What Is Nuclear Magnetic Resonance Spectroscopy of Proteins? A Guide

Nuclear magnetic resonance spectroscopy of proteins (protein NMR) is a technique that uses the magnetic properties of atomic nuclei to determine the three-dimensional structure and dynamic behavior of proteins in solution. Instead of producing a direct image like X-ray crystallography, NMR generates distance constraints between atoms, which are then used to calculate a model of the protein. It is one of the three main methods for determining protein structures at atomic resolution, alongside X-ray crystallography and cryo-electron microscopy.

How NMR Spectroscopy Works

NMR relies on the fact that certain atomic nuclei, such as hydrogen-1 (protons), carbon-13, and nitrogen-15, have a quantum property called spin. When placed in a strong, static magnetic field, these nuclei align either with or against the field. A radio-frequency pulse disturbs this alignment, and as the nuclei relax back to equilibrium, they emit radio signals at specific frequencies. The exact frequency absorbed or emitted by a nucleus depends on its local chemical environment—nearby atoms, bonds, and the three-dimensional arrangement of the molecule.

In a protein, each hydrogen atom sits in a slightly different chemical environment, so each proton gives a slightly different signal. The collection of these signals forms a spectrum, and the position of each peak (its chemical shift) reports on the local electronic surroundings. Nuclei that are close in space—within about 5 angstroms—can transfer magnetization to each other through the nuclear Overhauser effect (NOE). The strength of this NOE signal is directly related to the distance between the two nuclei. By collecting many NOE distance restraints, researchers can build a map of which atoms are near each other, and then use computational methods to fold that map into a three-dimensional structure.

Sample Preparation for Protein NMR

Preparing a protein sample for NMR is a critical and often time-consuming step. The protein must be highly pure and concentrated, typically between 0.1 and 3 millimolar, in a volume of 300 to 600 microliters Wikipedia. The sample is placed in a thin-walled glass tube and inserted into the magnet.

Most proteins studied by NMR are produced using recombinant DNA technology. The gene for the protein is inserted into bacteria (often E. coli), yeast, or insect cells, which then produce large quantities of the protein. This method allows researchers to control the isotopic composition of the protein. Natural proteins contain mostly carbon-12 and nitrogen-14, which are NMR-inactive or give poor signals. By growing the expression host in media enriched with carbon-13 and nitrogen-15, the resulting protein becomes “labeled” and its NMR signals become much easier to detect and assign. This isotopic labeling is essential for proteins larger than about 10 kilodaltons.

The protein is then purified, usually by chromatography, and dissolved in a buffer solution. The buffer often contains a small amount of deuterium oxide (heavy water) to provide a lock signal for the spectrometer and to reduce the strong water signal that would otherwise overwhelm the protein signals.

Data Collection and Spectral Assignment

Once the sample is ready, it is placed in the bore of a superconducting magnet. Modern NMR spectrometers use magnets with field strengths of 500 to 800 MHz (proton resonance frequency), with some instruments reaching 1 GHz or higher. Higher field strength improves sensitivity and resolution, making it possible to study larger or more complex proteins.

Data collection involves running a series of two-dimensional (2D) and three-dimensional (3D) NMR experiments. A simple 1D proton spectrum of a protein is too crowded to interpret—hundreds of overlapping peaks. Instead, researchers use multidimensional experiments that correlate the signals of different nuclei. For example, a 2D COSY (correlation spectroscopy) experiment shows which protons are bonded to each other through chemical bonds. A 2D NOESY (nuclear Overhauser effect spectroscopy) experiment shows which protons are close in space, regardless of whether they are bonded.

For larger proteins, 3D experiments are used to spread the signals into a third dimension, reducing overlap. Common 3D experiments include HNCO, HNCA, and CBCA(CO)NH, which correlate the backbone amide proton and nitrogen with the carbon atoms of the same or preceding residue. These experiments rely on the isotopic labeling mentioned earlier.

The first major analytical step is resonance assignment: linking each NMR signal to a specific atom in the protein sequence. This is done by following the through-bond correlations in a systematic way, residue by residue, along the protein backbone. Once the backbone is assigned, side-chain atoms are assigned using additional experiments. This process can take weeks to months for a medium-sized protein.

Structure Calculation from NMR Data

With the assignments in hand, the next step is to extract distance restraints from NOESY spectra. For each pair of protons that show an NOE cross-peak, the intensity of the peak is converted into an upper-bound distance, usually between 2 and 6 angstroms. A typical structure determination uses hundreds to thousands of such distance restraints.

Additional restraints come from other NMR parameters. Dihedral angle restraints can be derived from chemical shift data or from coupling constants (J-couplings), which report on the angles between bonds. Residual dipolar couplings (RDCs) provide information about the orientation of bond vectors relative to the magnetic field, which helps define the overall shape of the molecule.

All these restraints are fed into a structure calculation program, such as CYANA, XPLOR-NIH, or ARIA. The program starts with a random, extended chain and then applies the restraints as “springs” that pull atoms toward each other. The calculation also includes a physical force field that enforces proper bond lengths, angles, and van der Waals contacts. The result is an ensemble of structures that all satisfy the experimental data. The quality of the structure is judged by how well the ensemble converges (root-mean-square deviation, RMSD) and by how many restraints are violated. The final models are deposited in the Protein Data Bank Protein Data Bank.

Studying Protein Dynamics and Interactions

One unique strength of NMR is its ability to study protein dynamics over a wide range of timescales. While X-ray crystallography gives a static snapshot, NMR can reveal how a protein moves in solution. Fast motions (picosecond to nanosecond) of side chains and loops can be measured using relaxation experiments. Slower motions (microsecond to millisecond), which are often important for enzyme catalysis or allosteric regulation, can be detected using techniques like CPMG relaxation dispersion or chemical exchange saturation transfer (CEST).

NMR is also a powerful tool for studying protein-protein and protein-ligand interactions. When two molecules bind, the NMR signals of atoms at the binding interface change their chemical shifts, intensities, or line widths. By titrating one molecule into the other and monitoring these changes, researchers can map the binding site, measure the binding affinity (Kd), and even determine the structure of the complex. This is particularly useful in drug discovery, where small molecules are screened for their ability to bind to a target protein National Center for Biotechnology Information.

Limitations and Complementary Methods

NMR has practical limitations. The most significant is the size of the protein that can be studied. Traditional solution NMR is limited to proteins smaller than about 40 kilodaltons, because larger proteins tumble slowly in solution, causing broad and weak signals. Methods like transverse relaxation-optimized spectroscopy (TROSY) and deuteration (replacing protons with deuterium) have pushed this limit to about 100 kilodaltons for some systems. Solid-state NMR can study much larger proteins and even insoluble aggregates like amyloid fibrils, but it requires different sample preparation and experiments.

Another limitation is the time required for data collection and analysis. A full structure determination can take months, compared to days or weeks for a well-diffracting crystal. NMR also requires relatively high protein concentrations, which can be difficult to achieve for membrane proteins or proteins that aggregate.

For these reasons, NMR is often used in combination with other methods. NMR can provide dynamic information that complements the static structures from X-ray crystallography. Cryo-electron microscopy can give the overall shape of large complexes, while NMR fills in the atomic-level details of flexible regions.

Key Takeaways

• Nuclear magnetic resonance spectroscopy of proteins determines three-dimensional structures and dynamics by measuring distance and angle restraints between atoms in solution.
• Samples require high purity and concentration, with isotopic labeling (carbon-13, nitrogen-15) being essential for larger proteins.
• Data collection uses multidimensional NMR experiments to assign signals to specific atoms and to extract distance restraints from NOE cross-peaks.
• Structure calculation converts these restraints into an ensemble of models that satisfy the experimental data.
• NMR is uniquely suited for studying protein dynamics and molecular interactions, including drug binding.
• The main limitation is protein size (typically under 40 kilodaltons for solution NMR), but complementary methods like X-ray crystallography and cryo-EM can fill the gaps.