Promoting diversity in the cytoskeleton through STEM - Structural Transitions and Energetics of Microtubules

Abstract

All microtubules assemble from tubulin subunits but the impact of specific tubulin isotypes has been unclear. Reporting in Developmental Cell, (Chabaan et al. 2018) and (Ti et al. 2018) show that tubulin isotypes differ in their conformational flexibility that alter microtubule dynamics and architecture yet maintain “lock-and-key” interactions with their neighbors.

Microtubules are a critical component of the eukaryotic cytoskeleton. Microtubules undergo self-assembly as α-/β-tubulin heterodimers to form protofilaments that then associate laterally to form hollow tubes (Figure 1A). GTP binding and hydrolysis by the β-tubulin subunit fuels microtubule dynamics, which is used for exploring space and generating forces in the cell. Tubulin proteins have diverged little throughout evolution, with β-tubulins showing 75–85% sequence identity across most animals, plants, and fungi. Although some species contain only single genes encoding α- and β-tubulin, most eukaryotes have evolved multiple genes that express isotypes of high sequence and structural identity. Why are there so many tubulin isotypes and what are the sequence-structure-function relationships for each isotype? Studies on microtubules have largely been carried out with tubulin proteins purified from cow or pig brain, preparations that contain a mixture of tubulin isotypes that are post-translationally modified, thus making questions about isotype function challenging to address.

A) Microtubules are dynamic filaments that undergo phases of growth and shrinkage due to the addition or removal, respectively, of α-/β-tubulin subunits. Lateral interactions between neighboring subunits stabilize microtubules (light blue: H2-S3 loop; dark blue: H1-S2 loop; magenta: M loop). B) Summary of isotype-specific changes in protofilament number, tubulin structure, and microtubule stability.

Recent technological advances have ushered in a new era of research on the microtubule cytoskeleton, which promises to answer the question of why so many isotypes exist. Baculovirus technology has enabled the purification of recombinant, isotypically-pure tubulin dimers (Minoura et al., 2013; Pamula et al., 2016; Ti et al., 2016; Vemu et al., 2016) and the TOG domain of the microtubule associated protein (MAP) Stu2 can be used to purify tubulins from native tissue (Widlund et al., 2012). The ability to generate recombinant tubulins opens the door to understanding how tubulin mutations associated with human diseases or that alter GTP binding/hydrolysis can impact microtubule dynamics, structure, and function (Ti et al., 2016;Geyer et al., 2015). In this issue of Developmental Cell, work from the Brouhard and Kapoor groups -- Chaaban et al. (2018) and (Ti et al. 2018), respectively -- provides new information on the sequence-structure-function relationships of tubulin isotypes across species.

Chaaban et al. (2018) of the Brouhard group investigated the properties of C. elegans tubulin, as worm microtubules display fast growth rates and form microtubules containing fewer protofilaments than mammalian cells. Using a TOG affinity column, the authors isolated and studied a mixture of dimers comprised of the β-tubulins TBB-1 and TBB-2 and the α-tubulin TBA-1. (Ti et al. 2018) of the Kapoor lab purified recombinant dimers of either human α1B/β2B or α1B/β3 tubulins, the major tubulin isotypes expressed in neuronal cells. Notably, (Ti et al. 2018) use a purification scheme that generates recombinant tubulin of native sequence, thereby avoiding the potential effects of affinity tags or contaminating insect cell tubulin that plagued previous efforts.

Both groups discovered that their tubulin proteins produced microtubules with varying protofilament number: C. elegans tubulin formed 12 or 13 protofilament microtubules, human α1B/β3 tubulin formed 13 protofilament microtubules, and human α1B/β2B tubulin formed 14 or 15 protofilament microtubules (Figure 1B). This was striking as previous studies observed microtubules with mostly 14 protofilament microtubules (Atherton et al., 2018; Vemu et al., 2017; Vemu et al., 2016). These results suggest that isotypes differ in their structural “plasticity” and that these differences allow them to form microtubules with varying protofilament number. Both groups used cryo-electron microscopy (cryo-EM) to study the structure of the microtubules and found that, despite the differences in isotype flexibility and protofilament architecture, the “lock-and-key” contacts within the microtubule lattice remained intact. In particular, interactions between the H1-S2 loop and the M-loop, which form lateral contacts between dimers of adjacent protofilaments (Figure 1A), were observed in all of the microtubules.

The altered protofilament compositions were accompanied by significant changes in microtubule dynamics. C. elegans tubulin generates the most dynamic microtubules studied to-date: at all concentrations, it has a high on-rate and “ages” 4-fold faster than cow brain tubulin. Interestingly, molecular dynamics simulations indicate that the H1-S2 and M loops are ordered not only within worm microtubules, but also in solution. (Chabaan et al. 2018) propose that ordering of these loops “pre-pays” an entropic cost required for tubulin assembly into the microtubule lattice. That is, C. elegans tubulin is “primed” for polymerization (Figure 1B). In this respect, it is noteworthy that the H1-S2 loop of S. cerevisiae tubulin is also ordered, and that it exhibits a high on-rate compared to cow brain tubulin (Geyer et al., 2015).

(Ti et al. 2018) show that α1B/β3 GMPCPP microtubules depolymerize 4–6 more quickly than α1B/β2B GMPCPP microtubules, consistent with previous work highlighting catastrophe frequency as a parameter impacted by human tubulin isotypes (Pamula et al., 2016; Vemu et al., 2017; Vemu et al., 2016). The increased depolymerization rates were observed regardless of whether depolymerization was induced by dilution of the microtubules or by two different depolymerization factors (the kinesin-13 MCAK or the MAP chTOG). Significantly, both MCAK and chTOG bind equally to α1B/β2B and α1B/β3 microtubules, excluding the trivial explanation that MCAK and chTOG do not efficiently bind α1B/β2B microtubules. The authors propose that α1B/β2B microtubules contain fewer “buriable” residues than α1B/β3. This effectively makes α1B/β2B microtubules more flexible and stable and perhaps not as prone to depolymerization and breakage (Figure 1B).

One outstanding question with the work of (Ti et al. 2018) is the inability to distinguish whether differences in microtubule stability are due to the isotype itself and/or to the protofilament number. To address this, (Ti et al. 2018) mixed the isotypes and observed intermediate effects with respect to both protofilament number and stability. Similar findings are reported by (Chabaan et al 2018) and earlier by the Roll-Mecak lab (Vemu et al., 2017; Vemu et al., 2016). Based on these findings, (Ti et al. 2018) suggest that it is the isotype itself that controls multiple properties of microtubules including protofilament number, rates of growth and disassembly, and potentially binding to MAPs and motors.

Overall, the two studies demonstrate that tubulin isotypes differ in their structural plasticity and that these differences impact both the dynamics and the architecture of the resulting microtubule. These results open up many new questions. First, how do the isotype-specific properties identified in these purified systems relate to the templated assembly of microtubules in cells? Are isotypes with higher plasticity better substrates for templated assembly? Second, how does isotype composition influence the architecture and stability of microtubules within an individual cell? Isotypes with higher plasticity could be preferred in specialized structures that need to withstand forces, such as the axons of a neuron or the primary cilium of an epithelial cell. On the other hand, the stochastic incorporation of isotypes with different properties during polymerization may cause location-specific alterations in the microtubule lattice such as islands of GTP-containing tubulin subunits, stutters in the lattice, or shifts in protofilament number, all of which can influence microtubule integrity and function. Third, how does isotype composition impact the efficacy of microtubule-targeting agents used as therapeutics to treat a number of human pathologies? As with most fields, a technical advance brings not only an ability to refine and clarify previous work but also an ability to answer previously remote questions.