Analysis of the gene regulatory network underlying endomesoderm specification in S. purpuratus embryos: At present about 50 genes have been linked into this GRN.  The architecture of the network is emerging from an interdisciplinary approach in which computational analysis is applied to perturbation data obtained by gene expression knockouts and other methods, combined with experimental embryology.  A predictive model of the GRN has emerged which indicates the inputs and outputs of the cis-regulatory elements at its key nodes.  This model essentially provides the genomic regulatory code for specification of the endomesodermal territories of the embryo, up to gastrula stage

.Testing the cis-regulatory predictions of the GRN:  The GRN was constructed essentially by integrating the results of a massive perturbation analysis of expression of individual genes with spatial and temporal expression data for these genes.  It predicts the required specific regulatory inputs and outputs linking the genes within the GRN.  These predictions are subject to direct experimental cis‑regulatory test, and correction.  We have now authenticated the predicted cis-regulatory inputs into genes at a majority of the key nodes of the current GRN.  At these nodes are regulatory genes into which there are multiple regulatory inputs from genes elsewhere in the GRN, and multiple outputs to other genes in the GRN.  For some regions of the GRN the analysis is approaching maturity, in that it extends convincingly from maternal inputs to cell-type differentiation.  The best example is the GRN subregion determining skeletogenic micromere specification.  Overall, the results of these experiments are converting the GRN from a model proposition into a hard-wired map of the genomic control logic for this portion of development.  At present among cis-regulatory systems that are the subject of experimental analysis are those of the following genes:  alx1, tgif, hex, foxa, brachyury, rel

Completion of the repertoire of regulatory genes engaged in the endomesoderm GRN:  We used the data emerging from the genome sequence project to identify and assemble computationally all gene sequences that encode transcription factors.  The temporal patterns of expression of these genes were determined, and for those genes sufficiently expressed in the embryo, the spatial patterns as well.  Regulatory genes were identified in this manner that evidently play a role in endomesoderm specification, because they are expressed specifically in the endomesodermal territories at the relevant times, but that had not yet been incorporated into the GRN.  All of these genes are now being linked into the GRN by perturbation and cis-regulatory analysis; this project has been completed for the skeletogenic micromere lineage and is in process for the non-skeletogenic mesoderm and the endoderm.

Evolution, viewed as a process of change in GRN architecture:  Starfish and sea urchins shared a last common ancestor about 500 million years ago.  Thus, analysis of the GRN controlling endomesoderm specification events in the starfish embryo will reveal both the nature of functional change in the GRN, and conservation of features that are so essential that they have resisted alteration for half a billion years.  Examples of both have now been documented.  The underlying processes are of course change, or alternatively, conservation, of functional cis-regulatory features.  To study this we are examining starfish/sea urchin GRN differences at the cis-regulatory level.  A second ongoing project is an attempt to reprogram the development of the skeletogenic cell lineage in a primitive sea urchin, Eucidaris tribuloides, by inserting regulatory apparatus from S. purpuratus.  These echinoids diverged from a common ancestor in the Middle Triassic and generate their embryonic skeleton in different ways.  We term this Synthetic Experimental Evolution.

cis-regulatory evolution and interspecific recognition of cis-regulatory modules:  We are carrying out a sequence level evolutionary analysis of experimentally authenticated cis‑regulatory modules from nine different genes in four species of sea urchin.  At a 50 my divergence distance (S. purpuratus and Lytechinus variegatus) cis-regulatory modules and exons are the only conserved sequence elements.  Surprisingly, of authenticated transcription factor target sites, about a third are either novel additional occurrences of given sites in one or the other species, or are sites that have changed position within the conserved module.  However, some proximal site pairs recur repeatedly.  At the closer distance represented by two congeners for which genomic sequence exists (Allocentrotus fragilis and Strongylocentrotus franciscanus), cis-regulatory modules are marked by sharply decreased frequencies of large indels.  These are otherwise a major mechanism of divergence in unselected sequence.  A library of >100 sequence patches conserved between L. variegatus and S. purpuratus is under study for indel frequencies as a function of indel size in all four species.

Oral and aboral ectoderm GRNs:  We have recently attained draft GRNs for oral and aboral ectoderm specification including about 30 more regulatory genes.  This is part of an effort to extend the same kind of causal, system level GRN analysis to the whole embryo, and represents a major step toward that goal.  There is only one additional early embryonic territory, the apical neurogenic region, which is being studied in other sea urchin laboratories.  The aboral ectoderm generates a single cell type, but the oral ectoderm gives rise to several distinctly functioning domains:  mouth, columnar "facial" epithelium, neurogenic ciliary band, and the ectodermal signaling stripes which determine the location of the skeletal rods.  The approach is to obtain all the regulatory players expressed in oral and aboral ectoderm from the analysis of all genes encoding transcription factors predicted in the genomic sequence, and engage them in a provisional network by carrying out a matrix of perturbation experiments.  The network is anchored at the onset of the ectodermal specification process, of which the initial gene zygotically expressed on the oral side is nodal.  The cis-regulatory module controlling early oral ectoderm expression of nodal has been thoroughly characterized:  its target sites provide the direct links between the initial cytoplasmic anisotropy by which the future oral and aboral sides of the embryo are distinguished, i.e., differences in redox potential, and differential zygotic gene expression. 

Various explorations by new methods and approaches:  As always, we are trying to expand knowledge by use of novel technologies for analysis of the GRN and the genome.  Current applications of new technology include increasingly widespread use of in vitro reengineered BAC recombinants, which we are supplying to the whole sea urchin field; use of these in first attempts to "redesign" the process of embryonic development, by introduction of altered regulatory subcircuits in novel spatial domains; and extensive application of two-color gene transfer experiments in which the control version of a cis-regulatory construct drives expression of a reporter detectable in one color and a mutated version injected with it and incorporated in the same cells drives expression of a reporter detected in a different color.  We have also developed a completely novel method for blocking expression of any gene whenever and wherever desired, though this has so far been tested only in sea urchin embryo skeletogenic cells.  This method allows us to determine the function of regulatory genes that have multiple activity phases in one of the later phases, in embryos that develop normally up to that point.  Another new technology that has had a major impact is utilization of NanoString technology, a new instrument that permits direct automated counting of transcripts of one-several hundred genes in control and experimentally perturbed lysates of small numbers of embryos with high accuracy efficiency. 

Revolutionizing cis‑regulatory analysis:  We are seeking to improve the efficiency of cis-regulatory analysis by large factors.  First, we have shown that high coverage Solexa reads of Lytechinus BACs containing given genes can be used to identify conserved cis-regulatory modules by mapping them on the S. purpuratus genome sequence, using an elegant display apparatus, "Cis-Browser," developed by our colleague Sorin Istrail (Brown University).  Second, we have found that FACS sorted, disaggregated embryos expressing regulatory constructs producing GFP provide populations of expressing cells that can be analyzed by QPCR to determine spatial and temporal domain of cis-regulatory expression.  This obviates time consuming microscopic screening and can be done on, many samples at once.  Third, we have developed multiple tags for use as reporters so that many samples can be injected into embryos at once and analyzed together by NanoString or QPCR

Computational approaches to regulatory gene network analysis:  The GRN visualization software BioTapestry, developed by our collaborators Hamid Bolouri and Wm. Longabaugh at ISB, is now in wide use, and we are further expanding its capacities so that it will automatically generate allowed network architectures from machine readable time and space of expression data plus results of perturbation analysis.  A second-generation version with much enhanced capacities has been published.