(a) Scheme illustrating the domain structure of human chTOG. (b) Coomassie Blue-stained SDS-PAGE gel showing 1 μg of purified recombinant chTOG constructs. (c) Scheme of the experimental setup. (d) Dual colour TIRF microscopy kymographs showing Atto647N-labelled microtubules (magenta) growing in the absence or presence of 4 nM full-length chTOG-mGFP (green). (e) Plot of the mean growth speed as a function of the chTOG concentration. Data points, black; error bars are SEM; Hyperbolic fit (one-site binding) - magenta curve. Number of 25 s-growth intervals used to measure individual velocities for each chTOG concentration: 0 nM – 1,003; 5 nM – 966; 20 nM – 661; 50 nM – 499; 100 nM – 510; 200 nM – 600. Atto647N-labelled tubulin concentration in (d) and (e) was 7.5 μM. Data were pooled from two datasets. (f) Kymographs showing GMPCPP-stabilised Atto647N-labelled microtubules in the absence of soluble tubulin and either in the absence (left) or presence (right) of 100 nM chTOG. Scale bars as indicated.

(a) Scheme of the human TPX2 constructs used in this study: full-length TPX2, N-terminally truncated TPX2 containing amino acids 274 - 747 (TPX2^ΔN), and a minimal TPX2 construct containing amino acids 274 - 659 (TPX2^mini). Regions known to interact with Aurora A (AurA), importin α (Imp α) are indicated, together with predicted coil coils (CC) and nuclear localisation signal (NLS). (b) Coomassie Blue-stained SDS-PAGE gel showing 1 μg of purified recombinant TPX2 constructs. (c, d) Single channel and merged TIRF microscopy images showing how mGFP-TPX2 (green in merge) at 5 nM (c) and 0.35 nM (d) binds either all along (c) or preferentially to the growing ends and the GMPCPP segment (d) of a growing Atto647N-labelled microtubule (magenta in merge) (“−“ and “+” indicate end binding, the GMPCPP “seed” is marked by an arrowhead). (e) Kymographs depicting the time course of binding of 5 nM mGFP-TPX2 all along a growing microtubule. Att647N-labelled tubulin concentration was 7.5 μM. (f) Kymographs depicting the time course of 0.35 nM mGFP-TPX2 binding to a growing microtubule end and the GMPCPP “seed”. Att647N-labelled tubulin concentration was 12.5 μM. (g, h) Kymographs showing binding of (g) 10 nM mGFP-TPX2^ΔN and (h) 33 nM mGFP-TPX2^mini to dynamic microtubules (merged channels on the left, mGFP-TPX2 on the right). Atto647N-labelled tubulin concentration is 12.5 and 15 μM, respectively. (i, j) Kymographs showing that neither (i) 1 nM full-length mGFP-TPX2 nor (j) 10 nM TPX2^mini binds to shrinking microtubules. Atto647N-labelled tubulin concentrations were 5 μM and 7.5 μM, respectively. For all kymograph pairs: merged channel - left, mGFP-TPX2 channel - right. Scale bars as indicated.

(a) Modified box-and-whiskers graph for the microtubule growth speeds in the absence (control) and presence of different concentrations of full-length mGFP-TPX2 and mGFP-TPX2^mini, as indicated. Number of observed microtubule growth episodes per condition: control – n=150; mGFP-TPX2: 0.2 nM – n= 160, 1 nM – n=150, 5 nM – n=114; mGFP-TPX2^mini: 10 nM – n=153, 50 nM – n=121, 250 nM – n=154. All events are from one dataset each. (b) TIRF microscopy kymographs showing dynamic Atto647N-labelled microtubules in the absence (control) or presence of 5 nM mGFP-TPX2 or 250 nM mGFP-TPX2^mini. Scale bars as indicated. Bar graphs showing microtubule (c) catastrophe and (d) rescue frequencies, and box-and-whiskers graphs showing microtubule (e) growth and (f) depolymerisation speeds, for control, 5 nM mGFP-TPX2, and 250 nM mGFP-TPX2^mini, as indicated. Box-and-whiskers graphs show (g) microtubule lifetimes, (h) depolymerisation times and (i) depolymerisation lengths for the same conditions. Data for (c-i) were pooled from three datasets each. Total number of analysed events per condition: (c) catastrophes (total growth time in brackets): control – n=757 (457,280 s), mGFP-TPX2 –n= 209 (267,644 s), mGFP-TPX2^mini – n=385 (409,675 s); (d) rescues (total depolymerisation time in brackets): control – n=612 (22,062 s), mGFP-TPX2 – n=209 (10,708 s), mGFP-TPX2^mini – n=376 (13,130 s); (e) growth episodes: control – n=863, 5 nM mGFP-TPX2 – n=271, 250 nM mGFP-TPX2^mini –n=477; (f) depolymerisation episodes: control – n=756, 5 nM, mGFP-TPX2 – n=211, 250 nM mGFP-TPX2^mini –n=383. The same data were used in (g)-(i) (see ). Atto647N-labelled tubulin concentration was always 7.5 μM. Errors in bar graphs are SEM. For the modified box-and-whiskers plots the boxes range from 25^th to 75^th percentile, the whiskers span from 10^th to 90^th percentile, the horizontal line marks the mean value. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001 (only displayed for comparisons with control); determined for the comparison of mean values analysing raw data (Tukey’s test in conjunction with One Way ANOVA).

(a) Right: Images of 33 nM mGFP-TPX2^mini (green in merge) at growing Atto647N-labelled microtubule ends (magenta in merge) at the indicated tubulin concentrations. Left: Averaged mGFP-TPX2^mini intensity profiles at the end regions of microtubules growing in the presence of 33 nM mGFP-TPX2^mini and the indicated Atto647N-labelled tubulin concentrations. (b) Averaged mGFP-TPX2^mini intensity profiles at the end regions of microtubules growing in the presence of 33 nM mGFP-TPX2^mini and 20 μM Atto647N-labelled tubulin for three different time windows after start of growth. (c) Averaged mGFP-TPX2^mini intensity profiles for the indicated mGFP-TPX2^mini concentrations at 25 μM Atto647N-labelled tubulin. (d) Cartoons depicting the assay conditions and TIRF microscopy kymographs showing single molecule binding events of 5 nM mGFP-TPX2^mini (green in merge) to the ends of growing Atto565-labelled microtubules (blue in merge) in the absence or presence of 195 nM Alexa647-labelled SNAP-TPX2^mini (magenta in merge). (e) Dissociation rate constant k_off and association rate r_on, respectively, as determined for the conditions in (d) from the dwell and waiting time distributions of the single mGFP-TPX2^mini molecule binding and unbinding events (, ) as a function of the distance from the growing microtubule plus end. Note that the increase in k_off values and decrease in r_on values is less steep in the presence of excess Alexa647-labelled SNAP-TPX2^mini, agreeing with (4c). Lower panel: The normalised total fluorescence intensity of all mGFP-TPX2^mini binding events shows their steady state distribution. (f) Kymographs and instantaneous intensity line scans at different times, as indicated, showing strong temporal fluctuations of the fluorescence intensity of mGFP-TPX2^mini (25 nM, green in merge) at growing Atto647N-labelled microtubule ends (magenta in merge, 12.5 μM tubulin). (g) TIRF microscopy images showing 33 nM mGFP-TPX2^mini (green in merge, lower panels) accumulating strongly at the curved microtubule ends of Atto647N-labelled growing microtubules (magenta in merge, 30 μM tubulin). (h) Kymographs of 5 nM mGFP-TPX2 (green in merge) localising to the ends of Atto647N-labelled microtubules (magenta in merge, 25 μM tubulin) growing in the presence of the slowly-hydrolysable GTP analogue GTPγS. Scale bars as indicated.

(a) Scheme of the experimental setup for the ‘surface’ nucleation assay. (b) TIRF microscopy images from a time series showing Atto647N-labelled microtubules and/or tubulin particles on a glass surface with immobilised biotinylated Kin1^rigor mutant, chTOG or TPX2 after pre-incubation at the indicated concentrations. Insets on the right depict magnified images of the same surface at 7.5 min. (c) Time courses of the mean Atto647N-tubulin fluorescence intensities measured at the surface for the entire field of view for different chTOG, TPX2 and Kin1^rigor densities. Concentrations of biotinylated proteins used for surface incubation as indicated. The same Kin1^rigor curve is shown as a control for both chTOG and TPX2 plots. Note that the fluorescence signal typically represents the sum of several different tubulin species: soluble tubulin in the TIRF field (creating part of the background), immobilised individual tubulins (e.g. bound to chTOG), tubulin ‘stubs’ (bound to TPX2) and microtubules (bound to chTOG or TPX2). Raw intensities including background are shown. (d) Time series of TIRF microscopy images showing nucleation and growth of Atto647N-labelled microtubules on surfaces with immobilised biotinylated TPX2 (pre-incubated at 125 nM, top three rows) or, for controls, with a biotinylated Kin1^rigor (pre-incubated at 125 nM, bottom two rows) in either the absence or presence of 100 nM untagged chTOG, as indicated. Because of the high microtubule density at the TPX2 surface in the presence of chTOG, the same time series is also shown with reduced contrast for better visualisation of individual microtubules. Atto647N-labelled tubulin concentration was always 12.5 μM. Scale bars as indicated. t = 0 when the sample is placed at 30°C.

(a) Scheme of the experimental setup for the ‘solution’ nucleation assay. (b) Time series of TIRF microscopy images showing Atto647N-labelled microtubules that nucleated in solution in the presence of nucleation factors as indicated, followed after 1 min by binding to a neutravidin-coated surface via the biotinylated protein present as indicated. (c) TIRF microscopy images showing examples the concentration dependent nucleation efficiencies of chTOG- and TPX2-mediated microtubule nucleation in the ‘solution’ nucleation assay. (d) Time courses of mean Atto647N-tubulin fluorescence intensities measured for the entire field of view in ‘solution’ at the biotinylated chTOG and TPX2 concentrations as indicated. Note that the fluorescence signal typically represents the sum of several different tubulin species (see note in legend of ). Raw intensities including background are shown. Atto647N-labelled tubulin concentration was always 12.5 μM. Other protein concentrations and scale bar as indicated. t = 0 when the sample is placed at 30°C.

(a) Times series of TIRF microscopy images of the ‘solution’ nucleation assay where microtubules that nucleated in solution in the presence of different concentrations of biotinylated TPX2^mini and in the additional presence of untagged chTOG, followed after 1 min by binding to a neutravidin-coated surface. (b) Time series of TIRF microscopy images of the ‘surface’ nucleation experiments with immobilised biotinylated TPX2^mini (pre-incubated at 500 nM) in the absence (top row) or presence (bottom row) of untagged chTOG. Atto647N-labelled tubulin concentration was always 12.5 μM. Other protein concentrations and scale bars as indicated. t = 0 when the sample is placed at 30°C.

(a) Time series of TIRF microscopy images showing microtubule nucleation and growth on surfaces with immobilised biotinylated rigor kinesin (pre-incubated at 125 nM) always in the presence of 100 nM chTOG and Atto647N-labelled tubulin, without and with additional 500 nM importin α/β complex (first and second row, respectively), 100 nM mGFP-TPX2 (third row) or both 500 nM importin α/β and 100 nM mGFP-TPX2 (fourth row) as indicated. Plots showing the time course of the mean Atto647N-labelled microtubule (magenta) and mGFP-TPX2 (green) intensities for the same experiments are shown next to each individual times series of images. Note that the mGFP-TPX2 signal declines at later time points likely due to depletion of TPX2 from solution due to binding to the efficiently nucleated growing microtubules. (b) Time series of TIRF microscopy images showing that compared to a control (left), inclusion of importin α/β (500 nM) inhibits stabilisation of Atto647N-labelled nucleation intermediates by surface-immobilised biotinylated TPX2 (pre-incubated at 125 nM) (right image pair). Plot showing the time course of the mean Atto647N-labelled tubulin intensities on the surface (right). Atto647N-labelled tubulin concentration was always 12.5 μM. Scale bars as indicated. t = 0 when the sample is placed at 30°C. (c) Model of synergistic TPX2 and chTOG-stimulated microtubule nucleation and growth. Different reactions, all promoting nucleation and growth, are regulated by the distinct and complementary activities of TPX2 and chTOG.