CHARMM-GUI Membrane Builder for Lipid Nanoparticles with Ionizable Cationic Lipids and PEGylated Lipids

Abstract

A lipid nanoparticle (LNP) formulation is a state-of-the-art delivery system for genetic drugs such as DNA, mRNA, and siRNA, which is successfully applied to COVID-19 vaccines and gains tremendous interest in therapeutic applications. Despite its importance, a molecular-level understanding of the LNP structures and dynamics is still lacking, which makes a rational LNP design almost impossible. In this work, we present an extension of CHARMM-GUI Membrane Builder to model and simulate all-atom LNPs with various (ionizable) cationic lipids and PEGylated lipids (PEG-lipids). These new lipid types can be mixed with any existing lipid types with or without a biomolecule of interest, and the generated systems can be simulated using various molecular dynamics engines. As a first illustration, we considered model LNP membranes with DLin-KC2-DMA (KC2) or DLin-MC3-DMA (MC3) without PEG-lipids. The results from these model membranes are consistent with those from the two previous studies albeit with mild accumulation of neutral MC3 in the bilayer center. To demonstrate Membrane Builders capability of building a realistic LNP patch, we generated KC2- or MC3-containing LNP membranes with high concentrations of cholesterol and ionizable cationic lipids together with 2 mol% PEG-lipids. We observe that PEG-chains are flexible, which can be more preferentially extended laterally in the presence of cationic lipids due to the attractive interactions between their head groups and PEG oxygen. The presence of PEG-lipids also relaxes the lateral packing in LNP membranes, and the area compressibility modulus (KA) of LNP membranes with cationic lipids fit into typical KA of fluid-phase membranes. Interestingly, the interactions between PEG oxygen and head group of ionizable cationic lipids induce a negative curvature. We hope that this LNP capability in Membrane Builder can be useful to better characterize various LNPs with or without genetic drugs for a rational LNP design.