Beyond the Tip: Lattice Dynamics, Seams, and the Mechanism of Microtubule Fracture

TL;DR

The paper addresses bulk GDP microtubule lattice fracture and the role of multi-seam structures arising from monomer defects. It develops a kinetic Monte Carlo lattice model at the monomer scale, incorporating seam dynamics and vacancy-induced asymmetries, and calibrates the timescale against plus-end depolymerization to reproduce experimental fracture metrics. The results show seams accelerate longitudinal damage and that the intrinsic lattice anisotropy is modest, with $A \approx 1.5$ (potentially $A \approx 1.2$ in the presence of defects) and a GDP-lattice binding energy around $\Delta G_b \sim -45$ to $-50\,kT$. These findings challenge higher anisotropy predictions from tip-growth models and underscore the need to incorporate lattice dynamics and vacancy defects into growth theories to accurately describe microtubule assembly and fracture.

Abstract

The structural integrity of microtubules is paramount for cellular function. We present a theoretical analysis of their lattice fracture, focusing on the influence of multi-seam structures arising from monomer defects and aiming to provide a more accurate estimation of GDP lattice parameters. Our findings reveal that seams function as pre-existing pathways that accelerate damage propagation. Consequently, monomer vacancies destabilize the lattice due to the inherent structural loss of tubulin-tubulin contacts and the additive acceleration of fracture through multiple seams. Importantly, the comparison of our simulations with experiments on lattice fracture suggests that the intrinsic ratio of longitudinal to lateral binding energies is bounded at approximately 1.5, challenging previous predictions of lattice anisotropy from tip-growth models and highlighting the urgent need to incorporate into current growth models parameters obtained from lattice dynamics.

Figures (5)

• Figure 1: A: Image sequence showing a MT developing a damaged region, which is visible due to reduced fluorescence resulting from the loss of tubulin from the lattice. The MT eventually broke along the softened region (marked by the white arrow) and disassembled. Scale bar, $2\,\muup$m. See Ref. Biswas2025 for experimental details. B: Survival times of end-stabilized microtubules (experimental data from Refs. Triclin2021Biswas2025). C: Size of damage at fracture (experimental data from Ref. Biswas2025). Shown is the median and the interquartile range (IQR). The gray shaded regions in (B,C) indicate the range of admissible values we chose for comparison with our simulation results.
• Figure 2: A: Three dimensional model of the canonical microtubule structure (13 protofilament, 3-start helix). Monomer vacancies (horizontal red arrows) lead either to a lateral shift of an existing seam or introduces two new seams. Seam positions are marked by vertical red arrows and red dashed lines. B: Schematic of the kinetic Monte Carlo model illustrating the lattice structure. The dimer off-rate constant $k_{mn}$ (see Eq. \ref{['ec-detachment_rates']}) depends on the occupancy of neighboring lattice sites. Seam structures are marked as in (A). The transition between protofilaments $i=13$ and $i=1$ corresponds to a periodic boundary with a vertical shift of 3 monomers (as indicated by the blue annotated arrows) to assure the 3-start helical lattice structure.
• Figure 3: Path to fracture in multi-seam microtubules. A-C: Vacancy growth from a dimer vacancy (A) and a monomer vacancy situated at an existing seam (B) or in the B-lattice (C). Left: Illustrations of the lattice configurations in the vicinity of the growing vacancy at different propagation stages of vacancy growth. The colors distinguish the dimers by number of longitudinal and lateral neighbors. Right: Exemplary kymographs of the fracture process. The transition of seam structures are marked by white arrows. The $y-$axis represents the longitudinal axis of the microtubule and the color scale indicates the number of intact protofilaments. D,E: Examples of the longitudinal damage size at fracture (D) and time to fracture (E; in units of $\tau$) as a function of the initial lateral position (in protofilaments) of the initial defect. Symbols and error bars indicate the mean and standard deviation of the sample (1000 realizations). Parameters in (A-E) are $\Delta G_\mathrm{b}=-45\,$kT and $A=1.7$.
• Figure 4: A,B: Damage size at fracture ($L_\text{f}$, A) and time to fracture in units of $\tau$ ($T_\text{f}$, B) depending on the lattice binding energy $\Delta G_\text{b}$ for various lattice anisotropies $A$ (see legend). MTs are either defect free (perfect lattice) or contain randomly distributed monomer defects (spatial frequency 0.1 $\muup$m$^{-1}$). Shown is the median and the IQR range over 1000 simulations. The gray regions indicate the experimentally relevant range (see Fig. \ref{['fig-model_fracture']}B,C). C,D: Lattice parameters $\left(\Delta G_b,A\right)$ that reproduce the experimental fracture data for defect free MTs (perfect lattice, C) and MTs containing randomly placed monomer defects (spatial frequency 0.1 $\muup$m$^{-1}$, D). Green (light and dark) regions indicate parameters that reproduce ($L_\text{f}$); dark green regions indicate parameters that reproduce $L_\text{f}$ and $T_\text{f}$, assuming a tip depolymerization speed of 10-30 $\muup$m.min$^{-1}$ for the calibration of the time scale $\tau$.
• Figure 5: Simulated plus-tip depolymerization speed in units of $\muup$m/min depending on the binding energy for various values of the lattice anisotropy $A$ as indicated in the legend.