Modifications of the rate matrix required for the quantitative analysis of NOESY spectra of proteins

The introduction of rate matrix analysis to protocols for the refinement of solution structures of biopolymers on the basis of NMR-derived structural constraints has greatly enhanced the self-consistency of the general approach. However, current implementations of this strategy appear not to consider several issues arising from various commonly employed experimental conditions and sample characteristics which can prevent the quantitative interpretation of NOESY spectra of proteins. Here, a number of these effects are considered, including the influence of nonequilibrium populations generated by over-pulsing and solvent presaturation, the presence of heteronuclei, and the equilibration of exchangeable sites with deuterium in solvent. The modifications of the rate matrix that are required to treat these effects are summarized and simulations presented to illustrate the relative importance of each. The computational effort required to rigorously accommodate each effect is described along with several simple experimental modifications that reduce the computational burden.