Mechanisms underlying cell-cell communication and cell function during development
                       
Research in the lab addresses the mechanisms underlying interactions between cells that give rise to the formation of elaborate patterns and organs during development. While most of the pathways and genes that participate in patterning have been identified through genetic screens, several central issues in development remain open. We employ a wide spectrum of genetic, molecular and computational approaches, as we focus on two of these issues:

• One is the interphase between cell and developmental biology. In other words, how do the developmental signaling pathways operate in the context of the cell, its organelles and the extracellular milieu, and what general features of the cell impinge on developmental signaling. Another aspect of the same problem concerns the mechanisms by which basic processes of the cellular machinery, e.g. actin polymerization, are recruited to drive specific morphogenetic events.
• A second open direction concerns the way in which quantitative fluctuations in cell communication during development are buffered, such that the final outcome is fixed and robust to these fluctuations.

Both issues go beyond the classical genetic analysis, traditionally studied by gene loss-of function. In the case of the interphase with cell biology, the challenge is to dissect distinct developmental roles of general cellular components. To study robustness, the systems should be manipulated by altering the levels of their components, but not eliminating them. The ways in which both frontiers are addressed will be presented by describing three major research topics in the lab.

Actin-nucleating factors in morphogenesis

Actin-nucleating factors, including WASp, SCAR/WAVE and Dia/Formin, translate signal-transduction cues into cytoskeletal remodeling, to control a wide range of cellular functions, such as cell motility and endocytosis. In view of these pleiotropic roles, it is challenging to identify distinct morphogenetic functions for these pathways. We have used the ability to inactivate these pathways in defined tissues, as an entry point to address this issue.

1. Control of muscle fusion by branched actin nucleation

Individual myoblasts (muscle cells) fuse to form multi-nucleated muscle fibers. In the Drosophila embryo, muscles are formed by the specification of distinct muscle founder cells in each segment, defining the thirty-odd muscle types that will form. In order to gain muscle mass, each founder fuses to a defined number of fusion-competent myoblast cells. 
Our entry point to the process stemmed from the analysis of WIP/Verprolin, which serves as an anchoring protein for the WASp actin nucleator. WIP is expressed in embryos only in the muscles, before they fuse, and mutants in WIP are defective in muscle fusion. The definitive phenotype of WIP mutants triggered subsequent experiments that demonstrated the role of WASp in the same process. In the absence of WIP or WASp, myotubes associate with myoblasts, small fusion pores are formed, but these fail to grow in order to complete fusion. We are exploring the mechanism by which regulated actin polymerization drives the enlargement of the fusion pores.
Interestingly, analysis of mutants and inactivating constructs for members of a second actin-nucleating complex, SCAR/WAVE, also demonstrated a requirement in muscle cell fusion. However, in the absence of this pathway, fusion was arrested prior to the formation of the fusion pores. Thus, the two actin-nucleating factors appear to carry out distinct and consecutive roles during muscle cell fusion.

Given the evolutionary conservation of the WASp and SCAR/WAVE actin nucleators, it was tempting to speculate that these elements are involved in mammalian muscle fusion, a process whose molecular basis remains mostly unknown. We are currently addressing this issue in detail, by studying muscle development in  mice mutant for N-WASp, the major mammalian WASp-family element. Analysis both in vivo and in primary satellite-cell cultures, shows that disruption of N-WASp function does not interfere with the program of skeletal myogenic differentiation, and does not affect myoblast motility, morphogenesis and attachment capacity. N-WASp-deficient myoblasts, however, completely fail to fuse. Furthermore, our analysis suggests that myoblast fusion requires N-WASp activity in both partners of a fusing myoblast pair. These findings reveal a specific role for N-WASp during mammalian myogenesis. WASp-family elements appear therefore to act as universal mediators of the myogenic cell-cell fusion mechanism underlying formation of functional muscle fibers, in both vertebrate and invertebrate species.
New projects in the lab are examining later roles of WASp and Diaphoanous/Formin proteins in muscles of flies and mice, in forming and maintaining the mature structure of the sarcomeres, the intricate machinery that mediates muscle contraction.

Apical localization of filamentous actin (F-actin) is a prominent and common feature of epithelial tubes in multicellular organisms. However, the mechanism for nucleating polymerization of this actin layer and its roles are not known. We have demonstrated that the Diaphanous (Dia)/Formin actin-nucleating factor is required for generation of apical F-actin, in diverse types of epithelial tubes in the Drosophila embryo. Apical-basal polarity and microtubule organization of tubular epithelial cells remain intact in the absence of Dia function. The apical nucleation of actin filaments by Dia, which itself is apically localized both at the RNA and protein levels, is based on the apical localization of its activators, including Rho1 and two guanine-exchange factor proteins (Rho-GEFs). In the absence of apical actin polymerization, secretion through the apical surface to the lumen of tubular organs is blocked. We have further shown that apical secretion requires the Myosin V motor, implying that secretory vesicles are targeted to the apical membrane by Myosin V-based transport, along polarized actin filaments nucleated by Dia. This mechanism allows efficient utilization of the entire apical membrane for secretion, and appears universal to tubes of different function and diameter, such as the trachea, salivary gland, hindgut and Malpighian tubules.
Current work focuses on the mechanisms of apical localization of Dia in flies, and is also extending the studies to the role of Rho/Dia in apical secretion in tubular organs of mice, such as the acinar cells of the pancreas.


Robust developmental patterning

Development of multicellular organisms is dictated by cell-cell interactions mediated by conserved signaling pathways. A key challenge encountered by these pathways is to generate reproducible patterns, despite fluctuations in the doses of the signaling components, which commonly occur due to genetic and environmental alterations. To identify and characterize the mechanisms underlying such developmental robustness, we are combining experimental and theoretical computational approaches, in collaboration with the lab of Naama Barkai. We employ a systems-level approach to a restricted signaling module, where we assume that all the components within the module are known. The challenge is to obtain a quantitative understanding of the design principles of the module, in a situation where most of the quantitative in vivo parameters (e.g. concentration, affinity, diffusion rates) are not known.

We find that robustness of a pathway to fluctuations in the levels of its components provides a major restriction, and significantly limits the number of potential mechanisms that generate pattern. Thus, we initially check the pathway for its robustness to heterozygosity for mutations in the different components. If the pathway is robust, we describe by equations the possible interactions between components, and are liberal in giving values that span several orders of magnitude. For each of the cases the outcome is solved, and subsequently examined under conditions where component levels are halved. Only a very small fraction of solutions shows robustness. The restricted values of these solutions guide us towards possible cellular and molecular mechanisms of robustness, and allows to design experiments that will indeed examine these predictions.

1. BMP signaling in the early embryo
         
This approach was utilized to address the patterning of early embryos by BMP ligands. In this system, ligand levels are uniform in the dorsal part of the early Drosophila embryo, yet signaling is graded, and peaks at the dorsal midline. The key to this pattern is a secreted inhibitor of BMP, termed Sog, which is expressed in a lateral pattern flanking the region expressing the ligand. By using the computational screen and experimental verification, we demonstrated an active ligand shuttling mechanism, where the inhibitor traps the ligand, translocates it, and physically releases it in the dorsal-most region. This mechanism provides for robust patterning. We have recently shown that an elaboration on this conserved mechanism in a vertebrate organism, the Xenopus frog, allows not only to obtain robust dorso-ventral patterning, but also to scale patterning with modulations in embryo size. This mechanism may explain some of the cardinal outcomes of the classical Spemann experiments on vertebrate embryos.

2. Spätzle/Toll signaling in the early embryo

Animation of Self-organized shuttling

3. Scaling the BMP gradient

A central and largely open issue in developmental biology is how pattern is scaled with size. In other words, upon variation in the size of the embryo or developing organ, how is the morphogen gradient adjusted, to maintain proportional patterning. Using theoretical modeling, we examine how the activation gradient of the Dpp morphogen in the Drosophila wing imaginal disc scales with disc size. We predict that scaling is achieved through an expansion-repression mechanism, whose mediator is the widely diffusible protein Pentagone (Pent). Central to this mechanism is the repression of pent expression by Dpp signaling, which provides an effective size measurement, and the Pent-dependent expansion of the Dpp gradient, which adjusts the gradient with tissue size. We validate this mechanism experimentally by demonstrating that scaling requires Pent and further, that scaling is abolished when pent is ubiquitously expressed. The expansion-repression circuit can be readily implemented by a variety of molecular interactions, suggesting its general utilization for scaling morphogen gradients during development.
Current work in the lab examines scaling and the underlying mechanisms in dorsoventral patterning of the early embryo, both within the D. melanogaster species, and between different Drosophila species, where proportional patterning is maintained despite large differences in egg size.


Acknowledgements

BS is an incumbant of the Hilda and Cecil Lewis chair in Molecular Genetics.