 The way we determine those structures has changed a little bit over the years. Historically, the way we determine things was with an x-ray source, and that's still one of the most important techniques. And the way we do things with an x-ray relies on the diffraction properties in physics. So if we have something small like a crystal, now obtaining a crystal of a protein can be difficult because we need to overexpress the protein first, and that's not easy because many cells die if they express too much of a particular protein. Then we need to crystallize it, which is even harder. Most proteins don't want to be crystals. But assuming that you're lucky and have a crystal, you can put that crystal, a few milligrams or micrograms inside an x-ray beam. And today we don't even use x-rays. Well, we use x-rays, but not a traditional x-ray machine, but a shrinkage from light radiation facility, such as Max 4 in Lund in the South of Sweden, probably still the brightest one in the world. What you see here in the lower row here is an experimental setup when they mounted things. And in the middle here you see a typical diffraction pattern. The white part there is kind of the shadow of that mounting piece that we had the crystal on. And if you're lucky here, today we would not do this manually, but throw this in a computer and ask the computer to determine what is the corresponding electron density. Because those x-rays, they're not really interacting with the atoms, they're interacting with the electrons. And whilst we have that electron density, a skilled researcher will sit down and, with the aid of a computer, build the actual model of all the amino acids, the protein, which is the backbone you see here inside the blue density. And that's certainly possible to do for many proteins. But there are a couple of these limitations that I had up there. And in particular for membrane proteins, they have proven to be very difficult limitations. So many years ago there was another technique called cryo-electron microscopy that for a long time was the ugly duckling of structural biology together with nuclear magnetic resonance arguably. NMR is a very nice method in other ways. It occurs at room temperature so that you can measure dynamics. But NMR has never become really popular as a way to determine structures. Crye, for a long time, was jokingly referred to as blobology because the resolution was so low for technical reasons. It's simply difficult to detect electrons in these microscopes. I'll show you how it works in a slide. But then some 5, 6, 7 years ago there was a revolution in the field with a new generation of direct electron detectors combined with much better microscopes that suddenly enabled you to take literally microscope images of proteins, which is a bit of an oxymoron because you shouldn't be able to do that. Normally a microscope uses light, but the wavelength of light is a few hundred nanometers and you can't image something that is significantly smaller than the wavelength of light. This would be individual nanometers almost. Well, a bond is one angstrom. What happened, though, is that these techniques became so great that we could suddenly use them to determine structures down to three angstrom. Two to date is even occasionally around one angstrom. That's going to remain an exception. And there's this beautiful nature paper alluding to the television series saying the revolution will not be crystallized and one of the key advantages of cryo-electro-microscopy is we do not need a crystal and the amount of protein we need is frequently just a tenth or even less than that compared to X-ray crystallography. The people behind these techniques historically not so much, well, partly the direct electron detectors but both the techniques and in particular all the computational models required to do this you have Jacques Deboucher, Joachim Frank and Richard Henderson they get the Nobel Prize in Chemistry just a few years ago for this. I think it was 2017. The way these techniques work let me compare this to X-ray crystallography. That's probably the easiest part. In X-ray crystallography we would take a sample that has to be a crystal. If it's not the periodic crystal we're not going to get a systematic diffraction pattern and then we wouldn't see anything else. So if you have a crystal that's imperfect you're going to get a low resolution diffraction pattern and then you're not going to be able to determine the structure. But if you have a good periodic crystal you're going to get a diffraction pattern exactly like all the wavelength labs you might have had in upper secondary school or at least if you studied physics at university. It works great. There is only one problem. A unique crystal will give rise to a unique diffraction pattern but you can't go backwards because you're losing the faces when we're just determining the amplitudes here. And that's why we had this problem. I can't just tell a computer to invert this but I have to build a model and check if my model would produce this pattern. There are some techniques to go around that but again you still need a crystal and it's not one-to-one translational. In cryoelectron microscopy instead it's much easier. You just take a sample and freeze it and then we're shining a beam of electrons on it and they're having this detector at the end where we're just imaging the electrons. Like it was a microscope but instead of having light we're having electrons. Electrons is a particle but if you remember your quantum mechanics they have this wave particle duality. So a particle also has a wavelength and it turns out if we accelerate electrons to something like 300,000 volts they're going to have wavelengths that is a small fraction of an angstrom and then we can image with electrons. You've all seen electron microscope images. The problem with that in structural biology is that we can't use traditional electron microscopy because then we're going to damage our proteins. Remember, radiation damage, proteins too are sensitive to it. So we're going to need to use a very low dose and if we use a very low dose the images are going to be very noisy and they're going to be so noisy that you can't tell the protein from the background. And in that case we're going to need to use lots of computer power to take these images and map back what the structure is. So this is very much relied on modern computer algorithms. The other problem is that it's very important that I freeze this thing in blue here but don't have eyes because if I have a normal eyes I'm going to have a crystal and then I have diffraction from the eyes rather than the protein. Jacques Dubochier came up with this really cool idea of plant freezing that we drop this very quickly into liquid ethane so we freeze the water molecules so quickly that they don't have time to form eyes. They literally just stop moving. The reason I tell you this is that in less than the last 10 years this has gone from literally just being blobs where you could just detect the shape of a protein into being things well resolved enough that we can place individual amino acids and this is still an old structure but the best ones today it's so good that you can even see the holes in an aromatic ring. It's like night and day. We use this technique all the time in the lab but I can still hardly believe that it's true. This is a particular membrane protein called TRPV1. It's a pain and heat receptor that occurs in your cells. It's also an ion channel and it's a bit fun. This is the first high-resolution membrane protein that's determined with cryein by Yifan Sheng in San Francisco and it's a protein that also bind capsicin which is the small compound that's present for instance in chiles. And you see this relation, pain and heat that's when you eat chiles you frequently get these sensations. It's through the dining of capsicin to TRPV1.