Development of D-box peptides to inhibit the anaphase-promoting complex/cyclosome

Abstract

E3 ubiquitin ligases engage their substrates via 'degrons' - short linear motifs typically located within intrinsically disordered regions of substrates. As these enzymes are large, multi-subunit complexes that generally lack natural small-molecule ligands and are difficult to inhibit via conventional means, alternative strategies are needed to target them in diseases, and peptide-based inhibitors derived from degrons represent a promising approach. Here we explore peptide inhibitors of Cdc20, a substrate-recognition subunit and activator of the E3 ubiquitin ligase the anaphase-promoting complex/cyclosome (APC/C) that is essential in mitosis and consequently of interest as an anti-cancer target. APC/C engages substrates via degrons that include the 'destruction box' (D-box) motif. We used a rational design approach to construct binders containing unnatural amino acids aimed at better filling a hydrophobic pocket that contributes to the D-box binding site on the surface of Cdc20. We confirmed binding by thermal-shift assays and surface plasmon resonance and determined the structures of a number of the Cdc20-peptide complexes. Using a cellular thermal shift assay, we confirmed that the D-box peptides also bind to and stabilise Cdc20 in the cell. We found that the D-box peptides inhibit ubiquitination activity of APC/C^Cdc20 and are more potent than the small-molecule inhibitor Apcin. Lastly, these peptides function as portable degrons capable of driving the degradation of a fused fluorescent protein. Interestingly, we find that although inhibitory activity of the peptides correlates with Cdc20-binding affinity, degradation efficacy does not, which may be due to the complex nature of APC/C regulation and effects of degron binding of subunit recruitment and conformational changes. Our study lays the groundwork for the further development of these peptides as molecular therapeutics for blocking APC/C as well as potentially for harnessing the APC/C for targeted protein degradation.

Keywords: APC/C; E. coli; anaphase-promoting complex/cyclosome; biochemistry; chemical biology; degron; human; peptide inhibitor; protein–protein interaction; ubiquitin.

Figure 1.. Structure and function of anaphase-promoting complex/cyclosome (APC/C).

(A) Schematic of APC/C activity during mitotic exit, indicating the switch in co-activator from Cdc20 to FZR1. Most substrates contain variable degrons (D-box in green, KEN in yellow) present in IDRs. (B) Domain structuring of Cdc20 comprising an N-terminal IDR with the C-box, KEN-box, and CRY-box motifs, the central WD40 domain responsible for substrate recruitment via the degron binding sites and the C-terminal IDR containing the IR-tail. (C) Schematic of the structure of the Cdc20 WD40 domain (PDB: 4GGC) overlaid with those of the WD40 domain in complex with Acm1 D-box and ABBA motif peptides (PDB: 4BH6) (He et al., 2013) and the KEN-box peptide 4GGD (Tian et al., 2012).

Figure 2.. Biophysical characterisation of Apcin binding to Cdc20^WD40 by thermal shift assay (TSA) and surface plasmon resonance (SPR).

(A) Representative examples of thermal unfolding traces of Cdc20 ^WD40 in the presence of 1% DMSO as the vehicle control or Apcin at concentrations of 25, 50, and 100 µM. (B) Corresponding melting temperatures calculated from derivative plots of the thermal unfolding traces. Mean data from triplicate measurements are shown, with error bars representing standard deviations. (C) Reference-subtracted sensorgrams of biotinylated Cdc20^WD40 and Apcin. (D) Binding affinity determination of Apcin to Cdc20^WD40 domain by steady-state analysis of the sensorgrams.

Figure 2—figure supplement 1.. Verification of single-site biotinylation of purified Cdc20 WD40 domain (residues 161-477) using Sulfo-NHS-LC-LC-Biotin (Thermo Scientific, A35358).

The observed mass shift corresponds to one Sulfo-NHS-LC-LC-Biotin molecule added to the purified protein.

Figure 3.. D-box peptide mutations.

(A) Schematic showing the Acm1 D-box peptide bound to yeast FZR1 homologue Cdh1. R119 of the D-box forms H-bond interactions with D256 and E537 of Cdh1. L122 of the D-box buries into the canonical pocket on the surface of Cdh1 (PDB: 4BH6, He et al., 2013). (B) Melting temperature of Cdc20^WD40 in the presence of D-box peptides at 25, 50, and 100 µM concentrations, calculated from derivative plots of the thermal unfolding traces. Mean data from triplicate measurements are shown, with error bars representing standard deviations.

Figure 3—figure supplement 1.. Reference-subtracted surface plasmon resonance (SPR) sensorgrams and binding curves for (A) D1, (B) D3, (C) D4, (D) D5, (E) D10, (F) D19 binding to Cdc20.

Figure 3—figure supplement 2.. Linear amino acid sequence and chemical structures for peptides D1 and D2 are illustrated alongside the expected exact mass and molecular weights for each peptide.

LCMS data showing the m/z species for each purified peptide. Analytical HPLC chromatograms (10–30% solvent B over 15 minutes, 1 mL l/min, UV trace at 220 nm), with calculated purity of the peptides by the percentage area.

Figure 3—figure supplement 3.. Linear amino acid sequence and chemical structures for peptides D3 and D4 are illustrated alongside the expected exact mass and molecular weights for each peptide.

LCMS data showing the m/z species for each purified peptide. Analytical HPLC chromatograms (10–30% solvent B over 15 minutes, 1 mL l/min, UV trace at 220 nm), with calculated purity of the peptides by the percentage area.

Figure 3—figure supplement 4.. Linear amino acid sequence and chemical structures for peptides D5, D10 and D19 are illustrated alongside the expected exact mass and molecular weights for each peptide.

LCMS data showing the m/z species for each purified peptide. Analytical HPLC chromatograms (10–30% solvent B over 15 minutes, 1 mL l/min, UV trace at 220 nm), with calculated purity of the peptides by the percentage area.

Figure 4.. D-box peptides incorporating unnatural amino acids.

(A) Schematics of the two unnatural amino acids used. (B) Thermal stabilisation of the Cdc20^WD40 by the two highest affinity peptides D20 and D21 calculated from derivative plots in thermal shift assay (TSA) . Surface plasmon resonance (SPR) reference-subtracted sensorgrams and binding curves for (C) D20 and (D) D21.

Figure 4—figure supplement 1.. Linear amino acid sequence and chemical structures for peptides D7 and D12 are illustrated alongside the expected exact mass and molecular weights for each peptide.

LCMS data showing the m/z species for each purified peptide. Analytical HPLC chromatograms (10–30% solvent B over 15 minutes, 1 mL l/min, UV trace at 220 nm), with calculated purity of the peptides by the percentage area.

Figure 4—figure supplement 2.. Linear amino acid sequence and chemical structures for peptides D20 and D21 are illustrated alongside the expected exact mass and molecular weights for each peptide.

LCMS data showing the m/z species for each purified peptide. Analytical HPLC chromatograms (10–30% solvent B over 15 minutes, 1 mL l/min, UV trace at 220 nm), with calculated purity of the peptides by the percentage area.

Figure 5.. Crystal structures of Cdc20-D-box complexes.

X-ray crystal structures of peptides (A) D21, (B) D20, and (C) D7 bound to the canonical D-box binding pocket of Cdc20. Intermolecular hydrogen bonds between peptides and Cdc20 are shown by dashed lines. (D) Structural alignment of D21-bound Cdc20 and Acm1 D-box peptide bound to Cdh1 (PDB: 4BH6; He et al., 2013). Peptide backbones align with an RMSD of 1.007 Å. Modelled water molecules have been removed from images for clarity.

Figure 5—figure supplement 1.. Cdc20^WD40 is shown in cartoon (cyan) and surface (grey) representations.

All peptides are shown by stick representations (magenta). Intramolecular hydrogen bonds formed within peptides (A) D7, (B), D20, and (C) D21. All peptides form a hydrogen bond between the carbonyl of A2 to the amine of G5/S5. Peptides D20 and D21 form an additional H-bond between the carbonyl of A2 to the hydroxyl group of S5. 2F­_oF_c maps contoured at 1.0σ and modelled peptide atoms of (D) D7, (E), D20, and (F) D21. Unbiased F_oF_c maps from refinement steps prior to including the peptides in the model, contoured at 2.5σ. Maps shown are (G) D7, (H) D20, and (I) D21.

Figure 6.. D-box peptides bind to full-length HiBiT-tagged Cdc20 in the cellular context.

Representative cellular thermal shift assay (CETSA) data are shown for Cdc20-tranfected HEK293T cell lysates incubated with D-box peptides at a concentration of 100 µM.

Figure 6—figure supplement 1.. CETSA method development and validation using Apcin as a positive control.

(A) Cellular thermal shift assay (CETSA) of endogenous Cdc20 in HEK293T cell lysates by 100 µM Apcin compared with vehicle control (1% DMSO), analysed by densitometric analysis of western blots. Mean and standard deviation are calculated from two independent experiments. (B) CETSA of transfected Cdc20 with a C-terminal HiBiT tag spiked with 1% DMSO with and without including a centrifugation step following heat denaturation at each temperature set point. (C) Stabilisation of transfected Cdc20 with a C-terminal HiBiT tag by 100 µM Apcin compared with vehicle control (1% DMSO).

Figure 7.. Inhibition of APC/C^Cdc20-mediated ubiquitination of Cyclin B1 by D-box peptides and Apcin.

In vitro ubiquitination assays using reconstituted APC/C^Cdc20 with Cyclin B1 as the substrate for ubiquitination. Lead peptides and Apcin were titrated from 3 µM to 300 µM and showed concentration-dependent inhibition of Cyclin B1 ubiquitination compared to the vehicle control (0.7% DMSO).

Figure 8.. D-box variants can drive degradation in mitotic cells.

(A) Schematic of D-box-mNeon constructs used in fluorescence timelapse imaging. (B) mNeon fluorescence levels in individual cells plotted over time to show D-box mediated degradation of mNeon in mitosis. Fluorescence measurements from individual cells are normalised to fluorescence at metaphase then in silico synchronised to anaphase onset. Mean degradation curves are shown, with error bars representing SDs. (C) Degradation rate curves show rate of change in relative fluorescence of the D-box variants and reveal maximum degradation rate for each construct. Error bars are depicted as shaded regions and indicate SDs. (D) Levels of relative fluorescence in each cell at t=1 hour after anaphase onset. Degradation of each D-box construct was significant relative to D0 control, using Welch’s t-test. ****p≤0.0001. In (B–D), n=D0 (20) D1 (23), D2 (40), D3 (38), and D19 (34), with data pooled from two or more independent experiments.