 So looking at an actual membrane protein structure, there are a couple of things worth noticing here. The thickness here, as I told you, was roughly 30 angstroms, and that means that if you do your math, you can count, well, including the head groups, we're going to need roughly 20 amino acids and an alpha helix to get through here. We can't have hydrophilic amino acids because if they're hydrophilic, they're going to love to be water and that would be very bad to expose. So whenever we find stretches of roughly 20 hydrophobic amino acids that prefer to be in alpha helical shape in a genome, that's a very strong indication that that's a transmembrane part. And in fact, that has historically been one of the way how we found membrane proteins. We identify them from the genome sequence and just predict that they appear to be a transmembrane helix. But there are these loops. Do you see how the loops are stretching out? They would be water exposed there and there would be water exposed there. So it's not entirely hydrophobic. The parts that are here are hydrophilic. As we started to determine more and more membrane protein structures, initially many of us thought that if globular proteins are hydrophobic on the inside, but hydrophilic water soluble on the outside, if you then have membrane proteins that are hydrophobic on the outside, it would make sense that they're kind of opposite, right? They should be hydrophilic on the inside. That was a very good idea, but it just turned out to be wrong. Because when these insert, they need to insert one helix at a time. I'll get back to that. And it turns out that membrane proteins are really hydrophobic everywhere, most of them at least. There will be some exceptions as you will see later. So membrane proteins are not really stabilized by the hydrophobic collapse that globular proteins would be. They need to be hydrophobic to insert in the membrane, but then it's going to be a matter of packing all these helixes in an efficient way, which is mostly in our Jones interactions, actually, rather than the hydrophobic effect. There will also be some special effects up here. You might have a large domain bound here, and then somehow we need to create this membrane protein and insert it in the membrane, which is far from trivial. I can't dig a hole in the membrane and shove in the entire protein, because these proteins, they will not even be stable outside the membrane. They wouldn't fold until I have inserted them in the membrane. The other parts to think a little bit about what happens with the residue in here. In some cases, we can actually have slightly charged residues. Let me show you a movie that this was made by Anna Johansson, the student in the labs, several years ago. So if we zoom in here, do you see here how the lipids are moving? Because now we have actual lipids that are somewhat chaotic. I think this is probably 10 nanoseconds or so. The waters are moving everywhere. The waters love to form hydrogen bonds with the head groups. And in this particular case, I have a membrane protein that contains a lysine. A lysine is not something we would normally expect to see in a membrane protein, but we can actually put a charged side chain either here because this is effectively up in the loop, or even slightly further down because this side chain will be able to stretch out and interact with the water.