Characterization of IWR1-POMA

(A) IWR1-POMA induced TNKS degradation with a DC50 value of 60 nM and reached a nearly complete depletion of TNKS1 at 1.2 µM in HAP1 cells. The dose-response curve was presented with 95% confidence interval (CI). (B and C) IWR1 induced massive TNKS accumulation while IWR1-POMA promoted deep degradation in a dose and time-dependent manner in DLD-1 cells. (D) IWR1-POMA promoted a more complete degradation of β-catenin than IWR1 in HEK293 cells cultured with Wnt3A conditioned media. (E and F) TNKS substrates (light blue) accumulated and WNT/β-catenin-controlled proteins (red) down-regulated in DLD-1 cells treated with IWR1-POMA (3 μM). AXIN2 (yellow) is both a TNKS substrate and a WNT target. GSPT1/2 and ZFP91 (light brown) are the only off-targets identified. Comparative analysis showed that IWR1-POMA controlled WNT/β-catenin signaling more effectively than IWR1.

IWR1-POMA degraded TNKS to suppress both catalysis-dependent and independent WNT signaling

(A) 293T-TNKS1/2-DKO cells were transfected with FLAG-TNKS1 and STF plasmids and then treated with IWR1 or IWR1-POMA. In contrast to IWR1 that plateaued in suppressing the WNT/β-catenin signaling induced by TNKS1, IWR1-POMA allowed for a more complete control of the pathway activity. (B) 293T-TNKS1/2-DKO cells transfected with FLAG-TNKS2 and STF plasmids and then treated with IWR1 or IWR1-POMA similarly showed that catalytic inhibition led to residual WNT/β-catenin activity. (C) 293T-TNKS1/2-DKO cells were transfected with 3×FLAG-TNKS1-PD and STF plasmids and then treated with IWR1 or IWR1-POMA. IWR1-POMA suppressed the luciferase activities whereas IWR1 had no effect. (D) 293T-TNKS1/2-DKO cells transfected with FLAG-TNKS2-M1054V and STF plasmids and then treated with IWR1 or IWR1-POMA. TNKS2 exhibited robust scaffolding effects. All the data is presented as mean ± SEM, n = 3.

TNKS is required for AXIN puncta formation

(A) AXIN colocalized with TNKS and formed puncta in 293T cells transfected with AXIN1-mCherry plasmid. (B) AXIN distributed diffusely throughout the cytoplasm in TNKS1/2-DKO cells transfected with AXIN1-mCherry plasmid. (C) Introduction of TNKS1 restored AXIN puncta in TNKS1/2-DKO cells transfected with AXIN1-mCherry and TNKS1 plasmids. (D) TNKS2 also induced puncta formation in TNKS1/2-DKO cells transfected with AXIN1-mCherry and TNKS2 plasmids. (E) The catalytic function of TNKS1 is not required for puncta formation as demonstrated in TNKS1/2-DKO cells transfected with AXIN1-mCherry and TNKS1-PD plasmids. (F) Catalytically inactive TNKS2-M1054V also promoted AXIN puncta formation effectively in TNKS1/2-DKO cells transfected with AXIN1-mCherry and TNKS2-M1054V plasmids.

Chemically induced TNKS accumulation promoted AXIN puncta formation

(A) SW480 cells treated with DMSO, IWR1 (5 μM) or IWR1-POMA (1 μM) and stained with anti-AXIN1 antibody. (B) 293T-AXIN1-dsRed-KI cells treated with DMSO, IWR1 (3 μM) or IWR1-POMA (3 μM). (C) HeLa cells were transfected with AXIN1-GFP plasmid and then treated with DMSO, IWR1 (3 µM) or IWR1-POMA (3 µM). IWR1 promoted the formation of micrometer-sized AXIN puncta whereas IWR1-POMA dissolved them.

TNKS accumulation impeded the degradation of β-catenin

(A) HeLa cells were transfected with AXIN1-mCherry and mNeonGreen-β-catenin plasmids and then treated with DMSO, IWR1 (3 µM) or IWR1-POMA (3 µM). IWR1 promoted large AXIN1 puncta formation and IWR1-POMA dissolved the puncta. β-Catenin colocalized with AXIN1, indicating the proper assembly of the DC. (B) FRAP analysis provided support to the hypothesis that TNKS controls the dynamic assembly of the DCs. IWR1 treatment led to a slow recovery of the AXIN1-mCherry fluorescent signal after photobleaching. In contrast, IWR1-POMA treatment did not affect the mobility of AXIN1. The fluorescent recovery curves were fitted to one-phase association model and presented with 95% CI. (C) Western blot analysis of samples corresponding to Figure 5A confirmed the accumulation of TNKS by IWR1 and the depletion TNKS by IWR1-POMA. The deeper suppression of the Wnt/β-catenin signaling by IWR1-POMA is also supported by the reduced levels of total β-catenin. (D) FRAP analysis indicated that IWR1 limited the turnover of β-catenin. In contrast, IWR1-POMA accelerated the diffusion rate of β-catenin in the DC.

IWR1-POMA suppressed CRC cell growth

(A) DLD-1 cells were developed into 3D spheroids and then treated with DMSO, IWR1 (5 µM) or IWR1-POMA (5 µM) for 10 d. The spheroids treated with IWR1-POMA lost the tight, spherical structure while IWR1 had no effect. (B) Immunostaining showed that the DLD-1 spheroids treated with IWR1-POMA only contained a small core of living cells while those treated with IWR1 remained highly proliferative. (C) IWR1-POMA (1 μM) suppressed the growth of PDM-7 patient-derived CRC organoids of approximately 200 μm in diameter with an apoptotic phenotype while IWR1 (1 μM) had no effect. (D) IWR1-POMA (1 μM) reduced the level of β-catenin and suppressed the proliferation of PDM-7 organoids grown from single cells significantly more effectively than IWR1 (1 μM).

Crystal structures used to guide the design of PROTAC molecules

(A) The crystal structure of TNKS1 with IWR1-exo (PDB 4OA7). (B) The crystal structure of TNKS1 with XAV939 (PDB 3UH4). (C) The crystal structure of TNKS2 with IWR1 (PDB 3UA9). (D) The crystal structure of TNKS2 with XAV939 (PDB 3KR8).

Identification of active PROTAC molecules using CRISPR engineered HAP1 cells expressing a TNKS1-NanoLuc fusion protein

(A) The chemical structures of the PROTAC molecules. (B) The schematic diagrams of the domain structures of TNKS1, TNKS2 and TNKS1-NanoLuc. (C) The relative abundance of the endogenous TNKS1 was measured by the luciferase activity upon treating HAP1-TNKS1-NanoLuc cells with IWR1-P(n)-VHL, IWR6-P(n)-VHL, IWR1-P(n)-Poma, IWR6-P(n)-Poma, or IWR1-TP(n)-Poma. IWR1-R-Olena bearing a rigid linker of length and polarity comparable to IWR1-P4-Poma and IWR1-P5-Poma alleviated the hook effect but was less effectively in promoting TNKS1 degradation. Removing the oxygen atom from the linker of IWR1-P1-Poma gave IWR1-C6-Poma with a more hydrophobic linker, but there was no improvement in the degradation efficacy.

Additional characterization of IWR1-POMA

(A) TNKS did not recover in DLD-1 cells at least 36 h after removing IWR1-POMA. (B) IWR1 (3 µM) and pomalidomide (3 µM) prevented the degradation of TNKS by IWR1-POMA (3 µM) in DLD-1 cells. (C) IWR1, pomalidomide, and MG132 blocked the degradation of TNKS by IWR1-POMA in HAP1-TNKS-NanoLuc cells. (D) IWR1-POMA promoted TNKS degradation while IWR1 induced TNKS accumulation in SW480, HT-29 and HeLa cells.

Proteomic analysis of DLD-1 cells treated with DMSO, IWR1 or IWR1-POMA

(A) Western blot analysis of samples corresponding to Figure 1 E and 1F confirmed the accumulation of TNKS1/2 by IWR1 and the depletion by IWR1-POMA. (B) Correlation analysis showed high reproducibility between the two biological repeats. (C) IWR1-POMA selectively degraded TNKS without inducing appreciable perturbations to 7 other PARP family member proteins and 79 NAD(P)-dependent enzymes detected in this proteomic experiment.

Degradation of TNKS by IWR1-POMA in 293T cells

(A) 293T cells were transfected with different doses of Wnt3A plasmid and then treated with DMSO, IWR1 (3 μM) or IWR1-POMA (3 μM). The cytosolic fraction of the cell lysates was then examined by Western blot. IWR1-POMA promoted a more complete degradation of β-catenin than IWR1 and reduced the level of the WNT target AXIN2. (B) Western blot analysis confirmed the lack of TNKS expression in TNKS1/2-DKO cells and validated the expression of FLAG-TNKS1, 3×FLAG-TNKS1-PD, FLAG-TNKS2 and FLAG-TNKS2-M1054V after transfection. (C) Western blot analysis of samples corresponding to Figure 2A confirmed the degradation of TNKS1 by IWR1-POMA. (D) Western blot analysis of samples corresponding to Figure 2B confirmed the degradation of TNKS2 by IWR1-POMA. (E) Western blot analysis of samples corresponding to Figure 2C confirmed the degradation of TNKS1-PD by IWR1-POMA. (F) Western blot analysis of samples corresponding to Figure 2D confirmed the degradation of TNKS2-M1054V by IWR1-POMA.

TNKSi induced the formation of AXIN puncta

(A) Puncta count of samples corresponding to Figure 4C. (B) HeLa cells transfected with AXIN1-GFP and STF plasmids and then treated with DMSO, IWR1 (3 µM) or IWR1-POMA (3 µM). IWR1-POMA suppressed WNT/β-catenin signaling significantly better than IWR1, indicating that AXIN puncta formation is not required for the DC to promote β-catenin degradation. The data is presented as mean ± SEM with p-values calculated by two-tailed unpaired t-test. (C) HeLa cells transfected with GFP-AXIN1 followed by treating with DMSO, IWR1 (3 µM) or IWR1-POMA (3 µM). Together with Figure 4C, this experiment shows that the position of GFP tag does not affect puncta formation. (D) HeLa cells transfected with GFP-AXIN2 followed by treating with DMSO, IWR1 (3 µM) or IWR1-POMA (3 µM). This experiment shows that AXIN2 also forms puncta upon IWR1 treatment.

IWR1-POMA suppressed CRC growth through WNT inhibition

(A) IWR1-POMA (3 μM) suppressed DLD-1 and SW480 colony formation. (B) IWR1-POMA (3 μM) suppressed of WNT signaling more effectively than IWR1 (3 μM) in DLD-1 and SW480 cells. The data is presented as mean ± SEM with p-values calculated by two-tailed unpaired t-test. (C) Peptide abundance of WNT targets, corresponding to Figure 1E and 1F. IWR1-POMA (3 μM) controlled several WNT targets not regulated by IWR1 (3 μM) in DLD-1 cells. (D) Western blot analysis confirmed that c-MYC, Aurora A, CDK4 and cyclin D1 responded to IWR1-POMA (3 μM) but not IWR1 (3 μM) in DLD-1 cells.

IWR1-POMA suppressed CRC proliferation through on-target WNT inhibition

(A) HCT116 cells carrying a mutation in β-catenin that could not be processed by the DC were resistant to both IWR1 (3 μM) and IWR1-POMA (3 μM). (B) IWR1-POMA (3 µM) reduced the cytosolic β-catenin level more effectively than IWR1 (3 µM) in 293T cells. The level of cytosolic β-catenin in 293T-TNKS1/2-DKO cells that lack both TNKS1 and TNKS2 did not change with either drug treatment. (C) CC-90009 induced GSPT1 degradation in DLD-1 cells but had no effect on TNKS. GSPT2 was not detectable by Western blot, which is consistent with the reported GSPT levels determined by quantitative proteomic analysis in this cell line—102,567 ppb for GSPT1 and 2,495 ppb for GSPT2 (https://www.ebi.ac.uk/gxa/experiments/E-PROT-18/Results)52. (D) DLD-1R cells obtained from cultivating DLD-1 cells with CC-90009 have a dramatically reduced level of GSPT1 expression meanwhile maintaining normal TNKS expression. (E) IWR1 (3 μM) promoted TNKS accumulation and IWR1-POMA (3 μM) induced TNKS degradation in DLD-1R cells. (F) IWR1-POMA prevented colony formation of DLD-1R cells deficient in GSPT1/2.

IWR1-POMA demonstrated efficacy in CRC spheroid and primary organoid models

(A) The growth chart of DLD-1 and SW480 spheroids treated with DMSO, IWR1 (5 μM), or IWR1-POMA (5 μM). The sizes represent the apparent dimensions of the spheroids including the peripheral dead cells. (B) IWR1-POMA (5 µM) induced apoptosis in DLD-1 spheroids. (C) PDM-7 organoids grown from single cells preserved the heterogeneous nature of CRC tumors. (D) IWR1-POMA (1 μM) prevented the formation of PDM-7 organoids from single cells while IWR1 (1 μM) did not. (E and F) IWR1-POMA (1 μM) suppressed proliferation and induced apoptosis in PDM-7 organoids grown from single cells while IWR1 (1 μM) had little effect.