Exploiting RAS dimers for ligand discovery

Dillon Chu

Mentor: Dr. Trent Balius, RAS Com Chem Team, NCI RAS Initiative, Frederick National Lab for Cancer Research.

Date/Time: August 24, 2021 at 3:40pm

Abstract: The RAS proteins are important cancer targets as they are implicated in over 30 percent of all cancers. In a letter to the editor, researchers identified small molecules that had bound to KRAS-KRAS interfaces. This gave rise to the idea that if a small molecule is capable of inducing a non-functional dimer of KRAS, then there is potential to target functional KRAS dimers and identify molecules that will modulate dimer formation which might have interesting biological effects. The main goal of this project was to identify certain KRAS interfaces and to examine their characteristics through two primary methods: docking and molecular dynamic simulations.

In this project, molecular docking was used as a pose prediction tool. We used the DOCK 6 software to take a small molecule and place it in various configurations within a pocket on the receptor (protein). These poses of our molecule were assigned a score which quantified the quality of its fit to the receptor. The BI-2852 molecule was docked into a pocket on the monomer of RAS and dimer of RAS to examine possible differences. As expected, the docking to the monomer produced a pose that was different from the crystallographic pose while the docking to the dimer produced a pose that matched the crystallographic pose well.

Molecular dynamics is a method of simulating the movement of molecules. Here we simulated two systems: An RAS-RAS dimer without ligand and an RAS-RAS dimer with two ligands bound at the dimer interface (based on the 6GJ8 structure). We then performed two types of post-trajectory analyses over time in our simulations: i. measure key distances and ii. measured the root mean square deviation (RMSD) to the starting frame. The salt bridge formed between LYS5 of monomer 1 and GlU3 of monomer 2 as well as the symmetric interaction were measured by analyzing the MD simulations. The distances between the two residues were measured over time and the results were plotted. The distances of the two residues in the absence of a ligand had a much wider range and were incredibly sporadic. The distances in the presence of a ligand showed ranges with little variation and proved mostly consistent across the timeframe. Next, we analyzed our trajectory by examining the movement within monomers and one monomer relative to the other. This was performed by first fitting (or aligning) one monomer in a dimer and then calculating the RMSD to quantify the movement of the fitted monomer and the other monomer. As expected, when the monomer being examined was fit to itself, there was little variation in movement. In contrast, when the monomer being examined was not fit to itself, the monomer showed a great range of movement with the monomer possessing a ligand having a slightly less range of movement. The snapshots in time with the highest and lowest RMSD values were visualized and overlayed on top of one another in order to visualize the difference in monomer movement or “rocking”.

The project supports the idea that the KRAS dimer that exists in complex with BI-2852 relied on the small molecule for its stability. In the future, we intend to examine more interfaces that exist in the Protein Databank (PDB), identify possible pockets that accommodate ligands and identify ligands that may stabilize other interfaces. My work has laid the groundwork for these future goals.