An Intracellular Helix-constrained Peptide Library Screening Platform to Derive Functional Transcription Factor Antagonists

>60% of all multi-protein complexes feature helical interfaces, with >20% participating in gene regulation. Helical interaction inhibitors therefore have enormous potential to become a useful class of transcriptional modulator. Small molecules typically fail to abrogate these interactions owing to their limited ability to extend beyond interaction hotspots. In contrast, peptide-based inhibitors can block much larger interaction areas, making them a preferred approach for protein-protein interaction (PPI) inhibition.

However, peptide sequences corresponding to binding sites within a protein can lose their structure when created in isolation. For helical peptides, this can be compensated for with the use of 'helix-inducing constraints' which can be thought of as safety pins that hold the peptide together to retain helical structure. However, the beneficial effect of a constraint can be very hard to predict, leading to considerable time and expense. We will tackle this challenging problem by screening entire libraries of peptides inside cells where each members has been locked into a helical conformation. By constraining all peptides during the library screen we will identify those in which the constraint provides improved target affinity, resistance to proteases, and potentially the potential to cross biological membranes. We will ensure that antagonists are truly functional in blocking transcription factor-DNA binding by screening the helix-constrained libraries within living cells using our Transcription Block Survival (TBS) Assay. We will fully characterise peptides to establish how they acheive their favourble properties using a range of biochemical assays.

Finally, using knowledge gained we will create updated versions of our widely used isPCA/isCAN peptide library screening software that searches vast numbers of peptides to identify those most likely to bind with high affinity and selectivity to a given target (e.g. Chen et al, Nature 2019, Aupic et al, Nat Commun 2021, Daudey et al, Chem Sci 2021). Building on these we will further identify sequences within which 'modules' can be introduced that both tolerate constraints but that more importantly promote increased helicity, binding, biostability, and potentially membrane permeability. The software is anticipated to be widely used by the community to facilitate selective inhibiton of a wide range of disease relevant PPIs in which helical interfaces present. The computational tools will be made freely available and accessible online to the scientific community. We envisage that they will be widely adopted by protein biochemists, cell biologists and synthetic biologists.

The proposal focuses on a specific, well-defined protein-folding motif-the alpha-helical coiled coil, which forms highly specific interactions and is found in a wide range of disease relevant protein-protein interactions. The approach is synergistic and multidisciplinary: developing sequences and rules for peptides that can be constrained into alpha-helical conformations required for effective helical target engagement using two bZIP exemplars. Attention has recently turned to inhibiting disease-relevant PPIs using peptides. However, as helical peptides become progressively downsized to identify bioactive epitopes, they lose their intrinsic hydrogen-bond support network, leading them to become unstructured. We will address this using i-->i+4 and i-->i+7 helix-inducing constraints that covalently pin together residues one turn and two turns apart respectively on the solvent exposed face of an alpha helix.

The effect of constraint introduction is however challenging and highly unpredictable. To address this we will use two cell penetrant crosslinking agents to screen helix-constrained libraries within live cells. This will allow only the most appriate constraints that translate into improved target affinity to be selected. Moreover, we will screen using a transcription block survival assay to ensures that hits are functional antagonists. We will specifically target two structurally related yet unique oncogenic bZIP exemplars.

Finally, using knowledge gained we will create software that can identify candidate consensus sequences that are known to accept constraints that promote increased helicity, enhanced binding and potentially biostability, and membrane permeability. This will fast-track the search for sequences with desirable properties. The software will be very useful in in identifying inhibitors of a wide range of disease relevant PPIs in which helical interfaces present. The computational tools will be made freely available and accessible online.