What the Heck I am Doing Over Here
WARNING: This post contains scenes of extended pontification on a scientific subject. It is unsuitable for pretty much everyone.
It occurs to me that I haven't yet bored all of you by explaining exactly what it is I am doing over here. No doubt you have wondered on many occasions: What is this thing called Nuclear Magnetic Resonance that he goes on about for a second or so before start to drift off?
Maybe it would be better first to say what I'm using Nuclear Magnetic Resonance for and why anyone would even bother. I study proteins. Proteins come in all shapes and sizes, but they are all 'just one thing after another' as my very weird advance organic chem prof used to say. They are made up of a chain of small molecules called amino acids. There are 2o 'standard' amino acids. Each one has the same basic structure except for one of their 'arms' which varies. This arm can be long, short, acidic, basic, hydrophobic (water-hating) or hydrophilic (water-loving) and, when you make a chain of them, things start to get interesting. First, the non-variable part of the amino acid structure causes the chain either to want to twist into a sortof spiral (called an alpha-helix) or fold back on itself (called a beta strand, two or more strands side by side make a beta sheet). Whether the chain forms a spiral or a strand depends on the variable part of the amino acids, some are 'spiral friendly' and some are 'strand friendly'. The variable part of the amino acids also causes the spirals and strands to fold around each other in a specific way. Parts of the chain where there is a prevalence of amino acids with hydrophobic arms, for example, will try to minimize their exposure to water by clustering together in the middle of the protein.
And once you know how a protein is shaped, you know pretty much how it works. You can imagine that, depending on the sequence of amino acids, they can form pretty much any shape. Some will be long and and stiff like the stuff in your fingernails and hair, some are round-ish with hydrophobic pockets to ferry hydrophobic molecules around in your blood. So if we're going to mess with them, i.e. interfere with proteins we don't like or get proteins we do like to work better, we are going to have to know their shape. Lots of drugs work by messing with proteins. I'm on one right now. Omeprazole (Losec) is a molecule that fits into the protein that make acid in your stomach by pretending to be the molecule that it normally works on. The protein starts to work on Losec, but gets permanently 'jammed' in the middle of the job, taking it out of commission. So knowing the shape of a protein can help us find drugs like losec. But that's tricky, really, because for being so big, proteins are actually very, very small.
Enter Nuclear Magnetic Resonance (NMR). It wasn't an obvious choice at first. NMR is actually kindof like a radio. What you do is you take a sample and put it inside a coil of wire. This coil, it is important to note, is sitting in a very strong, permanent magnet. What does this do? Well, nothing, chemically, but it makes the nuclei in your sample (say hydrogen, H, or nitrogen, N) have a slight preference about the direction in which they are spinning. They are spinning, incidentally, at different speeds. Different nuclei (like H and N) spin at very different speeds, while nuclei that are the same spin at slightly different speeds because they can be shielded from the magnetic field to various degrees by the cloud of electrons that surrounds them. So basically every proton in a protein (and there are a lot of them) has a distinctive spin. What good does this do? Well, none, actually because we have no way to measure this spin. What we need to do is change the preferred alignment of the spin so that is 90 degrees to big huge permanent magnetic field. But how do we do that? I mean we're competing with a big huge permanent magnet here. Well, it turns out you can cancel out a really huge magnetic field with a really dinky one if you ensure that the field oscillates at the same frequency as your nuclei are spinning. That's resonance, baby, and it works.
Now we could apply our '90 degree' field and then vary the oscillation frequency so that we 'excite' (and thus detect) our spinning nuclei one at a time. But that would be boring. Instead we'll hit the nuclei with a very quick, high power field and excite them all at once. That way, we can measure them all at once. To visualize what the signal that we get, imagine a lighthouse light spinning around. As the light spins toward you it gets brighter and brighter, then fully bright, then dimmer then dimmer then fully dim, then brighter, then brighter etc. So we get a wave that goes up and down in intensity. Now imagine there are hundreds of lights in the tower and they are all spinning at different speeds. That is a very complicated set of waves and it doesn't really tell us anything to look at it. Fortunately, there is a neat trick in math called a Fourier Transform into which we can plop in a wave of any complexity and come out with a 'spectrum' of all the spin speeds that made the wave. I know, I know... way cool.
So now we can see all these spinning protons. So what? What does this tell us about the shape of a protein? Again, nothing. To get information about the shape, we have to take things two steps further. First, we have to limit the number of hydrogens we are looking at cause if we tried to look at all of them, we won't be able to tell one from the other (because their spin speeds are too similar). We do this by 'isotopically labeling' the Nitrogens in the protein. Every amino acid has one nitrogen in it's non variable part, and that nitrogen is attached to one hydrogen. So using some NMR magic, we can limit the spins we excite only to those hydrogens that are attached to our labeled nitrogens. Then we get really clever and do all kinds of NMR magic to measure the effect that one excited spin has on another. This is perfect because, if the spins are close by, they will affect each other really strongly, while if they are far apart, they'll hardly affect each other at all. So, by measuring the strength of this effect, we can get a series of inter-hydrogen distances. From these, using a computer, we can figure calculate all of the possible chain shapes that satisfy all of the hydrogen distances. We then take the average of these shapes (weighted average, actually, because some chain shapes are better than others) and do some more calculations to figure out in which direction the variable 'arms' are pointed. And voila! We have a protein structure!
It occurs to me that I haven't yet bored all of you by explaining exactly what it is I am doing over here. No doubt you have wondered on many occasions: What is this thing called Nuclear Magnetic Resonance that he goes on about for a second or so before start to drift off?
Maybe it would be better first to say what I'm using Nuclear Magnetic Resonance for and why anyone would even bother. I study proteins. Proteins come in all shapes and sizes, but they are all 'just one thing after another' as my very weird advance organic chem prof used to say. They are made up of a chain of small molecules called amino acids. There are 2o 'standard' amino acids. Each one has the same basic structure except for one of their 'arms' which varies. This arm can be long, short, acidic, basic, hydrophobic (water-hating) or hydrophilic (water-loving) and, when you make a chain of them, things start to get interesting. First, the non-variable part of the amino acid structure causes the chain either to want to twist into a sortof spiral (called an alpha-helix) or fold back on itself (called a beta strand, two or more strands side by side make a beta sheet). Whether the chain forms a spiral or a strand depends on the variable part of the amino acids, some are 'spiral friendly' and some are 'strand friendly'. The variable part of the amino acids also causes the spirals and strands to fold around each other in a specific way. Parts of the chain where there is a prevalence of amino acids with hydrophobic arms, for example, will try to minimize their exposure to water by clustering together in the middle of the protein.
And once you know how a protein is shaped, you know pretty much how it works. You can imagine that, depending on the sequence of amino acids, they can form pretty much any shape. Some will be long and and stiff like the stuff in your fingernails and hair, some are round-ish with hydrophobic pockets to ferry hydrophobic molecules around in your blood. So if we're going to mess with them, i.e. interfere with proteins we don't like or get proteins we do like to work better, we are going to have to know their shape. Lots of drugs work by messing with proteins. I'm on one right now. Omeprazole (Losec) is a molecule that fits into the protein that make acid in your stomach by pretending to be the molecule that it normally works on. The protein starts to work on Losec, but gets permanently 'jammed' in the middle of the job, taking it out of commission. So knowing the shape of a protein can help us find drugs like losec. But that's tricky, really, because for being so big, proteins are actually very, very small.
Enter Nuclear Magnetic Resonance (NMR). It wasn't an obvious choice at first. NMR is actually kindof like a radio. What you do is you take a sample and put it inside a coil of wire. This coil, it is important to note, is sitting in a very strong, permanent magnet. What does this do? Well, nothing, chemically, but it makes the nuclei in your sample (say hydrogen, H, or nitrogen, N) have a slight preference about the direction in which they are spinning. They are spinning, incidentally, at different speeds. Different nuclei (like H and N) spin at very different speeds, while nuclei that are the same spin at slightly different speeds because they can be shielded from the magnetic field to various degrees by the cloud of electrons that surrounds them. So basically every proton in a protein (and there are a lot of them) has a distinctive spin. What good does this do? Well, none, actually because we have no way to measure this spin. What we need to do is change the preferred alignment of the spin so that is 90 degrees to big huge permanent magnetic field. But how do we do that? I mean we're competing with a big huge permanent magnet here. Well, it turns out you can cancel out a really huge magnetic field with a really dinky one if you ensure that the field oscillates at the same frequency as your nuclei are spinning. That's resonance, baby, and it works.
Now we could apply our '90 degree' field and then vary the oscillation frequency so that we 'excite' (and thus detect) our spinning nuclei one at a time. But that would be boring. Instead we'll hit the nuclei with a very quick, high power field and excite them all at once. That way, we can measure them all at once. To visualize what the signal that we get, imagine a lighthouse light spinning around. As the light spins toward you it gets brighter and brighter, then fully bright, then dimmer then dimmer then fully dim, then brighter, then brighter etc. So we get a wave that goes up and down in intensity. Now imagine there are hundreds of lights in the tower and they are all spinning at different speeds. That is a very complicated set of waves and it doesn't really tell us anything to look at it. Fortunately, there is a neat trick in math called a Fourier Transform into which we can plop in a wave of any complexity and come out with a 'spectrum' of all the spin speeds that made the wave. I know, I know... way cool.
So now we can see all these spinning protons. So what? What does this tell us about the shape of a protein? Again, nothing. To get information about the shape, we have to take things two steps further. First, we have to limit the number of hydrogens we are looking at cause if we tried to look at all of them, we won't be able to tell one from the other (because their spin speeds are too similar). We do this by 'isotopically labeling' the Nitrogens in the protein. Every amino acid has one nitrogen in it's non variable part, and that nitrogen is attached to one hydrogen. So using some NMR magic, we can limit the spins we excite only to those hydrogens that are attached to our labeled nitrogens. Then we get really clever and do all kinds of NMR magic to measure the effect that one excited spin has on another. This is perfect because, if the spins are close by, they will affect each other really strongly, while if they are far apart, they'll hardly affect each other at all. So, by measuring the strength of this effect, we can get a series of inter-hydrogen distances. From these, using a computer, we can figure calculate all of the possible chain shapes that satisfy all of the hydrogen distances. We then take the average of these shapes (weighted average, actually, because some chain shapes are better than others) and do some more calculations to figure out in which direction the variable 'arms' are pointed. And voila! We have a protein structure!
posted by Derek at 1:13 PM
5 Comments:
To quote Neo: "Whoa!"
Your best blog entry yet. You're a born teacher. Even I understand it (I think).
Now let's have a similar one from Jane. It would be great to understand her work, too.
Hunh. I thought my explanation made clear the connection between how proteins are shaped and what they do. I go on at length about how knowing the shape of a protein lets you 'mess' with it, either stopping them from working, or making them work better. Lots of drugs work by stopping a protein. In particular the 'Losec example' was to illustrate just what you can do when you know how a protein is shaped. Structural studies by NMR take the hell of a long time, though (usually about 1.5 yrs if you know what you're doing). I'm just on the preliminaries, like assigning each peak to a particular proton.
hmm Nice description. We have a guy in the physics department here who does NMR on spider silk proteins to try and understand how the silk forms. The stuff is stronger than Kevlar but can be grown and is biodegradable.
One thing i think is a bit confusing is you seem to use spin to mean two different things, the intrinsic spin of particles and precession of the nuclei which is not quite the same. All particles have spin but the spin of all Nitrogen Nuclei is exactly the same, it's 1. It's the precession that is slightly different between the particles and is what you measure, not the spin. Really it's just semantics but it's a very important distinction.
Anyway, it was a very good introduction to NMR.
Shawn, you're absolutely right. NMR people just say 'spin' because what they are really measuring is a spinning magnetic field that results from a preferred alignment of for precession (after the application of our dinky field, this preferred alignment is on the transverse plane). When physics people say spin, they mean the quantum spin, which, as you say,is 1 for Nitrogen. Actually NMR works much better on spin 1/2 nuclei because the are dipolar (some profs still say that you can ONLY see spin 1/2 nuclei, but that's not true - we use deuterium all the time as a calibrant). This is actually the source of the 'NMR magic' that I mention in the post: When we work with proteins, we label them with N15 (spin 1/2) so that we only see the protons attached to them.
Post a Comment
<< Home