Alp7/TACC-Alp14/TOG generates long-lived, fast-growing MTs by an unconventional mechanism

Alp14 is a TOG-family microtubule polymerase from S. pombe that tracks plus ends and accelerates their growth. To interrogate its mechanism, we reconstituted dynamically unstable single isoform S. pombe microtubules with full length Alp14/TOG and Alp7, the TACC-family binding partner of Alp14. We find that Alp14 can drive microtubule plus end growth at GTP-tubulin concentrations at least 10-fold below the usual critical concentration, at the expense of increased catastrophe. This reveals Alp14 to be a highly unusual enzyme that biases the equilibrium for the reaction that it catalyses. Alp7/TACC enhances the effectiveness of Alp14, by increasing its occupancy. Consistent with this, we show in live cells that Alp7 deletion produces very similar MT dynamics defects to Alp14 deletion. The ability of Alp7/14 to accelerate and bias GTP-tubulin exchange at microtubule plus ends allows it to generate long-lived, fast-growing microtubules at very low cellular free tubulin concentrations.


A simple Alp14 model
The following model attempts to account for S. pombe microtubule dynamics in the absence and presence of Alp14. The constraints on the model are the following: • Microtubule growth in the absence of Alp14 (Figure 3 in main text) • Average microtubule growth in the presence of Alp14 (Figure 4 in main text) • Maximum microtubule growth with Alp14 and 6 µM tubulin is ~35 nm s -1 , independent of Alp14 concentration (including for de novo nucleated MTs) • Growth at 0.5 µM tubulin of 12 nm s -1 in the presence of Alp14 ( Figure 5) • At least one Alp14 per MT plus end in the Alp14 concentration ranges of the experiments Table S1. Rate constants for tubulin exchange at MT plus ends.
Microtubule polymerisation in the absence of Alp14 is modeled with a simple bimolecular reaction 1 , where T is a (GTP) tubulin dimer, M is the microtubule plus end, and (M + 1) is the microtubule plus end with an additional tubulin ligated (see Figure 6 left, main text). The forward rate (k 1 ) and backward rate (k -1 ) are fit to microtubule growth of single isoform S. pombe tubulin in PEM at 25°C in the absence of Alp14, with k 1 = 5.5 µM -1 s -1 and k -1 = 6.6 s -1 (see Figure 3A, main text).
With Alp14 present, a second, Alp14-facilitated pathway is added to the model ( Figure 6 right, main text),

Rate
where E is an Alp14 enzyme bound to the plus end of the microtubule and TE is a 1:1 tubulin-Alp14 complex bound to the plus end of the microtubule. The version of this model which we use has Alp14 able to move between protofilaments during growth, effectively averaging its effect across the entire MT plus end. We view this as justified by the data in Figure 4E: increasing Alp14 resident at plus ends results in higher growth rates. A fit using this model is shown in Figure 4, main text.
In principle, an even simpler Alp-14 facilitated pathway could combine the rates k ±2 and k ±3 into single accelerated on-and off-rates, given by This simpler pathway must be excluded due to the following experimental observation (Figs. 4,5 main text): At 0.5 µM tubulin, Alp14 microtubules grow at 12 nm s -1 (20 dimers s -1 ), implying an on rate of at least 40 µM -1 s -1 (higher when k off is included). At 6 µM, eq. 2alt would then predict at least 140 nm s -1 , much faster than the observed maximum of 35 nm s -1 and average of 10-20 nm s -1 . Hence, we must include the more complicated pathway (eq. 2), and would expect fast Alp14-tubulin complex formation (a fast k 2 ) followed by a slower incorporation of tubulin into the microtubule (k 3 ), which becomes rate-limiting at higher tubulin concentrations to explain the observed maximum.
This model ( Figure 6, main text, and equations 1 and 2) generates a set of five coupled differential equations, one each for T, M, (M+1), E, and TE as defined in (1) and (2) above. To solve, we take the perspective of a single microtubule plus end and assume steady-state: the growth rate is constant, the fraction of Alp14 occupied with tubulin and free from tubulin are constant, and the concentration of free tubulin remains constant. The five differential equations then simplify into a system of three linear equations: = (! !"! ) (total plus-end bound enzyme) These three linear equations are then solved for unoccupied enzyme E, occupied enzyme TE, and microtubule growth rate d(M+1)/dt via matrix inversion: The first term on the right-hand side corresponds to the Alp14-facilitated pathway, the second corresponds to the Alp14-independent pathway, with pf free the fraction of protofilaments without bound Alp14. This final term is exactly 1 for a model in which a given protofilament can simultaneously undergo Alp14 mediated and Alp14 independent GTP tubulin exchange; the resulting fit parameters remain almost unchanged (k 3 differs by approximately 10%).
How many Alp14 are bound to a microtubule tip? First, note that the data ( Figure 4C, main text) require some Alp14 acceleration at Alp14 concentrations as low as 6 nM. This sets an upper limit on the affinity of Alp14 for microtubules of roughly K D, Alp14-MT ~ 6 nM in order for at least a single Alp14 to be bound to the microtubule. On the other hand, 0.5 µM tubulin induces microtubule growth of 12 nm s -1 in the presence of Alp14, which requires a K D, Alp14-MT > 5 nM given the constraint of fitting to the Alp14-facilitated growth data of Figure 5A, main text. At lower K D values, the fit parameters for k ±2 and k ±3 to the 6 µM tubulin data cannot induce growth fast enough at 0.5 µM tubulin. Thus, the affinity of Alp14 for the microtubule tip must be strong, K D, Alp14-MT ≈ 6 nM.
How do the experimental data constrain the rates k ±2 and k ±3 ? k 3 is constrained by the linear effect of Alp14 concentration on microtubule growth at a constant tubulin concentration (Figure 3). This results in k 3 ≈ 6 s -1 ; deviations of more than 20% cannot be compensated by altering k ±2 and k -3 .
On the other hand, k ±2 and k -3 are interdependent in fitting the growth rate. As mentioned above, the constraints are growth as a function of Alp14 concentration at 6 µM tubulin, a maximum growth rate of ~ 35 nm/s at 6 µM tubulin (presumed to correspond to fully loading the MT tip with Alp14), and a minimum growth rate of 12 nm/s at 0.5 µM tubulin for a fully loaded MT tip (that is, 13 Alp14 per microtubule plus end). The one constraining parameter is K crit, 2 , the critical concentration of tubulin for microtubule growth in the presence of Alp14. This is ~0.08 µM, a 15-fold reduction from the 1.2 µM critical concentration for S. pombe tubulin without Alp14.
Thus, these three parameters (k ±2 and k -3 ) are under-constrained. Assuming k 2 = 50 µM -1 s -1 (just about the minimum of 40 µM -1 s -1 leads to a range of values for k -2 and k -3 of ~ 1 to 15 s -1 , with the two values inversely related. The final point of comparison is microtubule catastrophe. There are competing models for the details of microtubule catastrophe, but the consensus is that catastrophe follows breaching of the GTP-tubulin cap at the plus end of the microtubule. To be consistent with the increased catastrophe at low tubulin concentrations, then, Alp14 must increase the effect of cap breaches at a low solution tubulin concentration (0.5 µM tubulin shows a high catastrophe rate), but decrease the effect of cap breaches at a high solution tubulin concentration (6 µM tubulin shows a decreasing catastrophe rate with increasing Alp14 concentration).
The model predicts the rate of GTP-tubulin removal required to create a (single protofilament) gap at the end of the microtubule, the lifetime of such gaps, and the average number of gaps per microtubule tip under a given set of conditions. The quantitative predictions depend on the specific k ±2 and k ±3 values chosen, but the qualitative results remain consistent across the range of k ±2 and k ±3 values discussed above. Sample predictions are shown in Figures S1 below.   Percentage of time spent in each phase of MT dynamics.
Mean ± SEM; n, number of cells.