In recent years, world economy and technological advancement have been transformed by Genomics, which allows us to study, design and build biologically relevant molecules. Genomics is already deeply embedded in industries as diverse as pharmaceutical, food and agricultural, environmental and bio-tech in general. Fast and cheap tools for gene sequencing, protein expression and analysis are commonly used for high-throughput genomic-related studies. However, due to experimental difficulties and long time scales (e.g., protein crystallization), protein structure determination, and thus the fundamental structure function rationalization, cannot presently be performed at the same fast pace: a fact that is slowing down the discovery of proteins with new features, as well as ex novo design. These difficulties are particularly felt in the field of photobiology, where the crystal structure of Bovine rhodopsin (Rh, retina dim-light visual photo-receptor), still remains the only structure of a vertebrate photo-receptor sensor available for photobiological studies since the year 2000. Rhodopsins constitute a class of light-triggered proteins that can be found throughout the whole spectrum of living organisms, and represent the perfect blue-print for building light-activated bio-molecular machines. In principle, the problem of not having a sufficient number of rhodopsins molecular structures could be circumvented and overcome with the construction of accurate atomistic computer models of the set of studied photoreceptors, which would allow: (i) in silico fundamental structure-function characterization, (ii) thorough and detailed screening of mutant series, and even (iii) ex novo design. Nevertheless, such models should also be constructed using a fast, relatively cheap, reliable and standardized protocol, of known accuracy. In this thesis, we refine and test the Automatic Rhodopsin Modeling (ARM) computational protocol, which we demonstrate as being capable of helping to address the above issues. Such protocol has the primary target of generating congruous quantum mechanical/molecular mechanical (QM/MM) models of rhodopsins, with the aim of facilitating systematic rhodopsin-mutants studies. The cornerstone of this thesis is the validation of the ARM protocol as a successful attempt to provide a basis for the standardization and reproducibility of rhodopsin QM/MM models, aimed to study the behaviour of photoactive molecules. First, we validate the ARM protocol, which employs a CASPT2//CASSCF/AMBER scheme, for a benchmark set of rhodopsins from different biological kingdoms. We show that ARM is able to reproduce and predict absorption trends in rhodopsin protein sets, with blue-shifted values not much displaced (a few kcal/mol) from the observed data. Secondly, we present how to use this protocol towards a better design of novel mutations as applications for Optogenetics, an innovative biological tool aimed to visualize and control neuron signals through light. Two different microbial rhodopsins are studied: Krokinobacter eikastus rhodopsin 2 (KR2), a light-driven outward sodium pump, and Anabaena sensory rhodopsin (ASR), a light sensor. In both cases, the qualitative and quantitative information acquired from the ARM-obtained QM/MM models reveal nature (electrostatic or steric) and extent of the mutation-induced changes on the retinal configuration, which, in turn, are the cause of the shift in the absorption wavelength of the relative mutants. Finally, we explore the fluorescence of ASR mutants, particularly useful for the visualization of neuronal activity. The target of this work is to use QM/MM simulations to understand the opposite behaviour observed in two blue-shifted ASR mutants, where one presents a negligible fluorescence, while the other displays one order of magnitude enhanced fluorescence, with respect to the wild type protein. Our QM/MM models show that specific electrostatic and steric interactions control the character mixing of different electronic states, opening a path to the rational engineering of highly fluorescent rhodopsins. In conclusion, within the limits of its automation, the ARM protocol allows the study of ground and excited states of specific photoactive proteins: rhodopsins. This opens the way to an improved molecular-level understanding of rhodopsin photochemistry and photobiology. The results obtained highlight the importance of having a standardized, effective and automatic protocol, which renders this kind of studies more efficient and accessible, by drastically shortening the time required to produce accurate and congruous QM/MM models. For the above reasons the author of the present thesis believes that ARM stands as an important cogwheel in the virtuous cycle between experimental and theoretical work, aimed to prepare the photobiological tools for tomorrow’s needs.

MARIN PEREZ, M.D.C. (2019). Benchmarking and applications of a computational photobiology tool for design of novel and highly fluorescent rhodopsin proteins.

Benchmarking and applications of a computational photobiology tool for design of novel and highly fluorescent rhodopsin proteins

María del Carmen Marín Pérez
2019-01-01

Abstract

In recent years, world economy and technological advancement have been transformed by Genomics, which allows us to study, design and build biologically relevant molecules. Genomics is already deeply embedded in industries as diverse as pharmaceutical, food and agricultural, environmental and bio-tech in general. Fast and cheap tools for gene sequencing, protein expression and analysis are commonly used for high-throughput genomic-related studies. However, due to experimental difficulties and long time scales (e.g., protein crystallization), protein structure determination, and thus the fundamental structure function rationalization, cannot presently be performed at the same fast pace: a fact that is slowing down the discovery of proteins with new features, as well as ex novo design. These difficulties are particularly felt in the field of photobiology, where the crystal structure of Bovine rhodopsin (Rh, retina dim-light visual photo-receptor), still remains the only structure of a vertebrate photo-receptor sensor available for photobiological studies since the year 2000. Rhodopsins constitute a class of light-triggered proteins that can be found throughout the whole spectrum of living organisms, and represent the perfect blue-print for building light-activated bio-molecular machines. In principle, the problem of not having a sufficient number of rhodopsins molecular structures could be circumvented and overcome with the construction of accurate atomistic computer models of the set of studied photoreceptors, which would allow: (i) in silico fundamental structure-function characterization, (ii) thorough and detailed screening of mutant series, and even (iii) ex novo design. Nevertheless, such models should also be constructed using a fast, relatively cheap, reliable and standardized protocol, of known accuracy. In this thesis, we refine and test the Automatic Rhodopsin Modeling (ARM) computational protocol, which we demonstrate as being capable of helping to address the above issues. Such protocol has the primary target of generating congruous quantum mechanical/molecular mechanical (QM/MM) models of rhodopsins, with the aim of facilitating systematic rhodopsin-mutants studies. The cornerstone of this thesis is the validation of the ARM protocol as a successful attempt to provide a basis for the standardization and reproducibility of rhodopsin QM/MM models, aimed to study the behaviour of photoactive molecules. First, we validate the ARM protocol, which employs a CASPT2//CASSCF/AMBER scheme, for a benchmark set of rhodopsins from different biological kingdoms. We show that ARM is able to reproduce and predict absorption trends in rhodopsin protein sets, with blue-shifted values not much displaced (a few kcal/mol) from the observed data. Secondly, we present how to use this protocol towards a better design of novel mutations as applications for Optogenetics, an innovative biological tool aimed to visualize and control neuron signals through light. Two different microbial rhodopsins are studied: Krokinobacter eikastus rhodopsin 2 (KR2), a light-driven outward sodium pump, and Anabaena sensory rhodopsin (ASR), a light sensor. In both cases, the qualitative and quantitative information acquired from the ARM-obtained QM/MM models reveal nature (electrostatic or steric) and extent of the mutation-induced changes on the retinal configuration, which, in turn, are the cause of the shift in the absorption wavelength of the relative mutants. Finally, we explore the fluorescence of ASR mutants, particularly useful for the visualization of neuronal activity. The target of this work is to use QM/MM simulations to understand the opposite behaviour observed in two blue-shifted ASR mutants, where one presents a negligible fluorescence, while the other displays one order of magnitude enhanced fluorescence, with respect to the wild type protein. Our QM/MM models show that specific electrostatic and steric interactions control the character mixing of different electronic states, opening a path to the rational engineering of highly fluorescent rhodopsins. In conclusion, within the limits of its automation, the ARM protocol allows the study of ground and excited states of specific photoactive proteins: rhodopsins. This opens the way to an improved molecular-level understanding of rhodopsin photochemistry and photobiology. The results obtained highlight the importance of having a standardized, effective and automatic protocol, which renders this kind of studies more efficient and accessible, by drastically shortening the time required to produce accurate and congruous QM/MM models. For the above reasons the author of the present thesis believes that ARM stands as an important cogwheel in the virtuous cycle between experimental and theoretical work, aimed to prepare the photobiological tools for tomorrow’s needs.
2019
MARIN PEREZ, M.D.C. (2019). Benchmarking and applications of a computational photobiology tool for design of novel and highly fluorescent rhodopsin proteins.
MARIN PEREZ, MARIA DEL CARMEN
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11365/1070289
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo