Contribution of homeobox genes to the development of the oikoplastic epithelium, a major novelty of tunicate Larvaceans

The development of complex organisms is controlled by numerous genes, including developmentally regulated transcription factors (TF) encoded by multiple gene families. A well diversified group of TFs is the superclass of homeodomain proteins, among which some have a well studied and relatively well conserved function during the development of metazoans. The most popular homeodomain TFs are the Hox genes, core players in the AP axis patterning, with a peculiar genome organization in clusters which ensures the coordination of their spatio-temporal expression. However, the function of Hox and many other conserved homeobox genes has independently evolved in distinct metazoan lineages, contributing to changes of their body plan. This divergent evolution also led to their participation in lineage specific innovations. Such radical changes are governed by conserved genes whose expression patterns were importantly modified, as well as by new genes appearing in specific lineages. Gene duplication is one mechanism leading to genes with divergent functions. When changes in coding sequences are scored among distinct genomes, it clearly appears that some metazoan lineages have evolved more rapidly than others. This is the case for tunicates, compared to other chordates. Within tunicates, larvaceans (=appendicularians) also evolved more rapidly than others classes, as judged from genome data of Oikopleura dioica, a cosmopolitan species that has become easily bred in the laboratory. The rapid genome evolution in this group is also visible through profound changes of the complement of genes (loss and duplication of genes), their intron-exon organization or their order on the chromosomes. A very speaking example is the Hox cluster disintegration in this lineage. We contributed to show that in O. dioica, an important fraction of the homeobox genes has been lost, or at least fails detection by classical gene alignment tools. We also noticed that in the homeobox genes that remain present, a relatively large proportion have been duplicated once or several times. We then revealed, using in situ hybridization (ISH) in embryos and young larvae that the members of most amplified homeobox gene groups are expressed in the epithelium of the trunk. This happens very rarely for homeobox genes that have no duplicate in the genome. Since the expression in the trunk epithelium occurs when it is gradually transformed in the oikoplastic epithelium that later on generates the house (a large and complex filter-feeding apparatus typical of larvaceans), we postulated that duplicated genes play a role in the formation of this novel structure. If this is true, the question of whether and how gene duplication events have been involved in the evolution of the oikoplastic epithelium and the house also needed to be addressed. When we began our work, the function of O. dioica genes could be apprehended only with ISH and quantification of gene expression. We therefore devoted major efforts to experimentally modify their expression levels in adapting from other species a technique of RNA interference (RNAi) based on the introduction of double stranded RNA matching the gene sequence of interest. We contributed to validate this technique in knocking down the expression of three genes, chosen because of their easily predictable loss of function phenotype. These genes are Brachyury, which encodes a T-Box transcription factor and has a major role in the formation of the notochord, and two genes encoding essential enzymes of the cholinergic (ChAT=choline acetyltransferase) and GABA-ergic (GAD= glutamic acid decarboxylase) pathways, respectively. After injection of dsRNA for each of these three genes in defined conditions, we indeed observed an aborted tail development for Brachyury, a severe inhibition of tail locomotion for ChAT, and an uncoordinated swimming for GAD. Before applying this technique to homeobox genes expressed in the trunk epithelium, we morphologically described the epithelium development during the hours after hatching. We could observe the formation from a rather uniform cell layer of individual cellular fields that characterize the oikoplastic epithelium and we monitored the concomitant expression initiation of several house protein genes (oikosins) during this process. A variable number of cell divisions occur in each forming field, and we observe no migration of cells between distinct fields. Homeobox genes revealed in the initial screen were intensely and transiently expressed during the formation of the fields and we tentatively mapped their expression domain at an early and a late stage of the process. As mentioned above, most of these genes are duplicated and it is noteworthy that duplicates of a given gene are expressed in similar regions, usually together or near each other. Consequently, the epithelial expression very likely predated the duplication events. If these homeobox genes are involved in the transition towards an oikoplastic epithelium, their duplication was probably not essential for its emergence during evolution. Duplications may have contributed to increase the complexity of this structure through neofunctionalization, and/or resulted in a division of the original gene function among duplicates. To prove the involvement of homeobox genes in the formation of the oikoplastic epithelium we attempted RNA interference for genes that are expressed in epithelium at the early larval stage: three otx genes and two prop genes. The epithelial expression of prop genes was efficiently knocked down. We observed the malformation and disorganization of nuclei of epithelial cells around the dorsal midline, where both prop genes are normally expressed. The knockdown of prop genes also abolished the post-metamorphosis expression of one oikosin gene (oik41a) in the Anterior rosette field after metamorphosis, confirming their importance for cell differentiation in this region. In contrast, the attempts of RNAi for the three otx genes were not conclusive, partly because they provoked severe developmental abnormalities during embryogenesis. Overall, our expression study supports that multiple homeodomain transcription factors have been recruited in genetic pathways governing the development of the oikoplastic epithelium, which synthesizes and secretes diverse components of the larvacean house. We bring evidence for the involvement of at least two homeobox genes (prop duplicates) in this process, upstream of the ultimate cell differentiation in a specific region. Other transcription factors including Fox genes are also expressed in the trunk epithelium during early larval stages. Although most homeobox genes transiently expressed in the epithelium of the larva trunk are duplicated, there is no indication that duplications proper were at the origin of this evolutionary innovation