Fig 2: KIF4A binds to B56 via a conserved basic patch and an LxxIxE motif.(A) B56 interactors with the basic patch (blue) and LxxIxE motif (green) sequences shown; Δ indicates the number of residues between the basic patch and the LxxIxE motif. (B) Binding isotherm of WT KIF4A1192-1232 with B56γ. (C) Binding isotherm of KIF4A1192-1232,bpm (1208KKK1210 to AAA) with B56γ. (D) Cartoon representation of the effect of mutating the basic patch (BP) of the bp-dependent interactors on their interaction with B56 (AP, acidic patch). (E) Immunoprecipitation of YFP-B56α from cells stably expressing YFP-B56α and transfected with the indicated myc-tagged full-length KIF4A variants; asterisk indicates YFP, which was used as a control. The amounts of myc-KIF4AFL co-purified with YFP-B56α were normalized to the band intensity of YFP. The wt is set to 1. (F) Immunoprecipitation of YFP-B56α from cells stably expressing YFP-B56α and transfected with the indicated myc-tagged AIM1625-900 variants. The amounts of myc- AIM1625-900 co-purified with YFP-B56α were normalized to the band intensity of YFP. (G) Immunoprecipitation of YFP-B56α from cells stably expressing YFP-B56α and transfected with the indicated myc-tagged NHS1419-1631 variants. The amounts of myc-NHS1419-1631 co-purified with YFP-B56α were normalized to the band intensity of YFP.

Fig 3: The basic patch in B56-specific regulators binds B56 via a dynamic charge-charge interaction.(A) Overlay of the 2D [1H,15N] HSQC spectra of 15N-labeled KIF4A1192-1232,LE,PE in the presence (red) and absence (black) of B56γ (1:1 ratio); basic patch and LxxIxE residues labeled blue and green, respectively. (B) Overlay of the 2D [1H,15N] HSQC spectra of 15N-labeled KIF4A1192-1232,LE,PE in the presence (red) and absence (black) of B56γ3R E276R/E310R/E316R (1:1 ratio); basic patch and LxxIxE residues highlighted in blue and green, respectively. (C) [1H,15N] HSQC peak intensity ratios for spectra shown in A, B (black, red, respectively). (D) Cartoon representation of the effect of mutating the acidic patch (AP) of B56 on KIF4A:B56 binding (AP: acidic patch, BP: basic patch). (E) Immunoprecipitation of stably expressed YFP-B56α variants (wt, B56α2R: E335R/D338R, and B56α3R E301R/E335R/D338R and probed for endogenous KIF4A, PP2AC (PP2A catalytic subunit) and GFP (YFP-B56α). (F) Immunoprecipitation of transiently transfected myc-tagged KIF4A1001-1232 C-terminal variants (A1: K1208A; A2: 1208KK1209 to AA; A3: 1208KKK1210 to AAA; A4: 1208KKKK1211 to AAAA) from cells stably expressing YFP-B56α or YFP-B56γ. The amounts of myc-KIF4A co-purified with YFP-B56 were normalized to the band intensity of YFP.

Fig 4: SUMOylation does not affect ATPase activity of KIF4A. (A) In vitro SUMOylation reactions were performed on negative control samples and samples containing wild-type (WT) or SUMOylation-deficient (K460R and E458A/E461A) recombinant GST–KIF4A. Additional negative control assays were carried out in the absence of SUMO2 as indicated. Subsequently, an ATPase activity assay was performed in the absence or presence of MTs and absorbance was measured. The mean±s.d. percentage of relative ATPase activity for three independent experiments is shown. A two-sided Student's t-test was used to calculate P-values and the difference was considered not significant (ns) when P>0.05. (B) Aliquots of the in vitro SUMOylation samples, used for the experiment described in Fig. 3A, were used to verify SUMOylation efficiency by immunoblotting with an antibody against KIF4A.

Fig 5: KIF18A is required for the proliferation of chromosomally unstable cells.a Fold change in cell density (cells/mm2) after 96 h in the indicated cell lines following knockdown (KD) of kinesin proteins. Data are normalized to cells treated with control siRNA. Raw cell counts for proliferation assays are available in Source Data. n (wells of cells/independent experiments) = 24/3 (MCF10A, MDA-MB-468, and HCC1806), 64/5 (MDA-MB-231 control and KIF18A KD), 36/3 (MDA-MB-231 KIF18B, KIF4A, KIF22, and MCAK KD). b Representative images of MDA-MB-231 (from five experiments) and MCF10A cells (from three experiments) treated with either control or KIF18A siRNA. Scale bars are 100 microns. c Normalized fold change in cell density (cells/mm2) of MSI and CIN colorectal cancer cell lines after 96 h or until cells in the control populations reached confluence. n (wells of cells/independent experiments) = 88/6 (HCT116); 104/5 (LoVo control KD) and 128/5 (LoVo KIF18A KD); 92/5 (SW480); 44/3 (HT29 control KD) and 48/3 (HT29 KIF18A KD); and 32/3 (LS1034). d Normalized fold change in cell density (cells/mm2) of HeLa Kyoto cells after 96 h or until cells in the control populations reached confluence. n = 40 wells of cells from two independent experiments. Data were analyzed using a one-way ANOVA with post hoc Tukey’s test to compare the control to each of the knockdown conditions (a) or with an unpaired, two-tailed t test (c, d). All graphs show mean ± SD and individual data points. P values <0.05 from the indicated statistical tests are displayed on plots.