Embryonic expression of the Endo16 gene of Strongylocentrotus purpuratus is controlled by interactions with at least 13 different DNA-binding factors. These interactions occur within a cis-regulatory domain that extends about 2300 bp upstream from the transcription start site. A recent functional characterization of this domain reveals six different subregions, or cis-regulatory modules, each of which displays a specific regulatory subfunction when linked with the basal promoter and in some cases various other modules (C.-H. Yuh and E. Davidson (1996) Development 122, 1069-1082). In the present work, we analyzed quantitative time-course measurements of the CAT enzyme output of embryos bearing expression constructs controlled by various Endo16 regulatory modules, either singly or in combination. Three of these modules function positively in that, in isolation, each is capable of promoting expression in vegetal plate and adjacent cell lineages, though with different temporal profiles of activity. Models for the mode of interaction of the three positive modules with one another were tested by assuming mathematical relations that would generate, from the measured single module time courses, the experimentally observed profiles of activity obtained when the relevant modules are physically linked in the same construct. The generated and observed time functions were compared, and the differences were minimized by least squares adjustment of a scale parameter. When the modules were tested in context of the endogenous promoter region, one of the positive modules (A) was found to increase the output of the others (B and G), by a constant factor. In contrast, a solution in which the time-course data of modules A and B are multiplied by one another was required for the interrelations of the positive modules when a minimal SV40 promoter was used. One interpretation is that, in this construct, each module independently stimulates the basal transcription complex. We used a similar approach to analyze the repressive activity of the three Erzdo16 cis-regulatory modules that act negatively in controlling spatial expression. The evidence obtained confirms that the repressive modules act only by affecting the output of module A (C.-H. Yuh and E. Davidson (1996) Development 122, 1069-1082). A new hierarchical model of the cisregulatory system was formulated in which module A plays a central integrating role, and which also implies specific functions for certain DNA-binding sites within the basal promoter fragment of the gene. Additional kinetic experiments were then carried out, and key aspects of the model were confirmed.
