Structural mass spectrometry of the alpha beta-tubulin dimer supports a revised model of microtubule assembly

Abstract

The molecular basis of microtubule lattice instability derives from the hydrolysis of GTP to GDP in the lattice-bound state of alphabeta-tubulin. While this has been appreciated for many years, there is ongoing debate over the molecular basis of this instability and the possible role of altered nucleotide occupancy in the induction of a conformational change in tubulin. The debate has organized around seemingly contradictory models. The allosteric model invokes nucleotide-dependent states of curvature in the free tubulin dimer, such that hydrolysis leads to pronounced bending and thus disruption of the lattice. The more recent lattice model describes a predominant role for the lattice in straightening free dimers that are curved regardless of their nucleotide state. In this model, lattice-bound GTP-tubulin provides the necessary force to straighten an incoming dimer. Interestingly, there is evidence for both models. The enduring nature of this debate stems from a lack of high-resolution data on the free dimer. In this study, we have prepared alphabeta-tubulin samples at high dilution and characterized the nature of nucleotide-induced conformational stability using bottom-up hydrogen/deuterium exchange mass spectrometry (H/DX-MS) coupled with isothermal urea denaturation experiments. These experiments were accompanied by molecular dynamics simulations of the free dimer. We demonstrate an intermediate state unique to GDP-tubulin, suggestive of the curved colchicine-stabilized structure at the intradimer interface but show that intradimer flexibility is an important property of the free dimer regardless of nucleotide occupancy. Our results indicate that the assembly properties of the free dimer may be better described on the basis of this flexibility. A blended model of assembly emerges in which free-dimer allosteric effects retain importance, in an assembly process dominated by lattice-induced effects.

Figure 1

(A) Effect of urea-induced destabilization on total nucleotide occupancy, summing contributions from both the N- and E-sites, arising from simple equilibration with excess nucleotide (specific nucleotide shown in the legend). Values are normalized against the yield achieved in the absence of urea under GDP-saturating conditions, according to eq 1. (B) Results as in (A), adjusted for a reduced recovery of free dimer upon urea treatment. (C, D) Fractional E-site occupancy as a function of urea for tubulin equilibrated with excess GDP (C) and GTP (D), assuming full retention of N-site nucleotide occupancy. Error bars represent ±1 SD (see text for details).

Figure 2

Effect of E-site nucleotide occupancy on the deuteration levels of αβ-tubulin as determined by an H/DX-MS-based urea denaturation study. Deuteration values are relative to GTP-tubulin measured in the absence of urea. Total protein deuteration values were estimated by summation of peptides spanning the sequence (see Materials and Methods).

Figure 3

Reversibility assessment of the structural intermediate in GDP-tubulin suggested by Figure 2, evaluated at 0.25 M urea. (A) Peptides represent the set detectable at 10-fold dilution (190 nM tubulin) and demonstrating a significant increase in deuteration at 0.25 M urea. Experiments were done in triplicate, and error bars represent ±SD. (B) Representative isotopic clusters for VVEPYNSIL, demonstrating reversibility in deuteration level upon dilution.

Figure 4

(A) Structural representation of α/β-tubulin (α, pale green; β, cyan) at an oblique angle of the protofilament axis to highlight both the E-site and N-site nucleotides, with increases in deuteration arising from GMPCPP treatment relative to GTP superimposed in red. Nucleotides are represented as green spheres. (B) As in (A), with a face-on view of nucleotide at the E-site. (C) As in (A), with a face-on view of GTP at the N-site.

Figure 5

Alternative views of the differential deuteration level data for GDP/GTP at 0.25 M urea, mapped to PDB 1JFF (α-tubulin in pale green, β-tubulin in cyan). (A) Peptides with increased deuteration upon exchange with GDP, and in direct contact with E-site nucleotide, are shown in yellow. All others are shown in red. View is down the z-axis, corresponding to the microtubule protofilament axis viewed from the plus end, where y represents radial direction out from the center of the microtubule. (B) Color scheme as in (A), highlighting the intradimer region.

Figure 6

Differential deuteration data for GDP/GTP at 0.25 M urea, mapped to (A) PDB 1JFF and (B) PDB 1SA0. Data associated with the E-site are removed for clarity. GDP-induced increases in deuteration are again shown in red, with α-tubulin in pale green and β-tubulin in cyan. Peptides indicated in A represent the two sequences in contact with both inter and intradimer interface peptides. Colchicine is displayed in B as yellow spheres. (C) Differential SASA data for 1SA0 relative to 1JFF, restricted to the intradimer interface. Red represents >+50% change (colchicine removed for clarity).

Figure 7

View of the differential deuteration level data from the E-site for GMPCPP/GTP at 0.25 M urea mapped to PDB 1JFF as in Figure 5. Color scheme follows Figure 5.

Figure 8

Proposed model of nucleotide control over αβ-tubulin assembly from an intradimer perspective. The most flexible state of αβ-tubulin (left dimer, symbolized by greater spacing between α- and β-tubulin) is shown in equilibrium with a less flexible state (right dimer, symbolized by reduced spacing between α- and β-tubulin), with the relative population of states dictated at least in part by nucleotide occupancy, as shown. Intradimer flexibility is stabilized by the straight microtubule lattice upon assembly (symbolized by no spacing between α- and β-tubulin). On the basis of the significantly larger change in deuteration upon assembly, we suggest that the energy required for intradimer stabilization of GTP-tubulin (E₂) is only slightly less than that required to stabilize GDP-tubulin (E₁ + E₂).