Programmable Protein Stabilization with Language Model-Derived Peptide Guides

Introduction:: Over 20,000 proteins exist in the human proteome, and post-translational modifications (PTMs) including ubiquitination and phosphorylation further expand our toolbox in regulating critical processes that contribute to our overall health. Specifically, deubiquitination counteracts proteasomal degradation by removing ubiquitin protein chains, thus having several regulatory implications on DNA repair, cell differentiation, and folding catalysis processes. However, dysregulation of aberrantly altered proteins – such as improper degradation of tumor suppressors – can lead to several disease phenotypes (i.e., tumor survival, progression, and metastasis). Many of these proteins are considered “undruggable” by standard small molecule-based therapeutics (i.e., PROTACs, molecular glues, etc.) due to inaccessible binding pockets and disordered structure. While targeted protein degradation modalities have been developed, a counterpart “targeted protein stabilization” of key regulatory proteins represents a powerful tool to rescue specific proteins. As a more programmable, genetically-encoded strategy, designing functional peptide-guided stabilizers provides a unique opportunity for therapeutic interventions.

Materials and Methods:: We leverage protein language models, trained on millions of natural amino acid sequences, that takes an input protein sequence to conditionally generate binding peptides that capture diverse biochemical, structural, and functional features of a protein sequence alone (in other words, not needing protein structure). We experimentally link model-derived peptides to deubiquitinase (DUB) domains - specifically, a DUB fusion and a more minimal peptide-guided recruitment strategy - and systemically identify peptide-guided constructs that exhibit robust intracellular deubiquitination in vitro across diverse targets; we further show that stabilization of our targets show downstream functional effects, leading to a decrease in disease phenotypes (Figure 1, 2A).

Results, Conclusions, and Discussions:: Our stabilization platforms include (1) fusion of a potent DUB to a computationally-derived peptide, and (2) linkage of known peptide binders to the DUB USP11 to stabilize a protein via removal of ubiquitin and prevention of protein degradation. After testing different DUBs and peptide binders, we show that the DUB OTUB1 and binders FYLIR and LXLL show significant levels of protein stabilization in HEK293T cells (Figure 2B-C). Using a Structure-agnostic Language Transformer & Peptide Prioritization (SaLT&PepPr) model, we show statistically significant stabilization of the oncofusion protein EWS-FLI in our protein stabilization platforms, even in the presence of the natural E3 ubiquitin ligase TRIM8 (Figure 2B-C). Additionally, we demonstrate that our stabilizers induce apoptosis of Ewing’s sarcoma cells (A-673) when we stabilize endogenous EWS-FLI using several SaLT&PepPr-derived peptides in our DUB fusion and USP11-recruiting architectures (Figure 2D).
We seek to achieve CRISPR-like programmability for proteome-editing biologics – here, we develop a modular, genetically-encoded architecture for targeted protein stabilization by designing guide-like peptides linked to potent DUBs and target-recruiting peptide binders, enabling intracellular stabilization of various target proteins. We leverage SaLT&PepPr to derive binding peptides based on sequence only, enabling stabilization of even “undruggable” protein targets lacking previously accessible binding sites and disorderly structural characteristics. Our experimental results show statistically significant stabilization of our target expression, with clinical relevancy for modulation of downstream signaling pathways. Future work involves optimizing this architecture further for increased modularity and efficacy of stabilizing targets, as well as characterizing our architecture both computationally and experimentally for translational studies. Overall, our work expands accessibility of therapeutic interventions deemed untreatable by current small molecule-based therapeutic interventions.


Acknowledgements (Optional): :

References (Optional): :