900,000 Daltons

By Jason Socrates Bardi


"There is no greater pleasure for me than to practice and exhibit my art."

—Ludwig van Beethoven, 1800.


Even with the most powerful and sensitive of modern nuclear magnetic resonance (NMR) instruments, many interesting systems are impossible to solve because they are simply too big.

And structural biologists have sought for ways in which to breach, stretch, surpass, or otherwise blow away this size limitation. Some biologists have wanted to do this for decades.

"We have submitted proposals as early as 1979 to do basic research on ways to overcome the physics barrier to doing solution NMR on large structures," says Kurt Wüthrich, who is currently guest scientist in the Department of Molecular Biology at The Scripps Research Institute (TSRI) and is scheduled to become a full-time faculty member in 2004. "Methods development is an important direction of my laboratory."

In fact, since the 1960s, Wüthrich has pioneered the field of NMR structure determination. He is professor of biophysics and former chair of the department of biology at the Eidgenössische Technische Hochschule (ETH) in Zürich, Switzerland, where he was involved in the first de novo protein structure determinations with NMR. And he is author of one of the definitive books on the subject, "NMR of Proteins and Nucleic Acids" (1986).

But until recently, even Wüthrich had to accept size limits for studies of biological macromolecules in solution using NMR.

The Trouble With Large Systems

"It used to be that there was a molecular weight limit well below 50,000 daltons for the use of solution NMR," says Wüthrich (A dalton is equivalent to the mass of one hydrogen atom), and this was not an engineering limit, but a physics one.

Large structures such as many protein/protein, protein/DNA or protein/lipid complexes are impossible to investigate with conventional NMR because of their slow response to the thermal motion of the solvent molecules in the test tube.

Proteins in solution are constantly bombarded with solvent molecules and undergo random "Brownian" motion as a result. The larger the structure is, the more slowly it will respond to the impact and the more slowly it will reorient itself in solution. This slow response deteriorates NMR spectra and makes structure determination impossible.

"This we have overcome by using single-transition spectroscopy," says Wüthrich, referring to the new technique of transverse relaxation-optimized spectroscopy (TROSY).

TROSY is a technique that suppresses the transverse nuclear spin relaxation—the event that causes the deterioration of NMR spectra for larger molecular structures.

For any two coupled "NMR atoms" in a molecule, there will be a total of four energy transitions that can be detected by NMR. Conventional NMR averages all four, but this averaging deteriorates the signal when the molecules get too large. TROSY selects only one of the four transitions—the one that is insensitive to Brownian motion in a high magnetic field.

The optimal frequency for the TROSY effect is calculated to be near 1000 MHz, and since TSRI's new 900 MHz NMR is closer to this than any instrument ever before, Wüthrich will be able to effectively apply TROSY to very high molecular weight systems on the new instrument.

With TROSY, the size limit of structures that can be solved with NMR is expanded several-fold. "We can now do reasonably detailed structural investigations of proteins in structures of size up to about 150,000 daltons," says Wüthrich, who was the first to recognize and exploit TROSY at ultra-high magnetic field strengths with his research group in Switzerland.

In structures that are even larger than 150,000 daltons, though, TROSY is no longer effective on its own. But Wüthrich has pioneered other techniques, such as cross-correlated relaxation-enhanced polarization transfer (CRINEPT). When combined with TROSY, these result in highly effective experiments for very large structures.

"We can go up to one million [daltons], essentially," he says.

Solving Structures Great and Small

Besides developing methodology, much of Wüthrich's research is devoted to solving novel structures of biological molecules, and he works on many systems, from the very small to the very large.

He studies pheromone molecules and their receptors from insects and other organisms, for instance. Pheromones may be small organic molecules or very stable proteins of about 35 to 50 residues, which transmit a behavioral stimulus, often sexual. Animals emit pheromones in minute amounts and have receptor systems that can detect these.

Wüthrich has solved numerous protein pheromone structures from Mediterranean sea creatures and pheromone-binding proteins from other organisms, such as silkworms.

He also has a program looking at ways of solving membrane protein structures. Of all the relevant molecular structures in biology, membrane proteins have been solved the least—mainly because they are so hard to work with—and there is a great many membrane proteins that need to be solved. Less than one half of one percent of the structures contained in the Brookhaven National Laboratory Protein Data Bank are of integral membrane proteins, despite the fact that over a third of all proteins in the body are in the membrane.

Wüthrich's group studies the recombinant expression, reconstitution, and isotope labeling of integral membrane proteins for use in NMR. These are not particularly large proteins, typically 200 residues, but the experiments require the proteins to be suspended in large soap-like micelles, which mimic the membrane and preserve the fold of the proteins.

"You can't do conventional NMR on such things," says Wüthrich, "only TROSY enables such projects."

Wüthrich also is looking at chaperonin systems like GroEL—the very large protein complex that monitors the folding of newly synthesized proteins.

GroEL is a multimeric protein whose subunits—14 of them—come together to form two 7-subunit toroids that stack together. This 800,000 dalton structure binds misfolded proteins, which fit into the inside cavity of the toroid shape where they can refold. Since GroEL is such a large system, it cannot be studied by conventional NMR.

Combining CRINEPT and TROSY, however, says Wüthrich, "We can do GroEL and we have observed other, smaller proteins when bound to GroEL."

Prions Under the Magnet

One of the largest areas of study in Wüthrich's laboratory involves prion proteins. Mis-folded prion proteins have been suggested to cause bovine spongiform encephalopathy, or mad cow disease, and a form of the same disease in humans, called variant Creutzfeldt-Jakob Disease.

Prion proteins are expressed widely throughout the body and sit anchored onto the surfaces of cells in a wide variety of tissue, particularly on cells in neuronal tissue.

Infectious, malformed prion proteins start out with one shape, which is innocuous, and end up with another shape, which is observed in organisms suffering from a deadly "prion" infection. Infectious prions from an animal with mad cow disease, for instance, are believed to transmit the disease by initially causing normal prion proteins in the brain of a healthy cow to misform into the infectious form. Then these prions will act on more normal prion proteins to produce more and more misfolded proteins that accumulate and eventually lead to a sponge-like build-up and brain damage.

Wüthrich concentrates on comparative studies of the normal form of the prion protein in various species.

"[We want] to get the molecular basis of the species barrier," says Wüthrich. "Why are there no records of transmission from sheep to man, but there is mounting evidence that there is transmission from cattle to man?"

The assumption is that the more similar the prions are across species, the easier the transmission will be, as in cows to humans.

"Our results so far show that the healthy forms of the prion protein between man and cattle are identical in the folded part of the molecule," says Wüthrich. Other species, he adds, show differences even though the global fold is maintained.

Wüthrich does not stop at cows. He has solved the structure of the prion protein from chickens, for instance, and he is almost done with that of the turtle.

Prions are interesting also because much of the molecule is unstructured. The protein has a long tail that is highly flexible and is as much at ten times longer than the diameter of the folded part of the protein. "By studying evolutionarily widely divergent species, we hope to possibly target some clues as to the natural function of the prion protein, anticipating that the active site would be preserved," says Wüthrich.

Wüthrich is also interested in making preparations of prion protein aggregates that could be used to study the misfolded protein and the molecular basis of the aggregation. The needs are tantalizingly simple: a sample of isotope-labeled prion protein in solution that form repetitious aggregates. But the difficulty is preparing a sample that aggregates only a little, from two to forty proteins in a clump, as opposed to one that forms fibrils and crashes out of solution.

"If we had such preparations of aggregated prion proteins, we probably would have data from TROSY and CRINEPT experiments already," says Wüthrich.

Works In Structural Genomics

Since the start of his laboratory at TSRI in October 2001, Wüthrich has also been collaborating with the Joint Center for Structural Genomics (JCSG), a $30-million effort to develop high-throughput technology that could one day support efforts to find and catalog the structures of all proteins active in the human body. The JCSG is a multi-institution collaboration sponsored by the National Institutes of Health and led by TSRI Molecular Biology Professor Ian Wilson.

With the JCGS, Wüthrich is planning to use NMR as a tool to test sample preparations. What is the effect on the fold of a protein, for instance, when you add a histidine tag, typically several consecutive histidine residues that allow the protein to be highly efficiently separated on a column.

The idea is to use NMR as a screening tool to evaluate the quality of protein preparations from the automatic procedures—to check on the results and tighten the biochemistry used to prepare the samples. Choosing a biochemical technique exclusively for its amicability to the automation process may not be the best solution, since in the end the most important thing is having pure, relatively unmolested samples.

"We have the potential with our technique to make a major impact," says Wüthrich.


Go back to News & Views Index

Investigator Kurt Wüthrich is a pioneer in the field of NMR structure determination. Photo by Alan McPhee.


















NMR structure of the pheromone-binding protein from the silkworm Bombyx mori, BmPBP. Seven a-helices connected by linkers with nonregular secondary structure are shown in grey color and disulfide bonds are yellow. The core side chains are shown as bundles of 20 superimposed conformers, with red color for the side chains of the "regulatory" helix 7 and green color for all other core side chains.














NMR structure of the bovine prion protein, bPrP(23-230). A well-defined globular domain comprising residues 123-230 is green. An impression of the state of a flexible "tail" of residues 23-122 is provided by a superposition of 40 grey lines representing snapshots taken at short time intervals.