Insights into allosteric control of microtubule dynamics from a buried β‐tubulin mutation that causes faster growth and slower shrinkage

Abstract

αβ‐tubulin subunits cycle through a series of different conformations in the polymer lattice during microtubule growing and shrinking. How these allosteric responses to different tubulin:tubulin contacts contribute to microtubule dynamics, and whether the contributions are evolutionarily conserved, remains poorly understood. Here, we sought to determine whether the microtubule‐stabilizing effects (slower shrinking) of the β:T238A mutation we previously observed using yeast αβ‐tubulin would generalize to mammalian microtubules. Using recombinant human microtubules as a model, we found that the mutation caused slow microtubule shrinking, indicating that this effect of the mutation is indeed conserved. However, unlike in yeast, β:T238A human microtubules grew faster than wild‐type and the mutation did not appear to attenuate the conformational change associated with guanosine 5′‐triphosphate (GTP) hydrolysis in the lattice. We conclude that the assembly‐dependent conformational change in αβ‐tubulin can contribute to determine the rates of microtubule growing as well as shrinking. Our results also suggest that an allosteric perturbation like the β:T238A mutation can alter the behavior of terminal subunits without accompanying changes in the conformation of fully surrounded subunits in the body of the microtubule.

1. INTRODUCTION

Microtubules are hollow cylindrical polymers that organize the cytoplasm, provide tracks for motor‐based transport, and form the mitotic spindle that mediates proper chromosome segregation.1 Microtubules assemble from αβ‐tubulin heterodimers and exhibit GTPase‐dependent dynamic instability,2 the seemingly random switching between phases of growing and rapid shrinking. Dynamic instability allows rapid reorganization of the microtubule network, is necessary for function, and is targeted by anticancer drugs.

Microtubule dynamics is a nonequilibrium behavior that is a consequence of guanosine 5′‐triphosphate (GTP) hydrolysis in the polymer. The broad outlines of how GTP hydrolysis drives dynamic instability have been established for some time.2, 3, 4, 5, 6, 7, 8, 9 The microtubule lattice is constructed from two fundamental interactions: longitudinal (head‐to‐tail) contacts between αβ‐tubulin heterodimers form the “protofilaments” that are aligned with the long axis of the polymer, and lateral (side‐to‐side) contacts hold protofilaments together around the circumference of the tube.10, 11, 12 The GTPase site resides at the longitudinal interface between αβ‐tubulins.10, 13 Addition of a new subunit to the growing microtubule end completes the active site and thereby stimulates GTP hydrolysis. The guanosine 5′‐diphosphate (GDP) microtubule lattice is unstable but typically protected from depolymerization by the stabilizing cap, a region near the growing microtubule end where GTP hydrolysis has not yet occurred (and/or, perhaps, where P_i has not yet dissociated).2, 14, 15 Loss of this cap exposes the labile core of the microtubule, leading to rapid depolymerization.

In addition to the GTPase cycle, there is also a conformational cycle: αβ‐tubulin subunits transition through different conformations during and after their association with the growing microtubule end.16, 17, 18, 19, 20, 21, 22, 23 Unpolymerized αβ‐tubulin subunits have a “curved” conformation that is not compatible with the geometry of the “straight” microtubule lattice, and in which the α‐ and β‐tubulin subunits are related by a roughly 13° rotation.16, 18, 22, 24, 25 Partially straightened conformations occur at the microtubule end26, 27, 28 and are thought to reflect a balance between the preference of αβ‐tubulin to be curved,18, 22, 24, 25 and the tubulin:tubulin contacts that drive straightening.29, 30, 31 These intermediate conformations have not yet been described in atomic detail. αβ‐tubulin adopts a “straight” conformation when sufficiently surrounded in the microtubule lattice. Multiple conformations of straight αβ‐tubulin have been described, and these are thought to reflect different nucleotide states (GTP, GDP.P_i, GDP).19, 20, 23, 32 In contrast to the GTPase cycle, how this “conformational cycle” contributes to microtubule dynamics is much less well understood.

A longtime reliance on αβ‐tubulin purified from animal brains limited the ability to probe and perturb the αβ‐tubulin conformation cycle in a site‐directed way. Recently introduced methods for purifying recombinant αβ‐tubulin from overexpressing hosts33, 34, 35, 36 have allowed new ways to study how the αβ‐tubulin conformation cycle impacts microtubule dynamics. In an earlier study, we described a buried mutation in yeast β‐tubulin (β:T238A) that dramatically stabilized microtubules, stimulating spontaneous nucleation and reducing the rate of postcatastrophe shrinking by ~100‐fold.37 The mutation also perturbed the conformation cycle, apparently causing αβ‐tubulin in the GDP lattice to adopt a conformation closer to that of the GTP or GTP‐like lattice. These findings represented some of the first direct evidence that the conformation cycle could dictate aspects of microtubule dynamics.

In the present study, we sought to determine whether effects of the β:T238A mutation we had observed in yeast microtubules would generalize to microtubules from other species. Strong conservation of sequence and structure made it seem reasonable to expect similar consequences of the mutation in yeast and human αβ‐tubulin. At the same time, it also seemed possible that the mutation might yield different effects: yeast and human microtubules display different polymerization dynamics37, 38 and adopt somewhat different lattice structures.39, 40 This reflects a general problem for the field: although it is increasingly apparent that microtubule dynamics can vary significantly between organisms (or even for different isoforms within a given organism),37, 41, 42, 43 ascribing the change in dynamics to specific amino acid substitutions has generally not been possible (but see41 for one attempt).

Adapting a recently introduced method34 for purifying recombinant human αβ‐tubulin from baculovirus‐infected insect cells, we purified wild‐type (WT) and β:T238A human αβ‐tubulin to measure their in vitro polymerization dynamics. The β:T238A mutation in human αβ‐tubulin substantially reduced microtubule shrinking rate and catastrophe frequency, demonstrating a microtubule‐stabilizing effect similar to what we had observed for yeast tubulin. We made two observations that differed from what we had seen for yeast tubulin. First, human β:T238A microtubules elongated significantly faster than their WT counterparts. Second, as judged by EB1 binding, human β:T238A microtubules did not appear to adopt a more GTP‐like conformation in the body of the microtubule (GDP lattice).

Our findings identify a conserved consequence of the β:T238A mutation while also revealing species‐specific differences. On one hand, the stabilizing effect of the β:T238A mutation is conserved. Thus, the mutation likely perturbs a fundamental allosteric property of αβ‐tubulin that helps determine the stability of lattice contacts in all/most microtubules. On the other hand, the differential effects on growth rates and on EB1 binding to the lattice represent species‐specific consequences of the mutation. Faster growth of mutant human microtubules suggests that whatever allosteric property (probably assembly‐dependent straightening) is being perturbed contributes variably to elongation in different species. Likewise, failure to observe enhanced EB1 binding to mutant human microtubules suggests that the lattice conformation of mutant human microtubules is apparently not changing as much as it did in yeast, although we cannot yet rule out the possibility that the lattice changes in a way that human EB1 cannot detect. If the different binding of EB proteins to the body of mutant human or yeast microtubules reflects different, species‐specific conformational responses to the mutation, it would imply that the conformation of αβ‐tubulin in the body of the microtubule may not always reliably reflect how stably subunits will bind at the microtubule end. The conserved and species‐specific consequeces of the β:T238A mutation reported here provide new insights into the role of the αβ‐tubulin conformation cycle in microtubule dynamics.

2. RESULTS

2.1. Expression and purification of WT and β:T238A recombinant human αβ‐tubulin

The β:T238A mutation was previously identified in screens for drug‐ and temperature‐sensitive phenotypes in budding yeast.44 β:T238 is located on the “core helix” of the tubulin fold, a secondary structural element that bridges between the “top” and “bottom” of the protein and that is positioned differently in straight and curved conformations (Figure 1a). β:T238 is also solvent‐inaccessible (buried), strongly conserved from yeast to human, and positioned in a region of the tubulin sequence that shows little variation across evolution or isoforms (Figure 1b).

Figure 1.

β:T238A αβ‐tubulin: structural context and purification. (a) Cartoon showing straight (orange, PDB 6O2R) and curved (green, PDB 4I4T) conformations of β‐tubulin superimposed using Cα coordinates from the N‐terminal domain (residues 1–180). β:T238 (position indicated with spheres) resides on the “core helix” that is positioned differently in the two conformations. (b) Multiple sequence alignment showing that β:T238 (highlighted in yellow) is strongly conserved (predominantly T, occasionally C) across evolution and among several human β‐tubulin isoforms. The sequence surrounding β:T238 also varies little. The residue immediately following T238 is either a C or S in different isoforms (and homologs), but whether the identity of that residue affects microtubule dynamics has not been tested. (c) Coomassie Blue stained SDS‐PAGE showing purified, tag‐free wild‐type and β:T238A recombinant human αβ‐tubulin. (d) Mass spectrometry analysis of purified wild‐type and β:T238A αβ‐tubulin. Counts for α‐ and β‐tubulin were separately normalized to the most abundant species. The two major species correspond to TUBA1A (50,151 Da) and TUBB3 plus remnants from the TEV recognition site (51,227 Da and 51,197 for wild‐type and β:T238 respectively). Minor species detected are assigned as follows: a—α‐tubulin from the Tni expression host (49,936 Da); b—monoglutamlylated Tni α‐tubulin (50,065 Da); c—acetylated TUBA1A (50,194 Da)

To examine how the β:T238A mutation affected the dynamics of human microtubules, we expressed WT and β:T238A human αIb/β3‐tubulin in baculovirus‐infected insect cells, adapting previously published protocols.34 We chose to study αIb/β3‐tubulin because it was one of the earliest human isoforms characterized for polymerization dynamics.34 The expression construct for β‐tubulin contains a C‐terminal, TEV‐cleavable His₆‐tag. After Ni‐affinity and anion exchange chromatography (with intervening TEV cleavage of the His₆ tag), we routinely obtain ~1 mg of tag‐free human tubulin per liter of culture (Figure 1c). Intact mass spectrometry analysis (Figure 1d) revealed that ~70% of the purified recombinant tubulin contained human α‐tubulin and the rest contained endogenous α‐tubulin from insect cells. α‐tubulin proteins from human and insect cells share 97% sequence identity, so as noted previously34 it seems unlikely that substoichiometric amounts of insect‐cell α‐tubulin significantly alter the behaviors of recombinant MTs or their interactions with regulatory factors such as EB1.

2.2. The buried β:T238A mutation stabilizes human microtubules and increases their growth rate

We used time‐lapse differential interference contrast (DIC) microscopy to measure the polymerization dynamics of WT and β:T238A tag‐free human tubulin, using GMPCPP‐stabilized brain MTs as seeds (Figure 2a). Spontaneous nucleation prevented us from measuring mutant microtubule dynamics at the same concentrations as WT (not shown), but by using lower concentrations we were able to observe mutant microtubules growing and shrinking without appreciable spontaneous nucleation.

Figure 2.

Polymerization dynamics of wild‐type and β:T238A human microtubules. (a) Representative kymographs for wild‐type (10 μM) and β:T238A (2 μM) microtubules. Slower shrinking for β:T238A is evident, as is a rescue event. The assays were performed at different concentrations because β:T238A spontaneously formed microtubules at the concentrations where we measured wild‐type. (b) Quantification of growth rates for wild‐type (blue symbols) and β:T238A (red symbols) microtubules at the indicated concentrations of αβ‐tubulin (n = 25–66 for each point). Bars show s.d. and the line indicates a linear fit to the data. The concentration‐dependence (slope) of growth rates increases about four‐fold for β:T238A compared to wild‐type. (c) Quantification of shrinking rates for wild‐type (blue) and β:T238A (red) microtubules (n = 326, 106 respectively). Data are pooled from multiple concentrations. (d) Quantification of microtubule growth lifetimes at conditions where growth rates for wild‐type (8 μM, blue) and β:T238A (2 μM, red) microtubules (n = 186, 68 respectively) are approximately matched, plotted as a survival curve. These growth lifetimes correspond to catastrophe frequencies of 0.24 and 0.01 min⁻¹, respectively. Many mutant microtubules did not undergo catastrophe during the recording. (e) Lifetime data from panel D plotted as a cumulative distribution. Colors and concentrations as above. Only microtubules that showed a catastrophe are included here

The β:T238A mutation in human αβ‐tubulin changes microtubule polymerization dynamics in ways that resemble what we observed for yeast microtubules, with some differences. Whereas β:T238A yeast microtubules elongated with the same characteristic concentration‐dependent growth rates as WT,37 human β:T238A microtubules showed a substantially steeper concentration‐dependence of growth rates than WT (~3.5‐fold increase: slopes of 0.09 and 0.32 μm/min/μM for WT and mutant respectively) (Figure 2b). Analogous to what we observed for yeast tubulin, β:T238A human microtubules shrank significantly more slowly compared to WT (Figure 2a,c; ~16‐fold change: 45.2 vs 2.9 μm/min for WT and mutant; p < .0001), and showed a substantially lower frequency of catastrophe (Figure 2d; 0.24 vs. 0.01 min⁻¹ for wild type and mutant at respective concentrations of 8 μM and 2 μM, the lowest concentration of WT and the highest concentration of mutant we measured and where growth rates are comparable). We also observed rescues for mutant microtubules (1.08 min⁻¹ at 2 μM concentration), but not for wild type. These data are summarized in Table 1.

Table 1.

Summary of microtubule dynamics measurements for wild‐type and β:T238A microtubules. Error ranges correspond to standard deviation

	Wild‐type, 8 μM	β:T238A, 2 μM
Growth rate	0.58 ± .09 μm/min	0.62 ± .05 μm/min
Shrinking rate	45.2 ± 22.9 μm/min	2.9 ± 1.6 μm/min
Catastrophe frequency	0.24 min−1	0.01 min−1
Rescue frequency	none detected	1.08 min−1

2.3. β:T238A human microtubules do not accumulate substantially more GTP and GDP.P_i relative to WT

GTP stabilizes the microtubule lattice, so a mutation that causes slower shrinking and less frequent catastrophe might do so by perturbing GTPase activity, resulting in elevated GTP (or perhaps GDP.P_i) content in the mutant microtubules. In yeast microtubules, the β:T238A mutation did not significantly increase the amount of GTP or GDP.P_i in the microtubule lattice.37 To begin determining whether the human β:T238A microtubules were also stabilized without accumulating GTP or GDP.P_i in the lattice, we sought to directly compare the nucleotide content of WT and mutant human microtubules.

The nucleotides that incorporate into the microtubule with αβ‐tubulin are not exchangeable from the body of the lattice. We polymerized WT (10 μM) and β:T238A αβ‐tubulin (2 μM) in the presence of γ‐ or α‐³²P‐GTP and harvested the resulting microtubules by centrifugation. A higher concentration of WT αβ‐tubulin was necessary to get sufficient microtubule formation. After removing the supernatant (which contained unpolymerized αβ‐tubulin and unincorporated label), we denatured the pelleted microtubules to release bound nucleotides for analysis by thin layer chromatography (TLC) (Fig. 3a). The experiments using α‐³²P‐GTP revealed that β:T238A human microtubules do not contain more GTP than WT (6.0% for WT compared to 2.7% for the mutant; Fig. 3b). Experiments using γ‐³²P‐GTP showed that WT and mutant microtubules contained comparable amounts of GDP.P_i (2.2% for WT compared to 4.7% for mutant; the difference was not statistically significant; Fig. 3a,b). Thus, similar to what we had observed for yeast microtubules, the effects of the β:T238A mutation on microtubule dynamics cannot be ascribed to a measurable change in the nucleotide content of microtubules.

Figure 3.

Nucleotide content of wild‐type and β:T238A microtubules. (a) Representative image showing TLC analysis of exchangeable nucleotide content of microtubules grown with GTP doped with α‐³²P‐GTP (for analysis of GTP/GDP) or γ‐³²P‐GTP (for analysis of GTP/P_i). Microtubules were spontaneously assembled using 10 μM wild‐type or 2 μM β:T238A recombinant αβ‐tubulin. (b) Quantification of the TLC data (n = 3, errors represent s.d.); differences between wild‐type and β:T238A are not statistically significant

2.4. At matched growth rates, WT, and β:T238A human MTs have similarly sized stabilizing caps

As an alternative way to test for a change in GTPase activity, and as a way to probe whether the β:T238A mutation was affecting the conformation of αβ‐tubulin in the microtubule, we turned to EB1 binding assays. EB‐family proteins form “comets” that track growing microtubule ends by recognizing a conformation of αβ‐tubulin that occurs transiently as part of the GTPase cycle.15, 20, 32, 45 In addition to providing a way to detect conformational change in the microtubule lattice, characterizing the size and/or intensity of EB1‐GFP comets on growing microtubules also provides an indirect assay for GTPase activity that has better time resolution than the “TLC after sedimentation” assays described in the prior section.7, 46

We observed clear comets for both WT and mutant microtubules (Figure 4a; microtubules are not visible in these images because we did not fluorescently label the recombinant αβ‐tubulin). We were surprised to observe that EB1 binding was not elevated along the length of β:T238A human microtubules: this was a clear and striking difference from what we had previously observed for (His‐tagged) yeast microtubules37 (see Discussion). We quantified EB1‐GFP comet profiles by fitting a Gaussian function to intensity line scans along the microtubule, and we averaged several profiles together after aligning them to a common reference (Figure 4b). The fitted EB1 comet profiles were ~0.2 μm long; this distance is comparable to the point spread function of our microscope so it is not possible to reliably estimate the actual size of the comets. Accordingly, we used the total intensity from the EB1 comets as a proxy for their size. Control experiments demonstrated that integrated comet intensity depended on microtubule elongation rate (Figure 4c), as expected based on prior studies.7, 46, 47 Mutant microtubules also show a dependence of comet intensity on growth rate (Figure 4c). At growth rate matched conditions for WT (13 μM) and β:T238A (3.1 μM) microtubules (Figure 4d), the total EB1 comet intensity was not significantly different (Figure 4e).

Figure 4.

Lattice conformation of β:T238A microtubules probed using EB1‐GFP. (a) Representative kymographs of EB1‐GFP comets at the ends of growing microtubules, imaged by TIRF. 13 μM and 3.1 μM αβ‐tubulin was used for wild‐type and β:T238A respectively. Assays contain 43 nM EB1‐GFP, and microtubules are not labeled. (b) Averaged EB1‐GFP comet profiles (n = 30 and 20 for wild‐type [WT] and β:T238A respectively). The averaged line scans (solid lines) are shown with standard error (bars). The size and the height of the fluorescent intensity EB1‐GFP comet profile on growth rate matched WT (13 μM) and β:T238A (3.1 μM) microtubules are not statistically different. (c) Quantification of integrated fluorescence intensities (n > 10 for each) of EB1‐GFP (43 nM) comet on WT (black) and β:T238A (red) microtubules grown at multiple concentrations. The average integrated EB1‐GFP comet intensity (dots) is shown with the standard error (bars). Comet intensity for WT and mutant increases with microtubule growth rate. (d) Quantification of microtubule growth rates in the presence of EB1‐GFP (43 nM). Under these conditions, growth rates for WT (20.0 ± 0.7 nm/s, n = 26) and β:T238A (19.5 ± 0.5, n = 20) microtubules are not significantly different (p = .562). (e) Integrated fluorescence intensities of EB1‐GFP comet on WT (13 μM) and β:T238A (3.1 μM) microtubules. The differences in integrated intensities at growth rate matched conditions are not significantly different (p = .223)

These EB1‐binding assays reinforce the biochemical assays for nucleotide content by demonstrating that in human microtubules the β:T238A mutation does not measurably perturb EB1 comets and by inference GTPase activity. The failure to observe high‐affinity EB‐binding to the body of β:T238A human microtubules may also indicate that the mutation exerts its effects on microtubule dynamics without promoting a more GTP‐like conformation of αβ‐tubulin in the body of the microtubule, as it did in yeast. However, without structural information about the mutant microtubules we cannot yet reject the possibility that the β:T238A mutation causes a conformational change in the lattice that is “invisible” to human EB1.

3. DISCUSSION

3.1. Conserved consequences of the β:T238A mutation

The results reported here reveal that the buried β:T238A mutation in human αβ‐tubulin strongly stabilizes the microtubule lattice and reduces the frequency of catastrophe, without substantially changing the levels of GTP or GDP.P_i in the microtubule lattice. We previously observed similar effects of the β:T238A mutation on yeast microtubules,37 so reduced shrinking rate and catastrophe frequency without appreciable change in nucleotide content are likely to be conserved consequences of the mutation. Thus, despite their differing polymerization dynamics, preference for different protofilament number, and unequal fold change in shrinking rate, it appears that the stability of human and yeast microtubules is both controlled by some shared, allosteric property of αβ‐tubulin.

3.2. Species‐specific consequences of the β:T238A mutation

Two consequences of the β:T238A mutation differ substantially between the present study of human microtubules and our prior study of yeast microtubules.37 First, human β:T238A displayed growth rates that depended substantially (~3.5‐fold) more steeply on concentration than for WT (Figure 2b), whereas in yeast WT and mutant microtubules showed similar concentration‐dependence of growth rates.37 Second, as judged by EB‐binding experiments, αβ‐tubulin does not adopt a more GTP‐like conformation in the lattice of human β:T238A microtubules (Figure 4) whereas in mutant yeast microtubules it did.37 We discuss the implications of these differences below.

The concentration‐dependence of microtubule growth rates (slope in plots like Figure 2b) is related to the apparent on‐rate constant (k_on ^app) for microtubule elongation. How might the β:T238A mutation affect this parameter to make human microtubules grow faster? The mutation cannot directly alter the chemical composition of an interface, because it is buried. Instead, the mutation must modulate the strength of polymerization contacts indirectly, by alleviating some other property of αβ‐tubulin that normally opposes elongation. Given that the mutation resides on a secondary structure element that is differently positioned in straight and curved conformations of αβ‐tubulin, and because αβ‐tubulin “straightening” on the microtubule end is thought to incur an energetic cost that weakens interactions with the microtubule lattice,17, 18, 24, 25, 48 it seems likely that the mutation promotes assembly‐dependent "straightening" of αβ‐tubulin at/on the microtubule end (Figure 5). The mutation could in principle also affect the conformation of unpolymerized αβ‐tubulin, but we were unable to test for this because a tendency of β:T238A human αβ‐tubulin to aggregate prevented the conformation‐selective binding assay used in our prior study of the yeast mutant.37 The mutation could also operate through some mechanism other than curvature, for example, by inducing local structural changes that make longitudinal or lateral assembly interfaces more favorable,24, 41 or by altering the “flexibility” of αβ‐tubulin subunits as has been suggested in recent molecular dynamics studies.49, 50 Whatever the precise mechanism, that the buried β:T238A mutation can cause microtubules to grow faster provides new evidence for the intimate links between allosteric properties of αβ‐tubulin and microtubule growing, shrinking, and switching.

Figure 5.

Speculative model for how the β:T238A mutation affects microtubule growing and shrinking. (a) Assembly‐dependent “straightening” of αβ‐tubulin is thought to incur an energetic cost that acts against elongation. Pink: α‐tubulin, solid green: WT β‐tubulin, lined green: mutant β‐tubulin. (b) An effect of the β:T238A mutation to favor straightening on the microtubule end may explain the faster growing and slower shrinking of the mutant microtubules. Other models where separate consequences of the mutant explain effects on growing and shrinking are also possible. (c) Cartoon illustrating differences in EB‐protein binding to the GDP lattice of human and yeast β:T238A microtubules. The finding that human microtubules shrink slower without detectable changes in EB1‐binding may indicate that the conformation of αβ‐tubulin in the lattice is not always coupled to how strongly it will interact at the growing or shrinking end

Despite shrinking much slower than WT, human β:T238A microtubules did not show elevated EB1‐binding along their length. This contrasts with the same mutation in yeast tubulin, where the dynamics and the EB‐binding of yeast microtubules were both affected.37 In that prior study, we had inferred that the change in lattice conformation and the change in shrinking rate were coupled. Based on the results reported here, it now seems that the state of the lattice as detected by EB proteins will not always reliably reflect how αβ‐tubulin will behave biochemically (rates of association and dissociation) at the microtubule end (Figure 5a,b). Structural analysis of WT and β:T238A human microtubules will be required to resolve this issue.

3.3. Conclusions and perspective

It is remarkable that despite over four decades of study by a number of groups, fundamental aspects of how microtubules grow and shrink remain poorly understood. By comparing how the same mutation affects the dynamics of human and yeast microtubules, our work revealed conserved allosteric properties but also some species‐specific differences. It seems plausible that tubulins from different species (or even different isoforms from the same species) may be “tuned” differently in terms of how strongly their assembly‐dependent conformational changes contribute to microtubule growing and shrinking (Figure 5b). Interestingly, a recent study41 revealed that C. elegans microtubules grow faster but shrink comparably to mammalian ones, perhaps indicating a different kind of “allosteric tuning” than in yeast or human tubulin. It is striking that point mutations or evolutionary variation can apparently control microtubule growing and shrinking independently, or together. How these differing changes in microtubule dynamics are achieved at the structural and biochemical levels remains unknown, and is an important topic for future research on other mutants, different isoforms, and/or αβ‐tubulins from other species.

Understanding the structural and biochemical foundations of microtubule dynamics remains a major challenge, in large part because the microtubule end presents a transient, biochemically heterogeneous substrate that has been difficult to characterize structurally. Recent technical advances that allow recombinant αβ‐tubulin to be purified from overexpressing hosts33, 34, 35, 36, 51, 52 or endogenous αβ‐tubulin to be purified from sources other than mammalian brain,53, 54 promise to provide new sources of information about how evolutionary diversity, isotype variation, posttranslational modification, and point mutations affect microtubule structure and dynamics.

4. METHODS

4.1. Expression and purification of wild‐type and β:T238A human αβ‐tubulin

A construct to express human αβ‐tubulin (with nontagged TUBA1B gene and TUBB3 gene with a cleavable His‐tag at the C‐terminus)34 was a gift from Tarun Kapoor's lab (Rockefeller University). Recombinant baculovirus were generated following standard methods (Invitrogen). Tni cells (Expression Systems) were used for expression in ESF‐921 insect‐cell medium (Expression Systems). Approximately 42 h after infection, cells were harvested, resuspended in three volumes of lysis buffer (25 mM Hepes, pH 7.4, 30 mM Imidazole, 1 mM MgSO₄, 50 μM GTP), and lysed using a glass dounce. Lysate was clarified by centrifugation before loading onto a 5 ml Ni‐NTA column (TaKaRa), and then washed with high salt wash buffer (50 mM Hepes, pH 7.4, 30 mM Imidazole, 10 mM MgSO₄, 500 mM NaCl, 50 μM GTP) followed by low salt wash buffer (25 mM PIPES, pH 6.9, 30 mM Imidazole, 1 mM MgSO₄, 50 μM GTP). Proteins were eluted from the column using elution buffer (25 mM PIPES, pH 6.9, 300 mM Imidazole, 1 mM MgSO₄, 50 μM GTP). The His‐tag was removed by TEV protease (2 hr on ice using a TEV at 0.2 mg/ml final concentration). The tag‐cleavage reactions were further purified by ion exchange chromatography using a 4 ml source Q (GE Amersham) column. Peak fractions were pooled, concentrated to 15–20 μM, buffer exchanged to BRB80 with 50 μM GTP, flash‐frozen on liquid nitrogen in 100 μl aliquots, and stored at −80°C.

4.2. Microtubule dynamics assays

Microtubule dynamics were measured by time‐lapse DIC microscopy, as described previously.37, 55, 56 Temperature was maintained at 30°C using a Weather Station with enclosure fit to the Olympus IX81 microscope body. The microscope was controlled using MicroManager.57 Briefly, assays were performed in flow chambers, using GMPCPP seeds made from brain tubulin (5% biotinylated; PurSolutions) attached to coverslips using neutravidin (Invitrogen). Wild‐type or β:T238A mutant microtubule dynamics were measured in the following buffer (144 mM PIPES, pH 6.9, 3.6 mM MgCl₂, 50 mM KCl, 1 mM GTP, 0.1 mg/ml BSA and 0.07% methylcellulose, final concentration). We used a higher than typical concentration of PIPES because microtubule dynamics were not robustly observable at lower concentrations. A 30‐min video was taken for each assay, and 50 ms images were recorded on a Photometrics Prime95B camera and averaged in batches of 30 to improve signal to noise. Growing and shrinking rates were obtained from kymographs prepared using ImageJ.58 Three or more assays for each concentration of wild‐type or β:T238A mutant tubulin were tested with different batches of protein preps.

4.3. Nucleotide content of wild‐type and β:T238A microtubules

Wild‐type (10 μM) or β:T238A mutant (2 μM) were incubated in spontaneous assembly buffer (1x BRB80 with 2.5% DMSO, 5% glycerol, 100 μM GTP, and 0.33 μM radiolabeled α‐ or γ‐³²P‐GTP). After 30 min incubation at 37°C, samples were spun in a tabletop ultracentrifuge (Beckman Optima TLX using TLA120 rotor) at 80,000 rpm at 37°C for 10 min. The pellets were rapidly washed four times with warm BRB80, denatured by resuspending in 2 μl 6 M guanidine hydrochloride, and diluted with 18 μl water for TLC. Radioactivity of each sample was checked to determine the loading amount of TLC. TLC and data analysis were performed as described.37 Samples were loaded onto a Cellulose PEI TLC plate (Selecto Scientific) and TLC was performed with the mobile buffer containing 0.75 M Tris, 0.4 M LiCl, and 0.45 M HCl. Radiolabeled mixtures of GTP/GDP and GTP/P_i were used as markers. The TLC plate was exposed to X‐ray film after being air‐dried. Nucleotide contents were calculated as described previously.37

4.4. TIRF microscopy (MT lattice binding)

The flow chambers were prepared as described previously.37, 55, 56 Samples of WT or β:T238A αβ‐tubulin along with 43 nM EB1‐GFP in imaging buffer (BRB80 + 0.1 mg/ml BSA + 0.2 mM MgCl₂ +1 mM GTP + 50 mM KCl + 0.1% Methylcellulose + antifade reagents [glucose, glucose oxidase, catalase]) were flowed into the chamber. Interactions of EB1‐GFP with microtubules were imaged by total internal reflection fluorescence (TIRF) microscopy using an Olympus IX81 microscope with a TIRF ApoN 100×/1.49 objective lens, a 491 nm 50 mW solid‐state laser, and a Photometrics Prime95B camera. Reactions were temperature controlled at 30°C, and the microscope was controlled using MicroManager 2.0.57 Images of microtubules were taken every 10 s for 15 min with the exposure time of 10 ms. EB1‐GFP fluorescence intensity along microtubules and extending beyond their growing ends was obtained using the PlotProfile function in ImageJ.58 These line scans were aligned and averaged using a custom‐made function in MATLAB.56 Line‐scan profiles were fit using a Gaussian distribution with flat baseline for background fluorescence.

Ye X, Kim T, Geyer EA, Rice LM. Insights into allosteric control of microtubule dynamics from a buried β‐tubulin mutation that causes faster growth and slower shrinkage. Protein Science. 2020;29:1429–1439. 10.1002/pro.3842