Modeling magnetization transfer using a three-pool model and physically meaningful constraints on the fitting parameters.

A model for water-macromolecular magnetization transfer is presented which addresses the mechanism of coupling between the hydrogen populations and the extraction of physically meaningful parameters from experimental magnetization transfer data. Both physical exchange between bulk-solvent and site-specific hydration-layer hydrogens and intermolecular magnetic dipolar coupling between these specific hydration-layer-solvent and macromolecular hydrogens are explicitly included, leading to a three-pool model for magnetization transfer. It is shown that the three-pool model is well approximated by a two-pool model for coupling between the bulk-solvent and macromolecular hydrogens when the dipolar-coupled solvent hydrogens are a small fraction of the total solvent, and the solvent-macromolecular coupling constant includes both dipolar magnetic, kappa(dip), and physical exchange, kappa(ex), coupling rates. The model is also extended to multiple solvent systems. The model results in a set of coupled equations that predict magnetization transfer spectra as a function of temperature and composition. Physically meaningful constraints on the coupling and relaxation parameters are established for systems in which magnetization transfer has been observed including solvated cross-linked proteins and lipid bilayers. Using parameter estimates based on these constraints, empirical magnetization transfer spectra are well predicted by the model. It is found that the degree of magnetization transfer becomes independent of kappa(dip) and kappa(ex) when these parameters become greater than about 50 s(-1). In the semi-rigid cross-linked protein systems where the mobility of the macromolecular matrix is insensitive to temperature, the magnitude of the observed magnetization transfer is consistent with being limited by the intermolecular dipolar coupling and spin-lattice relaxation in the bulk-solvent phase.