La pedra rosetta de les proteïnes fluorescents: una descripció atomística

Abstract

Fluorescent Proteins have become a focus of scientific interest due to their applications in biomedical problems and bio-imaging. The growing interest in red fluorescent proteins demands a clear and systematic rationalization of the processes occurring in this complicated biosistems. In order to reach such a level of detail, a computational work has been done, using theories coming from Quantum Chemistry, Statistical Thermodynamics and Classical, Semi-classical and Quantum Dynamics. In a first step, and by means of molecular dynamics simulations integrated using the task-farming method, also developed in this work, in which we have applied Time Dependent Density Functional Theory calculations, we have been able to reproduce the absorption spectra of the selected proteins. The three green fluorescent proteins studied have been S65T/H148D-GFP, E222Q/H148D-GFP i wt-GFP. We have concluded that the proximity of a negative charge near the phenolic ring of the chromophore is presumably lowing the excitation energy, which increases the absorption wavelength. Regarding the second group of proteins, this time red fluorescent proteins, we have studied the mNeptune and mCardinal mutants, whose show their spectroscopic activity in the transparency window of the living mammal tissue. Using a similar methodology as the one used in the case of GFPs, we have been able to qualitatively reproduce the 4 nm red shift between the two mutants. In this sense, we have showed how the oxygen of the residue Phe62 is effectively interacting with the residue Gly41 in the case of mCardinal (interaction not found in mNeptune), which makes the former to be more red-shifted. In a second chapter, we have been able to reproduce and explain the experimental behavior observed for the mutant S65T-H148D, which is considered in this work as the Rosetta Stone of the fluorescent proteins, given that the knowledge obtained for it can be generalized to other proteins. In this sense, we have refuted the experimental supposition of the dramatic gain of acidity of the chromophore in the excited state and we have substituted it by a more dynamical view, in which there are structural factors that promote the proton transfer between the chromophore and the Asp148: (1) the spotted interaction between the residue Tyr146 and the chromophore, which stabilizes the photo-product and (2) the proper solvation of the proton acceptor, the residue Asp148. The solvation of this residue has to be adequate so it can accept the proton coming from the chromophore. Furthermore, by means of Quantum Dynamics calculations and non-adiabatic molecular dynamics simulations, we have been able to reproduce the time-scale for the proton transfer and the experimental emission time-resolved spectra. The first regime of relaxation has been associated to the main proton transfer and the second regime, which occurs in the picosecond time-scale, has been associated to the vibrational cooling, which involves (1) the separation of the residue Asp148 from the chromophore, (2) the loss of planarity of the second and (3) the interaction with the residue Tyr 148, which stabilizes the photoproduct.