De novo proteins to degrade unwanted proteins

The 2024 Nobel Prize in Chemistry recognized Dr. David Baker for his pioneering work in computational protein design, along with Drs. Demis Hassabis and John Jumper for their contributions to protein structure prediction. Dr. Baker, a professor at the University of Washington and a Cancer Consortium member, has spent decades developing computational algorithms that allow scientists to design novel proteins—proteins not found in nature. His work focuses on how proteins acquire their 3D shapes (and, by extension, their functions). This ability to design customized proteins ‘from scratch’ unlocks the potential to create highly targeted pharmaceuticals, vaccines, nanomaterials, and more.

To design novel proteins, Baker developed the Rosetta software starting in 1998. This tool assembles fragments from diverse protein structures with similar local sequences in the Protein Data Bank, simultaneously optimizing the protein’s sequence and structure. Without Rosetta, determining a protein’s structure could take years, but with this computational approach, researchers can achieve results far faster and with remarkable precision. It is worth noting that when Rosetta was first developed, its accuracy was fairly poor -it is only recently that these algorithms were accurate enough to begin taking the place of experimental structure determination. 

Now, let’s look at the real-world applications of Baker’s work, starting with Endocytosis-targeting proteins, or EndoTags. These proteins are engineered to enhance endocytosis, a cellular process where cells internalize extracellular  molecules, which are then broken down via the lysosome pathway. EndoTags are designed to bind receptor sites without interfering with the receptor’s natural ligands, “hijacking” the cell’s internal pathways to promote degradation of targeted proteins. The results were recently published in Nature. 

To test EndoTags, the team explored endocytic receptors with different tissue expression and downstream mechanisms. These targets included (1) Insulin-like Growth Factor 2 Receptor (IGF2R), which is found in most tissues and triggers endocytosis through receptor dimerization, (2) Asialoglycoprotein Receptor (ASGPR), which is primarily expressed in the liver and requires clustering for endocytosis, (3) Transferrin Receptor (TfR), which is expressed in the brain, liver, and muscles, and cycles between the cell surface and internal compartments, and (4) Sortilin, which is found in the central nervous system and similarly cycles through the cell’s internal structures.

Notably, each receptor posed a unique design challenge. For receptors like IGF2R, Baker’s team used Rosetta RIFdock to create a synthetic protein that binds to two IGF2R sites, promoting dimerization and triggering endocytosis. For ASGPR, clustering is needed to initiate internalization. Here, the team combined Rosetta ProteinMPNN and AlphaFold2, designing binders that mimic receptor clustering to stimulate endocytosis. For Sortilin and TfR, the team used Rossetta de novo binder design software to designed synthetic proteins that “ bind to the receptor at a site that does not compete for the natural ligand, which could have undesired side effects and reduce efficiency.” Gratifyingly, endocytosis was triggered using the EndoTags among all the receptors.  

Building on EndoTags, Baker’s team designed Lysosome Targeting Chimeras (pLYTACs). These synthetic proteins harness the lysosomal machinery to degrade disease-related proteins. As a proof of concept, they targeted EGFR, a protein often overexpressed in cancers, achieving an 85% reduction in EGFR levels. In this case, pLYTACs was composed of an EGFR binding protein fused with one of EndoTags. They also targeted PD-L1— an immune checkpoint protein that helps cancer cells evade immune detection. In this case, the EndoTag was fused with an anti-PD-L1 antibody-, resulting in a 77% decrease in PD-L1 within 48 hours in cells treated with this pLYTAC. 

To achieve higher specificity, the team combined EndoTags with Co-LOCKR, a technology that acts like an "AND" logic gate. This system activates only when two specific markers are present on a cell. They applied this technology to degrade EGFR only in cells also expressing HER, which is common in cancer cells but rare in healthy tissue. This approach reduced EGFR levels by around 80% in cells with both markers, sparing those with only one marker and minimizing potential side effects.

Going forward, the team aims to expand the range of EndoTag targets, potentially advancing treatments for cancer, autoimmune diseases, and neurological disorders. EndoTags may also prove invaluable in studying complex cellular processes, shedding light on the mechanisms underlying various diseases. “We aim to expand the range of targets for EndoTags, potentially enabling therapies for more diseases. We also plan to explore clinical applications of pLYTACs, with hopes of developing new treatments for cancer, autoimmune, and neurological disorders,” the authors commented. 

The spotlighted worked was funded by the Department of Defense, the Defense Threat Reduction Agency, the DARPA Synergistic Discovery and Design, the National Institutes of Health, the Audacious Project at the Institute for Protein Design,  the Nordstrom Barrier Institute for Protein Design Directors Fund, the Open Philanthropy Project Improving Protein Design Fund, the E. Schmidt, W. Schmidt and Schmidt Futures funding, the European Molecular Biology Organization, the Jane Coffin Childs Memorial Fund for Medical Research, a National Science Foundation Graduate Research Fellowship and the Stanford Center for Molecular Analysis and Design. 

Fred Hutch/University of Washington/Seattle Children's Cancer Consortium member Dr. David Baker contributed to this work.

Huang, B. et al., Designed endocytosis-inducing proteins degrade targets and amplify signals. Nature. 2024 Sep 25. Epub ahead of print.