 Yes, so I didn't quit my PhD 20 years ago. I decided I wasn't useless and it was somebody else. So I have been working with neutron reflectometry ever since then. Today I work at ESS actually in the deuteration platform DMAX. And even though Tommy promised you light politic enzymes in the title I think I would like to concentrate mainly on membrane deuteration and the creation of internal contrast with a few examples including some enzymes of how this is used in neutron reflectometry. I particularly would like to try to impress you with the ability of deuteration to allow us to see inside even single lipid bilayers. In a way that's quite different to what Tony will have just told you about NMR and neutron reflectometry are very complementary in the sense that they both give information about this environment. But the nature of the information is very different and neutrons can scatter from inside the membrane and because of their properties. This can of course also be done by x-rays, but it's the deuteration that gives neutrons the unique power to look at things with internal contrast. So that's the main theme of today's talk. Specifically we will look at what is contrast variation in membrane samples and why do you do it and how do you do it. This mainly relates to the deuteration of membrane lipids. So I'll tell you about how lipids can be deuterated and also what can be done today. What kind of lipids can you get and also a little bit about where you can get them from. Then this contrast variation and deuteration has quite an impact on your data analysis. You get much more out of it, but you also need to make some extra considerations because of it. So I'll walk you through some of the things that you need to know and some of the things you need to do in order to make the most of your deuteration. I have three examples that I hope will be illustrative so that you get an idea of the kinds of things that you can resolve by using this internal membrane contrast. I chose to do this so that you could get some understanding of how you might be able to use this in your research later on, instead of giving you a long literature review of too many examples to remember or know how exactly they were done. So I'll have all the way through this kind of a tutorial approach of trying to give you an understanding of also how things are done in addition to what results you can get. And contrast being the subject matter I pulled out my most favorite Farzai cartoon, which illustrates the point quite well that we are trying to do what Lola has achieved here. In many cases we are trying to make some things invisible, while other things we would like to create contrast to in order to see them better. In this case Harold was a bit unlucky because he was eaten by the monster but the picture shows very well what neutron contrast variation is about. It's about making the object of your interest visible. And there are three different ways in which we can do contrast variation for neutron reflectometry are the first to apply equally well to small angle scattering. The first one and most commonly use this solvent contrast variation, which means that we take exactly the same sample. And we measure it in solvents that are deteriorated to different degrees so you will have seen this probably in the science lectures too. And by doing so we can basically solve the face problem at least partially and determine the scattering density profile of the sample in this case a little bit membrane. And usually it's overall structure because you won't have any contrast between the things that are inside. And usually we get from this, where the solvent is where the hydration water is in relation to the membrane lipids. If you have more than one component in your membranes, and you want to distinguish between those two that's when the membrane deterioration really comes into play. And in multi component membranes by selectively situating some components or one component in turn, you can actually work out the composition of the membrane and also the distribution of the components, whether they're inside or above. But you still need to combine this with the solvent contrast variation to determine whether solvent is. So you can already begin to understand that there are many levels of analysis that you can do and many things to consider. There is also something called magnetic contrast variation that can be done with polarized neutron beams. And I'm not really going to go into this today but I think it's worth mentioning. It's possible to create a support structure for supportive lipid bilayers, in which underneath the surface there is a magnetic layer that's composed of a permaloid that actually has a different scattering length density to polarize neutrons depending on the spin orientation. So whether the spins are up or down relative to the surface. And this gives you two different measurement contrasts. Although, it doesn't really give you the ability to contrast match anything because the scattering length densities of the permaloid are quite high in relation to what you might have in a sample in this one is an antibody array for example. But it does give you two contrast and partially solves the face problem. It's particularly good for systems where you cannot for one reason or another exchange the solvent to do solvent contrast variation for example. Because something like an antibody array is so weakly bound to that some of it is removed if you change the solvent. But if you want to do internal contrast you still need to combine this with either a and or B to get a cleavage of what it is that the sample membrane contains on top of this magnetic layer. Polarized neutron measurements take longer than normal ones because you need to take two measurements basically, and the beam is usually also slightly less intense, but it's worth knowing that this possibility exists. So today I'll be mainly focusing on B, the membrane deuteration and what you can do with it. So for that we look at lipid deuteration, because that is probably the most common way in which things are do to it is for membrane samples. It's possible to deuterate lipids, either biologically by enzyme catalysis or by conventional organic synthetic chemistry. They all do different things and achieve different things so that's the thing here to concentrate on. There are many synthesis possible to make pure molecular species of for example phospholipids in milligram to up to gram quantities, partially or completely deuterated or contrast matched to for example D2O. But lipids are complex molecules and so are the synthesis roots so this is quite laborious. What do you see on the right here in red and some of the molecules that have been synthesized and are available. And they're not so many so mainly phosphatidylcholine, PC lipids, different types of glycerolipids tridiomeglycerides, and they are of course all made from fatty acids, where mainly saturated fatty acids are made. And I think oleic acid is the only deuterated unsaturated fatty acid that I know that's been synthesized chemically. And we heard that there was an example of a head group deuterated phosphatidylglycerolipid made at Anstel but I need to check whether this is actually the case or if I misunderstood something. Enzyme catalysis is something that we've taken as a thing at ESS because it can potentially shorten the synthesis roots to pure molecular phospholipids in deuterated form, because enzymes have very high regio and antiospecificity. So this means that you can bypass a lot of the steps that you do in normal chemical synthesis, and you can also immobilize enzymes on different types of support to reuse them as catalysts and increase the yield. So far we have managed to make POPC and POP in chain deuterated forms. And then biological deuteration lipids can be done in cell cultures because several types of microorganisms grow in deuterated media. It's important that this is minimal media as you probably will have heard from the other speakers already, so that you don't have to add deuterated amino acids into the growth medium which gets very difficult very soon. So here you can also produce something like 10 to 500 milligrams of biological lipid mixtures quite routinely, and they will have the native fatty acid chain distribution and all the different lipid classes that are present in the cell culture. They can be made completely deuterated, partially deuterated, or even in contrast matched forms. So this is the necessity of purification and separation of what it is that you need ultimately by chromatography. So other things can be used, but primarily a yeast and E. coli bacterial cultures have been used to produce phospholipid extracts, and also some glycerol lipids that they contain, and the sterols are quite easy to separate from the other non polar lipids. So here you can see the phospholipid classes and the sterols that have been purified from the rest. So purified phosphatidol coli mixtures have been made and I think also cardiolipin, and now in addition to purified aerogosterol and cholesterol. But the separation of the components is here what remains the challenge if you don't want to use the native whole lipid extract that comes from the cells. So let's look at some of the things related to organic synthesis. So these are made from different types of elements. The largest components are the fatty acyl chains. And the saturated ones are easy to make in a hydrothermal reaction in a power reactor that you can see here on the right side, it takes some time, could be up to a week or a couple of weeks to achieve a very high deuteration. And as an unsaturation or anything else that's interesting does not survive the conditions in this reactor so it has to be chemically synthesized afterwards. So that's why so few unsaturated fatty acids are available deuterated. You can see here the example of the oleic acid synthesis on the left here, and even that just making the oleic fatty acid is quite complicated you have to make two different fatty acids deuterated. Then you have to combine them specifically into a cis geometry. And the overall yield of this process is quite low and one of the intermediates this aldehyde is terribly unstable. So this is generally thought of as quite a difficult process. So it was published by Amstow already in 2013 for the first time. And they're very good at making this on a large scale. It's available from them. But recently biologically deuterated oleic acid that is actually pre-deuterated is also available from Sigma Ultrature or nowadays Merck, or something like 3000 euros per gram. We haven't tried how much of it you can get from them. It may be that there is supply issue if everybody suddenly starts ordering hundreds of grams. But the cost is not that significant, especially when you compare it to what you then have to do afterwards if you want to synthesize phospholipids from oleic acid and other fatty acids. So this is the example of the synthesis of POPC that was done at the National Australian Literation Facility in Sydney for this neutron study for which we actually travelled to Australia with Robin and Wolfgang to do the experiments. The synthesis of POPC with two different fatty acid chains means that you have to combine two different fatty acids, palmitic and oleic, with a glycerol that needs to be deuterated and with a choline head group that has to be synthesized in deuterated form until you get to the end. The percentages here are the yields of the different reactions and you realize that once you've gone through all of these processes, a gram of oleic acid does not give a gram of deuterated POPC. This is normally done on a scale of 100 milligrams or a few hundred milligrams at a time, and that is sufficient for many experiments. Why such a lipid is very nice when it is completely deuterated is that it's a very high contrast to almost anything that you would like to put in it where also the head groups have a contrast to even D2O, they are so deuterated. Then we started studying the use of lipolytic enzymes at ESS thanks to two of these European grants that we were part of, because enzymes are first of all nice in the sense that they make few byproducts, so you simplify purifications. The reaction conditions are usually mild. So this is nice and green. And because there's a highly specific this shortens the reaction sequence that you usually need to carry out and the mobilization of enzyme is quite well understood. So for lipids, the interesting thing is that there are several different lipolytic enzymes that acts on different parts of a phospholipid molecule where they either can be used to remove the choline, the entire head group. One or both of the fatty acyl chains. Sorry, I don't know why this is jumping. And quite specifically, they only act in those positions specifically. And so we were looking at developing methods by which we can specifically swap deuterated fatty acid chains for hydrogenous fatty acids in phospholipids due to make the synthesis faster. And commercially, there are lipases that attack the one position and phospholipase A2 that are available. Ours came from Novozymes as a gift. This sounds all very well, nice and well, and this example of deuterated POPC also looks quite simple that you first take a lysolipid that only has one chain. And then you put on a deuterated oleic acid chain in the two position, and then you remove the undeuterated palmitic acid chain that was there. And you replace it with a deuterated palmitol chain ending up with a chain deuterated POPC. But the reaction is complicated. We are making the enzymes work in two directions here, one the hydrolysis direction, which is what they normally do. And then the reverse direction in esterification so controlling the enzyme and water activity is something that is very strict in here and also lipids are soluble in and not soluble in water and the enzyme is so these are bi-phasic reactions. We also need to monitor what is going on and follow the how efficiently you've immobilized the enzyme for example and how well it still works afterwards. In other cases, the lipids that you buy contain significant byproducts. For example, commercial POPC is up to 20% OPPC where acyl migration has changed the positions of the chains. Ours is a lot purer than that because this enzymatic esterification is much, much more specific and less susceptible to this acyl migration. This involves doing a whole lot of different types of analysis to verify that everything is where it's supposed to be and everything is as deuterated as it's supposed to be. So we have after this also made a similar version of POPE chain deuterated by DSS. And biologically, I said yeast is one of the organisms that is used for doing this. The reason is that Pichia pastoris yeast is kind of industrially used yeast for protein production. And it tolerates D2O very well, and it can live on glycerol as the only carbon source and doesn't need any other foods other than minerals. So glycerol, deuterated glycerol is what is used to create the deuteration and lipids. The lipids are synthesized primarily from that. Depending on how much lipids you need, you can grow them either in flasks or in bioreactors and you can get something in the region of hundreds of milligrams to grams of deuterated lipids quite easily. And the reason why Pichia pastoris turns out to be interesting, in addition to being useful is that it has these de-saturated enzymes that most other yeasts have lost. And it can produce polyunsaturated fatty acids in quite reasonable amounts. You can see here on the right hand side the phospholipid composition and the fatty acid composition of both deuterated and normal hydrogenous yeast grown at 30 degrees, I think it was. The phospholipid composition is preserved in D2O, but the fatty acid composition is changed, so that fewer polyunsaturated lipids are produced and a lot more oleic acid chains are produced in D2O. We know that this is because of this kind of kinetic isotope effect, both on the growth rate of the cells, but also on the reaction rate of these de-saturated enzymes that progressively work from oleic acid to create first linoleic and then linoleic acids. But this difference can be reduced if you change the growth temperature of the cultures to lower temperatures. If you want to use lipids that are deuterated like this biologically, they come as a mixture of all the lipids that are in the cell, which you need to extract. The main problem here is that lipids are not overexpressed by the cells. And phospholipids, they are part of the cell structure. You cannot make too many of them before things go wrong. Some glycerolipids can be overexpressed in the sense that they are storage fats, and some kinds of organisms can grow a lot of oils. But in general, you have to grow a lot of cells to get a reasonable quantity of lipids and go through this procedure of extraction and chromatographic separation. There is quite detailed analysis to understand what it is that you have, because the growth conditions also when you harvest cells have an effect on the composition of the lipids. Here you see in this graph, quite a recent analysis of the fatty acid composition of de-chiavasteroid phospholipids, where you can see all the different classes that we've separated by thin layer chromatography. And each class has its own profile of fatty acids. The phosphoridolethanolamines being most unsaturated with cardiolipin and the PC fraction being the most saturated fraction, for example. So, knowing what you have in experiments is important both from a bio-physical and biochemical point of view, but also for the analysis that you have later. As is the degree of deuteration, so quite often part of this information comes, for example, from mass spectrometry. For example, the deuteration of ergosterol was verified by us like this. And the rest is often coming from NMR, looking at deuteration across the chains. E. coli, I forgot to mention one thing, which is that it is also possible to produce sterols like ergosterol, which yeast has naturally, it's not difficult to purify, but also cholesterol, because there is an engineered yeast strain that is used at the ILL, a recombinant pichia strain that is able to make cholesterol instead of ergosterol. So, due to per deuterated sterols are actually possible to make. Another modification that has been made in E. coli is that it's been adapted to produce PC lipids, which normally does not make at all, it normally only makes PE, cardiolipin and PG. And this is something that was done by the ILL in collaboration with Selma Marik, where they could also show that by selectively feeding the cultures, combinations of hydrogenated and deuterated glycerol and coli for the head groups. These kind of B2O contrast matched lipids can be produced that make completely invisible nano-discs for membrane proteins in D2O. But these lipids are also, they come as a mixture of all the different fatty acid chains and combinations where the PO chains are the predominant, there's also a fraction of this strange cyclo lipid that's specific to this organism. And so what you always get is still a mixture, even if you separate the PC fraction into the different chain lengths by reverse phase chromatography, and these two are near to impossible to separate from each other. Then, if we take a look at a little bit about what this is all for. It's really about the fact that in neutron experiments, the deterioration of your sample not only affects the signal that you have in your experiments, which is represented here, a simulation of different degrees of deterioration in a membrane. What you normally get out of an analysis like this is just the scattering length density profile, which is this black line here. But what you want to know is what is the structure of my lipid bilayer, so you have to be able to interpret this. And the idea of solvent contrast variation is that you identify the solvent and also where all the different components are. So in this hypothetical example of two lip two component membrane where one lipid is deuterated, you see in D2O, the hydrogenous lipids, you get one scattering curve. If you change the solvent to H2O, you see only the deuterated lipids, and you get a completely different scattering curve. And it's by analyzing those two together, that you derive the scattering length density profile, but then besides this there is still much more to do. And that comes to being able to interpret the scattering length density profile. Here, we're going to look a little bit on what is it that you need to know in order to be able to analyze your data and what do you need to do in order to get the results out and interpret them. So you need to know the scattering length density that you expect your molecule or your molecules if it's mixture to have. And that includes knowing the degree of deuteration, things are not always 100% deuterated, there can be 95% deuterated and that's more than enough. Sometimes they're selectively deuterated, like this lipid here is only chain deuterated, but not head deuterated. You need to know the molecular structure. You need to know either the molecular volume of the density, because the scattering length density is the sum of the nuclear scattering lengths, every atom, divided by the molecular volume, if it's the molecular scattering length density that you want. And the same thing for these different groups if you want to know the chain scattering length density you want to know the volume of the lipid chains. So how do you know this? There are quite a lot of examples of data that have been used and then many of much of it comes from for example x-ray diffraction and neutron diffraction measurements or density measurements. But it's becoming more and more common that the volumes of the different lipid groups are derived from molecular dynamic simulations. Here is probably the earliest example where this was done on purpose to compare to x-ray diffraction data to derive actually the component specific volumes of all the groups along the lipid chains and head groups in comparison to this experimental data if it's interesting for you to look at and they correlate quite well in the overall volume. So this kind of information is quite frequently used to compute scattering length densities. But it's important to understand that it is the right molecule and it is the right conditions and also what is the uncertainty inherent in the data that you're using as well as the analysis that you're doing. So according to this there are a number of things that you need to understand about the sample geometry and what you're analyzing. So in order to make sure that your analysis gives a physically realistic structure we normally use constraints that relate to the structure of the molecules that we are analyzing. If it's a lipid bilayer then the area or molecule for the head group and the lipid tails should be the same. Otherwise it's not a bilayer it's some kind of curved structure if the area is larger at the heads of the tails. So that's typically a constraint that is supplied. What does that have to do with scattering length densities you wonder. So basically it's also used in combination of constraining the amounts of solvent present because that is what affects the volume fractions of the lipids in different areas. It has something to do with fitting because this is what the data normally looks like from your analysis. This is real data relating to DMPC but I've made up the arrows to make a point. In this data you would like to understand that the structure is realistic so you can compute the area per molecule from knowing the molecular volume and taking the lipid volume fraction and dividing by the thickness of each layers. And that's all very well. The reason why you use two contrasts is that if you have two contrasts you can say that all the structural parameters must be the same. The only thing you are changing is the scattering intensity of the water. So that is a major constraint. But the interesting thing here in this table is the arrows the parameters are all the same but the numbers for the arrows are different between H2O and D2O. That relates to the sensitivity of each contrast to the parameters that you are fitting. The difference depends on what is the chemical composition or the deuteration of your molecules and the solvent contrast. So D2O is better at determining some things while H2O is better at determining other things in this case. And what matters in the end is the parameters that you're interested in so in this case the area per molecule is calculated to be the same everywhere but the arrows in the H2O data in this case are smaller. So that is what determines your fit, basically the best contrast. I should add here that the area per molecule is this one that actually includes all the solvent that is present and is used to ensure that the space filling and the bilate geometry also works from that point of view. It means that there are two tails per every head and not necessarily that there are exactly the same number of molecules on both sides of the membrane. I told you that it's possible to obtain deuterated biological lipid mixtures. How does one do that for those? Is it even possible to do this kind of analysis if we have a total lipid extract from some organism that makes all kinds of things? Yes it is. With a pinch of salt, which I will add at the end. So these are the Pikia pastores, phospholipids or some of them, where it makes several different head groups. And they are either deuterated or undeuterated depending on which ones you want to use. Or if you're really unlucky, your supervisors ask you to mix some of each and then you have both. And you know from analysis, you know the phospholipid composition. Unfortunately, the head groups are polar. And all of them except for the phosphazole choline group also exchange protons with the solvent. So they all have a different scattering length as in different solvents, depending on the degree of deuteration. So I'm going to bring you here for four different ones. But yes, if you know the composition, it is possible to calculate the average head group scattering length density, and also calculate how much it varies across different contrasts in the solvent. And this is important because the differences are large enough to be meaningful in your data analysis. It is not possible usually to try to analyze the data with a single number in all the contrasts. So if the fatty acid chain composition has been analyzed, you can also calculate a scattering density for the chains that are deuterated and non deuterated. Thankfully, this does not change with the solvents. And the picciapastore example is interesting because all of the polyunsaturated fatty acids mean that the scattering length density in both hydrogenous and deuterated molecules is actually higher than what you would get for POPC. So if you try to use values for POPC as the standard membrane phospholipid representative, it will go wrong. It will be next to impossible to fit data from picciapastores lipids using the scattering length density for this average lipid POPC. So yes, you can calculate all of this, but in the estimation of the scattering length density, you still need the information about the molecular volumes. So this relies on you being able to use a source for the molecular volumes that is at least of the same level of accuracy as the analysis that you're doing, and that's the take on point. Here we use the steps of molecular dynamic simulations values that were derived from POPC. So it is not exactly the same molecule. The accuracy of neutron reflection being able to determine the scattering length density where there is always an inherent error. The potential systematic molecular volume error does not count. It is considerably smaller or at least not larger. So that's the sort of consideration that you would want to take if you're using some of these assumptions. When it comes to analyzing data, it's more than possible to analyze data from even very complex mixtures as long as you know these numbers that you need to describe the molecules. So you see in an example, exactly this picciapastores lipid bilayer in a six layer structure because there is also an additional molecule called amphotericin B on top. So there are layers that describe scattering length density is a thickness of roughness and the solvent content of every layer. And they are all constrained to be the same in all construct contrast except the lipid head groups that change with the solvent. And in fact, the lipid chain region because it contains some of this amphotericin B, which exchanges with the solvent. So this is an example that I will show you towards the end. And in the end you refine a fit until all different solvent contrast fit hopefully equally well. And you do this calculation for the areas for molecule to the degree that it's possible to understand that you're still modeling a physically realistic structure. Then I promise some more enzymes. This is very old. This is in fact the experiment that Tommy was referring to one of my first at the ILO that was looking at the reaction of phospholipase A2 with lipid bilayers, because it is what's called an interfacial enzyme. The reaction rate of this reaction in which it splits lipids into two parts is something like 10,000 times faster on a lipid membrane surface compared to monomeric short chain lipids that are free in solution. So, in this situation understanding the crystal structure or the catalytic site mechanism of the enzyme does not nearly describe to you how its activity is regulated because it depends very much on the surface of the membrane. So I'm here to understand what happens during this reaction in a membrane when the membrane, when the enzyme interacts with the lipids and they are converted into lysolipids and fatty acids. The enzyme is very, very widely distributed in different biological systems. It's not digestion. It's part of the inflammatory cascade. It's in a lot of venoms like this cobra venom that I studied, and it also participates in cell signaling through these lipid messengers. So I think this was the first example, probably one of the few existing examples of intramolecular contrast variation, where a half deuterated POPC is in principle split into a deuterated half and a non deuterated half by the phospholipase A2. So by using this and the understanding that both lysolipid and the fatty acid should be more water soluble than the phospholipid. So the reaction should be able to tell us something about what happens to these two reaction product molecules. And so it does, it told us something that was a bit surprising, but with hindsight, maybe not so much. If one looks at the hydrolysis of pure POPC, the reflectivity drops in a way that tells us that about half of the lipid membrane is destroyed and solidifies the way by phospholipase A2. So what would suggest that half the membrane is digested? If we take this half deuterated molecule and do exactly the same measurement, the results and the curves look different. The results say that the amount of lipid doesn't really change. The membrane just gets a little bit thinner. So these are two analysis where we did not refine to the same structure because they were two different samples, but we refined to the same global phenomenon if you like. So the same thing has to happen in both samples. In case of the half deuterated molecule, well, yes, the thinning corresponds to about the thickness of the lipid head groups, but not much else seems to happen. So what this is consistent with the data from the unlabeled molecule is that it is the deuterated half of the molecule that is solubilised and taken away because it is invisible in the D2O contrast. And the only thing that you see is the hydrogenous fatty acid parts that are not solubilised and they say where they were before. Now, those measurements were quite long. You can see the measurement times for hours. So it's possible to do it today and already some years ago it's impossible to do much faster measurements, for example, at the ILL on the newer reflectometers like FIGARO, where this example is from like two-minute measurements per curve. So there is already an ability to some degree to follow enzyme kinetics function of time. What you want to do for structural analysis is that you want to measure quite a broad Q range, broad scattering angle range. So that you can make a structural model and unfit, for example, the shape changes to the changes in the membrane structure. At the ESS, my legacy will be the design of the Freya reflectometer that was actually made for this kind of purpose in the end. It's a very fast instrument that measures a very broad Q range simultaneously. So when it is finally available, it should be able to measure something like sub-second time scales when ESS is at its full power. In the early days, when ESS is at half power, I'm sure the speed will always already be at the second time scale. Then my second example is looking at membrane asymmetry, which was also related to a question that I caught towards the end of Tony's talk. Selective deterioration is a good way to actually look inside a membrane and see whether there is an asymmetry and other inhomogeneities in the distribution of molecules. So symmetric bilayers can be made on solid supports on purpose by Langmuir blodged deposition, for example. But sometimes it happens on its own because of the properties of the molecules and we've been doing studies that involve using cardiolipin for some time for the purpose of studying mitochondrial membrane models. Actually, in collaboration also with Gerhard Gröbner from UMO. Another different example, though, where cardiolipin is negatively charged and it's repelled by the negatively charged silicon dioxide supports that are used for the membranes. And if one mixes it with butyrated POPC, it's not very surprising that the distribution of the cardiolipin turns out to be asymmetric, so that there is more of the cardiolipin towards the solution, where there is no negatively charged surface than on the surface side. The asymmetry is not very pronounced. There's like 30% on the outside and 13% on the inside, where the nominal composition was 20-volume percent corresponds to the 10-mol percent of cardiolipin. But even such a low-level asymmetry is quite easily resolved in a measurement using four solvent contrasts and this one deuterated lipid. It happens in systems where we've looked at where ubiquinone goes. It's a very long hydrophobic molecule. It's actually longer than the thickness of the typical lipid bilayer. And it's an electron acceptor for many enzymes in mitochondria and this implies that it needs to be able to go to where all these many enzymes are. And with neutron reflection in the same membranes with deuterated POPC and cardiolipin, we can quite easily show that ubiquinone actually sits exclusively in the center of lipid bilayers in its own layer. That's about four or five angstroms thick and about 50% of the volume is definitely made of ubiquinone that we can see when we use deuterated POPC. So many details can be resolved even in relatively small layers inside a membrane when the compositional difference is large enough. So this could be potentially used to detect many other molecules. Something that I've been working on for quite a few years now is amphotericin B that is an antimicotic antibiotic. So it works against fungi and some parasites and mold, I believe. And it's a membrane binding antibiotic that inserts into lipid bilayers. It works by binding ergosterol that is only found in these fungal organisms. Why it's an interesting example here is that exchange is a very large number of protons with a solvent. The change of amphotericin scattering length density is pronounced going from H2O to D2O. It has good contrast to nearly all solvent contrasts except the silicon supports and the water matched to it where it is as good as invisible. So this scattering length density change across the solvent was actually large enough for us to be able to detect its insertion into membranes where the lipid scattering lengthness does not change with the solvent contrast. And here is the earliest example of that, where we looked at both these hydrogenous and third-dutrary epistoric lipids that we extracted, either having removed the natural ergosterol present or with it still there. And it's quite easy to build supported bilayers from these things nowadays, but it took some learning. So you might ask, as the first thing, why is there a difference in thickness between the red ones and the blue ones? Aren't they both yeast lipids? Yes, they are. But if you remember the analysis that I showed you which said that the deteriorated yeast make fewer polyunsaturated lipids, that's the difference. The polyunsaturated lipids that are present in the hydrogenous yeast normally they make the membrane quite a lot thinner. And the sterols, as they normally do, they thicken the membrane somewhat. But this effect is much more pronounced in the polyunsaturated membrane than in the largely monounsaturated, deuterated membrane. And for tersin B, what it does is this creates this kind of layer that is as thick as the membrane on the outside, but that is very hydrated. And it inserts into the lipid bilayers. It's visible in both hydrogenous and deuterated lipids. And the interesting thing is that we can see it extracting ergosterol into this amphoteric layer. The scattering intensity of this layer changes most dramatically when the ergosterol is deuterated ergosterol that is pulled out. And quite a significant thinning happens in the polyunsaturated membranes when the ergosterol is removed. And it goes more or less to the same value as the membrane that didn't have any ergosterol to begin with. So this is probably one of the most complex examples that I have studied. But you can see that many things can be unraveled. The last example is where we actually used some strains which certain genes were used to certain genes for up or down regulated to change the basically the function. In some cases, amphoteric resistant and some cases more susceptible to amphotericin. The lipid composition changes mainly related to the accumulation of squalene, which is the precursor to all sterols including ergosterol. So we were able to study where the squalene is also in the center of membranes and what it does to the ability of amphotericin to extract ergosterol. Where we could identify the deuterated ergosterol quite clearly also as a drop in reflectivity. It turns out that the squalene in the center of the membranes blocks amphotericin. It can't go through and it is only able to extract the ergosterol from the outside of the supported bilayers. And the way in which this happens is correlates with the resistance in the strains that were created by RNA interference in this case. So to conclude quickly, lipid deuteration is a powerful tool to understand membrane structure and function. Lots of different lipids can already be deuterated in different ways. There are different advantages to synthesis and biology, but both of them have challenges. And I would say that they are both equally costly. There is a certain cost involved in deuteration and in the end the way in which you do it doesn't really make any difference. But that analysis of your deuteration degree, the purity of your compounds and the molecular composition of mixtures is extremely important for you to be able to use this in your neutron scattering analysis. And deuteration really allows you to exploit the power of neutron reflectometry and sands to the full in that it can reveal all these details about the internal membrane structure and composition. But it's important to remember that the analysis is only as good as the information and assumptions that you put in yourself. So this is something to remember. I would just like to conclude by showing you what is already available at ESS because some of you said in your letters and motivation that you'd like to know what will be at ESS. The deuteration and crystallization service proteins is already in operation and has been since 2019. And you can see here the gang of people working. There's myself and Anna Leung who work in the chemical deuteration lab in MediCon Village. We've had help from Fatima this year and we have two postdocs on grants with Lund University, Yendi working on lipids and Jife on synthesis of surfactants. And then we have a biological deuteration lab that is actually based at the LP3 protein production platform at Lund University, where Zoe and the biotechnician do the biological cell growth and also purifications of proteins and crystallization of proteins including some testing and data collection with x-rays at Max4. The platform is run by the other Knecht that you may recognize. If you have any questions, we have run proposal calls already twice to basically learn how to manage proposals and what molecules we need to be able to provide. There will be at least one more of these before ESS becomes operational, but the data is not yet been determined. So if you have questions, you're more than welcome to email us at this address. We are not alone. There are many other deuteration facilities in the world and we're actually organized into TUNET as an international deuteration network that is still coordinated by me at ESS. But we have grown from a European grant collaboration in 2015 to a worldwide network where we have members from the American, the Australian and also the Asian neutron facilities and also some university laboratories where deuteration is used. For example, in drug studies, we have our own homepage and you're welcome to go and see who we are and what we do. And we also have a Twitter account where we occasionally publicize what we are doing. I would just like to conclude by saying that there is also a book that I wrote two years ago with my colleagues from ESS that describes some of these deuteration methods. Also lists a lot of the almost all of the instruments in the world where you can do deuteration, reflection and sense and other experiments in life science research. So it may be something that's interesting for you to check out. And I'd be happy to take any questions if you have. Thank you very much, Hannah for this very clear presentation. So do we have any, I have a lot of clapping hands here so so your contribution was very much appreciated. So is there any questions for Hannah, please go ahead. Any chance to get your favorite deuterated lipids. I have a question on. In one of the last slides that you showed you said that this am for terrorism be like it, it extracts and go stroll and puts it in a sponge phase or like it takes it out of the membrane and keeps it in the solution. I didn't follow that part. Yes, so we can go back to the presentation to show you. Yes, it binds Augusta roll this was known. But what was not known what happens does the complex stain the membrane it was thought to form pause. But what we could show was that it this is not the case it actually the Augusta roll comes out it's quite clear from the deterioration that that this this happens and then the sponge layer is just a name that was given to the membrane layer by some other people who discovered it without really using aqueous condition to show that it's something that just gets build on on the surface of the membrane and stays there with the Augusta roll. And what with the Lannister all does it like does it have any effect with. Yeah, Lannister all accumulates it's it's the last precursor to Augusta roll. In the synthesis we've done also some experiments with Lannister all specifically in membranes and it turns out that I'm for terrorism doesn't do anything to the Lannister all. And same thing with cholesterol it's able to go to the membranes and bind cholesterol but it doesn't is not able to pull the cholesterol out the same way as Augusta roll. Any more questions I think they call us a question. Hello. Hello yeah thank you so much for your presentation so I would like to ask you when you do. How big is the area of the membrane of the surface that you are probing and is it possible to like quantify your signal in some way like is it done to know like how how much or how many like. Lipids are contributing to respective like signal. Okay. Yes the area is quite large. In most places where you go today it's it's off the order of, I would say, three centimeters by five or six centimeters so it's quite a large area. And then the signal is averaged across the entire illumination area we don't reflectometry we don't have resolution to the lateral distribution at all. What we get from these experiments we get basically the density of lipid per unit area. The upper molecule if you like on the surface so you can you can calculate how much how many lipids are in the illumination area if you know what it is it's not 100% exact. Because the neutron beam is divergent. It is possible to calculate the foot the complete footprint and know it exactly but it's something you have to do separately just saying that the beam is this wide when it comes out of the beam hole. It doesn't give you the right answer but in that sense yes it's possible and the lipids that contributes the signal we're talking about nanograms in a normal experiment still it's not not that many. Thank you so much. I have one more question. So, in one of the experiments that you're using this cardiolipin like you know I saw that you use 10% of cardiolipin with your PC. How did you manage to make the bilier because I tried a lot but it the vesicles never seems to burst on. This is big of my specialty knowing how to deposit vesicles in the bilayers. So, for all negatively charged lipids, you need to do two things which as a physical chemist you might understand and you need to screen the repulsion between the vesicles and the surface. If they're both negatively charged and most often you do this by putting in sodium chloride at something like physiological concentrations or 100 million moles or so. And then the other thing that you need to do for negatively charged lipids is that you need to add calcium something divalent, divalent cation. And it's the exact mechanism of why it causes the membrane to form is this not certain but what I think is that it simply makes the lipid bilayer stable the connection to off the lipids between the surface to the surface more stable so that the vesicles are able to break open and stay there. Now this something like one to two million molar calcium should be more than enough. Some cases we've also used magnesium chloride. But it's whenever you use calcium and magnesium causes problems in that all the vesicles want to aggregate together. So you have to be a little bit careful in how you how you how exactly you do this. Okay, but do you add this in the buffer itself like. Okay, yeah, and then after you found the membrane you can change the buffer to whatever you like the man once it's formed the membrane stable. Okay, any more questions for now. When they want more of them to write the lipids, it's a chance to get 5% off if you get. I'm not sure I can produce them on the spot and by the way we don't we don't charge anything at the SS for the time being it's while we are still learning it's for free. 100% off. Very cheap. Now, I think it's, oh, there is a hand there. Yeah. I have a kind of, I guess it's quite a simple question about the calculating SLDs. We are saying that getting the correct density for each lipid is obviously very important, but the. Less common lipids that the molecular volumes aren't really well known or haven't been calculated. But it's okay as long as it's within her experimental error. What kind of allowance do you have in the molecular volume that would still be okay with an experimental. I think I worked that out once which is why I think I can say that it's if you look at the scattering length densities that if you're trying to from your measurement figure out what is my scattering length density. Then it's not unusual that the error in the number is something like plus or minus 0.1 or even 0.2. That's the quantity that you're interested in. What am I able to determine from my measurement and data analysis. And if you convert that number into a molecular volume difference in the lipid chains for example. And that that's what gives you a feeling so I think I worked it out that it's okay. The difference is up to something like 100 cubic angstroms between the value that you're using and the actual chain volume and this really makes no difference at all. There is a difference between whether what contrast you're measuring. So you always need to pick the most sensitive contrast so you get the smallest error to understand this but it's in that ballpark and often the values that you can use that probably are. Not that close or not very different. It's it's I think it's unlikely that you'll end up with much larger errors than that from using even the slightly wrong molecular dynamic simulations.