 Hello everyone again. This is the short talk and in this short talk as I already mentioned in the previous longer talk I will mainly concentrate on the protein protein binding and the applications of the chemical calculations for the calculation of relative free energy differences upon amino acid mutation in the context of protein protein binding and for that we will use exactly the same methodology so mainly chromax-based MD sampling and PMX for the setup of these calculations and the only thing that we need to take into account now is that the thermodynamic cycle that we will be looking at of course will be different so we are interested now in calculating the delta G values along the horizontal axis so in this case it would be calculating the change in the binding free energy upon amino acid mutation this which is in the top row there are delta G1 so for the protein which is in its hollow state so bound to a blue protein bound to the gray protein and as a reference to that we'll use the same mutation in the protein in its upper state so when it is unbound from the when the blue protein is not bound to the gray protein and in fact this blue and gray this complex of blue and gray proteins is exactly the application of which I would like to talk for a few minutes for so for this short talk I have selected two applications to showcase and the first one is TCR so T cell receptor interacting with the major histal compatibility complex bound to a peptide so this PEP MHC complex this system this complex formed by TCR and peptide MHC is a very important place a very important functional role in the cell so the gray protein which is known as MHC major histal compatibility complex it it's a function is to present on the to the cell surface fragments of peptides which could be either benign just native fragments of the protein of the native proteins in the cell or they could be also fragments from some pathogenic pathogenic infill infill trans into the cell so mainly some some viral or bacterial proteins that that could be detected in probably just mainly viral proteins that can be detected later by the by the T cell receptor so the blue protein which binds to the peptide MHC complex and identifies the potential the potential hazardous peptides and once it identifies that yeah this red peptide for example is in the picture it's a color red and depicted as a surface there so if this peptide is identified as dangerous then the further cascade of signal cascade would be initiated and then the cell would be destroyed all right and it would be of course very interesting to understand if our calculations could identify residues so residue mutations on the T cell receptor which which would be responsible for identification of these peptide these potentially dangerous peptides right and here a color to some important residues that have been previously already tested in this in an experimental SPR experiment published previously so it presents as a good test case and when we calculate run our alchemical free-energy calculations we see that yeah indeed we can get a very good agreement between the calculation and experiment only so on the left I'm showing just on the x-axis is the experimental Delta Delta G and on the y-axis the calculated one so we indeed can capture this effect of protein-protein interaction and the mutation induced free-energy changes on the right I'm showing exactly the same data just in a different representation in the bar plot so you can see that only three columns are I marked in red they are slight slightly giving slightly off values so slight inaccuracies but in principle these are only three out of what more than 20 almost 30 data values data points now we could also ask a slightly different question could we also would this method also work if we were to mutate the peptide right and here we have collaborated with the David Cole who presented us with some SPR data to benchmark our calculations and here we see a similar picture where we also have a fairly good agreement there are a few the data set of course is of course much smaller there are fewer residues to mutate here on the peptide and we have a few more more reddish points right at the columns however it is an interesting observation here you see one Lucy to Alanine at the position 3 is missing and the reason for it miss to be missing there is that after our calculation we noticed a very large disagreement with the experiment and we were puzzled by that so we came back to our collaborator and asked maybe they had some ideas about this and actually they went back to their essay and said oh indeed this data point is not reliable at all we should should not have sent it to you as as a as a benchmark point because it was there was something wrong with our with our essay for that so in fact it is the calculation in this case was even able to predict to identify some some problem with the experimental set up in that case alright so this was a very brief brief highlights of these TCR and peptide MHC complexes and for the second second case that I would like to show I would like to tell a little bit about the spike protein from SARS-CoV-2 virus binding interacting with the human ACE2 receptor so the spike spike glycoprotein is here I'm showing it in gray and red it's the same protein in gray and red it is protein responsible for mainly responsible for the the cough to SARS-CoV-2 virus entering the human cells how it does and how it starts this interaction it does it by flipping one of its receptor binding domains this RBD domain RBDs colors in red in this picture it flips it up and interacts forms a strong interaction with human ACE2 receptor which is there in blue if we look a little bit closer at it this there is a quite well-defined interaction interface and they're mainly from the human ACE2 receptor side there is a one helix that is interacting with the there is a helix interacting with that RBD and if we were to color all of those interface you would see that yeah they are quite fairly localized here I just called by the some distance criterion from the RBD and of course now if we we were to ask a question so which actually residues and how much their mutations in ACE2 would contribute to the binding between these two proteins we would like to if we were to ask such a question we would need to scan all of these residues colored here in yellow however of course scanning all of these residues and trying out all of them all of them potential mutations would result in something like 700 mutations in total so by means of chemical calculations it is possible however it would take us a long time to sample such a such a large data set so we here here I used a different approach of I divided into two tiers and firstly I ran a rather quick scan with Rosetta which is computationally cheaper than alchemical molecular dynamics based some calculations and with Rosetta I calculated all of the possible substitutions of those yellow residues into each other I mean acid type Rosetta is a it is not a molecular dynamics simulation but it has its own sampling set of sampling algorithms and it does not use a classical molecular mechanics force field but it uses an empirical knowledge based statistical potential and in the second round only from those predictions the most promising predictions made by Rosetta I selected about a hundred mutations to probe alchemical so let's immediately have a look how those results look like here I'm showing all of the data points about 700 data points on the x-axis would be mutations I'm not listing them because there would be far too many of them to write down but on the y-axis we see delta-delta G so blue delta-delta G so changing the binding free energy would mean that the the mutation this data point is stabilizing the interaction so most of the data points are either white and so they fall somewhere within one kilo calorie per mole region of a not of a very weak change in the binding free energy or are red there on the right side where they destabilized interactions and if we just here I'm just showing exactly the same plot just showing the most interesting and maybe the least interesting regions or or maybe also if we were to search for the mutations that destabilize the protein that would also be interesting the red ones but in this case we concentrated further on the blue on the blue data points there on the lower left and all of those calculations all of those possible mutations I then subjected to an alchemical scan so now these results are that I'm showing now are coming from the alchemical free energy calculations now every data point in the top plot there are there are about a hundred data points and in the lower plot it is exactly the same data I'm only showing those that would be potentially interesting in a sense that they should be stabilizing those mutations should be stabilizing the protein protein binding and you can see that now these data points are colored in either well reddish or bluish color and this color scale is another free energy calculation which actually reflects on the stability of the protein so one value delta delta g that we are looking at the on the y-axis is reporting on the free energy of binding but as you remember from the first from the first and second lecture this morning is that is that we are also able to calculate the actual stability so how much folding free energy of this of the mutation would affect the protein itself and the color scale then tells us that if the color if the color for example is red then the this mutation it would destabilize the protein if the color is blue then the protein itself so h2 would be stabilized so the most interesting mutations then to probe experimentally would be those that are both negative so very have very negative values in the delta delta g and also are blue or bluish so that they are in the experiment the protein would also fold properly and would be stable and unfortunately in we do not have a collaboration to probe exactly this complex experimental because it would be very interesting now to see how our predictions how well our predictions do in an experimental setup but we have established a collaboration to probe a slightly mode is experiment probe a slightly different system where only one helix is the main this main helix that I highlighted before is interacting with the with the RBD of the SARS-CoV-2 and in this case we can run exactly the same protocol by to identify those residues that would be interacting that those those residues that upon mutation would be a most mostly increasing both binding affinity and increasing the folding stability of the helix itself and here I color those our most promising candidates so the two or three bluish residues there on the helix are now being tested experimentally so I'm I hope that I can update have an update on this and maybe you can read in some near future scientific publication how how actually it all worked out all right these are my acknowledgments to people who collaborated on these projects and my funding thank you