PULSED DOUBLE-QUANTUM COHERENCE ELECTRON PARAMAGNETIC RESONANCE IN PROTEIN STRUCTURE DETERMINATION

Abstract

Electron paramagnetic resonance (EPR) or more specifically, pulsed dipolar EPR spectroscopy (PDS), combined with the site-directed spin labelling (SDSL) technique has emerged as a key technique in protein structure determination. The core concept is to filter out the weak dipolar interaction between a pair of spin labels by applying an appropriate pulse sequence and retrieve the inter-spin distance from the dipolar EPR signal. Double-quantum coherence (DQC) and double electron-electron resonance (DEER) are two such methods primarily used in studying the structure of proteins and other biomacromolecules. There are two main classes of spin labels used in PDS studies, (i) triarylmethyl (TAM) and (ii) nitroxides. DQC signal expression of nitroxide spin labels is extremely complex and without knowing the analytic form of the signal, the resulting spectra, especially in 2D, cannot be analyzed both accurately and efficiently. In the first part of the thesis, we derive analytic expressions of DQC signals for both TAM and nitroxide spin labels. These expressions are extremely useful in analyzing experimental signals using personal computers. Hence, we believe that this innovation is an important and necessary step in motivating the scientific community to use DQC more frequently in their studies. Another key challenge in PDS signal processing is the removal of intermolecular or background signal. An error in the process of background signal removal can translate into a critical error in obtaining the distance distribution. We have derived an analytic expression of the total DQC signal for spin, S=1/2, particles in frozen samples and this expression can be integrated over the spatial variables to derive the functional form of the signal. We have demonstrated the importance of the analytic expression in studying the spatial distribution of the spin-labeled proteins in frozen samples. In the last chapter, we present experimental studies that demonstrate the effect of the rate of freezing on the distance distributions derived from DEER experiments. In the same project, we have explored the effect of varying the amount of cryoprotectant and using different spin labels on the reconstructed distance distributions. We conclude that both slow freezing (>= 1 s) at 30% glycerol by weight and rapid freeze-quench (100 micro-s) at 10% glycerol result into reduced intermolecular spin-spin interactions and improved signal-to-noise ratio (snr). Additionally, we find that the effect of the conformational sub-states of the spin-labels on reconstructed distance distributions is averaged out in slow freezing, while the trapping of the conformational sub-states in rapid-quenched samples yields broadened distance distributions.