Fusion of nascent myoblasts to pre-existing myofibres is critical for skeletal muscle growth and repair. The vast majority of molecules known to regulate myoblast fusion are necessary in this process. Here we uncover, through high-throughput in vitro assays and in vivo studies in the chicken embryo, that TGFβ (SMAD2/3-dependent) signalling acts as a molecular brake on muscle fusion. While constitutive activation of the pathway arrests fusion, its inhibition leads to a striking over-fusion phenotype. This dynamic control of TGFβ signalling in the embryonic muscle relies on a unique receptor complementation mechanism, prompted by the merging of myoblasts with myofibres, each carrying one component of the heterodimer receptor complex. The competence of myofibres to fuse is restored through endocytic degradation of activated receptors. Altogether, this study shows that muscle fusion is a self-regulated process that relies on cyclic TGFβ signalling to regulate its pace.
MainMyoblast fusion occurs after muscle specification and early differentiation, themselves regulated by the Myogenic Regulatory Factors (MRFs) MYF5, MYOD and MYOG 1-3 . While MRF function is necessary for myoblast fusion 4 , terminal differentiation, including muscle contraction, can occur even if fusion is disrupted 5-7 . Work in Drosophila identified a number of molecules, mainly implicated in actin regulation, required for muscle fusion in this organism. A major advance was the demonstration that their homologs play similar function during vertebrate myogenesis 2,3,8 . Together with vertebrate-specific muscle fusion genes (e.g. Myomaker, Myomixer, JAM2-3) they constitute the "fusion machinery" necessary for the fusion of myoblasts to myofibres. Recent analyses we performed on myoblast fusion during embryonic development suggested that additional mechanisms must exert temporal and spatial control on the fusion machinery, modulating whether myoblasts and myofibres that are competent to fuse do so and at what pace 9 . The molecular underpinning of such control over fusion is completely unknown.