Revisiting ¹H^N CPMG relaxation dispersion experiments: a simple modification can eliminate large artifacts

Carr-Purcell-Meiboom-Gill relaxation dispersion experiments are commonly used to probe biomolecular dynamics on the millisecond timescale. The simplest experiment involves using backbone ¹⁵N spins as probes of motion and pulse sequences are now available for providing accurate dispersion profiles in this case. In contrast, ¹H-based experiments recorded on fully protonated samples are less common because of difficulties associated with homonuclear scalar couplings that can result in transfer of magnetization between coupled spins, leading to significant artifacts. Herein we examine a version of the ¹H^N CPMG experiment that has been used in our laboratory where a pair of CPMG pulse trains comprising non-selective, high power ¹H refocusing pulses sandwich an amide selective pulse that serves to refocus scalar-coupled evolution by the end of the train. The origin of the artifacts in our original scheme is explained and a new, significantly improved sequence is presented. The utility of the new experiment is demonstrated by obtaining flat ¹H^N dispersion profiles in a protonated protein system that is not expected to undergo millisecond timescale dynamics, and subsequently by measuring profiles on a cavity mutant of T4 lysozyme that exchanges between a pair of distinct states, establishing that high quality data can be generated even for fully protonated samples.