Targeted Molecular Dynamics On The Conformation Transition Mechanism Of Protein

Modern simulation techniques allow confirming experimental results or examining more closely the interactions responsible for the behaviour of proteins in the presence of some perturbation. They allow prediction of phenomena and may sometimes examine biological systems into more detail than experimental techniques. Moreover they give a better insight into which amino-acid residues are involved in complex phenomena.In this work we have used numerous techniques (MD, TMD, SMD, molecular docking, mixed QM/MM methods) which are now widely available, in order to examine conformational changes of some proteins, induced by a change in the surrounding pH, weak interactions with a ligand (such as H bonds, van der Waals or electrostatic interactions) and finally by the creation of a covalent bond with a ligand.Although these techniques are readily available, they need a great deal of computing power and time, since comparisons must be performed between different model systems, different types of calculations must be carried out (such as molecular docking, energy evaluations, free energy profiles,..), cross-checking and verifications are mandatory, and the results should be very carefully analysed to produce significant and reliable information.We first focused our interest on the opening process of the KcsA potassium channel, a widely studied system, but although recent literature indicated that the cytoplasmic domain plays a role in pH sensing, no simulation had been performed on the complete system.Using TMD we have carried out for the first time, a simulation of the opening of full-length KcsA, and focused our interest on the cytoplasmic domain.Studying different model systems at pH=4and pH=7we showed that stabilization occurs at pH=7for full-length systems, showing that specific amino-acids protonation in the cytoplasmic domain play a role in the opening process.We have been able to show the role of some of them such as His124,128and145, Glu130,134and135, Asp156and157as well as Glu146and Asp149. However, the role of His25invoked by several authors remains unclear.Our results also show that the ionisable residues located near the transmembrane domain act quickly, within the first nanosecond, so that they certainly help trigger the opening process. As the cytoplasmic sequence proceeds, the above-described events appear later in the TMD trajectory: these structural changes are gradually transmitted to help the KcsA gating. This confirms that pH sensing is due to both domains:the cluster region on the boundary of internal membrane, together with the cytoplasmic part.Although much work remains to be done to fully describe and understand this extremely complex system, we feel we have brought about new insights into the role of the cytoplasmic domain changes induced by protonation and a precise description of the involved specific amino-acids and their interactions, which finally results in the opening process.We were then interested in the interaction mode of inhibitors of the FGRF3protein kinase, involved in several bone pathologies so that efficient inhibitors represent a high therapeutic interest. As described in Chapter3, protein kinases evolve between an active and an inactive state characterized by the conformational change of a residue triad, Asp-Phe-Gly or DFG. Several inhibitors have experimentally been demonstrated, and we were interested in simulating the active to inactive state transition with the TMD technique, to our knowledge, for the first time that TMD. Owing to molecular dynamics, TMD, docking calculations have shown the existence of several intermediate conformations of the protein with which inhibitors may interact. In the case of type Ⅱ inhibitors we have shown that they are able to interact with an early open state of the kinase DFG-out cavity (DFGout’). These ligands do not prevent the inactivation process but rather participate, through favourable interactions, to the enzyme conformational transition. This induced-fit scheme should thus be considered for the kinase DFG-in/out transition. Another consequence of our work is that the design of new type Ⅱ inhibitors should take into account not only the ability of binding to the final DFGout structure but to the transient conformations as well.However, the possibility to obtain irreversible inhibition has been experimentally demonstrated for another kinase, FGFR1, quite similar to FGFR3, by the creation of a covalent bond between the inhibitor and the protein. This inhibitor reacts with the Cys488residue, located around the ATP binding site. For this reason, we studied this FGFR1protein-FRIIN inhibitor system with appropriate methods (i.e. quantum/classical mechanics mixed methods) in order to get some insight into the chemical reactivity of the inhibitor and the reaction pathway.Our results have first shown that there is no important conformational change induced by the formation of the covalent bond, as compared to the situation where the inhibitor only forms a complex with the protein. We were thus able to conclude that the first step of irreversible binding is the complexation of the compound inside the binding site and, then, the formation of the covalent bond. This result might be important for the design of future irreversible inhibitors since such molecules should first present a correct recognition pattern towards the protein target. Thus, an irreversible inhibitor should derive from a good competitive inhibitor.Then, molecular simulation permitted us to propose a mechanism for the chemical reaction of the irreversible inhibitor (FRIIN) with the FGFR1kinase. We have shown that only one mechanism, involving concerted addition and proton transfer is conceivable. This mechanism should be followed by a keto-enol tautomerization. The calculated energy profile showed that the thermodynamically permitted reaction is indeed irreversible.The results we obtained with FGFR1might be useful to pursue the research of irreversible inhibitors of the highly homologous FGFR3.