In cells, mRNAs are transported to and positioned at subcellular areas to locally regulate protein production. Recent studies have identified the kinesin-3 family member motor protein KIF1C as an RNA transporter. However, it is not clear how KIF1C interacts with RNA molecules. Here, we show that the KIF1C C-terminal tail domain contains an intrinsically disordered region (IDR) that drives liquid–liquid phase separation (LLPS). KIF1C forms dynamic puncta in cells that display physical properties of liquid condensates and incorporate RNA molecules in a sequence-selective manner. Endogenous KIF1C forms condensates in cellular protrusions, where mRNAs are enriched in an IDR-dependent manner. Purified KIF1C tail constructs undergo LLPS in vitro at near-endogenous nM concentrations and in the absence of crowding agents and can directly recruit RNA molecules. Overall, our work uncovers an intrinsic correlation between the LLPS activity of KIF1C and its role in mRNA positioning. In addition, the LLPS activity of KIF1C’s tail represents a new mode of motor-cargo interaction that extends our current understanding of cytoskeletal motor proteins.
Microtubules are cytoskeletal filaments that provide structural support for numerous cellular processes. Despite their high rigidity, microtubules can be dramatically bent in cells, and it is unknown how much force a microtubule can withstand before breaking. We find that the kinesin-3 motor KIF1C forms condensates that entangle and break neighboring microtubules. Combining computational simulations and experiments, we show that microtubule breakage is an emergent property that is dependent on a highly processive kinesin motor domain, the cluster properties, cytoplasmic viscosity, and microtubule anchors. We estimate a rupture force for microtubules in cells that is lower than previous estimates based on in vitro studies with taxol-stabilized microtubules. The absence of microtubule breakage under physiological conditions suggests that mechanisms exist to protect microtubule integrity, which may inform about physical constraints on the evolution of motor proteins. We suggest that release of either the motor–cargo or motor–microtubule interaction prevents the accumulation of mechanical stress upon the engagement of multi-motor clusters with microtubules.