Theoretical Study On The Interaction Between Divalent Metal Ions And Amino Acids In Aqueous Solution

Divalent metal ions are key cofactors that regulate the structure and functions of proteins.For example,Ca²⁺is a key signaling substance that triggers protein conformational changes.Mg²⁺and Zn²⁺ions are widely used as cofactors in enzymes.The biological function of metal ions is achieved through the special ion-binding structure of related biomolecules.Metal ions usually bound to protein molecules by coordinating with amino acid side chain groups such as ASP,GLU,CYS and HIS to form metal active sites,so as to initiate or control a series of subsequent biochemical reactions.The theoretical study of the interaction between amino acid side chains and metal ions is of great guiding significance for understanding the interaction between metal ions and proteins.Molecular dynamics simulation based on empirical force field is simple and efficient,but there is a problem of low accuracy in dealing with systems containing divalent metal ions.Molecular dynamics simulation based on first-principle calculation has high transferability and accuracy,but the calculation cost is expensive.In recent years,the proposal of interatomic potential energy surface model based on deep learning has enabled the realization of molecular dynamics simulations of large-scale systems with quantum precision.In this paper,the interaction mechanism between Mg²⁺,Ca²⁺,Zn²⁺and amino acid side chain analogs in aqueous solution was systematically studied by the combination of neural network potential energy surface,molecular dynamics simulation and umbrella sampling.The calculated potential of mean force not only reveal the binding process of each ion and amino acid side chain analogs,the most stable coordination structure,but also show the difference between different ions.In addition,we also use the neural network based potential of mean force as a standard to benchmark classical force fields.In this paper,we test two potential energy functions 12-6 LJ potential and 12-6-4 LJ potential.Overall,12-6-4 potential is more accurate than 12-6 potential,but there is still some room for improvement.The neural network based potential of mean force provides good guidance for the development of force field for metal ions.