Molecular mechanism of microtubule nucleation from gamma-tubulin ring complex

Determining how microtubules (MTs) are nucleated is essential for understanding how the cytoskeleton assembles. Yet, half a century after the discovery of MTs and αβ-tubulin subunits and decades after the identification of the γ-tubulin ring complex (γ-TuRC) as the universal MT nucleator, the underlying mechanism largely remains a mystery. Using single molecule studies, we uncover that γ-TuRC nucleates a MT more efficiently than spontaneous assembly. The laterally interacting array of γ-tubulins on γ-TuRC facilitates the lateral association of αβ-tubulins, while longitudinal affinity between γ/αβ-tubulin is surprisingly weak. During nucleation, 3-4 αβ-tubulin dimers bind stochastically to γ-TuRC on average until two of them create a lateral contact and overcome the nucleation barrier. Although γ-TuRC defines the nucleus, XMAP215 significantly increases reaction efficiency by facilitating αβ-tubulin incorporation. In sum, we elucidate how MT initiation occurs from γ-TuRC and determine how it is regulated.


Introduction 24
Microtubules (MTs) enable cell division, motility, intracellular organization and transport. MTs 25 were found to consist of αβ-tubulin dimers fifty years ago, yet, how MTs are nucleated in the cell 26 to build the cytoskeleton remains poorly understood 1-3 . 27 MTs assemble spontaneously from αβ-tubulin subunits in vitro via the cooperative assembly 28 of many tubulin dimers and hence this process displays a kinetic barrier 4-8 . Consequently, 29 spontaneous MT nucleation is rarely observed in cells 9,10 . Instead, the major MT nucleator γ-30 tubulin is required in vivo 9-11 . γ-tubulin forms a 2.2 megadalton, ring-shaped complex with γ-31 tubulin complex proteins (GCPs), known as the γ-Tubulin Ring Complex (γ-TuRC) 12-16 . γ-TuRC 32 has been proposed to template the assembly of αβ-tubulin dimers into a ring, resulting in nucleation 33 of a MT 15-21 . However, kinetic measurements that provide direct evidence for this hypothesis have 34 been lacking and several important questions about how γ-TuRC nucleates MTs have remained 35 unanswered. 36 In the absence of purified g-TuRC and an assay to visualize MT nucleation events from 37 single g-TuRC molecules in real time, recent studies used alternative MT assembly sources, such 38 as spontaneous MT assembly or stabilized MTs with blunt ends hypothesized to resemble the γ-39 TuRC interface. Based on these alternatives, competition between growth and catastrophe of the 40 nascent plus-end was proposed to yield the nucleation barrier in the cell 22,23 , but this has not been 41 examined with the nucleator g-TuRC. Recently, the MT polymerase XMAP215 was identified as 42 an essential MT nucleation factor in vivo, which synergistically nucleates MTs with γ-TuRC [24][25][26] . 43 Yet, the specific roles of XMAP215 and γ-TuRC in MT nucleation have yet to be discovered. 44 To explore the mechanism of MT nucleation, we reconstituted and visualized MT nucleation 45 by γ-TuRC live with single molecule resolution. We uncover the molecular composition of the 46 47 48 tubulin concentration (Fig. 1F), we calculated the rate of MT nucleation. The power-law 72 dependence on tubulin concentration (Fig. 1G) yields the number of ab-tubulin dimers, 3.7 ± 0.5, 73 that initiate MT assembly from g-TuRC (Fig. 1G). Thus, the cooperative assembly of 3-4 tubulin 74 subunits on g-TuRC represents the most critical, rate-limiting step in MT nucleation. 75 76

Efficiency of γ-TuRC-mediated nucleation 77
Based on the traditional, fixed, end-point assays for MT nucleation with large error margins, g-78 TuRC was believed to be a poor nucleator 14 . To measure the efficiency of γ-TuRC-mediated MT 79 nucleation, we compared it with spontaneous MT nucleation in our live TIRFM assay (Fig. 1H). 80 In contrast to γ-TuRC-mediated nucleation, a high concentration of 14 µM tubulin was required 81 for spontaneous assembly of MTs, after which both the plus-and minus-ends polymerize (Fig. 1H, 82 It has been widely proposed that the g-tubulin ring on γ-TuRC resembles the blunt plus-end of a 91 MT formed by a ring of ab-tubulins 20,22,27 . To test this proposition, we generated stabilized MT 92 seeds with blunt ends as described recently 22 and observed MT assembly from αβ-tubulin dimers 93 ( Fig. 2A). At a minimum concentration of 2.45 µM, approaching the critical concentration needed 94 for polymerization, a large proportion of pre-formed MT seeds assembled MTs immediately (Fig. 95 S2A-B, Fig. 2A and Movie S5). The measured reaction kinetics (Fig. 2B) as a function of the ab-96 tubulin concentration was used to obtain a power-law of the nucleation rate, 1.2 ± 0.4 (Fig. 2C). 97 This demonstrates that blunt MT seeds assemble tubulin dimers into a lattice in a non-cooperative 98 manner, where a single ab-tubulin dimer suffices to overcome the rate-limiting step resembling 99 the polymerization of a MT. Thus, the kinetics of γ-TuRC-mediated MT nucleation does not 100 resemble a blunt MT plus-end. 101 102

Molecular insight into microtubule nucleation by g-TuRC 103
We hypothesized that γ-tubulin's binding properties with ab-tubulin at the nucleation interface γ-104 TuRC could provide insight into the mechanism of nucleation. We purified γ-tubulin, which 105 assembles into higher order oligomers in physiological buffer 24 and strikingly, into filaments at 106 high g-tubulin concentrations (Fig. S2C). Because γ-tubulins have been shown to arrange laterally, 107 as observed previously in its crystallized form 28 , a plus-ends outward orientation of g-tubulin 108 molecules could form a nucleation interface. 109 Surprisingly, the γ-tubulin oligomers efficiently nucleated MTs from αβ-tubulin subunits 110 ( Fig. 2D and Movie S6) and even more strikingly, capped MT minus-ends while allowing newly 111 generated MT plus-ends to polymerize (Fig. 2E). This activity is similar to that of γ-TuRC, 112 suggesting that lateral γ-tubulin arrays on the nucleation interface of γ-TuRC are sufficient to 113 nucleate MTs. 114 Knowing that lateral γ-tubulin arrays in purified γ-tubulin oligomers and within g-TuRC 115 nucleate MTs, we hypothesized that the longitudinal affinity between γ-tubulin and αβ-tubulin at 116 the interface of γ-TuRC could be critical in regulating its nucleation efficiency. Using biolayer 117 interferometry, we compared the interaction of ab-tubulin dimers with themselves versus with g-118 tubulin. Specific interactions between probe-bound αβ-tubulin and increasing concentrations of 119 unlabeled αβ-tubulin were measured (Fig. 2F), which must be longitudinal based on the observed 120 protofilaments in the ab-tubulin sample by EM (Fig. S2D). In contrast, no significant binding 121 between monomeric γ-tubulin and αβ-tubulin was detected (Fig. 2F), suggesting that the 122 heterogenous longitudinal affinity between g-tubulin and ab-tubulin on the nucleation interface 123 may be weaker compared to αβ-tubulin with another ab-tubulin molecule that occurs when the 124 MT lattice polymerizes. In sum, the difference in interaction strength is the basis for the kinetic 125 barrier we observed with g-TuRC but not with a blunt MT plus-end, which we summarize with an 126 interface interaction model (Fig. 2G). 127 We next asked how 3-4 tubulin dimers formed the rate-limiting species during γ-TuRC 128 nucleation. In stochastic simulations, the 13 available binding sites on g-tubulin molecules within 129 γ-TuRC were allowed to be occupied at random with αβ-tubulin subunits. We then assessed how 130 many ab-tubulin dimers need to assemble on g-TuRC to obtain two αβ-tubulin molecules on 131 neighboring sites and form a favorable configuration with a lateral contact between the two αβ-132 tubulins (Fig. 2H). The simulations show that 3.7 ± 1 tubulin dimers assemble on γ-TuRC to form 133 the first lateral contact between two αβ-tubulins (Fig. 2H), in striking agreement with the critical 134 nucleus size we measured. In sum, our data shows that a lateral γ-tubulin array positioned by γ-135 TuRC promotes MT nucleation. The low g-tubulin:ab-tubulin affinity requires binding of 3-4 αβ-136 tubulin dimers to g-TuRC to form the first lateral contact between two αβ-tubulin dimers and 137 overcome the kinetic barrier before entering the MT polymerization phase. At low tubulin concentration of 3.5 µM and 7 µM, where either none or very little MT nucleation 157 occurs from γ-TuRCs alone respectively, the addition of XMAP215 induced many surface-158 attached γ-TuRCs to nucleate MTs resulting in significant increase in MT nucleation rate ( Fig. 3B-159 C and Movie S9). XMAP215 effectively decreases the minimal tubulin concentration necessary 160 for MT nucleation from γ-TuRC to 1.6 µM, which is very close to that needed for plus-end 161 polymerization. What is the sequence of events that leads to synergistic MT nucleation? By 162 directly visualizing γ-TuRC and XMAP215 molecules during the nucleation reaction, we found 163 that XMAP215 and γ-TuRC molecules first formed a complex from which a MT was nucleated 164 ( Fig. 3D and Movie S11). For 76% of the events (n=56), XMAP215 visibly persisted between 3 165 to over 300 seconds on γ-TuRC before MT nucleation, and with a 50% probability XMAP215 166 remained on the minus-end together with γ-TuRC (n=58). 167 Could XMAP215 accelerate nucleation by altering the critical tubulin nucleus that 168 assembles during γ-TuRC-mediated nucleation? Titrating tubulin at constant γ-TuRC and 169 XMAP215 concentrations ( Fig. S4A and Movies S10) yielded a similar power-law dependence 170 between the MT nucleation rate and tubulin concentration (Fig. 3E). The resulting critical nucleus 171 size of 3.2 ± 1.2 is very similar to that for γ-TuRC alone (Fig. 3E). Moreover, the C-terminus of 172 XMAP215 (TOG5 and C-terminal domain), which directly interacts with γ-tubulin but not with 173 αβ-tubulin 24 , does not enhance MT nucleation from γ-TuRC (Fig. S4B). Altogether, γ-TuRC 174 determines the critical nucleus of ab-tubulin dimers for MT nucleation (Fig. 2H). XMAP215, 175 which directly binds to g-tubulin via its C-terminal domain, does not appear to activate g-TuRC Here, we show that γ-TuRC-mediated MT nucleation is more efficient than spontaneous MT 184 assembly, requiring fewer tubulin dimers to form the rate-limiting reaction intermediate. This 185 explains why MTs do not form spontaneously in the cell and why γ-TuRC is essential, addressing 186 a long debate on g-TuRC's MT nucleation activity and requirement 36-38 . Spontaneous MT 187 assembly requires higher tubulin concentrations and occurs due to stronger longitudinally-188 interacting αβ/αβ-tubulin and weaker lateral interactions. In contrast, g-TuRC-mediated 189 nucleation, driven by the lateral adjacency of the g-tubulins on the nucleation interface, is sufficient 190 to overcome the intrinsically very weak ab-tubulin lateral interaction, thereby potentiating MT 191 nucleation. Thus, we propose that, in metazoans analogous to the S. cerevisiae γ-TuSC rings 15,16 , 192 GCPs within γ-TuRC restrict the number of laterally-arranged g-tubulin subunits, and position 193 them in the right geometry to template 13-pf MTs. Finally, our results show that 3-4 ab-tubulin 194 form the critical nucleus on g-TuRC, not 1 or 13 which would have been expected from previous 195 mechanistic hypotheses 20 . We find that on average 3-4 ab-tubulin dimers assemble on g-TuRC to 196 form the first lateral ab-/ab-tubulin contact and overcome the kinetic barrier that results from low 197 longitudinal affinity between g-:ab-tubulin on g-TuRC. However, alternative reaction 198 intermediates during nucleation from g-TuRC may exist. In the future, it will be important to 199 visualize the nucleation intermediates on g-TuRC, develop molecular simulations with 200 experimentally derived affinities at various interaction interfaces and evaluate whether additional 201 effects from tubulin straightening play a significant role in MT nucleation in the cell. 202 The intermediate level of MT nucleation efficiency afforded by g-TuRC allows other 203 factors to further modulate its efficiency. As such, XMAP215 accelerates MT nucleation from γ-204 TuRC, while not altering the geometry of the ab-tubulin nucleus on g-TuRC or directly activating 205 g-TuRC. Future studies will be necessary to define the modes by which XMAP215 contributes to 206 g-TuRC-mediated MT nucleation, such as increasing the probability of the g/ab-tubulin interaction 207 or promoting straightening of incoming tubulin dimers. Our findings suggest that influencing g/ab-208 tubulin interaction favorably or unfavorably may underlie a dominant mechanism for regulating 209 nucleation in the cell by other, yet unidentified nucleation factors. Additionally, g-TuRC's activity 210 is further regulated via accessory proteins such as CDK5RAP2, and NME7 2,20,39,40 . While the 211 mechanisms of these additional regulation layers are yet to be defined, the insights on MT 212 nucleation by γ-TuRC and XMAP215 provide an essential basis to build upon. Finally, this work 213 opens the door to reconstitute cellular structures in vitro using MT nucleation from γ-214 TuRC/XMAP215 to further our understanding of how the cytoskeleton is generated to support cell 215 coli Rosetta2 cells (EMD Millipore) by inducing with 0.5-1 mM IPTG for 12-18 hours at 16°C or 365 7 hours at 25°C. Wild-type XMAP215, MCAK and γ-tubulin were expressed and purified from 366 Sf9 cells using Bac-to-Bac system (Invitrogen). The cells were lysed (EmulsiFlex, Avestin) and 367 E. coli lysate was clarified by centrifugation at 13,000 rpm in Fiberlite F21-8 rotor (ThermoFisher) 368 and Sf9 cell lysate at 50,000 rpm in Ti70 rotor (Beckman Coulter) for 30-45 minutes. 369 EB1 and Stathmin were purified using His-affinity (His-Trap HP, GE Healthcare) by first 370 binding in binding buffer (20mM NaPO4 pH 8.0, 500mM NaCl, 30mM Imidazole, 2.5mM PMSF, 371 6mM BME) and eluting with 300mM Imidazole, followed by gel filtration (HiLoad 16/600 372 Superdex, GE Healthcare) into CSF-XB buffer (100mM KCl, 10mM K-HEPES, 5mM K-EGTA, 373 1mM MgCl2, 0.1mM CaCl2, pH 7.7 with 10% w/v sucrose). 374 TPX2 was first affinity purified using Ni-NTA beads in binding buffer (50mM Tris-HCl 375 pH 8.0, 750mM NaCl, 15mM Imidazole, 2.5mM PMSF, 6mM BME) and eluted with 200mM 376 Imidazole. All protein was pooled and diluted 4-fold to 200mM final NaCl. Nucleotides were 377 removed with a Heparin column (HiTrap Heparin HP, GE Healthcare) by binding protein in 378 250mM NaCl and isocratic elution in 750mM NaCl, all solutions prepared in Heparin buffer 379 (50mM Tris-HCl, pH 8.0, 2.5mM PMSF, 6mM BME). Peak fractions were pooled and loaded on 380 to Superdex 200 pg 16/600, and gel filtration was performed in CSF-XB buffer. 381 MCAK was first affinity purified by binding to His-Trap HP (GE Healthcare) in binding 382 buffer (50mM NaPO4, 500mM NaCl, 6mM BME, 0.1mM MgATP, 10mM Imidazole, 1mM 383 MgCl2, 2.5mM PMSF, 6mM BME, pH to 7.5), eluting with 300mM Imidazole, followed by gel-384 filtration (Superdex 200 10/300 GL, GE Healthcare) in storage buffer (10 mM K-HEPES pH 7.7, 385 300 mM KCl, 6mM BME, 0.1 mM MgATP, 1mM MgCl2, 10% w/v sucrose). 386 XMAP215-GFP was purified using His-affinity (His-Trap, GE Healthcare) by binding in 387 buffer (50mM NaPO4, 500mM NaCl, 20mM Imidazole, pH 8.0) and eluting in 500mM Imidazole. 388 Peak fractions were pooled and diluted 5-fold with 50mM Na-MES pH 6.6, bound to a cation-389 exchange column (Mono S 10/100 GL, GE Healthcare) with 50mM MES, 50mM NaCl, pH 6.6 390 and eluted with a salt-gradient up to 1M NaCl. Peak fractions were pooled and dialyzed into CSF-391 XB buffer. SNAP-tagged XMAP215 was first affinity purified with StrepTrap HP (GE Healthcare) 392 with binding buffer (50mM NaPO4, 270mM NaCl, 2mM MgCl2, 2.5mM PMSF, 6mM BME, pH 100mM KCl, 1mM MgCl2, 5mM K-EGTA, 10% w/v sucrose, 0.5mM TCEP, 1mM GTP, 10 µg/ml 435 LPC, pH 7.2) and fluorescent labelling was performed by incubating the beads with 1 µM Alexa-436 568 C5 Maleimide (A20341, ThermoFisher). Unreacted dye was removed with 10 ml CSF-XBg 437 buffer, beads were incubated with 25 µM NHS-PEG4-biotin (A39259, ThermoFisher) in CSF-438 XBg buffer for 1 hour at 4°C, and unreacted biotin removed with 30 ml CSF-XBg buffer. Labeled 439 g-TuRC was eluted by incubating 2-3ml of g-tubulin peptide (residues 413-451) at 0.4mg/ml in 440 CSF-XBg buffer with beads overnight. After 10-12 hours, g-TuRC was collected by adding 1-2ml 441 CSF-XBg buffer to the column, concentrated to 200 µl in 30k NMWL Amicon concentrator (EMD 442 Millipore) and layered onto a continuous 10-50 w/w % sucrose gradient prepared in a 2.2 ml ultra-443 clear tube (11x34 mm, Beckman Coulter) using a two-step program in Gradient Master 108 444 machine. Sucrose gradient fractionation of g-TuRC was performed by centrifugation at 200,000xg 445 in TLS55 rotor (Beckman Coulter) for 3 hours. The gradient was fractionated from the top in 11-446 12 fractions using wide-bore pipette tips and peak 2-3 fractions were identified by immunoblotting 447 against g-tubulin with GTU-88 antibody (Sigma). g-TuRC was concentrated to 80 µl in 30k 448 NMWL Amicon concentrator (EMD Millipore) and fresh purification was used immediately for 449 single molecule assays. Cryo-preservation of g-TuRC molecules resulted in loss of ring assembly 450 and activity. 451 452

Assessment of γ-TuRC with protein gel, immunoblot and negative stain electron microscopy 453
To assess the purity of g-TuRC, 3-5 µl of purified g-TuRC was visualized on an SDS-PAGE with 454 SYPRO Ruby stain (ThermoFisher) following the manufacturer's protocol. Biotinylated subunits 455 of g-TuRC were assessed by immunoblotting with Streptavidin-conjugated alkaline phosphatase 456 (S921, ThermoFisher). g-TuRC purification was also assessed by visualizing using electron 457 microscopy. 4 µl of peak sucrose gradient fraction of g-TuRC was pipetted onto CF400-Cu grids 458 (Electron Microscopy Sciences), incubated at room temperature for 60 seconds and then wicked 459 away. 2% uranyl acetate was applied to the grids for 30 seconds, wicked away, and the grids were 460 air-dried for 10 minutes. The grids were imaged using Phillips CM100 TEM microscope at 64000x 461 magnification. During the incubations, nucleation mix was prepared containing desired concentration of 500 αβ-tubulin (3.5-21 µM) purified from bovine brain with 5% Cy5-labeled tubulin along with 501 1mg/ml BSA (A7906, Sigma) in assay buffer, centrifuged for 12 minutes in TLA100 (Beckman 502 Coulter) to remove aggregates, a final 0.68 mg/ml glucose oxidase (SERVA, catalog # SE22778), 503 0.16 mg/ml catalase (Sigma, catalog # SRE0041) was added, and reaction mixture was introduced 504 into the flow chamber containing g-TuRC. representative set of curves is displayed in Fig. 1F. A straight line was fit to the initial (linear) 529 region of each N(t) versus t curve, rate of nucleation was obtained slope of each linear fit, and its 530 power-law relation with tubulin concentration was obtained and reported (Fig. 1G). 531

Spontaneous microtubule nucleation and data analysis 533
Spontaneous MT assembly was visualized similar to g-TuRC-mediated nucleation with the 534 following changes. The pluronic, casein and NeutrAvidin incubations were performed identical to 535 g-TuRC nucleation assay but instead of attaching g-TuRCs, sucrose-based buffer (of the same 536 composition as used for g-TuRC elution) was diluted 5-fold with BRB80, introduced in the flow 537 chamber and incubated for 10 minutes. Washes were performed with 1 vol of BRB80, nucleation 538 mix was added, and imaging was performed as described above. MTs nucleate spontaneously in 539 solution fall down on the coverslip due to depletion forces during the 10 minutes of visualizing the 540 reaction. The number of MTs nucleated in the field of view were counted manually and plotted in 541 Blunt MTs were prepared with GMPCPP nucleotide in two polymerization cycles as described 545 recently 22 . Briefly, a 50 µl reaction mixture was prepared with 20 µM bovine brain tubulin with 546 5% Alexa-568 labeled tubulin and 5% biotin-labeled tubulin, 1mM GMPCPP (Jena Bioscience) 547 in BRB80 buffer, incubated on ice for 5 minutes, then incubated on 37°C for 30 minutes to 548 polymerize MTs, and MTs were pelleted by centrifugation at 126,000 xg for 8 minutes at 30°C in 549 TLA100 (Beckman Coulter). Supernatant was discarded, MTs were resuspended in 80% original 550 volume of BRB80, incubated on ice for 20 minutes to depolymerize MTs, fresh GMPCPP was 551 added to final 1mM, incubated on ice for 5 minutes, a second cycle of polymerization was 552 performed by incubating the mixture at 37°C for 30 minutes, and MTs were pelleted again by 553 centrifugation. Supernatant was discarded and MTs were resuspended in 200 µl warm BRB80, 554 flash frozen in liquid nitrogen in 5µl aliquots, stored at -80°C and found to be stable for months. 555 To verify that these MT seeds have blunt ends, frozen aliquots were quickly thawed at 37°C, 556 diluted 20-fold with warm BRB80, and incubated at room temperature for 30 minutes to ensure 557 blunt ends as described previously 22 . MTs were pipetted onto CF400-Cu grids (Electron 558 Microscopy Sciences), incubated at room temperature for 60 seconds and then wicked away. 2% 559 uranyl acetate was applied to the grids for 30 seconds, wicked away, and the grids were air-dried 560 for 10 minutes. The grids were imaged using Phillips CM100 TEM microscope at 130000 x 561 magnification and most MT ends were found to be blunt. 562 To assay MT assembly from blunt MT seeds, MT assembly experiments similar to g-TuRC 563 nucleation assays were performed with the following variation. A lower concentration 0.05 mg/ml 564 NeutrAvidin (A2666, ThermoFisher) was attached, and washes were performed with warm 565 BRB80 prior to attaching MTs. One aliquot of MT seeds was thawed quickly, diluted to 100-fold 566 with warm BRB80, incubated in the chamber for 5 minutes, unattached seeds were washed with 1 567 vol of warm BRB80, and the slide was incubated at room temperature for 30 minutes to ensure 568 blunt MT ends. Wide bore pipette tips were used for handling MT seeds to minimize the shear 569 forces that may result in breakage of MTs. Nucleation mix was prepared as described above and a 570 low αβ-tubulin concentration (1.4-8.7 µM) was used. MT assembly from blunt seeds was observed 571 immediately after incubating the slide on the objective heater. Imaging and analysis were 572 performed as described above for to g-TuRC nucleation assays. However, the probability curves 573 for MT assembly were obtained (Fig. 2B) by normalizing for the total number of seeds observed 574 in the field of view. Rate of assembly was plotted against [tubulin concentration -C*], where C* 575 represents the critical tubulin concentration below which MT ends do not polymerize obtained 576 directly from experimental measurements (Fig. S2A-B). 577

Electron microscopy of γ-tubulin filaments in vitro 579
Purified γ-tubulin was observed to form higher order oligomers previously using analytical gel 580 filtration 24 . g-tubulin filaments were prepared by diluting pure g-tubulin to 1-5 µM to the buffer 581 50mM K-MES pH 6.6, 5mM MgCl2, 1mM EGTA, 100mM KCl. g-tubulin mixture were pipetted 582 onto CF400-Cu grids (Electron Microscopy Sciences), incubated at room temperature for 60 583 seconds and then wicked away. 2% uranyl acetate was applied to the grids for 30 seconds, wicked 584 away, and the grids were air-dried for 10 minutes. The grids were imaged using Phillips CM100 585 TEM microscope at 130000 x magnification and g-tubulin filaments were seen to form. At 500 586 mM KCl, g-tubulin filaments were not seen. 587 588 Nucleation of microtubules from purified γ-tubulin 589 MT assembly experiments from purified g-tubulin was performed similar to g-TuRC nucleation 590 assays described above with following variation. No avidin was attached to the coverslips, and 591 varying concentration of g-tubulin was prepared by diluting purified g-tubulin in a high salt buffer 592 (50mM K-MES pH 6.6, 500mM KCl, 5mM MgCl2, 1mM EGTA), centrifuging to remove 593 aggregates separately for 12 minutes in TLA100 before adding to the nucleation mix containing 594 15 µM αβ-tubulin (5% Cy5-labeled) with BSA, glucose oxidase and catalase as described above. 595 The reaction mixture was introduced into the flow chamber and imaged via TIRF microscopy. A 596 large number of MTs get nucleated immediately in the presence of 250 nM-1000 nM g-tubulin. 597 598

Measurement of affinity between purified γ-tubulin and αβ-tubulin 599
Interaction assays between αβ-tubulin and g-tubulin were performed with biolayer interferometry 600 using Octet RED96e (ForteBio) instrument in an 8-channel plate format. The plate temperature 601 was held at 33°C and the protein samples were shaken at 400 rpm during the experiment. First, 602 Streptavidin or anti-His antibody coated biosensors (ForteBio) were rinsed in interaction buffer 603 (50mM K-MES pH 6.6, 100mM KCl, 5mM MgCl2, 1mM EGTA, 0.05% Tween20, 1mM GTP). 604 100 nM biotin-labeled αβ-tubulin, or blank buffer, was bound to Streptavidin sensor, or 200 nM 605 His-tagged g-tubulin to anti-His sensor until loaded protein results in a wavelength shift (Δλ) of 3 606 nm. Unbound protein was removed by rinsing the sensor in interaction buffer, and interaction with 607 αβ-tubulin was measured by incubating the sensor containing αβ-tubulin, g-tubulin or buffer with 608 0-35 µM unlabeled αβ-tubulin in interaction buffer for 5 minutes. Δλ (nm) was recorded as a 609 measure of the amount of unlabeled αβ-tubulin that binds to the sensor. Longitudinal interaction 610 occurs between αβ-tubulin dimers and the resulting protofilaments were verified by visualizing 611 the αβ-tubulin sample stained with 2% uranyl acetate using electron microscopy as described 612 above (Fig. S2D). 613 614

Simulation of site occupation on γ-TuRC by αβ-tubulin dimers 615
A simulation was performed in MATLAB for occupation of sites on g-TuRC by αβ-tubulin dimers. 616 A circular grid was simulated with 13 empty positions that were occupied one per unit time 617 stochastically such that a new position was selected by uniform random number generator and 618 filled. If a previously filled position was selected, a different position was selected by the random 619 number generator. The sequence in which the sites were occupied was followed. For each 620 simulation, the total number of sites that were occupied when the first two neighboring sites are 621 filled was recorded. The simulation was repeated 10,000 times and the probability of occurrence 622 of first neighbor contact versus number of sites occupied is displayed in Fig. 2H. 623

624
Measuring the effect of microtubule associated proteins on γ-TuRC's activity 625 Effect of microtubule associated proteins (MAPs) was measured on g-TuRC's nucleation activity. 626 g-TuRC was attached on the coverslips using the setup described above and a control experiment 627 was performed with identical reaction conditions for each protein tested. Nucleation mix was 628 prepared containing 10.5 µM αβ-tubulin concentration (5% Cy5-labeled tubulin) as specified along 629 with 1mg/ml BSA and oxygen scavengers, and either buffer (control), 10nM GFP-TPX2, 100nM 630 EB1-mCherry, 5 µM Stathmin or 10nM MCAK was added. To test MCAK's effect, the assay 631 buffer additionally contained 1mM ATP. The reaction mixture containing tubulin and MAP at 632 specified concentration was introduced into the flow chamber containing g-TuRC, and MT 633 nucleation was visualized by imaging the Cy5-fluorescent channel at 0.5-1 frames per second. For 634 TPX2 and EB1, fluorescence intensity of the protein was simultaneously acquired. The number of 635 MTs nucleated over time was measured as described above and the effect of protein on g-TuRC's 636 nucleation activity was assessed by comparing nucleation curves with and without the MAP. 637 A similar set of experiments were performed to study the effect of XMAP215 on g-TuRC-638 mediated nucleation with the single molecule assays with the following differences. 20 nM of 639 XMAP215-GFP was added to nucleation mix prepared with 3.5-7 µM αβ-tubulin concentration 640 (5% Cy5-label) in XMAP assay buffer (80mM K-PIPES, 1mM MgCl2, 1mM EGTA, 30mM KCl, 641 0.075% w/v methylcellulose 4000 cp, 1% w/v D-(+)-glucose, 0.007% w/v Brij-35, 5mM BME, 642 1mM GTP). MTs nucleated from attached g-TuRC with and without XMAP215 were measured to 643 assess the efficiency of nucleation induced by XMAP215 (Fig. 3C). To assess if C-terminal of 644 XMAP215 increases nucleation efficiency, wild-type XMAP215 was replaced with a C-terminal 645 construct of XMAP215: TOG5-Cterminus-GFP in the described experiment. 646 To measure the kinetics of cooperative nucleation XMAP215 and g-TuRC, a constant 647 density of g-TuRC was attached as described above and nucleation mix nucleation mix was 648 prepared with a range of αβ-tubulin concentration between 1.6-7 µM (5% Cy5-label) with 20 nM 649 of XMAP215-GFP in XMAP assay buffer, introduced into reaction chamber and MT nucleation 650 was imaged immediately by capturing dual color images of XMAP215 and tubulin intensity at 0.5 651 frames per second. 652 653