https://www.nature.com/articles/s41467-023-42395-z

Recognition and reprogramming of E3 ubiquitin ligase surfaces by α-helical peptides

Key figures

• Figure 1: Defines the two-stage phage-display workflow that converts presenter-binding Helicons into focused trimerizer libraries and selects only target binders that require the presenter.
  
• Figure 2: Maps Helicon binding sites across HECT and Cullin-RING E3 ligase families, showing that α-helical peptides can occupy multiple functional E3 surfaces.
  
• Figure 3: Establishes CHIP and MDM2 as RING/U-box presenter proteins for downstream trimerizer discovery.
  
• Figure 4: Shows CHIP-dependent trimerizers that induce CHIP-TEAD4 and CHIP-PPIA ternary complexes and disrupt the YAP1-TEAD4 interaction in a CHIP-dependent manner.
  
• Figure 5: Demonstrates MDM2-dependent β-catenin recruitment by multiple Helicon clusters, including selectivity against MDM4 and SPR validation requiring all three components.
  
• Figure 6: Resolves residue-level ternary interfaces in MDM2-Helicon-β-catenin complexes, showing direct Helicon-target and presenter-target contacts.
  
• Supplementary Figure 5: Quantifies SPR-derived ternary complex affinities and percent complex formation for CHIP-TEAD4, CHIP-PPIA, and MDM2-β-catenin trimerizers.
  
• Supplementary Table 1: Lists the synthesized Helicon sequences, target assignments, cluster identities, staple modifications, and observed masses.
  

1) Thesis (one sentence)

To address the lack of general methods for inducing new PPIs without preexisting binders, in E3 ligase presenter systems, cysteine-stapled Helicon discovery and focused trimerizer library screening causes presenter-dependent ternary complex formation with new target proteins by fixing presenter-binding helix residues while diversifying the exposed face for cooperative target engagement, supported by phage-display enrichment, SPR/TR-FRET/FP biochemistry, and X-ray co-crystal structures.

2) Evidence card (three bullets only)

• Strongest result: (Fig. 5; Fig. 6) MDM2-focused trimerizer libraries yielded Helicons H330, H332, and H334 that induced MDM2-β-catenin complexes with direct FP EC50 values of 100 nM, 66 nM, and 67 nM, respectively, while X-ray structures of H330 and H332 showed distinct cooperative interfaces involving MDM2, Helicon, and β-catenin.
  
• Method enabler: (Fig. 1; Supplementary Table 2; research type + tools) The study used experimental platform development combining cysteine-stapled phage display, focused degenerate library design from sequence logos, next-generation sequencing, SPR, TR-FRET, FP, and crystallography to discover trimerizer peptides without requiring known binders to the target protein.
  
• Critical limitation: (Supplementary Fig. 2e; Supplementary Fig. 5b) Some screen hits can bind construct-specific or assay-biased surfaces, exemplified by CUL4B H316 binding a site judged non-physiological in the full CUL4 context and by CHIP-PPIA SPR responses of only 3% and 11% expected ternary complex formation for H326 and H328, likely due to immobilized PPIA/streptavidin steric hindrance.
  

Optional

Quote bank (2–4 short excerpts)

• Quote 1: “This limitation is particularly acute for E3 ubiquitin ligases, of which only a handful can be bound by small molecules.” (Introduction, page 1)
  
• Quote 2: “This method does not rely on rational design or previously known binding ligands, and does not require any preexisting structural information.” (Discussion, page 10)
  
• Quote 3: “Indeed, our efforts to directly discover trimerizers using fully naive screens have met with limited success to date.” (Discussion, page 11)
  

Key comparisons (1–3 lines)

• Compared to: Small-molecule molecular glues, PROTAC-like heterobifunctional molecules, and chimeric peptide controls.
  
• Win: Builds cooperativity from a presenter-binding peptide face and avoids needing preexisting target ligands; H321-H324 did not show the hook effect observed for chimeric P325 in CHIP-TEAD4 assays.
  
• Tradeoff: The paper establishes in vitro ternary-complex formation and structural mechanisms, but cellular penetration, ubiquitylation, and degradation remain downstream optimization steps.
  

Methods I might copy (protocol hooks)

• Construct design / Models: Naive phage libraries displayed ~10^8 14-mer cysteine-stapled Helicons; focused trimerizer libraries used 20-mer Helicons with conserved presenter-binding residues fixed and exposed residues fully or semi-randomized. Presenter/target constructs included CHIP23-303, CHIPTPR residues 23-154 or 24-154, MDM225-109 for screening, MDM217-111 and MDM414-111 for biochemical assays, TEAD4217-434, full-length PPIA residues 1-165, and β-catenin Armadillo domain residues 134-665.
  
• Conditions / Instruments: Trimerizer phage screening used bead-immobilized target protein with or without 10 μM soluble presenter; binding reactions were incubated 45 min at room temperature with rotation. Wash buffer contained 1X TBS, 1 mM MgCl2, 1% BSA, 0.1% Tween-20, 0.02% sodium azide, and 2% glycerol, with 10 μM presenter added during trimerizer washes. SPR used Biacore 8K or S200 instruments at 25 °C, typically in 1x HBS-P+ with 1% DMSO.
  
• Readout / Analysis: NGS used spike-in normalization, Phred score ≥18 filtering, Hit Strength as presenter-plus-target enrichment over target-alone controls, and hierarchical clustering with Hit Strength >5. CHIP-TEAD4 TR-FRET used 100 nM biotinylated TEAD4217-434, 150 nM Alexa Fluor 488-CHIP23-303, 2 nM terbium-streptavidin, 20 μL reactions, 11-point 3-fold Helicon dilution up to 20 μM, 60 min room-temperature incubation, and PheraStar reads at Ex 337 nm, Em 490/520 nm. SPR ABA ternary assays typically used 10 μM Helicon pre-equilibration, analyte serial dilutions, and Rmax-based percent ternary complex calculations.
  

Open questions / Theoretical implications (2–5 bullets)

• Can presenter proteins be treated as modular recognition scaffolds whose pre-bound interface is fixed while a second face is evolved or designed for a new target?
  
• How much weak binary affinity to the presenter or target is optimal before trimerizer behavior collapses into non-cooperative binding?
  
• Does the most productive trimerizer interface require direct presenter-target contacts, or can the peptide alone carry most of the interface energy?
  
• Can focused-library design from sequence logos provide a practical combinatorial shortcut for other induced-proximity systems where fully naive screens are too sparse?
  
• What design rules determine whether a ternary complex becomes merely stable in vitro versus productive for cellular function such as degradation or pathway inhibition?