Abstract
Carbonyl ^13^C relaxation experiments to study protein backbone dynamics have recently been developed. However, the effect of three‐bond ^13^C′–^13^C′ couplings on transverse relaxation measurements appears not to have been considered, and the potential to detect and quantify motions on the millisecond to microsecond time scale has not been fully explored. The present paper addresses these two issues. Simulations and experiments show that scalar couplings between adjacent backbone carbonyl carbon nuclei and between backbone and side‐chain carbonyl/carboxyl carbon atoms in Asp and Asn residues interfere with the accurate determination of transverse relaxation rates by Carr–Purcell–Meiboom–Gill or on‐resonance spin‐lock measurements. The use of off‐resonance radio‐frequency fields avoids efficient cross‐polarization, and offers a route towards accurate R~1ρ~ measurements. In addition, this approach yields dispersion in the transverse relaxation rate as a function of the effective field when conformational exchange is present. In the case of calcium‐bound calbindin D~9k~, ^13^C′ off‐resonance R~1ρ~ measurements yielded uniform values of R~2~ along the polypeptide chain, indicating homogeneous chemical shift anisotropies and restricted dynamics on the picosecond to nanosecond time scale. Variation of R~2~ as a function of the effective spin‐lock field strength was not observed for any residue, indicating the absence of large‐scale conformational changes of the protein backbone in the millisecond to microsecond time window. The absence of relaxation induced by internal motions on these wide‐ranging time scales reinforces the view that calcium‐loaded calbindin D~9k~ is extremely rigid. In contrast, for the C‐terminal tryptic fragment of calmodulin containing the E140Q mutation we observed widespread exchange broadening. From the carbonyl transverse relaxation dispersion profile of Asp129 the exchange rate was determined to be 28 000 s^−1^. Copyright © 2003 John Wiley & Sons, Ltd.