| Conformational studies for complicated systems have been for along time one of the most vital and time-demanding issues in bothexperimental and computational research. Due to the instantaneity, it isdifficult to capture high-energy intermediates by experimental approaches.Meanwhile, it is unrealistic to reconstruct a full-dimension potentialenergy surface for a complicated system with all-atomistic simulation,upon the intensive computational time required. Currently, MD-basedtransition state searching methods are scarce, for that rare events are notwell studied in sampling, and huge amount of molecular simulation areineffective and unnecessary. To overcome the dilemma of computerlimitation and reaction complexity, people usually employ transition statetheory to conduct a ‘supervised machine learning’-like method, in whichcore structures in reaction pathway are used for all-atomistic simulation.Currently, transition state theory has been widely applied in mechanismresearch in organic chemistry.In principles, the study of reaction dynamics for complicatedbiosystems demands a traversal search of transition structures from all thepotential reaction pathways. However, intelligent techniques of global TSscanning for structurally flexible molecules are yet lack. In this thesis, aprotocol pipeline was proposed and tested to accelerate exploration oftransition structure search. The method was designed by the following steps:1) Use CPMD Metadynamics simulation to extract corestructures and their featuring coordinates from a simplifiedsystem of target reaction;2) The grafting strategy, combined with fast QM methodand automatically conformation searching engine, to obtainreactants, products, and intermediates of target reaction;Furthermore, occurrence frequencies of the reaction pathways arestudied based on transition state theory, and thereby the resulting reactiondynamics have been verified by experimental data. This protocol has beenapplied in the reactivity studies of amide isomerization of a variety ofN,N’-benzamides, N,N’-(iso)nicotinamides and N,N’-pivalamides.In the above quantum calculation, a high-level DFT method andconductor-like polarizable continuum model was also employed tocalculate reaction rate-determining steps and the activation energies, toreveal the structure-activity relationship among energies, structures andchemical surroundings.As a parctical application of this protocol, the catalytic mechanismof M4L6supramolecule was studied. A M4L6supramolecule consists oftetrahedral cationic metals and six bis-bidentate ligands. As an artificialenzyme, M4L6provides a hydrophobic region to benefithydrophobicity-favored reactions for encapsulated compounds. To studyamide isomerization under enzymatic environments, a host-guest modelof M4L6-amide complexation is reconstructed with guest substratesunder both ground and transition states. |
