Multivalent protein-DNA nanostructures as synthetic blocking antibodies

点击翻译按钮获取中文摘要

Project Summary Protein-protein interactions (PPIs) drive countless processes in biology. The ability to block these interactions with high specificity is crucial for probing the basic science of these processes, as well as for developing imaging agents or novel therapeutics. However, most traditional molecules for blocking protein-protein interactions—like small molecules, peptides, or antibodies—rely on the precise targeting of the crucial interface or binding pocket, which can be difficult for some targets. Furthermore, none of these approaches can be easily tuned to match the valency or size of the target, and binding to patches on the protein not directly involved in PPIs can fail to block activity. Here, we propose to develop a nanoscale synthetic antibody (“nano-synbody”) consisting of a tunable DNA nanostructure bearing 2-3 peptide/protein ligands that can bind to distinct surfaces of a target protein and block its association with its partner through the steric bulk of the DNA structure. The individual peptide/protein binding agents will be derived from either known molecules, or found independently through methods like phage display. Critically, our method merges computational simulation—and in silico “evolution”—of these hybrid protein-DNA nano-synbodies, creating a library of structures and probing their association with the target. We aim to create a feedback loop, whereby the computational simulations yield candidate nano-synbodies that can be experimentally tested, further informing the next round of simulations. We will first apply this pipeline to a homo-trimeric nano-synbody against the SARS-CoV-2 spike protein trimer (Aim 1). This test bed will allow us to optimize the process and find a high-affinity blocking structure. Then, we will apply our method to nano-synbodies for blocking the assembly of fibrinogen into fibrin clots (Aim 2). The second Aim will involve phage display against fibrin to find novel binding agents, and thereby convert them into high-affinity hetero-trivalent structures. In both Aims, we will demonstrate the advantage of nano-structuring ligand presentation over simple oligomerization with flexible linkers. Taken together, our work will generate a new computational-experimental paradigm for the design of tunable, user-defined nanostructures that can present three or more peptides/proteins for binding to any protein, and blocking its association with its target. Crucially, our approach does not require binding directly to the interface, which should enable it to target a much larger range of proteins that may not be amenable to traditional approaches, large protein complexes, or mutants/variants of the targets that might escape single binding agents.