December 1, 1999;
A role for GATA-4/5/6 in the regulation of Nkx2.5 expression with implications for patterning of the precardiac field.
Interactions between the key regulatory genes of the cardiogenic pathway, including those from the GATA and Nkx2 transcription factor families, are not well defined. Treating neurula
-stage Xenopus embryos with retinoic acid (RA) causes a specific block in cardiomyocyte
development that correlates with a progressive reduction in the region of the presumptive heart
-forming region expressing Nkx2.5. In contrast, RA does not block expression of the GATA-4/5/6 genes, which are transcribed normally in an overlapping pattern with Nkx2.5 throughout cardiogenesis. Instead, GATA-4/5/6 transcription levels are increased, including an expansion of the expression domain corresponding to lateral
that is part of the early heart
field, but that normally is progressively restricted in its ability to contribute to the myocardium
. GATA-dependent regulatory sequences of the Nkx2.5 gene that implicate GATA-4/5/6 as upstream positive regulators were described recently. However, our experiments also indicate that GATA factors might normally antagonize transcription of Nkx2.5. To test this hypothesis we generated a dominant negative isoform of GATA-4 (SRG4) capable of inhibiting transcription of GATA-dependent target genes. Ectopic expression of SRG4 results in a transient expansion of the Nkx2.5 transcript pattern, indicating that a normal function of GATA factors is to limit the boundary of the Nkx2.5 expression domain to the most anterior ventral
region of the heart
field. Regulatory mechanisms altered by excess RA must function normally to limit GATA-4/5/6 expression levels, to define the region of Nkx2.5 expression and regulate myocardial differentiation.
[+] show captions
FIG. 1. GATA-4/5/6 transcription is not inhibited by RA treatment, which blocks myocardial differentiation. (a) Whole-mount in situ
hybridization was performed using probes specific to transcripts for cardiac troponin I, Nkx2.5, or GATA-4/5/6, as indicated on the right.
Embryos are oriented with anterior to the left, dorsal to the top. RA was added at stage 23 and embryos were allowed to develop until stage
35. Examples of control embryos, treated with DMSO alone, are shown on the left; examples of RA-treated embryos are on the right. The
arrow and arrowheads in the two upper right images indicate the small patch of differentiated myocardium that remains in these embryos,
at the most ventral position of the presumptive heart-forming region. Note that the pharyngeal endoderm, which does not express
GATA-4/5/6, remains positive for Nkx2.5 following RA treatment. The Nkx2.5 levels appear increased in pharyngeal endoderm of this
RA-treated embryo due to extended staining in order to observe the weak signal in the residual heart region. However, enhanced levels of
Nkx2.5 RNA are not observed generally in pharyngeal endoderm of RA-treated embryos (Drysdale et al., 1997). (b) An example of an
RA-treated embryo analyzed for xGATA-6 RNA to illustrate the induction of a posterior domain of transcription (arrow). A similar
activation is seen for GATA-5 but not for GATA-4. Our previous work implicated GATA-6 in the regulation of progenitor cells of the gut
(Gao et al., 1998), and so this domain is likely to correspond to the presumptive hindgut.
FIG. 2. GATA-4/5/6 transcription is maintained and enhanced in both the mesoderm and the endoderm of RA-treated embryos. (a) Sagittal
sections taken from embryos analyzed first by whole-mount in situ hybridization for Nkx2.5 or GATA-4/5/6 (top to bottom). Embryos are
oriented as in Fig. 1; the cement gland is visible as brown tissue (adjacent to the Nk in the upper left image). Sections from control
DMSO-treated embryos are shown on the left; samples from RA-treated embryos are on the right. The position of the folded heart (ht) is
indicated on the left. The myocardial (arrowhead) and associated endoderm layer (En), both of which are positive for GATA-4/5/6, are
indicated on the right. (b) Cross section of a similarly treated embryo analyzed for GATA-4 transcripts. Dorsal is to the top. The endocardial
(EC) layer is indicated for both control (top) and RA-treated (lower) samples.
FIG. 3. Earlier treatment of embryos with RA results in further expansion of the GATA-4 transcript pattern. (a) Embryos were analyzed
(and are oriented) as in Fig. 1a (except that the embryos were fixed at stage 30) for GATA-4 (left) or Nkx2.5 (right) transcripts following
continuous treatment of RA added at stage 10.5 (top), stage 14 (2nd from top), or stage 20 (3rd from top). Embryos on the bottom were
controls treated only with DMSO. Even in embryos with anterior truncations the GATA-4 gene is transcribed in an expanded region of the
lateral plate. GATA-5 and GATA-6 were also transcribed in these embryos, but did not show such an obvious expansion of the pattern. (b)
Cross sections are shown of embryos treated with RA beginning at stage 10.5 and analyzed for GATA-4 transcripts as in (a). Compared to
the control embryos (left), the GATA-4 transcript pattern is enhanced and expanded in the RA-treated embryos (middle). To confirm the
identity of the endoderm layer, embryos were also analyzed for transcripts from the Sox17a gene. This pattern was unaffected by RA
treatment (not shown). Sections were taken either from an anterior (A) position (top) or from a more posterior (P) position caudal to the heart
proper (bottom). (c) Embryos were treated with RA from stage 10.5 and RNA was harvested from embryos after 4 h (left lanes) or at the
equivalent of stage 30 (right lanes) and analyzed for GATA-4 or control EF-1a transcript levels by semiquantitative RT/PCR. Consistent
with the signals detected by in situ hybridization, the steady-state transcript levels for GATA-4 were increased 100% by RA treatment.
FIG. 4. SRG4 functions as a dominant negative isoform capable of inhibiting activation of GATA-4 targets. (a) Schematic illustration of
the domain structure of SRG4. The strong repression (SR) domain of Mxi.1 is fused to a region of GATA-4, containing both zinc fingers (cc
cc), that constitutes the DNA-binding and cofactor interaction domains of GATA-4. A control isoform (mtSRG4) was also generated,
containing a single amino acid change (leucine to proline, p) that abolishes the ability of SR to repress gene expression. SRG4 is predicted
to repress GATA-dependent target genes through interaction with normal GATA-binding sites containing a WGATAR consensus sequence.
(b) A gel mobility-shift experiment demonstrates that SRG4 binds with specificity to a consensus GATA cis-element. The GATA-binding
probe (P) was end-labeled and incubated with rabbit reticulocyte lysate charged with in vitro-generated RNA encoding SRG4. The reaction
included no competitor (0) or a 50-fold molar excess of unlabeled oligomer containing the same specific (sp) binding site or a similar
oligomer containing a mutation (mt) of the GATA motif. (c) A well-characterized GATA-dependent reporter (aD3-luc) was cotransfected
into QT6 fibroblasts with (1) or without (0) a GATA-4 expression vector, along with an empty vector (to maintain total amount of vector
DNA), mtSRG4, or SRG4 as indicated. The level of activation relative to the basal promoter activity is shown after normalizing for activity
of a cotransfected b-galactosidase reporter that is not regulated by GATA-4 expression. The results shown are from three independent
experiments, each performed in duplicate; red stacked bars represent standard deviation. SRG4 was equally capable of repressing activation
by xGATA-5 or xGATA-6 (not shown).
FIG. 5. Ectopic targeted expression of SRG4 results in an expanded pattern of Nkx2.5 transcription. Examples are shown of embryos
injected at the 8-cell stage into one left-side dorsal–vegetal blastomere with RNA encoding control LacZ RNA, mtSRG4, or SRG4, as
indicated. Embryos were fixed at stage 26 and analyzed by whole-mount in situ hybridization for Nkx2.5 transcripts; processed embryos
are viewed from the ventral side, anterior is up. The control embryo is slightly older so the right (R) and left (L) sides of the primitive heart
tube are fusing at the midline. Compared to the symmetric pattern seen here or in the prefusing primordia of the mtSRG4-injected embryo,
the SRG4-injected embryos display an expanded domain of Nkx2.5 transcription in the left primordium (arrows). (b) An embryo injected
in the left blastomere with SRG4 RNA was processed as in (a), and sections were taken through the region of the fusing primordia. Shown
are four sections taken from progressively more posterior regions (left to right). Note that in the most anterior regions the Nkx2.5 pattern
is not significantly changed, but left-side enhanced transcripts are detected in more posterior sections (arrows).
FIG. 6. Based on the data presented, we hypothesize that the levels or boundaries of GATA-4/5/6 expression delineate the region of the
heart field that will express Nkx2 genes and be competent for cardiomyocyte development. When GATA factors are overexpressed,
cardiomyocyte differentiation can be repressed. Inhibition of GATA-dependent gene expression results in an expanded pattern of Nkx2.5
transcripts, perhaps providing competence to a larger part of the heart field to contribute to the cardiomyocyte population. The model is
consistent with additional reports in the literature demonstrating the ability of enhanced levels of GATA-6 (Gove et al., 1997) or Nkx2.5
(Cleaver et al., 1996) to inhibit or expand, respectively, the population of differentiated cardiomyocytes (although note that this is a
relatively late effect and neither GATA-4 or GATA-6 overexpression alone is reported to alter Nkx2.5 transcript patterns). The domain of
Nkx2 expression in the pharyngeal endoderm is not indicated in the schematic, although we note that RA does not activate GATA
expression in the pharyngeal endoderm and also does inhibit Nkx2.5 transcription in this region.