All TOGether Now αβ-Tubulin is the polymerizing subunit of microtubules, which are dynamic polymers that have essential roles in cell division and intracellular organization. TOG domains are αβ-tubulin binding modules that occur in the evolutionarily conserved Stu2p/XMAP215 family of proteins and promote microtubule elongation. Ayaz et al. (p. 857 ) used crystallographic and biochemical experiments to reveal that the TOG1 domain interacts with guanosine triphosphate–bound αβ-tubulin in a conformation-selective manner, binding preferentially to a “curved,” microtubule-incompatible conformation. The binding mode apparently excludes analogous binding of a second TOG domain to the same heterodimer and may help to ensure polarized growth of microtubules.

Stu2p/XMAP215 proteins are essential microtubule polymerases that use multiple αβ-tubulin-interacting TOG domains to bind microtubule plus ends and catalyze fast microtubule growth. We report here the structure of the TOG2 domain from Stu2p bound to yeast αβ-tubulin. Like TOG1, TOG2 binds selectively to a fully ‘curved’ conformation of αβ-tubulin, incompatible with a microtubule lattice. We also show that TOG1-TOG2 binds non-cooperatively to two αβ-tubulins. Preferential interactions between TOGs and fully curved αβ-tubulin that cannot exist elsewhere in the microtubule explain how these polymerases localize to the extreme microtubule end. We propose that these polymerases promote elongation because their linked TOG domains concentrate unpolymerized αβ-tubulin near curved subunits already bound at the microtubule end. This tethering model can explain catalyst-like behavior and also predicts that the polymerase action changes the configuration of the microtubule end.DOI: http://dx.doi.org/10.7554/eLife.03069.001

Microtubule dynamics play essential roles in intracellular organization and cell division. They result from structural and biochemical properties of αβ-tubulin heterodimers and how these polymerizing subunits interact with themselves and with regulatory proteins. A broad understanding of the underlying mechanisms has been established, but fundamental questions remain unresolved. The lack of routine access to recombinant αβ-tubulin represents an obstacle to deeper insight into αβ-tubulin structure, biochemistry, and recognition. Indeed, the widespread reliance on animal brain αβ-tubulin means that very few in vitro studies have taken advantage of powerful and ordinarily routine techniques like site-directed mutagenesis. Here we report new methods for purifying wild-type or mutant yeast αβ-tubulin from inducibly overexpressing strains of Saccharomyces cerevisiae. Inducible overexpression is an improvement over existing approaches that rely on constitutive expression: it provides higher yields while also allowing otherwise lethal mutants to be purified. We also designed and purified polymerization-blocked αβ-tubulin mutants. These "blocked" forms of αβ-tubulin give a dominant lethal phenotype when expressed in cells; they cannot form microtubules in vitro and when present in mixtures inhibit the polymerization of wild-type αβ-tubulin. The effects of blocking mutations are very specific, because purified mutants exhibit normal hydrodynamic properties, bind GTP, and interact with a tubulin-binding domain. The ability to overexpress and purify wild-type αβ-tubulin, or mutants like the ones we report here, creates new opportunities for structural studies of αβ-tubulin and its complexes with regulatory proteins, and for biochemical and functional studies of microtubule dynamics and its regulation.

No abstract

AZD9291 (Osimertinib) is highly effective in treating EGFR-mutated non-small-cell lung cancers (NSCLCs) with T790M-mediated drug resistance. Despite the remarkable success of AZD9291, its binding pose with EGFR T790M remains unclear. Here, we report unbiased, atomic-level molecular dynamics (MD) simulations in which spontaneous association of AZD9291 with EGFR kinases having WT and T790M mutant gatekeepers was observed. Simulation-generated structural models suggest that the binding pose of AZD9291 with T790M differs from its binding pose with the WT, and that AZD9291 interacts extensively with the gatekeeper residue (Met 790) in T790M but not with Thr 790 in the WT, which explains why AZD9291 binds T790M with higher affinity. The MD simulation-generated models were confirmed by experimentally determined EGFR/T790M complex crystal structures. This work may facilitate the rational design of drugs that can overcome resistance mutations to AZD9291, and more generally it suggests the potential of using unbiased MD simulation to elucidate small-molecule binding poses.