Skip to main content

Characterization of membrane protein dynamics by hydrogen/deuterium exchange and time-resolved infrared spectroscopy, assisted by maximum entropy and Bayesian methods of analysis

Periodic Report Summary 2 - NAGOYA2BCN (Characterization of membrane protein dynamics by hydrogen/deuterium exchange and time-resolved infrared spectroscopy)

There have been scarce and controversial reports about to which level and in which time-scale membrane protein structure fluctuates. To clarify this question we studied the kinetics of hydrogen-deuterium exchange (HDX) of bacteriorhodopsin (bR), a membrane protein highly compact and stable. The HDX of side chain labile protons is limited by the kinetics of D2O access to the side chain vicinity, providing a way to test bulk water access to the protein interior. We first focused in the development of an HDX experimental setup with an improved time-resolution for H2O to D2O buffer exchange, reducing the deadtime to <5 s. During the second period we studied the HDX of bR at room-temperature and several pH values. All eleven tyrosines side chains of bR appeared to exchange with a time constant of 2-6 s, implying that bulk water can readily access most of the bR interior in (sub)seconds. This observation shows that bR suffers from relatively large fluctuations in its native tertiary structure. Similar HDX experiments where performed on the melibiose transporter (MelB) with similar results and conclusions.

In collaboration with Prof. Kandori in Nagoya (Japan) we studied structural fluctuations in bovine rhodopsin (Rho) by HDX. The HDX of Thr118, a residue in steric contact with the light-sensing retinal molecule, was selectively measured by IR difference spectroscopy. Using Bayesian inference characterized the dynamics and thermodynamics of a fluctuation that affects Thr118, which were found to correlate with previous estimates of the frequency and activation energy for the dark-activation of Rho. This suggests that this fluctuation may be involved in the false perception of photons in total darkness.

There is still no definitive agreement in the literature about the number of intermediate states formed in the proton pump cycle (photocycle) of bR. This number could be objectively obtained from the number of exponential components in time-resolved IR spectra, which we estimated from the lifetime distribution provided by the maximum entropy method (MaxEnt). But the noise and baseline errors in the data limit its usefulness. In this second period we focused in the combination of MaxEnt with singular value decomposition and derivatives. We obtained narrower and more robust lifetime distributions, increasing the resolution power to identify intermediates. Our preliminary results show that the bR photocycle consists of seven intermediates states, one more than habitually considered in quantitative models.

The decomposition of the experimental time-resolved spectra into the time-evolution of intermediates and their pure spectra would provide simultaneously dynamic and structural information on the photocycle intermediates. But this decomposition is overwhelmingly ambiguous, even if a kinetic scheme is assumed, as illustrated by the contradictory results in the literature. We have explored a kinetic scheme independent alternative. We confirmed that by appropriately probing some IR bands it is possible to estimate the kinetics of most of the intermediates, and without need of assuming any kinetic scheme. This approach might open new prospects for a robust quantitative kinetic analysis of the photocycle of bR and other proteins with photocycles.

We studied the structural changes in MelB mutants by Na+ and melibiose-induced IR difference spectroscopy. Four aspartic residues previously implicated in Na+ binding were mutated to cysteine. The results suggest that Asp55 and Asp59 are essential ligands for Na+ binding. Though Asp124 is not essential for Na+ binding, it is proposed that its interaction with the bound cation is critical for the full Na+-induced conformational changes resulting in the coupling of the ion and sugar binding sites; this residue may also be a sugar ligand. Finally, Asp19 does not participate in Na+ binding but it is a melibiose ligand.