HINGEID: AUTOMATED TOOL FOR IDENTIFICATION AND DESIGN OF HINGE-BENDING MOTIONS

Conformational changes play a fundamental role in biological processes. In this context, hinge-bending motions involving two or more domains have attracted attention due to their contribution to protein function. Indeed, the identification and characterization of hinge regions are of scientific interest, since mutations at these sites can directly affect protein function, as demonstrated in previous work by our group on the rational hinge design of the cellulase Cel9A-68. Moreover, it was recently demonstrated that hinge points can be inserted or enhanced through combinatorial mutagenesis. Here, we aim to tightly control sampling of functional states by introducing mutations in hinge points via combinatorial design. We developed an automated Python framework that requires a single structure as input and allows the user to: i. quickly identify hinge regions and ii. introduce mutations around this site using combinatorial design. The T4-Lysozyme (PDB: 2LZM) was selected as our test system. Hinge points were identified by a ProDy function based on the Gaussian Network Model (GNM). They correspond to residues with low fluctuation located between regions that move in opposite directions. Combinatorial design was then performed on residues around the identified hinge points using pyRosetta. The generated structures were modeled with the CHARMM36m force field, and their conformational properties were assessed by molecular dynamics (MD) simulations using NAMD3 with three replicas per construct. Comparison of GNM predictions with the curated DynDom database showed that the predicted points coincide with regions identified in experimental structures. To investigate the impact on interdomain dynamics, we designed constructs enriched with residues of high (PHI) or low (NHI) hinge propensity. Wild-type MD simulations confirmed agreement with experimental conformational landscapes, with PHI constructs showing slightly increased open-state sampling. MD simulations of NHI constructs are still running and will be analyzed soon. Although experimentally observed states were explored, inconsistencies were observed among replicas, with the system becoming trapped in specific conformational populations, indicating limited sampling. To overcome this issue, we will perform Metadynamics, using the conformational difference vector between the open and closed states as the collective variable. In the future, we plan to make this tool available to the scientific community.