Theoretical And Computational Studies Of Structure-Function Relationship In Light-Harvesting Protein And MUP-Ligand Complex

Biological macromolecules play indispensable roles in all life activities,and their specific three-dimensional spatial structures are the basis for their functions.Understanding the structure-function relationship is essential for the design of new drugs and artificial photochemical devices.The structures of biological molecules are often intricate,and some have not been well explained by experimental studies alone.With the rapid development of computer technology and theoretical methods,theoretical and computational investigations are becoming an effective research means.In this thesis,some effective theoretical methods are used for lightharvesting protein and protein-ligand complexes to reveal their structure-function relationship.Light-harvesting proteins transfer energy from the light-absorbing antenna protein to the reaction center with surprisingly high efficiency.However,the underlying molecular mechanisms of energy transfer are yet to be elucidated.To understand how the protein environment fine-tunes the optical properties and excitation energy transfer,the exciton model combined with the quantum master equation theory is used to study light-harvesting systems.The site energies of pigment molecules and the electronic couplings are used to construct the system Hamiltonian.The spectral density is used to describe the system-environment interaction.The semi-empirical method ZINDO/S-CIS,the point dipole approximation(PDA)method,and the Drude model are used for the effective evaluation of these three aspects,respectively.The absorption spectrum and population dynamics of the light-harvesting systems are calculated according to the dissipation equation of motion(DEOM)theory.In addition,the polarized protein-specific charge(PPC)scheme is used to provide a realistic description of the protein environment.Experimental studies have indicated that Fenna-Matthews-Olson(FMO)protein can almost perfectly perform energy transfer,but the excited energy transfer path is still controversial.Besides,the absorption spectra of the FMO complexes from various source organisms can be classified into two types.The FMO complex extracted from Prosthecochloris aestuarii 2K(P.aestuarii)has a spectrum of type 1,while that from Chlorobaculum tepidum(C.tepidum)is of type 2.The central bands of the absorption spectra are quite different in the two types of absorption spectra.However,questions remain unanswered regarding the origins of the spectral differences.Mutation-induced changes to the exciton structure and the absorption spectra provide a suitable means to investigate the critical role played by the protein.Based on the experimental results of mutation-induced changes of absorption spectra,a single-point mutation in the vicinity of BChl 6 is adopted for the two species of FMOs.Then quantummechanics/molecular-mechanics(QM/MM)calculations over the molecular dynamics(MD)simulation trajectories with polarized protein-specific charge scheme are performed for both wild-type FMOs and their mutants.Based on the MD trajectories,the site energies and the electronic couplings are calculated to construct the Frenkel exciton model Hamiltonian.The spectral densities are used to describe the system-bath interaction.The absorption spectra are calculated according to the recently developed DEOM theory.The present study shows that the single-point mutation in C.tepidum results in a switch of the absorption spectrum from type2 to type 1,which is consistent with the experimental results.In addition,the study predicts that the same mutation in P.aestuarii leads to the opposite transition(from type 1 to type 2).To find the molecular origin of the differences in the absorption spectra between FMO proteins from wild-type P.aestuarii and C.tepidum,we further analyze the contributions of individual pigments to the absorption spectra and compare these between wild-type FMOs and between wild-type FMO and its mutant.The results show that BChls 5,6,and 7 cause the difference in the absorption spectra of the two wild-type FMOs.Finally,population dynamics are calculated,and the energy transfer pathways are predicted.In drug design,one of the main goals is to find small molecules that bind tightly to the target protein.Accurate prediction of the protein-ligand binding free energy is of great significance for computer-aided drug design.Accurate free energy calculations require high accuracy of Hamiltonian,adequate sampling,and reliable free energy calculation methods.The Hamiltonian at the quantum mechanics level can accurately present the potential energy surface of the research system.However,the calculation for sampling under this potential energy surface is particularly expensive.Thus the sampling under the potential energy surface at the molecular mechanics level is more common.To obtain sufficient sampling,some enhanced sampling methods are used to improve the convergence,such as umbrella sampling(US),Metadynamics,and Alchemical sampling methods.After obtaining sufficient samplings,a reliable free energy calculation method needs to be selected.Common post-processing methods include thermodynamic integration(TI),thermodynamic perturbation(TP),and multistate Bennett acceptance ratio(MBAR).In this thesis,the Hamiltonian at the molecular mechanics level,the alchemical sampling method,the thermodynamic integration,and multistate Bennett acceptance ratio methods are used to calculate the binding free energy.In addition,the theoretically less accurate but computationally efficient end-point methods are used to calculate the binding free energy.Mouse major urinary protein(MUP)plays a critical role in regulating pheromones release and capture.Studying the relation of its structure and function contributes to clarifying the communication system of pheromones.Previous computational studies used early protein force fields and short simulation time to calculate binding thermodynamics,which may lead to nonconverged sampling and limited understanding of protein-ligand interaction patterns.Different free energy calculation methods are used to study their binding thermodynamics for 14 MUPligand complexes in the present study.Two AMBER protein force fields are used separately for molecular dynamics simulations,including the widely accepted AMBER14 SB force field and the newest AMBER19 SB force field,which are employed to access the development of the protein force fields.Then the MUP-ligand interaction networks to identify important residues of stabilizing the bound structure are further analyzed.The protein secondary structure and protein-ligand interaction networks under the two AMBER force fields are compared.The results show that the AMBER19 SB force field does not present any better performance than the widely used AMBER14 SB force field for RMSDs of the heavy atoms of the protein backbone and the binding thermodynamics.Thus the AMBER14 SB force field is still a good choice for predicting binding thermodynamics.At present,many researchers are devoted to developing different methods of understanding the structure-function relationships for specific biological systems.Some effective methods are adopted for pigment-protein and protein-ligand complexes.However,these methods also have some limitations,and further work is needed to be done in the future.Finally,a summary of the content of this thesis and prospects for future research are presented.