Electron spin-lattice relaxation in two heme iron and two blue-copper proteins at liquid helium temperatures

Abstract

The relaxation rates in frozen aqueous solutions of whale ferri-myoglobin azide, bovine ferri-hemoglobin azide, cupric azurin (P. aeruginosa) and cupric spinach plastocyanin were measured at 9.5 GHz using the pulse-saturation recovery method. Measurements covered a temperature range of 1.4 K to as high as 22 K, with corresponding relaxation rates up to 10$\sp5$/sec. Improvements in the equipment and the methods of analysis have enabled more stringent tests of the temperature dependence of the rates. In particular, several models proposed in the literature to explain the anomalous temperature dependence of the Raman rates in proteins are shown to be insufficient, including two fractal models. In addition, it is shown that any model based exclusively on the protein structure fails due to the diversity of the data under various solvent conditions. A general functional form consistent with a crossover in the vibrational properties is proposed instead, similar to the localization crossover in amorphous materials.

The effect on the relaxation rate of several cosolvents and solutes is also examined. The effect on the direct process is much more pronounced than on the Raman region. The differences are shown to be consistent with changes in the velocity of sound at room temperature caused by the addition of cosolvents and solutes.

Finally, the EPR recovery form is analyzed. We propose that the deviations in the recovery from an exponential form are due to a distribution of relaxation rates. The source of the distribution is most likely sample heating in the lower temperatures and a distribution of conformations frozen in near the paramagnetic site in the higher temperatures. It is not likely that it is caused by spin-spin interactions. The exact form of the distribution is unclear, but the most successful functional form for the recoveries is a stretched exponential with an exponent ranging from 0.5 to 1.0. However, a simple exponential fit to a limited portion of the recovery was the experimentally most viable and consistent method of extracting a rate.