UC San Diego

Atomic-resolution membrane protein structures can be determined by solid-state Nuclear Magnetic Resonance (NMR) spectroscopy, and the unique advantage of the approach is that membrane proteins reside in near-native lipid bilayer environment at physiological pH and temperature, which minimizes the potential distortions of the protein structure caused by the environment. Here, the full-length mercury transporter protein, MerF, is the focus of the structural studies, and the protein is an essential part of the bacterial mercury detoxification system that has been exploited as a potential engineering target for mercury bioremediation strategies. The backbone structures of the full-length MerF are determined in two environments, (i) magnetically aligned bicelles by oriented-sample (OS) solid-state NMR and (ii) proteoliposome by rotationally aligned (RA) solid-state NMR; and notably, both environments provide the planar lipid bilayer environment for the protein. The structural study of MerF in aligned bicelle has initially been challenging for the OS solid- state NMR, and consequently, methods have been developed to tackle the two major obstacles, the spectral resolution and resonance assignments. New pulse sequence, MSHOT-Pi4/ Pi, has demonstrated a reduction of the 1H resonance line width by more than a factor of two, a significant improvement in spectral resolution. New resonance assignment method, Dipolar Coupling Correlated Isotropic Chemical Shift (DCCICS) Analysis, has been developed that is able to transfer resonance assignment from isotropic NMR methods to OS solid-state NMR spectra. The combined usage of several resonance assignment strategies and special tactics, such as applying DCCICS to the new high- resolution proton-evolved local field experiments for terminal and loop residues, has resulted in the complete assignment of all backbone immobile residues of the full- length MerF protein in magnetically aligned bicelle. Meanwhile, RA solid-state NMR is developed in the lab as a new method that combines the strength of magic-angle- spinning (MAS) solid-state NMR in obtaining resonance assignment and the concept of molecular alignment from OS solid-state NMR in obtaining angular restraints. In applying to the structural study of MerF, the method is further incorporated with multi-contact cross polarization and sequential backbone "walk" with three three- dimensional experiments, and the first structure of full- length MerF is determined with the method. In comparison to the previously determined structure of the truncated MerF (MerFt), the full-length structure reveals that the protein truncation has caused large conformational rearrangement at a place more than ten residues away from the truncation site, which serves as an example to demonstrate the importance of studying the full-length unmodified proteins by structural biologists. Additionally, the structure reveals that both mercury-binding sites are located at the intracellular side of the membrane, hinting at the observation of a conformation that allows intramolecular transfer of mercury ions. Subsequently after the complete assignment of MerF in OS solid-state NMR, the MerF structure determined by RA solid-state NMR is further improved by incorporating additional angular restraints from OS solid-state NMR and by the new treatment of dihedral restraints derived from the experimental study of C-terminal dynamics. Lastly, as a side project, the theoretical foundation of MSHOT-Pi4 pulse sequence is further explored. The observation that the pulse sequence selectively improves the resolution of membrane protein samples but not of standard single crystal sample has been analytically generalized as the principle of "motion-adapted" pulse sequence, where it is found that the interference between sample's spatial rotational motion and the radio-frequency pulse rotation in the quantum spin space is the cause of the selectivity. As a related endeavor, the mechanisms of dilute spin exchange and the magic-angle ¹H spin-lock pulses have been analyzed theoretically and demonstrated in standard and biological samples. Mixed-order proton-relay mechanism is proposed to be the main contributor to dilute spin exchange in stationary aligned sample, and once more, the difference of pulse performance between standard and biological samples is observed that may be a consequence of several causes including sample motion. In conclusion, the development of various methods in OS and RA solid- state NMR are likely to find their usage in future structural studies of membrane proteins; the theoretical principle of motion-adapted property opens up new avenue to develop pulse sequences for membrane protein samples; and the atomic-resolution backbone structures of MerF contribute information for structural biologist and for the mechanistic study of mercury transportation