Analysis of the regulatory gene network underlying endomesodermal specification in S. purpuratus embryos:
At present about 50 genes have been linked into this network. The architecture of the network is emerging from an interdisciplinary approach in which computational analysis is applied to data from gene expression knockouts, genomic sequence, and gene discovery projects, combined with experimental embryology. Regulatory and downstream genes required for skeletogenesis and for territorial specification have been isolated utilizing high-density arrayed cDNA libraries. A predictive model which indicates expected inputs and outputs of key cis-regulatory elements has been built and is being verified as the relevant cis-regulatory systems are found. Many direct experimental tests that can be carried out by altering gene expression in given embryonic cells can now be deduced from the architecture of the model. Many individual projects reported below are contributing to understanding of this network.

Analysis of cis-regulatory systems of territorially expressed genes:
We are using in vitro mutagenesis and reconstruction of regulatory systems with synthetic DNA fragments in order to unravel the cis-regulatory "information processing system" by which genes determine where they are in the embryo, and whether to be expressed. Computational models are essential for interpretation and as guides to experimental analysis of these systems. Among genes expressed in the definitive territories and cell types of the embryo that are being studied in this way are: the endo16 gene; the wnt8 gene; the brachyury gene; hbox12, a gene encoding a homeodomain regulator of the very early skeletogenic lineage; krox1, a gene encoding a Zn finger transcription factor; foxa, a gene encoding a sea urchin HMF3b winged helix-type factor; gatae, an endodermal regulatory gene; delta, which encodes a Notch ligand; and otx, a transcriptional regulator which acts in many domains of the embryo, but which also plays a key role in endomesoderm specification.

Isolation, cloning, and functional analysis of transcription factors:
We have developed technologies for purification of transcription factors from embryo nuclear extract, followed by microsequencing, and hence cloning. Transcription factors that interact with target sites of known function can thus be characterized, and their embryonic provenance determined. Current research targets include several different transcription factors and a repressor that serves as a terminus of a signal pathway, all identified in studies on the endo16 gene; and a transcriptional regulator of the cyIIIa aboral ectoderm-specific actin gene

Sea urchin homeobox gene cluster:
We have isolated the sea urchin HOM-C genes and shown that they reside in a single genomic cluster. Their genomic organization and pattern of expression during development are of particular comparative interest, since echinoderms (together with hemichordates) are the surviving sister group of all other major deuterostome members of the Animal Kingdom, i.e., chordates. We showed that whole cluster of hox genes is expressed in a colinear fashion in the rudiment of the adult body plan, and only two of the ten genes are used in the embryo at all. A current objective is to discover the downstream targets of the hox11/13b gene, which is expressed in the embryonic endoderm.

Regulatory evolution:
We are engaged in comparative assessments of cis-regulatory evolution in echinoderm species, comparing some key regulatory linkages in the endomesoderm gene network of a starfish to those discovered in the sea urchin. Starfish and sea urchins have evolved separately almost since the Cambrian. The endomesodermal gene network provides the opportunity of determining precisely which regulatory linkages have changed during evolution, and what are the consequences for the developmental process

Computational approaches to regulatory gene network analysis:
Regulatory gene networks for development cannot be informatively treated as equilibrium or steady-state systems. With Hamid Bolouri of the Institute for Systems Biology, Seattle, WA, a new approach to mathematical description of these developmental regulatory systems is being developed, with which to describe its unidirectional progression through successive spatial regulatory states. In addition, several genome analysis tools are being development and a new project to build probability models of target sites for all known transcription factors in the endomesoderm network is under way.

Sea urchin genome project
The Sea Urchin Genome Project web site (http://sugp.caltech.edu/) is the distribution center for sequence and annotation information related to the sea urchin genome and the macroarray libraries. We have maintained and extended this part of the facility that is housed in the Center over the last year in preparation for the incoming whole genome sequence.
As genomic BAC sequence from the S. purpuratus, L. variegatus and other echinoderm genomes accumulates, it is annotated using the software package SUGAR, and is now being rewritten to be useful with the Cartwheel queueing system. These programs also reside on the SUGP machines and efficiently interdigitate with the existing packages.
The most unique, information-rich component of the Sea Urchin Genome Project web site is the database of macroarray filter information. Here all of the sequence mentioned above and the gene annotation information collected in the process of screening these library filters, for whatever purpose, are stored. This includes sequence collections from complex probe screens such as those used for the identification of genes in the endomesoderm specification pathway, the results of homology screening strategies, and random EST projects. Because the data is coupled to a filter location that contains an individual clone from the library, the clone is immediately recoverable. As more clones are characterized in a library, that library becomes more valuable. Eventually, the several well-characterized libraries can be used to confirm ab initio gene predictions and confirm gene catalogs for the sea urchin.
Personnel: The staff who support the computational aspects of the Center are: Kevin Berney, C. Titus Brown, Ramon Cendejas and Ian Lipsky.