Transcription factors of the Forkhead (Fox) family have been identified in many metazoans, and play important roles in diverse biological processes. Here we define the set of fox genes present in the sea urchin genome, and survey their usage during development. This genome includes 22 fox genes, only three of which were previously known. Of the 23 fox gene subclasses identified in vertebrate genomes, the Strongylocentrotus purpuratus genome has orthologues of all but four (E, H, R and S). Phylogenetic analysis suggests that one S. purpuratus fox gene is equally related to foxA and foxB of vertebrates; this gene defines a new class. Two other genes appear to be specific to the sea urchin, with respect to the genomes so far sequenced. Fox genes orthologous with those of vertebrates but lacking in arthropod or nematode genomes may be deuterostome-specific (subclasses I, J1, J2, L1, M and Q1), while the majority are pan-bilaterian. All but one of the S. purpuratus fox genes (SpfoxQ1) are expressed during embryogenesis, most in a very specific temporal and spatial manner. The sea urchin fox genes clearly execute many different regulatory functions, and almost all of them participate in the process of embryonic development.

The genomic cis-regulatory systems controlling regulatory gene expression usually include multiple modules. The regulatory output of such systems at any given time depends on which module is directing the function of the basal transcription apparatus, and ultimately on the transcription factor inputs into that module. Here we examine regulation of the Strongylocentrotus purpuratus tbrain gene, a required activator of the skeletogenic specification state in the lineage descendant from the embryo micromeres. Alternate cis-regulatory modules were found to convey skeletogenic expression in reporter constructs. To determine their relative developmental functions in context, we made use of recombineered BAC constructs containing a GFP reporter and of derivatives from which specific modules had been deleted. The outputs of the various constructs were observed spatially by GFP fluorescence and quantitatively over time by QPCR. In the context of the complete genomic locus, early skeletogenic expression is controlled by an intron enhancer plus a proximal region containing a HesC site as predicted from network analysis. From ingression onward, however, a dedicated distal module utilizing positive Ets1/2 inputs contributes to definitive expression in the skeletogenic mesenchyme. This module also mediates a newly discovered negative Erg input which excludes non-skeletogenic mesodermal expression.