Study On The Conformational Dynamics Of STING Protein Using Molecular Dynamics Simulation And Markov State Model Analysis

The Stimulator of Interferon Genes(STING)protein binds cyclic dinucleotides(CDNs)to activate the production of type I interferons and inflammatory cytokines,inducing the body to enter an antiviral state and enhancing the immune response against tumors.The activity and function of the STING protein can be regulated by certain compounds,which can affect the occurrence and regulation of immune responses.For example,STING agonists can effectively treat pathogen infections and cancer,while antagonists may be able to treat autoimmune or auto-inflammatory diseases.The development of structurally diverse and more drug-like STING ligands has important roles and application prospects,which requires a better understanding of the interaction mechanisms between STING protein and small molecules,as well as the microscopic processes of small molecule regulation of STING protein conformational changes.The information provided by static crystal structures alone is insufficient to describe these interactions and dynamic changes.In this paper,we systematically studied the mechanism of action between the STING protein and its small molecule ligands using various molecular dynamics simulation methods,including the following aspects:Firstly,we used molecular dynamics simulations combined with hydrogen bond analysis and the MMGBSA method to investigate the mechanism of small molecule agonist-specific binding to the STING protein for nine different skeletal types.We found that the protein binding pocket is large enough to accommodate small molecules with different skeletal shapes.By decomposing the free energy into each amino acid,we found that the main binding sites for agonists are two α-helices and residues above the LID region in the pocket.Although the specific residues involved in binding differ,the residues R238 in the LID region,Y163,Y167,T263,and P264 at the bottom of the pocket play important roles in all systems.Among them,the residue R238 may be the key for small molecules to specifically activate the STING protein.Using dynamic cross-correlation matrix analysis,we also found that the correlation between the two chains in the "closed" activation systems is mostly positive,while the correlation between the two monomers in the "open" system is negative.Next,in order to study the dynamic mechanism of antagonist dissociation from the STING protein pocket,we first used molecular dynamics simulation combined with a Markov state model to analyze the interaction mechanism between the antagonist and STING protein.We found that the antagonist induced a transition of STING and stabilized it in a more open conformational state.Through MMGBSA free energy calculations and hydrogen bond occupancy analysis,we identified the important residue T263,where the number 2 nitrogen atom on the antagonist’s large ring and the phenyl ring play a key role.We found that during the interaction between Astin C and STING,the information transmission between residues decreased,and the residues in the pocket lost contact with those in the LID region.Subsequently,the RAMD simulation combined with MD-IFP(molecular dynamics combined with interaction fingerprint)provided us with several possible pathways for dissociation and some key residue barriers.The metadynamics simulation results provided us with the energy barrier and free energy changes required for the antagonist to overcome and leave the pocket.Finally,in order to understand the conformational communication within the ligand-binding domain of STING protein,we used molecular dynamics simulations and Markov state model analysis to explore the conformational dynamics of STING with different ligands starting from the same open conformation.Our results showed that agonists and antagonists induce different conformational changes and thus have different mechanisms of action.We found that the agonist small molecule can induce the protein to transition towards a "closed" state,ultimately stabilizing it in a closed "activated" conformation,while the antagonist can only stabilize the protein in an open conformation.To further investigate how the agonist small molecule c GAMP induces this conformational change,we used accelerated molecular dynamics simulations combined with Markov state model to elucidate the activation mechanism.We found a most probable activation pathway G1-G2-G4-G5,and through a detailed analysis of the residue interaction network along this transition path,we identified several key residues that could affect the conformational change by influencing residue communication.Comparing the system with antagonist small molecules,we also found that the information transfer between pocket residues and LID residues was significantly increased,which could explain the reason for the different conformational changes induced by the two ligands.The study of these mechanisms not only provides us with a clearer understanding of the STING protein,but also provides useful information for future rational drug design based on the STING protein,with the hope of helping us understand and design more effective drugs that target the STING protein.