Kif18A uses a microtubule binding site in the tail for plus-end localization and spindle length regulation

Abstract

The mitotic spindle is a macromolecular structure utilized to properly align and segregate sister chromatids to two daughter cells. During mitosis, the spindle maintains a constant length, even though the spindle microtubules (MTs) are constantly undergoing polymerization and depolymerization [1]. Members of the kinesin-8 family are important for the regulation of spindle length and for chromosome positioning [2-9]. Kinesin-8 proteins are length-specific, plus-end-directed motors that are proposed to be either MT depolymerases [3, 4, 8, 10, 11] or MT capping proteins [12]. How Kif18A uses its destabilization activity to control spindle morphology is not known. We found that Kif18A controls spindle length independently of its role in chromosome positioning. The ability of Kif18A to control spindle length is mediated by an ATP-independent MT binding site at the C-terminal end of the Kif18A tail that has a strong affinity for MTs in vitro and in cells. We used computational modeling to ask how modulating the motility or binding properties of Kif18A would affect its activity. Our modeling predicts that both fast motility and a low off rate from the MT end are important for Kif18A function. In addition, our studies provide new insight into how depolymerizing and capping enzymes can lead to MT destabilization.

Figure 1. Kif18A regulation of spindle length is dependent on K-fibers

(A) HeLa cells transfected with Luciferase (Control), or the indicated siRNAs were stained for MTs (white) and DNA (blue). Scale bar, 10 μm. (B) Western blot of cells treated with Luciferase, Kif18A, HSET, or Kif18A+HSET siRNAs and then probed with anti-Kif18A, anti-HSET, or anti-tubulin antibodies. (C) Quantification of the spindle lengths from at least three independent experiments. For each knockdown condition, a total of at least 60 cells were scored for spindle length, and dot plots showing the lengths are graphed with the mean ± SD indicated by the bar and whiskers. (D) The percentages of cells containing bipolar spindles with aligned chromosomes for each condition were determined from > 100 total cells in three independent experiments, and the mean ± SEM is graphed. Asterisk indicates p < 0.05 relative to control.

Figure 2. Overexpression of GFP-Kif18A domain deletion proteins does not alter spindle length or chromosome alignment

(A) HeLa cells were transfected with the indicated GFP-fusion construct (green) and stained for MTs (magenta) and DNA (blue). Scale bar, 10 μm. The boxed region represents the region used to generate the linescans, which show the localization of Kif18A (green) relative to DNA (blue) and MTs (magenta). (B) Spindle lengths from at least three independent experiments in which a total of 30 spindles for each Kif18A domain construct are graphed as dot plots with the mean ± SD indicated by the bar and whiskers. An asterisk indicates p < 0.05 relative to control. (C) The average percentages of transfected cells with aligned chromosomes for each construct are indicated as the mean ± SEM from > 100 total cells in three independent experiments.

Figure 3. The tail domain of Kif18A binds MTs in vitro and in vivo

(A) Schematic diagrams of Kif18A constructs. (B) GST-Kif18A(593–898), GST-Kif18A(727–898), or GST-Kif18A(802–898) (1.3 μM) were incubated with increasing concentrations of pre-assembled MTs (0–2 μM) (T) for 15 min at RT in the absence of nucleotide. Soluble proteins were separated from MT-bound proteins by ultracentrifugation, and equivalent amounts of supernatant (S) and pellet (P) were analyzed by SDS-PAGE and Coomassie Blue staining. (C) The amount of protein in the (S) and (P) were quantified, and the binding curves were fit to the one-site quadratic MT binding equation. Each point on the curve represents the mean ± SEM from at least three individual experiments. (D) Summary table of binding data in (C). (E) HeLa cells were transfected with GFP fusions of Kif18A(802–898) or Kif18A(1–801) (green) and stained for DNA (blue) and MTs (magenta). The boxed region represents the region used to generate the linescans, which show the localization of Kif18A (green) relative to DNA (blue) and MTs (magenta). Scale bar, 10 μm. (F) The average spindle lengths for each construct relative to control are graphed as dot plots with the mean ± SD indicated by the bar and whiskers for > 30 cells in at least three independent experiments.

Figure 4. Mathematical modeling predicts MT plus-end affinity is important for regulation of MT length

Simulations were performed in which 20 dynamic MTs were nucleated, and 50 Kif18A motors were added with a variety of different dynamics parameters, including the k_on^*, k_off,lattice, k_off,end, and motor velocity. (A) A MT depolymerase was defined as a motor that increases the catastrophe frequency. (B) A MT capping protein was defined as a protein that reduces MT growth velocity to 0 nm s⁻¹. (C–H) Average MT length ± SEM of 75 trials recorded at the end of each 60 min simulation. The control value is indicated by the dashed line. The average MT length decreases with increasing V_motor for a MT depolymerase (C) or a capping protein (D). The average MT length decreases with increasing k_on^* for a MT depolymerase (E) or a capping protein (F). Decreasing the k_off,end from the end has only a modest effect on the activity of a MT depolymerase (G); however as k_off,end is increased, a MT capping protein has a decreased ability to destabilize MTs.