Theoretical Calculation Methods And Application Exploration In Protein Function Research

This work focuses on the theoretical calculation methods and application exploration in protein function research,and performs three main researches:1.In order to reduce the computational cost of current explicit solvent molecular dynamics(MD)simulations,a fast-slow method for simulating biomolecules in explicit solvents is proposed in this research.The method divides the entire system into two parts: the core layer(solute molecules)and the peripheral layer(solvent molecules).The core layer is treated with standard MD methods,while the peripheral layer is treated with a slower dynamic integration method to reduce computation.Four different simulation models,including gas-phase,implicit solvent,fast-slow explicit solvent,and standard explicit solvent,are compared in test of several small proteins.The results show that gas-phase and implicit solvent models cannot provide a realistic solvent environment,nor can they generate reliable protein dynamic structures.However,the fast-slow model can reproduce solvent effects that are similar to those of the standard explicit solvent model in terms of simulation stability,solvent distribution,and sampling space,while achieving an order of magnitude acceleration.Therefore,the fast-slow model proposed in this work provides an effective method for accelerating MD simulation of biomolecules in explicit solvents.2.The dynamic recognition mechanism and binding mode in the binding process of coronavirus spike glycoprotein(S protein)and angiotensin converting enzyme 2(ACE2)in host cells have been lack of clear understanding for a long time,which seriously restricts the progress of research and development of corresponding vaccines and inhibitors.Aiming at the problem,this work systematically compares and analyzes the receptor binding domains(RBDs)of six coronaviruses’ S proteins through molecular dynamics simulations and binding free energy calculations.The results show that the affinity and stability of SARS-Co V-2’s RBD are stronger than those of other coronaviruses when binding to ACE2,and the diversity of the Loop 2(Y473-F490)region is an important factor affecting stability and binding affinity.In addition,the solvent-accessible surface area(SASA)and binding free energy of different RBD subunits indicate the existence of an "anchor-lock" recognition mechanism when the S protein binds to ACE2.The Loop 2 structure on RBD acts as the anchor that is first recognized by ACE2,and then the Loop 3(G496-V503)structure locks ACE2 to another end of RBD,while the charged or long-chain residues in the β-sheet 1(N450-F456)region enhance this binding.The "anchor-lock" binding mechanism proposed in this paper is verified by umbrella sampling simulations of the dissociation process.This work provides an important theoretical basis for the development of new vaccines and inhibitors for SARS-Co V-2.3.In response to the current problem of linear epitopes not considering threedimensional structures,a new spatial epitope search algorithm is designed based on peptide chip detection technology in this research.The method first uses random peptide chips to obtain peptides with differential signals in positive sample serum and control group serum and simulates these differential signal peptides as epitope peptides of antigen proteins.Then,based on the three-dimensional structure of the antigen protein,a residue contact network is constructed,and the epitope peptides on the chip are fragmentarily matched with the residue contact network.The matching residue is assigned a differential signal value as its score.The residues with higher matching scores are recursively grouped,and then constructed a continuous and complete spatial epitope using automatic stitching program.The two epitopes predicted by the algorithm on the S protein of SARS-Co V-2 have been verified by published research and structure,which solves the problem of incomplete three-dimensional structure of the current linear epitope prediction.