Molecular Grammar of Microtubule-Wetting Condensates

Microtubule condensate interactions are fundamental for cell division, vesicle transport and cellular locomotion. Accordingly, they represent a large number of attractive drug targets. Due to the size of microtubuli and the slow timescale of condensate structural relaxation, there has not been a systematic investigation at the molecular level as to what binding patterns (molecular grammar) enable condensate binding to microtubuli. We provide a protocol that is able to predict whether any given disordered protein sequence will bind to microtubuli. This protocol is suitable for high-throughput screening. Our pattern analysis allows us to establish two categories of strongly interacting subsequences that enable binding to microtubuli: positively charged hydrophobic clusters and alternating charge sequences. Their overall optimal balance is analyzed and preferential regions of interaction on microtubuli are identified and validated with known experimental results. Our results enable rapid prototyping of proteins that target the microtubule surface, i.e. they predict whether unstructured proteins will wet the microtubule interface.