Cloning and early dorsal axial expression of Flik, a chick follistatin-related gene: evidence for involvement in dorsalization/neural induction.

We have cloned and sequenced a chick gene, Flik (follistatin-like) that appears to be the homolog of the mammalian TSC36. The ORF encodes a secreted protein of approx 38 kDa, containing a single cysteine-rich domain that shows a strong relationship with the second of the four from which Follistatin is constructed. The remainder of the Flik protein shows no strong family affinities. We describe here the normal expression pattern of the gene during primitive streak and neurula stages. We also give revised data for early neurectodermal expression of chick follistatin (Connolly et al., 1995, Developmental Genetics 17(1), 65-77) and follow the new ectopic expression of both genes during the induction of a second neural axis, after grafting of Hensen's node into a peripheral position in a host blastoderm. Both genes mark the organizer (node and/or mesodermal head process) and early neural plate, and could thus be involved in intercellular signaling during mesodermal dorsalization and neural induction. Flik expression appears earlier than that of follistatin however, and unlike that of follistatin, it is maintained strongly in the dorsal midline with intensity smoothly declining into presumptive lateral regions. We show that both genes are upregulated in host tissue in the neighborhood of node grafts, but whereas follistatin is transcribed after 8-10 hr in host epiblast that has formed new neural plate, Flik is expressed within 4 hr in this region, sometimes detectable before the first structural changes (columnarization) of neuralization. Thus, although ectodermal Flik expression is later confined within neural plate, and mesodermal expression concentrated in dorsal axial tissue, its early distribution is consistent with the idea that the encoded protein may first be involved in generating a graded system that positions the boundaries of both neural and dorsal axial mesodermal territories. The results are discussed in relation to this hypothesis and to other recent findings regarding control of vertebrate dorsoventral patterning.