Development Of Quantum Mechanical Fragmentation Method For Large Complex Biomolecules

The ab initio quantum chemistry theories have witnessed large development in the past half century,which makes people try to apply theoretical calculation methods into more and more large systems.Related applications are becoming more common in the fields of biological macromolecules and materials.The scope of application is also from the initial experimental results verification analysis to the current predictions,and even to guide the experiment.With the development of computational power,the theoretical level of calculation and the complexity of the computing system are getting higher and higher.Because the computational scale of quantum chemical methods has a high-order exponential relationship with the size of the system,people often only use lower theoretical levels or reduce the size of the computing system when calculating the properties of biological macromolecules,thus sacrificing theoretical calculation accuracy.The linear-scaling quantum mechanical?QM?fragmentation method opens the new door for the application in complex biological macromolecular systems,and has received widespread attention in recent years.The fragmentation method could reproduce the results of high-order theoretical level calculation of the whole system of biomacromolecules,so that the properties of macromolecular energy,chemical shift,and kinetics can be accurately calculated.Firstly,the total energies of RNAs were performed by efficient linear-scaling quantum mechanical calculations using EE-GMFCC method.In EE-GMFCC,the nucleotide-centric fragment with large caps or small caps?termed EE-GMFCC-LC and EE-GMFCC-SC,respectively?QM energy was calculated and deducted by the energies of concaps to obtain the total energy of RNA.In energy calculation,we also considered two-body QM interaction energy between non-neighboring nucleotides,which were in close contact.Levels at Hartree-Fock?HF?,density functional theory?DFT?and second-order many-body perturbation theory?MP2?were applied in tests of calculating the total energies of several RNAs,using the EE-GMFCC-LC and EE-GMFCC-SC methods,respectively.The results show that compared with EE-GMFCC-LC method,the efficiency of the EE-GMFCC-SC method was about 3 times faster and with minimal accuracy sacrifice.The EE-GMFCC-SC method also applied HF and DFT calculations on relative energy of 20 different conformers of two RNA systems,respectively.Both single point and relative energy calculations demonstrate that the EE-GMFCC method has only a few kcal/mol deviations from the full system results.Next,AF-QM/MM approach was used to perform ¹H,¹³C and ¹⁵N NMR chemical shift calculations on RNAs.The influence of density functionals,force fields,ensemble average and explicit solvent molecules on NMR chemical shift calculations were also investigated.By comparing the performance of a series of density functionals,we found that the mPW1PW91 functional is one of the best functionals for predicting RNA ¹H and ¹³C chemical shifts.The study also shows that the performance of the force fields in describing H-bond strength could be validated by AF-QM/MM calculated imino proton chemical shifts.Compared to the FF10 force field,the polarized nucleic acid-specific charge?PNC?model significantly improves the accuracy of imino hydrogen and nitrogen NMR chemical shift prediction,which underscores that the electrostatic polarization effect is critical to stabilizing the hydrogen bonds between base pairs in RNAs.Furthermore,by adding explicit water molecules the result shows that the accuracy of the chemical shift of amino proton could be improved.Finally,the AF-QM/MM method was applied for NMR chemical shift calculations of protein-ligand complexes.In the AF-QM/MM approach,the protein binding pocket is automatically divided into capped fragments?within?200 atoms?together with ligand.Each fragment were calculated in parallel using density functional theory?DFT?to predict NMR chemical shifts.Meanwhile,the Poisson-Boltzmann?PB?model was used to add the solvent effect into calculations,which properly accounts for the electrostatic polarization effect from the solvent for protein-ligand complexes.The NMR chemical shifts of neocarzinostatin?NCS?-chromophore binding complex calculated by AF-QM/MM accurately reproduce the large-sized system results.The ¹H chemical shift perturbations?CSP?between apo-NCS and holo-NCS predicted by AF-QM/MM are also in excellent agreement with experimental results.Furthermore,the protein-ligand binding conformation could be validated by DFT calculated chemical shifts of the chromophore and residues in the NCS binding pocket.The new scoring function combining the CSP of the atoms in the binding pocket with the Glide scoring function can accurately distinguish the native ligand pose from decoy structures.Therefore,the AF-QM/MM approach provides an accurate and efficient platform for protein-ligand binding structure prediction based on NMR derived information.In summary,the innovations of this thesis mainly have the following three aspects:1.The EE-GMFCC method realized the full QM calculation of RNA energy calculation,and further developed a more computational efficient fragmentation scheme.2.The paper systematically studied the chemical shift prediction of RNA,and gave solutions to improve the prediction accuracy of labile hydrogen,which is helpful for RNA structure prediction and base pair spacing measurement.3.The developed fragmentation method for protein-ligand NMR chemical shift predictions combined with the Glide scoring function can accurately distinguish the natural ligand pose and docking structure of NCS complexes.