 So hopefully David introduced you to the concept of lipids and the properties of lipids. This is actually quite old. We've known that lipid was an important component of cellular membranes for almost 100 years. The actual structure was published in the 1930s when Schmidt-Bear and Ponder in 1938. I think it was used to polarize light on red blood cells to realize that you have this ordering and an interior lipid phase that's hydrophobic and then things facing the water that are hydrophilic. This was confirmed in the 1950s with electron microscopy and eventually David Robertson did a huge amount of amazing work in the 1980s. I'll see if I have a couple of pictures about that. Here we have the red blood cell. Do you see the thin layer here? This is actually not one but two layers of lipids. And I think I have some micrographs. Yes, here again you see an electron micrograph where you really magnify things so that you can see the lipid detail. But for natural reasons most of the early studies on membranes were topped out. We started to have a membrane and see how far we could penetrate this either with light microscopy or electronic microscopy. Which is fine in many ways but just looking at this we see that there is a membrane but we do not see all the atomic detail. Exactly what are the properties of the molecules here. For a long time we've drawn this in textbooks as these plain simple lipids with two straight tails. Today we can do this in slightly modified ways by using computers and though computer simulations are severely limited in other ways that I'll be happy to tell you about at some point they are really useful as a computational microscope so we can look at the real behavior of these molecules. So this is a movie of a lipid from a simulation. It's called POPC this particular one not super important. You have two large fatty acid chains here they're derived from fatty acids originally and they're very hydrophobic. They're connected to a group here that we call a glycerol. It was originally derived from a glycerol when we created the lipid and then eventually we have a head group here that is very polar. Hopefully David told you about the fact that it's even so polar that we occasionally call it switter ionic. You have a full minus chart here and a full plus chart here. Do you see the lipid tails here though that they are much more zigzaggy than the traditional pictures you might see in textbooks. Let's have a look at that. I have an old movie for you. It's a bit corny but I think it's a great way to introduce lipids in particular in the context of how they interact with proteins and other stuff in the memory. You won't hear the speaker voice here. So if we zoom in on this animation you will see the lipid bilayer. This lipid bilayer is really working like a two-dimensional liquid the so-called Singer-Nicholson model when things diffuse freely in a two-dimensional space. The lipids themselves though are very flexible and this flexibility is what gives these self-assembling and in particular self-repairing ability that if we have small defects the lipids will heal very quickly and there will not be any water going in although you don't see the water molecules here. The membrane is not limited to lipids though that's important. You have a lipid bilayer as one component but in addition to that we have cholesterol these small gray parts here. We have some sugars sticking out and in particular what we do not show you here is that we have a ton of proteins actual membrane proteins embedded. Depending a little bit on how we count it can be up to 30% of the mass of a membrane that's actually the membrane proteins rather than lipids and that's what I'm going to spend most of the time today talking about. A better way of viewing the lipid bilayer itself rather than those simple I'm not even going to draw them but rather than those simple models where we have a head group and then two straight tails at this this is a snapshot taken from an actual simulation that's probably even me who read this a few years ago now it's most of my students. Do you see the chaotic nature here in the hydrophobic part? That is because all these long hydrocarbons while it might look plain nice and simple to put them with straight chains that would actually be very unfavorable from an entropic point of view. Think of this as throwing out a string the probability that the string will be completely straight it's virtually zero although it could technically happen. This in particular means that the hydrophobic part of the bilayer will not have any ability to form hydrogen bonds or anything so it's going to be difficult for us to put anything charged in here I'll get back to that in a second. This head group part on the other hand that's almost the opposite not only do they like water the large charges here means that they're almost more hydrophilic than water itself they will love to interact with charged and the hydrophilic polar things which is going to create a very specific pattern. Hydrophobic parts go in the center of the membrane while hydrophilic polar even charged parts will not just accept but love to interact with the head groups and then eventually we might have a large water soluble domain or something outside it. It's an amazing chapter just to understand transport through the membrane some molecules such as a oxygen or carbon dioxide will actually be able to go spontaneously diffuse through the membrane because they're not polar or anything. David will likely tell you a little bit about that in the class but again the obvious example is your red blood cells. All the oxygen I'm breathing when I'm standing here that's diffusing freely through my membranes and the carbon dioxide that is then being transported back to the lungs the reason that goes into the same red blood cells is that it can diffuse pretty much freely through the membrane. But this is roughly where we're going to stop looking at plain lipid membranes that consist just of lipids and can start considering all the complications. Do you see the parts that we have here first we have the blue part the blue large part roughly there is a large membrane protein actually it's not just a membrane protein you have one so-called domain that sits straight through the cellular membrane here then you have one part that's sitting outside one part that's sitting inside we might have some small I think these yellow components or so that could be cholesterol you might have a green component out here that's a so-called monotopic membrane protein I think so it's a protein that's anchored a bit into the membrane but then it sticks out on the other side and this is a far more realistic picture than the others these things are influenced by the membrane but they also do influence the membrane themselves cholesterol for instance have a tendency to rigidify the membrane which is very important to some of ourselves and occasionally plants too. What I'm going to tell you about today is a bunch of concepts these are roughly the ones we're going to cover I think it's a great idea to go back revisit this be make sure that you know what I'm speaking about when I write this down if you do that you master this lecture and then you should be able to pretty much breeze through the study questions at the end but I'll go through the roughly one by one hopefully you've studied amino acids by now but that's going to be where we have to start the amino acids and proteins that's going to be the stuff that interacts with the membrane Gunnar and a few others will talk a little bit about membrane insertion which is a complicated process in itself but if you trust me that the backbone of say helices can insert in membranes that's going to mean that the entire difference will come from the different properties of different amino acids and that's entirely due to the sidechase that's the one difference we have in particular we have some non-polar aliphatic groups like smaller sidechains alanine is not quite non-polar but it's small enough to the side chain is effectively non-polar you have valine leucine isoleucine etc they're great in the cells that they don't have any large aromatics rings it's a fairly floppy flexible chain and there are pretty much no charges at all in the chain they're going to love to be in the membrane interior we have some polar sidechains they're not charged but they are polar they will not love to be in those lipids environment but they might be able to get by they're under some circumstances but they will likely prefer the head groups then we have on the other side here you have positively and negatively charged sidechains there is no way they're going to enter the membrane at least not under normal circumstances they will need to be in head groups or facing water and then we have some special cases for instance the aromatic sidechains some of them are hydrophobic some of them might have one hydrophobic and hydrophilic part which makes them very interesting because they could kind of like both the head groups and the tail region of the lipids at the same time that's going to help us so i think the best way to illustrate that is to go back to our friend the lipid this is i have to confess that this is not an isolated lipid if this was isolated the chains would stretch in both directions here but this is a snapshot of one lipid from a large membrane running in a simulation it's just that i've hidden all the other lipids for you but if we consider a simple sidechain such as valine where would this valine molecule like to go well the entire sidechain here is hydrophobic i actually could not put that amino acid the way it looks here in a membrane because the zwitterionic backbone here right the nitrogen and the oxygens they would have very large charges but if we presume that those were part of a helix these parts would literally be part of the helix and have hydrogen bonds to other partners and then we could kind of cut this off so i only have to consider the sidechain that sidechain is going to love to interact with the lipid chain so roughly from here up to here it's going to be soluble like like solves like and in particular there's not going to be any hydrogen bonds perturbing through anything beautiful another sidechain that we considered the same way again let's forget the backbone part this is arginine it's a beautiful sidechain in many ways but in terms of solvating it inside membranes it's as horrible as it gets you have plus one charge here typically that charge is going to mean that there is no way it's going to go here this will have to interact with the head group region in particular the phosphate there that's negatively charged or even those carbonils you have the oxygens here oxygens typically have negative partial charge in molecules so that's will likely be a place where that one likes to interact then on the other hand we could pick something more complicated tryptophan it doesn't get larger and more complicated than that in terms of amino acids where will this sidechain want to be well you have the hydrophobic part here that would likely prefer to be down here then you have this part here right the NH group that's going to be polar so that would prefer to be up here so this molecule would somehow prefer to be maybe oriented this way so that the polar part can interact with the head group and the hydrophobic part can interact with the chains and that means that we will not really be able to push this either down or up so it's kind of like casting an anchor here in addition to that you see how large this sidechain is it's going to be virtually impossible for this sidechain to start to rotate because remember I'm surrounded by these lipid chains there are several other lipids that I've hidden here so this sidechain will have to be placed like my hand this way so that it can intercalate squeeze in between two lipid chains if I were to place it that way it would cut straight through like half a dozen lipid chains so this is going to be a very particular sidechain that both is ordered and it stays put in one place I will come back and show you that in a few lipids so already here we start to see that there are patterns with which how amino acids interact with lipids and in particular different amino acids will prefer different parts of the lipid environment instead of looking at a single lipid we can do the same studies for the entire membrane the most obvious way to do that is with neutroscattering and the reason for that is that the membrane itself is fluid right that I can't determine a crystal of it if you do this with neutroscattering you can use deuterium exchange to find out what are the different parts in a membrane and roughly where are they located in a cross section so if we do that you would see that in the entire middle part of the membrane maybe up to 30 angstrom or so it's purely hydrophobic there is not a single party here that can form a hydrogen bond then we start having this carbonyl groups so the carbonyl groups are those oxygens connecting the central part to the chains they're not particularly strongly charged they they certainly don't have a plus or minus charge but they have a bit of partial charge so suddenly we start to be able to form maybe a few hydrogen bonds or so and then you have what it says in red and green here sorry if you're colorblind I didn't choose the colors the phosphate and culling groups they are the actual negatively charged part phosphate and positively charged culling in this particular lipid the switter ionic component where charged things will love to interact and eventually you get the water but here too you see there is no water whatsoever in the core of the membrane you might think that this is a plot that belonged better in the lipid properties lecture you're probably partly true but you need to keep this in mind when you ask yourself what amino acids will be stable where what will they do to the protein and what will that do to the membrane environment so I guess it's time to start looking at that this thickness we can determine quite accurately now either with neutroscattering or you can frequently actually see membranes in electron microscopes so this is a much higher magnification picture than the other one in the previous part we saw kind of two layers with a well two completely separate membranes separate by a few nanometers this is part of the horny layer of skin so the small lines here the stacks we see are actually multiple laminar layers of lipid membranes that are starting to form skin that's the chapter itself that I'd be happy to tell you about at some point not relevant for lipid protein interaction most membranes have roughly this thickness but not all there is a bit of a tendency so first this is hard to determine and we don't know this exactly but if you look at different organisms to first approximation they have roughly a similar membrane thicknesses again ballpark of 30 angstrom in 25 maybe in the hydrophobic region if we look inside a cell though there appears to be some slight differences where the endoplasmic reticulum in particular the first membrane where we create proteins the protein factory in their cells appears to be slightly thinner and then there is some sort of tendency all the way out to the plasma membrane around our cells which is slightly thicker I would take this I would take this with a grain of salt there are other studies that somewhat contradict this but in theory at least if there is a systematic difference between the thickness of the membrane cells might be able to use that as a way to sort proteins so that thin proteins that are not alpha helices that are not so long for instance they would go into thin membranes while longer alpha helices would go in the thicker membranes so the obvious way to understand how a membrane protein interacts with the membrane that would be to just determine the structure of the membrane protein the problem with that is that there are two problems first is really really really difficult to determine structures of membrane proteins second when we actually do that in the few cases we do the only reason we manage is that we typically take them out of the membrane and put them in a much simpler environment that simpler environment is not really going to help us understand the original lipid interactions that much but we have to start somewhere the one of the first membrane proteins people studies was bacterial rhodopsin rhodopsin this is an exceptionally interesting class of proteins but in the interest of time I won't take you through exactly what they do but it's very similar to the rhodopsin molecule we have in our eyes that's detecting light bacteria rhodopsin occurs in bacteria in particular in a membrane called the purple membrane and the membrane is so stuck with proteins that it's like 50 percent of the mass here is protein rather than lipids so it's almost like lipids embedded in protein that made it a great candidate to try to stabilize. Hartmut Michel came up with a great idea that you can rip this out of a membrane and stabilize it pretty much with detergent I'll come back to that at the very end of the lecture so you don't have to understand why for now then you could turn this into a crystal you could ground the crystal down if you ground the crystal down that's where you see this purple color it actually comes from the mixture of the protein and lipids that's why we call it the purple membrane and based on that they were eventually able to determine the structure of the first membrane protein and they got a noble prize for this discovery too that you could stabilize membrane proteins the structure itself consists of seven alpha helices transmembrane so they go from one side and then a small loop out to the other side small loop down to the other side that this makes a lot of sense the alpha helix is going to be a very nice stable structural element because all those peptide bonds you know about they would they are polar they would normally hate to face the lipid environment but this way they face each other you form hydrogen bonds in each alpha helix hydrogen bonds in the second alpha helix beta sheet proteins well there are a few membrane proteins that are beta sheets but in general we could not put a beta sheet in a lipid environment because the end of the edge of the beta sheet would face the lipids right the only way we do manage that is by turning the entire beta sheet into a barrel and closing it up on itself so there is effectively no edge but this when this first appeared we were excited that it's simple we just need to understand these helices and then if the helices pack we should be able to understand how membrane proteins work in practice it turned out to be a little bit more difficult but we haven't quite reached that part of history yet so let's just be content and assume that this is going to be something easy to understand you might think that if you remember globular proteins how are globular proteins stabilized globular proteins are stabilized by having hydrophilic water-liking residues facing the outside and then you have a rapid hydrophobic collapse where the hydrophobic residues will face the inside so if I draw that to you in water you would rapidly have a chain here where you had some positive and negative charges and some hydrophobic ones that would rapidly go to the state we have all the hydrophobic parts on the inside while the charges are on the outside important part of protein folding when we first saw these proteins first you realize if the lipids are entirely hydrophobic and then you see the amino acids here a ton of them are hydrophobic so some of the first simple models were that membrane proteins were kind of the opposite it makes sense if that's hydrophilic on the outside hydrophobic on the inside these should be hydrophobic on the outside and then maybe hydrophilic on the inside unfortunately that's not quite true membrane proteins tend to be hydrophobic everywhere or let me modify membrane proteins tend to be hydrophobic everywhere in the parts that go through the membrane with a few exceptions I'll come back to these parts on the other hand can be quite hydrophilic and as Gunnar will tell you there are going to be patterns here when in general we tend to have positive residues on the inside which determines how they sit in the membrane that makes membrane protein folding exceptionally difficult to understand Gunnar will spend an entire lecture on it I think I will touch a little bit on it but it means that the stabilization of membrane proteins is much more delicate than water this will to first approximation just be Lenard Jones packing fundervolse interactions there are going to not going to be any hydrogen bonds keeping things together in general these healers is just diffuse and find each other and hopefully they pack well enough that they will maintain their stability the membrane helps with this because the hydrophobic hydrophilic parts of these head groups means that I can't take this protein and pull it out of the membrane then I would expose hydrophobic parts I also can't move it down because this hydrophilic regions here would then be exposed to the lipid environment so there is a symbiosis here the membrane helps stabilize the protein and the protein might help create a bit of a structure in the membrane as we will see so in an ideal situation the alpha helices we would like to assert would perfectly match the membrane or maybe that isn't so ideal after all it turns out that we can learn a lot by studying what would happen in a non-ideal case guess about that and see if that is confirmed by experiments so let's assume that I have a small membrane and now I'm sorry I'm going to draw this in the way you really shouldn't draw membranes this is not how membranes look but you don't want to have me spend the next 10 minutes just drawing proper lipid chains what if I try to insert some membrane protein here alpha helices that are too long they're not going to look like that because there would be hydrophobic parts here sticking out that will definitely not happen so if I have several helices like that the first thing that would happen is that they would aggregate that I can draw on here it's at least better to have three helices together this way if they mismatch so that there is only the parts on the outside matching rather than having three different parts at different positions of the membrane matching I might also end up in a situation where these lipids somehow start to stretch out their chains a bit so that I create a longer region here where the membrane is locally a bit thicker again now that I just draw the chains in parallel fashion that's not going to be so obvious but remember that I said the chains are in general a bit zigzaggy while here they would be straight then I might create an environment where I get these three helices to fit just perfectly and we will see that that actually happens now and then so that's one way to stabilize this so-called hydrophobic mismatch that's an important concept so I'll write it down hydrophobic mismatch the other part that can happen is that I didn't draw this specific but there are different types of helices do you remember that there is a helix called 310 helix so if I stretch the helix harder it would that I would literally distort the backbone a bit it's not going to be as advantageous as a normal helix but I get a slightly longer helix so if the membrane is too thin or short I might be able to alter the backbone of a helix so 310 helix would be slightly longer or in this case I would likely prefer so-called pi helix so that I make the helix itself slightly shorter to fit the membrane the other part that could happen if I'm thinking still thinking about too long helices I could take my membrane let's see here I need to do the math correctly here slightly fewer lipids here and take that entire helix and draw it in a tilted fashion that would also enable you to keep a longer helix in the membrane without distorting the lipids too much you can probably imagine that this is going to be pretty bad because it's now distorting the lipids instead but maybe if I have five or six of these together it could stabilize things a bit if you look back at that bacterial rhodopsin structure you will kind of see that a few of those helices are tilted a bit so that makes sense the same way we can consider what happens if we have helices that are not really thick enough there too you could have an aggregation effect that if we anyway need shorter lipids it's not ideal but it's at least better to have three sorry short helices it's at least better to have three short helices together than at different places the other alternative we can assume here we can have the membrane curl up even more here I will actually draw six saggy chains I'm not good at drawing lipids so you see here by making the chains six saggy I effectively create at least a locally thinner membrane that could then match my helices and here too I might be able to stretch out the backbone of the helix itself to make that slightly thicker the other thing I just might be able to do is that if nothing else works if the helix is simply not long enough to go through the membrane well at some point I will have to give up and by giving up I mean that the helix will eventually sit on just one side of the membrane it will become a sort of interfacial helix that's going to be bad for other reasons in this case I'm going to need half the helix here to like water and the other half of the helix to like the membrane but again if I don't have any other choice the key take-home message here is that do you see how lipids can adapt to the protein for instance here but you also have the protein adapt to the lipids and I'm going to keep coming back to that it goes both ways the problem though here is that this is just based on my hand waving you have no idea whatsoever if this is true and I can't really determine x-ray structures of this because I'm looking we're looking at transient features of lipids floating around in individual lipid transient features of protein floating around in lipid environments I can't crystallize that lipid environment it is possible to access this indirectly in particular with NMR nuclear magnetic resonance and Antoinette Kilian and Ole Moritzin spent almost 15 or 20 years on this in the 1980s I'll come back to that in a second and show you how they did that but basically they systematically try to replace residues here and see what happens at what point will we start to have tilting in the helix for instance but before that let's look at what this would do to a real protein so if we now apply this to a real protein let's show you one what happens when a protein inserts depending both on the protein and the lipid environment we might have this you could imagine that the lipid is a relatively rigid environment and in that case we would have this scenario that the protein would adapt to fit the bilayer do you see here how all those helices are tilting of it in fact tilting helices is great for packing helices that has to do with the way the side chains are organized in an alpha helix if you put them in a parallel fashion side chains would in general bump into each other but if they're slightly tilted the ridges of one helix will fit into the grooves of another helix and here the lipid environment might actually help that the other alternative is that the protein is quite rigid if the protein is quite rigid we would rather end up on the right where we have the lipids the membrane will adapt to fit the protein it turns out that if we now have the protein undergo a conformational transition for instance with changing pH changing pressure or something this way we could actually either have the cell influence say a protein to open a channel or something or a protein influence a membrane say you start to bend the membrane which is how we achieve a whole lot of functionality in the cell so in general these properties are true the only question is how Anthony Tkelyan and others were able to study them the way they did that was that they started from plain simple alpha helices because they're easy to insert in membranes and then we systematically changed the competition composition here w tryptophan a alanine l leucine and there is something else i'm not showing here simply because it's not really part of the helix but the black part up here is a proline do you remember proliens proliens are strong helix breakers so when i put the proline here that's going to mean the end of the alpha helix and then there will be so small coil but still that's important if the helix has a clear beginning and a clear end that's going to mean that it will not for as if the helix is too short one way to that that would be for the helix to become slightly longer right but by putting a proline at the start and a proline at the end i ensure that the length i gave this helix is going to stay that way it's not going to be able to extend the helix to fit better w tryptophan that was that particular residue that i mentioned that really will anchor things to the head groups so by putting a tryptophan there i can almost force that part to stay in the head groups i can't pull it down and i can't push it up a and l alanine is a small residue formerly it's a hydrophobic side chain but it's so small that it can really go either in water or oil l leucine is a longer hydrophobic side chain but plain simple nothing special with it if i introduce more leucines more of the red parts here i'm going to make the helix more hydrophobic if i introduce more alanine i will make it more hydrophilic so by changing the a and l composition here and also the length i can choose a how hydrophobic my helix is will i have the hydrophobic part mixed or will it be hydrophobic on one side hydrophilic on the other and then i can also adjust the length while ensuring things stay in the head group breathing and this is roughly what i'm doing a bit they looked at the length distributions of helices and that's how they saw these parts that in particular with the anchors the trip to fancy right with the anchors you can actually get helices that are surprisingly short maybe just 12 residues longer so to still go straight through the membrane normally i would need say 20 residues or so for it to be stable but this one works it's just that it's going to distort the lipids heavily but nmr tells us it does go straight through the membrane at some point when the healer starts becoming maybe 2021 22 residues it's starting to be a stretch and when you go all the way out to 24 we can see in the nmr experiment because we're starting ordering that this lipid is now no longer straight but it has tilted relative to the membrane that's not good per se we will distort the here the lipids more but again that is the only way to keep the helix stable so be it and most of the simple understanding of the parts i have showed you where things go were determined this way so what if we now take that helix but instead of drawing it the way we have it here let's assume that i take a helix that's roughly 50 aline and 50 lucene but i put all the alanis on that side and all the lucenes on that side well to first approximation i would have the same average composition as the helix but my helix up here would be had roughly hydrophilic here and hydrophobic there so that particular helix would actually prefer to sit this way in the bilayer with the lucenes the hydrophobic side facing down and the alanis the hydrophilic side facing up as i already hinted to you this occurs in real proteins too let's start with mechanosensitive channels so mechanosensitive channels these are ion channels that david will likely talk a little bit more about later they're usually surprisingly simple these are typically just channels that are very large holes that open up to let water through in cells and one of the reasons to let water through in cells is simply to adjust pressure if there is too high pressure in a cell we need to let out some water or the cell would rupture on the other hand the channels these holes can't be open all the time because then the cells would be leaky so what we'd have are channels that under normal circumstances they sit embedded in this bilayer up here and they are closed they are relatively thick actually so they're not super happy in the membrane and the membrane is not super happy with them do you see here how the membrane has had to become slightly thicker around the channels to be able to host them what then happens is that if i start to stretch the membrane why would i stretch the membrane well if the pressure increases in the cell eventually this will lead to pressure building up on the membrane right that's going to lead to a tension in the membrane as this tension builds up some of these channels will now start to open do you see the channel on the right they're opening up so the channel itself will undergo a conformational transition but this is really caused by when this membrane is stretched out the lipids will have to become even thinner again the volume of the lipids is roughly constant so if i want the area to be larger when i'm adding tension that means the thickness has to go down well then the protein will have to adapt to that otherwise it would eventually no longer be in the membrane the way these proteins adapt to it is that when they become thinner the actual channel opens up as we've seen there on the light blue part when that happens short term that's good the pressure is actually relieved a bit so that it might be sufficient for that channel to open up or maybe not maybe the pressure is starting to build up really rapidly in the cell now if the pressure builds up even more rapidly we're going to be down here and then additional channels will open up so now even the second channel has to open up we didn't relieve the tension enough here and now i have two channels where water will be able to flow out of the cell once that water has flown out of the cell what happens well the pressure in the cell is lower the pressure is reduced as the pressure is reduced inside the cell the tension in the membrane is reduced and then we start going back here maybe one channel will close the pressure might become even lower because we're still having one channel open here and eventually you're not going to go back here when the pressure is relieved and the channels are now closed again this is mechanosensitive channels the you might you might want to have seen the abbreviation at least it's called usually called msc mechanosensitive channel and then there's occasionally a small and a large one which is called msc s and mcl there are other proteins where this happens not just channels but say a protein that would like to induce curvature in a membrane what if i insert a protein that has an almost conical shape so here the conical shape of the membrane up on the upper right there do you see what that does that will induce a local change of the curvature in the membrane which is of course per se it's not good but if we have lots of these proteins expressed the proteins will definitely not be happy in water if they're hydrophobic so this is going to be a balance it's at least better for the cell to insert those proteins in the lipids but when i do that that will distort the lipid bilayer if i have lots of proteins like that i might actually start have a large global curvature of the entire bilayer and that's what you see in some cells for instance certain bacteria and everything that you have a shape to the cellular bilayer by expressing specific proteins in certain parts of the cell so we're back to this theme that lipids certainly decide what proteins will be stable in the bilayer or the membrane but the proteins themselves when they insert that they in turn also influence the shape and the properties of the bilayer either by opening it up mechanosensitive channels or by inducing curvature the only somewhat irritating part is that i keep hand waving but we do not need to hand wave one advantage today is that we have computer simulations we can actually look exactly at what happens to specific helix and specific residues inside a membrane and i'll show you a movie of that that one of my students anna you once on did many years ago this is of course a highly simplified system and i'm only showing one helix and then we're zooming in and making the lipids transparent here this particular helix is mostly hydrophobic but there is one residue here that is not not only is it not hydrophobic it's even polar this is a lysine sidechain i always get dizzy from this movie sorry this lysine sidechain is at the end of the helix so and being at the end of the helix do you see where the lysine ends up well the lysine is not really in the membrane technically it's a residue in a transmembrane helix but the lysine sidechain itself is very very happy up here in the head group region and even the waters those dotted yellow lines you're seeing here are hydrogen bonds that it's forming not even hydrogen bonds we tend to call them salt bridges when they are with a charged region so all those charges in lysines dislocated this far up will be paired so we're not going to have that poor lysine sitting in the membrane all by itself that one was quite happy so don't assume that the second you see a charged residue in a helix means that it can't go in a membrane it will matter where we put it in the helix apparently a better way of showing you this could be this this is a helix i think there were some lysines in this but i don't show them in general what a helix can do the central part here i cannot have any water in but if there are some slightly hydrophobic hydrophilic residues here at least at the end i can distort the system a little bit to pull in a bit of water here that is not good per se on the contrary it's bad per se but it's better to put a little bit of water here to make sure we at least have some hydrogen bonds rather than taking say a polar residue and pull it into a completely hydrophobic environment a little bit of distortion is better than having unpaired charges or hydrogen bonds so what happens though if we take that lysine that i had before and try to pull it further down in the membrane i have a movie of that now i have moved this one or two residues down and here the backbone part of this is certainly in the hydrocarbon part roughly here where my hand is right but you see what that lysine is doing the lysine due to its long neck the side chain is actually stretching out so this side chain is certainly not whisking around and sampling the interior parts of the membrane down here it's always sticking up to pair those hydrogen bonds that's an interesting concept let's continue let's pull it even further down so now i have actually introduced two of them do you see what happens here each of those lysines they're almost in the middle of the membrane now and yet they are not because the actual charged part of the lysine side chain is located by stretching this out it's located recently far up they have to pull in a bit of water or do you see what it's interacting with it's pulling down a lipid a bit so it's taking the carbonyl groups of the lipid and i think that's roughly what happens here too so the positively charged side chain can interact with those negatively partially negatively charged oxygens in the carbonyl but that's going to create a force on the side chain and if i now have two of these residues on opposing sides of the helix do you see what the helix is doing it's tilting and this you can even see in experiments that there is going to be a systematic tilt depending on where i placed hydrophilic residues do you notice how i'm almost contradicting myself because i said that charged residues really couldn't go in membranes and on one hand i'm right because they are not really in membranes but they are in membranes in the sense that among these 20 residues in the helix i'm certainly having two residues that are arginines embedded relatively far into helix it's not going to be good to have it this way but it is possible we can continue this pattern um you can take many different hydrophobic or hydrophilic residues and just check what happens in a simulation for some of them if i had a full charge that's so bad that i need to have water around it at any cost but if i take something that's merely polar um i think what is this it's a methionine i think the methionine here is a bit polar it's certainly not going to be happy here but it's not so unhappy that we can afford the cost of taking water all the way in to solve it that sidechain so for methionine apparently the methionine will rather make do and accept that it's not going to have a perfect hydrophilic environment what happens if i introduce another rescue say that i take that arginine do you see the difference here the arginine is so charged that at any cost we can't afford to keep that charge in the membrane i have to pull in water even though this means destroying the membrane locally you're not seeing the lipids here but trust me they are around and this is an interesting concept the reason why this happens is that taking a full charge and solvating it in oil it's basically not going to happen i'll hide myself for a second here but this is a hydrocarbon environment this would be pure oil but this is also how the lipid environment will look to a charge if i take this environment and try to put a sole charge in here we can actually measure experimentally roughly how costly that would be that would be roughly 20 kilocalories per mold i'm not sure if you have any gut feeling for that energy but this energy is eat my left true territory there is no way on earth this will happen spontaneously it's going to be once in a billion or so forget about it the residue the arginine residue would rather would rather deprotonate itself and becoming neutral paying a lot of energy for that than to trade to introduce the charge in the middle of the membrane this itself leads to an interesting pattern you might think that positive and negative charges are same size opposite sides of the same coin which is kind of true but they behave very different in the membrane lysine and arginine for that matter they're positively charged side chains they will have this pattern that i showed you that they're sticking out a bit we're going to call that snorkeling in a few slides that particular means that they can interact with water but they don't have to stretch as far as the water they can interact with those negatively charged oxygens in the carbonyl groups which is the charged group located closest to the interior in the diptylary but note that oxygen is negative it can pair up with a positively charged head group but it cannot pair up with a negatively charged amino acid such as aspartic acid aspartic or glutamic acid they have negatively charged side chains instead which again would seem just like the positive charge one just with different sign this helix or these side chains here they cannot interact with the carbonyl groups they have to stretch that all the way out to the water or the phosphates do you see what the difference is this is a far larger distortion of the entire helix and the membrane around it this is going to be much much much costier so it's harder to introduce negatively charged residues on the inside of membranes than positively charged ones now friend of order would say can't we just forget about that by not trying to introduce any charged ones at all but it's going to be functionally relevant to introduce them at some points and i'll show you why in 20 minutes or so just remember when we need a charge in the membrane it's better to have positive charges that we work with the negative charges and finally just to compete this here we have a residue that was not charged but tyrosine kind of similar to tryptophan that it has an aromatic ring do you see how these rings locate this is not a simplified model or my illustration this is a snapshot from a computer simulation the rings orient in the way so that they intercalate between the lipid chains they would never be in this orientation because that would require them to push away five or six different lipid chains so this is a very particular pattern for aromatic side chains positively charged side chains and negatively charged side chains let's minimize that a bit and take this into the lipid environment here the reason for the positive charges is that they interact with these groups the negatively charged sizes on the other hand had to stretch all the way up here or even the water and the tyrosines in particular they need to orient this way to avoid interfering too much with the chains let's get back to this whole snorkeling sticking out concept and if i look at that snapshot this is roughly what it would look like that lysine side chain do you so do you remember seeing in the simulation how this pretty much never changed it stay stuck in a way well it's not exactly stuck but we can calculate the average angle this would be zero degrees so the question is how much are we sticking up here on average this concept is called snorkeling and we occasionally see that in structures of proteins at charged or positively or negatively charged or polar side chain will prefer to snorkel out to the aqueous the hydrophilic parts let's look at this an example of this you don't have to trust me your basis in computer simulation this is a relatively large protein with a transmembrane domain it doesn't matter what it does i think it's one nik if you want to have a look at the pdb i'm really only interested in the transmembrane domain here this is the transmembrane domain we have two chains up here and we have a few gray lipids i'm not going to talk about the lipids but that's how you see that it's the transmembrane part do you also see that the transmembrane part consists entirely of alpha helices let's magnify that one and look just as that part and i'm going to ignore the lipids for a second here there are two interesting residues in this particular protein first we have a lysine do you see the shape of that lysine this is not based on a molecular simulation this is a structure in an x-ray structure in the protein data bank this x-ray structure the reason why we can see that is that it's a diffraction pattern from a crystal of billions of billions of billions of billions of identical copies and in every single such copy this lipid will stretch out because we have all the lipids here the lipids is also effectively bound to the structure that's why we see it that's because it's a short side chain that's snorkeling i'll write that here we snorkeling but we have another side chain up here do you see what this side chain does this is an aromatic ring it's a phenylalanine so it's a hydrophobic aliph sorry hydrophobic aromatic residue but this wine is pointing in the opposite direction which makes sense this is something that is purely hydrophobic it will definitely only one interact with the lipid chain so this particular residue we say that it is anti snorkeling these are important concepts to understand we see them in proteins now and then and in particular it's something that we can use in protein design later on in particular to understand how the membrane proteins work so the problem with showing you snorkeling and anti snorkeling in a single structure is that there are too many fluctuations and there might just be one or two residues that exhibit that property the advantage of trying to find a numerical measure such as literally measuring this angle is that i can create averages either an average in a molecular simulation or maybe averages of structures now 20 years ago that wasn't particularly useful because we only had a handful of primary protein structures today we have hundreds if not thousands and that means that not only are there enough structures available the resolution is starting to become so high that we definitely see the side chain snorkeling and in many cases we might even see the lipids they interact with i'll show you this first in a simple simulation so this is just two examples a methionine side chain and an arginine side chain and the exact values here are not important but you see how there is in particular for the arginine there is a clear snorkeling effect what do i mean by n terminal and c terminal well there are two sides to this helix right if i put the helix to the membrane there is the beginning of the helix and there is the end of the helix and it turns out they're not quite identical if you look at a helix and let's say that we go from the n side to the c side so in the direction of the sequence if you look at that bond from the c alpha to the c beta in general that bond points just so slightly towards the n terminal again when we first look at it it might look like it's stretching straight out but it is pointing a bit backwards the structure down here which side is the end terminus well we can see that do you see that that bond is pointing that way so the end terminus here the start of the chain is at the bottom and this helix is pointing up that means that if i now take this light scene that would like to point outwards if i put this close to the end terminus well i'm already pointing a little bit in the right direction after the first bond so it's going to be easier for the arginine to be located closer to the end terminus than it is to put it at the c terminus because at the c terminus here i first the first bond by definition goes a bit in the wrong direction and then i would have to snorkel in the other direction and that's one more effect there will be systematic patterns in the occurrence of amino acids and membrane proteins we'll look at that in a second but i need to show you trip defending so i already told you about the aromatic side chains but i want to introduce this concept properly aromatic side chains such as phenylalanine tyrosine but in particular tryptophan is going to have this strong effect that the mere size of the aromatic rings means that they have to stand up in the bilayer tryptophan is even cooler than the other ones because tryptophan has this one large hydrophobic part the six member ring here that will have to stick down and then we also have this polar part here the n-h group the dashed line you see here is the n-h group making a hydrogen bond so that hydrogen bond is going to have to go up to an oxygen either in another lipid or in water but that hydrogen bond means that we literally can't pull it down because if i pulled it down i would break the hydrogen bond and that's very costly i also can't pull this up because i try to pull this up i would take that entire aromatic ring and expose that to water or something that's not gonna happen so tryptophan in particular it's an anchor it's a double anchor it's an anchor upwards so that i can't pull it down and it's an anchor downwards that i can't pull it up so just to illustrate this concept this is called anchoring and it occurs particularly for tryptophan and that's why we had the w in the warp helices it's an awesome way to ensure that a part of my helix actually occurs in the head group so we're almost done looking at how different amino acids behave in a helix but i want to take one more step this far we only looked at the head group region and that's partly because the hydrophobic part is boring right that we just want hydrophobic residues there but maybe maybe i can try to introduce something maybe not charged but at least polar in the center of the membrane let's pick serine it has a small OH group this serine side chain do you see what it's doing the red one well i did put it in the middle of the membrane and it's not going to be happy about that but what this side chain effectively does this is in a computer simulation but do you see how the OH group there is effectively sharing the hydrogen bond to the backbone and the helix so while it's not going to be happy it's not going to be an unpaired hydrogen bond sticking out in the membrane if you know your amino acids there is another residue very similar to serine called threonine also has an OH group and we see the same effect with threonine so for both these residues they're not charged so that while they are polar it is possible to assert them in membrane i will be paying a bit but in many cases i can compensate for that if i make the rest of the helix really hydrophilic lots of loose scenes on the other side chains that will balance things so that in total it will still be better to put it in the membrane and then once they are in the membrane they will have this pattern where sure the individual side chain is not happy but it can get by because it's at least sharing a hydrogen bond there is a reason why i'm showing you that but i will have to keep you on hold for like 30 seconds there while i'm going to introduce one more concept before you understand the beauty of it so in the early days of studying protein biogenesis when gunnar von heiner for instance proposed a positive inside rule another group studying helix was don engelman and don engelman studied a very simple helix called glycophorin glycophorin A this is a single spanning membrane protein that is a helix an individual helix that just goes straight through the membrane fairly hydrophobic and boring from that point of view the interesting thing with glycophorin is that it has a very peculiar pattern a motif in the amino acid sequence and that means that you have pretty much any amino acids and then you must have a glycine and then you can have three other amino acids but it has to be exactly three and then you must have a glycine again and then the chain continues it turns out that this forms an alpha helix if you know something about the alpha helix is how many residues are the pertern in a typical alpha helix 3.6 so in practice these two glycines are going to be roughly on the same side of an alpha helix you can also show in experiments that these alpha helixes would actually form dimers and membranes and here the beauty is that if I have one alpha helix here and one alpha helix here and I would like them to interact the problem is that there are side chains and there are side chains in the way right they will I will bump into my peer partner side chains here unless they are glycines the glycine will effectively if I try to draw a helix here just in sort of space filling fashion this will almost create a depression on the side of that helix similarly for this other helix there will almost be a depression on the side of that helix too and you can imagine what now happens if I push these helixes together right these depressions will fit each other and they will form a very nice stable dimer you can see that if you make an NMR experiment so here's an example of an NMR structure of glycophorin A and this gx3g motif is up here so either you call it gxxg or gx3g the reason why this was a landmark discoveries first is that when this was published there was a lot of hope by many people me included that maybe this is the first motif and now we're going to gradually start finding all the motifs and be able to understand membrane protein packing just by finding the motifs unfortunately that has turned out not to be the case and that's why I mentioned to you earlier that membrane protein packing and structure stabilization is much more complicated than we think there are pretty much no other motifs but for the glycophorin helices we found something else what if some of these x residues here or at least residues close to the gx here what if they are serins or threins so they're not charged so I can introduce them in the membrane but if I now have one serine here and one serine here they can start to form hydrogen bonds with each other so the hydrogen on this serine will interact with the oxygen on that other serine so sure introducing the serins per se in the membrane that's not going to be good but both serine asparatine and threonine once I have done that they're going to create a strong driving effect to dimerize helices so that we they can form hydrogen bonds with each other instead of stealing their partner in the chain so surprisingly introducing something that's somewhat polar in the membrane the first step of introduction is not good but once you've done that they can create a strong driving effect that helps stabilize membrane protein structure so to complete the statistics on amino acid occurrence let's look at that c alpha c beta bond again again do you see here that there's a systematic pattern where the bond is pointing the first bond out from the alpha helix is always pointing towards the n-terminus this will lead to many systematic differences I already mentioned in particular that it's going to be easier for things to snorkel out from the n-terminus but it's also going to be easier to snorkel in from the c-terminus you don't necessarily have to believe me there there's a lot of statistics we've been able to do with PDB structures just calculating average sequences you can also calculate average occurrences where inner helix do things tend to occur do they occur in an n-terminus or c-terminus and this pretty much spears out so this is not just based on simple models but a ton of statistics from the protein data bank the other statistics we can do with the protein data bank is really double check my initial statement when I said that most residues in a protein are hydrophobic that turns out to be true so if I just plot this on sort of hydrophobicity all the the average hydrophobicity goes from modest to average in the head group region in the center of the membrane I'm up here virtually everything is hydrophobic and then it becomes average or even hydrophilic again but the interesting pattern is the lower part here do you see that there is no difference whatsoever between buried and exposed residues they are all hydrophobic everywhere initially we I think we went a bit wrong there as a community because some of the first structures we determined were for instance channels or things but a hole and these holes frequently contain water so those are the exceptions where it makes sense to have a few amino acid res amino acids that are might be polar or even charged but in the grand scheme of things when calculating average over tens of tens of thousands of residues they disappear so membrane proteins are hydrophobic everywhere which is a royal pain when it comes to predicting them so building membrane proteins from hydrophobic residues with only alpha helices is a very simple way to create them and that's why the first structures we found looked that way bacteria rhodopsin 1990 seven transmembrane helices all alpha helical and very hydrophobic aquaporin 1997 well it's not quite helix but you have two re-entrant helices and they pair up to effectively form one long helix I will count that as one helix that's fine exception confirming the rule but as time has gone on we've seen more and more proteins that are very complicated glutamate transporter we have a horizontal helix here multiple re-entrant regions two proteins are likely interacting with each other I won't even start to go into details so why does this happen well likely because when we started determined structures we started with the lowest hanging fruit and the lowest hanging fruit where the easy was that were easy to predict and understand be aware that there is a selection bias in chemistry and in science in general we start by finding the easy stuff doesn't mean that everything is easy so are all membrane proteins helical no there are some beta sheets too but in the interest of time I'm not really going to go through it in terms of statistics 95 percent is helix to understand membrane protein interactions between lipids starting just with the helices is going to be fine enough for now can I predict what parts will face the membrane versus what parts will face the interior that was a really good question when we first started seeing structures so we're first so the first protein structures we thought that there should be the opposite of globular ones that's not true membrane proteins are the first approximation hydrophobic everywhere the only weak signal we have is that we can look at information contents in bioinformatics what is information contents well that means that if I have a protein and I look at many sequence alignments there will be some patterns here that residues exposed on the surface here will change more quickly while residues that are buried are going to be more conserved the reason for that is simply if I have an alanine here it's going to be very unlikely that I can replace that for a tryptophan but if I'm doing it exposed on the surface here it might be possible so there is a clear signal that the buried residues have more information when looking at sequences while the fully exposed ones have less so if there's one thing you should remember here it's that membrane proteins are not simply globular proteins that are oriented inside out they are hydrophobic everywhere which makes them beautiful but also very difficult to understand because it's going to be hydrophobic packing everywhere so if it's hydrophobic everywhere can't we say anything about the residue occur of course we can based on all the things that we've gone through you now know that small aliphatic linear hydrocarbons if they are hydrophobic they will primarily be on the inside hydrophobic part charged residues they will primarily be facing water in this interface region all bets are kind of off the one thing that we can say is most common to see aromatic residues here but the patterns will serve you very well you should just be aware that there are exceptions to them Gunnar will later tell you about the positive inside rule I'm not going to go through that in detail but if you for a second buy that there are going to be more positive residues on the inside and more negative on the outside we can calculate that statistics from the protein data bank this over 10 years old results and we definitely see that trend but then we can subtract the overall trend the difference between positive and negative ones and see once we've removed that where do they occur and then you see a pretty fun peek that you have it's not entirely is to see here but you have an excess concentration of these charged amino acids both charges in the head group regions so charges actually prefer to interact with the head group regions over water partly because there are so many other charges in the head group regions the lipid head groups are more charged than the water molecules so after we've looked at the interior and the head group region I really want to stress my main love story that the interface region here the interior part is super important that's the reason why I've the entire barrier but the pure hydrophobic region there also makes it surprisingly boring head up here when we have water yeah that's fine but that's going to be like globular proteins the really fascinating part is the small region here where we can kind of do both so this is an amphiphatic region meaning that it can be either polar or hydrophobic so it's amphiphatic and that means that almost all these concepts we brought up happen there we have the anchoring anchoring anchoring of aromatic compounds happening there we have the snorkeling we have the anti snorkeling and we have this surface helices remember if the helix itself is amphiphatic it would have to lie down and then it's going to be present in this region so this region is frequently the most interesting in proteins because we can go either down or up depending on the amino acid so I hope I've convinced you now that it is possible to stabilize different type of amino acid side chains at different places inside the membrane but what we haven't talked about is how do they end up there in the first place and why can we insert them gunnar we'll talk more about membrane protein biogenesis exactly how this protein happens but here and at one other place I'm going to have like a 30 second background to it so don't worry if you don't follow this all proteins are created from a string of RNA and then it will go through the ribosome and the ribosome will have an exit tummy here we will have a protein being formed from that peptide and again you will learn much more about this later here this ribosome sits in something called the endoplasmic reticulum and actually attaches to the membrane and this membrane has a channel that it's binding to it's a very fascinating channel it's a channel of a type that I bet you've never heard of this is a protein channel we don't really call it a protein channel we call it the translocum we know the structure of that since roughly 15 years this is what it looks like from the side and this is what it looks like from the top then you can probably almost see if you look at this one from here there is a hole there in the middle what happens then depends on the helix and our hydrophobic it is if this is a very hydrophilic helix it's going to slide straight out and end up here and this is actually going to be the inside of the cell but if it's a very hydrophobic residue the interior of the translocum here is moderately hydrophobic so then it's going to stay there for a while and then magic happens that magic is that if you're looking here between the blue and the green helix in the middle there that entire translocum channel can open up that way and if this channel opens up what will happen is that this helix again if it's hydrophobic enough can diffuse out in the membrane one helix at the time good i will likely tell you much more about this but once we've done that the next question is how will they stabilize each other and how will we ensure that this helix is stable in the membrane for instance a helix with some charges or something how can it end up there let's have a closer look at it so just to check that we're on the same wavelength i'm going to do a small quiz for us me too here i have a plain simple helix with some residues there green blue here i have another plain simple helix boring just polyalanine which one of this will go into the membrane when i'm asking the translocum to process it to make things easier i'm going to add some charges to those residues i wish i could do this in person with you but hopefully you're all going to say that that's not likely to go into the membrane and that's also what i would have guessed if i was just shown the sequence and didn't know the answer and yet that is possibly the most important helix in your life it's the reason you exist this is part of a voltage gated iron channels and in particular the parts that senses the voltage and those channels are for instance responsible for closing the egg once they're personalized if this didn't happen you wouldn't exist and neither would i it's also responsible for every heartbeat in our bodies so kind of important why is that helix a membrane protein helix well beats me i almost said this is the entire channel and this is the helix there are four copies of it actually you will probably hear more about these channels later but why on earth is this charged well this is a protein that should react in response to a voltage change and that's an electric field change it makes a lot of sense to have a charge because a charge in an electric field will feel a force and if you now want to create a machine getting into movies the first step that's of course why it helps create all these nerve current sorry iron currents that in turn leads to nerve impulses and everything and yet we don't know about that molecular mechanism and how it happens the first person to realize the importance of the voltage sensor itself i would argue is yasushi or kamura so yasushi is a japanese researcher he realized that the same type of helices occur in a few small organisms co9 test analysis one of them but in these organisms he didn't only see a full voltage gate the voltage gated ion channel he saw a small voltage sensor only protein that is a protein that consisted only of those four helices forming the voltage sensor occasionally they occur in pairs and everything but that's not so important but the point is that the large tetrameria i showed you they had six helices times four but it turns out that we can cut off just the four first helices of the voltage sensor and it still senses voltage and can conduct eju protons there is also a peculiar pattern here that the first helix is hydrophobic second is sorry first helix is hydrophobic second is hydrophobic the third helix well maybe halfway but it goes in because it appears already there the fourth helix definitely not it's even hydrophilic with all those charges so somehow the previous three helices helped stabilize the fourth helix i think i have an illustration of the structure of it yes i do the helix here in the rear the blue one is the one that has all those arginine side chains that is responsible for somehow moving up and down and people fought a decade about this before we eventually converged on models so one way or another they will have to change the structure to open the central ion channel there are a few different models but we and some others believe that the helix should move roughly straight up and down maybe rotate a bit there were people assumed that there should be a paddle moving straight through the membrane and there have also been ideas that maybe you just had some sort of tilting mechanism that you don't really move up and down but you're exposed alternatively either to the inside or the outside i would argue that we won and the people eventually agreed on our model but i bet other people will disagree with me how do we know that that's true well the idea here is that if it's moving that way those positively charged residues should be stabilized by the lipids right that's why it would not work with negatively charged residues because there are there are much fewer positive charges deeply buried in the membrane rod mckinnon came up with a very smart way of testing this so they took this bolted sensor and then they expressed it in a special membrane that was not the normal mix of all this twitter ionic lipids but that you had in particular only a dotap lipid it's a bit strange lipid with only a single positive charge and this simple plain proteins that worked great when expressed in pc head groups if you just had them in dotap head groups the protein doesn't work it can't open now you could argue that might be due to the lipid complete misfolding or something but the neat test is that if you then just titrate and add a few drops of pc normal head groups to it it starts working again so it's the presence of negative charges fairly far down in the lipids that helps stabilize this bolted sensor either in the open upstate or in the closed downstate which is a remarkably cool biological process we know much more about this today people have been able to do neutroscattering studies molecular dynamic simulations and in principle i would argue that we have a pretty darn good consensus about the closed downstate too that we would normally only see when there is a polarized membrane and you can't really get that in a crystal we also have good models for what happens when we say mutate this ratio we know that this is all electrostatics so in theory if i start to alter some loops in the protein well if i introduce or remove charges there i'm going to change the local electrostatic environment and that means that it's going to be either easier or harder to move my four charges up or down and that makes sense if i measure this in electrophysiology systematically it bears out exactly where i place these charges so it's a plain simple electrostatic problem and yet one of the physically or physiologically most important processes for us let's have a second look at what they learned you already saw this one let's look at it from the top do you see this helix with the arginines exposed here now you could say that in a real vaulted gated ion channel you're going to have the pore domain and stuff here so this won't really be exposed to water but remember Yasocio Kamura's result that they can exist in isolation just this one two three four helices and in that case you would expose the side chance here to the protein there is no way that can have sorry the side chance here to the lipids there is no way that can happen the way we solve that is exactly the lipid head groups so here the helix is in its upstate and in the upstate these two residues up here in particular they have to form hydrogen bonds or salt bridges even to lipid carbonals or even the phosphate groups in the lipids but that's going to make them quite happy maybe the lipids have to stick down a little bit but they don't have to interact with hydrophobic residues essentially they snort a lot of it these two other residues here in the middle they are facing the red residues here that is aspartic acid so they form salt bridges and are quite happy if we now move this entire helix down they trade places these residues that were previously out here they're now going to face the red ones while the residues that were previously facing the red ones they are now pointing down and interacting with lipid head groups on the other side it's a remarkably beautiful example of how lipids stabilize membrane proteins and membrane proteins get their function by moving between different states and using that and in fact this is not limited to fundamental results in biochemistry and biophysics we use this to understand some pretty severe diseases it turns out that quite a few mutations in this voltage sensors leads to arrhythmia effects and we might be able to treat that with compounds that alter the motions of the helix in addition certain forms of epilepsy also influence how these helices move mutants in places where we shouldn't have muted them in the voltage sensor exactly how this happens we don't know yet it's a very active field of research but if we believe that this is caused by malfunctioning that the helix in particularly if the residue the voltage sensor can't open enough maybe we could find a way to help it open more how would you do that well remember what i said about the electrostatics if i could just put these are positive for positive charges if i could just put a negative charge up here i might help pull them up a bit but i can't just add a negative charge in a molecule randomly and make it appear in the right place so the trick is that what if i had something that was almost a lipid in fact it is almost a lipid it's a lipid chain a fatty acid chain but this fatty acid chain now has a negative charge on it so if i pick the right fatty acid chain it could diffuse to the bilayer bind to the protein and then i would now have a negative charge here that would pull it up where do i find such a fatty acid well there are a couple of different alternatives i could take a plain saturated stretched out fatty acid this one is not going to work why no we don't quite know this one on the other hand would work it's the same number of carbons here it's only here i have six double bonds i think we need at least three the difference i think that again you have to trust me a bit here is that for any type of molecule that should bind this is actually a very flexible molecule because it can rotate around all the bonds if i take a very flexible molecule and move that to a very restricted state i'm going to lose a lot of entropy and that means that it's costly in terms of free energy so it's not going to bind very well this is a more rigid molecule meaning it will not lose so much freedom when i bind it and that's going to make the binding energy better that's usually what we see for small drag compounds samir asti who worked with us and my colleagues in lynch herping fredrik ansara a few years ago she started this with computer simulations and in fact she show showed how these particular different fatty acids would diffuse around in a membrane and in particular these ones with many double bonds they're called poofas poly unsaturated fatty acid if you then take those poofas and throw them in a simulation where we also have this entire voltage gated iron cell what's going to happen is beautiful they will spontaneously diffuse and find the regions where we had that charged s4 helix i haven't told them to but they spontaneously end up there not just that we can show that they preferentially bind to a handful of residues and if we now take those residues and mutate them in the lab those proteins are no longer binding poofas so it appears that we have a handful of residues around this s4 helix that are responsible for the entire binding pattern we can even create some molecular motorspace on that even if we don't have structures bound and here do you see them in red the red s itins here are the ones located right next to s4 where we have poly unsaturated fatty acids bound but if the poofa is bound its negative charge is going to be placed right smack on top of this s4 helix that should change its local electrostatic environment and if you look at these channels they open earlier we shift them by say 10 or 20 millivolts so it's indeed easier for the channel to open which is exactly the effect we were after does that work you bet it does there have been clinical studies validating this uh how do you eat the fatty acid well this is the extra beautiful part normally in a clinical study you would need lots of permissions and take this through many steps to check whether you're allowed to administer a particular drug to humans but this is kind of a lipid how are lipids produced well our lipids are produced from fatty acids but not in our genome or anything but based on what we eat and we eat fatty acids so this is just the type of fat you eat if we eat large amounts of unsaturated fats in particular this type of omega-6 fatty acids or so under certain conditions it appears to completely obliterate the symptoms of epilepsy here this is just one form of epilepsy i don't try this in general there are drawbacks you're going to need to eat a lot of these molecules and again eating large amounts of fat is very unhealthy per se but what fredrik and colleagues in lyncher picking up try to do and they're still working on is that we try to find smaller molecules with the same binding properties so that they should also get the channels to open easier but in contrast to the fatty acids they should have orders of orders of magnitude higher efficiency so that we don't have to eat so much of them and that appears to work really well in early preclinical testing we'll see if it turns to drugs in a decade or so so that was a lot of talk about ion channels and voltage sensors the reason for that is that science frequently works by somebody starting digging into a problem and then we expand around that and this very much expanded around yes or she or camora's work today we know that there are several other membrane proteins where arginine occurs surprisingly close to the membrane this one is called five lepoxygenase activating protein and appeared a few years ago here i have a very very thin transmembrane layer much thinner than i would normally expect maybe 15 20 angstroms thick here and do you see here that we have arginines almost inside the membrane i don't know exactly what the roles of those arginines is but they're likely related to function otherwise we wouldn't see them there you should just be aware that while we normally would not expect to see charged amino acids they can occur if they are important to function so when you see it you should guess that this could be relevant for function and go after that so does this make sense as far then it's both good and bad you've understood a lot of things but you also forgotten something i've explained how why we might have residues in the inside how they are stable once they are there and how they might be stabilized by other helices but at some point those helices will have to go in good i will talk more about that too in membrane protein biogenesis but just to give you the background here they have been able to measure how much it actually costs to insert different helices in a biological environment i won't bother you with the method good i will actually cover that but the net outcome of that is that you can get a cost of inserting different side chains as a function of the side chain this makes a lot of sense functionally the hydrophobic ones are easy to insert the hydrophilic ones are difficult to insert as we would expect but don't you remember that i said it might cost 15 to 20 k cal to insert the charge here the y axis says 3.5 so something is off here either in this prediction or in the insertion or in the membrane why well remember how i stressed at the start of this lecture that a membrane is not a lipid bilayer and this is a membrane while those pure 20 k cal so is valid for a purely hydrophobic environment in a lipid bilayer that difference comes from the 25 or 30 percent or so of the membrane that is not a lipid bilayer but that is proteins so if there is one mental you you should have when i say membrane instead of lipid bilayer is this one a real membrane looks roughly something like this this is a fake system just as a model to try to understand how the interactions work i've just placed helices so that 30 of the atoms are helix rather than being lipids but you see how much stuff there is now in the center that is not really lipids sure these helices might have hydrophobic side chains but the backbone and everything here they're not as hydrophobic as the lipids so maybe when we're inserting these helices helix by helix previous helices in the membrane either the ones that came before me in my sequence or simply other helices are already there can help stabilize a charged residue and that definitely appears to be the case we down we and others have down simulations with that looking just at the isolated side chain from arginine the guanidinium ion and we love to interact with individual ions and even at the outskirts of the translocal this side chain will be quite happy to interact both with waters and in particular with slightly charged areas of the lipids and i'm not going to say if it's exactly 3.5 but between four and five kcals we can explain and understand why it's surprisingly cheap to insert charges in real membranes although it would be quite expensive to insert them in pure lipid bilators and yet there is something nagging here i'm going to keep back coming back to that isolated charge plus one charge that i need to put here in vacu that would cost 17 kcals or so between 5 and 20 and i know i'm gonna both i and others will keep inventing this magic translocal it will fall down on us magic happens in the translocal of course we don't have to go from a to c that would be expensive we go from a to b to c that might make sense but there is something deeply troubling here that one of the most famous physicists in the world have noticed and that physicist is homer simpson there is a favorite episode in homer simpson when it tells his daughter lisa young lady in this household we obey the laws of thermodynamics and we do so in my class too the problem is that the laws of thermodynamics tell you that you can't destroy or create energy so that when you go from a to c the only thing that can matter for the thermodynamics equilibrium is what the relative energies are it does not matter whether you take the left or the right stair to get there and in particular that means that if a helix is not really stable in the membrane putting it through one or five thousand translocans is not going to change that it will still not be stable that potentially leads to some interesting questions and maybe some very deep results so most proteins will be quite happy in the membrane but assuming that i have that arginine helix why doesn't this slide just a little bit out well in theory maybe it could but what would happen here is that i would now expose the end of the helix where we have unpaired hydrogen bonds to the interior of the membrane that's going to be costly i could do the same thing for polylucine or something there the driving force would likely not be particularly strong to pull it out but i would have the same problem here i can't really pull something out because it's anchored in the head group regions and if you have a tryptophan or something this effect is even stronger so in general it's true we have this situation that membrane proteins are 99% hydrophobic but when it comes to these handful of residues that are exceptions it might actually be that a few of them might not formally be stable in the membrane in the sense of thermodynamic stability if you wait an infinitely long time but they are perhaps merely kinetically stable that the translocan has forced them into this position now that we are in this position just short technically it would be cheaper to move it out here but then i would have to go over an energy barrier that is so high that that would take a hundred years before it happens essentially a kinetic stability and of course in a hundred years i'll be dead so i don't really care for all intents and purposes i would be stable anyway it's just an interesting thought and an active area of research we don't really know it but remember that in general hydrophobicity fully explains insertion so while much of i have said depends on early models indirect experiments and everything we have gone through a revolution the last 30 years there are so many membrane proteins for which we now have not just structures but high resolution structures that are so great that i can see lipids in parallel to that we can do simulations on time and length scales where i can actually simulate membrane proteins and see how a lipid will bind to them and then we can do this this is an old study of ours actually but here we have part of a voltage gated ion channel where the already solid parts are examples of lipids we found in a new x-ray structure while these licorice parts here are lipids from the simulation that we started them far away and then they've used and found their binding site so it appears both the simulation and the x-ray structure predict that you would have lipids very tightly bound to certain part of the structure if they were not tightly bound we would not see them in the x-ray crystal because they have to be the same confirmation in every copy of the crystal we can test that even more in simulations because in simulations i can check how fast the lipids are moving and diffusing around the membrane protein remember singer nickles on protein should diffuse in a sea of lipids that's not quite true if you ask the simulations the protein here is white i don't care about the proteins just the lipids the lipids far out here in red they are the ones that are diffusing freely as if they were in a lake but as i'm getting closer to the protein here we turn into yellow green blue and eventually black territory and that means that the lipids are hardly moving at all in fact we have at least one some of these in this cavities here the lipids are so tightly bound that they never move they're literally part of the structure and even around that we typically have one if not two layers of lipids that there's some sort of coat around the protein so it's technically true that the protein is diffusing in lipids but it is diffusing together with one or two layers of lipids around it the lipids are effectively part of the protein look at the size here the dark part here is almost twice the size of the protein itself we can look at that another way by checking the average jump length of lipids when they diffuse and see how that varies as we get further away from the protein things move faster and faster and faster but it's pretty much just in the outskirts of this box right that we reach the region where they behave like bulk lipids lipids that are in a pure lipid bilayer anywhere close to the protein we're going to have a significant influence property here and we can we can look at this as probability of moving a certain length and here too we see that we need to get almost a factor two away from the protein before we behave like bulk lipids simulations are cool this might not be numerically exactly correct but the properties definitely make sense and it shows that we have a much larger surrounding around proteins than we would expect just from looking at those structures this concept of how liquid a membrane is is interesting so i'm going to spend a little bit of time on that this is an old simulation of why not a pure po dppc bilayer a pure bilayer is going to be super flexible and it will even have these undulation motions if i don't subject it to attention it's literally like a film very floppy i'll move this up so that it doesn't perturb my view of you so much but what if i take this and add some other molecules in particular cholesterol cholesterol occurs in virtually all vertebrate cells but not in prokaryotes just adding 10 percent of cholesterol completely changes the properties of the system the system becomes tighter it becomes more packed and as a consequence this also becomes a bit thicker if we move from i think that's from 10 to 20 percent at 20 percent it's an even stronger effect and at 30 percent of the lipids being cholesterol it's extreme this is a completely flat planar bilayer now and it's very rigid it hardly doesn't move at all and there's almost no diffusion left of the lipids this is a very typical property of cholesterol cholesterol tends to rigidify membranes and cell walls which is occasionally useful in some case we want the structure implants it definitely makes sense to have more rigid structures and in other cases such as my blood vessels eventually the depositions of cholesterol and everything leads to a very rigid system that i might not want but note here how the specific composition of lipids cholesterol itself is a lipid too how the specific composition of lipids will significantly influence the physiochemical properties of a lipid bilayer i haven't even started adding proteins yet is cholesterol important for any function you bet you will hear more about many different membrane proteins i hope but this is a particular love story of my ligand gated iron channels they are imperative in the nervous system where when a nerve signal reaches the end of one nerve cell and has to jump to the next one this happens by releasing small compounds neurotransmitters that diffuse over say 10 micrometers or so and then they bind to another receptor and if the right neurotransmitter finds the right receptor this ligand gated iron channel that is a channel responding to a ligand will open up and create a new nerve impulse in the next nerve cell and we get nerve signal propagation for a long time we hardly knew anything about these receptors but today we do they're fascinating because they kind of Dr. Jekyll and Mr. Hyde of receptors they can be either anionic or cationic channels sorry the wrong word there no anionic and cationic they can bind any of a bunch like a dozen different small neurotransmitters and depending on what receptor i have i will either hyperpolarize or depolarize the membrane in addition to that they're very sensitive to small compounds such as anesthetics alcohol it's in fact the main charges of those compounds but they don't compete with this ligand and in fact these compounds can't open the channel they can only amplify or dampen this property i won't take you through all the details of them but i will show what they do with cholesterol so normally under normal circumstances one of these channels say the 5-HT3 serotonin receptor it's going to work great and we can measure that in the lab by adding a bit of the molecule opening it and then you see this current and then we after a while it will close and then we add a bit of the compound again and it's repeated if i now take the same channel and do repeat the experiments but in a membrane deficient in cholesterol this is what happens do you see there's almost no current at all remaining so this is a channel that is not just that cholesterol can somehow influence the channel it's a channel where the normal functionality will not work at all unless you have cholesterol present in the membrane and that's kind of interesting because this is a nerve system channel that will only occur in vertebrates vertebrates will always have cholesterol expressed but then we have bacteria and bacteria do not have cholesterol expressed which is correlated again with not having a nervous system 20 years ago we didn't have any good structures of these channels today we do the first structures were determined by Nigel Anouin very low resolution cryeum results by today's standards it was amazing then we were thrilled when this came out so this is the two-dimensional class just of the transmembrane section and what made us ecstatic here is that we could see that it's a pentamer five different subunits and each subunit has four helices going straight to the membrane there is one helix in each subunit pointing towards the poor the red part there and then one helix pointing furthest out from the membrane we couldn't see any cholesterol there though that's the acetylcholine receptor by the way but armed with that there were several groups that had really interesting hypothesis where the cholesterol might bind it could go perhaps between the helices in different places maybe in the outer parts maybe attached on the surface and Grace Branigan and Michael Klein did a beautiful set of simulations where they just made attempts placing cholesterol in different places and then they tried to reason what would that lead to and then comparing that to Nigel Anouin structure I'm not showing you this because it's right in fact it turned out to be wrong but it's still a beautiful hypothesis and I think this is how we should develop science come up with that hypothesis test it against experiment discard the ones that are already incompatible and as new experimental data comes out then we might have to discard more of our models or maybe even all of our models but it's a role model for how science should be done why am I mentioning this well because today we have structures where we see that cholesterol remember that fourth helix the one furthest away from the poor a few years ago we started getting more and more functional indications that if you do mutations in that helix under some circumstances it appears that we influence how the channel works and that could be related to cholesterol binding and then just a year or two ago colleagues of ours Chris Ulens and John Benziger they found out that there is a cholesterol binding site in this fourth helix in the transmembrane domain beautiful work that confirmed all of these things that we had thought a long idea for a long while in fact we too have been able to determine the structure of this in collaboration with Ryan Hibbs group where Jochuan Trang and my team has done simulations of the structure this is a acetylcholine receptor just as Nigel Angwis was and in this particular receptor we have cholesterol bound here in the transmembrane name where it says coal the chs cholesterol hemisexidate not only do we see cholesterol binder we see a whole range of other molecules bound in particular lipid but not where you think it might bind in this particular structure we have an intracellular domain inside the cell here that is somehow responsible for keeping all five subunits apart and what Jochuan and Ryan's team noticed is that we likely have a density with a lipid stabilizing that so this is a single lipid located outside of the bilayer because it's interacting the lipid is stabilized by the protein but the protein is also likely stabilized in its pentameric conformation by this lipid pretty cool and here too we have overall stabilization of this channel either in open or closed states depending on whether we have all these compounds present or not i could talk in hour about that but i won't in the interest of time but they're amazing channels but i will show you one last thing about them a relative of these channels is the GABA receptor and it's arguably the most important target for anesthetics so the GABA receptor itself it was only two years ago we started having structures of them one important not anesthetic but sedative that binds to GABA is diatopam you've likely never heard of diatopam but you might have heard of Valium which is the market name for diatopam what Jochuan and Ryan's team showed that if you take GABA and bind diatopam to it diatopam will rigidify the entire channel and cause it to be slightly narrower you probably see here right going to the state and that likely explains some of the effect that will have an urinary system we're changing the conductance in these channels because we now have diatopam bound if you do this experiment and happen to take too much diatopam value that can be pretty severe and then you might end up in the emergency room and then they have to give you an antidote and the antidote to this would be flamazenil which is another small compound GABA plus flamazenil when we're binding flamazenil to the same protein do you see here how flamazenil appears to have the opposite effect it makes the it makes the entire transmembrane domain slightly looser and pushes the subunits apart so it makes sense flamazenil will have an opposite effect and counter the effect of diatopam so this was merely small sedatives and everything but this works for anesthesia too we know less about anesthetics but the mechanisms are likely exactly the same I would argue that anesthesia is possibly the most important miracle of modern medicine you might think of advanced cancer treatments and everything but the reason we can't treat so many disorders now it's that we can't perform surgery it used to be the case that a good surgeon was a fast surgeon and that's no longer true we can operate on nine-year-olds today because we can sedate them controllably and also revive them this is a picture from the Morton Auditorium at Massachusetts General Hospital in October 1846 when Edward Abbott here is having a tumor removed from his neck and he survived to tell the world about it at the time we they didn't know what actually caused you to go unconscious and I would argue that that was the case for probably over a hundred years after this a very large part of modern anesthetics is mostly trial and error there was one very important result though already a hundred years ago that there is a surprising correlation not just correlation almost a perfect correlation over orders of magnitude between how hydrophobic and molecule is and how good it is as an anesthetic based on that everyone assumed that anesthesia just must just somehow work on our lipid membranes right and change the fluidity or something in our lipid membranes and then magic happens and then you fall asleep it's not a bad model based on what we knew about lipid membranes and everything at the time until some other experiments happened the last 30 years so first we started to see a few anesthetics that broke this rules and they were some of the most efficient ones the second part is that people found the ligand gaited ion channels and noticed that there are certain mutations you can make in these channels that make it impossible to sedate mice with the mutants doesn't matter how much anesthetic you have they are not susceptible to it and that's of course a very strong indication that we have a binding site for the anesthetic in a particular protein involved in our nervous system rather than acting on the membrane today I think we know roughly what those binding sites are based for instance Orion hip's work but if I draw this membrane I would one could argue that both explanations are still kind of relevant so if I have my membrane it is definitely clear that if I have an anesthetic that's very hydrophobic the anesthetic will go in here in no time no question about it and then of course I have this ligand gaited ion channel sitting here with some sort of binding site and depending exactly what happens here this anesthetic will enter this binding site or not depending on how large it is how solubility is in the lipid bilayer and how easily accessible this site is that in turn is likely related to for instance the cholesterol compounds binding here so there's a competition for this binding site from cholesterol anesthetics other small molecules it's a wealth of fun information there that I won't have time to go into why do I think that's important well because I glossed over one important anesthetic one of the most efficient anesthetics is actually xenon xenon is known for many things but not its great binding properties xenon is a small completely hydrophobic molecule that doesn't really want to interact with anything xenon will instantly partition into the lipid bilayer the question is why should xenon bind for a long time we thought that there might be some influence of the fluidity of the membrane or local pressure although Olaf Andersen's group showed some 10 years ago that that's likely not the case it's hard to measure but they were one of the groups that used to argue for it and if even they are giving up on the idea I wouldn't believe it either xenon is subject to a hydrophobic effect though but once it's already in the membrane there's not going to be a strong hydrophobic effect to go into the protein so I think the blunt answer here is I don't know why xenon is so efficient as an anesthetic do a PhD with us or somebody else and find out if you're interested it's a super interesting result that is likely intimately related to how lipids stabilized proteins and how proteins change their conformations depending on the lipid surrounding so in the last few slides you've increasingly seen me talk about not just a small helix but an entire protein the lipids small molecules other types of lipids it's increasingly a large complex of many things interacting right this turns out to be important for membranes in general and in fact if we just think about two random particles in a large membrane that should interact if they diffuse around freely according to the Singer Nicholson model well if they have to interact in general it would take a very long time before two random particles find each other in a cell it would be much more convenient if I could somehow group them together and force them to be close to each other this is a theoretical concept that physicists and people doing modeling proposed a long time ago essentially called a lipid raft so the idea with the lipid raft is roughly using many of the things that I brought up here that what if I had a domain in a membrane that was not quite as fluid as the fluid mosaic model but maybe having excess cholesterol and others here stiffen it up so that proteins here don't really diffuse so much relative to each other that would mean that these two proteins would stay close to each other and it could facilitate and speed up reactions that were dependent on having multiple partners together the only cavities was very hard to measure this experimentally because you can't really determine a structure of it I would still say that this is not common we might see it now again in cells there is some advanced light microscopy for instance that appear to see some more well-ordered regions in cells but be aware of the concept of lipid rafts 10 years from now we might tell that it wasn't until the late 2020s that we found all the lipid raft structures but today we don't have that much experimental data on it on the other hand remember my comment about each protein having a layer or two of lipids around the right this agrees kind of well with the idea of having a few lipids that don't really diffuse freely but these lipids stick together with a protein or two and this region would then diffuse freely around in the membrane so it's not an alternative to the fluid mosaic model it's more of a modification that we might have locally more rigid regions that diffuse together in the large sea of the remaining lipids so in fact a structure determination techniques have become better and better we see more and more structures that are some sort of complexes not quite lipid rafts but not that far away from it this is an example of a deep protein coupled receptor that you will hear more about later on in the class but briefly it's a protein very similar to that first bacterial rhodopsin structure I talked about also seven transmembrane helices here but here do you see how it's oriented as a dimer and we even have a bunch of cholesterol bound to it in the middle and here we're then showing it in a surrounding of lipids so this is almost a micro lipid raft per se although it's just one small dimer of a protein but again something we see in the experiment so there are more and more examples where we can see not just the protein but even lipids in experimental data and that has led to a bunch of interesting things what if we could determine general lipid interactions in experiments well not so fast that's not going to work it works if the lipids are so rigidly bound that they're really part of the structure so that they're part of the structure the same way in billions of copies of the protein in the crystal but the lipid that just interacts with the protein in different ways in each copy that would average out you would never see it and I would also like something that's not in one of these idealized membranes but in a live cell ideally that's not easy but there are some new very powerful techniques based on mass spectrometry this is not a methods class but mass spectrometry briefly works by cutting proteins into small segments and then we accelerate them in an electric field and based on where they land on a detector that gives them a relation between the charge and the mass of the fragments and if I then compare that to the amino acid sequences I expected to find I can frequently in this database identify what amino acid fragments I had if you translate that to membrane proteins and lipids the trick is that if I am able to break the protein in parts but still keep the lipids attached to it native mass spectrometry I can occasionally use this to detect at least which one of a few lipids candidates appears to be the one that's bound Carol Robinson the UK is the master behind behind these techniques but we also have an exceptionally skilled junior group Michael Landray at the Carolinus Gay Institute and PsyLife Lab who is a pioneer in this field together with Carol and again when this works I can see what I can see what lipids are bound I can see how tightly bound there are how rapidly they appear to exchange with the environment and in a few cases if we combine this with measuring what lipids I have present in the cell in the first place I might be able to use a lipidomics approach and from scratch identify are there any lipids specifically binding to this particular membrane protein and which ones in that case it's a potentially very promising technique that I think is going to take off in ways we can't even imagine the next few years so since it's so important to determine structures of membrane proteins I'm going to share with you how we determine that because that is based on understanding lipid protein interactions itself in general a membrane protein is going to be hydrophobic on the sides and that means that I can't just solvate it in water in the real membrane that works because we have lipids we have an entire lipid membrane around it right but if I were to just tear this out of the membrane I would break up the protein so that's not going to work and yet to be at least be able to take an x-ray crystal of this I need to somehow crystallize it I can't crystallize a membrane because the membrane is essentially oil but there are other molecules in particular micelles detergents detergent micelles it's essentially soap if I put many detergents like that in water their chains will face each other but since there's typically just one chain they're going to form small spherical micelles instead of bilayers if I now extract the membrane protein here but then I add a lot of micelles what I end up with is roughly the situation so I have the membrane protein with the hydrophobic parts and on all these sides I will now have micelles bound and I'm going to have enough micelles bound now that all of it will effectively be water soluble here I already had water soluble parts but the micelles per se are now also water soluble so this suddenly becomes a water soluble complex if you're doing cryo-em we can throw that directly under a microscope but you can also crystallize it this is what Hartmut Michel did with the bacterial rhodopsin and if I draw this in a schematic way if you now have lots of copies of that the part that are water soluble here they're going to be just fine they will pack and the parts that are fat soluble well in this region we would have those micelles I will leave it as an exercise for you to draw the remaining 100 billion molecules here this can form a nice stable periodic crystal and that's how Hartmut Michel got the structure of bacterial rhodopsin and eventually a noble prize for their work in 1988 if I recall correctly that one we can use directly in cryo-em and over the years we have developed this taste of trying to get something ideally even closer to the membrane this is not horribly bad but it's certainly not the membrane environment right and there could be important differences between these two so if I really wanted a membrane environment I would like something like this and I'm going to show you a way we can do that so we look at a membrane protein from the top in a membrane what we did in the detergent that we ripped this one out and then we just took the protein but what I would I really like to do is ripping out a slightly larger part maybe include a small disc of the membrane around it now this is of course a pipe dream because this part would now have the edges exposed to water here it wouldn't really work so I can't do it that way but I can get pretty darn close this is a lifted nano disc what I've done here is that to tell the truth I have ripped the protein out first but then we reconstituted not in detergent micelles but in actual lipids these lipids are then mixed with protein here and this protein is an amphiphatic protein so that on the inside both the blue and the green helices here are hydrophobic they're gonna they and their side tears will completely cover the lipids I just don't show the side tears here the outside of them on the other hand are hydrophilic so this entire small disc here will be water soluble but I can now put with a bit of black retain my protein in the middle here and make sure that the protein has a completely or almost native membrane around it so it's not going to be fully fully fully native in the sense that I actually have to do roughly what I did with detergent rip this out reconstituted and stabilize it but it's certainly a much more realistic environment than detergent if we need it so when I spoke about this a few years ago five years ago I had a very talented student in this class who got excited and said she wants to do this for our ligand gated iron channels my only problem is at the time we were not doing structural biology but thank god that student Urska Rofstein she didn't take no pranaxer so she decided to do this anyway with us and I'll show you some of the work she did not with nanodiscs with pure detergent so Urska took a bacterial protein called glyc and she overexpressed it and then reconstituted it in detergent and then she put it under a cryea microscope that we have here at PsyLifeLab and this is what a micrograph looks like once you've done all your homework if you average and average and magnify those small dots there they're going to look roughly like this you can probably count the five subunits here we see the pentameric iron channel from the top and the hole we see is the actual ion pore you're not going to be able to get a structure from only these we also need some side views and the side views would look roughly like these when you got them later and here you can I'm still in shock and awe that we can actually see atomic structure with a microscope even if it's an electronic microscope some 10 years ago we wouldn't have been able to use cryium to get that type of detail because it was just the resolution was too low we would get rough blobs and then we should somehow try to fit this into blobs but with structures of this class Urska has been able to determine electron densities with a resolution of roughly three angstrom and with that type of resolution she can literally trace out the individual alpha helices in the transmembrane region the beta sheets up here in the extracellular domain and then examine what happens if this undergoes a change in pH which for this particular channel corresponds to the opening awesome work that I'm so impressed to see and that's thanks to Urska that we're now doing structural biology in the group so some of the stuff I've told you today might be fundamental research questions and everything but just to drive home the message that there are some very direct pharmaceutical applications here I'm going to show you two maybe even 2.5 viruses in general are sadly very much on topic right now viruses is a very simple organism I won't take you through the entire class on them but briefly a virus works by somehow delivering its internal genetic material to my cells so that my cells produce more viruses and for that to happen they need to find my cells anchor to them and somehow create the delivery this particular virus it's the COVID-19 virus SARS-CoV-2 internal it looks something like this schematically you have these red spike proteins that are the anchors that find my receptors then you're having some sort of coat a lipid environment here and then you have the internal interior where they have the RNA which is very fragile and that the virus wants to protect the way this protein attaches to my cells the COVID protein in particular is by attaching the so called ACE2 receptors that are common in lugs heart blood vessels again I won't go into the pharmaceutical details here and the process works by first literally anchoring the protein there it will undergo a sequence of changes to somehow deliver its genetic material into the host cell the exact sequence of these events will depend a bit from virus to virus but schematically it's roughly the same for all of them so first you need to anchor yourself to the host otherwise you will diffuse away and then on average you would never refuse because it takes a while for fusion to happen second when you're anchored we have to tighten that anchor so we push away the water and really have full opposition the membranes should be right next to each other ideally you would somehow like to kick the host door open a bit maybe perturb the structure of the cellular membrane a bit because normally the cellular membrane would be self-repairing right they would not fuse but if they're a bit perturbed at some point they might start to heal against each other so that these two become one membrane that's deliberately a bit fussy wait until the next slide when this happens under some conditions first you will have the outer layer of each membrane fused with each other and that means we have something called a hemifused state it's now highly curved and everything and it's the interiors are not yet in contact but this is an intermediate state that is somewhat unstable so usually a short while later we have full fusion and then we're in the lower right here and at the full fusion situation i can deliver my RNA if i'm a virus to the host cell interior the way this hand waving part in the middle works is usually by having some sort of method to increase the probabilities of the membrane's fusing i'll show you that in a second but first i'm going to show a schematic movie that a former postdoc actually at the time peter casson who's now a professor at university of virginia did some 15 years ago so this is a gigantic simulation of two full vesicles it's roughly one million atoms both lipids of water here two full layers and then a similar three-dimensional vesicle here also one million atoms it's too expensive to include all the proteins and everything here but what peter came up with he's adding a small chemical linker with roughly 10 bonds or so to force them to stay together and when i hit the play button here you will see quickly that they push the water away the head groups are interacting and then you're going to see that the first two layers fusing and a short while after that you saw the inner layers fusing too it was a bit fast but you probably agreed that it followed all the mechanical steps that i hinted it would follow i kind of i'd kind of seen the movie already the way the two membranes are actually encouraged to fuse we don't know that exactly but we somehow know that there are certain proteins hemagglutinin if it's flu that contains small segments of amino acids that are semi hydrophilic semi hydrophobic they seem to prefer to go into the membrane but they're not really transmembrane helices and it appears that these drill down in lack of a better word and the host cells membrane and then they perturb that structure a bit we again we don't know exactly what the structure is these are based on nmr experiments guesses what the structure might be but when these are drilling down here are my normal lipids pointing down if i'm perturbing the structure at some point i might have a lipid or two pointing sideways and if i have the same type of perturbation in my viral membrane here at some point the hydrophobic parts here will start to get in touch and that is really going to be the part that initiates the actual fusion so this fusion peptides are exceptionally important for the virus why do we care well if you want to create antiviral drugs which is a hot topic right now not vaccines but drug that's specifically bind to a compound on the virus and would inhibit its function first you could try to find something that might inhibit the fusion peptides but even if you do manage to find something which might not be that hard viruses mutate so fast so they will likely very quickly have a new strain that is no longer sensitive to that one but the trick with targeting the most important functional regions of the virus is that if the virus the virus is exceptionally dependent on the sequence in this fusion peptides for fusion to happen so the virus does not have so much freedom to mutate away and change this to anything it wants well it can change them but then it's no longer going to be as infectious so targeting regions of viral fusion proteins that are important for the fusion itself is likely a very good strategy to develop new antiviral drugs there is of course another particular virus that we can't help to measure right now or rather not so much the virus but this do you know what this is this is the sequence that's going to save a fair share of us this year this is the mRNA sequence of the Pfizer-BioNTech mRNA vaccine against COVID-19 you can read the sequence you might find that there is a little psi here the reason for that if we use uracil it will actually set off our immune system that would target the mRNA itself but by replacing you for pseudo uradine which is the psi this is silent it will not trigger the immune system that is just a minor detail so we just now need to inject this in each and every cell in our body wait a second that's going to be problematic the RNA is very fragile so i can't eat it either and in fact if i just inject it it might very well end up in the fat or breakdown before it has time to reach the cells so we need to find a way for this mRNA to be delivered to the right place the way that is done in these vaccines is actually using lipids this type of thing called LLNP lipid nanoparticle we don't know exactly what the structures of lipid nanoparticles are yet there's been a lot of trial and error and experiments to optimize their properties but briefly we have a bunch of lipids a coat here that solvates some cholesterol other lipids there are some surfactant proteins there are maybe some lipids here that are pH dependent so that we deliver them to the right cells and then there are some relatively large water cavities here the blue parts maybe where the mRNA itself the long string would be bound and you saw in the previous slide it's a fairly long string so these are going to be large particles they work well so i wouldn't worry about taking it at all but it's still fascinating that we have something so exceptionally important then we hardly know anything about the internal structure that is of course because the mRNA vaccines was just a distant future possibility until a year ago and i bet i even know that there are a fair number of pharmaceutical companies that are now spending a lot of time and research money on understanding and characterizing lipid nanoparticles as delivery mechanisms primarily for mRNA vaccines but potentially also for other drugs imagine if you could tailor a cancer drug so it's only delivered to the tumor cells and not anywhere else in the body super exciting research i just wish that i had a structure i could show you but that's going to be up to you to fix the final medical example i'm going to show you is just a severe if not worse not so many decades ago premature infants that were born roughly before week 26 or 30 or so would die simply because they couldn't breathe the reason for that is that the babies do not yet express some protein some genes that are required for breeding because again the body does not expect to have to breathe until week 40 when you're born it took quite a while before researchers understood exactly why this happened but it has to do with the structure in our lungs in particular the alveoli responsible for the very end of the cells where we had the air exchanging air molecules with the hemoglobin and the inside this alveoli will have to undergo some structural changes when we breathe in they have to expand and contain more air and when i so and when i so they're going to need to collapse a bit right and contain a much smaller surface and that's mean that the membrane around them will have to expand and compress all the time that takes a lot of energy and nature has optimized this in ways i will show you in the next slide so avoid this taking too much energy by expressing some proteins helping but these are not expressed if you're 25 weeks old fetus it's not entirely easy to treat that but we would like to create some sort of artificial protein or something that had properties similar to this i'm going to need to show you what this pulmonary surfactant is and how it works so in your lungs there are two phases when we compress the membrane that means that you can end up with extra membrane and what this surfactant appears to do that is binding one membrane layer to the next membrane later almost like folding up laundry or something so this takes care of the excess membrane but in a very nice and tidy way then when we breathe in that i expand the lungs then i'm going to need that extra layers of lipids the laundry and then they rapidly let go of each other again so that keep things in order but make sure that there is always spare membrane available and occasionally that's even in some separate vesicles or so this works when we're adults because a few weeks before we're born it works when you when it works from your born because a few weeks before you're born we start expressing this proteins that are required for respiration and that is occasionally why you hit the baby on the bottom to get make sure that it takes its first breath taking that first breath is the first expansion that might require a bit of extra force what we would need to cure this is that we would need a protein that has exactly the surfactant properties and acts in roughly this way and then we could just give that to the baby and in that case the babies would survive you know what that is not science fiction you do that with premature babies now and you've done it for probably two decades and that's why it's routine babies will survive for week 24 or so here's an example of a chest x-ray roughly 30 minutes after administering the pulmonary surfactant and then a couple of hours after administering it part of this might be exposure but in particular you can see that there is now air in the lungs of the infant that is likely going to be healthy and lip because again now the baby is born and a few weeks later i bet it's going to express its own pulmonary surfactant it's a very temporary measure but without this measure they die because they could not breathe on their own there are a whole lot of people alive today that can that have their lives to thank for pulmonary surfactant so that's a huge number of fun concepts we're going through i will not repeat all of them but these are the conceptual things that i told you we'll go through i think i covered all of them to take you through that think about the concepts we've covered here and then i have a number of study questions for you try to answer these for yourselves if there's something you hesitate about go back to the corresponding concept or made me or write a question in atina and we will handle it from there but hopefully they should add in your studies it was great seeing you