Amide proton spin-lattice relaxation in polypeptides. A field-dependence study of the proton and nitrogen dipolar...

The proton nuclear magnetic resonance (NMR) spin-lattice relaxation of all six amides of deferriferrichrome and of various alumichromes dissolved in hexadeutero-dimethylsulfoxide have been investigated at 100, 220, and 360 MHz. We find that, depending on the type of residue (glycyl or ornithyl), the amide proton relaxation rates are rather uniform in the metal-free cyclohexapeptide. In contrast, the (1)H spinlattice relaxation times (T(1)'s) are distinct in the Al(3+)-coordination derivative. Similar patterns are observed in a number of isomorphic alumichrome homologues that differ in single-site residue substitutions, indicating that the spin-lattice relaxation rate is mainly determined by dipole-dipole interactions within a rigid molecular framework rather than by the specific primary structures. Analysis of the data in terms of (1)H-(1)H distances (r) calculated from X-ray coordinates yields a satisfactory linear fit between T(1) (-1) and Sigmar(-6) at the three magnetic fields. Considering the very sensitive r-dependence of T(1), the agreement gives confidence, at a quantitative level, both on the fitness of the crystallographic model to represent the alumichromes' solution conformation and on the validity of assuming isotropic rotational motion for the globular metallopeptides. An extra contribution to the amide proton T(1) (-1) is proposed to mainly originate from the (1)H-(14)N dipolar interaction: this was supported by comparison with measurements on an (15)N-enriched peptide. The nitrogen dipolar contribution to the peptide proton relaxation is discussed in the context of {(1)H}-(1)H nuclear Overhauser enhancement (NOE) studies because, especially at high fields, it can be dominant in determining the amide proton relaxation rates and hence result in a decreased effectiveness for the (1)H-(1)H dipolar mechanism to cause NOE's. From the slope and intersect values of T(1) (-1) vs. Sigmar(-6) linear plots, a number of independent estimates of tau(r), the rotational correlation time, were derived. These and the field-dependence of the T(1)'s yield a best estimate approximately 0.37 ns, in good agreement with 0.38 ns [unk] [unk] 0.41 ns, previously determined from (13)C and (15)N spin-lattice relaxation data.