<TITLE: Virus Club
ACADEMIC DOMAIN: medicine
DISCIPLINE: virology
EVENT TYPE: post-graduate seminar presentation
FILE ID: USEMP15A
NOTES: continuation of and continued in USEMD310, seminar also includes presentation USEMP15B

RECORDING DURATION: 17 min 45 sec

RECORDING DATE: 10.5.2007

NUMBER OF PARTICIPANTS: 35

NUMBER OF SPEAKERS: 2

S1: NATIVE-SPEAKER STATUS: Romanian; ACADEMIC ROLE: research student; GENDER: female; AGE: 24-30

S2: NATIVE-SPEAKER STATUS: Finnish; ACADEMIC ROLE: senior staff; GENDER: male; AGE: 51-over

SS: several simultaneous speakers>


<S1> okay everyone is ready </S1>
<S2> almost but <SS> @@ </SS> alright go ahead </S2>
<S1> good afternoon everyone today i will speak to you about the tale of two spikes in bacteriophage PRD1 i am pretty sure that some of you at least are quite familiar with the bacteriophage PRD1 so those of you might concentrate on more on your sandwiches while i will give a brief introduction of the system , PRD1 is the monovirus of the tectiviridae family er it's infecting gram-negative bacteria such as salmonella enterica escherichia coli er it has a icosahedral er symmetry with a er pseudo T equals 25 lattice er an internal lipid membrane and er linear double-stranded DNA genome , the structure of PRD1 has been investigated with erm cryo-electromicroscopy methods erm x-ray and three-dimensional reconstruction , er if we take a closer look PRD1 presents approximately 18 phage-specific proteins amongst which er we can see the P3 major pro- protein er P31 which occupies the vertices er <P:05> er P30 which is a protein situated underneath the P3 layer and it cements er the facets er neighbouring each other and of course P2 and P5 which are the vertex the s- basically the spike proteins er it has been noticed that P- er PRD1 is structurally similar with human adenovirus and er this similarity is er comprised basically there are many similarities but for example P3 er molecular fold is very similar to the hexon fold of the adenovirus , if we take a closer look at the structure itself we can see the er asymmetric unit called PRD1 which is er mainly er composed of er four P3 trimers , er the P31 pentamer which i mentioned before er and because this image was obtained as a er an atomic model of PRD1 and er it has been obtained by imposing icosahedral symmetry the P2 and P5 er spike proteins have been averaged out , now the spike of PRD1 is quite important for the life-cycle of the virus because erm it is involved in the bining binding step which is one of the initial steps of infection this triggers a cascade of events which er lead to the peptidoglycan er digestion er injection of genome replication and transcript- transcription of the proteins erm procapsid assembly packaging of the genome and finally the mature virion er lysis (xx) . erm it has been noticed that er PRD1 presents two types of vertices the majority of which er is represented by binding vertices so both the P2 and P5 proteins together with P31 the other type of vertices is represented by erm , a DNA translocation in vertex and it is believed that er it's just one such vertex per binding . now if we take a closer look to the proteins they are the purified proteins er of the vertex of PRD1 the P2 receptor-binding protein P5 trimeric fibre and P31 the pentameric base have been investigated erm by x-ray crystallography and small-angle x-ray scattering or SAXS and x-ray structures are available for P2 C-terminus of P5 and P31 for a P5 we have also available a SAXS model for the entire protein , unfortunately the structures alone cannot tell us too much about the in situ situation how these three proteins relate to each other in the virus and also this cannot be tested in just you know in (xx) because these proteins are not interacting as they would do in the virus so we needed to develop some new methods er to look at this situation and basically see how they interact in situ . based on the available information we were able to develop four schematic models on how these three proteins are connecting to each other in the virus in panel one A we can see the P2 receptor-binding protein binding to the P5 the trimer and everything and P5 binds to the P31 pentamer , the same situation is in panel one B but the difference between these two images is that here the N-terminus of the P5 is embedded into the capsid as opposed to here where it actually sits upon the P31 pentamer , now if we look at the panel two A we can see that the situation is a bit different from the SAXS model it has been suggested that P2 actually does not bind it has no interaction with the C-terminus of P5 so , we then modelled these two proteins as two different spikes , again here is the same situation this is P2 this is P5 and the difference is as before P5 N-terminus here is embedded into the capsid as opposed to here where it just sits on top of the P31 pentamer . now to see exactly which one of those models is the correct one we have developed a several erm a series of methods er through which i will briefly go , first of all we needed a full-length model of the P5 which was er obtained by sequence alignment and homology modelling er combined with data from x-ray crystallography and SAXS next we needed some pure PRD1 and we focused on four types of particles of course the wild type , then a mutant er which has a mutation in the five P5 protein that makes the spike er a bit stiffer because we know that in the wild type the spikes are rather flexible and then two other mutants which are lacking either one or two of the spike proteins , the next step was represented by cryo-electromicroscopy and image processing . with the cryo-electromicroscopy things are not that difficult to prepare the sample but you just need steady hands the basic idea is that you have a guillotine like device , er a pair of tweezers and the holey carbon grid you apply a small amount of the sample on the grid and then blot it with some filter paper to get a thin layer of the sample and then (xx) the grid will plunge in the liquid ethane bath er underneath the guillotine er the grids are then stored in er special folders or (xx) in liquid in liquid nitrogen . image reconstruction process is rather complicated so i have appealed to this representation which to me seems quite suggestive and i really hope that i won't find myself talking like this (xx) whistler's mother now @but hopefully that you understand something@ so first of all you start with the electromicroscopic er micrograph and imagine that these are the viral particles and as it happens also in the micrograph they are in er various positions or orientations so first of all you need to select each individual particle , and then you need to determine their orientation so basically as you see here all the houses are pointing with roof upwards if you think of the p- PRD1 the viral particle and you imagine that we focused only on the vertex , you can imagine that the chimney here is basically the vertex pointing up so basically we selected only the images which have vertex pointing up , then you need to classify these images , and then you get classes basically the classification is done on the similarities between images so you pick this this and this because they are similar and they form one class and these these and these and so on , and then these classes need to be averaged and if you think of this it's needed because it improves the signal so if you're wearing glasses and you take them off you have this kind of blurry vision you can see th- things you see the shapes but not very well and when you put the glasses on everything becomes sharper and clearer so this is basically basically the man- main idea also with this , so now you have all the pieces that you can put together and get a really nice three-dimensional reconstruction of the house @in this case@ or the viral particle . so i have showed you the methods now let's see what the results were basically first of all we got an atomic model of the P5 because , we just had all the pieces which needed to be put together so in panel A this grey surface is the SAXS model which has been fit with the C-terminal fragment of the P5 to give the orientation of the molecule , in panel B the atomic structures of the P5 which are known have been fitted first separately from the SAXS modelling and then they have been altogether fitted into the SAXS model and you can see that they fit quite nicely but still there is some unresolved volume here and we believe that this is because P5 might have two different conformations and this would illustrate the er conformation with the N-terminal with an extended N-terminus . next we determined that the spike protein P5 is actually di- binding directly to the penton-based P31 protein and the images are not all of them showed here but we have determined that , by applying five-fold er icosahedral symmet- er five-fold symmetric reconstruction erm on the wild type on the P2 P5 minus mutants on P2 minus mutant and on the P5 delta 8G mutant er we will get a series of representations that did actually show us a bit of what h- what's happening at the vertex , so here if you apply five-fold symmetry on on the vertex you average out most of the P5 which is a trimer but still you can see that there is something there , if you now look at the close-up of these two images , you have a knob-like density here which represents this image and we observed the same situation for the P5 delta 8G and for the P2 minus mutant , as a- apart from the P2 P5 minus mutant where the knob- the density is gone and also some of the connections of the P5 er with the capsid surrounding it so this is how we determined that actually P5 minus is directed to the P31 (pentamer) . then to our surprise after er looking at the , two-dimensional average classes we found that PRD1 presents not one but two spikes , so if you look at the wild type and the P5 delta 8G which is similar to the wild type but as i have mentioned previously has a stiffer spike , you can see the presence of two , two vertex er two spikes and these are sc- schematic models which will help you interpret the images so in this situation because in the wild type the vertex is rather flexible you can see that the orientations of these are rather fuzzy but if you look at this you er see one one spike which is kind of pointing upwards and it has a knob-like er head and a smaller one which is er at an angle with the first one , so of course then next we needed to assign which one is P2 and which one is P5 so we imaged the P2 minus mutant and we discovered that one of the spikes is missing the one which was at an angle so you get only this one which is pointing upwards and has this big head domain so we assigned that the one missing which is this one should be P2 and to check if this er other one which is left is actually P5 we imaged the P2 P5 minus mutant and we discovered that they are spikeless . so based on what i have shown you previously we were able to make a three-dimensional reconstruction of the spike complex , so our data now supports the fact that there are two separate spikes and you can see here the C-terminus domain and the shaft er of the P5 the rest of the er P5 in averaged out in the reconstruction and also you can see the P2 this is a side-view of the vertex and because P2 is rather flexible there were several ways of fitting er the P2 er into the model , the second panel shows the same structure but viewed from from the top with the P5 trimer and P2 as a (model) <P:05> now if we look at panel C , we see a reconstruction of the P5 together with P2 and their relationship er with the capsid so basically we have the P5 which might probably form a heteropentamer together with the P31 but the N-terminus of P5 is definitely embedded into the capsid the P2 has the tail embedded somewhere into the capsid so leaving the head domain er to be the receptor-binding part of the protein , and the next panel shows a possible representation erm of the P31 P5 pentamer heteropentamer , so to sum everything up basically we have studied the architecture of PRD1 er vertices by cryo-EM and three-dimensional reconstruction and we have discovered that it presents not one but two spikes and these spikes are formed by the P5 and P2 proteins , and last but not least @i@ i would like to thank the following group members <NAME> my supervisor <NAME> <NAME> and <NAME> er for basically doing most of the stuff er <SS> [@@] </SS> [@involved in the project@] and for having the patience of this guy over here even when things were moving just as fast so thank you very much for your attention </S1>
<APPLAUSE>
