Defects in cardiac valve morphogenesis and septation of the heart chambers constitute some of the most common human congenital abnormalities. Some of these defects originate from errors in atrioventricular (AV) endocardial cushion development. Although this process is being extensively studied in mouse and chick, the zebrafish system presents several advantages over these models, including the ability to carry out forward genetic screens and study vertebrate gene function at the single cell level. In this paper, we analyze the cellular and subcellular architecture of the zebrafish heart during stages of AV cushion and valve development and gain an unprecedented level of resolution into this process. We find that endocardial cells in the AV canal differentiate morphologically before the onset of epithelial to mesenchymal transformation, thereby defining a previously unappreciated step during AV valve formation. We use a combination of novel transgenic lines and fluorescent immunohistochemistry to analyze further the role of various genetic (Notch and Calcineurin signaling) and epigenetic (heart function)pathways in this process. In addition, from a large-scale forward genetic screen we identified 55 mutants, defining 48 different genes, that exhibit defects in discrete stages of AV cushion development. This collection of mutants provides a unique set of tools to further our understanding of the genetic basis of cell behavior and differentiation during AV valve development.

Endodermal organs such as the lung, liver and pancreas emerge at precise locations along the primitive gut tube. Although several signalling pathways have been implicated in liver formation, so far no single gene has been identified that exclusively regulates liver specification. In zebrafish, the onset of liver specification is marked by the localized endodermal expression of hhex and prox1 at 22 hours post fertilization. Here we used a screen for mutations affecting endodermal organ morphogenesis to identify a unique phenotype: prometheus (prt) mutants exhibit profound, though transient, defects in liver specification. Positional cloning reveals that prt encodes a previously unidentified Wnt2b homologue. prt/wnt2bb is expressed in restricted bilateral domains in the lateral plate mesoderm directly adjacent to the liver-forming endoderm. Mosaic analyses show the requirement for Prt/Wnt2bb in the lateral plate mesoderm, in agreement with the inductive properties of Wnt signalling. Taken together, these data reveal an unexpected positive role for Wnt signalling in liver specification, and indicate a possible common theme for the localized formation of endodermal organs along the gut tube.

The G2 DNA damage checkpoint ensures maintenance of cell viability by delaying progression into mitosis in cells which have suffered genomic damage. It is controlled by a number of proteins which are hypothesized to transduce signals through cell cycle regulators to delay activation of p34cdc2. Studies in mammalian cells have correlated induction of inhibitory tyrosine 15 (Y15) phosphorylation on p34cdc2 with the response to DNA damage. However, genetic studies in fission yeast have suggested that the major Y15 kinase, p107wee1, is not required for the cell cycle delay in response to DNA damage, although it is required for survival after irradiation. Thus, the target of the checkpoint, and hence the mechanism of cell cycle delay, remains unknown. We show here that Y15 phosphorylation is maintained in checkpoint‐arrested fission yeast cells. Further, wee1 is required for cell cycle arrest induced by up‐regulation of an essential component of this checkpoint, chk1. We observed that p107wee1 is hyperphosphorylated in cells delayed by chk1 overexpression or UV irradiation, and that p56chk1 can phosphorylate p107wee1 directly in vitro. These observations suggest that in response to DNA damage p107wee1 is phosphorylated by p56chk1 in vivo, and this results in maintenance of Y15 phosphorylation and hence G2 delay. In the absence of wee1, other Y15 kinases, such as p66mik1, may partially substitute for p107wee1 to induce cell cycle delay, but this wee1‐independent delay is insufficient to maintain full viability. This study establishes a link between a G2 DNA damage checkpoint function and a core cell cycle regulator.