NMR is no longer the second fiddle. It now plays a bigger role than ever in protein structure determination.
In 1993, when Frank Sonnichsen, PhD, and colleagues solved the structure of a type III antifreeze protein using NMR—nuclear magnetic resonance (Protein Sci. 1993 Sep;2(9):1411-28.), this was an unusual thing to do. The whole procedure took about a year, and it was considered second-best to X-ray crystallography. Traditionally, NMR has been chosen for protein structure studies when crystallography could not be used—for example, if the protein could not be crystallized, or if it was thought that the structure in solution was different from the solid state structure of the crystal. Since then, a culture of NMR proteomics has emerged within the field of structural proteomics, populated by a unique breed of scientist. A scientist who commits to protein NMR must be willing to play second fiddle to crystallographers on the conference circuit, and take comfort in the knowledge that his instrument is surely more expensive than an X-ray machine.
The weather is beginning to change, however, with more powerful magnets and more sensitive probes. NMR is becoming a solid first choice for distinct classes of proteins and types of protein studies. Sonnichsen’s interest in antifreeze proteins remains as keen now as it was in 2003. Originally discovered in fish, then in insects and plants, antifreeze proteins are structurally very diverse. Rather than being one single class of proteins, they evolved independently from several unrelated proteins, as they were adapted by different organisms for their serendipitous property of binding to ice crystals to prevent seed-formation, which otherwise would lead to fatal freezing.
Sonnichsen, an associate professor in the Department of Physiology and Biophysics at Case Western Reserve University, Cleveland, Ohio, has pursued mechanisms of action of the antifreeze proteins and structure is a key window into mechanism. Says Sonnichsen, “The protein fold determines what the protein can do. Knowing the structure can tell you a lot about the mechanism, for example the geometry of an active site, how a reaction is catalyzed, and why it is specific.” Fifteen years after his first antifreeze protein structure, he continues to contribute to structural knowledge of antifreeze proteins with a new publication on RD3 antifreeze proteins. (Biochemistry. 2008 May 7 [Epub ahead of print])
Building a better mousetrap
Physically, the NMR scanner is much like the more familiar medical MRI scanner—a large magnetic “donut.” For chemical and biological applications, the “donut” is turned ninety degrees so that it is perpendicular to the floor, and the “donut hole” is shrunk down to about two inches. The sample is inserted into a glass tube from the bottom, and held in place at the top by a probe, which is connected to the spectrometer part of the instrument. The probe is an interface between the spectrometer and the magnet. It delivers radio frequencies that excite the nuclei of the atoms from ground state to excited state. The return to ground state can be observed via emission spectroscopy.
NMR spectroscopy has some major limitations. One is that the technique, in general, is not very sensitive compared to other types of spectroscopy. A relatively large amount of material is needed to give a detectable signal. Another common obstacle is that it is a time- and labor-intensive study that takes months to produce a single protein structure. However, new technologies are overcoming these limitations, bringing sensitivity and speed of data acquisition and analysis closer to X-ray crystallography and other proteomic methods.
The most significant technological advance is the field strength of the magnet. An entry-level magnet for protein NMR would be in the 400 mHz range, but magnets up to 900 or 950 mHz are available through Varian, Inc., Palo Alto, Calif., and Bruker Biospin, Billerica, Mass. Both manufacturers anticipate that magnets will become more and more powerful in the future. Unlike a common refrigerator magnet, these are sophisticated electromagnets. The largest magnet ever sold by Varian went to the Pacific Northwest National Laboratory (PNNL). It was a unique wide-bore magnet—the donut hole was about five inches instead of two—and it was so massive that it had to be shipped through the Panama Canal. The magnet is about 20 feet tall and 10 feet wide, filled with liquid nitrogen and liquid helium. The supercooled interior is so efficient that it does not need a continuous supply of electricity, just occasional top-offs of nitrogen and helium. A room the size of a basketball court is needed to contain its magnetic field. The PNNL magnet is used for biosolids studies, a field with some promise for the future of proteomics, although still in its infancy. Biosolids has the exciting potential to reveal clues about proteins that exist in fixed or “solid” states within the cell, such as ion channel receptors or other cell surface proteins.
Although not every laboratory can afford—or could even use—such a large magnet, the more common laboratory instruments are many times more powerful than the older instruments, such as that used in 1993 by Sonnichsen’s team. A tantalizing technology on the horizon is high temperature superconductors. Says Mark Dixon, PhD, software marketing manager for Varian: “If and when they produce high temperature superconductors, and if they make that into a pliable wire, the whole magnet technology is transformed overnight. When that technology comes online, our magnets come down to benchtop size.”
Another exciting technical development is the cold probe. The probe is cryogenically-cooled to the temperature of liquid helium, which causes the thermal noise in the electronics to fade almost to nothing. That makes the instrument more sensitive and raises the signal-to-noise ratio.
Using stronger magnets, cryogenic probes, and other technical improvements in the instrument, it is possible to successfully characterize the structure of very small amounts of proteins—in the range of about 200 micrograms. This can be a very important fact for a rare or difficult-to-prepare protein. (And aren’t they all?)
For the impatient spectroscopist
John Markley, PhD, is principal investigator on a Protein Structure Initiative (PSI) project at the University of Wisconsin, Madison, called the Center for Eukaryotic Structural Genomics. The specialized center is part of a system of centers mandated by the National Institute of General Medical Studies (NIGMS)—a part of the National Institutes of Health—to determine the structures of as many proteins as possible. The PSI is in the second phase of its existence, following a five-year pilot phase during which PSI centers developed new technologies and new methods for protein structure determination.
For streamlined protein production, Markley produces mRNA from cDNA coding for the target protein and uses an extract from wheat germ that contains all of the cellular machinery (including ribosomes) needed to convert mRNA into protein, with added energy compounds N-15 and C-13-labeled amino acids for raw materials. The advantages of the cell-free system are that protein production is rapid and there is no cell lysis step in the prep. The in vitro reactions take place in a volume of a few milliliters, and the purification is a simple affinity column step. From beginning to end, the automated process from DNA to purified labeled protein takes 36 hours.
Markley’s data collection method is also much faster than standard methods—by as much as ten-fold. Says Markley: “What we’re doing now is using a computer to analyze the results as they come in. On the basis of this analysis, the computer decides on the fly what to do next, what data to collect, and under what conditions. The analysis involves determining where multidimensional NMR signals are located and associating these locations with particular groups of atoms. Instead of collecting a whole bunch of data and analyzing it after the fact, we can do it at a much more efficient rate by making course corrections at every step along the way.” He compares traditional data collection methods to wandering in a forest blindfolded to find the trees. “In our approach, we’re able to look and record what we see as we go along. It allows us to figure out the most efficient strategy for finding [the trees]. We want to find where the peaks are located. We want to associate those with individual atoms.”
In this case, the “trees” are “tilted planes” of data in the three-dimensional protein space. The computer algorithm chooses the angle of the next plane to be collected. It estimates whether the next plane will improve the model, and if not, it terminates the current experiment and begins a new one. (Eghbalnia, H.R., Bahrami, A., Tonelli, M., Hallenga, K., Markley, J.L. (2005). High-resolution iterative frequency identification for NMR as a general strategy for multidimensional data collection. J Am Chem Soc. 127, 12528?12536.) This method is known as HIFI NMR (High-resolution Iterative Frequency Identification for NMR). “In our approach, we focus directly on the goal of associating spectral frequencies with individual atoms in the covalent structure of the protein, which we know already from its gene sequence. This shortcut speeds up the process of structure determination. Our current approach of combining HIFI and PINE should reduce the once-epic task of solving a protein structure through NMR to just one or two weeks.”
With the newest advancements in technology, there is no reason for NMR spectroscopists to play a backup role to crystallographers. Rather, each technique has its unique and complementary strengths. Protein NMR excels at characterizing “live” and “wiggly” proteins, and those with large, random domains that do not crystallize prettily. The NMR studies can now be accomplished in a timeframe competitive with crystallography, and are critical to any large-scale protein structure studies. Additionally, anticipated advances in technology could very well revolutionize the way that NMR is used in structural biology, making it an important method to watch for the future.
About the Author
Catherine Shaffer is a freelance science writer specializing in biotechnology and related disciplines with a background in laboratory research in the pharmaceutical industry.
This article was published in Drug Discovery & Development magazine: Vol. 11, No. 6, June, 2008, pp. 34-36.
Filed Under: Genomics/Proteomics