Skip Over Navigation Links

The Structures of Life

Chapter 3: The World of NMR: Magnets, Radio Waves, and Detective Work

Did you ever play with magnets as a kid? That's a large part of what scientists do when they use a technique called nuclear magnetic resonance (NMR) spectroscopy.

An NMR machine is essentially a huge magnet. Many atoms are essentially little magnets. When placed inside an NMR machine, all the little magnets orient themselves to line up with the big magnet.

By harnessing this law of physics, NMR spectroscopists are able to figure out physical, chemical, electronic, and structural information about molecules.

Next to X-ray diffraction, NMR is the most common technique used to determine detailed molecular structures. This technique, which has nothing to do with nuclear reactors or nuclear bombs, is based on the same principle as the magnetic resonance imaging (MRI) machines that allow doctors to see tissues and organs such as the brain, heart, and kidneys.

One of the largest structures determined by NMR is malate synthase G, with a mass of 82 kilodaltons.
One of the largest structures determined by NMR is malate synthase G, with a mass of 82 kilodaltons.
The largest structure determined by X-ray crystallography is the ribosome. The Protein Data Bank includes many structures of ribosomes, the largest more than 2,000 kilodaltons.
The largest structure determined by X-ray crystallography is the ribosome.The Protein Data Bank includes many structures of ribosomes, the largest more than 2,000 kilodaltons.

Currently, NMR spectroscopy is only able to determine the structures of small and medium-sized proteins. Shown here to scale is one of the largest structures determined by NMR spectroscopy compared to the largest structure determined by X-ray crystallography (the ribosome). Images courtesy of Catherine Lawson, Rutgers University and the RCSB Protein Data Bank

Although NMR is used for a variety of medical and scientific purposes—including determining the structure of genetic material (DNA and RNA), carbohydrates, and other molecules —in this booklet we will focus on using NMR to determine the structure of proteins.

Methods for determining structures by NMR spectroscopy are much younger than those that use X-ray crystallography. As such, they are constantly being refined and improved.

The most obvious area in which NMR lags behind X-ray crystallography is the size of the structures it can handle. Most NMR spectroscopists focus on molecules no larger than 60 kilodaltons (about 550 amino acids). X-ray crystallographers have solved structures up to 2,500 kilodaltons—40 times as large.

But NMR also has advantages over crystallography. For one, it uses molecules in solution, so it is not limited to those that crystallize well. (Remember that crystallization is a very uncertain and time-consuming step in X-ray crystallography.)

NMR also makes it fairly easy to study properties of a molecule besides its structure—such as the flexibility of the molecule and how it interacts with other molecules. With crystallography, it is often either impossible to study these aspects or it requires an entirely new crystal. Using NMR and crystallography together gives researchers a more complete picture of a molecule and its functioning than either tool alone.

NMR relies on the interaction between an applied magnetic field and the natural "little magnets" in certain atomic nuclei. For protein structure determination, spectroscopists concentrate on the atoms that are most common in proteins, namely hydrogen, carbon, and nitrogen.

Before the researchers begin to determine a protein's structure, they already know its amino acid sequence—the names and order of all of its amino acid building blocks.What they seek to learn through NMR is how this chain of amino acids wraps and folds around itself to create the three-dimensional, active protein.

Solving a protein structure using NMR is like a good piece of detective work. The researchers conduct a series of experiments, each of which provides partial clues about the nature of the atoms in the sample molecule—such as how close two atoms are to each other, whether these atoms are physically bonded to each other, or where the atoms lie within the same amino acid. Other experiments show links between adjacent amino acids or reveal flexible regions in the protein.

The challenge of NMR is to employ several sets of such experiments to tease out properties unique to each atom in the sample. Using computer programs, NMR spectroscopists can get a rough idea of the protein's overall shape and can see possible arrangements of atoms in its different parts. Each new set of experiments further refines these possible structures. Finally, the scientists carefully select 10 to 20 solutions that best represent their experimental data and present the average of these solutions as their final structure.

Back to Top

NMR Spectroscopists Use Tailor-Made Proteins

 Only certain forms, or isotopes, of each chemical element have the correct magnetic properties to be useful for NMR. Perhaps the most familiar isotope is 14C, which is used for archeological and geological dating.

You may also have heard about isotopes in the context of radioactivity. Neither of the isotopes most commonly used in NMR, namely 13C and 15N, is radioactive.

Like many other biological scientists, NMR spectroscopists (and X-ray crystallographers) use harmless laboratory bacteria to produce proteins for their studies. They insert into these bacteria the gene that codes for the protein under study. This forces the bacteria, which grow and multiply in swirling flasks, to produce large amounts of tailor-made proteins.

To generate proteins that are "labeled" with the correct isotopes, NMR spectroscopists put their bacteria on a special diet. If the researchers want proteins labeled with 13C, for example, the bacteria are fed food containing 13C. That way, the isotope is incorporated into all the proteins produced by the bacteria.

Back to Top

NMR Magic Is in the Magnets

A Slam Dunk for Enzymes

Dorothee Kern

NMR spectroscopy is ideal for studying how enzymes change shape as they do their jobs. Take it from Dorothee Kern, a former professional basketball player who is now an NMR researcher at Brandeis University. Read about her work at

Most NMR spectroscopists use magnets that are 500 megahertz to 900 megahertz. This magnet is 900 megahertz. Varian NMR Systems
Most NMR spectroscopists use magnets that are 500 megahertz to 900 megahertz. This magnet is 900 megahertz. Varian NMR Systems

The magnets used for NMR are incredibly strong. Those used for high resolution protein structure determination range from 500 megahertz to 900 megahertz and generate magnetic fields thousands of times stronger than the Earth's.

Although the sample is exposed to a strong magnetic field, very little magnetic force gets out of the machine. If you stand next to a very powerful NMR magnet, the most you may feel is a slight tug on hair clips or zippers. But don't get too close if you are wearing an expensive watch or carrying a wallet or purse—NMR magnets are notorious for stopping analog watches and erasing the magnetic strips on credit cards.

NMR magnets are superconductors, so they must be cooled with liquid helium, which is kept at 4 Kelvin (-452 degrees Fahrenheit). Liquid nitrogen, which is kept at 77 Kelvin (-321 degrees Fahrenheit), helps keep the liquid helium cold.

Back to Top

The Many Dimensions of NMR

This one-dimensional NMR spectrum shows the
chemical shifts of hydrogen atoms in a protein from streptococcal bacteria. Spectrum courtesy of Ramon Campos-Olivas, National Institutes of Health
This one-dimensional NMR spectrum shows the chemical shifts of hydrogen atoms in a protein from streptococcal bacteria. Spectrum courtesy of Ramon Campos-Olivas, National Institutes of Health

To begin a series of NMR experiments, researchers insert a slender glass tube containing about a half a milliliter of their sample into a powerful, specially designed magnet. The natural magnets in the sample's atoms line up with the NMR magnet just as iron filings line up with a toy magnet.

The researchers then blast the sample with a series of split-second radio wave pulses that disrupt this magnetic equilibrium in the nuclei of selected atoms.

By observing how these nuclei react to the radio waves, researchers can assess their chemical nature. Specifically, researchers measure a property of the atoms called chemical shift.

Every type of NMR-active atom in the protein has a characteristic chemical shift. Over the years, NMR spectroscopists have discovered characteristic chemical shift values for different atoms (for example, the carbon in the center of an amino acid, or its neighboring nitrogen), but the exact values are unique in each protein. Chemical shift values depend on the local chemical environment of the atomic nucleus, such as the number and type of chemical bonds between neighboring atoms.

The pattern of these chemical shifts is displayed as a series of peaks in what is called a one-dimensional NMR spectrum. Each peak corresponds to one or more hydrogen atoms in the molecule. The higher the peak, the more hydrogen atoms it represents. The position of the peaks on the horizontal axis indicates their chemical identity.

A two-dimensional NMR spectrum of a protein with labeled spots. The laboratory of Xiaolian Gao, University of Houston
A two-dimensional NMR spectrum of a protein with labeled spots. The laboratory of Xiaolian Gao, University of Houston.
Click for larger image.

The overlapping peaks typical of onedimensional NMR spectra obscure information needed to determine protein structures. To overcome this problem, scientists turn to a technique called multi-dimensional NMR. This technique combines several sets of experiments and spreads out the data into discrete spots. The location of each spot indicates unique properties of one atom in the sample. The researchers must then label each spot with the identity of the atom to which it corresponds.

For a small, simple protein, computational programs require only a few days to accurately assign each spot to a particular atom. For a large, complex protein, it could take months.

To better understand multi-dimensional NMR, we can think of an encyclopedia. If all the words in the encyclopedia were condensed into one dimension, the result would be a single, illegible line of text blackened by countless overlapping letters. Expand this line to two dimensions—a page—and you still have a jumbled mess of superimposed words. Only by expanding into multiple volumes is it possible to read all the information in the encyclopedia. In the same way, more complex NMR studies require experiments in three or four dimensions to clearly solve the problem.

Back to Top


NMR Tunes in on Radio Waves

Illustration of a NMR Radio TowerEach NMR experiment is composed of hundreds of radio wave pulses, each separated by no more than a few milliseconds. Scientists enter the experiment they'd like to run into a computer, which then sends precisely timed pulses to the sample and collects the resulting data.

This data collection process can require as little as 20 minutes for a single, simple experiment. For a complex molecule, it could take weeks or months.

NMR's radio wave pulses are quite tame compared to the high-energy X-rays used in crystallography. In fact, if an NMR sample is prepared well, it should be able to last for many years, allowing the researchers to conduct further studies on the same sample at a later time.

Back to Top

Spectroscopists Get NOESY for Structures

To determine the arrangement of the atoms in the molecule, scientists use a multi-dimensional NMR technique called NOESY (pronounced "nosy") for Nuclear Overhauser Effect Spectroscopy. This technique works best on hydrogen atoms, which have the strongest NMR signal and are the most abundant atoms in biological systems. They are also the simplest—each hydrogen nucleus contains just a single proton.

The NOESY experiment reveals how close different protons are to each other in space. A pair of protons very close together (typically within 3 angstroms) will give a very strong NOESY signal. More separated pairs of protons will give weaker signals, out to the limit of detection for the technique, which is about 6 angstroms.

From there, the scientists (or, to begin with, their computers) must determine how the atoms are arranged in space. It's like solving a complex, three-dimensional puzzle with thousands of pieces.

Back to Top

The Wiggling World of Proteins

Examining ribosomal structures in detail will help researchers better understand the fundamental process of protein production. It may also aid efforts to design new antibiotic drugs or optimize existing ones. Courtesy of Catherine Lawson, Rutgers University and the RCSB Protein Data Bank

Although a detailed, three-dimensional structure of a protein is extremely valuable to show scientists what the molecule looks like, it is really only a static "snapshot" of the protein frozen in one position. Proteins themselves are not rigid or static—they are dynamic, rapidly changing molecules that can move, bend, expand, and contract. NMR researchers can explore some of these internal molecular motions by altering the solvent used to dissolve the protein.

A three-dimensional NMR structure often merely provides the framework for more in-depth studies. After you have the structure, you can easily probe features that reveal the molecule's role and behavior in the body, including its flexibility, its interactions with other molecules, and how it reacts to changes in temperature, acidity, and other conditions.

Back to Top

Untangling Protein Folding

A hundred billion years. That's the time scientists estimate it could take for a small protein to fold randomly into its active shape. But somehow, Nature does it in a tenth of a second.

Most proteins start out like a loose string flopping around in a lake, possibly with short coiled sections. The molecules contort quickly into various partially folded states before congealing into their final form. Because the process is so fast, scientists cannot study it directly. But NMR is well suited to certain studies of protein folding.

By changing the temperature, acidity, or chemical composition of a protein's liquid environment, spectroscopists can reverse and interrupt protein folding. By capturing a protein in different stages of unraveling, researchers hope to understand how proteins fold normally.

H. Jane Dyson and Peter Wright, a husband-and-wife team of NMR spectroscopists at the Scripps Research Institute in La Jolla, California, used this technique to study myoglobin in various folding states.

Myoglobin, a small protein that stores oxygen in muscle tissue, is ideal for studying the structure and dynamics of folding. It quickly folds into a compact, alpha-helical structure. Dyson and Wright used changes in acidity to reveal which regions are most flexible in different folding states. The first two "structures" below each represent one of many possible conformations of a floppy, partially folded molecule.

Adapted with permission from Nature Structural Biology 1998, 5:499-503
Adapted with permission from Nature Structural Biology 1998, 5:499-503

Understanding how proteins fold so quickly and correctly (most of the time) will shed light on the dozens of diseases that are known or suspected to result from misfolded proteins. In addition, one of the greatest challenges for the biotechnology industry is to coax bacteria into making vast quantities of properly folded human proteins.

Back to Top

Student Snapshot: The Sweetest Puzzle

Chele DeRider. Credit: Jeff Miller, University of Wisconsin-Madison
Courtesy of Jeff Miller, University of Wisconsin-Madison

“Getting a protein structure using NMR is a lot of fun,” says Chele DeRider, a graduate student at the University of Wisconsin-Madison. “You're given all these pieces to a puzzle and you have to use a set of rules, common sense, and intuitive thinking to put the pieces together. And when you do, you have a protein structure.”

DeRider is working at UWMadison's national NMR facility. She is refining the structure of brazzein, a small, sweet protein. Most sweet-tasting molecules are sugars, not proteins; so brazzein is quite unusual. It also has other remarkable properties that make it attractive as a sugar substitute. It is 2,000 times sweeter than table sugar—with many fewer calories. And, unlike aspartame (NutraSweet®), it stays sweet even after 2 hours at nearly boiling temperatures.

In addition to its potential impact in the multimillion-dollar market of sugar substitutes, brazzein may teach scientists how we perceive some substances as sweet. Researchers know which amino acids in brazzein are responsible for its taste—changing a single one can either enhance or eliminate this flavor—but they are still investigating how these amino acids react with tongue cells to trigger a sensation of sweetness.

The plum-sized berries of this African plant contain brazzein, a small, sweet protein. H.M. Hadik
The plum-sized berries of this African plant contain brazzein, a small, sweet protein. Courtesy of H.M. Hadik
“Getting a protein structure using NMR is a lot of fun…You start out with just dots on a page and you end up with a protein structure.”
Chele DeRider
Graduate Student
University of Wisconsin-Madison

DeRider became interested in NMR as an undergraduate student at Macalester College in St. Paul, Minnesota. She was studying organic chemistry, but found that she spent most of her time running NMR spectra on her compounds. “I realized that's what I liked most about my research,” she says.

After she finishes her graduate work, DeRider plans to obtain a postdoctoral fellowship to continue using NMR to study protein structure and then to teach at a small college similar to her alma mater.

Back to Top

Got It?

Give one advantage and one disadvantage of NMR when compared to X-ray crystallography.

What do NMR spectroscopists learn from a NOESY experiment?

Why is it important to study protein folding?

Next Chapter

This page last reviewed on October 31, 2013