The electron spin resonance (ESR) spectra of spin-labeled proteins are sensitive to dynamics, but discrimination between the various dynamics is often difficult. Here, we report an improvement in ESP. spectral sensitivity to local backbone dynamics of a protein by a methodology that performs ESR measurement when the protein is confined in the nanochannels of a mesoporous material. An extensive set Of ESR data, which includes the spectra of a nitroxide-based side chain from buried and solvent-exposed sites of a T4 lysozyme (T4L) protein, were obtained over a range of temperatures, 200-300 K, to explore the dynamics of T4L under nanoconfinement. Spectra were simulated by perform, ing theoretical fits to the data using the microscopic ordering with macroscopic disordering model. Two principle dynamic modes, which differ in mobility and ordering,, are required to account for the spectra at temperatures >240 K. We show that the mobile one correlates only with the local backbone dynamics of bhried sites, whereas the other reflects the difference in local hydration dynamics between the labeling sites in T4L. The assignment of the mobile, component is supported by the X-ray crystallography data of T4L. Collectively, this-study has, demonstrated the validity of such a methodology for improving ESR sensitivity to buried sites in,a protein.