The systematic difference between T-2 values obtained from CPMG and T(1)rho
experiments was observed for backbone N-15 nuclei of bacterial ribonucleas
e barnase. Theoretical consideration suggests that the observed difference
is caused by off-resonance effects of 180 degrees pulses of the CPMG pulse
train. Namely, at off-resonance conditions T-1-dependent secondary echo coh
erence pathways considerably contribute to the signal decay in the CPMG exp
eriment and result in systematic (up to 10%) offset-dependent overestimatio
n of N-15 T-2 measured by the CPMG technique. Under certain circumstances o
ff-resonance effects result in dependence of N-15 T-2 on CPMG frequency, wh
ich might be erroneously interpreted as conformational exchange on the mill
isecond time-scale. A procedure for numerical correction of (1)5N T-2 (CPMG
) data is proposed.