Feedforward regulatory logic controls the specification-to-differentiation transition and terminal cell fate during Caenorhabditis elegans endoderm development

ABSTRACT The architecture of gene regulatory networks determines the specificity and fidelity of developmental outcomes. We report that the core regulatory circuitry for endoderm development in Caenorhabditis elegans operates through a transcriptional cascade consisting of six sequentially expressed GATA-type factors that act in a recursive series of interlocked feedforward modules. This structure results in sequential redundancy, in which removal of a single factor or multiple alternate factors in the cascade leads to a mild or no effect on gut development, whereas elimination of any two sequential factors invariably causes a strong phenotype. The phenotypic strength is successfully predicted with a computational model based on the timing and levels of transcriptional states. We found that one factor in the middle of the cascade, END-1, which straddles the distinct events of specification and differentiation, functions in both processes. Finally, we reveal roles for key GATA factors in establishing spatial regulatory state domains by repressing other fates, thereby defining boundaries in the digestive tract. Our findings provide a paradigm that could account for the genetic redundancy observed in many developmental regulatory systems.

provided SKN-1 activates med-1 and -2, which have both a maternal and zygotic component.In MS, POP-1 represses the expression of end-1 and -3 genes.MED-1 and -2 then directly activate the expression of mesoderm-specifying factor, TBX-35.In E, however, Wnt, MAPK and Src signaling from neighboring P2 cell ultimately leads to the phosphorylation of POP-1 (indicated by *), altering its nucleocytoplasmic localization (high nuclear level in MS and low nuclear level in E; bottom panel) and converting it from a repressor to an activator of gut fate (Shetty et al., 2005).
Hence, MED-1/2 and POP-1* provide redundant inputs to turn on end-1/3.END-1 and -3 subsequently activate the expression of ELT-7 and -2, both of which promote gut morphological differentiation.Removing med-1 or med-2 alone does not affect end-3 expression at the 4E embryonic stage.Table S3.Worm strains used in this study.* indicates strain that contains endogenously tagged reporter generated by CRISPR/Cas9.
Click here to download Table S3 Table S4.Model parameters and outputs of the endoderm GRN.Expression of each factor is determined by the concentration of its activators multiplied by a coefficient a, representing the strength of the inputs.SKN-1 expression follows a square wave in EMS blastomere.POP-1 expression follows a square wave in the E cell.Feedback coefficients f become nonzero once their respective factors surpass a certain threshold .
Click here to download

Fig. S1 .
Fig. S1. C. elegans mesendoderm development.(A) At the four-cell stage, EMS (blue) undergoes asymmetrical division to produce anterior MS (green) and posterior E (orange) blastomeres.Newly hatched L1s contain 20 intestinal cells arranged in nine rings (Ints).Int1 contains four cells, while the remaining eight rings contain two cells each.The MS cell gives rise mesodermal cell types, including the posterior pharynx.(B) Simplified mesendoderm GRN showing three sequential tiers of paired redundant GATA transcription factors (2 redundant MEDs  2 redundant ENDs  2 redundant ELTs).Maternally

Fig. S4 .
Fig. S4.Mutants lacking alternate GATA pairs do not show apparent gut defects.(A) Mutants lacking alternate GATA pairs contain fully differentiated lumen (top row) and gut granules (bottom row) along the length of the animals.The same set of double mutants show wildtype expression of (B, C) immunoreactive IFB-2 and (D, E) AJM-1.

Fig. S5 .
Fig. S5.The expression of endogenously tagged END-3 reporter in the med mutants.