CLASP2 recognizes tubulins exposed at the microtubule plus-end in a nucleotide state-sensitive manner

Interface-acting nucleotide controls polymerization dynamics at microtubule plus- and minus-ends

GTP-tubulin is preferentially incorporated at growing microtubule ends, but the biochemical mechanism by which the bound nucleotide regulates the strength of tubulin:tubulin interactions is debated. The 'self-acting' (cis) model posits that the nucleotide (GTP or GDP) bound to a particular tubulin dictates how strongly that tubulin interacts, whereas the 'interface-acting' (trans) model posits that the nucleotide at the interface of two tubulin dimers is the determinant. We identified a testable difference between these mechanisms using mixed nucleotide simulations of microtubule elongation: with a self-acting nucleotide, plus- and minus-end growth rates decreased in the same proportion to the amount of GDP-tubulin, whereas with interface-acting nucleotide, plus-end growth rates decreased disproportionately. We then experimentally measured plus- and minus-end elongation rates in mixed nucleotides and observed a disproportionate effect of GDP-tubulin on plus-end growth rates. Simulations of microtubule growth were consistent with GDP-tubulin binding at and 'poisoning' plus-ends but not at minus-ends. Quantitative agreement between simulations and experiments required nucleotide exchange at terminal plus-end subunits to mitigate the poisoning effect of GDP-tubulin there. Our results indicate that the interfacial nucleotide determines tubulin:tubulin interaction strength, thereby settling a longstanding debate over the effect of nucleotide state on microtubule dynamics.

Higher-order protein assembly controls kinetochore formation

To faithfully segregate chromosomes during vertebrate mitosis, kinetochore-microtubule interactions must be restricted to a single site on each chromosome. Prior work on pair-wise kinetochore protein interactions has been unable to identify the mechanisms that prevent outer kinetochore formation in regions with a low density of CENP-A nucleosomes. To investigate the impact of higher-order assembly on kinetochore formation, we generated oligomers of the inner kinetochore protein CENP-T using two distinct, genetically engineered systems in human cells. Although individual CENP-T molecules interact poorly with outer kinetochore proteins, oligomers that mimic centromeric CENP-T density trigger the robust formation of functional, cytoplasmic kinetochore-like particles. Both in cells and in vitro, each molecule of oligomerized CENP-T recruits substantially higher levels of outer kinetochore components than monomeric CENP-T molecules. Our work suggests that the density dependence of CENP-T restricts outer kinetochore recruitment to centromeres, where densely packed CENP-A recruits a high local concentration of inner kinetochore proteins.

Microtubule rescue control by drugs and MAPs examined with in vitro pedestal assay

Microtubules are essential cytoskeletal polymers, which exhibit stochastic transitions between assembly and disassembly, known as catastrophes and rescues. Understanding of catastrophes, rescues, and their control by drugs and microtubule associated proteins (MAPs) has been informed by in vitro reconstitutions of microtubule dynamics. In such experiments microtubules are typically observed on a flat surface of the coverslip. In contrast, we have recently proposed a modified setup in which microtubules assemble from stabilized seeds, overhanging from microfabricated pedestals, so that their dynamic extensions are fully isolated from contact with the coverslip. This assay allows to eliminate potential artifacts, which may substantially affect the frequency of microtubule rescues in vitro. Here we use the pedestal assay to study the sensitivity of microtubules to paclitaxel, one of the best-known inhibitors of microtubule dynamics. By comparing observations in the conventional and the pedestal assays, we find that microtubule dynamics are substantially more sensitive to paclitaxel when the polymers can contact the coverslip. We interpret this as a consequence of the coverslip-induced microtubule assembly perturbation, leading to formation of lattice with defects, and thereby enhancing the efficiency of paclitaxel binding to microtubules in the conventional assay. To test this idea, we use vinblastine, another small-molecule inhibitor, which had been previously shown to cause microtubule growth perturbations. We find that in the pedestal assay vinblastine sensitizes microtubules to paclitaxel to the level, observed in the conventional assay. Interestingly, a minimal fragment of MAP called CLASP2, a previously characterized rescue factor, has a strong effect on microtubule rescues, regardless of the type of assay. Overall, our study underscores the role of microtubule damage in promoting rescues and highlights the utility of the in vitro pedestal assay to study microtubule dynamics modulation by tubulin inhibitors and MAPs.

Abl2 repairs microtubules and phase separates with tubulin to promote microtubule nucleation

Abl family kinases are evolutionarily conserved regulators of cell migration and morphogenesis. Genetic experiments in Drosophila suggest that Abl family kinases interact functionally with microtubules to regulate axon guidance and neuronal morphogenesis. Vertebrate Abl2 binds to microtubules and promotes their plus-end elongation, both in vitro and in cells, but the molecular mechanisms by which Abl2 regulates microtubule (MT) dynamics are unclear. We report here that Abl2 regulates MT assembly via condensation and direct interactions with both the MT lattice and tubulin dimers. We find that Abl2 promotes MT nucleation, which is further facilitated by the ability of the Abl2 C-terminal half to undergo liquid-liquid phase separation (LLPS) and form co-condensates with tubulin. Abl2 binds to regions adjacent to MT damage, facilitates MT repair via fresh tubulin recruitment, and increases MT rescue frequency and lifetime. Cryo-EM analyses strongly support a model in which Abl2 engages tubulin C-terminal tails along an extended MT lattice conformation at damage sites to facilitate repair via fresh tubulin recruitment. Abl2Δ688-790, which closely mimics a naturally occurring splice isoform, retains binding to the MT lattice but does not bind tubulin, promote MT nucleation, or increase rescue frequency. In COS-7 cells, MT reassembly after nocodazole treatment is greatly slowed in Abl2 knockout COS-7 cells compared with wild-type cells, and these defects are rescued by re-expression of Abl2, but not Abl2Δ688-790. We propose that Abl2 locally concentrates tubulin to promote MT nucleation and recruits it to defects in the MT lattice to enable repair and rescue.

Interface-acting nucleotide controls polymerization dynamics at microtubule plus- and minus-ends

GTP-tubulin is preferentially incorporated at growing microtubule ends, but the biochemical mechanism by which the bound nucleotide regulates the strength of tubulin:tubulin interactions is debated. The 'self-acting' (cis) model posits that the nucleotide (GTP or GDP) bound to a particular tubulin dictates how strongly that tubulin interacts, whereas the 'interface-acting' (trans) model posits that the nucleotide at the interface of two tubulin dimers is the determinant. We identified a testable difference between these mechanisms using mixed nucleotide simulations of microtubule elongation: with self-acting nucleotide, plus- and minus-end growth rates decreased in the same proportion to the amount of GDP-tubulin, whereas with interface-acting nucleotide, plus-end growth rates decreased disproportionately. We then experimentally measured plus- and minus-end elongation rates in mixed nucleotides and observed a disproportionate effect of GDP-tubulin on plus-end growth rates. Simulations of microtubule growth were consistent with GDP-tubulin binding at and 'poisoning' plus-ends but not at minus-ends. Quantitative agreement between simulations and experiments required nucleotide exchange at terminal plus-end subunits to mitigate the poisoning effect of GDP-tubulin there. Our results indicate that the interfacial nucleotide determines tubulin:tubulin interaction strength, thereby settling a longstanding debate over the effect of nucleotide state on microtubule dynamics.

CLASPs stabilize the pre-catastrophe intermediate state between microtubule growth and shrinkage

Cytoplasmic linker-associated proteins (CLASPs) regulate microtubules in fundamental cellular processes. CLASPs stabilize dynamic microtubules by suppressing microtubule catastrophe and promoting rescue, the switch-like transitions between growth and shrinkage. How CLASPs specifically modulate microtubule transitions is not understood. Here, we investigate the effects of CLASPs on the pre-catastrophe intermediate state of microtubule dynamics, employing distinct microtubule substrates to mimic the intermediate state. Surprisingly, we find that CLASP1 promotes the depolymerization of stabilized microtubules in the presence of GTP, but not in the absence of nucleotide. This activity is also observed for CLASP2 family members and a minimal TOG2-domain construct. Conversely, we find that CLASP1 stabilizes unstable microtubules upon tubulin dilution in the presence of GTP. Strikingly, our results reveal that CLASP1 drives microtubule substrates with vastly different inherent stabilities into the same slowly depolymerizing state in a nucleotide-dependent manner. We interpret this state as the pre-catastrophe intermediate state. Therefore, we conclude that CLASPs suppress microtubule catastrophe by stabilizing the intermediate state between growth and shrinkage.