The contribution of native protein complexes to targeted protein degradation

Abstract

Targeted protein degradation (TPD) destroys proteins of interest (POIs) by hijacking the cellular proteolytic machinery. Most proteins in cells exist and function as part of multi-protein or macromolecular complexes, thereby allowing a single protein to control multiple biological processes. Therefore, when a small molecule degrader induces proximity between an E3 ligase and the POI, the macromolecular context of the POI potentially influences the degradation outcomes of the POI and of the complex components. Here, we explore degradation of the eight CK1α-SACK1(A-H) (formerly known as FAM83A-H) complexes initiated by molecular glue degraders primarily designed to target Ser/Thr kinase CK1α. We demonstrate that lenalidomide-derived degraders DEG-77 and SJ3149, which selectively target the CK1α isoform, co-degrade multiple SACK1(A-H) proteins. We show that the degradation of SACK1(A-H) proteins by DEG-77 and SJ3149 requires CK1α, the CUL4A^CRBN E3 ligase complex and the proteasome. In cells derived from palmoplantar keratoderma patients harbouring the CK1α-binding deficient SACK1G^R265P mutation, DEG-77 targets CK1α and mitotic SACK1D but not SACK1G^R265P, highlighting the requirement for CK1α-SACK1(A-H) interaction to achieve co-degradation. Our study underscores the importance of POI context in TPD and reinforces the potential for selectively targeting specific protein complexes for degradation.

Figure 1:. SACK1 domain-containing proteins form complexes with CK1α at the endogenous level in DLD1, U2OS, A549 and TOV-21G cell lines.

(A) Depiction of subcellular distribution of serine/threonine protein kinase CK1α and interacting SACK1 domain-containing proteins. SACK1(A-H) direct CK1α and other CK1 isoforms to distinct subcellular compartments to regulate CK1 biology. (B&C) DLD1, U2OS, A549 (B) and TOV-21G (C) cells were lysed, and extracts normalised to 2 mg total protein and subjected to anti-CK1α IPs. IgG IPs were used as negative controls. Input extracts and IPs were resolved by SDS-PAGE, transferred to PVDF membrane and analysed by immunoblotting using the indicated antibodies.

Figure 2:. DEG-series of CK1α degraders lead to co-depletion of CK1α-SACK1(A-H) complexes to varying degrees.

(A) Wild-type DLD1 cells were incubated with DMSO or 100 nM of each DEG compound (−35, −54, −52, −48, −64, −61, −77) as indicated for 24 h before cell lysis. Extracts were resolved by SDS-PAGE, transferred to PVDF membrane and immunoblots were incubated with the indicated antibodies. (B) Wild-type DLD1, U2OS and A549 cell lines were incubated with increasing concentrations of DEG-77 (1-1000 nM) or 10 μM lenalidomide for 24 h before cell lysis. 20 μg extract protein was resolved by SDS-PAGE and transferred to PVDF membranes for immunoblotting analysis using the indicated antibodies. (C) As in (B), except TOV-21G cells were cultured for 24 h with increasing concentrations of DEG-77 (1-1000 nM) before cell lysis. (D) As in (B), except wild-type DLD1, U2OS, and A549 cells were incubated with 100 nM DEG-77 for different time periods (0.5 - 24 h) before cell lysis. (E) As in (C), except TOV-21G cells were incubated with 100 nM DEG-77 for different time periods (0.5 - 24 h) before cell lysis. All blots are representative of at least 3 biological replicates.

Figure 3:. SJ-series of CK1α degraders lead to co-depletion of CK1α-SACK1(A-H) complexes.

(A) Wild-type DLD1 cells were treated with SJ compounds (SJ7095, SJ0040, SJ3149) at increasing concentrations from 1-1000 nM for 4 h before cell lysis. (B) Wild-type DLD1, U2OS and A549 cell lines were incubated with increasing concentrations of SJ3149 (1-1000 nM) or 10 μM lenalidomide for 24 hours before cell lysis. 20 μg extract protein was resolved by SDS-PAGE and transferred to PVDF membranes for immunoblotting analysis using the indicated antibodies. (C) As in (B) except wild-type DLD1, U2OS, and A549 cells were incubated with 100 nM SJ3149 for different time periods (0.5 - 24 h) as indicated before cell lysis. All blots are representative of at least 3 biological replicates.

Figure 4:. Comparison of CK1α degraders for their ability to co-deplete the interacting SACK1 proteins.

(A) Wild-type DLD1 cells were incubated with different imide-derived CK1α degraders: lenalidomide (10 μM), BTX161 (10 μM), DEG-48 (100 nM), DEG-77 (100 nM), SJ7095 (1 μM) and SJ3149 (100 nM) or DMSO control for 24 h before cell lysis. 20 μg extract protein was resolved by SDS-PAGE, transferred to PVDF membrane and analysed by immunoblotting using the indicated antibodies. Blots are representative of at least 3 biological replicates. (B) Quantification by densitometry of protein signals shown in (A) using Fiji 1.53q (ImageJ). Each graph depicts the abundance of CK1α or SACK1(A-H) proteins normalised to GAPDH loading control relative to DMSO treatment control (n = 3, error bars represent mean ± SD).

Figure 5:. The degradation of SACK1 proteins by imides DEG-77 and SJ3149 requires CK1α, CRBN and the proteasome.

(A) The postulated mechanism of action by which imide-derived molecular glue degraders (MGDs) DEG-77 and SJ3149 induce CK1α-SACK1(A-H) co-degradation. By binding to substrate receptor cereblon (CRBN), both MGDs primarily recruit CK1α to CUL4^CRBN E3 ligases and, as CK1α exists in complex with SACK1(A-H) proteins, it co-recruits SACK1(A-H) proteins to the CUL4^CRBN E3 ligase complex. Depending on the proximity of the recruited CK1α-SACK1(A-H) complex to the CUL4^CRBN E3 catalytic site, specific CK1α-SACK1(A-H) complexes are potentially ubiquitylated and targeted for degradation via the proteasome. (B&C) Wild-type, CRBN^−/− and CSNK1A1^−/− DLD1 cells were incubated with 100 nM DEG-77 or DMSO (B) or 100 nM SJ3149 or DMSO (C) for 6 h before cell lysis. 20 μg extract protein was resolved by SDS-PAGE and transferred to PVDF membrane for immunoblotting analysis. (D&E) As in (B), except wild-type DLD1 cells were pre-treated with either MG132 (20 μM) or bafilomycin A1 (50 nM) for 2 hours before treatment with DEG-77 (100 nM) or DMSO (D) or SJ3149 (100 nM) or DMSO (E) for a further 4 h before lysis. For (B-E), blots are representative of at least 3 biological replicates.

Figure 6:. CK1α-SACK1(A-H) co-degradation mediated by DEG-77 and SJ3149 requires a direct interaction between CK1α and SACK1(A-H) proteins.

(A) Depiction of CK1α and SACK1G interaction that drives WNT signalling in normal cells (as in control fibroblasts, FIB03) and how this becomes disrupted with the pathogenic SACK1G^R265P mutation identified in a patient with palmoplantar keratoderma (as in PPK patient fibroblasts, FIB04). (B) Patient-derived fibroblast cells (FIB04 represents the PPK patient cells harbouring the pathogenic SACK1G^R265P variant while FIB03 represents an age and sex-matched control cell line) were treated with 100 nM of DEG-35, DEG-48, DEG-77 or DMSO for 24 h before cell lysis. 10 μg extract protein was resolved by SDS-PAGE and transferred to PVDF membrane for immunoblotting analysis. (C) SACK1G^−/− DLD1 cells and those retrovirally transduced to stably express either SACK1G-GFP or SACK1G^R265P-GFP were treated with DEG-77 (100 nM) or DMSO for 6 h before lysis. 20 μg extract protein was resolved by SDS-PAGE and transferred to PVDF membrane for immunoblotting analysis. For (B&C), blots are representative of at least 3 biological replicates.