GATA-4/5/6 GENE REGULATION AND HEART DEVELOPMENT

We are interested in understanding how developmental events are determined at the transcriptional level. For the past several years, we have sought to gain insights into heart development by analyzing GATA-4/5/6 transcription factor gene expression and function. These genes are initially expressed in the cardiogenic regions of early stage embryos and continue to be expressed in the heart, and in the gut, during later stages of development. Although these genes are expressed throughout the heart, we have determined that this regulation is effected by sets of modular enhancers that function in different regions of the heart and/or at different stages of heart development. For example, one GATA-6 enhancer is activated in the posterior heart field and remains active in several derivative cell types in the heart, whereas another enhancer is transiently activated in the anterior region of the developing heart. We also mapped a GATA-5 enhancer that provides evidence for a possible contribution of epicardial cells to valve morphogenesis. We are analyzing these enhancers further to gain molecular insights into several interesting facets of heart development that may have clinical relevance.

A modest number of heart-restricted transcription factors have been identified; various combinations of these factors have been shown to regulate the expression of numerous heart-specific genes. Among this set of heart-restricted factors, Nkx2.5 and three GATA factors (GATA-4/5/6) stand out as being expressed very early in the cardiogenic lineage. Several years ago, we set out to determine how the genes for two of these factors (GATA-5/6) are regulated to gain insights into the transcriptional hierarchies that function upstream of, or otherwise modulate, GATA gene expression in the heart.

At the time of our last report, we had assayed two overlapping GATA-6 genomic fragments for their ability to direct the expression of a lacZ reporter gene in F0 (transient) transgenic mouse embryos. Two qualitatively distinct heart-specific elements were indicated by this preliminary analysis (He and Burch, J. Biol. Chem. 272:28550, 1997.) The activity of the more proximal element was restricted to myocardial cells in the atrioventricular (AV) canal at the single developmental stage assayed (embryonic day 10; E10). Interestingly, cells in this region are known to play a critical role in the process of valve formation. The more distal element functioned in two other segments of the primitive heart at this developmental stage (E11); namely, the primitive ventricle and outflow tract.

This past year, we carried out additional F0 transgenic assays to further delineate these GATA-6 elements. We determined that these elements are enhancers that can also function in conjunction with a heterologous promoter. We established several GATA-6/lacZ transgenic mouse lines in order to assay the activities of the proximal and distal enhancers at different stages of development; the GATA-6 promoter does not possess activity on its own when assayed in transgenic mice.

The proximal enhancer was found to be activated very early in the cardiogenic program (by E7.5), specifically in the posterior region of the cardiogenic field. This heart region-specific expression pattern was surprising because most genes become heart region-specific after first being expressed throughout the primitive heart. Whereas one mouse gene is known to be restricted from the outset to the anterior heart-forming region, we believe that our GATA-6/lacZ transgene is the first example of a complementary expression pattern. Virtually nothing is known about how the cardiogenic field is patterned at this early stage. We hope to exploit our novel transgenic mice to gain molecular insights into this phenomenon.

The proximal enhancer remains active in restricted regions of the heart as development ensues. The transgenic lines that contain this enhancer display the expected pattern of AV canal-restricted expression pattern at E11. Interestingly, over the next few days of development, this expression pattern becomes restricted to cells that elaborate a number of neuronal markers and comprise a component of the conduction system. These results are suggestive of a lineage relationship, although this remains to be proven.

An analysis of deletion constructs revealed a 0.3 kilobase (kb) region that is necessary for the activity of the proximal enhancer. This critical region is embedded within a 0.6 kb region that is sufficient to direct expression in the AV canal. Our goal is to map and identify essential factor binding sites in the critical 0.3 kb region of this enhancer. Our working model is that the activity of this enhancer is dictated by the combination of a heart-specific factor(s) and a factor(s) that is unique to this region of the heart, but is not necessarily heart specific.

Interestingly, the distal GATA-6 enhancer directs expression in the anterior portion of the very early heart-forming region (by E7.5). Indeed, this GATA-6/lacZ transgene is one of the earliest markers for these cardiogenic cells. Much to our surprise, however, this enhancer does not remain active after chamber formation, except in small and dispersed clusters of heart cells. We have not yet precisely mapped this enhancer.

Since the evolutionarily related GATA-5 and GATA-6 genes are co-expressed in many tissues, including heart and gut, we reasoned that a transgenic analysis of GATA-5/lacZ constructs might reveal enhancers analogous to the GATA-6 enhancers described above. This proved, however, not to be the case. Rather, we have found evidence for two distinct types of enhancers upstream of the GATA-5 gene. The more distal of these enhancers is gut-specific and is remarkable in that it affords a very early marker of definitive endoderm. The other enhancer functions in the septum transversum, in the epicardium, which derives from the septum transversum and constitutes the outer layer of the heart, and in cells within the heart that appear to derive from the epicardium. Remarkably, cells that express the GATA-5/lacZ transgene can be found within the cushions where the valves will form, suggesting that epicardially derived cells may play a role in valve formation. In support of this concept, the heart valves of GATA-5/lacZ neonates stain brilliantly. While these results are striking, it remains to be determined whether the transgene is a true lineage marker in this context.

Whereas most other families of non-LTR retrotransposons have poly(A) tracts or other simple repeats at their 3' ends, the 3' ends of CR1 elements conform to the consensus [(CATTCTRT)(GATTCTRT)_1-3]. It is also noteworthy that CR1-like elements have been found in other vertebrate classes, such as amphibia, reptiles and fish, and even in a few invertebrate species. Thus, CR1 elements are prototypes for a novel and ancient family of vertebrate non-LTR retrotransposons.

Our derivation of a 4.6 kb full-length CR1 consensus sequence revealed that the pol-like ORF contains an endonuclease domain in addition to a reverse transcriptase domain (Haas et al., Gene 197:305, 1997.) This suggests that CR1 elements retrotranspose by a "nick and prime" mechanism similar to several other non-LTR elements that have been analyzed. However, the details must be distinct to account for the fact that CR1 elements have novel 3' ends. We also delineated another ORF that encodes what appears to be a novel bipartite domain comprised of a zinc finger followed by a leucine zipper. This motif is a conserved feature of CR1 and CR1-like elements. We speculate that this domain may function as a dimeric, or multimeric, nucleic acid binding domain, perhaps recognizing the conserved 3' untranslated region (UTR) to facilitate priming.

We made a concerted effort over the past year to find a master element that, by definition, can direct new retrotransposition events. To this end, we devised a PCR-based protocol to screen for CR1 elements whose ORF1 and ORF2 are not only intact, but have several domains that are known to be conserved among CR1 and CR1-like elements from different vertebrate classes. This strategy has yielded several promising clones. It is noteworthy that one of these CR1 elements encodes a reverse transcriptase that can function in a yeast assay.

It has been inferred, on the basis of sequence comparisons, that at least six master elements have contributed to the large number of CR1 elements in the chicken genome. Our previous analysis yielded a consensus sequence for one of these presumptive master elements (denoted CR1-B). This past year we derived the consensus sequence for another presumptive master element (denoted CR1-F). Whereas the coding and 3' UTR sequences for these CR1-B and CR1-F elements are conserved, the 5' UTR sequences are unrelated. We infer that the multiplicity of master elements is due to intact ORF1/ORF2/3' UTR copies of one master element having retrotransposed downstream of other functional promoters. A similar proposal has been advanced to account for the distinct subfamilies of mouse L1 non-LTR elements.

JIANG, Y., TARZAMI, S., BURCH, J.B.E., EVANS, T. Common role for each of the cGATA-4/5/6 genes in the regulation of cardiac morphogenesis. Dev. Genet. 22:263-277, 1998.

^a   C. MacNeill: Present address-Department of Obstetrics and Gynecology, The Pennsylvania State University, College of Medicine, Hershey, PA 17033

^b   A. Wessels: Department of Cell Biology and Anatomy, Medical University of South Carolina, Charleston SC 29425

Illustrations or unpublished data in these reports should not be used without permission of the author.

Fox Chase Cancer Center

Scientific Report 1998