Microtubules do not start growing spontaneously in cells; they rely on supporting protein machinery to nucleate tubulin building blocks and kick off polymerization. To find their origins, Efimov et al. tracked the growth of new microtubules in living cells. They found that the majority of new microtubules radiated out from a central position in the cell that is consistent with centrosomes.
A significant proportion of nucleation events, however, appeared to occur at the Golgi. When the team obliterated centrosomes using laser beams, they found that new microtubules indeed continued to sprout from the Golgi. Golgi, but not centrosomal, nucleation required a microtubule-binding protein called CLASP.
Microtubules grew out from Golgi in one direction—away from the nucleus and toward the cell's periphery. In migrating cells, this directionality corresponded with the direction of migration, as microtubules from the Golgi grew toward the cell's leading edge. Golgi-mediated nucleation might therefore support the asymmetric architecture and secretory vesicle trafficking of polarized motile cells.