The Role of Microtubule Retrograde Flow in Neuronal Development

The microtubule array in developing neurons consists of more than a hundred microtubules in each neurite. This microtubule array has been thought to act as stationary tracks, in which only individual microtubules are dynamic and could be transported. Using live cell imaging combined with photoconversion of microtubule patches I reveal that instead the whole microtubule array flows retrogradely into the soma. Before axon development, this microtubule retrograde flow (MT-RF) fuels the fluctuating state of neurons by decreasing microtubule density, leading to the longknown hallmark of microtubule density cycles instead of stable increases in microtubule density. MT-RF then slows down in the axon, right after initial axon formation, reducing microtubule density cycles to allow a more stable increase in microtubule density and thereby driving axon extension. I also observed that MT-RF later slows down in dendrites. By combining live-cell imaging with pharmacological and chemogenetic recruitment to the plasma membrane, I show that Dynein at the plasma membrane can fuel MT-RF. To understand the molecular effect of MT-RF on the microtubule array, I developed a biophysical model of microtubules in neurites across developmental stages and combined it with quantitative immunocytochemistry and microtubule dynamics data. Surprisingly, my model revealed that MT-RF slowdown is needed for both axon and dendrite development, uncovering a unified model of axon and dendrite development, in contrast to the prevailing idea that axon and dendrite development are triggered by different programs. In dendrites, my model and experiments show that MT-RF slowdown enables an efficient increase in stable microtubules. In axons, my model predicts that without slowdown of MT-RF, iii microtubules cannot reach close to the axon tip, blocking axon growth. MT-RF slowdown together with only three properties of the axonal and dendritic microtubule array, stability and two types of nucleation, is sufficient to explain the stable microtubule distribution in axons and dendrites, which drives neuronal polarization. In this thesis I found a novel type of cytoskeletal mechanism at an unprecedented scale in neurons and characterized its molecular function across developmental stages to find that it unifies the fluctuating state before axon development as well as axon and dendrite development. This provides new research avenues from the role of MT-RF in other processes and non-neuronal cell types to understanding how similar instead of different processes in axons and dendrites could underly neurodevelopmental.