A focus of interest in this laboratory has been the study of genes that are involved in the formation of the neural crest and of its derivatives such as pigment cells, pharyngeal arches, and others. We found that the BTB domain-containing protein Kctd15 that is first expressed in the embryo at the neural plate border, may have a role in regulating the domain in which neural crest forms as overexpression of Kctd15 strongly inhibits neural crest specification. We have found that Kctd15 inhibits the activity of transcription factor AP-2, a regulatory molecule known to be critical for the specification of the neural crest in different animals. Mechanistic studies on the interaction between AP-2 and Kctd15 have been carried out. Ap-2 is a major regulator of neural crest formation, being involved in the initial induction of neural crest precursor cells as well as in later stages such as their migration and differentiation into multiple derivatives. Kctd15 is capable of binding AP-2 in co-immunoprecipitation experiments, and inhibits the activation of an AP-2 reporter both in cultured cells and in zebrafish embryos. In studying the mechanism of this inhibition we found that Kctd15 does not affect the level of AP-2 in the cell, does not inhibit dimerization of AP-2 which is required for activity, and does not affect nuclear localization of AP-2. Also, AP-2 remains bound to its cognate sites in chromatin in the presence of Kctd15. To probe the mechanism further we studied the activation domain of AP-2. Kctd15 binds to the activation domain, and specifically requires proline 59 for this interaction. An AP-2 mutant in which this proline is changed to alanine (P59A) is active in the reporter assay but cannot bind Kctd15, and its activity is not effectively inhibited by Kctd15. We conclude that Kctd15 inhibits AP-2 activity by specific binding to its activation domain which precludes its function. In further studies on Kctd15 we found that this protein is SUMOylated in vivo, both by SUMO1 and SUMO2/3. We identified the site of SUMO attachment and showed that mutation of the relevant lysine residue abolished SUMOylation. Non-SUMOylated Kctd15 was fully effective in inhibiting AP-2 function, but a stable fusion protein of Kctd15 and SUMO was not effective. We conclude that the non-SUMOylated form of Kctd15 functions in AP-2 inhibition, but leave open the possibility that the SUMOylated form has a role in a different biological function of Kctd15. In the project aiming to characterize pineal gland transcriptional regulation we characterized RNA populations in the pineal gland as compared to the rest of the brain at different stages of zebrafish development. Among genes identified in our DNA microarray studies as differentially expressed in the pineal gland we noted a gene encoding a homolog of the Unc119 protein family. Whereas two Unc119 homologs are known in humans, the gene we isolated is the third family member noted in zebrafish; accordingly we named this protein Unc119c. The unc119c gene is specifically expressed in the pineal, with a low level of expression in the retina. In fish the pineal is a photosensitive organ and shows many similarities in gene expression to the retina. Unc119 proteins interact with small GTPases of the Arl3 family, and we have shown that Unc119c binds to Arl3l2 when both are coexpressed in heterologous cells. Using morpholino antisense oligonucleotide (MO) mediated knock-down of Unc119c expression we found that this protein is required for the formation of the habenular commissure (HC) in the zebrafish. The HC crosses the midline in close proximity to the pineal gland. Knock-down of the Unc119c binding partners Arl3l1 or Arl3l2 also affect HC formation. We hypothesized that Unc119c might be involved in protein trafficking or secretion of a guidance factor involved in HC formation. Based on the literature and expression pattern we picked Wnt4a as a candidate target gene; Wnt4a has been reported by others to be required for HC formation, a fact confirmed in our experiments. We found that Wnt4a accumulation and secretion from cultured HEK293T cells is stimulated by Unc119c and Arl3l1/2. Thus we propose that Unc119c is required in the pineal gland to stimulate Wnt4a secretion, while Wnt4a in turn acts as a guidance cue for the HC as it traverses the zebrafish forebrain. The March family of transmembrane E3 ubiquitin ligases is involved most prominently in the immune system. In fact, the first family member was discovered as a cellular homolog of a viral protein involved in immunosuppression. March 8, the founding member of the family, is known to catalyze the ubiquitination and consequent downregulation of cell surface molecules such as MHC II, B7.2 (CD86), and others. However, essentially nothing was known about possible functions of March 8 in cell types not part of the immune system, or in embryonic development. We found that March 8 is expressed in the early embryo both in zebrafish and the frog Xenopus laevis, and we proceeded to study its possible functions at these stages. Both inhibition of expression and overexpression of March 8 proved deleterious to development. The effects of overexpression were most striking, leading to a loss of cell-cell adhesion leading to cell dissociation in zebrafish embryos and in Xenopus animal explants (so-called animal caps) where the effect could be most clearly visualized. In searching for a molecular basis of this cell dissociation we noted that March 8 could bind to E-cadherin in co-immunoprecipitation experiments, and that an excess of March 8 led to a reduction of cell surface E-cadherin in zebrafish embryos and in human cultured cells. As E-cadherin is the major cell-cell adhesion factor in the early embryo we believe that cell dissociation by March 8 is due to its effect on E-cadherin surface localization. We hypothesize that the biological function of March 8 in the embryo is the modulation of cell-cell adhesion, which must be maintained within a certain range to allow normal development to proceed.
Showing the most recent 10 out of 29 publications
