Boundary-sensing mechanism in branched microtubule networks

The self-organization of cytoskeletal networks in confined geometries requires sensing and responding to mechanical cues for dynamic adaptation. Here, we show that branching of microtubules (MTs) via branching MT nucleation combined with dynamic instability constitutes a boundary-sensing mechanism within confined spaces. Using a nanotechnology platform, we observe the self-organization of a branched MT network in a channel with a narrow junction and a closed end. Our observations show that branching MT nucleation occurs in the post-narrowing region only if it exceeds a certain length before terminating at the channel's closed end. We further show that this length dependency is tunable and forms the basis for mechanical feedback that adapts MT networks to local geometry, enabling tunable MT network formation in response to physical boundaries in confined spaces. After experimental characterization of boundary-sensing feedback, we propose a minimal model and conduct numerical simulations. We investigate how this feedback, wherein growing MTs dynamically sense their environment and provide nucleation sites for new MTs, sets a length/time scale that steers the architecture of MT networks in confined spaces. This "search-and-branch" mechanism has implications for the formation of MT networks during neuronal morphogenesis, including axonal growth and the formation of highly branched dendritic networks, as well as for plant development, MT-driven guidance in fungi, and engineering nanotechnologies.