MCAK preferentially binds to growing microtubule ends

(A) Representative kymographs of single-molecule GFP-MCAK (green, 10 nM, measured in terms of monomer concentration unless otherwise stated) binding events on dynamic microtubules (red, tubulin: 16 mM) growing from GMPCPP microtubule seeds (blue) in the presence of 1 mM ATP. A plus-end binding event and a lattice binding event were marked by the white and blue arrowheads, respectively. Scale bars: vertical, 5 s; horizontal, 2 μm.

(B) Representative intensity profiles of a single GFP-MCAK molecule (green, 10 nM) and a growing microtubule end (red, tubulin: 16 mM) at one frame during the dwell time of a binding event (corresponding to the white arrowhead in panel A). The peak position of a GFP-MCAK molecule was determined by Gaussian fitting (green dashed line).

(C) Position probability distributions of GFP-MCAK binding sites (green, n=152 events) along the microtubule longitudinal axis. The red curve represents the fitting curve of the growing microtubule end as the positional reference. The MCAK binding region was defined as the area within the FWHM (double black arrow, 162 nm) of the fitted green curve.

(D) Statistical quantification of on-rate (kon) of GFP-MCAK’s binding to the plus-end versus lattice of dynamic microtubules in the presence of 1 mM ATP (n=66 microtubules from 3 assays).

(E) Statistical quantification of RE/L in the presence of 1 mM ATP (n=66 microtubules from 3 assays).

(F) Statistical quantification of the dwell time of GFP-MCAK’s binding to the growing end versus lattice of dynamic microtubules in the presence of 1 mM ATP (n=966 events from 3 assays for the plus end; n=702 events from 3 assays for lattice).

(G) Mean-squared displacement (MSD) of GFP-MCAK plotted against the time interval (t, 0.1s per frame). The diffusion coefficient (D) was calculated via linear regression of the equation (<x2>= 2Dt), yielding 0. 023 μm2 s−1 (error bars represent SEM, n=203 trajectories).

In panel D, E and F, all the data were presented as mean ± SEM. All comparisons were performed using two-tailed Mann-Whitney U test with Bonferroni correction, n.s., no significance; ***, p<0.001.

MCAK binds to the entire GTP cap of growing microtubule ends

(A) Representative kymographs showing the individual binding events of GFP-MCAK (1 nM, with 1 mM ATP, left), EB1-GFP (10 nM, middle) or XMAP215-GFP (1 nM, right) at growing microtubule ends (red, tubulin: 12 μM). The plots showed the representative intensity profiles of single molecules (green) and growing microtubule ends (red) at one frame during the dwell time of the individual binding event (white arrowheads). Peak positions were determined by Gaussian fitting (green dashed line). Scale bars: vertical, 5 s; horizontal, 2 μm.

(B) Position probability distributions of the binding sites of GFP-MCAK, EB1-GFP (black, n=184 events) and XMAP215-GFP (blue, 142 events) along the microtubule longitudinal axis. The fitting curve of the growing microtubule end (red) was used as the positional reference. The binding region was defined as the area within the FWHM of the fitted curves. The GFP-MCAK molecules localized to the left of XMAP215-GFP’s binding region was defined as proximally distributed (grey bar), and those localized to the right of EB1-GFP’s binding region was defined as distally distributed (orange bar).

(C) Averaged intensity profiles of the growing microtubule ends decorated with GFP-MCAK (green, 123 microtubules), EB1-GFP (black, 184 microtubules) and XMAP215-GFP (blue, 142 microtubules). Data were presented as mean ± std, and no significant difference was observed between the three conditions (Kolmogorov-Smirnov test, EndEB1 vs EndMCAK, p=0.68; EndMCAK vs EndXMAP215, p=0.68; EndEB1 vs EndXMAP215, p=1). In panels B and C, the red arrows indicated microtubule growth direction.

(D) Representative kymographs showing concurrent binding events of XMAP215-GFP (25 nM) and MCAK-RFP (20 nM) at growing microtubule ends (tubulin: 12 μM) in the presence of 1 mM ATP. The plots showed the representative intensity profiles of XMAP215-GFP (blue), MCAK-RFP (green) and growing microtubule ends (red) at one frame during the dwell time of the binding event (cyan arrowhead). Peak positions were determined by Gaussian fitting (dashed line). The spacing between the two peaks represents the instantaneous distance between MCAK and XMAP215 (double black arrow). Scale bars: vertical, 5 s; horizontal, 2 μm.

(E) Probability distribution of the distance between XMAP215-GFP and MCAK-RFP (62 events from 3 assays) at growing microtubule ends. In this plot, the position of XMAP215-GFP was taken as the origin. A negative distance indicated a more proximal localization.

The growing end-binding preference of MCAK is dependent on its nucleotide state

(A) Representative kymographs of single-molecule GFP-MCAK (green) binding events on dynamic microtubules (red, tubulin: 16 μM) growing from the GMPCPP microtubule seeds (blue) in the presence of 1 mM AMPPNP (left), 1 mM ADP (middle) and APO (right) (the nucleotide-free state). The plus-end binding events were indicated by white arrowheads. Scale bar: vertical, 5 s; horizontal, 2 μm.

(B) Statistical quantification of on-rate (kon) of GFP-MCAK binding to the plus end (E) versus the lattice (L) of dynamic microtubules in the presence of 1 mM AMPPNP (n=29 microtubules from 2 assays), 1 mM ADP (n=21 microtubules from 2 assays) and APO (n=36 microtubules from 3 assays). The cyan and purple dashed lines denoted the on-rate value of MCAK binding at the plus end and lattice in the ATP condition (from Fig. 1), respectively.

(C) Statistical quantification of RE/L in the presence of 1 mM AMPPNP (n=29 microtubules from 2 assays), 1 mM ADP (n=21 microtubules from 2 assays) and APO (n=36 microtubules from 3 assays). Dashed line: the mean value of RE/L in the ATP condition from Fig. 1.

(D) Statistical quantification of the dwell time for GFP-MCAK binding at the growing ends and lattice of dynamic microtubules in the presence of 1 mM AMPPNP (n=231 events from 2 assays for the plus end; n=142 events from 2 assays for lattice), 1 mM ADP (n=327 events from 2 assays for the plus end; n=1184 events from 2 assays for lattice) and APO (n=71 events from 3 assays for the plus end; n=573 events from 3 assays for lattice). The cyan and purple dashed lines denoted the dwell time value of MCAK binding at the plus end and lattice in the ATP condition (from Fig. 1), respectively.

In the panel B, C and D, all the data were presented as mean ± SEM. All the comparisons were made against the corresponding value in the ATP condition from Fig. 1 using two-tailed Mann-Whitney U test with Bonferroni correction, n.s., no significance; ***, p<0.001.

MCAK strongly binds to GTPγS microtubules in a nucleotide-dependent manner

(A) Representative projection images of GFP-MCAK binding to GTPγS microtubules (red) versus GDP microtubules (cyan) in the presence of 1 mM ATP (left, 10 nM GFP-MCAK), 1 mM AMPPNP (left middle, 1nM GFP-MCAK), 1 mM ADP (right middle, 10 nM GFP-MCAK) and at the APO state (right, 0.1 nM GFP-MCAK). Note that the binding on the GDP microtubules was compared to that on the GTPγS microtubules in the same flow cell. The GFP-MCAK binding to the GTPγS and GDP microtubules was indicated by the red and cyan arrowhead, respectively. Scale bar: 5 μm.

(B) Statistical quantification of the normalized fluorescence intensity of GFP-MCAK (1 nM) on different microtubules in the presence of 1 mM ATP (95 GTPγS microtubules, 28 GDP microtubules from 2 assays), 1 mM AMPPNP (133 GTPγS microtubules, 56 GDP microtubules from 3 assays), 1 mM ADP (111 GTPγS microtubules, 44 GDP microtubules from 2 assays) or at the APO state (52 GTPγS microtubules, 36 GDP microtubules from 3 assays). All data were normalized to the binding to GTPγS microtubules binding in the AMPPNP condition.

(C) The binding intensity ratios of GFP-MCAK at GTPγS microtubules to GDP microtubules under various nucleotide conditions. ATP: n=28, from 2 assays. AMPPNP: n=56 GTPγS/GDP from 3 assays. ADP: n=44 from 2 assays. APO: n=36 from 3 assays in APO state. The dashed line represented 1. Statistical comparisons were performed between the ratios and 1.

In panels B and C, all the data were presented as mean ±SEM. All comparisons were performed using two-tailed Mann–Whitney U test with Bonferroni correction, ***, p<0.001.

MCAK binds to the ends of GMPCPP microtubules in a nucleotide state-dependent manner

(A, B, C, D) Representative fluorescence projection images and intensity profiles showing GFP-MCAK binding to the ends and lattice of GMPCPP microtubules in the presence of 1 mM ATP (10 nM GFP-MCAK), 1 mM AMPPNP (1 nM GFP-MCAK), 1 mM ADP (10 nM GFP-MCAK) or at the APO state (0.1nM GFP-MCAK). Red arrowhead: the signal of MCAK enriching at the ends. Intensity profiles (right panels) showed spatial intensity signal distributions of the microtubule (black curve) and GFP-MCAK (green curve). Red and black bars indicated the end and the lattice, respectively. Scale bar=2 μm.

(E) Statistical quantification of the normalized binding intensity of GFP-MCAK (1 nM) at the end versus lattice regions of GMPCPP microtubules in the presence of 1 mM ATP (40 microtubules from 3 assays), 1 mM AMPPNP (47 microtubules from 3 assays), 1 mM ADP (45 microtubules from 3 assays) or at the APO state (52 microtubules from 3 assays). All the data were normalized to the binding intensity of GFP-MCAK at GMPCPP microtubule lattices in the AMPPNP condition. All the data were presented as mean ± SEM. Statistical analysis was performed using two-tailed paired t-test by Bonferroni correction, ***, p<0.001.

(F) The end-to-lattice binding intensity ratios of GFP-MCAK at GMPCPP microtubules under various nucleotide conditions. The ratios were presented as mean ± SEM. The dashed line represents 1. Statistical comparisons were performed between the ratios and 1 using two-tailed Mann-Whitney U test with Bonferroni correction, ***, p<0.001.

The α4 helix and the L2 loop of MCAK contribute to the preferential binding to the EB cap

(A) Structural model of the MCAKsN+M-tubulin complex (PDB ID: 5MIO; X-ray diffraction). Top: Domain organization of MCAK. Bottom: Enlarged view of α4 helix and L2 loop, regions mediating the interaction between MCAK and tubulin, with the key sites (K524 and V298) were highlighted in red. The residues in tubulin that potentially interact with K524 and V298 were highlighted in green.

(B) Representative kymographs of single-molecule binding events of GFP-MCAKK524A (10 nM) and GFP-MCAKV298S (30 nM) at growing microtubules (red, tubulin: 16 μM) in the presence of 1 mM ATP. The end-binding events were indicated by red arrowheads. Scale bars: vertical, 5 s; horizontal, 2 μm.

(C) Statistical quantification of on-rate (kon-P) of GFP-MCAKK524A (31 microtubules from 3 assays) and GFP-MCAKV298S (23 microtubules from 2 assays) at growing microtubule ends in the presence of 1 mM ATP. The data were presented as mean ± SEM and were compared to that of wild-type GFP-MCAK (from Fig. 1).

(D) Statistical quantification of dwell time of GFP-MCAKK524A (18 binding events from 3 assays) and GFP-MCAKV298S (39 binding events from 3 assays) at growing microtubule ends in the presence of 1 mM ATP. The data were presented as mean ± SEM and were compared to that of wild-type GFP-MCAK in the presence of ATP (from Fig. 1). The dashed line: the mean dwell time of wild-type GFP-MCAK.

(E) Representative fluorescence projection images of 5 nM GFP-MCAKK524A and 10 nM GFP-MCAKV298S binding to GTPgS (red arrowhead) versus GDP (cyan arrowhead) microtubules in the presence of 1 mM AMPPNP. Scale bar: 5 mm.

(F) Statistical quantification of normalized binding intensity of 1 nM GFP-MCAKK524A (144 GTPgS microtubules and 62 GDP microtubules from 3 assays) and GFP-MCAKV298S (124 GTPgS microtubules, and 59 GDP microtubules from 3 assays) on different microtubules. All the data were normalized to the binding intensity of wild-type GFP-MCAK at GTPgS microtubules in the AMPPNP condition (from Fig. 4). Data were presented as mean ± SEM.

(G) The binding intensity ratios of GFP-MCAKK524A or GFP-MCAKV298S on GTPγS microtubules to that on GDP microtubules in the presence of 1 mM AMPPNP. GFP-MCAKK524A: n=62 from 3 assays in AMPPNP state. GFP-MCAKV298S: n=59 GTPγS/GDP binding ratios from 3 assays in AMPPNP state. Dashed line: the GTPγS/GDP ratio of wild-type GFP-MCAK. The data were presented as mean ± SEM and were compared to that for wild-type GFP-MCAK (from Fig. 4).

In panel C, D, F and G, all comparisons were performed using two-tailed Mann-Whitney U test with Bonferroni correction, n.s., no significance; *, p<0.05; **, p<0.01; ***, p<0.001.

Functional specification of MCAK and XMAP215 at growing microtubule ends

(A) Representative kymographs of dynamic microtubules (tubulin: 10 μM) under four conditions: control (no MAPs), 20 nM MCAK (1 mM ATP), 50 nM XMAP215 or both (1 mM ATP). Scale bars: vertical, 100 s; horizontal, 2 μm.

(B) Probability distribution of the microtubule lifetime under the conditions indicated in the panel A. Lines represent gamma function fits. Control: n=443 microtubules from 7 assays. 20 nM MCAK: n=621 microtubules from 7 assays. 50 nM XMAP215: n=237 microtubules from 3 assays. 20 nM MCAK+50 nM XMAP215: n=283 microtubules from 3 assays.

(C) The probability-based catastrophe frequency versus microtubule lifetime showing how the likelihood of catastrophe depended on the age of microtubules. Data derived from panel A.

Cartoon schematics depicting the working model of MCAK at growing microtubule ends

(A) A hypothetical model for the binding cycle of MCAK, given its binding preference on the EB cap. Arrows indicate the transitions in nucleotide-dependent MCAK functional states. T: MCAK‧ATP. D: MCAK‧ADP.

(B) Cooperation schematics of MCAK, EB1 and XMAP215 at growing microtubule ends. Pathway 1: the direct binding of MCAK to microtubule ends. Pathway 2: the indirect binding of MCAK to microtubule ends via EB1.

Biochemical and functional characterization of purified MCAK, its variants, EB1 and XMAP215

(A) Domain architecture schematics of GFP-MCAK and its mutants, MCAK-RFP, EB1-GFP, and XMAP215-GFP used in this study.

(B) The western blot result showing the removal of His-tag in purified GFP-MCAK. In the figure, the molecular weight marker lane and the protein sample lane were merged. The molecular weight of GFP-MCAK monomer is 110 kDa. Anti-his antibody was used for detection.

(C-I) Size-exclusion chromatography (Superose™ 6 Increase 5/150 GL) profiles of GFP-MCAK, GFP-MCAKK524A, GFP-MCAKV298S, MCAK-RFP, GFP-MCAKsN+M, EB1-GFP and XMAP215-GFP-his6. The molecular weight of GFP-MCAK, GFP-MCAKK524A, GFP-MCAKV298S and MCAK-RFP monomer were 110 kDa. The molecular weight of GFP-MCAKsN+M was 73 kDa. The molecular weight of EB1-GFP monomer was 60 kDa. The molecular weight of XMAP215-GFP-his6 was 260 kDa. The elution volumes of the peak for GFP-MCAK, GFP-MCAKK524A, GFP-MCAKV298S, MCAK-RFP, GFP-MCAKsN+M, EB1-GFP and XMAP215-GFP-his6 were 1.85 ml, 1.84 ml, 1.91 ml, 1.90 ml, 2.25 ml, 2.11 ml and 1.78 ml, respectively. In panel C, the elution volume of the peak corresponded to the molecular weight range between 120 kDa (EB1-GFP dimer) and 260 kDa (XMAP215-GFP-his6), suggesting that GFP-MCAK exists as a dimer under this experimental condition.

(J) SDS-PAGE gels of elution fractions corresponding to peaks from size-exclusion chromatography profiles of GFP-MCAK, GFP-MCAKK524A, GFP-MCAKV298S, MCAK-RFP, GFP-MCAKsN+M, EB1-GFP and XMAP215-GFP-his6 in panels C-I.

(K) Fluorescence intensity distributions of single GFP-MCAKsN+M (blue, n=327 binding events from 3 assays) and GFP-MCAK (red, n=509 binding events from 3 assays) molecules were fitted by Gaussian fitting. The background intensity was subtracted. The concentrations of the two proteins were both 10 nM. Dynamic microtubules (tubulin: 16 μM) were polymerized in the presence of 1 mM ATP. The intensity range-1 (blue, μ±2σ=196±79 A.U.) and range-2 (red, μ±2σ=300±145 A.U.) were used to determine if a fluorescence spot represents a single monomer and dimer, respectively.

(L) The apparent off-rate (koff) of GFP-MCAK at growing microtubule ends in the presence of ATP (circle), AMPPNP (down triangle), ADP (up triangle) and APO (diamond). Data of dwell time were from Fig. 1F. koff was calculated by fitting the dwell time of individual GFP-MCAK binding events to a single exponential function.

(M) Statistical quantification of the growth rates of dynamic microtubules with GFP-MCAK in the presence of 1 mM ATP (57 microtubules from 3 assays), AMPPNP (38 microtubules from 2 assays), ADP (24 microtubules from 2 assays) and APO (34 microtubules from 3 assays). Data were from the experiments shown in panel Fig. 1A, which were presented as mean ± SEM. Statistical analysis was performed using two-tailed Mann-Whitney U test with Bonferroni correction, n.s., no significance, ***, p<0.001.

Single-molecule localization analysis at growing microtubule ends

(A) Representative kymograph showing a binding event of GFP-MCAK (1 nM, with 1 mM ATP, left) at the growing microtubule end (red, tubulin: 12 μM). Schematic illustration of MCAK binding to the microtubule end (right). The red cross represented the reference location point. The green dots represented the position of MCAK at every frame during its dwell time. The orange cross indicated the average position of MCAK during the dwell time, which represented the location of MCAK over this period. Scale bars: vertical, 2 s; horizontal, 2 μm.

(B) Snapshot of a single GFP-MCAK molecule (green), and the microtubule (red) at the timepoint marked by a black arrow in panel A (left). Intensity profiles of the microtubule (red) and the single molecule (green) at this moment. The peak position of the single molecule was determined using Gaussian fitting (green dashed line), which represented the location of the molecule at this moment. Scale bar: 2 μm.

Projection of single-molecule fluorescence images

Representative images showing how to generate a summation image using 1000 frames of raw images. The binding intensity of GFP-MCAK at GDP versus GTPγS microtubule was indicated using green and yellow arrowheads, respectively. The concentration of GFP-MCAK here was 1 nM. The experiment was performed in the presence of 1 mM AMPPNP. Scale bar: 5 μm.

EB1-GFP strongly binds to GTPγS microtubules

(A) Representative projection images of EB1-GFP binding to GTPγS microtubules. The binding at GDP microtubules was compared to that at GTPγS microtubules in the same flow cell. In GFP channel, the fluorescence intensity of EB1-GFP binding to GTPγS and GDP microtubules were indicated by red and cyan arrowhead, respectively. Scale bar: 5 μm.

(B) The binding intensity ratios of EB1-GFP on GTPγS microtubules to that on GDP microtubules under various nucleotide conditions. 36 GTPγS/GDP binding ratios from 3 assays. Dashed line on the plot represented 1. All the data were presented as mean ± SEM. Statistical comparison was performed between the ratios and 1. The comparison was performed using two-tailed Mann–Whitney U test with Bonferroni correction, ***, p<0.001.

MCAK preferentially binds to the ends of taxol stabilized GDP microtubules

(A) Representative projection images and intensity profiles showing GFP-MCAK (10 mM) binding at the ends of taxol stabilized GDP microtubules in the presence of 1 mM ATP. Red arrowhead: the end-binding of MCAK events. Intensity profiles (right panels) showed spatial intensity signal distribution of microtubule (black curve) and GFP-MCAK (green curve) along microtubule longitudinal axes from left to right. Red and black bars above the intensity profiles indicated end and lattice regions, respectively. Scale bar=2 μm.

(B) Statistical quantification of the binding intensity of GFP-MCAK at the end and lattice of taxol stabilized GDP microtubules in the presence of 1 mM ATP (35 microtubules from 2 assays). Data were scaled relative to the binding intensity of GFP-MCAK at the lattice of taxol stabilized GDP microtubules. Data were presented as mean ± SEM. Statistical analysis was performed using two-tailed pair t-test by Bonferroni correction, ***, p<0.001.

(C) The ration of the binding intensity of GFP-MCAK at the end to that on the lattice of GDP microtubules in ATP state. The data were presented as mean ± SEM. Dashed line on the plot represents 1. Statistical comparisons were performed between the ratios and 1 using two-tailed Mann-Whitney U test with Bonferroni correction, ***, p<0.001.

GFP-MCAKK525A and GFP-MCAKV298S are depolymerizing-deficient and ATPase activity-reduced mutants

(A) Representative kymographs showing depolymerization of GMPCPP-stabilized microtubules by GFP-MCAK, GFP-MCAKK525A and GFP-MCAKV298S in the presence of 1 mM ATP. Scale bars: vertical, 100 s; horizontal, 2 μm.

(B) Statistical quantification of microtubule depolymerization rates of GFP-MCAK (n=21 microtubules from 3 assays), GFP-MCAKK525A (n=11 microtubules from 2 assays) and GFP-MCAKV298S (n=52 microtubules from 3 assays). Data were presented as mean ±SEM. Statistical analysis was performed using two-tailed Mann-Whitney U test with Bonferroni correction, ***, p<0.001.

(C) Microtubule-stimulated ATPase activities of wild-type GFP-MCAK (6 assays), GFP-MCAKK525A (4 assays) and GFP-MCAKV298S (4 assays). Data were normalized by the microtubule-stimulated ATPase activity of the wild-type. Statistical analysis was performed using two-tailed Student’s t test, ***, p<0.001.

A simple model for the end-binding of MCAK and EB1.

MCAK can bind to microtubule growing ends through both the direct (left) and EB-dependent (right) pathways. The dissociation constants were K0, K1, K2 and K3, respectively. MTE: growing microtubule end.

The binding kinetics of single-molecule EB1-GFP binding to growing microtubule end

(A) Statistical quantification of on-rate (kon) of EB1-GFP’s binding to the plus end of dynamic microtubules (data calculated from Fig. 2, n=73 microtubules from 3 assays).

(B) The apparent off-rate (koff) of EB1-GFP at growing microtubule ends (data calculated from Fig. 2, n=153 bonding events from 3 assays). koff was calculated by fitting the dwell time of individual GFP-MCAK binding events to a single exponential function.