Separating the contributions to 15N transverse relaxation in a fibronectin type III domain

In proteins, dynamic mobility is an important feature of structure, stability, and biomolecular recognition. Uniquely sensitive to motion throughout the milli- to picosecond range, rates of transverse relaxation, R2, are commonly obtained for the characterization of chemical exchange, and the construction of motional models that attempt to separate overall and internal mobility. We have performed an in-depth study of transverse relaxation rates of backbone 15N nuclei in TNfn3(1-90), the third fibronectin type III domain from human tenascin. By combining the results of spin-echo (CPMG) and off-resonance T1 rho experiments, we present R2 rates at effective field strengths of 2 to 40 krad/s, obtaining a full spectrum of 16 independent R2 data points for most residues. Collecting such a large number of replicate measurements provides insight into intrinsic uncertainties. The median standard deviation in R2 for non-exchanging residues is 0.31, indicating that isolated measurements may not be sufficiently accurate for a precise interpretation of motional models. Chemical exchange events on a timescale of 570 microseconds were observed in a cluster of residues at the C terminus. Rates of exchange for five other residues were faster than the sampled range of frequencies and could not be determined. Averaged 'exchange free' transverse relaxation rates, R2(0), were used to calculate the diffusion tensor for rotational motion. Despite a highly asymmetric moment of inertia, the narrow angular dispersion of N-H vectors within the beta sandwich proves insufficient to define deviations from isotropic rotation. Loop residues provide exclusive evidence for axially symmetric diffusion (Dpar/Dper = 1.55).