Biomolecular processes such as folding, binding, catalysis, and allosteric site modulation are made possible through protein dynamics, which is characterized by interconversions among different protein conformations. At the lowest energy level is the ground-state structure, and at the highest energy level is the transient structure(s). The energy difference ΔG determines the population of each conformational state, while the activation energy determines the interconversion rate kex (sum of forward and reverse kinetic rates). X-ray crystallography usually characterizes the ground-state structure or conformation that is enriched during crystallization, and classic nuclear magnetic resonance (NMR) techniques determine the structure of predominant species in solution. Our group develops paramagnetic NMR techniques and characterizes lowly populated protein transient structures; in conjunction with other biophysical and biochemical methods, we seek to understand the relationship between protein dynamics and function.
Paramagnetic Relaxation Enhancement (PRE)
Biomolecular NMR often focuses on diamagnetic systems, as paramagnetic contamination deteriorates the spectra. In my laboratory, we deliberately introduce a paramagnetic probe at a specific site on the protein surface (often via a cysteine residue). The paramagnetic probe causes relaxation enhancement (hence the term PRE) to nuclei up to 40 Å away, with an <r–6> distance dependence. Placing the paramagnetic probe strategically, we can detect and directly visualize protein transient structures that are otherwise invisible. For example, we characterized protein–protein encounter complexes between two signaling proteins in a bacterial phosphotransferase system. The encounter complexes facilitate the formation of stereospecific complexes and had previously only been inferred through theoretical simulation or mutagenesis analysis.
When the interconversion rate between different protein conformations is extremely fast, the observed PRE value approaches the population-weighted average over PREs in all states. However, when the dynamic fluctuation is not so fast (i.e., the µs–ms regimen, in which a lot of biomolecular events take place), the observed PRE falls between the value for the ground-state structure and the value for the population-weighted average. We devised a scheme to determine the exact PRE value for each conformation and the associated exchange timescale by introducing multiple probes at the same conjugation site. Depending on the interconversion rate, the ratio between the observed PREs from different probes can be smaller than the ratio between their intrinsic paramagnetic strengths. As PRE affords redundant measurements for residues participating in the same dynamic process, we can then extrapolate the exchange timescale and achieve a spatiotemporal characterization of such interconversions. We call this improved version of PRE differentially scaled PRE, or DiSPRE. Applying the DiSPRE, we captured an alternative complex between two interacting proteins, which occurs ~4 percent of the time and interconverts with the specific complex at ~1,100 s–1 (Figure 1).
A paramagnetic probe is conjugated to protein via multiple rotatable bonds and takes up a quite large conformational space. As a result, protein dynamics on a spatial scale smaller than the distribution of the paramagnetic center cannot be seen. We are designing rigid paramagnetic probes that afford a more fixed observation point. When a paramagnetic probe is introduced, other structural information such as residual dipolar coupling and pseudocontact shift can also be obtained. In conjunction with PRE, we are utilizing paramagnetic NMR to better depict protein structure and dynamics.
Transient Protein–Protein Interactions
NMR is uniquely suited to characterize transient protein–protein interactions. These transient complexes are often too weak to be crystallized, or multiple modes of interactions may exist, and only one conformation is enriched in the crystallization process. While developing and improving the paramagnetic NMR technique, we successfully applied it to characterize the structure and dynamics of a number of ultraweak complexes in solution.
HIV-1 protease is expressed as part of the Gag-Pol precursor and possesses very low activity. The protease has to be activated before it can release other proteins (matrix, capsid, nucleocapsid, integrase, and reverse transcriptase) from the precursor. We found that the protease precursor, though mostly monomeric in solution, transiently dimerizes with a KD of 3–6 mM. Refined against intermolecular PREs, we obtained an ensemble structure of the transient protease dimer. The two subunits adopt a variety of relative orientations, and a subset of "correct" conformation would allow transient insertion and processing of its N-terminal tail, leading to self-activation.
Ubiquitin is a model system widely used in biomolecular NMR. Using PRE NMR and analytical ultracentrifugation, we found that ubiquitin noncovalently dimerizes with a KD of ~5 mM, with an interface overlapping that for binding to its partner proteins. The two subunits adopt an array of relative orientations in the dimer (Figure 2); the dynamic fluctuation at the quaternary level thus complements its tertiary dynamics, which may regulate the accessibility of ubiquitin binding partners.
Functional Excursions of Glutamate Receptor Ligand-Binding Domain
Ionotropic glutamate receptors (iGluRs) play a central role in excitatory neurotransmissions in the brain. Upon binding to a glutamate or pharmacological ligand, an iGluR shuttles between resting, open, and desensitized states, macroscopically characterized by an inward current lasting for several milliseconds to several hundred milliseconds. As the gating module of iGluR, the extracellular ligand-binding domain (LBD) functions as a dimer with a KD in the millimolar range. The crystal structure of the LBD dimer has been obtained through mutation, cross-linking, and stabilization with small molecules, yet it is not known how the dimer forms under native conditions. In addition, for such a transient ultraweak dimer, it is likely that it samples an ensemble of conformations in solution. Intrigued, we are studying the structure and dynamics
The effort to visualize protein transient structures is in part supported by the National Natural Science Foundation of China and the Human Frontier Science Program.
As of January 17, 2012