Kurt Wüthrich (2015) - NMR in Biology, Chemistry and Medicine

Well, thank you. NMR is an ideal theme for an interdisciplinary meeting considering that there have been explicitly Nobel Prizes for NMR in physics, chemistry, and medicine, and several dozen Nobel prizes in chemistry and physics depended heavily on applications of NMR. Let me start. Well, what do we expect today from NMR. We expect to get structures like this. Here you see a drug, you see a receptor of a drug. In chemistry, you want to characterize, this is a simple inorganic compound, you want to do measurements on dynamic processes. This compound. And then you go to the hospital and you get images like this one, which is of my left knee. You see, if you want to play football for an extended period of time, you have to have some ways to look at your joints from time to time. Ok, so that's what we expect today of NMR: Getting structures in chemistry, in biology, learning a lot about rate processes, conformational equilibria, and when we have some pains in our joints or in our head, then we want to get an MRI image. Now let me start with a piece of physics. In 1896, a Dutch physicist who was then a graduate student was looking at optical spectra of material. And then he observed that when he placed the samples into a magnetic field, there were so far unknown splittings in the optical spectra. This is now referred to as Zeeman Effect. There seems to be an interesting story about it because the head of the institute had strictly forbidden the students to place their samples into magnetic fields. Therefore the student lost his job. And when he got the Nobel prize six years later, the head of the department was dismissed and he became head of the department. That's what I read. (laughs) The splitting between these eigenstates of nuclear spins one have is very small. And it took half a century until transitions between the two states could actually be observed. And that was due primarily to the fact that the US Army had a lot of engineers and physicists working on the development of radar during the Second World War. And when these engineers and physicists came back from Army service, immediately there were very rapid advances in paramagnetic resonance spectroscopy based on klystron technology, and in radio frequency technology. And so in 1946, within about one month, two groups reported observations of the induced transitions between these very short energy differences. So one of these was a Swiss citizen, Felix Bloch, who worked at Stanford, the other one was Edward Purcell who worked at Harvard. So that was the Physics Prize in 1952. And I have known Felix Bloch quite well because he spent his last years in Zürich. He visited with us, and they never thought that NMR would be good for anything but more accurate measurement of gyromagnetic ratios and the likes. Then came a discovery made by others that the frequency of the NMR lines depended on the chemistry. So if you, this is a simple compound, a lumichrome, actually I am looking at the nitrogens in here and because the nitrogens in this compound are chemically different, you see distinct lines. This is called the chemical shift, and that enabled chemists to analyse the products of their synthesis to investigate natural products and so on and get to structures of small molecules. Then, this happened in the late 1950s and the early 1960s, but NMR had very low sensitivity, and because of the limitations of the magnetic field, that would be applied, also relatively low resolution. And only towards the end of the 1960s it became possible to use NMR also for large molecules. Now the background here is not meant to be here, but what you see here is a painting of one of the first protein molecules that had been studied by crystallography. You see in the 1960s, there were no computers that would be able to draw images of molecules, so this is an artist's painting based on the atomic coordinates. There is a museum in New York with about 300 paintings by Irving Geis, who did all these early illustrations. Now because we have a large number of chemically different spins one have in such a molecule, we get a large number of lines. You see in this particular protein, I don't want to go into details, but in this particular protein a few lines are very well separated along the frequency axis, but here there are hundreds of lines which are strongly overlapped. On the one hand, we use these lines to learn what we needed about spin physics in large molecules in solution, and we despaired about the fact that we couldn't work with this part of the spectrum because we had learned from here that we could get structures if we did things right. So we had to solve a problem and jointly with Prof. Ernst, who unfortunately is not here, at this meeting, we developed multidimensional NMR, I mean initially COSY, NOESY, SECSY, and FOCSY in 1982. The point is that, from the physics point of view, we introduce additional artificial time domains in addition to the running time. We would use tricks to introduce a second time axis, a third, fourth, today usually five to seven time axes. And even at two dimensions, we now have a frequency plane, and you see out of the overlapping spectrum, we get well-separated lines. So problem one, to deal with large molecules, had been solved. And Richard got the Nobel Prize in Chemistry in 1991. Now there was a second problem to be solved when trying to work with large molecules. And you may, you probably all attended the lecture by Professor Ramakrishnan on day one. And he said, I noted it down, That means, in crystallography, you essentially take a photograph of the resting molecule. It's a bit more complicated, but that's what it is. And in order to keep the molecules in place, actually crystallography of proteins is now done at liquid nitrogen temperatures, so that everything is frozen stiff. But of course proteins are in body fluids in solution, they move, they aren't frozen down. So we have Brownian Motion. I come back to that a bit later. What happens is that these molecules, let's assume you have a small, or a large protein, they undergo translational diffusion, and they also undergo rotational diffusion. So when we try to talk about that... It actually started this way, that we were able to find errors in crystal structures of proteins, that was a lot of fun. And so, but then the crystallographers, our dear colleagues, said, These motions happen on the nanosecond timescale and this will always give funny pictures." So you have to think a bit, and then you come to the solution. All you need are parameters that are invariant on the rotation and translation. And if you find the sufficient number of parameters that are invariant on the rotation and translation, then you only need to measure these and develop mathematical tools that can then visualize the structures from these rotation and translation invariant parameters. And one of these parameters is the distance between two atoms. If you have a given structure, then the distance between each pair of atoms is invariant on the rotation and translation. Ok, that's a key. So we measure NOE, so called nuclear Overhauser effect, this is a two dimensional spectrum, it's actually very old, this is from 1983, and each peak in here tells me that there is a pair of hydrogen atoms in a three dimensional protein that is at a distance of less than five angstrom. That was 2D NMR, then we had of course to assign each of these peaks to two particular points. Let me do this again. I usually demonstrate what happens with my belt. Ok? If I have thousands of these peaks and I don't know where those short distances are, then I will never find out what the structure is. But if I know that one of these peaks tells me that there is a short distance between the two ends then I have a ring. If I have another assigned short distance, then you have a three dimensional structure. If you have a helix, then you find that there is a short distance between every fourth residue along the helix. Ok? So that's how it works, except when you have that many peaks, then you need a computer, not a belt, to do the game. And then you get the three dimensional structure of the protein. Now let's move on to medicine. This is not a complicated molecule, this is water. And in water you have two spin one half nuclides and two protons and you cannot distinguish between the two. So you get only a single line. This seems sufficiently boring not to spend too much time with it. I actually spent several years of my early scientific life looking at only this line. There's an awful lot of information in this line. But we didn't have the key idea on how to make use of this simple spectrum. You see, this is very different from working with big molecules, which have lots of lines. Here you have only one line. Now comes the idea. Since the frequency of the transitions, of the NMR transitions, depends on the strength of the magnetic field, you can apply a gradient across microscopic objects. Now the frequency, and if you only see the water, always see the water, now all of sudden you are able to distinguish between water next to the right ear and water next to the left ear because the frequency is proportional to the magnetic field which varies across a macroscopic object. That means outside of the head we have this splitting, here we have a bit higher frequency, and here we have still a bit higher frequency. So we can, sorry, we get three lines: this is the water next to the left ear, this is the water next to the right ear. Now you do the same with field gradients from the back to front and from top down and then you have a very similar situation as we have with the distance constraints in macromolecules. We just have to develop imagery construction techniques and then we get an image of the head. And this is Paul Lauterbur and Peter Mansfield here in front of an MRI, who got the Physiology or Medicine Prize in 2003. Ok, so that's how NMR developed across the disciplines and became a useful and also economically important part of the game. I mean, there are ten thousands of high tech jobs in hospitals using MRI, and the business, as you know, there is an MRI at every corner in the Western world, and this is also economically an important business. Now let me go back to Brownian Motion and try to understand why all this works. And here we go much further back in history. Brownian Motion is named after a botanist in England, who in 1828 had some pollen. He was a botanist, so he worked with pollen, the pollen feel into a liquid, and then he took the microscope and under the microscope he saw that the pollen were moving in the liquid. And all he saw were such movements. He saw that small pollen would change more frequently direction of the transverse diffusion, and larger pollen would change direction more slowly. Now he published this, but he didn't understand what was going on. It took almost 80 years until that guy was very, very bored in Berne. He was in the Patent Office, he had actually a standing desk where he had to write down I don't know what, but he must have been very bored so he was thinking about Brownian motion. And one of his important papers in 1905, in May 1905, treated the translational diffusion of particles suspended in liquids. That was not all. I think that while he was working on the translation diffusion it must have occurred to him at some point that in addition to what Mr. Brown saw in 1826, there must also be rotational diffusion, because it's clear, it's a random process. You have thermally agitated solvent molecules that hit the suspended particles and that causes this random translational diffusion and therefore there must also be rotational diffusion. So that must have dawned in him during the summer of 1905 and in December 1905 he published a second paper. These papers are 40 pages long, very detailed descriptions in German, so it's easy for us to read, and we get the Stokes-Einstein Relation from this. Now the Stokes-Einstein Relation essentially says that there is a correlation time which depends on the radius of the suspended particle. This is all for spheres - I mean, if it's not spherical, then the mathematics get a bit more difficult. So for a sphere, the correlation time is proportional to the radius, third power, is proportional to the viscosity and inverse proportional to the temperature. And this correlation time, essentially in very qualitative terms, says something about the time that goes by after which the system has lost memory of what has happened to him before. Ok? And this correlation time determines all the key NMR parameters, in particular the nuclear Overhauser effect, which we used for the distance measurements, and the relaxation times. Those of you who have been in an MRI, you may remember that you got T1 weighted images of your head or your knees, or T2 weighted images. So this is now all, it's used all over the place. And all these parameters depend on this correlation time. Now we immediately know why, well let's not go too complicated yet. Now we all of a sudden understand why MRI works. We see the resonance of the water, but we don't have a background from the bones or from the muscles because there, the correlation time is so long that the signals are decayed before we ever have a chance to record them. And that's why we can make sharp images of the human body looking at the water, then clearly what we see is the soft tissue, rather than the bones, and now we understand thanks to Einstein. Ok? Now there is a little more to it. When you look at the time evolution of single transition basis operators, you get apparently difficult equations that describe the time evolution of the coherence. But then there are, do you see there are the relaxation terms? Now these are those terms that depend on the correlation time. And you go closer, then you see that there is a term, a relaxation term which has a sum of two quantities and another one which has a difference between two quantities. Now when this theory was developed in 1963, it was first described by Japanese physicist Shimizu, the magnetic fields were at about 60 megahertz for observatio ofl protons. And I think I have another picture here. Do I? No, I don't. Alright, I didn't want... What happens is that this difference, when this theory was developed, this difference was almost the same as the sum. Ok? In the meantime, the magnetic field strength has increased 15 fold, and the chemical shift anisotropy, the second term increases roughly proportional to the field strengths. Now in 1997, these two, we had exactly the field strength where these two were equal, and so this term became zero, and all we had to do is pick out the one transition where this term is zero and then we uncouple the NRM spectrum from the Brownian motion. See, the theory was developed almost 50 years before the technology was available to exploit it properly. The result was a method that was call TROSY, or transverse relaxation optimized spectroscopy, and there was a strict limit to the size of molecules that we could observe until about 1997. We all of a sudden now went to recording spectra of particles up to approximately 1,000,000 in molecular weight because now the NMR spectra were uncoupled from the Brownian motion. And I think that gets me to the end of my talk. I tried to illustrate to you that NMR really goes across the three disciplines that are represented here in this meeting. It has been a highly interdisciplinary venture to develop all this technology. You should consider that there were no program libraries around where we could get, packages of programs, of routines, as it is today. Everything had to be written by my students, in the case of macromolecular NMR, in the case of MRI, the software had to be developed by students. So it was, I mean, in my case, theoretical physicists, experimental physicists, we would build our own programmes and so on. And so it was a lot of fun and a really interdisciplinary effort that eventually went from biology all the way to computation science and theoretical physics. Thank you.

Kurt Wüthrich (2015)

NMR in Biology, Chemistry and Medicine

Kurt Wüthrich (2015)

NMR in Biology, Chemistry and Medicine

Abstract

For the discovery of the physics phenomenon of nuclear magnetic resonance (NMR), Felix Bloch and Edward Purcell were awarded the Nobel Prize in Physics in 1952. NMR has then been used in a wide range of fundamental studies in physics, and in the 1960s it also became an important analytical tool in chemistry. Based on novel concepts and advances in NMR instrumentation and informatics tools, exciting developments in the early 1970s laid the foundations for magnetic resonance imaging (MRI), which is today a key technique in medical diagnosis, and for the use of NMR spectroscopy in modern structural biology. These achievements have been recognized by Nobel prizes to four scientists, and a fundamental understanding of their achievements is greatly helped by work of Albert Einstein in 1905. NMR thus is one of many research areas where the results of basic research during the first half of the 20th century provided the basis for technological breakthroughs in the second half of the 20th and into 21st century, which now support all branches of chemistry, structural biology, drug discovery and clinical diagnosis.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content