ИСТИНА

Antibodies are immune system proteins that recognize the surfaces of foreign molecules (antigens) for subsequent elimination from the organism during an adaptive immune response. Continued exploitation of antibodies for therapeutic purposes relies on more efficient ways to develop these molecules. Antibody structures have the potential to be useful during drug development, allowing the implementation of rational design procedures. The most challenging part of the antibody structure prediction is to model the H3 loop, which in addition is often the most important region in an antibody's binding site. The majority of H3 loop prediction approaches are based on computationally expensive decoy generation. The biggest challenge in such ab initio modelling remains selection of best loop models among those generated. The use of coarse grained (CG) models in a variety of molecular modelling techniques has proven to be a valuable tool to probe the time and length scales of systems beyond what is achievable with traditional all atom (AA) models. The recently released CG forcefield Martini3 with an expanded ability to include specific interactions such as hydrogen bonding and electronic polarizability may have a practical use for antibody structure modelling. We aimed to demonstrate the potential application of CG molecular dynamics (MD) for the specific task of H3 loop conformation sampling. We used pairs of apo- and antigen-bound antibody Fab-fragment structures that have differing H3 loop conformations. The molecular dynamics simulations based on CG forcefield Martini3 were set up starting from an antigen-bound structure with antigen removed. In order to enhance the sampling of the H3 loop conformational space we used Hamiltonian replica-exchange (HREX) MD with the H3 loop assigned to be the hot region. The overall Fabs’ structure remained stable along the trajectories in HREX replicas. The H3 loop region was observed in conformations close to the unbound antibody crystal structures, thus proving the CG MD suitable for antibody modelling.