Gently does it for crystallography
Very cold, weak electron beams have been used to collect large numbers of diffraction patterns from protein microcrystals...
Tamir - We study proteins that live in the surrounding membrane of a cell. They form gateways for sugars and other molecules like ions and water to go into a cell or out of it, and we arte trying to study how they're built and based on how they're built we're trying to understand how they work and what goes wrong in disease. And one of the major problems that we've had in the lab is that to form crystals from which you can deduce a structure is very difficult for such membrane-embedded proteins. And so, often what we got were crystals that were way too small for structural studies. And what we were trying to figure out is, is there another way of getting to the protein structure out of crystals that are really very small?
Chris - So, how were you trying to image them when you were making these crystals or, all be it, small ones? How were you then trying to get the structure?
Tamir - The most common way of determining a structure from a crystal is by a method called x-ray crystallography. Now, you can think of a crystal as a three-dimensional object that has repeat structural motifs in it and when you shoot an x-ray beam through it, those repeat structural motifs act like a diffraction grating and they split the x-rays and scatter them in a predictable way. And based on the scatter that you get, you can then ask what kind of a structure would give me such a pattern?
Chris - Is the downside that you need a lot of crystal matter there in order to get that reproducible pattern in order to deduce what it's made of?
Tamir - The downside is that the x-rays have a lot of destructive energy in them because it's such a strong beam. So you need large crystals for x-ray crystallography to be able to withstand this large amount of radiation that we're giving it.
Chris - So, how have you tried to get around that?
Tamir - Well, so what we tried to do is rather than using x-rays, is using electrons. And the thinking behind it was that electrons interact with matter much better than x-rays, and so maybe we could get away with using much smaller crystals.
Chris - Why has no one done that before because we've had electron microscopes for donkeys years, people have been imaging many, many things in enormous detail with electron microscopes? So, why had no one tried to do what you did?
Tamir - People have tried to do that before. Many of my colleagues have put crystals in an electron beam before. The issue is when people use electron microscopy for biological specimens, they use a procedure that's called, low-dose cryo-EM. That means that they try to limit the amount of electrons that hit the sample, because if too much electrons hit the sample, then just like with x-rays, the sample will get destroyed. The low-dose procedure limits the dose to about 20 electrons per square angstrom. Now, if you take a crystal and shoot it at that dose rate of 20 electrons per square angstrom, the crystal gets destroyed after a single shot. So, what we said is, well, if a single shot at 20 electrons per square angstrom would destroy the crystal, what would happen if instead of using 20 electrons per square angstrom, can we use 10? Would we still get meaningful data? And the answer was, yes. So then, we got greedy and we said, okay, well can we cut it down more? And we got it down to 0.01 electrons per square angstrom per second. And so, that's about 200 times less in dose than what my colleagues were using and yet we still got meaningful data up to what is called, atomic resolution. Once we knew that we can get so much data out of a single crystal, all we had to do is that while we're exposing it, we start tilting it. And it's that tilt that gives you the different views, if you like, the different patterns from all the different angles and because you have all these different angles in three dimensions, you can calculate back and get a structure.
Chris - And what sort of quality of image can you get out because, obviously, we can get some idea of the crystal's structure with an x-ray, a pattern? But with electrons, is there anything else that you can see beyond just working out what the crystal structure is?
Tamir - Well, in x-ray diffraction, what you end up with is, is an electron density map. In microED, what you end up with is a column potential map, and so that means that you could look at charge. So, if you have a process that is charge-coupled, for example a proton-coupled transporter, you could figure out which amino acid residues in a protein get proteinated or deproteinated. In other words, you can see where the charges are. And that would really begin to help you explain the mechanism by which a protein works, and that you can't do by x-ray crystallography.