In the study of protein backbone dynamics by 15 N relaxation global correlation time of the molecule being determined measurements, an initial estimation of the isotropic global correlafrom the T 1 /T 2 ratio (1, 2). tion time, t m , is usually obtained from the average T 1 /T 2 ratio of Nuclear spins with an effective internal correlation time, nuclear spins that do not exhibit slow internal motion and with T 2 t e , much faster than the global correlation time of the molevalues not significantly shortened by chemical or conformational cule have a spectral density function, as adopted by the exchange processes. Different methods have been used for identimodel-free formalism ( 14), of fication of the rates of internal motion. However, the number of nuclear spins included in the t m estimation is often larger than
[1]
the number that ultimately can be fitted to a single-order parameter, S 2 , implying that some nuclear spins involved in the initial t m estimation actually have an effective internal correlation time, where S 2 is the order parameter, reflecting the extent of t e , not as fast as assumed. As a consequence, t m is underestimated, spatial restriction of the NH bond, and t m is the global since internal motion reduces the T 1 /T 2 ratio. This situation becorrelation time of the molecule. Using Eq. [1], t m can be comes more obvious if the molecule has a large t m value because determined from the average T 1 /T 2 ratio, provided that nuthe reduction in T 1 /T 2 ratio arising from internal motion is more clear spins with slow internal motion or with T 2 values that significant than for molecules with smaller t m and the same degree are significantly shortened due to chemical or conformaof internal motion. This Communication describes a more reliable tional exchange are excluded (1). As internal motion results method for identifying nuclear spins which should be excluded in a reduction in both the T 1 /T 2 ratio and the NOE (see from the t m estimation because of insufficiently rapid internal below), whereas a significant contribution from exchange motion. This results in an improved t m value, giving a much better processes leads to an increase in the T 1 /T 2 ratio, alternative agreement between the number of nuclear spins fitted successfully to a single-order parameter, S 2 , and those used in the t m estima-methods have been used to identify nuclear spins for the tion.