During embryogenesis, the vertebrate heart is the first organ to form a linear-tube structure that begins to function. The heart tube is divided into a ventricle and an atrium (DeHaan,1965). The ventricle differs from the atrium by morphological, physiological, histological, and molecular criteria, including chamber-specific gene expression programs (Franco et al.,1998; Yelon and Stainier,1999).
Several members of heart contractile-protein gene families show chamber-specific expression patterns. The myosin heavy chain (MHC) has been identified as a large contractile protein of myocardial-myosin thick filaments. Many aspects of mhc-gene expression that code MHC proteins during heart development are evolutionarily conserved. In mammals, the thick filaments of cardiac muscle are composed entirely of cardiac α-MHC and β-MHC (Hoh et al.,1979; Chizzonite et al.,1982). Xenopus and chicken ventricular MHC (vMHC) are more homologous to the mammalian MHC isoform, MyH15, than to mammalian β-MHC. However, vMHC is a functional ortholog of mammalian β-MHC as they are exclusively expressed in the ventricle, and the MyH15 gene seems to be no longer transcriptionally active (Garriock et al.,2005).
In zebrafish, mhc is recognized as an atrial isoform (amhc) and a ventricular isoform (vmhc; Yelon,2001). During zebrafish embryogenesis, the vmhc-expression pattern divides the myocardial precursors into ventricular and atrial subpopulations (Yelon et al.,1999). The medially restricted vmhc expression in zebrafish is the earliest known indicator of molecular diversification of myocardial precursors (Yelon et al.,1999). In lateral plate mesoderm, vmhc begins to express at a medial subpopulation of myocardial precursors that includes several discrete tissue types such as the heart, endothelium, blood, connective tissue, smooth muscle, and the chondrogenic portion of the limbs. The medial vmhc+ cells are believed to be the ventricular precursors, and the lateral vmhc− cells are thought to be atrial precursors (Yelon,2001). Whole-mount in situ hybridization of zebrafish embryos revealed that vmhc mRNA is specifically expressed in the trunk as well as in ventricular muscle (Peterkin et al.,2003; Holtzinger and Evans,2007).
GATA-binding proteins (GATA) are required for the normal expression of chamber-specific genes and heart formation. During the differentiation of the heart, GATA4, -5, and -6 are expressed in the mesoderm and endoderm (Molkentin,2000; Pikkarainen et al.,2004). Recent studies using antisense morpholino oligonucleotides (MOs) of GATA factors suggested that zebrafish and Xenopus GATA4, -5, and -6 have redundant functions in the regulation of heart development and cardiac-specific gene expression (Holtzinger and Evans,2007; Peterkin et al.,2007). In mice, GATA4/6 double-heterozygous mutant embryos (GATA4+/−GATA6+/−) displayed cardiovascular defects and down-regulation of myocyte enhancer factor 2C (MEF2C) and β-MHC expression (Xin et al.,2006). Furthermore, loss of GATA4 and -6 (GATA4−/−GATA6−/−) completely blocked heart development in mice (Zhao et al.,2008). Clinical studies have indicated that mutation of human GATA4 and a mutation preventing GATA4-TBX5 complex formation are implicated in human congenital heart defects (Garg et al.,2003). Thus, in several species, GATA4 and -6 are more likely to be conserved in the regulation of whole heart development. To date, little is known about the regulation and maintenance of chamber-specific differences.
We isolated and analyzed the zebrafish vmhc promoter to understand the mechanism of ventricle-specific expression. We show that the GATA-binding site is essential for vmhc expression in the zebrafish ventricle, and that GATA4 and -6 are involved in vmhc expression.
Isolation and Expression of a Ventricle-Specific Promoter
Using a nested polymerase chain reaction (PCR), we isolated the 5′-flanking region of the vmhc gene. PCR primers were designed from the zebrafish genomic DNA GeneBank BX000385 sequence (Supp. Table S1, which is available online). The 5′-flanking region was deduced by comparing the genomic DNA sequence to the cDNA sequence from NM_001112733 (GeneBank), which contains the most distal 5′-untranslated region (UTR) among vmhc cDNA sequences. The 2.2-kb genomic DNA fragment includes an 1,834-bp 5′-upstream region, three exons, and two introns (accession no. EU786118). Using the Genomatix MatInspector program (http://www.genomatix.de), we found that the 5′-upstream region contains putative transcription-factor binding sites for serum response factor (SRF), GATA, Smad, and Nkx, all of which have been implicated in the control of heart development (Fig. 1). When the 2.2-kb genomic DNA fragment was fused to the open reading frame (ORF) of enhanced GFP (EGFP; pZVMHC-1834) and transiently expressed in zebrafish embryo using microinjection, the fragment was sufficient to induce ventricle-specific expression of GFP from the construct.
Generation of the vmhc Promoter: gfp Transgenic Zebrafish Line
To confirm that the ventricle-specific expression of the vmhc promoter construct was consistent with endogenous expression, we generated a transgenic line using a pZVMHC-1834 construct and injected NotI-linearized pZVMHC-1834 into fertilized eggs. Ventricle-specific GFP expression was displayed in 25–30% of the surviving embryos. The GFP-expressing embryos were selected and raised to adulthood. After 3 months, the surviving founders were mated with wild-type (WT) zebrafish to find germline-transmitted transgenic zebrafish. A stable ventricle-specific GFP-expressing zebrafish line was established from the 50 founders tested. The transgenic line was mated and produced normal offspring. The GFP-labeled ventricles of both homozygotic and heterozygotic transgenic zebrafish were highly resolved throughout their life span, which confirmed normal atrial and ventricular contractions (Supp. Movie S1). Using in situ hybridization, we identified a difference between endogenous vmhc and transgenic GFP expression by the zebrafish vmhc promoter region (Fig. 2). The transgene was expressed with a same pattern to the endogenous vmhc in the heart region (Fig. 2A,C,E,G,I,K). In trunk muscle, where the endogenous vmhc transcripts were detected, gfp transcript was not observed in transgenic fish (Fig. 2B,D,F,H,J,L). Fluorescence microscopy revealed that the GFP protein from the transgene was expressed in the primitive heart tube, 21.5 and 24 hours postfertilization (hpf), when the cardiac precursors form heart tube (Fig. 2M,N). At 48 hpf, GFP signals were clearly shown at the ventricle of transgenic zebrafish embryo (Fig. 3A–C). Using the monoclonal antibody MF20, which recognizes sarcomeres (Bader et al.,1982) and allows clear visualization of the heart, we confirmed that a green fluorescent signal appeared specifically in the ventricle at 48 hpf (Fig. 3D–F). At 2 months postfertilization, ventricle-specific GFP expression was maintained in heterozygotic F3 progeny (Fig. 3G–I). GFP expression began at 3.5 days postfertilization (dpf) in jaw, eye and trunk muscles, and at 1 month postfertilization, we detected additional GFP signals in muscles of pectoral fin, dorsal fin, anal fin, caudal fin (data not shown).
The −301 to −277 Region Is Important for vmhc Promoter Activity
The 5′-flanking region contained several putative transcription factor-binding sites related to heart development, such as Nkx2.5, GATA, and Smad (Fig. 1). To determine the regulatory elements for ventricle-specific expression in the proximal vmhc promoter, we serially deleted the 5′-upstream region containing the promoter linked to the ORF of EGFP (Fig. 4).
The ventricle-specific promoter activity was estimated by the intensity of the transiently expressed GFP-positive signals in 72-hpf embryos (Fig. 4). Fluorescence was detected in embryo hearts injected with pZVMHC-1834, pZVMHC-508, pZVMHC-348, pZVMHC-301, pZVMHC-1834ex1, and pZVMHC-301ex1 constructs. However, the embryos injected with the pZVMHC-276 (n = 176) or pZVMHC-240 (n = 250) constructs showed no fluorescence in whole body, including the heart (Fig. 4). These results suggest that the region from −301 to −277 contains regulatory elements for vmhc promoter activity.
A Proximal GATA Site Is Essential for Ventricle-Specific vmhc Promoter Activity
The −301 to −277 region contains several putative cis-acting elements, such as binding sites for Smad and GATA, that are important for heart formation in the mouse model (Peterkin et al.,2005; Qi et al.,2007; Song et al.,2007). To investigate which putative cis-acting element(s) are responsible for ventricle-specific vmhc promoter activity in the −301 to −277 region, we introduced point mutations into the putative Smad- and GATA-binding sites of the pZVMHC-301 construct (Fig. 5A). A single nucleotide substitution of the GATA-binding site (C→T at −287) completely abolished GFP expression in the ventricle (Fig. 5D), whereas a mutation of the Smad-binding site (GTC→CAT at −300) did not affect expression (Fig. 5C). These results indicate that the putative GATA-binding site is a critical cis-acting element for vmhc expression.
GATA4 and -6 Are Involved in the Regulation of vmhc Expression
To further confirm whether GATA factors are involved in vmhc promoter activity, we used MOs directed against GATA4 (MO4), GATA5 (MO5), and GATA6 (MO6). MO4, MO5, and MO6 were individually or simultaneously injected into fertilized eggs derived from vmhc:gfp, and vmhc expression was analyzed by measuring ventricular GFP expression. GFP expression was not significantly changed when MO4, MO5, or MO6 was separately injected compared with mock-injected eggs (Fig. 6A–D). However, when GATA4 and -6 were depleted by simultaneous injection of both MOs, the double morphants showed a significant decrease of ventricular GFP expression (Fig. 6F). When GATA4, -5, and -6 were depleted by simultaneous injection of all three MOs, the morphants represented heartless phenotype (Fig. 6H), whereas other combinations of MOs did not show significant decrease of GFP expression (Fig. 6E,G).
Zebrafish is an excellent vertebrate model for cardiovascular development (Fishman and Stainier,1994; Fishman and Chien,1997; Alexander et al.,1998). Heart-specific GFP-expressing zebrafish were established using a cmlc2 promoter, and they have been successfully used to study zebrafish heart development (Huang et al.,2003; Wang et al.,2006; Chi et al.,2008). Although the cmlc2 promoter drives downstream gene expression throughout the entire heart, this promoter is unlikely to explain chamber-specific gene expression. Therefore, we isolated the vmhc promoter, which appears to be a valuable tool to study chamber formation and function. Our most important findings are that the nucleotide fragment from −301 to +26 (−301/+26) is sufficient for ventricle-specific vmhc promoter activity during embryonic heart development, and that GATA plays an essential role in its expression.
Generation of Ventricle-Specific Fluorescent Zebrafish
Using the 2.2-kb genomic DNA fragment linked to GFP, we established a transgenic line specifically for expressing GFP in the ventricle of the zebrafish heart. The expression of the transgene (vmhc:gfp) was also restricted to the ventricle from the embryonic stage through adulthood. In contrast to myocardium-specific GFP-expressing transgenic mice that suffered from severely dilated cardiomyopathy (Huang et al.,2000), our ventricle-specific fluorescent zebrafish developed normally and produced offspring that stably expressed GFP. Therefore, the vmhc promoter could be a tool to investigate ventricle-specific expression of foreign genes during zebrafish embryonic heart development, and the ventricle-fluorescent zebrafish would be useful to specifically monitor ventricular function.
The difference of expression patterns in trunk muscle between endogenous vmhc and transgene remains to be identified. There could be several explanations. One is that our vmhc promoter contains the positive regulatory region only for the ventricle, not for the trunk muscle. Another is that vmhc transcripts detected in the trunk muscle by in situ hybridization is originated from a duplicated gene with trunk muscle-specific promoter activity.
GATA Site as a Critical Upstream Element of vmhc Promoter Activity
Based on the finding that deletion of the 5′-upstream region up to −276 completely abolished the fluorescent signal in the ventricle, the promoter analysis revealed that the −301/−277 region is critical for vmhc promoter activity. The −301/−277 region of the promoter contains putative Smad- and GATA-binding sites. The function of these putative regulatory elements in myocardial differentiation is highly conserved (Molkentin,2000; Liberatore et al.,2002; Peterkin et al.,2005; Qi et al.,2007; Song et al.,2007). A single nucleotide substitution (−287/C→T) of the putative GATA-binding site of vmhc promoter region led to a complete loss of promoter activity, whereas a mutation of the Smad-binding site did not affect its activity.
Essential GATA-binding sites have been detected in the promoter of myocardium-specific expressing genes in vertebrates (Searcy et al.,1998; Lien et al.,1999; Small and Krieg,2003; Latinkic et al.,2004). These results suggest that GATA factors positively regulate vmhc expression as activators for ventricle-specific expression.
GATA4 and -6 Are Redundant in Their Regulation of vmhc Expression
To confirm the involvement of the GATA factors with the vmhc promoter, we measured transgenic vmhc promoter-derived GFP expression using MOs against GATA4, -5, and -6. Only a simultaneous injection of GATA4 and -6 into fertilized eggs diminished transgenic vmhc expression. A recent study using GATA4/6 double heterozygous mutant mice showed diminished β-MHC and MEF2C expression in response to GATA4 and -6, which suggests a combined role in chamber-specific mhc expression (Xin et al.,2006). Furthermore, the β-MHC and MEF2C expression level were unaltered in mice depleted of GATA4 alone (Zeisberg et al.,2005). These results suggest that GATA4 and -6 are redundant for the ventricle-specific expression of β-MHC, which is a functional ortholog of vMHC (Garriock et al.,2005). More recently, GATA factors were shown to be functionally redundant for the differentiation of zebrafish and Xenopus cardiomyocytes (Holtzinger and Evans,2007; Peterkin et al.,2007). Our finding that GATA4 and -6 are functionally redundant in ventricle-specific vmhc expression in zebrafish embryos is in general agreement with these findings.
Although the mutation of the GATA-binding site of the vmhc promoter completely blocked GFP expression, we observed reduced but considerable fluorescent signal and vmhc ventricular expression in the GATA4 and -6 double morphants (Fig 6F and Fig. S1). These results indicate that the GATA-binding site (−287 to −290) are not exclusively dependent on GATA4 and -6, suggesting that GATA5 or other factors may share functions with the vmhc promoter. However, it was difficult to test their effects to vmhc expression because simultaneous depletion of GATA4, -5, and -6 resulted in heartless phenotype.
GATA Factors and Heart Development
Knockdown of GATA5 and -6 did not completely abolished myocardial formation in our experimental condition (Fig. 6G), while GATA5 and -6 double-deficient zebrafish appeared to be heartless phenotypes (Holtzinger and Evans,2007). These results suggest that GATA5 might be involved in the myocardial precursor number control at the early stage as suggested by Trinh et al. (2005) and GATA4 and -6 might function to regulate the expression of functional genes such as vmhc at later stage.
GATA4, -5, and -6 are involved in vertebrate heart development. GATA4-deficient zebrafish failed to develop hearts but had normal heart tube formation and differentiation of the chambers (Holtzinger and Evans,2005). GATA6 has been implicated in the maturation of cardiac mesoderm in Xenopus and zebrafish embryos (Peterkin et al.,2003). In the mouse myocardium, conditional deletion of GATA4 resulted in abnormal morphogenesis of the right ventricle (Zeisberg et al.,2005), and GATA4/GATA6 double heterozygous mutant embryos had diminished β-MHC and MEF2C expression and cardiovascular defects (Xin et al.,2006).
Because all GATA morphants tested here showed abnormal heart morphogenesis (Fig. 6), each GATA factor is likely to have a nonredundant function during heart formation. However, only embryos co-injected with GATA4 and -6 MOs showed reduced vmhc expression. Thus, we conclude that GATA4 and -6 play a key role in the regulation of vmhc expression in the ventricle with certain level of redundancy.
A zebrafish AB strain was maintained and staged as previously described (Westerfield,1995). Embryonic stages were recorded as hr postfertilization (hpf) and days postfertilization (dpf; Kimmel et al.,1995). Transgenic zebrafish embryos displaying ventricle-specific GFP expression were confirmed by fluorescence using a Leica MZ10F fluorescence stereomicroscope.
Cloning of the vmhc Promoter
A nested PCR was performed using the vmhc-5 forward primer, the vmhc-3 reverse primer, and genomic DNA from the zebrafish AB strain as a template. The reaction product was purified using QIAquick Spin Columns (Qiagen), amplified with nested PCR primers (vmhc-1834-XhoI and vmhc-R-BamHI), and then subcloned into the XhoI-BamHI site of the pEGFP-1 vector (Clontech). High fidelity Platinum DNA Taq polymerase (Invitrogen) was used for all PCR reactions. DNA sequence analyses were conducted by Solgent Co. Ltd (Daejeon). The primers are described in Supp. Table S1.
vmhc Promoter-GFP Constructs
To identify the regulatory elements in the proximal region of the zebrafish vmhc, we constructed GFP reporter plasmids containing a serially deleted 5′-upstream region. The serial deletion constructs, pZVMHC-1834, -348, -301, -276, -240, were amplified from the indicated upstream sites to +394 of the third exon containing the ATG start codon. pZVMHC-508 was generated with the BamHI restriction enzyme site at −508. Point mutations were introduced into pZVMHC-301 by PCR with base-substituted primers (Supp. Table S1).
Whole-Mount In Situ Hybridization and Immunostaining
Whole-mount in situ hybridization with digoxigenin (DIG) -labeled riboprobes was performed according to a previously described protocol (Alexander et al.,1998). The probes for in situ hybridization were synthesized from wild-type zebrafish cDNA using specific primers (Supp. Table S1). After fixation, embryos older than 24 hr were treated with 6% H2O2 and 10 μg/ml Proteinase K. Hybridization was performed using a DIG RNA labeling kit (Roche) at 65°C in 50% formamide buffer with DIG-labeled antisense riboprobes synthesized by an in vitro transcription reaction with T7 RNA polymerase.
Whole-mount immunostaining was performed as previously described (Yelon et al.,1999), using the monoclonal antibodies MF20 (Bader et al.,1982). MF20 was obtained from the Developmental Studies Hybridoma Bank. A secondary antibody against mouse IgG tagged Alexa-594 (Invitrogen) was used.
Each GFP reporter construct plasmid (10 ng/μl) was injected into the cytoplasm of single-cell fertilized eggs incubated at 28°C in 0.03% sea salt containing a low concentration of methylene blue. Injection conditions for GATA4 (5 ng/nl MO), GATA5 (5 ng/nl MO), and GATA6 (2 ng/nl MO; Open Biosystems) were optimized to a concentration that showed consistent phenotypes. The morpholinos designed to target GATA4, GATA5, and GATA6 were previously described and characterized with respect to specificity (Holtzinger and Evans,2005; Peterkin et al.,2003; Trinh et al.,2005). Control was injected with injection buffer (Mock).
Total RNA for reverse transcription (RT) reactions was isolated from collected embryos using Trizol Reagent (Invitrogen). The RT reactions were conducted in 20 μl of reaction mix containing 1 μg of RNA for 90 min at 42°C. The PCR reaction was followed with 2 μl of RT sample in 50 μl. The PCR primers are described in Supp. Table S1.
We thank the members of the Jo and Huh Laboratories. We also thank the Korea Zebrafish Organogenesis Mutant Bank for providing zebrafish lines.