Nuclear Magnetic Resonance (NMR) spectroscopy has become an extremely familiar analytical tool in chemistry and biochemistry laboratories. Even researchers with little exposure to the technique recognize that NMR can provide a great deal of information about everything from the acetone content of a poorly prepared undergraduate chemistry lab sample to the structure and dynamics of complex biomolecules. In recent years the analytical potential of NMR has expanded, offering researchers a growing arsenal of experiments for both solution and solid state applications. Most recently, NMR has entered the arena of high-throughput screening, providing new methods for drug discovery and structural genomics research.
For a technique that offers such complex information, the basic principles behind NMR are straightforward. The nuclei of certain atoms, for example, 1H, 13C, and 15N, exhibit a physical property known as spin. These nuclei can be viewed as tiny magnets that, when placed in an external magnetic field, can orient themselves in two possible ways, with spin vectors aligned in the direction of, or directly against, the field. For nuclei with a nuclear quantum spin number of 1/2, such as those listed above, these two orientations correspond, respectively, to a low energy state and a high-energy state. Transitions between the two states occur spontaneously, but infrequently.
However, if the sample is irradiated with energy equivalent to the energy difference between the two states--in the radio frequency, or RF, range--transitions will occur more frequently. These induced transitions form the basis of NMR spectroscopy. When the magnetization vectors associated with the transitions are rotated perpendicular to the applied field, they precess about the direction of the field and induce a current in the receiver coil, which is recorded and plotted as a function of time. The resulting sine wave decays with time due to spin dephasing, and the signal is recorded as a free induction decay (FID), which is then converted into a frequency domain spectrum.2
A number of factors help NMR spectroscopists elucidate the structure of a given molecule. The chemical shift, or the frequency at which each nucleus resonates when a magnetic field is applied, provides information about the chemical environment of a nucleus. Spin-spin coupling, or J coupling, provides information about the relationship between nonequivalent nuclei connected through bonds. The Nuclear Overhauser Effect (NOE), or the interaction between the dipole moments of two nuclei in spatial proximity, provides information about the distance between nuclei and is one of the parameters studied in multidimensional NMR.1
In order for a signal to be detected, there must be a difference between the populations of nuclei in the lower and upper energy states. According to the Boltzmann equation, the difference in population, and thus the strength of the signal, increases with the strength of the external magnetic field. As a result, a stronger magnetic field will increase sensitivity.1 In addition, as the strength of the magnetic field increases, the difference between the resonance frequencies of different nuclei will also increase, thus resulting in better resolution of peaks.
Varian's Infinityplus system
Higher and Higher
One of the most exciting recent developments in NMR technology is the introduction of NMR at the highest field strength currently available, 900 MHz. Two major suppliers of NMR equipment, Bruker Instruments of Billerica, Mass., and Varian Inc. of Palo Alto, Calif., introduced systems with 900 MHz magnets suitable for high-resolution NMR within the past year, and JEOL USA Inc. of Peabody, Mass., has a 900 MHz NMR under development. Bruker was the first company to successfully energize a 900 MHz magnet in 1999, and Varian was the first to present high-resolution NMR data from the 900 MHz magnet (see www.varianinc.com). However, the path to development has been far from smooth. The world's first commercial 900 MHz NMR, developed jointly by Varian and Oxford Instruments PLC of Oxfordshire, U.K., was scheduled to be delivered in the summer of 1998 to the Pacific Northwest National Laboratory (PNNL) in Richland, Wash.3; owing to technological complications with the magnet, delivery is now set for May or June of next year.
In addition to creating a magnet that demonstrates the demanding levels of spatial and temporal stability that are necessary for high resolution NMR, the developers of the 900 MHz magnet faced two obstacles. The higher the magnet's field, the greater the amount of energy stored in the magnet; a 900 MHz magnet contains approximately 15 MJ of stored energy. As the energy increases, forces on the wire conductor build up outward from the center of the magnet.
"When you pack more and more energy into these magnets, the forces that develop in them are very severe, and the magnet essentially wants to tear itself apart," says Raymond Shaw, vice president of NMR systems at Varian. "You have to make these magnet structures very strong to withstand the stresses that are built up as you put current into these magnets," Shaw says.
The second problem is that a magnet can suddenly release all of its energy, a process known as quenching. A magnet must be designed so that in the event of quench, all of the released energy can be dissipated without destroying the magnet's coil. To solve these problems, magnet designers have developed new superconducting materials for stronger coils, created new coil designs, and devised ways to protect a magnet during a quench. For example, Oxford Instruments has devel-
oped a Niobium-Tin-based material called UltraSn™ for use in superconducting coils capable of carrying current densities up to 200 A/mm2 and has devised a real-time energy management system to protect the magnet during a quench.
Complicating the basic technological problems associated with high-field NMR is the cost of the instruments themselves, which may limit accessibility to the technology to only a small number of national laboratories and large, well-funded universities. A standard 600 MHz NMR costs roughly $800,000, but the 900 MHz sells for about $5 million. "It won't be the kind of machine that a small college is going to buy," says Shaw. Universities are thus likely to acquire the instruments jointly, and collaboration and shared research time will become more prevalent.
The Department of Energy's Environmental Molecular Sciences Laboratory (EMSL) national user facility, located at PNNL, which will house the first 900 MHz NMR system, offers scientific users access to its NMR facilities, including the higher field instrument, according to David Koppenaal, associate director of the Macromolecular Structure and Dynamics Directorate of EMSL. Despite the high cost, orders are coming in for the magnets. Mark Chaykovsky, Bruker's vice president of marketing and sales, analytical NMR, says that his company has thus far received seven orders for its Avance™ 900 system.
Why the interest in high-field NMR? The spectra of complex biomolecules contain a large number of peaks, many of which are close together or overlap. Higher field magnets, or higher frequency instruments, offer better peak resolution, enabling analysis of larger and larger molecules. Also, in NMR, sensitivity increases almost with the square of the magnetic field, so when magnetic field strength is doubled, sensitivity increases about fourfold. Data can thus be acquired faster, or alternatively, samples can be run at lower concentrations in the same experimental time. The latter advantage is particularly important to the study of large biomolecules, which are often difficult to express and purify in large quantities and can aggregate and precipitate out of solution at high concentrations. Finally, high-field NMR can lead to the development of new NMR experiments that exploit properties exhibited by molecules at high magnetic fields. For example, Varian's Web site shows the spectrum of the protein YciH from a TROSY (transverse relaxation-optimized spectroscopy)4 experiment run on the 900 MHz NMR; the data show significant line narrowing. According to Koppenaal, this result is theoretically expected at high fields: "The resulting line width [should be] narrowest at an optimal field strength; theoretically that optimal field strength is calculated at 1.1 GHz for proton-nitrogen pairs. This preliminary data bears out that we're moving in the right direction."
JEOL's Eclipse+ system
But higher fields are not the only means of improving the sensitivity of NMR. Advances in cryogenic probe technology also have broadened the capabilities of NMR. Bruker, for example, has pioneered the development of cryogenic probes, and similar probes are under development by JEOL. The Bruker CryoProbe™ technology provides three to four times the sensitivity of conventional probes and reduces acquisition times by a factor of 16. The probes have been used in high-throughput applications with a high degree of success.
One application for the use of these probes involves a technique for ligand screening developed several years ago by Stephen Fesik, NMR spectroscopist at Abbott Laboratories in Abbott Park, Ill. This technique, known as SAR (structure-activity relationships) by NMR, uses the chemical shift changes observed in 15N-heteronuclear single quantum correlation (15N-HSQC) spectra to identify small molecules that bind to target proteins.5 CryoProbes allow NMR-based screening to occur at high-throughput levels, with threefold decreases in the amount of sample needed.6 In addition, the technology has been beneficial to the field of structural genomics, as it makes high-throughput protein structure determination possible.7
"What used to take two months of data acquisition using conventional probes for determining a protein structure now takes four to seven days," says Fesik. In his view, CryoProbe technology overshadows the advances made by high-field NMR, as "even [with] an 800 MHz NMR, the sensitivity improvements that you get over a 500 MHz NMR are not as great as the sensitivity improvements that you get with the CryoProbe, and the 800 costs a lot more."
The ability to study samples at lower concentrations also has been put to use in the study of natural products. Other than the proton, the nuclei that NMR can "see"--in particular, 13C and 15N--are not abundant in nature, and researchers studying natural molecules often artificially enhance their samples, a procedure that can take months to complete. CryoProbes eliminate the enrichment step: "People can literally start taking samples of compounds out of nature and put them into the NMR without having to go through four to six months of chemistry," says Chaykovsky.
Bruker NMR also offers MicroCryo Probes™ for 3 mm sample tubes. Traditionally, NMR samples are placed in 5 mm tubes, with a high solvent-to-sample ratio. With the increased sensitivity offered by accessories such as the CryoProbes, researchers have noticed more solvent impurities in their spectra. The 3 mm tubes and MicroCryoProbes allow spectroscopists to study mass-limited samples, such as natural products and metabolites, in less solvent.
A flat-coil probe from Doty Scientific
For those seeking complete integration of analytical techniques, Bruker offers a combined LC-NMR-MSn system. The system features Bruker NMR Avance™ NMR technology, the Bruker Daltonics esquire3000™ Ion Trap Mass Spectrometer (MS), and industry-standard HPLC. By combining mass spectrometry, a highly sensitive technique, with the less-sensitive NMR, the LC-NMR-MSn system helps to streamline the drug discovery process.
Because acquisition times for NMR can run as long as 24 hours, isolating the correct molecule for analysis is crucial. With LC-NMR-MSn, researchers can inject the sample, identify a target by its molecular weight, isolate it, and analyze it directly by NMR. Samples injected into the system are tracked by computer, eliminating the ambiguity that can arise when a sample is sent to three separate instruments for analysis: "Part of the problem with doing research [on mixtures], when you send the sample to the NMR group, then the LC group, then the mass spec group, you want to make sure that everyone is looking at the same compound when comparing data. With our design ... you know all of the spectra are of the same compound in the separation," says Chaykovsky.
One area in which the LC-NMR-MSn has been advantageous is the study of drug metabolites. John Shockcor, principal research scientist at DuPont Pharmaceuticals Co. of Wilmington, Del., uses the system to determine the structure of metabolites isolated from body fluids.8 Shockcor notes that the system allows his group to accurately correlate NMR and mass spectrometry results, which is often difficult to do when using HPLC to identify compounds. "More often than not, one chromatograph doesn't operate like another ... peaks are very close together and can sometimes switch places. This way ... all three techniques are looking at the same peak," he explains. The LC-NMR-MS also permits samples to be analyzed more quickly; for example, Shockcor's group has characterized 13 metabolites from bile in less than two days with one injection. Moreover, by combining the three experiments in one system and in one room, more efficient communication between members of a research group is achieved. "It makes for a nice working environment for scientists doing structure elucidation," says Shockcor.
Advances in Automation
Another time-saving development from Bruker NMR is the BEST™, or Bruker Efficient Sample Transport, system. This method incorporates a Gilson automated liquid handling system to uptake samples directly from a rack or 96-well plate and send them directly to be analyzed by NMR. By eliminating the need to place each sample in an individual NMR tube and then place each tube in the machine, the BEST system provides high-volume, high-throughput NMR.
JEOL also offers "tubeless NMR" with the Eclipse+ system, which features JEOL's Delta™ software with Flex Automation™ for high-throughput applications such as combinatorial chemistry screening and biofluid analysis. The system automates the entire NMR process, including sample injection, data acquisition, and probe flushing. While the Eclipse+ is available from 300 MHz to 900 MHz, NMR product manager Doug Meinhart notes that the system is geared mainly toward high-throughput users in the 300 MHz to 500 MHz range, JEOL's main customer base. Another highlight of JEOL's automation products is the AutoTune™ Broadband Probe, which offers automated, computer-controlled probe tuning for multinuclear and multisample applications.
Bruker's Avance System
State-of-the-Art Solid State NMR
Like solution NMR, solid state NMR has experienced an explosion in recent years, particularly in the study of membrane-associated proteins, which can't be studied by solution NMR methods. Solid state experiments typically involve higher RF power levels than corresponding liquid state NMR, and to accommodate these power levels, the magnet bore, the cylindrical hole in the middle of the magnet in which the sample is placed, needs to be larger. Liquid NMR uses narrow bore (54 mm diameter) magnets; solid state NMR requires wider-bore (63 or even 89 mm diameter) magnets. The higher stored energy of these wide bore magnets means that they are significantly more difficult to build, and as a result high-field solid state NMR lags behind liquid state in terms of available field strength. The highest field currently available for a wide bore magnet is 800 MHz.
Varian's Infinityplus™ spectrometer was developed specifically for solid-state NMR; it offers the highest field available for this technique (800 MHz), thus making analysis of large biomolecules using solid-state NMR a possibility. This will prove critical to the human genome project, as many of the known gene sequences encode for membrane-associated proteins.
"The human genome encodes for about 40,000 membrane associated proteins ... and we have the structure for only about 20 such proteins from all sources," explains Tony Watts, professor of biochemistry at Oxford University, Oxford, U. K. One area of research explored by the Watts group is the elucidation of the structures of membrane proteins bound to activating molecules such as drugs and natural ligands. The proteins are studied in their active state in a natural membrane environment. The magnet for the 800 MHz Varian Infinityplus spectrometer in Watts' NMR facility is currently being energized, and when fully functional, the instrument's higher sensitivity will help his group work with smaller amounts of material than is possible with currently available lower frequency instruments. "In our area, where protein amounts are always at a premium, high sensitivity is a major issue," Watts says.
Bruker and Varian offer accessories for high resolution Magic Angle Spinning (hr-MAS) experiments in which the sample is spun at a high speed at a well-defined angle to the main magnetic field. Orienting the sample in this manner is a technique that was developed originally for solid state NMR to minimize the spectral line broadening introduced by intermolecular dipolar coupling or sample inhomogeneity.2 The appeal of hr-MAS is that it allows the study of systems that were previously not accessible with NMR. Some particularly interesting recent applications of this technology include the study of biological tissues and combinatorial chemistry samples that have been isolated on polymer beads.
Another company offering products for solid state MAS is Doty Scientific of Columbia, S.C. Doty, a leader in probe technology, sells MAS probes for high-field magnets up to 1 GHz. The probes are available with a 4, 5, or 7 mm spinner and can be double- or triple-tuned. As sales manager Laura Holte notes, Doty's triple-tuned MAS probes are available in narrow as well as wide-bore formats, which makes them unique in this field: "Other manufacturers aren't able to accomplish three resonances simultaneously in a narrow bore. The conventional approach of triple-tuning a single circuit is impractical for narrow bores. Doty's use of multiple, orthogonal coils has greatly simplified the H/X/Y circuit and also results in lower heating of the coils and thus of the sample." Doty is currently developing a MAS probe with a 2 or 2.5 mm spinner diameter, with spinning speeds projected at 40 kHz or higher. Currently available spinners can reach speeds of 35 kHz; faster speeds provide better line narrowing and resolution. One complication resulting from higher spinning speeds is the creation of frictional heat by spinning rotors, which transfer heat to the samples themselves; Doty's design minimizes frictional heating.
Doty also has responded to the needs of the membrane protein research community by developing flat (rectangular) coil probes for the study of membrane protein structure in lipid bilayers. Samples are loaded onto glass plates, and the plates are stacked and inserted into the coil. The membrane proteins are thereby aligned perpendicular or parallel to the magnetic field, which allows high resolution NMR spectra to be obtained using higher power RF pulse sequences. Researchers themselves often build these probes, but Doty introduced the first commercially available flat-coil probe with improved RF pulse performance.
Recent developments in NMR are clearly paving the way for exciting new applications in solution- and solid-state research, and manufacturers are continuing to improve on NMR's high-throughput capabilities. And as higher field strength magnets are designed, new experimental methods are likely to follow. The cutting-edge techniques associated with NMR will be essential tools in the structural genomics era.
Aileen Constans can be contacted at email@example.com.
Technical expertise was provided for this article by Juliette Lecomte of The Pennsylvania State University.
1. T.L. James, "Fundamentals of NMR," in D. Gorenstein, ed., Nuclear Magnetic Resonance (NMR), Online Biophysics Textbook, Biophysical Society, 1999, biosci.umn.edu/biophys/OLTB/NMR.html.
2. J.P. Hornak, "The Basics of NMR," Online Textbook, 1997, www.cis.rit.edu/htbooks/nmr/inside.htm.
3. R. Service, "NMR researchers look to the next generation of machines," Science, 279:1127-8, 1998.
4. K. Pervushin et al., "Attenuated T2 relaxation by mutual cancellation of dipole-dipole coupling and chemical shift anisotropy indicates an avenue to NMR structures of very large biological macromolecules in solution," Proceedings of the National Academy of Sciences, 94:12366-71, 1997.
5. S.B. Shuker et al., "Discovering high-affinity ligands for proteins: SAR by NMR," Science, 274:1531-4, 1996.
6. J.P. Shockcor et al., "Combined HPLC, NMR spectroscopy, and ion-trap mass spectrometry with application to the detection and characterization of xenobiotic and endogenous metabolites in human urine," Analytical Chemistry, 68:4431-5, 1996.
7. P. Hajduk et al., "High-throughput nuclear magnetic resonance-based screening," Journal of Medicinal Chemistry, 42:2315-7, 1999.
8. A. Medek et al., "An approach for high-throughput structure determination of proteins by NMR spectroscopy," Journal of Biomolecular NMR, in press.
Do-It-Yourself: Homebuilt NMR Spectrometers and Probes Provide an Alternative for Demanding Applications
New products from the major NMR vendors have made the technique faster and more sensitive. But some experiments simply cannot be run on commercially available equipment. For example, many spectroscopy groups that study membrane proteins by solid state NMR make their own equipment because commercial instrumentation is not robust enough--or is too expensive--for their experimental needs. Such experiments are particularly demanding of probes. As Sean Burns, postdoctoral research associate in the laboratory of Stanley Opella, professor of chemistry at the University of Pennsylvania, explains, "Most solid state NMR experiments, if you run them properly, require putting enough power into the probe that it's either near or above the limit where you'll damage it. So, many solid state NMR labs routinely repair their own probes. They run things right at the technical limit of what the apparatus can handle, they break it, and then they start over."
This process can become prohibitively costly if commercial probes are used, so Burns and his colleagues at the Resource for Solid State NMR of Proteins build their own flat coil probes for the study of proteins in membrane bilayers. The Resource for Solid State NMR of Proteins also operates a Homebuilt NMR Mini-Site (cherry.chem.upenn. edu/resource/), which provides schematic diagrams for spectrometer design as well as links to various vendors of NMR equipment.
Another factor influencing the use of homebuilt equipment is speed. Commercial pulse programmers tend to be so sophisticated that they actually hinder the ability to run certain experiments. Kathy Valentine, facility director, notes that although these systems can run the necessary pulse sequences, "their pulse programming is complicated with hidden timing events such that it becomes difficult to do the exact pulse sequence you are coding." Burns adds that the rate at which radio frequency (RF) signals need to be sent to the probe approaches the rate at which components switch: "If there's a one microsecond gap in the RF pulse train that you're sending to the probe, that can cause big problems for our experiments."
Homebuilt NMR spectrometers thus allow researchers to run experiments that could not normally be run using conventional equipment. One such procedure is the PISEMA (polarization inversion spin exchange at the magic angle) experiment, which correlates a nitrogen chemical shift and a proton-nitrogen dipolar coupling, resulting in one crosspeak per amino acid residue.1 This experiment requires that frequency switches be performed in hundreds of nanoseconds. Though commercial spectrometers are able to run PISEMA experiments, this was not the case several years ago. "We've been doing that experiment for about eight years or so," says Valentine, "It's become accepted as a useful experiment, and the community is interested in using it, so now the commercial spectrometer [vendors] are responding to that need. So having had the homebuilt spectrometer back then allowed us to push forward the technique."
Which brings up an important point: As the market for a technique expands, commercial vendors can quickly equal or surpass the technical capabilities of do-it-yourself spectrometer builders. Valentine explains that the Resource for Solid-State NMR of Proteins currently purchases consoles from Tecmag of Houston, which can handle the RF requirements of their experiments and which can be purchased for about the same amount of money that it would take to build a similar spectrometer. But for cutting-edge research, homebuilt spectrometers are frequently the only option: "If you're working on something in a small market segment that commercial people aren't interested in ... you basically build your own stuff or you don't do the experiments," says Burns.
1. A. Ramamoorthy, S.J. Opella, "Two-dimensional chemical shift/heteronuclear dipolar coupling spectra obtained with polarization inversion spin exchange at the magic angle and magic-angle sample spinning (PISEMAMAS)," Solid State Nuclear Magnetic Resonance, 4:387-92, 1995.