Dynamical deductions from nuclear magnetic resonance relaxation measurements at the water-protein interface

R.G. Bryant, W.M. Shirley
1980 Biophysical Journal  
Nuclear magnetic resonance (NMR) measurements provide both structural and dynamical information about the molecules in which nuclear resonances are observed. This manuscript addresses NMR relaxation of water protons in protein powder systems. Inclusion of magnetic communication between the water proton spins and protein proton spins leads to a clearer view of water molecule dynamics at the protein surface than has been previously available. We conclude that water molecule motion at the protein
more » ... urface is somewhat slower than in the solute free solvent, but it is orders of magnitude faster than motions in a rigid ice lattice even in samples hydrated to levels well below what is generally thought to be the full hydration complement of the protein. The NMR relaxation data on lysozyme powders support a model that leaves adsorbed water very fluid at the protein surface with reorientational correlation times for the water shorter than nanoseconds. An understanding of water-protein interactions is crucial to a detailed understanding of protein structure and catalysis. An important aspect of this interaction involves the solvent motion both in semisolid systems, such as tissues, and in the region immediately adjacent to a solute particle in a solution of a macromolecule. Our concepts of structure, however, draw most heavily from models that are based on crystalline low molecular weight solids. This basically sound strategy has been recently used to address structural aspects of water-peptide interactions (1). The structural information obtained from such systems is easy to visualize because the structure is static, usually geometrically simple, often highly symmetrical, and asthetically pleasing. There is therefore a great temptation to use the language and pictures associated with truly solid structures to describe structures in liquids or in liquids associated with solids. Some time ago Klotz (2) extended such an idea first proposed by Frank and Evans (3), who suggested that water adjacent to a protein be viewed as an "ice-like lattice." This approach conveys a structural picture, to be sure, but, as will be shown, errs by many orders of magnitude in implying the time scale for describing the motion of the oxygen atoms in the water under consideration. The consequences of such sometimes useful analogies involve semantic as well as conceptual problems that may be resolved to some extent by using the somewhat more cumbersome concepts of liquid structure characterization. NMR relaxation measurements provide dynamical information that is readily related to such descriptions. The underlying ideas that have led to extensive applications of NMR relaxation to the study of surface systems (4-6) are apparent in Eq. 1. BIOPHYS J. e Biophysical Society * 0006-3495/80/10/003/14 $1.00 3 BRYANT: I appreciate Dr. Resing's comments, as he has made substantial contributions to our understanding of NMR relaxation at surfaces. There is no question that the analysis of sums of exponentials may be difficult. While our present spectrometer, which is better than the one used to collect the data on which this discussion is based, gives us precision of a per cent in amplitude vs time, the errors in Tls or TIF are easily several percent. I do not believe we may presently measure T, w to within an error of 20%. Certainly Fig. 4 is not any better than that, and there we may have underestimated RI w or overestimated RIW-'. Now to answer Dr. Resing's question about a failure of the equation RI w = PI -' [RI., -(1 -Pw)RIp], which is predicted by RIT >> RlW = Rp" since we find clearly that RT >> RlW >> RIP, there is a rapid mixing of magnetization and spin lattice rates. It is possible then to write, RIS = PWRIw + PpRIp -PWRIw + (1 -Pw)RI which when solved for Rlw gives Dr. Resing's equation. The difficulty is now to define Pw and Pp in terms of accurately known quantities. If the spin systems may be assumed to be in equilibrium within themselves, then the Pi are simply ratios of populations or concentrations of protons in each system. As motions slow down at low temperature, the distinction between spin systems may become difficult to make and the model will fail. Nevertheless, it is interesting and informative to approach this question from several limiting cases. If we assume RT iS very large 'L. B. McCuster and K. Seff. 1978. Private communication.
doi:10.1016/s0006-3495(80)84912-5 pmid:7248450 pmcid:PMC1327249 fatcat:7mxofj5xzna5beoftojemcwwfe