The intersection of artificial intelligence and biomedicine has yielded a novel approach to designing molecular interactions, as scientists at the Laboratory of Protein Design and Immunoengineering have successfully harnessed deep learning to create new proteins that bind to complexes involving small molecules like hormones or drugs.
By leveraging their proprietary MaSIF pipeline, researchers can rapidly scan millions of proteins to identify optimal matches based on chemical and geometric surface properties, effectively engineering novel protein-protein interactions that play a crucial role in cell regulation and therapeutics. This technological advancement has far-reaching implications for developing targeted therapies, including the potential to fine-tune the activity of cell-based cancer treatments like CAR-T therapy, thereby enhancing their safety and efficacy.
By designing proteins that can interact with specific protein-drug complexes, scientists may be able to create “on” or “off” switches for cellular functions, paving the way for more precise control over therapeutic outcomes.
Introduction to Deep Learning in Protein Design
The field of protein design has witnessed significant advancements in recent years, particularly with the integration of deep learning techniques. Researchers at the Laboratory of Protein Design and Immunoengineering (LPDI) have made notable contributions to this area, including the development of a deep-learning pipeline for designing new proteins that interact with therapeutic targets. This pipeline, known as MaSIF, has been shown to rapidly scan millions of proteins to identify optimal matches between molecules based on their chemical and geometric surface properties.
The LPDI team, led by Bruno Correia, has published several studies demonstrating the capabilities of MaSIF in designing novel protein-protein interactions that play key roles in cell regulation and therapeutics. One of the recent advancements of this technology involves using MaSIF to design novel protein binders that interact with known protein complexes involving small molecules like therapeutic drugs or hormones. These bound small molecules can induce subtle changes in the surface properties of these protein-drug complexes, acting as “on” or “off” switches for the fine control of cellular functions.
The concept of designing proteins that interact with specific targets is complex and involves a deep understanding of protein structure and function. Proteins are made up of amino acids, which are linked together to form a polypeptide chain. This chain then folds into a specific three-dimensional structure, determining the protein’s function. The surface properties of a protein, including its shape, charge, and hydrophobicity, play a crucial role in its interactions with other molecules. MaSIF’s ability to capture these surface features and predict optimal matches between proteins and small molecules has significant implications for the design of novel therapeutics.
Mechanism of MaSIF
MaSIF works by generating “fingerprints” for surface features like positive and negative charge, hydrophobicity, shape, etc. These fingerprints are then used to identify complementary surfaces from a database, allowing researchers to digitally graft protein fragments onto larger scaffolds and select binders predicted to fit best with their targets. The difference in the recent study is that the team assumes the surface features of a protein change if a small molecule binds to it, creating a neosurface. MaSIF was able to capture this difference with a high degree of sensitivity.
The team experimentally validated their novel protein binders against three drug-bound protein complexes containing the hormone progesterone, the FDA-approved leukemia drug Venetoclax, and the naturally occurring antibiotic Actinonin, respectively. The protein binders designed using MaSIF successfully recognized each drug-protein complex with high affinity. This was possible because MaSIF is based on general surface features that apply to proteins and small molecules alike, allowing researchers to map the small molecule features onto the same descriptor space that MaSIF was trained on for proteins.
Applications of MaSIF
One exciting potential application of this work is the fine control of cell-based cancer treatments like chimeric antigen receptor (CAR-T) therapy. CAR-T therapy involves engineering a patient’s T cells to target their cancer better, but after being re-introduced into the patient, engineered cells may attack the wrong targets or exhaust their ability to fight cancer. In a proof-of-concept experiment, the EPFL team showed that a Venetoclax-inducible system designed with MaSIF effectively switched on tumor-killing activity of CAR-T cells in vitro.
The ability to precisely control the spatiotemporal activity of cell-based therapies with small molecule switches could significantly improve the safety and efficacy of these treatments. This work has significant implications for developing novel cancer therapies and highlights the potential of deep learning techniques in protein design.
External Link: Click Here For More
