Structural model for differential cap maturation at growing microtubule ends

Abstract

Microtubules (MTs) are hollow cylinders made of tubulin, a GTPase responsible for essential functions during cell growth and division, and thus, key target for anti-tumor drugs. In MTs, GTP hydrolysis triggers structural changes in the lattice, which are responsible for interaction with regulatory factors. The stabilizing GTP-cap is a hallmark of MTs and the mechanism of the chemical-structural link between the GTP hydrolysis site and the MT lattice is a matter of debate. We have analyzed the structure of tubulin and MTs assembled in the presence of fluoride salts that mimic the GTP-bound and GDP•P_i transition states. Our results challenge current models because tubulin does not change axial length upon GTP hydrolysis. Moreover, analysis of the structure of MTs assembled in the presence of several nucleotide analogues and of taxol allows us to propose that previously described lattice expansion could be a post-hydrolysis stage involved in P_i release.

Keywords: GTP-cap; dynamic instability; microtubule; molecular biophysics; none; structural biology; taxol.

Figure 1.. Structure of tubulin bound to GDP-phosphate analogues.

(A) The T₂R-TTL complex includes one RB3 molecule (orange), one TTL molecule (pink) and two tubulin heterodimers: α-tubulin (dark gray, GTP-bound, chains A and C), β-tubulin (light gray, GDP-bound, chains B and D). (B) Composite omits maps of BeF₃^- (left) and AlF₃ (middle and right): mFo-DFc maps (blue, contour level 1.0) of nucleotide and Mg²⁺ ions combined with 2mFo-DFc maps (green, contour level 3.0) of the related phosphate analogues. (C) GTP (PDB 5xp3, black) and GDP- (PDB 4i55, gray) bound β-tubulin highlighting the localization of secondary structural elements surrounding the nucleotide-binding site according to Löwe et al. (2001) and alternative positions of T5 loop depending on the nucleotide-bound state. (D) Chain D alignment of GDP-BeF₃^- structure (blue) with GMPCPP-bound (PDB 3ryh, orange), GTP-bound (black) and GDP-bound (gray) structures showing BeF₃^-/γ-phosphates co-localization, differences on the position of Mg²⁺ ions depending on the nucleotide-bound state (4.5 Å apart), and T5 loop in a GTP-like (‘out’) conformation. (E) Chain B alignment of GDP-AlF₃ (red) structure with GMPCPP-bound (orange) and GDP-bound (gray) structures highlighting that AlF₃ sits out of the γ-phosphate site at 2.77 Å from the Mg²⁺ ion (in a position equivalent to that on the GDP-bound structure). (F) Chain D alignment of GDP-AlF₃ (red) structure with GTP-bound (black) and GDP-bound (gray) structures showing the dual conformation of T5 loops.

Figure 2.. Phosphate analogues sustain tubulin activation and MT stabilization.

(A) Time course assembly of 30 μM GDP-tubulin (gray line) with either 1 mM GTP (black line) or 1 mM GDP and increasing BeF₃^- concentrations (1 mM, 2 mM, 4 mM, 6 mM, 8 mM and 10 mM; from light to dark blue). (B) Critical concentration (Cr) measurements of GDP-tubulin with 1 mM GTP (black square) or 1 mM, 2 mM, 3 mM, 4 mM, 5 mM, 6 mM, 7 mM and 8 mM BeF₃^- (blue circles). (C) Wyman plots showing the effect of increasing Mg²⁺ concentrations (4 mM, 5 mM, and 6 mM) on tubulin assembly (37°C) in the presence of 1 mM GTP (black squares) or 1 mM GDP plus 3 mM BeF₃^- (blue circles). Positive slopes indicate that at increasing Mg²⁺ concentrations the ion incorporation into the filament is higher, and similar slope values (under BeF₃^- and GTP) indicate a similar Mg²⁺ dependency of the polymerization. (D) Time course assembly of 30 μM GDP-tubulin with either 1 mM GTP (black line) or 1 mM GDP, 3 mM HKF₂ and increasing concentrations of AlCl₃ (0 μM, 50 μM, 100 μM, 250 μM, 500 μM and 1 mM; from light to dark red). (E) Time course assembly of 30 μM tubulin in the absence of free GTP (black line) and in the presence of increasing concentrations of BeF₃^- (100 nM, 500 nM, 1 μM, 10 μM and 50 μM; from light to dark blue) showing the stabilization effect of the analogue. (F) Time course assembly of 30 μM tubulin in the absence of free GTP and 0 mM HKF₂ (black line) or in the absence of free GTP with 3 mM HKF₂ and increasing concentrations of AlCl₃ (0 μM, 50 μM, 100 μM, 250 μM, 500 μM and 1 mM from light to dark red). Aluminum contamination in nucleotides and glass induce MT stabilization even when no AlCl₃ was added (light red). Arrows in graphs (A), (D), (E), and (F) indicate when samples were incubated at 4°C.

Figure 2—figure supplement 1.. Electron micrographs of MTs polymerized in the presence of BeF₃^- and AlFx.

(A) Electron micrograph of MTs grown in the presence of GDP-tubulin and 6 mM BeF₃^-. The inset shows length estimations of MTs (units are μm, n > 50) from samples containing 30 μM GDP-tubulin with 2 mM, 4 mM, and 6 mM BeF₃^- (blue circles), or 1 mM GTP (black square). (B) Electron micrograph of MTs grown in the presence of GDP-tubulin with 3 mM HKF₂ and 1 mM AlCl₃. Bar is 200 nm.

Figure 3.. Fiber diffraction of MT models systems.

GDP-BeF₃^--MT (blue), GDP-AlF_x-MT (red), GMPPCP-MT (salmon), GMPCP-MT (yellow), GMPCPP-MT (orange), GDP-Tx-MT (brown) and GDP-MT (gray). (A) Top; representative image (GMPCPP-MTs) of meridional diffraction displaying the meridional plane from l = 1 (4nm) layer line and related harmonics (l = 2 to 4) for longitudinal metric calculations. Bottom; meridional intensity patterns, where arrows indicate the 4 nm and 8 nm peaks. The inset shows the best fit of 1 nm band experimental intensities in a Lorentzian normal distribution, highlighting positional differences between all tested MT growing conditions (peaks maxima, arrows). (B) Top; representative image (GMPCPP-MTs) of equatorial diffraction highlighting the equatorial plane (l = 0) for lateral metric calculations. Bottom; equatorial intensity patterns showing the corresponding Bessel functions from J₀₁ to J₀₄+J_N1. The inset shows the J_N1, calculated as described in M and M, displaying the differences in peak maxima (arrows) that occur under various nucleotide polymerization conditions. The red dash line on (A) and (B) top images shows planes used for intensity line plotting in q_x space and further metric calculations. (C) Estimation of the number of PFs per MT and percentage of each subpopulation within the solution from fiber diffraction experiments (left) and cryo-EM images (right).

Figure 3—figure supplement 1.. Shear-flow aligned fiber diffraction experiments.

(A) 2X samples were incubated at 37°C and mixed (1:1) with pre-warmed 2% methylcellulose. The mixture was displayed onto a mica ring set on a ceramic holder and was inserted into the shear-flow device. Temperature and rotation was remotely controlled. (B) The sample was exposed to the X-ray beam and small angle diffracting images coming from the diamond window were collected in a 1M PILATUS detector. With an exposure time of 150 s, we could collect a maximum of 4 images per sample before any sign of radiation damage. (C) Images were buffer subtracted and averaged for data analysis. Each dataset (with a minimum of 16 images) was analyzed for the equatorial and meridional signals. The intensity profile was measured along distance from the center of the diffraction and subsequently analyzed. The equatorial profiles allowed the estimation of the diameter of the MTs, the PF number and the inter-PF distances. To get this information the J₀₃ and J₀₄+J_N1 signals were deconvolved according to MTs with different putative PF numbers and the fitted curves were integrated and correlated with the raw data. The meridional profile shows peaks at the 8 nm and/or the 4 nm families of layer lines. We calculated the Lorentzian distribution of the 1 nm line (a Bessel harmonic of the 4 nm) to find differences in the axial spacing between α- and β-tubulin.

Figure 3—figure supplement 2.. Shear-flow aligned fiber diffraction images.

Dataset images from MTs assembled under different conditions, showing the meridional diffraction. Insets centered on the equatorial diffraction highlight the presence (GDP-Tx-, GMPCPP- and GMPCP-MTs) or absence (GDP-, BeF₃^--, AlF_x- and GMPPCP-MTs) of the 8 nm layer line.

Figure 3—figure supplement 3.. Shear-flow aligned fiber diffraction images of BeF₃^-- and AlF_x-MTs in the presence of taxol.

All these dataset images show the presence of the 4 nm layer lines and the clear absence (inset) of the 8 nm layer line.

Figure 4.. Cryo-EM of GDP-, BeF₃^--and GMPCPP-MTs.

(A) Straightened images of microtubules with 13 PFs (N) and 3-start monomer helices (S), denoted 13_3 (N_S) MTs. For each condition: raw image (left) and filtered image using the J₀ and J_N layer lines in the FFT of the MTs (right). Filtered images of 13_3 GDP-MTs display 2 dark inner fringes running parallel to the MT axis and slightly offset from it (arrow), which correspond to PFs from the top and bottom surfaces superposed in projection. In BeF₃^-- and GMPCPP-MTs, the 2 fringes make moiré patterns offset from the MT axis on one side and the other separated by blurred regions (noted ‘0’ for no internal fringes). The periodicity (L) of these moiré patterns provides a direct measure of their PF skew angle θ (Equation 8). (B) Comparison between the FFTs of the 13_3 GDP- vs. the 13_3 BeF₃^--MT (top), and the 13_3 GMPCPP- vs. the 13_3 BeF₃^--MT (bottom) in A. The monomer spacing along PFs (a in Equation 9) is given by the position of the J₃ layer line in Fourier space, that is 40.95 Å, 40.65 Å, and 42.19 Å for the 13_3 GDP-, BeF₃^-- and GMPCPP-MTs in A, respectively. (C) Blow up of the equator and ‘4 nm layer lines’ in 13_3 GDP- (top), BeF₃^-- (middle) and GMPCPP-MTs (bottom), corresponding to the boxed regions in B. In 13_3 GDP-MTs, J₀ and J_N overlap on the equator, and J₃ and J₁₀ overlap on the ‘4 nm layer line’ since the PFs are parallel to the MT axis. In 13_3 BeF₃^-- and GMPCPP-MTs, J₁₃ is away from the equator due to the PF skew. In BeF₃^--MTs, J₁₀ is closer to the equator than J₃, indicating that the PFs are left-handed (negative skew), while in GMPCPP-MTs, J₁₀ is farther apart from the equator than J₃, indicating that the PFs are right-handed (positive skew), (Chrétien et al., 1996). (D). 3D reconstructions of the 13_3 MTs in A using TubuleJ (Blestel et al., 2009). The 3D reconstructions were elongated to the same size as the original images and presented front face at a slight angle with respect to the MT longitudinal axis to emphasize the PF handedness in 13_3 GDP-MT (θ = 0°), BeF₃^--MT (θ = - 0.20°), and GMPCPP-MT (θ = + 0.30°). (E) Left: average PF skew angles of 13_3 GDP-MTs (θ = + 0.10 ± 20°, n = 15), BeF₃^--MTs (θ = - 0.27 ± 0.07°, n = 9), and GMPCPP-MTs (θ = + 0.33 ± 0.11°, n = 12). The average PF skew angle of 13_3 GDP_MTs must be lower since a majority of MTs did not show long enough moiré patterns that could be measured at the high magnification used. Therefore, their average PF skew angle is likely closer to 0°. Middle: average monomer spacing along PFs in GDP-MTs (a = 40.93 ± 0.24 Å, n = 66), BeF₃^--MTs (a = 40.82 ± 0.20 Å, n = 103), and GMPCPP-MTs (a = 42.10 ± 0.13 Å, n = 25). All N_S microtubule types were included in the analysis. Right: inter-PF monomer tubulin rise in GDP-MTs (r = 9.37 ± 0. 13 Å, n = 64), BeF₃^--MTs (r = 9.70 ± 0. 12 Å, n = 83), and GMPCPP-MTs (r = 9.45 ± 0. 07 Å, n = 24) determined using Equation 9. MTs with modified lateral interactions (essentially observed in 13_4 and 14_4 MTs in BeF₃^- conditions) were not included in this analysis (Chrétien and Fuller, 2000).

Figure 4—figure supplement 1.. additional cryo-EM data.

(A) Turbidimetric analysis of MTs assembled in BRB80, 35°C, and in the presence of 1 mM GTP (tubulin concentration 40 μM, gray), 5 mM BeF₃^- (tubulin concentration 40 μM, dotted blue), 5 mM BeF₃^- in the presence of BeF₃^--MTs seeds (tubulin concentration 40 μM, blue), and 0.1 mM GMPCPP (tubulin concentration 10 μM, orange). In the presence of BeF₃^--MTs seeds, the turbidimetric curve drops below zero due to the depolymerization of the seeds. Note the lowest depolymerization rate of BeF₃^-- and GMPCPP-MTs with respect to GDP-MTs after switching to 4°C. (B) Comparison between 12_3 BeF₃^--, 14_3 GDP- and 14_3 GMPCPP-MTs. In each panel: left: raw straightened image; right: J₀+J_N filtered image. The moiré period L provides the absolute value of the PF slew angle according to Equation 8, that is 0.53°, 0.65° and 0.53° for 12_3 BeF₃^--, 14_3 GDP and 14_3 GMPCPP-MTs, respectively. (C). Top: FFTs of the 14_3 GDP- versus 12_3 BeF₃^--MTs in B. The monomer spacing along PFs is similar in the 14_3 GDP-MT (41.16 Å) and the 12_3 BeF₃^--MT (40.86 Å). Bottom: FFTs of the 14_3 GDP- versus 14_3 GMPCCP-MTs. The monomer spacing along PFs is lower in the 14_3 GDP-MT compared to the 14_3 GMPCPP-MT (42.13 Å). (D) Blow up of boxed regions in B. Top: 12_3 BeF₃^--MT; middle: 14_3 GDP-MT; bottom: 14_3 GMPCPP-MT. J_N is systematically away from the equator since all MTs have skewed PFs. 12_3 BeF₃^--MTs have right-handed PFs since J₉ is farther apart from the equator than J₃, while 14_3 GDP- and GMPCPP-MTs have left-handed PFs since J₁₁ is closer to the equator than J₃. This analysis provides the sign of the PF skew angles, that is +0.53°, −0.65° and −0.53° for 12_3 BeF₃^--, 14_3 GDP- and 14_3 GMPCPP-MTs, respectively. (E) 3D reconstructions of the MTs in B presented face-on at a slight angle with respect to the MT axis to emphasize the PF skew. The inter-PF subunit rise r for each MT is given by Equation 9. 12_3 BeF₃^--MTs display a higher inter-PF subunit rise compared to GDP- and GMPCPP-MTs (see also Figure 4).

Figure 5.. Cap model derived from MT model systems.

GTP-bound state (BeF₃^-, blue), transition state (AlF_x, red), expanded state (GMPCPP, GMPCP, GDP-Tx, orange), and GDP-bound state (gray). (A) Schematic GTPase related conformational changes within the MT lattice. Tubulin activation upon GTP binding (T_A) induces polymerization. During assembly (MT_A), the formation of lateral contacts favors tubulin straightening, which allows GTP hydrolysis. GTPase activity drives MT through a transitional state (MT_T), where the P_i is at the nucleotide-binding site before it is released. Expansion (orange) may be an intermediate transient step between GDP•P_i and GDP states, which may facilitates P_i release (MT_E) and would be blocked in the presence of taxol. GDP-MT (MT_M) shrinks through a ‘peeling-off’ disassembly in which tubulin reverts to the curved conformation, which is inactive (T_I) in the GDP-bound state. (B) MT model illustrating specific lattice features of the GTPase cycle. This mosaic structure shows that: (i) the GTP-bound tip (blue) contains curved PFs/sheets that come together into a straight lattice due to the formation of lateral contacts, (ii) the post-hydrolysis GDP•P_i lattice (red) retains overall MT structure, (iii) hypothetically, lattice undergoes an energy-consuming expansion phase (orange) that contributes to P_i release, and (iv) in the GDP state (gray) subtle changes on the PF skew distinguish the metastable compact lattice or, (v) lattice reverts into its previous lower energy state (compaction), preventing the structure from returning to the cap architecture.