 Hello everyone. Thank you for joining online and also big thanks to the organizers for inviting me to give this talk for you today. As you might have guessed from the title already, my talk is going to be about microtubules, which are essentially the nanoscopic Swiss army knife of any eukaryotic cell. My name is Maxim and I'm happy to share our recent research about the mechanical properties and energetics of these astounding nanomachines of the cell. Even before we could observe microtubules directly by eye, we knew that microtubules are very much not like any other structure in the cell. What captures the eye when we look at them is the marvel of their dynamicity and adaptability. That is something that is not frequently observed for other polymers in general. The dynamic behavior has become one of the central foundations of cell physiology. For example, it has allowed us to develop explanations for how cells search capture and divide their chromosomes during my toes is by pulling them apart mechanically, or how cells develop polarized and asymmetric shapes and protrusions by basically pushing around cell boundaries, or even how neurons can supply the axons with nutrition using microtubules as stable characteristics. And there are many more other ways the cell can employ microtubules in fact. Now, structurally, microtubules are tiny migrant sized cell cylinders made of tubulin dimers. This feature makes them as rigid as plexiglass, but also allows us to stay largely dynamic at their tips, and the tips are where all the magic actually happens. Now, microtubules grow by stacking GTP bound tubulin dimers on top of one another, and form so-called protoflaments, protoflaments on these long tubulin strands, which then connect laterally to form the tubular lattice. And the process of tubulin stacking at the tip has the consequence that it strongly accelerates the hydrolysis of GTP to GTP inside the microtubules, and sometimes up to the point that microtubules become unstable, collapse abruptly, and shrink to zero way. This can be seen here nicely on this experimental chemograph. Now, this collapses so fast and so directional that cells evolve mechanisms to extract mechanical work from this disassembly process. And one example here is the East and one kinetical complex that couples microtubules to chromosomes doing metosis, and it is thought the elastic energy transmitted to this complex by disassembly microtubules makes it slide toward the opposite end of the microtubule, which is in a pulling force that acts directly on the chromosomes. Now, the traditional view of microtubule dynamics found in textbooks postulates that the ends of growing microtubules are blunt and straight, while those of shrinking microtubules are always strongly flared, and therefore also substantially different in shape. Now, the shape difference is the, to be precise, the change in shape is believed to make them inconsistent with growth and also lead to catastrophe in the end. However, this widely prevailing textbook picture is accumulating more and more controversies in the recent years. And from the most recent highlights, cryo electron thermography studies have actually shown that the tips of growing and shrinking microtubules are striking the similar and almost always flared. Now this paradox is raising rather old but fundamental questions in the microtubule community. Namely, if the shapes are so similar, how do microtubules know if they should grow a shrink at a given moment of time, or even more, how can they choose between pushing and pulling modes of operation if the tip structure is essentially the same. Well, now, unfortunately, marketable tips are extremely transient structures, and the timers of dynamics are not easily minimal to experiment at a high resolution. Now, because the dynamic information is hard to obtain experimentally, molecular dynamic simulations might be uniquely suitable method to breach the readout from static high resolution structural data and the conformational dynamics of tubulin aligamers at the tip. To address this question, we used atomistic molecular dynamic simulations in explicit solvent to investigate how this complex system behaves at microsecond time scales. Now we started off with accurate atomistic models of microtubule tips comprising 14 protofilaments in circumference and being roughly 60 nanometers in length and performed relaxation simulations starting from the straight microtubule structures as shown here on the slide. Now, a great advantage of using such a setup to study the microtubule plus and dynamics is that all factors determining the microtubule stability, as well as the kinetics of this plane process can be assessed simultaneously in one molecular structure. Of course, we did not perform the simulations only once, but multiple times to get the statistics about the confirmations and interactions of individual protofilaments at the tip. And when analyzing the final structures, we found several interesting features of these microtubule tip confirmations. First, microtubule tips in which all protofilaments would be perfectly straight like the textbook picture says, are highly unstable in our simulations and it doesn't matter in which nucleotide stated microtubules are currently, which is kind of consistent with the cryo and thermography data already. Now, second, the relaxation process has no obvious pattern. That means that microtubules can in principle crack open at almost any interface and then relax in multiple possible ways. So, individual protofilaments always play in clusters, so they stick together, sometimes in duplets, sometimes in triplets and so on. And these clusters seem to be structural intermediates that are no longer straight like in the initial starting structure, but are still stabilized bilateral interactions. Similarly, it seems that all the shapes of the microtubule tips are always the result of a complex competition between the elasticity of individual protofilaments on one hand, and on the other hand, the lateral interactions between them. And therefore, to get a deeper insight into the tip stability, we need to take a closer look at those components separately. Now, the first component of the microtubule tip stability, as I said before, is the dynamics of individual protofilaments at the microtubule tip and how they are regulated by the nucleotide state. As I mentioned before, also, our large scale microtubule simulations do not provide this information directly in the limited simulation time. In this end, we can see that a smaller subsystem consisting of only one short protofilament restraint at the microtubule, at the minus end, and carried out multiple molecular dynamics simulation of the system in both nucleotide states. Now, due to the stochastic nature of the simulations, they, of course, contain both slow conformational changes which are of interest for us, and fast thermal vibrations. Therefore, we use the principle component analysis to extract those degrees of freedom that contribute most to the total atomic displacement in our simulations, which is shown here on the right. Now, it turned out that only two modes are fully sufficient to account for more than 90% of the conformational dynamics in our protofilament simulations. The first mode being twist bending motion with a non-radial component, and with the second mode being a sort of tangential motion of the protofilament perpendicular to the radial plane of the microtubule, and we call this motion sort of tangential swing. Now, the principle component modes as such, nice, provide the qualitative picture about the protofilament motions, but it would be also nice to have some energetic insights, right? And so to that end, let's now use the progress along this principle component modes as reaction coordinates and project all the simulations we have onto them. So if we do this for the GDP and GDP performance simulations separately, we can estimate the underlying free energy landscapes that guide these motions in our simulations. Now, because we know where the straight protofilament conformations and the relaxed protofilament conformations are exactly located on this free energy profiles, we can estimate the energetic cost required to form a straight GDP protofilament at the microtubule tip as a function of the nucleotide state. So if we do so, we then observe that substantially more free energy is required to form a straight GDP protofilament, which correlates with previous experimental and computational studies and so we kind of knew that before. But what is more interesting is that not only the flexibility seems to change upon GDP hydrolysis, so these two values, but also the dominant motions themselves, so basically the way the protofilaments move, depending on the nucleotide state. So, for example, whereas for the GDP protofilament system, we observe roughly a 50 to 50 distribution between the twist bending and swing component. This distribution shifts to roughly 70 to 20% for the GDP protofilament system, because the GDP protofilament seems to have a much higher rigidity, tangential rigidity in this direction. So in other words, what it means for the entire microtubule tip, that basically means that GDP microtubule tips would have it actually much higher, not only to form a straight protofilament lattice because the elements are much more rigid, they are protofilaments are much more rigid, but also to form protofilament, protofilament pairs, clusters, through spontaneous collisions with neighbors, if we assume that this tangential swing motion is exactly responsible for this type of lateral counter Now the second component of the microtubule tip stability is the lateral interaction between the protofilaments. Now, unlike the tendency of the protofilaments to twist bend and splay at the tip, the lateral interactions have a counter balance and stabilization effect on the microtubule Now to estimate the interaction free energies, we performed accurate umbrella sampling simulations using again reduced simplified subsystem consistent of only two straight protofilaments, where we applied an external potential to the center of mass difference of each time repair in this protofilaments to mimic the effect of protofilaments playing in the microtubule tip. So what do we see there. Well, in both cases, we observe that the free energy raises steeply as we try to pull the protofilaments apart so it makes sense that's their native contact there And also we observe that the rupture happens at roughly the same center of mass separation at around 6.5 nanometers. But what is more interesting is that the energy required to cause the protofilaments to disengage depends strongly on the nucleotide state and in particular the contact between GDP protofilaments is by almost seven KBT stronger than the same system in the GDP state Now this might already sound very counter-intuitive, right? We know that pure GDP microtubules are more unstable, but on the other hand they seem to have more stable lateral contacts. So does it make sense? Well, such a statement is inconclusive without also considering the viscoelastic dynamics of these protofilaments at the tip and without taking together all energetic factors that are at play in the microtubule tip Now of course, our tip models as I said before are too large and too complex to obtain the full free energy landscape of the system to calculate it, but because we kind of know the elementary contributions of individual protofilaments to the total tip stability because we calculate them, we could still try to get an idea of how this complex free energy landscape might look like. And to test this, we can again consider a simplified subsystem consisting of only two neighboring protofilaments at the plus microtubule tip So for example, these two guys here. And for convenience, we can also use the same collective motions of protofilaments that I introduced several slides before. So in particular protofilaments can twist bands, so this will be this collective variable one and they can also do this tangential swing motion. This would be the collective variable to involve cases and the difference in this tangential swing parameter between the two protofilaments would then determine how likely it is that the two protofilaments would come in contact and interact with each other, given some degree of the twist banding. So if we now take two copies of the simulated protofilament ensembles, place them next to each other, according to the microtubule geometry, and then reweight their motions according to the lateral interaction energies from the previous slide, and of course due to for all possible confirmation and pairs, we would obtain the free energy landscape for the motion of this couple double protofilament system. So if you actually do so and plot the reweighted free energy landscapes. So in the United States of GDP and GDP, we can make some interesting conclusions already. So first, there is always at least two free energy minimum corresponding to states with fully split separated protofilaments, and also states with coupled, partially clustered protofilaments. And in both cases, these states seem to be separated by sort of free energy barrier located around here. But what is more interesting is that the bound nucleotide state GDP or GDP seems to affect both the barrier states between these two states, and also the relative depth of this free energy minimum. And interestingly, in this case, the GDP protofilament system would kind of more easily cross the barrier back and forth, meaning that the protofilament cluster in this case can quickly form and dissolve. But for the GDP protofilament system, it becomes increasingly harder and harder to cross this barrier, and also to populate this coupled clustered state. Now this already gives us some sort of idea how the free energy landscape for the entire market you will take might look like, namely as a complex multi dimensional well 28 dimensional free energy landscape in this particular case, with one global state corresponding to the fullest plate state where all protofilaments are separated, and then also numerous intermediate states with metastable protofilament clusters of different sizes. And then depending on the nucleotide state, these intermediate states may or may not be easily accessible. In conclusion, I would like to summarize the major findings of our work, which we believe, suggest a new mechanism of microtubule growth and catastrophe that does not require differences in the shapes of growing and shrinking microtubule used as postulated by the textbook mechanism, rather this, this new model solely relies on the statistical mechanics of protofilament fluctuations and the formation of intermediate protofilament clusters. So first, regardless of where the microtubules grow or shrink, the tips statistically alternate between states with fully explained and separated protofilaments states with metastable protofilament clusters, and states with where all the protofilaments are perfectly straight. Now these transitions are very rapid, much faster than tubulin binding unbinding or even GDP hydrolysis, explaining why it is so hard actually for the experimentalist to observe this fluctuations at a high resolution. The content, well, to be precise, the average content of GDP tubulin at the microtubule tip controls the probabilities of adopting one of those, well, actually any of those states, but not only that, also the transition rates between the states as we saw for the protofilament system in the previous slide. So in other words, microtubules are on in this in this model in this formalism are only able to grow as long as their collective probability to form the straight lattice is relatively high. And only if the transitions from this fully split state via this metastable states are frequent enough so that new dimers can find the straight structure stabilize it and elongate the existing microtubules otherwise this full structure is doomed to collapse. And so with this, finally, I would like to thank my lab and in particular Hannah McGriffin enough and making this work possible, but also Evan Nogales and her former PhD from a postdoc who is young for Kylie providing and for useful discussions on microtubule model construction, and of course you for your attention, and I will be happy to take any few questions.