Reverse Engineering the Programming Logic of Cytoskeletal Dynamics

Author: Larios-Colorado, David Antonio

Year: 2026

Degree: Dissertation (Ph.D.)

Advisor: Thomson, Matthew

Committee Members: Elowitz, Michael B.; Phillips, Robert B.; Parker, Joseph; Thomson, Matthew

Option: Biology

Abstract

Eukaryotic cells generate mechanical force through cytoskeletal filaments actively reorganized by families of molecular motors. Despite extensive characterization of filament-motor self-organization, how variations in motor sequence and structure translate into filament organization dynamics remains poorly understood. Here we develop ActiveDROPS (Active Dynamic Reprogrammable self-Organizing Protein System), a cell-free approach to reconstitute microtubule dynamics driven by genetically encoded kinesin variants in bacterial lysate droplets. Across twelve diverse kinesin-1 homologs, microtubule dynamics collapse onto three classes: "Slow-Sustained" flows that activate near 8-10 h and persist to ~30 h, "Fast-Burst" contractions initiating within minutes and dissipating within ~1 h, and "Multiphase" progression through nematic, rotational, and contractile flows over ~30 h. Microtubule gliding assays and molecular dynamics simulations on AlphaFold-predicted structures show that Fast-Burst motors couple high ATP hydrolysis rate with strong microtubule binding, while Slow-Sustained motors couple low turnover with weak binding. By recombining two motors across these classes, we generated a "Fast-Sustained" chimera with rapid microtubule flows lasting ~15 h, revealing the protein domain configurations that specify the velocity (Slow/Fast) and duration (Sustained/Burst) of macroscopic dynamics. These results uncover a constrained modular logic by which kinesin sequence specifies microtubule self-organization, providing a framework for dissecting the mechanical repertoire available to cellular systems.