Mergers and Acquisitions Unequal Partnerships in Drosophila Myoblast Fusion

not known. Interestingly, although the founder specification programs act upon the fusion process, the converse does not seem to be true. If the ability of myoblasts to fuse is blocked (for example by mutations in dock180/mbc), then the founder cells still make tiny Manfred Frasch*‡ and Maria Leptin†‡ *Department of Biochemistry and Molecular Biology Mount Sinai School of Medicine One Gustave L. Levy Place muscles with the correct specifications (Rushton et al., New York, New York 10029 1995; Erickson et al., 1997). However, as predicted by †Institute of Genetics the founder cell hypothesis, the unfused fusion-compeUniversity of Cologne tent cells (now incompetent) remain undifferentiated. Weyertal 121 An important feature of the fusion process is its inherD-50931 Cologne ent asymmetry. Thus, founder cells can only fuse with Germany fusion-competent cells and not with each other, even when they arise as sister cells from a common progenitor and abut each other. Similarly, fusion-competent cells Like muscle fibers in vertebrates, the body wall muscles have never been observed to fuse with each other. Since both types of cells actively participate in fusion, it has in the Drosophila embryo are generated by cell fusion been a mystery how this asymmetry of the fusion profrom groups of undifferentiated myoblasts (reviewed in cess is achieved. What is it that makes the founder cell Baylies et al., 1998; Frasch, 1999). Several ways can attract fusion-competent cells, and what determines the be imagined by which groups of cells within a field of directionality of fusion? myoblasts might be assigned to different muscles. For An asymmetric distribution of cell–cell interaction molexample, all the cells that generate a particular muscle ecules might be expected to be involved, and the most might be specified and assigned to fuse to each other. recent papers in fact describe just that (Bour et al., However, this is not how it works in Drosophila, nor 2000; Ruiz-Gómez et al., 2000). The new findings and probably in other species. Instead, the myoblasts are the remaining unanswered questions will make for an divided into two types of cells. One type, the muscle exciting new phase in the investigation of this particular founder cells, have the ability to recruit a number of problem as well as myoblast fusion in general. Both cells of the second type, the fusion-competent cells, to molecules are novel members of the immunoglobulin fuse with them and trigger differentiation into a muscle. superfamily and appear to act during the earliest steps This minireview focuses on two new papers that deof myoblast fusion. The first, named Dumbfounded (Duf; scribe the identification of cell type–specific cell surface Ruiz-Gómez et al., 2000), is expressed in the founder but molecules with essential roles in this myoblast fusion not in the fusion-competent cells, whereas the second, process (Bour et al., 2000; Ruiz-Gómez et al., 2000 [this named Sticks-and-stones (Sns; Bour et al., 2000), is issue of Cell]). specifically expressed in fusion-competent cells. Both Drosophila Myoblast Fusion: An Asymmetric Process proteins are essential for myoblast fusion and null mutaInvolving Differentially Expressed Ig tions in either the duf or the sns gene cause a complete Domain Molecules failure of fusion. Thus, these are the first fusion compoAs was first observed in the grasshopper (Ho et al., nents that have been found to be differentially expressed 1983), in Drosophila each muscle is “seeded” by a single on the surfaces of the two types of somatic mesodermal founder cell (Bate, 1990; Dohrmann et al., 1990). There cells that participate in fusion. By contrast, previously are no fixed lineage relationships between a particular identified components of the fusion machinery, includfounder myoblast and its prospective fusion partners, ing Drosophila Rac1, Myoblast city (Mbc/Dock180), and nor are groups of specific fusion-competent myoblasts Blown fuse (Blow), are intracellular and, as far as is allocated to fuse with a particular founder cell. Rather, known, not differentially expressed in the two myoblast it appears that founder cells simply select the fusionpopulations (Luo et al., 1994; Doberstein et al., 1997; competent cells that happen to be their closest neighErickson et al., 1997). Duf and Sns appear to act at an bors as preferred fusion partners. Muscle size is then early step of the myoblast fusion process in helping to largely determined by the number of cells that have determine the polarity of the fusion event and probably contributed to the syncytium, and thus the extent of also in the fusion process itself. fusion must also be imposed by the founder cell. This A Role for Duf in the Recognition and Attraction implies coordination of the generic fusion program and of Fusion-Competent Cells by Muscle Founders the specification programs that determine founder cell Myoblast fusion normally occurs in a series of steps that identities. The latter are distinct for each of the z30 can be distinguished morphologically and genetically founder cells within a hemisegment and also determine (Bate, 1990; Doberstein et al., 1997; Paululat et al., 1999). other characteristics of future muscle fibers, such as Fusion-competent cells first extend filopodia toward a their shape, attachment sites, and specific innervations. nearby founder cell. Upon cell contact an area of tight There is evidence that founder cell identity is imparted by membrane contact is established, at which both cells specifically expressed transcription factors (reviewed deposit vesicles and other electron-dense material. The in Baylies et al., 1998; Frasch, 1999), but how these area of membrane contact increases as the two cells line factors influence the extent of muscle fusion is currently up side by side, and the plasma membranes separating them eventually vesiculate, leading to the fusion of the two cells. Mutations exist that interfere with this process ‡ E-mail: [email protected] (M. F.), [email protected]