla is not affected in the homozygous mutant (Fig. 2F,G). Together, these results indicate that vclb is disrupted in v12 mutants.

Fig. 2.

The vclb gene is disrupted in the v12 mutant. (A) The zebrafish vclb gene is a homolog of human vinculin (VCL). (B) vclb mRNA expression as revealed by in situ hybridization is similar to EGFP expression patterns in the somites at 24 hpf and the heart at 48 hpf. Arrow points to the heart. (C) vclb is an alternative transcript corresponding to the vcla transcript variant without exon 19. No other alternative splicing transcripts are detected. odc (ornithine decarboxylase 1) is a loading control. RT–, without reverse transcriptase. (D) Diagram indicating that the transposon insertion in the v12 mutant causes fusion of EGFP with the N-terminal 261 amino acids of Vclb (NVclb). (E) NVclb-EGFP fusion protein expression detected with a GFP antibody in heterozygous and homozygous v12 mutant heart at 30 dpf. Size markers (kDa) are shown to the right. Gapdh provides a loading control. (F,G) Intact vclb, but not vcla, mRNA expression is blocked in the v12 mutant at 36 hpf as assayed by RT-PCR (F) and qRT-PCR (G). The expression level of vcla is the same among wild type, heterozygotes and v12 mutant. The qRT-PCR data are presented as mean±s.e.m. Two-tailed unpaired t-test; NS, not significant; ***P<0.001.

Fig. 2.

The vclb gene is disrupted in the v12 mutant. (A) The zebrafish vclb gene is a homolog of human vinculin (VCL). (B) vclb mRNA expression as revealed by in situ hybridization is similar to EGFP expression patterns in the somites at 24 hpf and the heart at 48 hpf. Arrow points to the heart. (C) vclb is an alternative transcript corresponding to the vcla transcript variant without exon 19. No other alternative splicing transcripts are detected. odc (ornithine decarboxylase 1) is a loading control. RT–, without reverse transcriptase. (D) Diagram indicating that the transposon insertion in the v12 mutant causes fusion of EGFP with the N-terminal 261 amino acids of Vclb (NVclb). (E) NVclb-EGFP fusion protein expression detected with a GFP antibody in heterozygous and homozygous v12 mutant heart at 30 dpf. Size markers (kDa) are shown to the right. Gapdh provides a loading control. (F,G) Intact vclb, but not vcla, mRNA expression is blocked in the v12 mutant at 36 hpf as assayed by RT-PCR (F) and qRT-PCR (G). The expression level of vcla is the same among wild type, heterozygotes and v12 mutant. The qRT-PCR data are presented as mean±s.e.m. Two-tailed unpaired t-test; NS, not significant; ***P<0.001.

vclb deletion causes lethality at juvenile stages

The viability of v12 mutants is comparable to that of wild-type and heterozygous embryos for the first 21 days of development. After 21 days post fertilization (dpf), the mutants die off and few survive beyond 2 months (Fig. S2A). The overall morphology of the mutants is indistinguishable from that of the wild-type or heterozygous siblings until 3.5 dpf, when pericardiac edema is obvious in the mutant (Fig. S2B,C). At 30 dpf, v12 mutants exhibit a smaller body size (about two-thirds the size of control siblings, Table S1) with craniofacial malformations and pleural effusion (Fig. S2D,E). Adult vclb heterozygotes are viable, fertile and indistinguishable from wild type, suggesting that the NVclb-EGFP fusion does not affect the wild-type allele through a dominant-negative effect.

To confirm that vclb deficiency causes the observed defects, we used CRISPR-Cas9 technology (Chang et al., 2013; Hwang et al., 2013) and generated a new vclb mutant with a 17 bp deletion comprising nucleotides 51 to 67 downstream of the translation initiation site (Fig. S3A). The resulting vclbdel17 mutants show similar cardiac defects to v12 mutants (Fig. S3B-O′). Furthermore, the new mutant allele failed to compensate the vclb allele in the V12 line. The v12/vclbdel17 double heterozygotes exhibit the same cardiac defects as the individual mutants (data not shown). Together, these results confirm that vclb mutation causes the v12 mutant phenotypes.

vclb regulates epicardial proliferation and differentiation in the ventricles

Pericardiac edema and pleural effusion in the v12 mutant indicate that cardiac defects occur at juvenile stages. In the v12 mutant, vclb mRNA is greatly reduced in the heart during embryonic stages (Fig. S4A-D). However, the cmlc2 (myl7) and nkx2.5 genes, which are involved in cardiac progenitor specification, have comparable expression levels in control siblings and v12 mutant embryos (Fig. S4E-H). During ventricle morphogenesis, the expression levels of vmhc and nppa in v12 mutants are also comparable to those in control siblings (Fig. S4I-P). Early development of endocardial and epicardial cells appears normal in the v12 mutant, as indicated by a Tg(flk1:mcherry) (flk1 is also known as kdrl) transgenic reporter (Dong et al., 2012) and tbx18 RNA in situ hybridization (Begemann et al., 2002) (Fig. S5). These results suggest that the early specification of cardiac progenitors and the initial development of cardiac chambers are largely normal and that the lethality is due to subsequent chamber differentiation and maturation defects in v12 mutants.

Next, we examined hearts from v12 mutants and control siblings at different larval stages. The mutant ventricles change from pyramid-like, as in control siblings, to spherical, resembling those from human DCM patients (Fig. 3A,B). The mutant hearts retain a relatively normal sarcomere structure, but gaps can be seen between some muscle fibers (Fig. S6A-D). In addition, the density of cardiomyocytes increases in the mutant, as indicated by Mef2 (cardiomyocyte nuclear specific) antibody staining (Fig. S6E,F), indicating cardiomyocyte proliferation. Sub-epicardium hemorrhage occurred in ∼30% of v12 mutant ventricles (Fig. 3C,D, Fig. S7). Moreover, at 30 dpf, epicardial cells in v12 mutants exhibit cobblestone-like shapes (Fig. 3F,H) that clearly differ from the squamous epithelium of serrated cell membrane morphology in wild-type siblings (Fig. 3E,G). In wild-type ventricles, only a single layer of epicardial cells expressing Aldh1a2 exists (Fig. 3I). By contrast, in the v12 mutant, in addition to one layer of Aldh1a2-positive epicardial cells, two to three extra layers of vimentin-positive cells fill in the sub-epicardial space (Fig. 3E,F,I-L′). We also used a Tg(tcf21:dsRed) transgene (Kikuchi et al., 2011) to label the epicardial lineage and found that these extra cells were DsRed-positive but negative for cardiac myosin heavy chain as detected by MF20 antibody staining, suggesting that they were fibroblasts derived from epicardium (Fig. 3I,J,M-N′). An overall increase in Pcna-positive cells is obvious in the v12 mutant at 15 dpf compared with wild type, but the increase is especially pronounced in the epicardium, confirming the overproliferation observation (Fig. S8). We conclude that Vclb deficiency causes overproliferation and differentiation of epicardial-derived cells.

Fig. 3.

Overproliferation of Tcf21-positive epicardial-derived cells in v12 mutants. (A,B) Ventricular morphology (dorsal view) of the v12 mutant and a wild-type sibling at 30 dpf. Note the spherical ventricles in the v12 mutant (B) as compared with the pyramid-like heart in wild type (A). (C-F) Histology of the v12 mutant (D,F) and its wild-type sibling (C,E), showing the malformed heart and sub-epicardial hemorrhage (arrow) in the mutant. Epicardial cells (epi) change from wild-type squamous epithelium (E) to oval-shaped in the v12 mutant (F). (G,H) β-catenin immunostaining marks the cell membrane of epicardial cells. Cell size is reduced and the cell boundary is smoother in the mutant (H) compared with the wild-type sibling (G). (I,J) MF20 staining shows that the overproliferated cells in the sub-epicardium of the v12 mutant are negative for muscle-specific myosin heavy chain. (K-L′) The overproliferated cells are positive for vimentin (a marker of mesenchymally derived cells) in the v12 mutant (L,L′) compared with its control sibling (K,K′). Aldhla2 marks the epicardium. (M-N′) Aldh1a2 colocalizes with Tg(tcf21:dsRed) transgenic reporter in the epicardial cells of the wild-type sibling at 30 dpf (M,M′). By contrast, both overproliferated cells in the sub-epicardium region and the outermost layer of epicardial cells are positive for DsRed staining. Only the outermost layer of epicardial cells shows strong Aldh1a2 staining in the mutant (N,N′). A, atrium; V, ventricle. Scale bars: 100 μm in A-D; 20 μm in E-N.

Fig. 3.

Overproliferation of Tcf21-positive epicardial-derived cells in v12 mutants. (A,B) Ventricular morphology (dorsal view) of the v12 mutant and a wild-type sibling at 30 dpf. Note the spherical ventricles in the v12 mutant (B) as compared with the pyramid-like heart in wild type (A). (C-F) Histology of the v12 mutant (D,F) and its wild-type sibling (C,E), showing the malformed heart and sub-epicardial hemorrhage (arrow) in the mutant. Epicardial cells (epi) change from wild-type squamous epithelium (E) to oval-shaped in the v12 mutant (F). (G,H) β-catenin immunostaining marks the cell membrane of epicardial cells. Cell size is reduced and the cell boundary is smoother in the mutant (H) compared with the wild-type sibling (G). (I,J) MF20 staining shows that the overproliferated cells in the sub-epicardium of the v12 mutant are negative for muscle-specific myosin heavy chain. (K-L′) The overproliferated cells are positive for vimentin (a marker of mesenchymally derived cells) in the v12 mutant (L,L′) compared with its control sibling (K,K′). Aldhla2 marks the epicardium. (M-N′) Aldh1a2 colocalizes with Tg(tcf21:dsRed) transgenic reporter in the epicardial cells of the wild-type sibling at 30 dpf (M,M′). By contrast, both overproliferated cells in the sub-epicardium region and the outermost layer of epicardial cells are positive for DsRed staining. Only the outermost layer of epicardial cells shows strong Aldh1a2 staining in the mutant (N,N′). A, atrium; V, ventricle. Scale bars: 100 μm in A-D; 20 μm in E-N.

vclb deficiency causes coronary vessel disorganization

Epicardial cells are involved in coronary vessel development (Dettman et al., 1998; Reese et al., 2002; Vrancken Peeters et al., 1999). Overproliferation of epicardial-derived cells in the v12 mutant could lead to abnormal coronary vessel development. To better examine possible roles of vclb in coronary vessel development, we first investigated its expression pattern in v12 and sibling hearts using the NVclb-EGFP fusion reporter. NVclb-EGFP is ubiquitously distributed on the ventricle surface (epicardium) and EGFP-positive cells are most abundant in plexus-like structures and in the sub-epicardial region. The EGFP signal is particularly elevated in the mutant (0-9.2 μm from the surface) compared with heterozygotes (0-4.6 μm from the surface) (Fig. 4A-D). In the inner myocardium, EGFP expression is located at the boundary between endocardial cells and muscle cells in both mutants (13.8 μm from the surface) and heterozygotes (9.8 μm from the surface) (Fig. 4E,F). Expression of NVclb-EGFP in epicardial cells was also confirmed by co-immunostaining of myosin heavy chain (MF20) and EGFP in cryosections of ventricles from heterozygotes and v12 mutants (Fig. 4G-H‴). Expression of NVclb-EGFP at the muscle-endocardium boundary was also shown by co-immunostaining of mCherry and EGFP in cryosections of ventricles from heterozygotes and v12 mutants in the Tg(flk1:mcherry) background (Fig. 4I-J′).

Fig. 4.

Ventricular expression patterns of the NVclb-EGFP fusion protein. (A,B) GFP expression in the ventricle surface (0 μm) at 52 dpf, showing the formation of plexus-like structures in the v12 mutant (B) compared with heterozygotes (A). (C,D) GFP expression patterns in the sub-epicardium of ventricles at 4.6 μm (C) and 9.2 μm (D) from the surface. Strong expression of GFP fusion protein is shown in the mutant (D) compared with heterozygotes (C). (E,F) GFP is detected at the boundary between endothelial cells and muscles of ventricles at 9.8 μm (E) and 13.8 μm (F) from the surface. (G-H‴) GFP is ubiquitously expressed in the ventricle. GFP expression is detected in several layers of GFP+/MF20 epicardium-derived cells in the mutant (H″,H‴) compared with a single layer of epicardial cells in heterozygotes (G″,G‴). (I-J′) Tg(flk1:mcherry) transgenic reporter is used to label endothelial cells. GFP is detected at the boundary between endothelial cells (red) and myocardium. endo, endocardium; epi, epicardium; myo, myocardium. Scale bars: 10 μm in A-F,I-J′; 20 μm in G,G″,H,H″.

Fig. 4.

Ventricular expression patterns of the NVclb-EGFP fusion protein. (A,B) GFP expression in the ventricle surface (0 μm) at 52 dpf, showing the formation of plexus-like structures in the v12 mutant (B) compared with heterozygotes (A). (C,D) GFP expression patterns in the sub-epicardium of ventricles at 4.6 μm (C) and 9.2 μm (D) from the surface. Strong expression of GFP fusion protein is shown in the mutant (D) compared with heterozygotes (C). (E,F) GFP is detected at the boundary between endothelial cells and muscles of ventricles at 9.8 μm (E) and 13.8 μm (F) from the surface. (G-H‴) GFP is ubiquitously expressed in the ventricle. GFP expression is detected in several layers of GFP+/MF20 epicardium-derived cells in the mutant (H″,H‴) compared with a single layer of epicardial cells in heterozygotes (G″,G‴). (I-J′) Tg(flk1:mcherry) transgenic reporter is used to label endothelial cells. GFP is detected at the boundary between endothelial cells (red) and myocardium. endo, endocardium; epi, epicardium; myo, myocardium. Scale bars: 10 μm in A-F,I-J′; 20 μm in G,G″,H,H″.

We then examined whether vclb plays a role in coronary vessel development using the Tg(fli1a:EGFP) transgenic reporter to label coronary endothelial cells (Roman et al., 2002). Owing to much stronger signals of Tg(fli1a:EGFP) than of NVclb-EGFP (Fig. S9), we were able to follow coronary vessels by Tg(fli1a:EGFP) in the compound Tg(fli1a:EGFP); v12(NVclb-EGFP) zebrafish.

Similar to the previous mouse study (Red-Horse et al., 2010), in zebrafish the coronary vessels first appear on the dorsal surface of the ventricles near the atrium-ventricle connection at ∼28 dpf (Fig. 5A,E). The coronary vessels then spread over the entire dorsal surface and extend gradually to cover the ventral surface from 35 to 55 dpf (Fig. 5B-D,F-H). New vessels initially appear as isolated blood plexuses that subsequently connect to each other to form a network. This is similar to zebrafish coronary vessel development as described by Harrison et al. (2015). Compared with wild-type siblings, the v12 mutant has more blood plexuses and vessels at initiation stages (Fig. 5I,M). The blood vessel density in v12 mutants is higher than that in the control siblings (Fig. S10E). However, v12 mutants show abundant but disorganized vessels in the ventricle when the wild-type siblings have already formed well-connected coronary vasculatures (Fig. 5J-L,N-P). Blood cells can be seen in some of the coronary vessels of the v12 mutant (Fig. S11). At 55 dpf, co-immunostaining with EGFP and MF20 antibodies showed that Fli1a-EGFP-positive coronary vessels are in contact with the myocardium in the wild-type ventricle, evidently in the compact myocardium (Fig. 5Q,Q′, arrows). By contrast, coronary vessels in the mutants are surrounded by extra epicardial-derived cells in the sub-epicardial space and most of the vessels are not in contact with cardiac muscle (Fig. 5R,R′, arrows). Together, these results strongly suggest that vclb deficiency causes disorganization of the coronary vessels and that abnormal epicardial cell proliferation/differentiation contributes to the coronary vessel defects in v12 mutants.

Fig. 5.

vclb deficiency causes disorganization of coronary vessels. The Tg(fli1a:EGFP) transgenic line (endothelial cells labeled by EGFP) is used to trace coronary vessel development. The Tg(tcf21:dsRed) transgenic line is used to label epicardial cells. (A-H) In wild-type hearts, coronary vessels first appear on the dorsal surface of the ventricle and are close to the atrium-ventricle connection at ∼28 dpf. Coronary vessels then spread, extend, and wrap the entire ventricle wall from 35 to 55 dpf. Dorsal views (A-D) and ventral views (E-H) are shown. (I-P) In the v12 mutant, coronary vessels appear at the same place as in wild-type siblings, but more vessels are detected at the initial stages. Wild-type coronary vessels are well connected to each other, but the overproliferated mutant vessels remain isolated and disorganized. Dorsal views (I-L) and ventral views (M-P) are shown. (Q-R′) In the v12 mutant, coronary vessels are surrounded by epicardial-derived cells (R,R′) rather than by cardiac muscles as in the wild-type sibling (Q,Q′) at 55 dpf. MF20 stains cardiac muscles. Arrows indicate vessels that are (Q,Q′) or are not (R,R′) in contact with cardiac muscle. (S-V″) At 40 dpf, Tg(fli1:EGFP)-labeled endothelial tubes in wild-type ventricles are well aligned with Tg(tcf21:dsRed)-labeled epicardial-derived perivascular cells (S,S′,U-U″). By contrast, most of the endothelial tubes in v12 mutants are not well aligned with perivascular cells (T,T′,V-V″). Note the overproliferated endothelial and perivascular cells in the mutant. Results are shown in whole-mount view (S-T′) and cryosection (U-V″). Scale bars: 100 μm in A-P; 10 μm in Q-R′; 20 μm in S-V″.

Fig. 5.

vclb deficiency causes disorganization of coronary vessels. The Tg(fli1a:EGFP) transgenic line (endothelial cells labeled by EGFP) is used to trace coronary vessel development. The Tg(tcf21:dsRed) transgenic line is used to label epicardial cells. (A-H) In wild-type hearts, coronary vessels first appear on the dorsal surface of the ventricle and are close to the atrium-ventricle connection at ∼28 dpf. Coronary vessels then spread, extend, and wrap the entire ventricle wall from 35 to 55 dpf. Dorsal views (A-D) and ventral views (E-H) are shown. (I-P) In the v12 mutant, coronary vessels appear at the same place as in wild-type siblings, but more vessels are detected at the initial stages. Wild-type coronary vessels are well connected to each other, but the overproliferated mutant vessels remain isolated and disorganized. Dorsal views (I-L) and ventral views (M-P) are shown. (Q-R′) In the v12 mutant, coronary vessels are surrounded by epicardial-derived cells (R,R′) rather than by cardiac muscles as in the wild-type sibling (Q,Q′) at 55 dpf. MF20 stains cardiac muscles. Arrows indicate vessels that are (Q,Q′) or are not (R,R′) in contact with cardiac muscle. (S-V″) At 40 dpf, Tg(fli1:EGFP)-labeled endothelial tubes in wild-type ventricles are well aligned with Tg(tcf21:dsRed)-labeled epicardial-derived perivascular cells (S,S′,U-U″). By contrast, most of the endothelial tubes in v12 mutants are not well aligned with perivascular cells (T,T′,V-V″). Note the overproliferated endothelial and perivascular cells in the mutant. Results are shown in whole-mount view (S-T′) and cryosection (U-V″). Scale bars: 100 μm in A-P; 10 μm in Q-R′; 20 μm in S-V″.

In zebrafish, coronary endothelial cells are covered by pericyte-like mural cells and myosin light chain kinase-positive smooth muscle cells (González-Rosa et al., 2011; Hu et al., 2001; Kim et al., 2010). Previous studies showed that Tcf21-positive epicardial cells contribute to progenitors of coronary vessel perivascular cells in zebrafish (Kikuchi et al., 2011). To test whether the overproliferation of epicardial cells affects interactions between coronary endothelial cells and perivascular cells, we used Tg(fli1a:EGFP) and Tg(tcf21:DsRed) transgenes to label endothelial (EGFP) and epicardial (dsRed) cells (Kikuchi et al., 2011), respectively. Unlike wild-type embryos, in which endothelial cells are well aligned with dsRed-positive cells, most v12 mutants show disorganization of increased numbers of perivascular cells, probably owing to mispatterned coronary vessels (Fig. 5S-V″). From these results, we conclude that vclb deficiency causes disorganization of the coronary vasculature. Moreover, vclb deficiency causes overproliferation of epicardial-derived cells and the disorganization of endothelial cells and perivascular cells.

Vclb deficiency activates ERK and FAK in epicardium and endocardium

The overproliferation of epicardial cells and endothelial cells in v12 mutant hearts indicates that they are mitotically active. There are a number of mechanisms underlying cell proliferation. For example, growth factor-activated cell proliferation can be mediated by ERK (mitogen-activated protein kinase, MAPK) activity (Seger and Krebs, 1995). We examined ERK activation in zebrafish heart using a phosphorylation-specific ERK antibody. We found a stronger phosphorylated (p) Erk1/2 (Mapk3/1) signal in coronary endothelial cells than in the endocardium of wild type at 55 dpf (Fig. 6A,A′), suggesting that relatively high ERK activity in coronary endothelial cells might be required for coronary vessel development. When treated with 10 μM MEK1 inhibitor U0126 from 28 dpf, coronary vessel development was significantly impaired in the wild-type siblings at 40 dpf (Fig. S10E). pErk1/2 levels were markedly increased in the epicardium and endocardium in v12 mutants compared with control siblings at 15 and 55 dpf (Fig. 6A-F′). It should be noted that ERK activation occurs at 15 dpf, long before the formation of coronary vessels, indicating that the activation is not a result of myocardial hypoxia. This was confirmed by examining RNA expression of hypoxia markers hypoxia-inducible factor 1a (hif1a) and egl-9 family hypoxia-inducible factor 3 (egln3, the zebrafish homolog of prolyl hydroxylase 3) (Kaelin and Ratcliffe, 2008; Manchenkov et al., 2015) in v12 and control sibling hearts (Fig. S12).

Fig. 6.

pERK and pFAK are upregulated in the epicardium and endocardium of v12 mutant ventricles. (A-B′) pErk1/2 expression (green) is increased in the endocardium of v12 mutants at 55 dpf compared with wild-type sibling. (C-D′) pErk1/2 expression is increased in the epicardium and endocardium of v12 mutant at 15 dpf. MF20 staining (red) indicates cardiac muscles. (E-F′) Tg(tcf21:dsRed) transgenic reporter indicates epicardium. Upregulated pErk1/2 colocalizes with RFP signal. (G) Co-immunoprecipitation shows that NVclb-EGFP is not associated with paxillin, which normally interacts with wild-type Vclb, when ectopically expressed in HEK293 cells. IP, immunoprecipitation; WB, western blot. Size markers (kDa) are shown to the left. (H-I′) The localization of paxillin at the boundary between epicardium and myocardium is partially lost in the v12 mutant (L,L′) compared with wild-type siblings (H,H′). (J-M) pFAK(Y576/577) expression is upregulated whereas total FAK expression remains unchanged in the epicardium and endocardium of the v12 mutant at 45 dpf (L,M) compared with wild-type sibling (J,K). Scale bars: 20 μm.

Fig. 6.

pERK and pFAK are upregulated in the epicardium and endocardium of v12 mutant ventricles. (A-B′) pErk1/2 expression (green) is increased in the endocardium of v12 mutants at 55 dpf compared with wild-type sibling. (C-D′) pErk1/2 expression is increased in the epicardium and endocardium of v12 mutant at 15 dpf. MF20 staining (red) indicates cardiac muscles. (E-F′) Tg(tcf21:dsRed) transgenic reporter indicates epicardium. Upregulated pErk1/2 colocalizes with RFP signal. (G) Co-immunoprecipitation shows that NVclb-EGFP is not associated with paxillin, which normally interacts with wild-type Vclb, when ectopically expressed in HEK293 cells. IP, immunoprecipitation; WB, western blot. Size markers (kDa) are shown to the left. (H-I′) The localization of paxillin at the boundary between epicardium and myocardium is partially lost in the v12 mutant (L,L′) compared with wild-type siblings (H,H′). (J-M) pFAK(Y576/577) expression is upregulated whereas total FAK expression remains unchanged in the epicardium and endocardium of the v12 mutant at 45 dpf (L,M) compared with wild-type sibling (J,K). Scale bars: 20 μm.

Previous studies indicated that loss of vinculin enhances the interaction between paxillin and FAK, which results in upregulation of FAK phosphorylation and in turn activates ERK in mouse F9 embryonic carcinoma cells (Subauste et al., 2004; Xu et al., 1998a). We investigated whether similar events occur in v12 mutant zebrafish hearts, especially in coronary and endocardial endothelial cells. The major paxillin-binding motif of vinculin is known to be located in its C-terminal region (Wood et al., 1994). Indeed, when expressed in HEK293 cells, the remaining NVclb-EGFP fusion protein is not associated with zebrafish paxillin (Fig. 6G), suggesting that paxillin is released from the Vclb protein complex in vivo. Immunostaining results also show that the majority of paxillin is no longer localized at the boundary between epicardium and myocardium (Fig. 6H-I′). Release of paxillin may enhance its interaction with FAK to increase FAK phosphorylation. As expected, FAK phosphorylation at Y576/577 is increased in epicardial cells, coronary and endocardial endothelial cells in v12 mutants compared with wild-type siblings, whereas total FAK is comparable between them (Fig. 6J-M). It should be noted that FAK is not activated in the myocardium, even though Vclb is ubiquitously expressed in the ventricle (Fig. 4).

We identified a monoclonal antibody against human VCL that recognizes endogenous Vcla proteins but not NVclb-EGFP proteins (Fig. 7A). We noted that Vcla expression levels are similar between control siblings and v12 mutants (Fig. S13). Whereas NVclb-EGFP is expressed in both myocardium and epicardium (Fig. 7B-B‴), Vcla is mainly expressed in the myocardium (Fig. 7C-C‴). The sequences of Vcla and Vclb are highly conserved and the proteins are likely to be functionally redundant. The lack of FAK activation in the myocardium of v12 mutants might therefore be due to Vcla expression there (Fig. 7C-C‴).Thus, complete lack of vinculin activity in the epicardium causes the elevation of pFAK and pERK.

Fig. 7.

Vcla expression may alleviate defects in myocardium in the v12 mutant. (A) The monoclonal anti-VCL antibody fails to recognize the NVclb-EGFP fusion protein and C-terminal Vclb but recognizes Vclb (aa227-1065) (left panel). All three Vclb variants can be detected by anti-FLAG antibody (right panel). The anti-VCL antibody recognizes the middle region of Vclb, which is absent from NVclb-EGFP, and recognizes Vcla protein (aa227-837), indicating that the VCL antibody only recognizes Vcla protein in v12 mutants. Arrow points to the endogenous VCL protein of HEK293 cells. (B-B‴) Immunostaining reveals that NVclb-EGFP colocalizes with Aldh1a2 in epicardial cells in v12 mutant ventricles at 50 dpf. (C-C‴) Immunostaining shows that Vcla is mainly expressed in the myocardium but not in epicardial cells in v12 mutant ventricles at 50 dpf.

Fig. 7.

Vcla expression may alleviate defects in myocardium in the v12 mutant. (A) The monoclonal anti-VCL antibody fails to recognize the NVclb-EGFP fusion protein and C-terminal Vclb but recognizes Vclb (aa227-1065) (left panel). All three Vclb variants can be detected by anti-FLAG antibody (right panel). The anti-VCL antibody recognizes the middle region of Vclb, which is absent from NVclb-EGFP, and recognizes Vcla protein (aa227-837), indicating that the VCL antibody only recognizes Vcla protein in v12 mutants. Arrow points to the endogenous VCL protein of HEK293 cells. (B-B‴) Immunostaining reveals that NVclb-EGFP colocalizes with Aldh1a2 in epicardial cells in v12 mutant ventricles at 50 dpf. (C-C‴) Immunostaining shows that Vcla is mainly expressed in the myocardium but not in epicardial cells in v12 mutant ventricles at 50 dpf.

Previously, vinculin was shown to be crucial for myocardium morphogenesis. In Vcl knockout mice, the myocardial structure is severely reduced (Xu et al., 1998a). Furthermore, VCL mutations have been identified in a subset of human DCM patients. Mechanistically, disruption of intercalated disc structure and an abnormal distribution of cell-junction proteins in cardiac muscle were found in one of these DCM patients and in cardiomyocyte-specific Vcl knockout mice (Olson et al., 2002; Zemljic-Harpf et al., 2007). In addition to cardiac muscle, VCL and metaVCL are also expressed in non-muscle cells (Belkin et al., 1988; Koteliansky et al., 1992). Vinculin functions in other tissues remain largely unknown. We show here that two vclb mutants have identical multiple cardiac defects, especially in epicardial differentiation and subsequent coronary vessel development. Vclb deficiency increases the phosphorylation of Erk1/2 and FAK in endothelial cells and epicardial cells, resulting in the overproliferation of epicardial-derived cells and abnormal coronary vessel development. Our results exemplify crucial functions of the cytoskeletal protein vinculin in heart development.

The human VCL gene is located on chromosome10. Alternative splicing of exon 19 generates the two isoforms VCL and metaVCL. Human DCM-related mutations all map to exon 19 of VCL (Olson et al., 2002). It should be noted that human DCM patients are heterozygous for the VCL mutations, suggesting that homozygous mutants are lethal. metaVCL, which contains additional amino acids from exon 19, is specifically expressed in cardiac muscle and so its major function is likely to be cardiac muscle specific. In Vcl knockout mice both vinculin isoforms are disrupted, making it difficult to distinguish any different roles of the two isoforms (Xu et al., 1998a). By contrast, vcla and vclb are two separate genes in zebrafish. While vcla has two splicing isoforms similar to its human homolog, the zebrafish vclb transcript resembles the human short isoform. In v12 mutants, vcla expression remains high in the myocardium (Fig. 7), which masks the loss of Vclb due to their functional redundancy. The different genomic organizations of the vinculin genes in human and zebrafish explain why we were able to uncover Vclb-specific functions in zebrafish endocardial/endothelial and epicardial cells. We speculate that the phenotypic difference between complete knockout (embryonic lethal) and cardiomyocyte-specific deletion (survival to adulthood) of the Vcl gene in mouse might be due to its expression and essential functions in endocardial/endothelial and epicardial cells. Specific deletion of the mouse Vcl gene in these cells can be used to test this hypothesis.

FAK promotes angiogenesis in tumor formation (Zhao and Guan, 2011). FAK kinase activity and Grb2/ERK signaling are required for VEGF secretion acting downstream of activated Src kinase in breast carcinoma cells (Mitra et al., 2006). The blood vessel hyperplasia in the v12 mutant suggests that ERK/FAK activation might play a role in the process. Indeed, during coronary vessel formation, levels of Erk1/2 phosphorylation were upregulated in coronary endothelial cells compared with the endocardium (Fig. 6A). This suggests that activation of ERK is required for epicardial- and endocardial-derived cells to form coronary vasculature, which is consistent with the role of ERK activation in cell proliferation and migration (Cho and Klemke, 2000). When treated with 10 μM MEK1 inhibitor U0126 from 28 dpf, coronary vessel development was significantly impaired in wild-type siblings at 40 dpf, but U0126 at this concentration could not alleviate the coronary vessel hyperplasia in v12 mutants (Fig. S10E). Concentrations above 10 μM cause lethality within 24 h, which might be caused by other organ failures owing to their higher sensitivity to the MEK1 inhibitor. Coronary vessel development includes both vasculogenesis and angiogenesis. Vclb deficiency does not appear to affect initial vasculogenesis, but impacts blood vessel connection such that they fail to form a well-organized network of coronary vasculatures. It is likely that coronary progenitor cells and ECM adhere to form cell-cell adhesions and focal adhesions at the end of cell migration. During this process, vinculin not only functions in the assembly of focal adhesion structures, but also helps to downregulate ERK phosphorylation to maintain a balanced level of coronary angiogenesis, possibly through modulating paxillin-FAK interaction. In v12 mutants, this balance is disrupted and hyperactivated ERK causes overproliferation of endocardial- and epicardial-derived cells, including pericytes, leading to abnormal alignment and interaction of endothelial cells and pericytes in the vessel wall, as well as improper interaction of vessels with the myocardium.

Previous results indicated that overexpression of the 258 amino acid N-terminal fragment of Vcl protein in vinculin null cells causes more stable focal adhesions than full-length Vcl (Carisey et al., 2013; Humphries et al., 2007; Xu et al., 1998b). We note that v12 mutants retain the N-terminal 261 amino acids, but the vclb heterozygotes show no obvious developmental defects. To exclude possible roles of the N-terminal fragment in v12 mutants, we also generated a vclb null mutant, vclbdel17, using CRISPR-Cas9 technology. vclbdel17 mutants exhibit similar cardiac phenotypes to v12 mutants, including pleural effusion and epicardial overproliferation (Fig. S3). Thus, the phenotypes of the v12 mutant are caused by loss of function rather than by other mechanisms such as dominant-negative interference.

Epicardial cell activation also plays a key role in heart regeneration. Adult zebrafish hearts can regenerate after damage, making it an excellent model for the study of heart regeneration (Major and Poss, 2007; Poss et al., 2002). During zebrafish heart regeneration, reactivated epicardial cells migrate into wound sites and differentiate into fibroblasts and pericyte-like cells (González-Rosa et al., 2012; Kikuchi et al., 2011; Lepilina et al., 2006). FGF and H2O2 signaling are crucial to fine-tune the phosphorylation of ERK and to control both myocardial and coronary vessel regeneration (Han et al., 2014; Lepilina et al., 2006). The role of vinculin-pFAK-pERK in regulating epicardial and coronary vessel development, as reported here, will enhance our understanding of the molecular basis of heart regeneration.

Zebrafish strains and gene trapping

The zebrafish AB line was used in this study. Zebrafish husbandry and embryo manipulations were performed as described (Westerfield, 2000). The experimental procedures in this study have been approved by the Animal Ethical Committee of the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences. The gene-trap vector and procedure were described previously (Liu et al., 2012). Over 300 injected embryos were raised and outcrossed with wild-type fish. EGFP expression was examined under a fluorescence microscope at different developmental stages until 96 hpf. Sixteen gene-trap lines with obvious EGFP expression were identified. The V12 line, which is one of the trapped lines, was maintained by crossing the heterozygotes with wild-type fish to avoid inbreeding depression for all experiments described here.

Tg(flk1:mcherry) (Dong et al., 2012) and Tg(fli1a:EGFP) (Roman et al., 2002) were obtained from Dr Feng Liu (Chinese Academy of Sciences, Beijing). Tg(tcf21:dsRed) (Kikuchi et al., 2011) was obtained from Dr Geoffrey Burns (Massachusetts General Hospital, Boston). vclbdel17, which carries a 17 bp deletion near the ATG of the vclb gene, was created using CRISPR-Cas9 technology as previously reported (Chang et al., 2013; Hwang et al., 2013).

TAIL-PCR, RT-PCR and RT-qPCR

To identify the entrapped gene in the v12 mutant, we used TAIL-PCR to clone the genomic sequences upstream and downstream of the transposon insertion site as reported (Liu and Chen, 2007). These genomic sequences were used to design genotyping primers. The primers used in TAIL-PCR and genotyping are shown in Table S2. For RT-PCR, total RNA was isolated from ten embryos each of wild type, heterozygote and v12 mutant at the indicated stages using TRIzol reagent (Life Technologies). The first-strand cDNA was synthesized with M-MLV reverse transcriptase (Life Technologies). RT-qPCR was performed in the Bio-Rad CFX96 system. The full-length vclb mRNA sequence was cloned (GenBank KT862534) and used to design the primers shown in Table S3. The primer sequences of egln3 were described previously (Manchenkov et al., 2015).

Quantitative analysis of cardiomyocyte and coronary vessel density

All quantifications were performed using high-resolution confocal images and ImageJ software. Calculations employed GraphPad Prism software. Statistical analysis was performed by unpaired, two-tailed Student's t-test. For all bar graphs, data are represented as mean±s.e.m. P<0.05 was considered significant.

For measurement of cardiomyocyte density, two randomly chosen fields were used to quantify the number of Mef2-positive nuclei and MF20-positive areas in immunostained ventricle sections of 35 dpf samples. We calculated the ratio of Mef2-positive nuclei to the MF20-positive area of the field (175 μm×175 μm), and calculated the density as the number of Mef2-positive nuclei (cardiomyocytes) per 1 mm2. Each group contained five hearts. Measurement of coronary vessel density was performed as described previously (Wilhelm et al., 2016). Randomly chosen fields were used to quantify vascularization at the ventricle surface (40 dpf). Endothelial coverage was determined by calculating the ratio of the GFP-positive area to the total area of the field (200 μm×200 μm), and was expressed as a percentage of the area covered by GFP-positive endothelial cells. Each group contains five hearts with four fields (two on the dorsal and two on the ventral side) from each heart.

RNA in situ hybridization

Digoxigenin-labeled or fluorescein-labeled probes were synthesized using an in vitro transcription system (Roche). Whole-mount in situ hybridization was performed as described (Thisse and Thisse, 2008). BM Purple (Roche) was used to visualize RNA signal at room temperature. Cryosection in situ hybridization was performed as described previously (Simmons et al., 2007).

Western blotting and immunofluorescence

To detect the NVclb-EGFP protein and Vcl protein expression, primary antibodies used in western blotting were rabbit polyclonal anti-GFP (1:2000; Abcam, ab290) and mouse monoclonal anti-VCL (1:5000; Sigma, V9131). For immunoprecipitation experiments, the primary antibodies used in western blot were rabbit polyclonal anti-HA (1:2000; EasyBio, BE 2008), mouse monoclonal anti-FLAG (1:2000; Sigma, F3165) and anti-FLAG M2 affinity gel (1:100; Sigma, A2220). Whole-mount immunofluorescence assay of dissected heart was performed as described (Yang and Xu, 2012). Embryos were photographed with an Olympus FV1000 confocal microscope.

For immunofluorescence staining, samples were fixed with 4% paraformaldehyde overnight at 4°C, washed with PBST (PBS with 0.1% Tween 20), and embedded in paraffin. Sections (5 μm) were then rehydrated and incubated overnight at 4°C with primary antibody in blocking solution followed by an Alexa Fluor 594 goat anti-mouse or Alexa Fluor 488 goat anti-rabbit secondary antibody (1:1000; Life Technologies). DAPI was used to stain nuclei. Primary antibodies were rabbit monoclonal anti-pFAK(pY576/577) (1:100; Invitrogen, 44-652G), rabbit polyclonal anti-β-catenin (1:200; Abcam, ab6302), rabbit monoclonal anti-pErk1/2 (1:400; Cell Signaling Technology, 4370), rabbit polyclonal anti-Aldh1a2 (1:200; GeneTex, GTX124302), rabbit polyclonal anti-Mef2 (1:50; Santa Cruz, sc-313), mouse monoclonal anti-α-actinin (sarcomeric) (1:200; Sigma, A7811), anti-myosin heavy chain monoclonal MF20 (1:200; Developmental Studies Hybridoma Bank, Iowa), mouse monoclonal anti-VCL (1:200; Sigma, V9131), mouse monoclonal anti-Pcna (1:100; Sigma, clone PC10), rabbit polyclonal anti-Pcna (1:100; Santa Cruz, sc-7907), mouse monoclonal anti-mCherry (1:100; EasyBio, BE2026), mouse monoclonal anti-GFP (1:100; EasyBio, BE2001) and rabbit monoclonal anti-GFP (1:100; Cell Signaling Technology, 2956). Sections were photographed with the Olympus FV1000 confocal microscope.

We thank members of J.Z. and J.X. laboratories for help and discussions; Dr Koichi Kawakami for the original Tol2 system; Dr Chao Liu for the modified Tol2 system; and Dr Feng Liu for providing Tg(fli1a:EGFP) and Tg(flk1:mcherry) transgenic lines.

Author contributions

F.C., L.M., Q.W., X.G. performed experiments; F.C. , J.X. and J.Z. analyzed data; F.C. prepared figures; F.C., J.X. and J.Z. designed experiments and wrote the paper.

Funding

This work was financially supported through grants from the National Natural Science Foundation of China [31471359, 31590830]; the Ministry of Science and Technology of the People's Republic of China [2013CB945000]; and the Chinese Academy of Sciences [XDA01010108].

Data availability

The complete cDNA sequence of Danio rerio vinculin b is available at GenBank under accession number KT862534.

Begemann
,
G.
,
Gibert
,
Y.
,
Meyer
,
A.
and
Ingham
,
P. W.
(
2002
).
Cloning of zebrafish T-box genes tbx15 and tbx18 and their expression during embryonic development
.
Mech. Dev.
114
,
137
-
141
.
Belkin
,
A. M.
,
Ornatsky
,
O. I.
,
Glukhova
,
M. A.
and
Koteliansky
,
V. E.
(
1988
).
Immunolocalization of meta-vinculin in human smooth and cardiac muscles
.
J. Cell Biol.
107
,
545
-
553
.
Braren
,
R.
,
Hu
,
H.
,
Kim
,
Y. H.
,
Beggs
,
H. E.
,
Reichardt
,
L. F.
and
Wang
,
R.
(
2006
).
Endothelial FAK is essential for vascular network stability, cell survival, and lamellipodial formation
.
J. Cell Biol.
172
,
151
-
162
.
Carisey
,
A.
and
Ballestrem
,
C.
(
2011
).
Vinculin, an adapter protein in control of cell adhesion signalling
.
Eur. J. Cell Biol.
90
,
157
-
163
.
Carisey
,
A.
,
Tsang
,
R.
,
Greiner
,
A. M.
,
Nijenhuis
,
N.
,
Heath
,
N.
,
Nazgiewicz
,
A.
,
Kemkemer
,
R.
,
Derby
,
B.
,
Spatz
,
J.
and
Ballestrem
,
C.
(
2013
).
Vinculin regulates the recruitment and release of core focal adhesion proteins in a force-dependent manner
.
Curr. Biol.
23
,
271
-
281
.
Chang
,
N.
,
Sun
,
C.
,
Gao
,
L.
,
Zhu
,
D.
,
Xu
,
X.
,
Zhu
,
X.
,
Xiong
,
J.-W.
and
Xi
,
J. J.
(
2013
).
Genome editing with RNA-guided Cas9 nuclease in zebrafish embryos
.
Cell Res.
23
,
465
-
472
.
Cho
,
S. Y.
and
Klemke
,
R. L.
(
2000
).
Extracellular-regulated kinase activation and CAS/Crk coupling regulate cell migration and suppress apoptosis during invasion of the extracellular matrix
.
J. Cell Biol.
149
,
223
-
236
.
Dettman
,
R. W.
,
Denetclaw
,
W.
, Jr.
,
Ordahl
,
C. P.
and
Bristow
,
J.
(
1998
).
Common epicardial origin of coronary vascular smooth muscle, perivascular fibroblasts, and intermyocardial fibroblasts in the avian heart
.
Dev. Biol.
193
,
169
-
181
.
Dong
,
W.
,
Yang
,
Z.
,
Yang
,
F.
,
Wang
,
J.
,
Zhuang
,
Y.
,
Xu
,
C.
,
Zhang
,
B.
,
Tian
,
X.-L.
and
Liu
,
D.
(
2012
).
Suppression of Rap1 impairs cardiac myofibrils and conduction system in zebrafish
.
PLoS ONE
7
,
e50960
.
Fincham
,
V. J.
,
James
,
M.
,
Frame
,
M. C.
and
Winder
,
S. J.
(
2000
).
Active ERK/MAP kinase is targeted to newly forming cell-matrix adhesions by integrin engagement and v-Src
.
EMBO J.
19
,
2911
-
2923
.
Geiger
,
B.
(
1979
).
A 130K protein from chicken gizzard: its localization at the termini of microfilament bundles in cultured chicken cells
.
Cell
18
,
193
-
205
.
González-Rosa
,
J. M.
,
Martin
,
V.
,
Peralta
,
M.
,
Torres
,
M.
and
Mercader
,
N.
(
2011
).
Extensive scar formation and regression during heart regeneration after cryoinjury in zebrafish
.
Development
138
,
1663
-
1674
.
González-Rosa
,
J. M.
,
Peralta
,
M.
and
Mercader
,
N.
(
2012
).
Pan-epicardial lineage tracing reveals that epicardium derived cells give rise to myofibroblasts and perivascular cells during zebrafish heart regeneration
.
Dev. Biol.
370
,
173
-
186
.
Han
,
P.
,
Zhou
,
X.-H.
,
Chang
,
N.
,
Xiao
,
C.-L.
,
Yan
,
S.
,
Ren
,
H.
,
Yang
,
X.-Z.
,
Zhang
,
M.-L.
,
Wu
,
Q.
,
Tang
,
B.
, et al. 
(
2014
).
Hydrogen peroxide primes heart regeneration with a derepression mechanism
.
Cell Res.
24
,
1091
-
1107
.
Harrison
,
M. R. M.
,
Bussmann
,
J.
,
Huang
,
Y.
,
Zhao
,
L.
,
Osorio
,
A.
,
Burns
,
C. G.
,
Burns
,
C. E.
,
Sucov
,
H. M.
,
Siekmann
,
A. F.
and
Lien
,
C.-L.
(
2015
).
Chemokine-guided angiogenesis directs coronary vasculature formation in zebrafish
.
Dev. Cell
33
,
442
-
454
.
Hauck
,
C. R.
,
Hsia
,
D. A.
and
Schlaepfer
,
D. D.
(
2000
).
Focal adhesion kinase facilitates platelet-derived growth factor-BB-stimulated ERK2 activation required for chemotaxis migration of vascular smooth muscle cells
.
J. Biol. Chem.
275
,
41092
-
41099
.
Hauck
,
C. R.
,
Hsia
,
D. A.
and
Schlaepfer
,
D. D.
(
2002
).
The focal adhesion kinase--a regulator of cell migration and invasion
.
IUBMB Life
53
,
115
-
119
.
Hu
,
N.
,
Yost
,
H. J.
and
Clark
,
E. B.
(
2001
).
Cardiac morphology and blood pressure in the adult zebrafish
.
Anat. Rec.
264
,
1
-
12
.
Humphries
,
J. D.
,
Wang
,
P.
,
Streuli
,
C.
,
Geiger
,
B.
,
Humphries
,
M. J.
and
Ballestrem
,
C.
(
2007
).
Vinculin controls focal adhesion formation by direct interactions with talin and actin
.
J. Cell Biol.
179
,
1043
-
1057
.
Hwang
,
W. Y.
,
Fu
,
Y.
,
Reyon
,
D.
,
Maeder
,
M. L.
,
Tsai
,
S. Q.
,
Sander
,
J. D.
,
Peterson
,
R. T.
,
Yeh
,
J.-R. J.
and
Joung
,
J. K.
(
2013
).
Efficient genome editing in zebrafish using a CRISPR-Cas system
.
Nat. Biotechnol.
31
,
227
-
229
.
Ilić
,
D.
,
Furuta
,
Y.
,
Kanazawa
,
S.
,
Takeda
,
N.
,
Sobue
,
K.
,
Nakatsuji
,
N.
,
Nomura
,
S.
,
Fujimoto
,
J.
,
Okada
,
M.
and
Yamamoto
,
T.
(
1995
).
Reduced cell motility and enhanced focal adhesion contact formation in cells from FAK-deficient mice
.
Nature
377
,
539
-
544
.
Kaelin
,
W. G.
, Jr.
and
Ratcliffe
,
P. J.
(
2008
).
Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway
.
Mol. Cell
30
,
393
-
402
.
Kikuchi
,
K.
,
Gupta
,
V.
,
Wang
,
J.
,
Holdway
,
J. E.
,
Wills
,
A. A.
,
Fang
,
Y.
and
Poss
,
K. D.
(
2011
).
tcf21+ epicardial cells adopt non-myocardial fates during zebrafish heart development and regeneration
.
Development
138
,
2895
-
2902
.
Kim
,
J.
,
Wu
,
Q.
,
Zhang
,
Y.
,
Wiens
,
K. M.
,
Huang
,
Y.
,
Rubin
,
N.
,
Shimada
,
H.
,
Handin
,
R. I.
,
Chao
,
M. Y.
,
Tuan
,
T.-L.
, et al. 
(
2010
).
PDGF signaling is required for epicardial function and blood vessel formation in regenerating zebrafish hearts
.
Proc. Natl. Acad. Sci. USA
107
,
17206
-
17210
.
Klemke
,
R. L.
,
Cai
,
S.
,
Giannini
,
A. L.
,
Gallagher
,
P. J.
,
de Lanerolle
,
P.
and
Cheresh
,
D. A.
(
1997
).
Regulation of cell motility by mitogen-activated protein kinase
.
J. Cell Biol.
137
,
481
-
492
.
Koteliansky
,
V. E.
,
Ogryzko
,
E. P.
,
Zhidkova
,
N. I.
,
Weller
,
P. A.
,
Critchley
,
D. R.
,
Vancompernolle
,
K.
,
Vandekerckhove
,
J.
,
Strasser
,
P.
,
Way
,
M.
,
Gimona
,
M.
, et al. 
(
1992
).
An additional exon in the human vinculin gene specifically encodes meta-vinculin-specific difference peptide. Cross-species comparison reveals variable and conserved motifs in the meta-vinculin insert
.
Eur. J. Biochem.
205
,
1218
.
Lepilina
,
A.
,
Coon
,
A. N.
,
Kikuchi
,
K.
,
Holdway
,
J. E.
,
Roberts
,
R. W.
,
Burns
,
C. G.
and
Poss
,
K. D.
(
2006
).
A dynamic epicardial injury response supports progenitor cell activity during zebrafish heart regeneration
.
Cell
127
,
607
-
619
.
Lifschitz-Mercer
,
B.
,
Czernobilsky
,
B.
,
Feldberg
,
E.
and
Geiger
,
B.
(
1997
).
Expression of the adherens junction protein vinculin in human basal and squamous cell tumors: relationship to invasiveness and metastatic potential
.
Hum. Pathol.
28
,
1230
-
1236
.
Liu
,
Y.-G.
and
Chen
,
Y.
(
2007
).
High-efficiency thermal asymmetric interlaced PCR for amplification of unknown flanking sequences
.
Biotechniques
43
,
649
-
656
,
652, 654 passim
.
Liu
,
C.
,
Ma
,
W.
,
Su
,
W.
and
Zhang
,
J.
(
2012
).
Prdm14 acts upstream of islet2 transcription to regulate axon growth of primary motoneurons in zebrafish
.
Development
139
,
4591
-
4600
.
Major
,
R. J.
and
Poss
,
K. D.
(
2007
).
Zebrafish heart regeneration as a model for cardiac tissue repair
.
Drug Discov. Today Dis. Models
4
,
219
-
225
.
Manchenkov
,
T.
,
Pasillas
,
M. P.
,
Haddad
,
G. G.
and
Imam
,
F. B.
(
2015
).
Novel genes critical for hypoxic preconditioning in zebrafish are regulators of insulin and glucose metabolism
.
G3
5
,
1107
-
1116
.
Missinato
,
M. A.
,
Tobita
,
K.
,
Romano
,
N.
,
Carroll
,
J. A.
and
Tsang
,
M.
(
2015
).
Extracellular component hyaluronic acid and its receptor Hmmr are required for epicardial EMT during heart regeneration
.
Cardiovasc. Res.
107
,
487
-
498
.
Mitra
,
S. K.
,
Hanson
,
D. A.
and
Schlaepfer
,
D. D.
(
2005
).
Focal adhesion kinase: in command and control of cell motility
.
Nat. Rev. Mol. Cell Biol.
6
,
56
-
68
.
Mitra
,
S. K.
,
Mikolon
,
D.
,
Molina
,
J. E.
,
Hsia
,
D. A.
,
Hanson
,
D. A.
,
Chi
,
A.
,
Lim
,
S.-T.
,
Bernard-Trifilo
,
J. A.
,
Ilic
,
D.
,
Stupack
,
D. G.
, et al. 
(
2006
).
Intrinsic FAK activity and Y925 phosphorylation facilitate an angiogenic switch in tumors
.
Oncogene
25
,
5969
-
5984
.
Moiseyeva
,
E. P.
,
Weller
,
P. A.
,
Zhidkova
,
N. I.
,
Corben
,
E. B.
,
Patel
,
B.
,
Jasinska
,
I.
,
Koteliansky
,
V. E.
and
Critchley
,
D. R.
(
1993
).
Organization of the human gene encoding the cytoskeletal protein vinculin and the sequence of the vinculin promoter
.
J. Biol. Chem.
268
,
4318
-
4325
.
Munoz-Chapuli
,
R.
,
Gonzalez-Iriarte
,
M.
,
Carmona
,
R.
,
Atencia
,
G.
,
Macias
,
D.
and
Perez-Pomares
,
J. M.
(
2002
).
Cellular precursors of the coronary arteries
.
Tex. Heart Inst. J.
29
,
243
-
249
.
Olivey
,
H. E.
,
Compton
,
L. A.
and
Barnett
,
J. V.
(
2004
).
Coronary vessel development: the epicardium delivers
.
Trends Cardiovasc. Med.
14
,
247
-
251
.
Olson
,
T. M.
,
Illenberger
,
S.
,
Kishimoto
,
N. Y.
,
Huttelmaier
,
S.
,
Keating
,
M. T.
and
Jockusch
,
B. M.
(
2002
).
Metavinculin mutations alter actin interaction in dilated cardiomyopathy
.
Circulation
105
,
431
-
437
.
Parsons
,
J. T.
(
2003
).
Focal adhesion kinase: the first ten years
.
J. Cell Sci.
116
,
1409
-
1416
.
Pascoal
,
S.
,
Esteves de Lima
,
J.
,
Leslie
,
J. D.
,
Hughes
,
S. M.
and
Saúde
,
L.
(
2013
).
Notch signalling is required for the formation of structurally stable muscle fibres in zebrafish
.
PLoS ONE
8
,
e68021
.
Peng
,
X.
,
Ueda
,
H.
,
Zhou
,
H.
,
Stokol
,
T.
,
Shen
,
T. L.
,
Alcaraz
,
A.
,
Nagy
,
T.
,
Vassalli
,
J. D.
and
Guan
,
J. L.
(
2004
).
Overexpression of focal adhesion kinase in vascular endothelial cells promotes angiogenesis in transgenic mice
.
Cardiovasc. Res.
64
,
421
-
430
.
Polte
,
T. R.
,
Naftilan
,
A. J.
and
Hanks
,
S. K.
(
1994
).
Focal adhesion kinase is abundant in developing blood vessels and elevation of its phosphotyrosine content in vascular smooth muscle cells is a rapid response to angiotensin II
.
J. Cell Biochem.
55
,
106
-
119
.
Poss
,
K. D.
,
Wilson
,
L. G.
and
Keating
,
M. T.
(
2002
).
Heart regeneration in zebrafish
.
Science
298
,
2188
-
2190
.
Raz
,
A.
and
Geiger
,
B.
(
1982
).
Altered organization of cell-substrate contacts and membrane-associated cytoskeleton in tumor cell variants exhibiting different metastatic capabilities
.
Cancer Res.
42
,
5183
-
5190
.
Red-Horse
,
K.
,
Ueno
,
H.
,
Weissman
,
I. L.
and
Krasnow
,
M. A.
(
2010
).
Coronary arteries form by developmental reprogramming of venous cells
.
Nature
464
,
549
-
553
.
Reese
,
D. E.
,
Mikawa
,
T.
and
Bader
,
D. M.
(
2002
).
Development of the coronary vessel system
.
Circ. Res.
91
,
761
-
768
.
Roman
,
B. L.
,
Pham
,
V. N.
,
Lawson
,
N. D.
,
Kulik
,
M.
,
Childs
,
S.
,
Lekven
,
A. C.
,
Garrity
,
D. M.
,
Moon
,
R. T.
,
Fishman
,
M. C.
,
Lechleider
,
R. J.
, et al. 
(
2002
).
Disruption of acvrl1 increases endothelial cell number in zebrafish cranial vessels
.
Development
129
,
3009
-
3019
.
Seger
,
R.
and
Krebs
,
E. G.
(
1995
).
The MAPK signaling cascade
.
FASEB J.
9
,
726
-
735
.
Shen
,
T.-L.
,
Park
,
A. Y.-J.
,
Alcaraz
,
A.
,
Peng
,
X.
,
Jang
,
I.
,
Koni
,
P.
,
Flavell
,
R. A.
,
Gu
,
H.
and
Guan
,
J. L.
(
2005
).
Conditional knockout of focal adhesion kinase in endothelial cells reveals its role in angiogenesis and vascular development in late embryogenesis
.
J. Cell Biol.
169
,
941
-
952
.
Simmons
,
D. G.
,
Fortier
,
A. L.
and
Cross
,
J. C.
(
2007
).
Diverse subtypes and developmental origins of trophoblast giant cells in the mouse placenta
.
Dev. Biol.
304
,
567
-
578
.
Staudt
,
D.
and
Stainier
,
D.
(
2012
).
Uncovering the molecular and cellular mechanisms of heart development using the zebrafish
.
Annu. Rev. Genet.
46
,
397
-
418
.
Subauste
,
M. C.
,
Pertz
,
O.
,
Adamson
,
E. D.
,
Turner
,
C. E.
,
Junger
,
S.
and
Hahn
,
K. M.
(
2004
).
Vinculin modulation of paxillin–FAK interactions regulates ERK to control survival and motility
.
J. Cell Biol.
165
,
371
-
381
.
Thisse
,
C.
and
Thisse
,
B.
(
2008
).
High-resolution in situ hybridization to whole-mount zebrafish embryos
.
Nat. Protoc.
3
,
59
-
69
.
Turner
,
C. E.
(
2000
).
Paxillin and focal adhesion signalling
.
Nat. Cell Biol.
2
,
E231
-
E236
.
Vogel
,
B.
,
Meder
,
B.
,
Just
,
S.
,
Laufer
,
C.
,
Berger
,
I.
,
Weber
,
S.
,
Katus
,
H. A.
and
Rottbauer
,
W.
(
2009
).
In-vivo characterization of human dilated cardiomyopathy genes in zebrafish
.
Biochem. Biophys. Res. Commun.
390
,
516
-
522
.
Vrancken Peeters
,
M.-P. F. M.
,
Gittenberger-de Groot
,
A. C.
,
Mentink
,
M. M. T.
and
Poelmann
,
R. E.
(
1999
).
Smooth muscle cells and fibroblasts of the coronary arteries derive from epithelial-mesenchymal transformation of the epicardium
.
Anat. Embryol.
199
,
367
-
378
.
Westerfield
,
M.
(
2000
).
The Zebrafish Book: A Guide for the Laboratory Use of Zebrafish (Danio rerio)
.
Eugene, OR
:
University of Oregon Press
.
Wilhelm
,
K.
,
Happel
,
K.
,
Eelen
,
G.
,
Schoors
,
S.
,
Oellerich
,
M. F.
,
Lim
,
R.
,
Zimmermann
,
B.
,
Aspalter
,
I. M.
,
Franco
,
C. A.
,
Boettger
,
T.
, et al. 
(
2016
).
FOXO1 couples metabolic activity and growth state in the vascular endothelium
.
Nature
529
,
216
-
220
.
Wood
,
C. K.
,
Turner
,
C. E.
,
Jackson
,
P.
and
Critchley
,
D. R.
(
1994
).
Characterisation of the paxillin-binding site and the C-terminal focal adhesion targeting sequence in vinculin
.
J. Cell Sci.
107
,
709
-
717
.
Wu
,
B.
,
Zhang
,
Z.
,
Lui
,
W.
,
Chen
,
X.
,
Wang
,
Y.
,
Chamberlain
,
A. A.
,
Moreno-Rodriguez
,
R. A.
,
Markwald
,
R. R.
,
O'Rourke
,
B. P.
,
Sharp
,
D. J.
, et al. 
(
2012
).
Endocardial cells form the coronary arteries by angiogenesis through myocardial-endocardial VEGF signaling
.
Cell
151
,
1083
-
1096
.
Xu
,
W.
,
Baribault
,
H.
and
Adamson
,
E. D.
(
1998a
).
Vinculin knockout results in heart and brain defects during embryonic development
.
Development
125
,
327
-
337
.
Xu
,
W.
,
Coll
,
J. L.
and
Adamson
,
E. D.
(
1998b
).
Rescue of the mutant phenotype by reexpression of full-length vinculin in null F9 cells; effects on cell locomotion by domain deleted vinculin
.
J. Cell Sci.
111
,
1535
-
1544
.
Yang
,
J.
and
Xu
,
X.
(
2012
).
Immunostaining of dissected zebrafish embryonic heart
.
J. Vis. Exp.
,
e3510
.
Zamir
,
E.
and
Geiger
,
B.
(
2001
).
Molecular complexity and dynamics of cell-matrix adhesions
.
J. Cell Sci.
114
,
3583
-
3590
.
Zemljic-Harpf
,
A. E.
,
Miller
,
J. C.
,
Henderson
,
S. A.
,
Wright
,
A. T.
,
Manso
,
A. M.
,
Elsherif
,
L.
,
Dalton
,
N. D.
,
Thor
,
A. K.
,
Perkins
,
G. A.
,
McCulloch
,
A. D.
, et al. 
(
2007
).
Cardiac-myocyte-specific excision of the vinculin gene disrupts cellular junctions, causing sudden death or dilated cardiomyopathy
.
Mol. Cell. Biol.
27
,
7522
-
7537
.
Zhao
,
X.
and
Guan
,
J.-L.
(
2011
).
Focal adhesion kinase and its signaling pathways in cell migration and angiogenesis
.
Adv. Drug Deliv. Rev.
63
,
610
-
615
.
Ziegler
,
W. H.
,
Liddington
,
R. C.
and
Critchley
,
D. R.
(
2006
).
The structure and regulation of vinculin
.
Trends Cell Biol.
16
,
453
-
460
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information