LBL scientists are using NMR spectroscopy to probe structure- function relationships in complex biological molecules
If biological molecules were not so enormous, David Wemmer's life would be a lot easier--and a lot less interesting.
Wemmer, a chemist who got started in nuclear magnetic resonance (NMR) research as a student in Alex Pines' group at LBL in the '70s, spends his time these days applying NMR spectroscopy to the life sciences--especially, to the new frontier of structural biology.
Modern structural biology tries to relate the structure of molecules to their biological function and to understand how such structure affects fundamental cellular processes such as enzyme catalysis, mutagenesis, and DNA synthesis. The tools of this new discipline include NMR and x-ray crystallography as well as photochemical and molecular genetic techniques.
The group Wemmer heads, a part of LBL's Structural Biology Division, has space in the Calvin Lab on campus and also in the Tritium Labeling Facility on the Hill. Besides pursuing its own vigorous research program on the structure, function, and interaction of biological molecules, the team serves as a Lab- wide resource group, using NMR spectroscopy to solve problems in the life sciences for a variety of LBL collaborators and outside users.
The root cause of many of these problems is the fact that biological molecules (chiefly nucleic acids and proteins) can be up to a thousand times bigger than the simple organic and inorganic compounds that Wemmer encountered back in his student days. It follows that scientists seeking to apply NMR to these giant systems must devise ingenious new strategies to cope with a host of complexities.
Of the many analytical techniques available to the biological researcher, only two--NMR spectroscopy and x-ray crystallography--permit scientists to relate the detailed structure of a molecule (such things as its shape, the number and kind of projections, appendages, and folds, or the sequence of its amino acids) to its function in the organism.
The two techniques are often used together, since their areas of application are complementary. Not every molecule can be studied with x-ray crystallography, because not all can be induced to form crystals. (And even some that can are so distorted by the process that they bear little resemblance to their structure in nature.) Similarly, not every molecule can be studied with NMR.
Nuclear magnetic resonance is a phenomenon that occurs in the nuclei of many--though not all--elements when they are placed in a magnetic field. In such elements, the spinning protons and neutrons in the nucleus oscillate like tiny gyroscopes and line up with the magnetic field.
In NMR spectroscopy, a sample is placed in a magnetic field, which forces the spins of the nuclei into alignment. The sample is then bombarded with radio-wave pulses. As the nuclei absorb energy from a pulse, the spins topple out of alignment with the magnetic field, and as they lose the absorbed energy they line up again. By measuring the specific radio frequencies that are emitted by the nuclei, and the rate at which realignment occurs, scientists can gain detailed information on the atomic structure and motion of the atoms within the sample.
But the molecules and chains important in biology tend to be both huge and complicated. For example, a typical protein molecule may weigh 25,000 grams per mole or more and may contain thousands of protons with overlapping resonances. Making up the proteins are amino acids; and, to further confuse the picture, 10 or more copies of the same amino acid may occur together on a chain. The result is what Wemmer calls the "assignment problem" -- How do you know which molecule or which part of a complex molecule the NMR signal is coming from?
For many years, the assignment problem effectively inhibited the use of NMR spectroscopy in biological research and the life sciences, although NMR imaging techniques (based on measuring the density of various parts of a sample) have found a fertile ground in those fields. Over the past 10 years, however, things have changed, and groups like Wemmer's today are resolving the structure of biological molecules to a level of precision that has never been attempted before.
One of the things that made this possible has been the availability of on-line computers, which handle complex data- acquisition and analysis, as well as reconstructing the structure of molecules or chains from even the most complex spectroscopic data.
A second advance that has opened the door to the use of NMR in structural biology is the development of "multidimensional" NMR. In conventional one-dimensional NMR spectroscopy, Wemmer explains, each proton has both a characteristic frequency and also a characteristic coupling pattern--variations in that basic frequency that are related to the proton's proximity to its neighbors. Using these two constants, scientists have been able to analyze the structure of thousands of inorganic and organic molecules. But they found themselves stymied by the complexity of biological molecules.
"The solution," says Wemmer, "has been multidimensional NMR --a technique originally suggested by Belgian chemist Jean Jeener in 1971 and developed primarily by Richard Ernst, Ray Freeman, and others. In this technique, we look at the frequencies of a pair of spins--two protons that are somehow connected to each other. They can be paired either structurally--as part of the same functional unit--or geographically--simply close to each other in space. When two protons, rather than one, are associated with a given NMR signal, we can calculate the distance between the two members of the pair; then the computer reconstructs the structure of the whole molecule from the experimentally measured distances of thousands of such pairs.
"If the bonding between the two protons in a pair is structural, it tells us what kind of molecule we're looking at--for example, an amino acid rather than a nucleic acid. If the bond is geographical, it tells us which molecules are neighbors, how the contacts between neighboring amino acids work, and so on.
"For example, complex folding--including hairpin turns, helices, and so forth--is frequently found in proteins. If we see that certain amino acids are found close together in a particular molecule, even though they are not actually neighbors in the sequence (not structurally joined), this is an indication of folding."
Another recent development is the use of tritium-labeled tracers in NMR spectroscopy. This technique was pioneered in LBL's Tritium Labeling Facility, which is headed by Wemmer and staffed by Hiromi Morimoto, Phil Williams, and Manouchehr Saljoughian.
In order to be suitable for NMR spectroscopy, a nucleus must exhibit what is known as a "magnetic moment." Magnetic moment arises from the ratio of charge to angular momentum (spin) in a nucleus; this ratio, in turn, depends on the unique combination of protons and neutrons in a particular isotope. In some cases, the two cancel each other, and there is no magnetic moment. This situation occurs almost randomly in the periodic table and is difficult to predict even from detailed nuclear data.
As it happens, the best of all isotopes for NMR spectroscopy is tritium, the radioactive isotope of hydrogen with three nucleons (a proton and two neutrons) in its nucleus. Tritium's superiority for NMR has nothing directly to do with the property that has made it so valuable as a tracer in other types of research--its radioactivity. It is due simply to an advantageous addition of the magnetic moments of protons and neutrons, which gives tritium the largest magnetic moment of any nucleus.
"Added to that," Wemmer points out, "is the fact that there is no natural background of tritium, since all the tritium on the Earth's surface decayed away long ago. When you detect a tritium NMR signal, you can be sure it's coming from the label that you placed in the molecule."
Despite its natural advantages, health and safety considerations associated with tritium's radioactivity delayed its exploitation in NMR spectroscopy. At LBL, however, the availability of the Tritium Labeling Facility, with its specialized systems for handling and waste disposal of radioactive materials, has made the use of tritium in NMR spectroscopy safe and routine.
Samples are prepared by chemical synthesis of a precursor, in most cases, with the actual introduction of tritium at the last stage. In some cases, the labeled molecule may then be bound to a macromolecule of interest or may be enzymatically incorporated into a macromolecule.
As one enthusiastic tritium user, Structural Biology's Richard Newmark, wrote, "With a tritium label, we can boost NMR sensitivity by a factor of 10 and reduce the data-collection time by a factor of 200. I can collect data in one minute, using tritium, that would take an hour and a half using carbon-13."
Out of all the millions of biological molecules that are potential candidates for NMR investigation, how do spectroscopists choose the few to be analyzed in depth--a procedure that may take (in the case of a complex protein containing 100 amino acids or more) a year of study of a single sample?
That depends, says Wemmer. In NMR developmental work, when the principal objective is to explore a new method or refine a technique, the team may pick a molecule simply because it is convenient to work with. On the other hand, when the focus is on a particular biological question, the crucial molecule may have to be identified before it can be analyzed.
Most biological molecules have very specific and narrowly defined functions, Wemmer explains. For example, a given protein molecule may function as a catalyst, enhancing or speeding up the production of an amino acid or releasing a hormone. To identify that molecule in a sample of biological fluid or tissue is a fairly straightforward matter of successive chemical separations of the sample, followed by testing to see which fraction can still perform the function.
"This method delivers the molecule you're looking for," says Wemmer, "but it doesn't tell you a thing about how the molecule does what it does, or what part of the molecule plays the major role. That's where NMR spectroscopy comes in."
One such recent experiment in Wemmer's lab, with team members Kalle Gehring, Philip Williams, Jeffrey Pelton, and Hiromi Morimoto, focused on the general problem of how one biological molecule recognizes another. The specific question involved a protein that is responsible for transporting maltodextrins--important sugars--into bacterial cells.
It was known that the protein seeks out long sugar chains and binds to them, but it was not known how it recognizes these chains. The LBL team's NMR analysis, reported recently in the journal Biochemistry, showed that the protein particularly recognizes the terminal glucose residue of the maltodextrin and in fact distinguishes whether an oxygen (hydroxyl) group at the end is "up" or "down," relative to the sugar. The "up" form binds more tightly and is probably the form responsible for stimulating the bacteria to swim toward the source of the maltodextrin.
Another recent project, reported in the Journal of the American Chemical Society in July 1992, has to do with how a small molecule recognizes particular sequences of DNA and binds to them. This work, with co-workers Tammy Dwyer, Bernhard Geierstanger, Yadagiri Bathini, and J. William Lown, has important implications for the development of drugs used in fighting cancer and other diseases.
"Most of the cancer-fighting pharmaceuticals now available are either natural products, derivatives of such products, or synthetic versions of them--substances that have been found by trial and error to be useful," says Wemmer. "In the future, the arsenal may also contain `designer drugs' devised to enhance or inhibit particular biological functions. But in any type of drug development, synthetic or natural, it will be important that the drug contain the smallest possible molecules. That's because small molecules are taken up more easily by the body and are not chemically altered by digestion if taken orally.
"All the big drug companies have research programs aimed at developing such `small molecule' drugs," says Wemmer, "but there are problems in making a small molecule do what a bigger molecule can. We've studied one particular type of small molecule, known as Distamycin-A analogs. These molecules were known to have the ability to search out and bind to areas of DNA that have certain characteristics--for example, areas rich in AT pairs. But little was known about how they did it."
Through multidimensional NMR spectroscopy, Wemmer's team was able to identify the method these molecules use. It was then possible to change the structure of the molecule to target a new site on the DNA for the molecule to attach to and induce it to interact even more strongly than the original molecule.
"Though we had no specific drug in mind when we did this work," says Wemmer, "it's clear that the ability to manipulate bonds in this manner could have important applications for pharmaceutical drug development."
Another one of Wemmer's current projects, with co-worker Ann Caviani Pease, has to do with the important biological molecule known as ribonucleic acid, RNA.
In the world of the cell, it is the common wisdom that DNA carries information, while proteins perform chemical activities. But only recently has it been understood that RNA--long thought to be nothing but an intermediary or "messenger" between DNA and the cell--actually does both things, carrying out certain chemical reactions for which it also carries the code. What is not understood is how RNA performs this feat.
"We know that proteins carry out their functions through enzymes," says Wemmer, "and we know that RNA also has enzymes associated with it, but until recently no RNA enzymes have been looked at closely. The RNA enzyme we have been studying--one associated with cleavage of the RNA strand--is one of the first to be analyzed. Earlier researchers had studied the area where cleavage takes place and had proposed a "hammerhead" model for the shape of the structure involved. Our studies confirmed that the hammerhead structure is basically correct, and we have been able to add details about the folding that seems to precede cleavage."
Another recent success from Wemmer's lab was an experiment, with Milton Werner, to determine how a biological inhibitor binds to the enzyme that it inhibits. As is so often the case in biology, the molecule of interest in this case was too large to analyze even by multidimensional NMR, which has a practical limit of a molecular weight of about 25,000 (equivalent to about 2500 protons).
"In order to use NMR to analyze molecules this big," says Wemmer, "one needs to find ways to take apart the system and look at small pieces of it, one at a time. We developed new techniques that enabled us to isolate the part of the molecule that performs the inhibition function, and we also saw exactly how it does it--by binding to specific sites in the enzyme and making them unavailable for action."
The ability to take apart large molecules is proving particularly important in the study of catalytic antibodies, the synthetic enzymes for which LBL researcher Peter Schultz was recently honored with the E.O. Lawrence Award. In catalytic antibody synthesis, a chemist synthesizes a structurally similar but stable transition state analog of a molecule. The synthetic molecule is injected into a mouse, whose immune system produces an antibody that will bind loosely to the desired substrate. This catalytic antibody accelerates the desired reaction (see LBL Research Review, Spring 1988).
Though NMR spectroscopy does not figure in this initial step, Wemmer's team has worked with Schultz to help determine whether the strategy used to create a given antibody was the one envisioned in the original design of the transition state analog. Since most catalytic antibodies are far too big for direct NMR analysis (molecular weight 150,000 or so), it becomes necessary to split up the molecule and look only at the region near the binding site--still formidable at a molecular weight of 50,000.
The collaboration with Schultz is only one of many joint efforts that Wemmer is involved in. Because of the power and versatility of the NMR technique, Wemmer's team is frequently called upon by colleagues at LBL or UC to determine the structure of some molecule of interest. Recently, for example, they have studied the structure of pigments of photosynthetic bacteria (with Structural Biology Division biochemist Alex Glazer) and helped to identify a new co-factor of an enzyme (with Energy and Environment Division enzymologist Judith Klinman). Other current collaborations include analysis of transcription factors with UC biochemists Robert Tjian and Hillary Nelson, as well as a new biomolecular engineering project with Peter Schultz.
Besides this full agenda of research collaborations, just keeping up with new spectroscopic techniques requires constant effort, Wemmer says, because the NMR field is evolving so rapidly.