It was hypothesized that a cascade of transcriptional regulators may control dorsoventral development of the Drosophila embryo. In presumptive mesoderm, one target gene of the dorsal morphogen was shown to be the twist gene. Previous genetic studies and the expression patterns of NK-4 and NK-3 homeobox genes, together with those of twist during embryogenesis, suggested that these genes may regulate each other. In order to understand the mechanism underlying mesodermal cell differentiation, we investigated the transcriptional control of these genes. Because our nucleotide sequence analysis of the 5' upstream region of the NK-4 promoter revealed two clusters (E1, 3 copies of ACATATG; E2, 3 copies of CACTTGA) of E-box sequences (CANNTG) which are putative binding sites for HLH motif transcription factors such as MyoD and twist, and because nuclear extracts from Drosophila embryos showed binding activities to these E-box clusters, we have tested whether twist is a direct activator of the NK-4 gene. When we measured CAT activities in extracts from CV1 cells cotransfected with CAT reporter plasmids containing various 5' upstream regions of the NK-4 promoter and the twist expression vector, we observed a 5-fold increase in CAT activities from extracts transfected with F1 (150 bp, from -950 to -801) constructs. Further mutation analysis within this region and DNaseI footprinting assays with a purified GST:Twist fusion protein strongly suggested that the E1 cluster is responsible for the direct activation of the NK-4 gene by twist. We have also demonstrated that E-box sequences within the E2 cluster were responsible for autoregulation by NK-4 itself. From the cotransfection assay with the NK-4 expression vector and reporters driven by the NK-3 promoter, we have determined that NK-4 is an activator of the NK-3 gene. Taken together, these results demonstrate that NK-4 is a direct target of twist which is a mesoderm determinant in Drosophila embryo, and that the NK-4 protein, in turn, autoactivates the NK-4 gene itself and, finally, upregulates the NK-3 gene, suggesting that a cascade of these gene regulators may play an important role in mesodermal cell specification during embryogenesis.
