 ... for a change in the program, so Eric May, that was supposed to give the talk just before the last, will be here on Friday, is not here today due to problem with his flights, so his talk is moved to the first on Friday morning, so Mike's talk will be the first one and last after the coffee break. Ok, so without further ado we have the talk by Mauricio Mateo, we talk about assembly pathway of a small spherical virus capsid. Thank you. Good afternoon everybody, first of all I would like to thank the organizers for this opportunity being here with you and in this very, very nice workshop. So I will start by telling you that I am a biochemist who has been interested since many, many years ago in viruses, in particular in trying to relate their structure, their structure and with their functions. And to do that we have been using different biological, biochemical and biophysical techniques. Since about 20 years ago my group has been interested mainly in studying assembly and disassembly biophysical properties, and this is mainly because I am here, and conformational stability and dynamics of vise particles. And we are currently using four different virus models to do this depending on the specific question we would like to ask. We are more interested in basic knowledge, in trying to understand how viruses work, but we have also some view on trying to use this knowledge to develop vaccines, antivirals, new biomaterials and nanoparticles. So, during this meeting Alex Balwena, who is a senior postdoc in my group, is presenting a poster on what we think is a very interesting experiment or study in assembly, kinetics, dynamics, mechanical properties of the protein lattice that makes the mature capsid of the human immunodeficiency value. So, if you are interested, please see Alex's poster and he will gladly explain to you this part of our work. But in this talk I will focus on some of our recent results using the minute virus of mice, or sorry about that, or NVM for short. So, in the first part of my talk I will talk about assembly and disassembly in vitro, and this work has been done mainly by Maria, who is a PhD in my group, with essential contributions by Alex Balwena again, and Miguel Angel Fuertes here. And in the second part of my talk I will talk about some relationships between mechanical properties and biological functions in these same values, and this is related with the assembly process. And this work has been done mainly by Pablo Carrillo, who already defended his PhD in my group, again with the help of Maria, Alex, and Alicia here has helped in both types of study. I must say that Alex and Miguel Angel are physicists, Maria and Alicia and Pablo are biologists, and as I said I am a biochemist, so in my group we have very interesting discussions in those occasions when we manage to understand each other, which is not always the case. So, I will start with the first part of the talk. We are, as I said, interested in studying the pathway of self assembly of a very, very simple spherical virus capsid. And this is because these steps are two comparatively poorly known stages of the viral infectious cycle, and also because it is very clear that they constitute very interesting targets for developing new antiviral drugs. So, the one problem to do this is that understanding self assembly of virus capsid requires a lot of knowledge on the assembly pathway through the identification of as many assembly intermediates as possible. And fortunately for structurally simple virus capsid, assembly intermediates and pathways have been predicted in much detail using theoretical approaches. For example, Dennis Rappaport here mentioned some early studies from his, in which he managed to predict a lot of details on the assembly of icosahedral T1 capsid. I will refer to this later on. So, for example, using, as he did, using coarse grain molecular dynamic simulations. The problem is that these predicted pathways for assembling simple viruses are frequently very difficult to confirm experimentally because of the very transient existence of most intermediate states. Usually when you look experimentally to this process, you normally see assembly of simple virus capsid, very simple virus capsid as two state process. So, in this study we attempted to use a combination of AFM and electron microscopy EM to experimentally detect transient assembly, this assembly intermediates for most simple T1 icosahedral capsid, that of the minute virus of mice, MVM. Before I start with this, telling you about this study, I must explain some facts on MVM, which are important to help to interpret the results. So, first MVM is one of the smallest and structurally simplest viruses known. It's 25 nanometers in diameter, the icosahedral T1 capsid is composed of 60 structurally identical protein subunits. There are no conformational switches, there are no quasi equivalents. The capsid proteins associate in the first stage into stable trimmers in the cytosol. And these trimmers are transported into the cell nucleus, and 20 of these trimmers in the nucleus constitute the stable building blocks, or CBVs, capsid building blocks, which each capsid is self-assembled in the cell nucleus. And I must stress this thing, this part here, there is no co-assembly of capsid and viral nucleic acid. In this case, more or less exceptionally, the viral nucleic acid is single-stranded DNA, not RNA, single-stranded DNA. Only after the capsid is assembled in the cell nucleus, then the single-stranded DNA is packaged through one of the capsid pores at the 5-fold symmetry axis. So, what we did first is try to disassemble and then reassemble the capsids. We purified capsids formed in the cells, and we had to, first of all, to dissociate these capsids into as much as possible into CBVs, into these trimeric subunits, and then we had to reassemble them. But not only that, we had to manage to slow down the process in order to be able to visualize the intermediates, because if you use, for example, a lot of the nature and agent, like Guanyidim in chloride, what you end up is you have either capsid building blocks or complete capsid, but nothing in between because they are very transient. So, to make a long story short, finally, we got a very simple procedure in which we had a moderate amount of Guanyidim in chloride, and then we managed, if we incubate for a certain amount of time, we managed to almost completely dissociate these capsids here, and then just by removing the dissociating agent by dialysis or by gel filtration, depending on the experiment, we managed to reassociate these capsids and efficiency under appropriate conditions of pH and ionic strength, which are normal physiological conditions, pH around 7, and ionic strength about 0.15 molar, sodium chloride, you managed to obtain reassembly efficiencies that can reach up to 80%. So, most of the capsids are reassembled. And the reassembly process followed a sigmoidal kinetics that was dependent on protein concentration as expected for a nucleation and growth process, and by using different techniques, we found that the reassemble capsids, for example, these are AFM images with high profiles, EM images, etc., these reassemble capsids are indistinguishable from the original ones. So then we look at the profiles after this assembly before reassembly, and we observed only small peaks, whose height corresponded to those expected for trimeric CVVs, but to make sure that these were only trimeric CVVs, we did a gel filtration analysis, before this assembly and after reassembly, we got a single peak with the size corresponding to capsids as expected, and after this assembly we got a single peak with the size corresponding exactly to trimers of the capsid subionics, these CVVs, these capsid building blocks. So we now could explore the intermediates between these trimeric CVVs and the complete capsids. And we obtained many, many images during this assembly and also during the reassembly processes at different times, and we analyzed these images, and it's important to notice that we both used EM, like in this case, and also AFM, because they are complementary techniques that allow a much thorough characterization of the types of intermediates we observed. So by looking at the EM images, we can see here that you can observe different incomplete capsids that goes here in this assembly for almost complete to very small fragments, and when you look at the difference in the angles subtended by these arcs, the difference between one and the other could be just one trimer. So it appears that you have here almost all possible intermediate course. It's just a simplification, but there are no discontinuities in the process, as far as we can tell. When we look at reassembly process, most interesting one, of course, we observed exactly the same types of intermediates. I don't have time to go into details, but we didn't see any difference by looking at this assembly or reassembly. Then we had to put some numbers on this, and we took a lot of images, and we did countings of different types of particles. We grouped the particles into different classes, for example, sorry, again, this could be class one, class two, class three, and we counted the number of particles as a function of time, and what we observed as we expected is that smaller fragments during this assembly occurred later, for example, these fragments in red, this is the curve, and larger fragments in blue occurred earlier during this assembly, and the opposite was true when we explored assembly. Smaller fragments occurred first, red aligned here, and larger fragments occurred later, blue aligned here. But EM is not able to characterize in enough detail what types of intermediates you have during the process, so we resorted to AFM, just as one example, by looking at topography and the height of different particles, and taking advantage of the spikes, the protrusions that this virus has at the threefold symmetry axis, you could see here that there is a trimer missing. This is a complete particle, you have one, two, three, four, five protrusions around a fivefold symmetry axis, and you have only four here, so one of the trimmers is missing. And by doing this with many, many particles again, during different times, during reassembly and during this assembly, you can see this with detail, but I can just explain you that during this assembly we managed to observe complete capsids, capsids missing just one trimer, one CVV, capsids probably missing one pentamer of CVVs, capsids missing more than one pentamer of CVVs, isolated pentamers of CVVs, looking at the height, topography, and even smaller fragments that by doing different analysis, we know that are CVVs, free CVVs. And when we looked at the reassembly process, we observed again exactly the same types of intermediates. And this reversible process was very efficient, sorry, I forgot about that, is just to check that by counting different types of particles using AFM instead of EM, we have the same results. So this assembly experiment in which different types of particles observed by AFM are counted, and again, particles that are more complete occurred earlier than particles that are less complete as expected during this assembly. Yes, it's only trimmers. Why? Because monomers are not stable, are not folded. The trimmers, well, the monomers have very, very long loops that intertwined to form a trimer. If you just dissociate the trimer, the protein is unfolded. We have trimmers in solution, yes. Sorry, two? Yes. No, no, no, we didn't that. I mean, I understand. Of course, it could be done, but we didn't that. So what I was telling you is that this reassembly process is very efficient. It's reversible, and occasionally, but only very occasionally, you have some off-pathway intermediates, like these filaments here, but they are extremely rare. So, in conclusion, MVM assembly may proceed as a nucleation and growth process in which trimmers are sequentially added to the growing capsid. Pentamers of trimmers, and capsids missing just one pentamer of trimmers, or only one trimmers, are conspicuous assembly intermediates. The reversible assembly disassembly process appears to follow the principle of microscopic reversibility with exactly the same types of intermediates being observed in either direction. And, very importantly, this study provides experimental verification of many features and many, many details that were predicted long time ago on the virtual self-assembly of an empty T1 icosahedral capsid from 20 triangular subunits, exactly like our actual trimmers. And this was done, as predicted by molecular data, this was done, as I said before, by Denis Rapaport, who is here with us. So, going to the second part of my talk, I would like to tell you something about our analysis of structural determinants and biological relevance of a mechanical property of virus capsids, as you have heard Pedro and Bouter telling very interesting things about this, using very different viruses. We were interested in trying to relate precise structural determinants and the possible biological relevance of capsid stiffness, in particular, using NVM as a model of a very simple virus. And in this talk, I will talk about the role of amino acid residues, the mechanical role and the biological role of amino acid residues that are specifically located at interfaces between protein subunits. And specifically between trimmers. So, before that, I must say that we have been studying, during a lot of time, about 12 years ago, relationships between virus atomic structure, mechanical stiffness and biological function. And we started to do this because we started a collaboration with Pedro de Pablo's group and we have been doing many experiments and many studies together. I'm not going to talk on these collaborative studies today, but we did a lot of them. And for us, it was also very important that Pedro and his group taught us how to determine, how to use an AFM and how to determine mechanical properties of viruses. So, all the work we have been doing, trying to relate mechanical properties with other biology and structure and so far and so on, have been because of this collaboration that we started with Pedro many years ago. So, what we specifically do to do this type of studies is, first of all, we inspect the atomic structure of the virus. In this case, in VM, we have been using other viruses as well. We revise previous structural biological information and make some hypothesis on the possible effects of amino acid substitutions on stiffness and structure and or function, depending on the specific study. Then, we genetically engineer capsids or virions carrying individual point mutations. So, one per capsid subunit. In this particular case, we have mutated amino acid here to alanine, this one, this is an asparagine, we have mutated it to alanine, so we have removed just one amide group in each capsid subunit and then we check mechanical properties, function, structure, whatever we are interested in in this particular case. So, we use AFM imaging to determine the orientation of individual particles. As I said, this is very easy in this particular virus because we can clearly see these protrusions. This particle, for example, is oriented with a 5-fold symmetry axis on top. We perform non-inventations with AFM tip, just as Pedro and Bouter told us. And compare the elastic constant at different regions of the capsid, regions around 2-fold symmetry axis as 2, 3-fold symmetry axis as 3 or 5-fold symmetry axis as 5. And then we analyze these changes in mechanics and, as I said, we relate them to other properties or structure or biological function or whatever. So, by doing this so far, we have determined the effects on mechanical stiffness as many as 34 biologically relevant point mutations. Each one or each of these remove or replace individual chemical groups, as I exemplified before, in the virus capsid. For example, looking at DNA banding sites or around capsid pores or at interzability interfaces. And I will refer here to this last study in this study is still ongoing. We asked whether different amino acid side chains that establish non-covalent interactions between capsid subunits could have different roles on various mechanics. So, what we did is to choose individual side chains in particular 12 at intertrimer interfaces shown here in green. This is a trimer, this is a pentamer of trimer. You can see here the interfaces between the trimers and the selected residues to be mutated to allanine to remove these groups. And these groups are involved in different types of non-covalent interactions like hydrogen bonds, carbon-carbon hydrophobic contacts, vulnerable interactions, a combination of several of these interactions together, etc. As I said, we trunked them to allanine and purified the capsids, the mutant capsids and compare their stiffness with the natural capsid stiffness. And these are the results I won't go into any details. These are just histograms showing the elastic constants and comparing for different axes of symmetry, different mutants in each case comparing the results with the wild type, the reference results. But what I would like to do is just to summarize the general results of this study by saying first that we found, contrary to what we hoped, we found no relationships between variations in stiffness and either type of chemical group removed, irrespective of if it was a polar or polar, charge or uncharge, small or large, etc. We didn't find any correlation, any relationship between type, between stiffness and type of estimated strength of inter-sabionite interactions removed. We didn't find any relationship between distance, between stiff and capsid regions and the mutated residue, but we did find a clear relationship between increased stiffness at S2 regions at the interfaces here and also S3 regions at the center of its trimer far from the interfaces and the location of the mutated residues in structural elements of the capsid. In particular, mutations in secondary structural elements at the core of each inter-trimer interface are buried in the central part of each interface invariably led to a higher stiffness and average stiffness increase was as much as 50%. In contrast, mutations also at the interfaces, these first ones are the red ones and the mutations at exposed loops protruding from the main inter-trimer interfaces not at the cores, but at the interfaces and involved also in interactions, these ones had a very low effect on average stiffness and the average stiffness increase was only 8%. Moreover, most mutations stiffened S3 and S2 capsid regions and virtually no mutations reduced capsid stiffness at any region. So, these mutants are not natural mutants, these mutations are in conserved residues and in nature these residues don't change. We forced them to change and the invariant result was that either the stiffness was the same or it increased even dramatically. In no case, we found a mutation that was able to reduce capsid stiffness and also every mutation that the stiffness S3 regions and all but one that the stiffness S2 regions in the capsid were invariably associated with reductions in virion infectivity. So, we hypothesized, of course this is a correlation, I know, in biology many things are only correlations because it is very difficult to prove a cause effect relationship. So, we hypothesized that the linkage exists between a stiffened S2 and S3 and S2 capsid regions and reduce virus infectivity and we challenged this hypothesis and we verified it by analyzing the effect and capsid stiffness of an independent set of four unrelated mutations that we knew beforehand that reduced virus infectivity but were not located at inter-trimer interfaces but at a different place at intra-trimer interfaces. So, we predicted that if by stiffening these regions, we are in some way reducing infectivity this infectivity reducing mutations should stiffen the capsid for all of the cases. But then we had to explain how this stiffening effect could be related to a reduction in infectivity and what we found to make again a long story short is that the mutations that the stiffened S3 and S2 capsid regions invariably impaired capsid assembly. This suggests the mechanism by which stiffening these regions may impair virus infectivity. What we think is very simple is that we should not be able to associate with each other with a high affinity. If you re-jedify, if you increase the stiffness of the trimer and this you can look at the center of the trimer or at the periphery then trimers won't fit so well in assemblies impaired and because assemblies impaired the viral titers are reduced. So, the last thing how much time do I have? We asked was what structural changes elicited by these stiffening mutations are responsible for this mechanical effect. So, we started to address this question by analyzing the structural effects of one mutation only. This one I referred to before as per gene to alanine mutation at position 170 here close to a capsid pore we chose this mutation just as a representative example of all of the mutations we have been studying and all of them were stiffening the capsid. So, we knew that this particular mutation stiffens the capsid both locally at the capsid pores but also at regions distant from the mutation side and as a part of a collaboration with us, Nuri Aberdaguer's group in Barcelona and we compared it with the crystal structure of the wild type capsid which has been solved before. So, the result was that the mutation causes quite subtle but significant structural differences not only locally here at S5 at the pore regions but also in many other structural elements located at distant capsid regions including the S3 regions and the S2 regions regions are on the two-fold axis. Moreover, when we look at the differences in structure between the mutant and the wild type capsid if the movie works you can see something that we think is very interesting here is the wild type structure is more expanded and the other one is the mutant structure is more compacted and the ideal difference between the two there is a structural compactness and we think that this compaction increases the interactions and increases the stiffness not only at the pore sites at the site of the mutation but everywhere in the capsid in addition we looked at the B factors at the temperature factors of the two structures of the two crystal structures we found that the wild type is very flexible and when we compare the B factors the average B factors sorry, the normalized B factors of the wild type with those of the mutant we found that the mutant has lower B factors precisely in these flexible regions in the wild so I think this is another way to see, to look at the same thing we are rigidifying the capsid we are restricting restricting the movements because we are compacting the structure where increasing the interactions and this is something that affects depending on the mutation different biological function in the case of the mutations we have tested in the study I have mentioned before is because this loss of flexibility impairs assembly and because of that infectivity so just to conclude this second and last part many amines and acid side chains buried at interzabunit interfaces in the natural ambient capsid contribute to keep a low stiffness at S2 and S3 capsid regions the limited stiffness of S3 and probably S2 regions may facilitate interzabunit association during capsid assembly fast enhancing virus yields two stiffening mutations the only ones we tested so far led to significant structural differences and reduced B factors in many capsid elements and to a subtle overall compaction of the capsid I explained the results with one mutation only but we did the same with the second mutations and the results were very similar and finally very very small chemical changes causing subtle structural differences can lead to large biological harmful variations in capsid stiffness the last slide I would like to briefly comment on is a summary on the relationship between S2 and S3 but also previous results from our group some of them obtaining collaboration with Pedro de Pablo and what we can say is that NVM despite its remarkable structural simplicity appears to have a complex a very complex mechanical behavior first an empty capsid with a biologically restricted stiffness at different regions at interzabunit interfaces at poor regions to facilitate the biologically required through poor translocation of functional components this has been published before and secondly a video I mean a capsid with the DNA inside in which segments of the DNA genome are used as buttresses to locally stiffening specific regions relative to the anti-capsid and during that this stiffening is able to impair a heat induced values inactivating transition and the flexibility of other regions the poor regions that are still required to allow through poor translocations that are needed for effective so that's it thank you very much for your attention