Functional Dissection of the N-Terminal Domains of KIF14: A Mitotic Kinesin Required for Cell Abscission

Abstract

Kinesins are the largest superfamily of microtubule-binding motor proteins in eukaryotic cells. All kinesins contain a motor domain that hydrolyzes ATP to power their directional movement along microtubules, or to catalyze the release of tubulin subunits from microtubule ends. These activities support essential functions such as vesicle and organelle transport, chromosome segregation during mitosis and meiosis, and regulation of microtubule dynamics for cell growth and division. The mitotic kinesin KIF14 is distinguished from other kinesins by a long N-terminal extension (NTE) (355 residues) of its motor domain that enables it to bind and bundle actin filaments. The ability of KIF14 to bind to actin is unusual because myosins have long been considered as the only cytoskeletal motors to act directly upon actin filaments. The actin binding ability of KIF14 appears to correlate with its co-localization within the actin-rich contractile ring at the cell midbody where it plays a vital, but undefined, role during cytokinesis. Accordingly, loss or disruption of KIF14 activity leads to cytokinesis failure. Conversely, KIF14 overexpression is associated with poor prognosis for multiple types of cancer. The first research aim of this thesis was to locate the binding sites of a panel of KIF14-selective inhibitors identified from a high-throughput screen. Unable to obtain crystal structures of the KIF14-inhibitor complexes, we used molecular docking to predict their interaction modes computationally. The docking software “Glide” identified a unique binding pocket between helix 2 and 3 of the KIF14 motor domain, which is directly adjacent to the site of ATP hydrolysis. Visual inspection and computational analysis of each inhibitor’s binding conformation enabled the design of a library of derivatives of these inhibitors; some of which were predicted to have improved affinity and specificity for KIF14 upon repeating the molecular docking workflow. The second aim was to understand how the NTE of KIF14 binds actin and affects actin filament structures. By systematically dissecting the human KIF14 NTE into a series of short polypeptide fragments, we were able to locate distinct microtubule- and actin-binding regions. By a combination of pyrene-actin polymerization assays and TIRF microscopy, we further identified the regions of the NTE that stimulate F-actin polymerization and F-actin bundling. We also obtained evidence that the NTE promotes actin polymerization by acting as an actin nucleation factor and an F-actin stabilizer. Interestingly, these abilities of the NTE are influenced by the microtubule-binding interactions of KIF14’s motor domain. These new insights provide the first building blocks of a molecular model that explains how KIF14 influences actin dynamics at the midzone during cytokinesis, and could help guide the development of strategies to inhibit KIF14 activities for the benefit of people living with cancer.