Lattice defects induced by microtubule-stabilizing agents exert a long-range effect on microtubule growth by promoting catastrophes

Microtubules are dynamic cytoskeletal polymers that spontaneously switch between phases of growth and shrinkage. The probability of transitioning from growth to shrinkage, termed catastrophe, increases with microtubule age, but the underlying mechanisms are poorly understood. Here, we set out to test whether microtubule lattice defects formed during polymerization can affect growth at the plus end. To generate microtubules with lattice defects, we used microtubule-stabilizing agents that promote formation of polymers with different protofilament numbers. By employing different agents during nucleation of stable microtubule seeds and the subsequent polymerization phase, we could reproducibly induce switches in protofilament number and induce stable lattice defects. Such drug-induced defects led to frequent catastrophes, which were not observed when microtubules were grown in the same conditions but without a protofilament number mismatch. Microtubule severing at the site of the defect was sufficient to suppress catastrophes. We conclude that structural defects within the microtubule lattice can exert effects that can propagate over long distances and affect the dynamic state of the microtubule end.

Microtubules are cytoskeletal polymers that rapidly switch between phases of growth and shortening, and this behavior, termed dynamic instability, plays a crucial role in the formation, maintenance, and reorganization of microtubule arrays during cell division, migration, and differentiation (1, 2). The transition from growth to shrinkage, an event called catastrophe, is known to occur when the protective cap of guanosine triphosphate (GTP)–bound tubulin subunits is reduced or lost, but the underlying mechanisms are still the subject of investigation (3, 4). One interesting property of microtubules is that the frequency of catastrophes depends on microtubule age: Microtubules that are growing for a longer time have a higher chance to switch to depolymerization (5, 6). Changes occurring at the microtubule end, such as loss of individual protofilaments or end tapering, have been shown to promote catastrophe (7–9). In principle, it is also possible that the catastrophe frequency at the plus end is affected by structural features in the microtubule lattice farther away from the tip, but this possibility has so far remained untested.Structural studies have established that tubulin can form tubes with different protofilament numbers (10), dependent on the species, nucleation template, presence of different microtubule-associated proteins, and other properties of the polymerization reaction (e.g., ref. 11; reviewed in ref. 12). An important consequence of the structural plasticity of the microtubule lattice (13) is the formation of lattice defects, such as sites where a microtubule gains or loses one or more protofilaments (11, 14–17). A recent cryoelectron tomography analysis showed that in some cell types, such as Drosophila neurons, variations and transitions in protofilament number are readily detectable (18) and are thus likely to be physiologically relevant. Switches in protofilament number can be introduced during microtubule growth, and their presence may affect microtubule dynamics in different ways. For example, defects can be repaired through tubulin incorporation, and the resulting islands of GTP-tubulin can trigger microtubule rescue (19–22). On the other hand, the presence of defects could potentially also induce catastrophes (as proposed in ref. 14), since conformational properties of the microtubule lattice might propagate over some distances (23).To study the relationship between lattice defects and microtubule catastrophes, one should be able to directly correlate the presence of defects with the dynamics of microtubule ends. We recently found that fluorescent analogs of microtubule-stabilizing agents (MSAs) can be used to induce microtubule lattice defects that can be visualized by fluorescence microscopy. When present at low concentrations, MSAs preferentially bind to microtubule plus ends that enter a “precatastrophe” state (24), which is manifested by the gradual loss of the GTP cap and reduced recruitment of end-binding (EB) proteins that detect GTP-bound microtubule lattice (25–27). Strong accumulation of MSAs at precatastrophe microtubule ends leads to the formation of stabilized patches of microtubule lattice, where the tube is incomplete and keeps incorporating GTP-tubulin but is not fully repaired (24). When microtubules switch to depolymerization, such persistent lattice defects, which coincide with the hotspots of MSA binding, can induce repeated rescues and, therefore, they were termed “stable rescue sites” (24).Here, we used MSA-induced lattice defects to address two questions. First, what prevents complete repair of an MSA-induced persistent lattice defect? And second, does the presence of such a persistent defect affect the dynamics of the microtubule plus end? Since different MSAs are known to affect the number of protofilaments (15, 28–33), we hypothesized that persistent lattice defects could be associated with the changes in protofilament number and thus could not be fully repaired for geometrical reasons. We tested this idea by generating stable microtubule seeds with one MSA and then elongating them in the presence of another MSA, with the same or different preference for protofilament number. Use of fluorescent MSAs allowed us to directly follow drug binding. We found that precatastrophe microtubule ends accumulated MSAs in all conditions; however, the outcome of drug binding was different. When there was no mismatch in protofilament number between the seeds and the elongation conditions, drug accumulations were short in duration and length, and microtubule growth beyond such sites was processive. In contrast, when, based on the MSA properties, a mismatch in protofilament number could be expected, large and persistent drug accumulations were formed. The existence of such mismatches was confirmed by cryoelectron microscopy (cryo-EM) and by measuring microtubule growth rate, which became higher with increasing protofilament number. When microtubule ends extended beyond a mismatch-containing lattice defect, they displayed elevated catastrophe frequency. Laser-mediated severing of a microtubule at the site of the persistent defect reduced catastrophe frequency at the plus end. Our data demonstrate that local perturbations in microtubule structure can affect the state of the dynamic end at a distance of several micrometers.