Retinoic acid-mediated gene activation is important for normal vertebrate development. The size and nature of retinoic acid make it difficult to identify the precise cellular location of this signaling molecule throughout an embryo. Additionally, retinoic acid (RA) signaling is regulated by a complex combination of receptors, coactivators, and antagonizing proteins. Thus, in order to integrate these signals and identify regions within a whole developing embryo where cells can respond transcriptionally to retinoic acid, we have used a reporter transgenic approach. We have generated several stable lines of transgenic zebrafish which use retinoic acid response elements to drive fluorescent protein expression. In these zebrafish lines, transgene expression is localized to regions of the neural tube, retina, notochord, somites, heart, pronephric ducts, branchial arches, and jaw muscles in embryos and larvae. Transgene expression can be induced in additional regions of the neural tube and retina as well as the immature notochord, hatching gland, enveloping cell layer, and fin by exposing embryos to retinoic acid. Treatment with retinoic acid synthase inhibitors, citral and diethylaminobenzaldehyde (DEAB), during neurulation, greatly reduces transgene expression. DEAB treatment of embryos at gastrulation phenocopies the embryonic effects of vitamin A deprivation or targeted disruption of the RA synthase retinaldehyde dehydrogenase-2 in other vertebrates. Together these data suggest that the reporter expression we see in zebrafish is dependent upon conserved vertebrate pathways of RA synthesis.

The retinoic acid receptors (RARs or rars) and the thyroid hormone receptors are members of the steroid receptor superfamily that interact with their DNA response elements (for RARs: retinoic acid response elements or RAREs) in the regulatory regions of promoters in the absence of their ligand. In this ligand minus configuration, it has been suggested that the RAR provides a binding site for a corepressor (SMRT or N-CoR) that also brings in other proteins to repress the gene. In the presence of the ligand, the receptor goes through an allosteric change eliminating the corepressor binding site and providing a coactivator binding site. In this manuscript we describe the isolation of the zebrafish corepressor, smrt. We show that its association with the zebrafish rar aa is sensitive to retinoic acid and that the corepressor mRNA is present in 8 cell zebrafish embryos – a time at which the embryonic genome is not active. We suggest that this rar–corepressor complex may be part of an embryonic, epigenetic switch that keeps retinoic acid responsive genes off before retinoic becomes available to the embryo.