Investigating the molecular basis of Xotx1, Xotx2 and Xotx5 differential actions during Xenopus laevis development

Otx genes are a class of Vertebrates homeobox genes homologous to the orthodenticle gene of Drosophila melanogaster. In this study we focus on three members of the Otx class in Xenopus laevis: Xotx1, Xotx2 and Xotx5. These three homeoproteins show a high level of homology and exploit both common and differential actions during Xenopus laevis development. During retinal histogenesis, Xotx2 drives progenitor cells to a bipolar fate, while Xotx5 guides retinal precursors toward a photoreceptor fate; analogously, Xotx2 and Xotx5 play a similar role in cement gland induction, while Xotx1 is unable to induce this structure; all three transcription factors seem to be involved in regulating the head organizer activity and convergent extension gastrulation movements. It has been demonstrated that Xotx2 and Xotx5 specific action in frog retina is due to a small amino acid stretch, highly divergent between the two transcription factors and localized downstream of the homeodomain, named retinal specificity box (RS box). Since the specific actions of different transcription factors can be due to their interaction with different cofactors, we have hypothesized that the RS box specific sequences could make XOTX2 and XOTX5 able to interact with different cofactors, thereby leading to the activation of different specific downstream differentiation pathways. To investigate this, we performed two parallel two-hybrid screens, to search for XOTX2 and XOTX5 specific interactors, in order to clarify their divergent action during Xenopus retinogenesis. Several candidate interactors of the two homeoproteins have thus been isolated, but all these potential cofactors were found able to interact in vitro with both XOTX, and also with XOTX1. However, since XOTX proteins exploit also common actions during Xenopus development, the existence of common XOTX interactors is also feasible; besides, a protein that is able to interact in vitro with several partners, may interact in vivo with only one or few of them simply because it colocalize with them, but not with the others. Thus, we decided to go further with our investigation about identified XOTX hypothetical partners. We performed an extensive in silico-analysis, to find out any homologies with described sequences and we thus selected some of the clones for further analysis: Xusf1, Xusf2, Xgrn1, Xgrn2 and a hypothetical peptide named c29. Furthermore, we mapped the specific domain(s) involved in the interaction with each selected cofactor to XOTX N-terminus. An almost partial co-localization of hypothetical partners and Xotx has been found by comparing their expression profiles. After deeply analyzing the data base search results and the expression profiles, we decided to focus our attention on two XOTX hypothetic interactors: USF2, a described transcription factor of bHLH type and C29, a hypothetical so far uncharacterized peptide. We decided to better characterize their molecular interaction with XOTX transcription factors in vitro by GST-pull down assays, as well as their in vivo possible function by performing gain- and loss-of-function experiments. We have predicted in silico the secondary structure of C29 and its subcellular localization; we have demonstrated C29 capability to localize into the nucleus, and we have obtained preliminary data about C29 potential role in vivo. Besides, we here describe a possible antagonistic action of XOTX2/XOTX5 and USF2 both in vitro and in vivo. Moreover, it is known that Xotx2 and Xotx5 induce cement gland in Xenopus laevis ectoderm, while Xotx1 does not. Different transcription factors can exert differential actions also on the basis of sequence divergence. Sequence analysis shows the presence of histidine rich and serine rich regions in XOTX1, that are absent in both XOTX2 and XOTX5. We have investigated the molecular basis of XOTX2/5 and XOTX1 differential action in cement gland formation, and we have demonstrated that it is due to the presence/absence of those XOTX1 specific regions. Besides, we have characterized XOTX molecular domains involved in cement gland promoting action, and we have gained some preliminary data concerning XOTX domain(s) involved in neural tissue induction and in regulation of gastrulation movements.