NuMA’s microtubule-binding region is located close to its C-terminus. (A) Schematic of mScarlet-tagged NuMA^FL and C-terminal fragments, indicating the main structural parts of NuMA and functional domains identified in the C-terminal region (microtubule-binding domain [MTBD] 1 [Du et al., 2002] and 2 [Gallini et al., 2016; Chang et al., 2017], clustering domain [Okumura et al., 2018], LGN oligomerization domain [Pirovano et al., 2019], LGN binding domain [LGNBD] [Pirovano et al., 2019], nuclear localization signal [NLS] [Chang et al., 2017]). (B) Coomassie Blue-stained SDS-PAGE showing the pellet (P) and supernatant (S) fractions of microtubule co-sedimentation assays. 1 µM mScarlet-tagged NuMA^FL or C-terminal fragments were incubated alone or with Atto647N-labeled GMPCPP-microtubules (0.5 µM polymerized tubulin) at 30°C for 15 min. Upon centrifugation, in the absence of microtubules, all proteins remained in the supernatant (left gel). When incubated with microtubules, they co-sedimented with the microtubule pellet (green arrow) at different degrees. NuMA^{C-term L} was found exclusively in the pellet; NuMA^FL predominantly in the pellet; NuMA^{C-term S2} partially in the pellet; NuMA^{C-term S1} entirely in the supernatant. Arrows indicate the expected molecular weight for each NuMA construct. (C) Representative TIRF microscopy images of 40 nM mScarlet-tagged NuMA constructs binding to surface-immobilized Atto647N-labeled GMPCPP-microtubules. (D) Background-corrected and microtubule length-corrected mScarlet intensity of the different NuMA constructs binding to microtubules as shown in C (mean of medians ± SEM). Each big circle represents the median value of one replicate; each small circle represents the intensity on a single microtubule; n = 45 microtubules for all conditions; P values by Welch’s t test: 0.0364 (FL versus C-term L), 0.1346 (FL versus C-term S1), 0.0530 (FL versus C-term S2). (E) Clustering propensity: sums of the background-corrected intensities of all outliers of the mScarlet NuMA intensity distributions for the different NuMA constructs shown in C nonspecifically bound to the glass surface (mean of medians ± SEM). Each big circle represents the median value of one replicate; each small circle represents the clustering at one surface area of one replicate; n = 9 areas for all conditions; adjusted P values by Welch’s ANOVA test with Holm-Sidak’s post-hoc test for multiple comparisons: 0.0297 (FL versus C-term L), 0.2367 (FL versus C-term S2), 0.0358 (C-term L versus C-term S2). (F) Representative TIRF microscopy images of 40 nM mScarlet-tagged NuMA constructs binding to surface-immobilized Atto647N-labeled GMPCPP-microtubules in the presence of 10 µM Atto647N-tubulin. Red arrowheads indicate selective end binding; yellow arrowheads indicate examples of NuMA^{C-term L} clusters nonspecifically adsorbed to the surface. (G) Background-corrected and microtubule length-corrected mScarlet intensity of the different NuMA constructs binding to microtubules in the presence of 10 µM Atto647N-tubulin as shown in E (mean of medians ± SEM). Symbols as in D; n = 45 microtubules for all conditions; adjusted P values by Welch’s ANOVA test with Holm-Sidak’s post-hoc test for multiple comparisons: 0.2825 (FL versus C-term L), 0.0143 (FL versus C-term S2), 0.0184 (C-term L versus C-term S2) (H) Microtubule end selectivity: percentage of microtubules showing mScarlet signal exclusively at one end, of all microtubules with an mScarlet signal, for mScarlet-NuMA constructs binding to microtubules in the presence of tubulin as shown in C (mean ± SEM). Each color represents a replicate; n = 126, 71, 161 microtubules; adjusted P values by Welch’s ANOVA test with Holm-Sidak’s post-hoc test for multiple comparisons: 0.1339 (FL versus C-term L), 0.0028 (FL versus C-term S2), 0.1732 (C-term L versus C-term S2). All data are from three biological replicates. All assays were executed in NuMA microscopy buffer. Source data are available for this figure: SourceData F2.