Factors determining the reliable description of global tumbling parameters in solution NMR

An accurate description of global tumbling of a protein is essential for correct analysis and interpretation of internal dynamics and thermodynamics. The accurate fitting of global tumbling parameters is affected by the number of experimental relaxation data points available for analysis, the distribution of data points over the domain of the function describing the tumbling, the measurement error associated with the data, the error associated with use of an approximate functional form, and errors in the protein structure. We present an analysis of the influence of these factors on the error in global tumbling parameters and the corresponding error in the calculated T(1)/T(2) values. We find that reduction of experimental and approximation error can compensate for a less-than-ideal quantity or distribution of data points, and that accurate parameters can be obtained for proteins with highly anisotropic distributions of bond vectors, as illustrated using the helical bundle protein G-CSF. This indicates that proteins with anisotropic distributions, such as the helical bundle class of proteins, should not summarily be excluded when selecting proteins for dynamic and thermodynamic analyses of (15)N backbone relaxation measurements.