Transcription factors are key regulators of cardiogenesis, including those from the GATA, Nkx2, and TBX gene families, as well as SRF, Isl1, and MEF2 (Harvey,2002; Srivastava,2006; Chien et al.,2008). GATA proteins are transcription factors, which regulate diverse processes during embryogenesis and in adults (Patient and McGhee,2002). The GATA4, 5, and 6 genes are all important in vertebrates for endoderm formation (Weber et al.,2000; Afouda et al.,2005) and are central regulators of heart development (Gove et al.,1997; Reiter et al.,1999; Peterkin et al.,2003,2007; Holtzinger and Evans,2005; Afouda et al.,2008; Haworth et al.,2008; Takeuchi and Bruneau,2009). They have overlapping spatial and temporal expression patterns and regulate cardiomyogenesis in a partially redundant manner (Holtzinger and Evans,2007; Peterkin et al.,2007). In addition, recent studies have provided evidence that they regulate each other (Afouda et al.,2005,2008). It is, therefore, challenging to dissect the specific contribution different individual GATA factors make to cardiogenesis.
Evidence about the requirements of GATA factors for vertebrate cardiogenesis differs depending on the species in which the investigations have been carried out. In mouse, loss of either GATA4 or GATA6 results in developmental arrest due to a requirement for these factors in the extra-embryonic endoderm (Kuo et al.,1997; Koutsourakis et al.,1999). Studies using chimeras to circumvent this problem suggested that GATA4 has no evident function in cardiogenesis (Narita et al.,1997a) and that GATA6 is not required for specification of the myocardium (Kuo et al.,1997; Molkentin et al.,1997; Morrisey et al.,1998; Koutsourakis et al.,1999). In contrast, loss-of-function studies in zebrafish and Xenopus show that GATA6 is required for the maintenance and differentiation of cardiac progenitors (Peterkin et al.,2003). One possible explanation for these apparently contradictory results in mouse chimeras could be that GATA gene function in heart development is cell non-autonomous, which would allow for some rescue of cardiogenesis in gata4 or gata6 mutant cells from surrounding wild-type cells.
In Xenopus, GATA4 is sufficient for driving not only induction of cardiac marker gene expression, but differentiation of beating tissue as well (Latinkic et al.,2003; Afouda et al.,2008). Recent studies in mouse have confirmed the importance of this factor and show that GATA4 is among three factors (GATA4, Tbx5, and Baf60c) that together are sufficient for transdifferentiation of non-cardiogenic mesoderm into beating cardiomyocytes (Takeuchi and Bruneau,2009). In addition, GATA4 together with Mef2c and Tbx5 is able to reprogram post-natal cardiac or dermal fibroblasts directly into differentiated cardiomyocytes-like-cells (Ieda et al.,2010).
There is also a controversy about the importance of GATA5 for heart development. In contrast to zebrafish where a critical role for GATA5 in specification of the myocardium was demonstrated (Reiter et al.,1999), gene knockout studies in mice suggested that GATA5 was dispensable for embryonic development (Molkentin et al.,2000). These mouse findings seemed to be supported by initial Xenopus studies reporting only a minor role for GATA5 in heart formation (Peterkin et al.,2007). However, a recent report claims instead that Xenopus GATA5 is essential for early heart development (Haworth et al.,2008).
This discrepancy could be due to the different tools used for studying GATA5 in Xenopus heart development. Peterkin et al. (2007) used translation blocking morpholinos against GATA5 when they observed no significant effect on cardiogenesis, while Haworth et al. (2008) designed new splice-blocking morpholinos for their investigation to suggest an important role for GATA5 in heart formation. While these observations may cast doubt on the efficiency of the original translation blocking GATA5 MO, we may also raise questions about the specificity of the newly designed splice morpholino, particularly since Haworth et al. (2008) failed to reverse the effects caused by their splice-blocking GATA5 morpholino in attempted rescue experiments. It is clearly important to settle this controversy about the role of GATA5 in Xenopus cardiogenesis, not least because of the implications that solving this question in Xenopus has for the role of GATA5 in vertebrate heart development in general.
Wnt genes are among the non-cell-autonomous targets of GATA factors (Alexandrovich et al.,2006; Afouda et al.,2008). The so-called canonical pathway involves β-catenin (Hoppler and Kavanagh,2007); alternative, β-catenin-independent pathways are known as non-canonical pathways, one of which involves rho-GTPase, cdc42, and PKCδ and may result in activation of Jun N-terminal Kinase (JNK) (Eisenberg and Eisenberg,2006). Activation of Wnt11b is implicated in early cardiogenesis, since Wnt11b induces expression of cardiogenic markers such as Nkx2-5 and GATA4, as well as differentiation into beating cardiomyocyte tissue (Eisenberg and Eisenberg,1999; Pandur et al.,2002). Furthermore, we have shown that, in turn, GATA4 and 6 induce Wnt11b gene expression directly, suggesting a possible regulatory loop between GATA factors and Wnt signaling molecules (Afouda et al.,2008).
In this investigation, we uncovered some functionally distinct roles for different GATA factors during heart formation. Using translation-blocking (Afouda et al.,2005; Peterkin et al.,2007) and splice junction-blocking morpholinos (Haworth et al.,2008), we uncover differences in the requirements of the GATA genes for different steps in cardiogenesis. We show that loss of GATA4 and 6 result in the inhibition of early (cardiogenic) and late (cardiomyocyte differentiation) maker gene expression, in contrast to GATA5 whose inhibition affects only differentiation markers. We use rescue experiments to uncover functional activity differences between the GATA proteins in cardiogenesis, which correspond well with the observed requirements for the different GATA genes. We show that while the effects resulting from loss of GATA4 and 6 functions can be rescued by activating non-canonical Wnt11b signaling, the loss of GATA5 function cannot be rescued by non-canonical Wnt signaling. We, therefore, show that the differential requirements for GATA factors during cardiogenesis are mediated by non-canonical Wnt11b signaling.
Wnt11b Differentially Mediates Function of GATA Factors During Cardiogenesis
Before examining in more details the interplay between the three cardiogenic GATAs and Wnt11b, we first re-examined the individual loss of GATA4-6 using previously reported MO tools. We used the conventional translation blocking MOs (tlMOs) against all three GATA factors as described in Afouda et al. (2005) and Peterkin et al. (2007) and in addition the newly designed splicing MOs (spMOs) against GATA4 and 5 (Haworth et al.,2008). When MOs against GATA4, 5, or 6 were injected into the presumptive cardiogenic mesoderm, different effects on cardiac differentiation were observed. Because some of the early cardiac markers are also expressed in non-cardiac tissues such as endoderm, only expression of late differentiation markers was monitored to assess the effects on cardiac development. The spMOs against GATA4 and 5 and the tlMO against GATA6 cause profound defects ranging from mild to severe reduction in expression of the cardiac differentiation marker TnIc as monitored by whole-mount in situ hybridization (Fig. 1A, B, C, E, G, J) as well as cardia bifida (data not shown). In contrast, when tlMOs against GATA4 and 5 are injected, the GATA4 tlMO causes a more severe phenotype than the GATA5 tlMO, which shows only minimal effects on cardiac differentiation marker gene expression (data not shown, as in Peterkin et al.,2007). The efficiency of both GATA4 and GATA5 spMOs was assessed by monitoring the status of the endogenous full-length genes by RT-PCR as previously shown (Haworth et al.,2008). Injection of GATA4 or GATA5 spMOs resulted in viable embryos and caused significant reduction in the level of the expected full-length mRNAs (Fig. 1I). The effects of depletion of individual GATA factors on cardiac differentiation were also analyzed with RT-PCR using two cardiomyocyte-specific differentiation markers MLC2 and TnIc (Fig. 1K). We designed the conditions for the RT-PCR reactions to be within the linear range to allow qualitative and quantitative analysis. We observe a comprehensive but not a complete reduction of cardiac differentiation markers because RT-PCR analysis is conducted on a pool of embryos, which presumably individually had been differently affected. A closer look at the RT-PCR results shows that the GATA6 tlMO causes more reduced expression of differentiation markers than either GATA4 or GATA5 spMOs (Fig. 1K) and the GATA4 tlMO causes stronger effects than the GATA5 tlMO, which had minimal effect on marker gene expression (data not shown; Peterkin et al.,2007). These data suggest different requirements of these three GATA factors during cardiogenesis, which indicates that they carry out slightly different functions.
The cardiogenesis-promoting activity of GATA factors (Latinkic et al.,2003; Afouda et al.,2008) is shared by Wnt11b (Eisenberg and Eisenberg,1999; Pandur et al.,2002; Terami et al.,2004). We, therefore, investigated whether Wnt11b could rescue the effect of depletion of individual GATAs on cardiogenesis. MOs inhibiting individual GATA factors were injected as before into the prospective cardiac mesoderm, but together with Wnt11b mRNA. The heart phenotypes in GATA4 and 6 morphants were rescued by Wnt11b (Fig. 1D and H). Interestingly, little or no rescue was observed in GATA5 spMO-injected embryos (Fig. 1F). The in situ hybridization data were corroborated with RT-PCR analysis performed on pools of injected embryos to monitor MLC2 and TnIc expression as above (Fig. 1K). Wnt11b significantly rescues the expression of cardiac differentiation markers only in GATA4 and 6 morphants, but not in GATA5 spMO-injected embryos, consistent with the in situ data (Fig. 1D and H). These results from whole-embryo experiments suggest that Wnt11b mediates functions for GATA4 and 6 during cardiogenesis, but not those for GATA5.
Different Requirements for GATA4, GATA5, and GATA6 During Cardiogenesis
GATA factors and Wnt11b have previously been shown to have functions during embryogenesis prior to heart development, such as germ layer induction and gastrulation (Tada and Smith,2000; Weber et al.,2000; Afouda et al.,2005). Such functions could indirectly influence heart development. In order to investigate in detail the specific function of each GATA factor specifically in cardiogenesis and the differential use of Wnt11b signaling downstream of GATA factors, we used stem-cell-like Xenopus animal cap explants. This assay allows us to study cardiogenesis in isolation to minimize the interference with other developmental processes when using intact whole embryos (Pandur et al.,2002; Ariizumi et al.,2003; Afouda et al.,2008; Afouda and Hoppler,2009). A newly developed protocol for inducing cardiogenesis in Xenopus animal cap explants relies on treatment with the peptide growth factor Activin (Ariizumi et al.,2003). In that study, the efficiency of induction was shown to depend on the concentration, treatment time, and number of cells present (Ariizumi et al.,2003). Expression of Activin mRNA in Xenopus animal cap explants has been shown to induce cardiogenesis as monitored by gene expression analysis by RT-PCR (Pandur et al.,2002). We have improved this technique, which allows us to uncover a gene regulatory network directing cardiogenesis (Afouda et al.,2008) and achieve a high efficiency differentiation into rhythmically beating cardiomyocyte tissue (Afouda and Hoppler,2009).
Our refined protocol involves delivery into all cells of the explant of the appropriate low amount of Activin they need to be fated toward the cardiogenic mesoderm by injecting 50 fg of Activin mRNA into the animal cap tissue prior to harvesting of explants. We then confirm initiation of cardiogenesis by monitoring induction of gene expression of early, first heart field markers such as GATA4, GATA6, Wnt11b, Nkx2-5, Tbx5, and later, cardiomyocyte differentiation markers such as MHC, MLC2, and TnIc (Fig. 2A) and also by phenotypic differentiation into functional beating cardiomyocyte tissue (Fig. 2B). In order to test whether the different GATA factors are differentially required during cardiogenesis, we used MOs designed to inhibit GATA4-6 expression in Activin mRNA-injected animal cap tissue. We initially used translation-blocking (tlMOs) and then splicing junction-blocking morpholinos (spMOs). We observe that cardiogenesis is quantitatively and qualitatively substantially reduced when either GATA4 or GATA6 factor function is inhibited by tlMOs (Fig. 3). The GATA5 tlMO has only a very weak effect (Fig. 3C). A closer look suggests that inhibition of GATA4 function has a stronger effect than inhibition of GATA6 function on early heart-field markers (such as Nkx2-5) and differentiation markers (such as TnIc) (Fig. 3A). A more complete inhibition of cardiogenesis is observed when both GATA4 and GATA6 are inhibited (Fig. 3A), suggesting some redundancy in the functions of these GATA factors during cardiogenesis. The phenotypic analysis of differentiation into functional beating cardiomyocyte tissue (Fig. 3B, D; and see Supp. Movies S1–4, which are available online) is in perfect agreement with these observed effects on cardiogenic gene expression. These findings in embryonic explants confirm those from whole-embryo experiments (Fig. 1) in suggesting that these GATA factors differ in their precise requirements in cardiogenesis. These factors may function at different steps of cardiogenesis.
GATA5 Functions in Regulating Cardiogenic Differentiation Markers
Our results so far are in broad agreement with published data about the roles for GATA factors in cardiogenesis, but further demonstrate that there are differences between them. Using previously described translation-blocking MOs (tlMOs) against all three cardiogenic GATA factors (Afouda et al.,2005), we did not uncover any major role for GATA5 in cardiogenesis, consistent with Peterkin et al. (2007). The efficacy of this GATA5 tlMO was, however, called into question with a recent study using splice-blocking MOs (spMOs) in Xenopus, which suggests a role for GATA5 in early cardiogenesis (Haworth et al.,2008). To corroborate our findings, we decided to use these newly described spMOs to inhibit GATA4 and GATA5 function (4spMO, 5spMO, Haworth et al.,2008) in Activin-expressing animal cap explants. Depletion of GATA4 with the spMO shows strong reduction in expression of both early and late cardiac markers (Fig. 4A). This effect is also reflected in a strong reduction of differentiation into beating cardiomyocytes (Fig. 4B). Using the GATA5 spMO has striking and strong effects resulting in reduction of expression of “late” differentiation markers (such as MHC, MLC2, and TnIc) but has no effect on the expression of “early” cardiogenic markers (such as GATA4 and6, Nkx2-5, Wnt11b, and Tbx5) (Fig. 4A). This specific strong reduction of differentiation markers is consistent with the observed dramatic reduction of differentiation into rhythmically beating cardiomyocyte tissue (Fig. 4B).
We, therefore, uncovered some important qualitative differences between 5spMO and 4spMO (Fig. 4). If they work as intended to deplete GATA5 and GATA4, respectively, then our experiments suggest some qualitative differences between GATA4 and 5 function in cardiogenesis. GATA4 function would be required for early heart field specification (which is a prerequisite for subsequent differentiation), and GATA5 more specifically for later cardiomyocyte differentiation. The similar effect of these spMOs at the phenotypical level (Fig. 4B) would, therefore, be explained by the fact that both 4spMO and 5spMO cause strongly reduced expression of cardiomyocyte differentiation markers (Fig. 4A). These results are consistent with our loss-of-function experiments in whole embryos (Fig. 1) and would further demonstrate that individual GATA factors regulate different steps of cardiogenesis.
The GATA4 and particularly the GATA5 spMO cause more severe phenotypes than their tlMO equivalents. However, depletion of GATA4 with this spMO causes stronger but clearly similar effects on cardiogenesis than those already shown in experiments using the tlMO (Fig. 3). The results with the GATA4 spMO, therefore, confirm the conclusions drawn from the previous experiments using GATA tlMO. The results with GATA5 spMO are, however, clearly different from those obtained with the GATA5 tlMO (Fig. 3). The obvious differences between the results obtained with the GATA5 spMO and the GATA5 tlMO may confirm previous doubts expressed about the efficiency of the original GATA5 tlMO (Haworth et al.,2008); however, we may also raise questions about the specificity of the GATA5 spMO and whether off-target effects could explain the observed effects on cardiomyocyte differentiation, particularly since we also observe some tissue disintegration at later stages when using the GATA5 spMO (see below). We decided that the specificity of these spMOs needed to be confirmed in rescue experiments (see below), particularly since Haworth et al. (2008) failed to reverse the effects caused by their splice-blocking GATA5 morpholino in their attempted rescue experiments.
GATA4 and GATA6 Functions Are Required Before GATA5 Activity During Cardiogenesis
We use depletion followed by rescue approaches to test the specificity of our GATA MO tools, additionally to help shed light on the question of whether different GATA factors can replace each other functionally and to inform about their individual contribution toward the complex regulation of cardiogenesis. We inhibited specific GATA functions with MOs in Activin-injected explants (as in Figs. 3 and 4) and reinstated different GATA function with hormone-inducible GATA proteins (Afouda et al.,2005) that were activated after the explants were excised. The use of inducible versions of GATAs will enable us to have a temporal control over the activity of the over-expressed proteins that could then be activated after the endogenous gene product has been inhibited by the injected MOs. We only ever observe a rescuing effect when GATAGR constructs are activated with dexamethasone (Fig. 5A) confirming that they are strictly inducible.
The reduction of all cardiogenic markers resulting from the inhibition of GATA function by the combined GATA4 and GATA6 tlMOs can be robustly rescued by either GATA4GR or GATA6GR mRNAs (Fig. 5A). There is even a weak rescue observed with GATA5GR mRNA, suggesting that over-expressed GATA5 can substitute, although only badly, for GATA4 and 6 function (see below). When the spMO was used to reduce endogenous GATA4, GATA4GR was able to rescue (Fig. 5B), and remarkably GATA5GR also rescues to some extent, especially cardiomyocyte differentiation markers. These data confirm that GATA5 protein, when activated at the stage when GATA4 is required in early cardiogenesis, can substitute to some extent for GATA4 function. A clear requirement for GATA5 in cardiogenesis was suggested when the spMO was used, as previously seen (Fig. 4), resulting in strong reduction specifically of differentiation markers (Fig. 5B). Injection of GATA5GR (50 pg mRNA) and early activation after the explants were excised (stage 8) was, however, unable to rescue the effect of the GATA5 spMO, consistent with the failed rescue observed by Haworth et al. (2008) and with the possibility that GATA5 spMO causes off-target effects.
The fact that at early stages GATA5 protein is not able to rescue a GATA5 loss of function phenotype, may not necessarily discredit the GATA5 spMO, since this result could also be explained by a function for endogenous GATA5 specifically at later stages of cardiogenesis, which is already suggested by a specific requirement of GATA5 for expression of late, differentiation markers in Figures 1, 4, 5. We, therefore, tested whether GATA5 needs to be activated at later stages in order to rescue GATA5 morphants. We carried out experiments in which GATA5GR (50 pg of mRNA) was expressed with 5spMO and Activin in explants with GATA5GR activated with dexamethasone early during gastrulation (at stage 8) or later during organogenesis (at stage 20) (Fig. 5C). GATA5GR was finally able to comprehensively rescue the effects caused by the GATA5spMO, but only when activated at later stages (stage 20, Fig. 5C). The assertion about the timing of the requirement of GATA5 during cardiogenesis is also supported by the fact that we observe some tissue disintegration in 5spMO-injected explants at later stages (data not shown). This observation could be interpreted either as a non-specific off-target effect of the 5spMO or that GATA5 is required for the cell adhesion process that has been described during heart morphogenesis (Garriock et al.,2005). Although the molecular mechanisms that underlie this phenomenon are currently unclear to us, we believe that this cannot be linked to a non-specific effect of 5spMO since this effect is also clearly rescued by resupplying GATA5 (data not shown). This might be indicative of a possible but not yet described role for GATA5 in regulating cell adhesion. In contrast, GATA4GR or GATA6GR rescue the effect of 4spMO or 6tlMO, respectively, when the over-expressed proteins are activated either at stages 8 or 20.
However, unlike GATA5GR, either GATA4GR or GATA6GR are also able to rescue expression of differentiation markers in GATA5 spMO explants, even when activated at early stages (Fig. 5B), suggesting that high overexpression of GATA4 and 6 can artificially circumvent any requirement for GATA5 function during later stages (see also Fig. 9). Additionally, we have already observed that GATA5GR, when activated at early stages, is to some extent able to substitute for GATA4 and 6, though somewhat inefficiently (Fig. 5A, B). We, therefore, also tested whether activation at early stages of a higher amount of GATA5GR could cause a rescue of GATA5 spMO phenotype. When higher amounts of GATA5 mRNA were injected (100 pg), we observed that any cardiac phenotype caused by GATA MOs is substantially rescued even when dexamethasone is added at stage 8 (data not shown). This observation is consistent with the notion explained above that at higher amount of GATA5GR can functionally replace the early activity of GATA4 and 6 during heart induction, and at high expression levels artificially circumvent the normal requirement for GATA5 during later stages.
These rescue data, firstly, indicate that the observed GATA5 spMO phenotypes are undoubtedly specific. Secondly, they indicate that the spMOs are clearly more efficient than their translation-blocking counterparts and, therefore, can be considered reliable tools for depleting GATA4 and GATA5 function in Xenopus. Thirdly, they also indicate that GATA5 function in cardiogenesis is different from that of GATA4 and 6, which precede the function of GATA5. This temporal rescue approach that we have adopted has allowed us to uncover a late function of GATA5 during cardiogenesis (cardiomyocyte differentiation) and could explain why the previous attempts to rescue the GATA5 spMO phenotype by GATA5 mRNA, which causes GATA5 to be active early, were unsuccessful (Haworth et al.,2008).
Our data shown that these three GATA factors have different roles during cardiogenesis. The results from Figure 5B and C suggest that endogenous GATA5 function is required after a prerequisite for GATA4 and 6 activities to allow differentiation to proceed. The GATA5 rescue observed when endogenous GATA4 is inhibited (Fig. 5B) supports this idea, since it specifically pertains to differentiation markers. Taken together, we conclude that GATA4, 5, and 6 are differently required within the complex gene regulatory network in cardiogenesis. The question about the different abilities for GATA factors to induce cardiogenic markers now becomes evident and needs to be addressed.
Different Abilities of GATA4, 5, and 6 to Induce Cardiogenic Marker Gene Expression
To test whether the different abilities of GATA4-6 are sufficient for inducing cardiogenic marker gene expression, we injected GATA4GR, GATA5GR, and GATA6GR mRNAs (hormone-inducible version of GATA4, 5, and 6) (Afouda et al.,2005) into animal caps and activated them with dexamethasone (Dex). All three factors were efficiently synthesized in explants (Fig. 6A), including both the short and the long isoforms of GATA6 that have previously been characterized (Brewer et al.,1999; Peterkin et al.,2003). We observed that approximately equivalent amounts of different GATA proteins are synthesized; yet we observe consistent differences in the induction of cardiogenic gene expression (Fig. 6B). GATA4 induces cardiogenic markers more effectively than GATA6, which in turn is more efficient than GATA5. A more detailed observation of the gene induction profile shows more subtle but reproducible qualitative and quantitative differences in the abilities of these factors to induce cardiogenic markers; GATA4 is a more potent inducer of early cardiogenic markers (such as Nkx2-5, Wnt11b, and Tbx5) and GATA6 is more potent at inducing cardiomyocyte differentiation markers (such as MLC 2 and TnIc). GATA5 is even less efficient at inducing early markers, but a potent inducer of late markers, although not as potent as GATA6. We, therefore, conclude that although all three factors, GATA4–6, have overlapping activities in that they are able to induce cardiogenic markers, they differ in their contributions to the expression of different cardiogenic genes in this assay.
The observed differences in the abilities of these factors to induce cardiogenic markers are consistent with our loss of function results above (Fig. 1K). The data from these experiments suggest that GATA4 (and with less functional activity also GATA6) function during the early phases of cardiogenesis before GATA5 and 6, which are clearly required for differentiation. These results confirm that we have uncovered functionally different roles for GATA4–6 during cardiogenesis especially during the formation of the first heart field (FHF). The power of the animal cap explant system is highlighted here since the results obtained using this assay (Fig. 6) validated those from whole embryo experiments (Fig. 1), without the concerns about indirect effects caused by germ layer induction and gastrulation defects. It will be interesting to investigate the molecular mechanisms that underlie the activity differences and the differential requirements for GATA4-6.
Wnt11b Activity Mediates the Differential Requirements Of GATA Factors Functions During Cardiogenesis
Our experiments carried out in whole embryos show that Wnt11b signaling can only rescue GATA4 and 6 but not GATA5 morphants (Fig. 1). A complete inhibition of cardiogenesis was observed in our animal cap explant assays with combined injection of both GATA4 and 6 tlMO (Fig. 3). We decided to test whether this effect could be rescued by experimentally reinstating Wnt11b expression. Indeed, Wnt11b mRNA robustly rescues reduced cardiogenesis caused by inhibition of both GATA4 and GATA6 as monitored by gene expression (Fig. 7A) as well as by observing beating cardiomyocyte differentiation (Fig. 7B). Possible functions for Wnt11b activity downstream of GATA5 were also investigated. Strikingly, Wnt11b mRNA is unable to rescue any effect of GATA5 spMO either on late cardiogenic gene expression (Fig. 7C) or functional cardiomyocyte differentiation (Fig. 7D). We, therefore, conclude that non-canonical Wnt11b activity is among the mechanisms used by only some cardiogenic GATA factors to exert their functions during cardiogenesis.
Non-Canonical Wnt Signaling Mediates the Differential Requirements of GATA Factors Functions During Cardiogenesis
Because Wnt11b activation of non-canonical Wnt signaling pathway was previously reported to be required for cardiogenesis (Pandur et al.,2002), we investigated the involvement of Dishevelled, a key molecule mediating Wnt signaling (Wallingford and Habas,2005; Eisenberg and Eisenberg,2006). We fused a Dishevelled truncation known to specifically activate the non-canonical Wnt pathway (Rothbacher et al.,2000; Wallingford and Habas,2005) to the ligand binding domain of the human glucocorticoid receptor (GR). The obtained fusion construct called dshΔDIXGR was first tested and found to specifically activate non-canonical Wnt signaling (Afouda and Hoppler, unpublished data; see also Smalley et al.,2005). Interestingly, reduction of cardiogenic marker expression caused by inhibition of GATA4 and GATA6 is rescued when the non-canonical pathway is activated, both in animal cap and dorsal marginal zone explants (Fig. 8A). In contrast, reduced cardiac differentiation marker expression caused by GATA5 inhibition cannot be rescued when the non-canonical Wnt signaling is activated, either from the early stages when the explants were excised (Fig. 8B and C) or at later stages during organogenesis (Fig. 8D and E). Our experiments use for the first time research tools that are able to activate non-canonical Wnt signaling in an inducible manner to demonstrate that it is able to promote cardiogenesis. Our results suggest that non-canonical Wnt11b signaling is involved in mediating the differential requirements for GATA factors during cardiogenesis.
Mounting evidence indicates that early embryonic development of the heart is comparable in different vertebrate species and that the fundamental molecular mechanisms regulating cardiogenesis are also evolutionarily conserved among vertebrates (Bruneau,2002; Zaffran and Frasch,2002). Thus, the Xenopus embryo has proven to be an excellent model system for analyzing vertebrate cardiogenesis. We exploited the pluripotency of animal pole cells from Xenopus blastulae to investigate the functional differences between cardiogenic GATA transcription factors.
Advantages of Using Activin-Injected Xenopus Animal Caps as an Efficient and Reliable Cardiogenic Assay
Ariizumi and colleagues have developed an experimental protocol based on treating animal cap cells with Activin A solution that allows driving differentiation into beating cardiomyocyte tissue (Ariizumi et al.,2003). Ectopic transplantation of such an induced cardiac tissue into a host embryo remarkably results in the formation of a functional heart with a developed atrium and ventricle connected to the host vascular system (Ariizumi et al.,2003). However, this protocol presents some limitations because it requires the key step of dissociation and re-aggregation of the animal cap explant. In addition, the concentration of Activin, the number of cells in the re-aggregates, as well as, the treatment time are all critical factors that need to be just right for successful cardiogenesis. Our refined protocol used here represents a simple yet reliable method based on injection into all of the animal cap cells of the very low amount (50 fg) of Activin mRNA that is needed to be fated into cardiac mesoderm. By doing so, we bypass the requirement for any dissociation and re-aggregation steps as in the Ariizumi protocol.
Although this protocol is an artificial one, we believe that it does carry some advantages that merit mentioning. It allows obtaining beating cardiomyocyte tissue from a single explant in which the functional activities of factors can be studied without interfering with early embryogenesis such as gastrulation. Using this assay, we were able to carry out rescue experiments to confirm the specificity and efficacy of the MO tools, which gave us more insight into the roles of GATA factors and Wnt11b in cardiogenesis (Figs. 5, 7, 8). Therefore, the cardiogenic protocol used here is powerful for dissecting molecular mechanisms that underlie formation of the different cell types in the vertebrate heart. This ultimately will be relevant for developing therapeutic strategies for repairing damaged adult heart muscle.
Redundancy and Functional Differences Between GATA Factors During Cardiogenesis
Our investigation addresses the previous controversy about GATA5 function in vertebrate cardiogenesis. Knockout and overexpression experiments have revealed important roles for GATA5 during Zebrafish cardiogenesis (Reiter et al.,1999), but not in mouse as suggested by the GATA5 mutant mice (Molkentin et al.,2000). Furthermore, initial analysis of GATA5 function in Xenopus using a translation-blocking MO (tlMO) had suggested no role in cardiogenesis (Peterkin et al.,2007), which, however, was called into question in a recent study that used a newly designed splice-blocking MO (spMO) (Haworth et al.,2008) to suggest an important role for Xenopus GATA5 during early embryogenesis. The GATA5 knockout mutant mice mentioned above may not represent a true null due to the formation of a truncated protein that contains the DNA-binding domain, as reported in Peterkin et al. (2007). Consistent with that, mouse GATA5 was shown to be important for formation of endothelial-endocardial cells during cardiogenesis (Nemer and Nemer,2002). It is, therefore, important to address the questions about the efficiency of MO tools used in Xenopus for depletion of GATA5 and about what the Xenopus experiments are able to tell us about the conserved role of GATA5 during vertebrate cardiogenesis.
Our experimental approach used stage-specific rescue of MO-mediated depletion of cardiogenic GATA factors to uncover the requirement for GATA5 function during late cardiogenesis (see below). Thus, our results have clarified the issues surrounding the efficiency of the translation and splicing-blocking MOs as tools for depleting GATA5 function and informed us about the role of GATA5 in cardiogenesis. Our results demonstrate a prominent role for GATA5 during cardiogenesis in both Xenopus and Zebrafish and, therefore, suggest a conserved role during vertebrate heart development. Notwithstanding a role in early embryogenesis (Haworth et al.,2008), our results argue for a function for GATA5 during cardiomyocyte differentiation, consistent with our model depicted in Figure 9.
In previous studies, we found that either GATA4 or GATA6 can initiate cardiogenic gene expression, but only GATA4 induces full cardiogenesis capable of leading to differentiation of beating cardiomyocyte tissue (Latinkic et al.,2003; Afouda et al.,2008). In this study, we uncovered preferential regulation of heart-field markers (such as Nkx2-5, Wnt11b, and Tbx5) or cardiomyocytes differentiation markers (such as MHC, MLC2, and TnIc) by different cardiogenic GATA transcription factors, which argues for qualitative activity differences (Fig. 6). While GATA4 (and with less functional activity also GATA6) function in early heart field specification, both GATA6 and particularly GATA5 function specifically in later cardiomyocyte differentiation. Encouragingly, these conclusions are supported by both our gain- and our loss-of-function results (Figs. 3 and 4), and to some extent by the loss-of-function experiments in whole embryos (Fig. 1). Our experiments are, however, also consistent with the previously described partially redundant roles for GATA factors in the regulation of cardiogenesis (Holtzinger and Evans,2007; Peterkin et al.,2007).
Differential Requirements for GATA Factors Are Mediated by Non-Canonical Wnt Signaling
Surely several molecular mechanisms could explain the differential requirements for GATA4, 5, and 6 in cardiogenesis that we have uncovered. Our data strongly support that non-canonical Wnt11b signaling is among the molecular mechanisms that underlie the different functions for these factors during cardiogenesis. We found that loss of GATA4 and 6 functions can be rescued to a large extent by injection of Wnt11b mRNA (Figs. 1 and 7A,B). In contrast, inhibition of cardiac differentiation as results from loss of GATA5 function cannot be rescued by injection of Wnt11b mRNA (Figs. 1 and 7C,D). Interestingly, because Wnt11b function in cardiogenesis is mediated by non-canonical Wnt signaling (Pandur et al.,2002), the observation that experimental activation of non-canonical Wnt signaling is also only able to rescue the effects of GATA4 and 6 depletion but not those from GATA5 confirms the notion that Wnt11b mediates different functions of GATA factors during cardiogenesis (Fig. 8). It also suggests that no other potential non-canonical Wnt signaling ligand, such as Wnt11r (Garriock et al.,2005) mediates the different requirement for GATA5. In support of this assertion, experimental activation of non-canonical Wnt signaling during organogenesis, when Wnt11r is supposed to be active in cardiogenesis (Garriock et al.,2005), is still unable to rescue the effect of GATA5 depletion (Fig. 8D and E).
Our data confirm the previously suggested cell non-autonomous functions of GATA4 and GATA6 during cardiogenesis (Narita et al.,1997b; Peterkin et al.,2003; Afouda et al.,2008). In our previous investigation, we proposed a regulatory pathway controlling cardiogenesis, integrating cardiogenic GATA transcription factors with canonical and non-canonical Wnt signaling (Afouda et al.,2008). The data presented here provide additional evidence for a model in which non-canonical Wnt signaling mediates the function during cardiogenesis of GATA4 and 6, but not of GATA5 (Fig. 9).
Whatever the molecular mechanism involved in mediating these different functions of GATA factors, this will probably involve differences in the co-factors that associate with each GATA transcription factor, subtle differences in their DNA-binding site preferences (Sakai et al.,1998), and additional mechanisms connected to the N-terminal protein domain that is present in GATA6, but not in GATA4 and 5 (Brewer et al.,1999). The crucial role of GATA4 in mammalian cardiogenesis has been highlighted in recent studies. Takeuchi and Bruneau (2009) show that GATA4 acts together with Tbx5 and Baf60c to direct ectopic differentiation of mouse mesoderm into beating cardiomyocytes. And more recently Ieda et al. (2010) have reported that GATA4 is among a combination of three developmental transcription factors (i.e., with Mef2c and Tbx5) that rapidly and efficiently reprogramme post-natal cardiac or dermal fibroblasts directly into differentiated cardiomyocyte-like cells.
Our simple cardiogenic assay using Activin-injected animal cap explant tissue, which is stem-cell-like, can be exploited to address further molecular mechanisms and cellular processes that underlie the different activities of GATA factors in cardiogenesis.
Expression Constructs and Morpholinos
GATA4GR, GATA5GR, GATA6GR, and Activin (Afouda et al.,2005) as well as Wnt11b (Tada and Smith,2000) mRNA expression constructs have been previously described. GATA4, 5, and 6 translation-blocking morpholino antisense oligonucleotides (tlMOs) (Afouda et al.,2005; Peterkin et al.,2007) and GATA4 and 5 splice junction-blocking morpholinos (spMOs) (Haworth et al.,2008) have been previously characterized. Twenty nanograms (GATA4, GATA5) and 10 ng (GATA6) of tlMOs, and 50 ng (GATA4, 4spMO) and 8 ng (GATA5, 5spMO) of spMO were injected per embryo.
The dshΔDIXGR is a hormone-inducible construct of the Dishevelled protein without the DIX (nucleotides 1 to 480) domain (Afouda and Hoppler, in press; Smalley et al.,2005). All injected RNAs were synthesized using mMESSAGE mMACHINE kits (Ambion, Austin, TX).
Embryos and Explants
Xenopus embryo harvesting, explant culture, and embryo injection were as in Afouda et al. (2008). Cardiomyocyte differentiation of explants was monitored with molecular analysis of gene expression by RT-PCR (see below), whole-mount RNA in situ hybridization (see below), and phenotypic analysis by differentiation into rhythmically beating tissue as in our previous study (Afouda et al.,2008).
RNA Expression Analysis by RT-PCR
RNA expression of marker genes associated with heart development was determined by RT-PCR as previously described (Afouda et al.,2008). To allow qualitative and quantitative analysis, RT-PCR was performed so as to be within the linear range (see linearity control in Figs. 1–8). Genes known as early heart-field markers (cardiac fate) such as Nkx2-5 (Tonissen et al.,1994), GATA4 and 6 (Jiang and Evans,1996) and Tbx5 (Horb and Thomsen,1999) or as cardiomyocyte differentiation markers, such as MHC (Logan and Mohun,1993), TnIc (Drysdale et al.,1994), and MLC2 (Chambers et al.,1994) were used for assessing cardiac development. Primers for detection of the splicing defects caused by GATA4 or GATA5 spMOs are as previously described (Haworth et al.,2008). Sequences of the oligonucleotide primers and PCR conditions are listed in Table 1. All experiments were repeated at least three times for reproducibility, except Figure 8D,E.
Table 1. Sequences of Primers Used for RT-PCR Analysis of Gene Expression
Primer sequences (5′→3′)
No of cycles
PCR was carried out with 0.3 M MgCl2 for TnIc and 0.1 M MgCl2 for all other markers.
Spatial expression of the heart muscle differentiation marker TnIc was analyzed by whole-mount RNA in situ hybridization with the protocol previously described (Afouda et al.,2005). A digoxigenin-labelled RNA probe was prepared with T7 polymerase, using High Yield Megascript kit (Ambion) from NotI-linearized plasmid template for TnIc (Drysdale et al.,1994).
We thank the anonymous reviewers and John Monaghan (University of Aberdeen) for valuable comments and suggestions. We also thank Stuart A Duncan and Yvonne Turnbull for technical assistance.