Fig 1.
TNKSi treatment can induce axin expression and axin puncta associated with β-catenin degradation.
A. Western blot analysis of different cell lines treated +/- for 24 h with 2.5 μM of the TNKSi XAV939 revealed an increase in axin expression in different cell lines, and rescue of the degradation of β-catenin in SW480 CRC cells where the wnt pathway is disrupted by APC mutation. β-actin levels are shown as loading controls. Quantification of band intensities from two separate blots was performed and normalized to actin. The values shown are mean ± SD. B. SW480 cells were treated with 5 μM of the TNKSi G007-LK for 24 h, then immunostained with specific antibodies to detect different components of the β-catenin degradation complex located at axin puncta. The following combinations of proteins were co-stained: axin and total β-catenin, APC and Axin, GSK3β and Axin and phosphorylated and non-phosphorylated β-catenin. Cells were imaged using a DeltaVision microscope system. C. SW480 control cells or cells treated with 5 μM of the TNKSi G007-LK for 24 h were stained for fluorescence. In control cells a disperse staining pattern is observed whereas in TNKSi treated cells the TNKS-1 and -2 (red) specifically colocalize at TNKSi induced axin puncta (green) as shown by immunofluorescence microscopy.
Fig 2.
TNKS expression is required for TNKSi-dependent assembly of axin puncta.
The effect of silencing expression of TNKS1, TNKS2 or both (TNKS1/2) by siRNA was examined on TNKSi induced puncta formation in SW480 cells. A. Western blot and quantification showing the differences in TNKS expression after knockdown of the individual TNKSs. The band intensities represent the mean ± SD (normalized to actin) of two independent experiments. B. Immunofluorescence staining of SW480 cells untreated or treated for 24h with 5 μM of the TNKSi G007-LK. The individual knockdown of TNKS1 or TNKS2 did not significantly alter axin puncta formation, but the combined silencing of both TNKS1 and TNKS2 (siTNKS1/2) caused a clear reduction of cells showing axin and TNKS-positive puncta after G007-LK treatment, as shown by the cell images. C. A general scoring for axin puncta formation by microscopy revealed a ~40% decrease in the % cells showing visible axin puncta but only after silencing of both TNKS-1 and -2 (one-way anova bonferroni post-test, ** p<0.01 shows significance in comparison to the control). Only small differences were observed after knockdown of the individual TNKS. D. A more detailed comparison of the impact of TNKS expression on puncta was performed by scoring the range of puncta present per cell (categories: <5, 5–15 or >15 puncta). Here small reductions in the number of puncta per cell were seen with the individual TNKS siRNAs, but again only the double knockdown caused a significant reduction (two-way anova bonferroni post-test * p<0.05 and ** p<0.01 shows significance relative to control) in the number of puncta per cell.
Fig 3.
TNKSi induces the binding of TNKSs to axin.
A. SW480 cells were untreated or treated for 6 h or 24 h with 5 μM G007-LK (+/- 6 h with 20 μM MG132), then cell extracts were harvested and subjected to immunoprecipitation (IP) analysis with an axin antibody to detect axin-protein complexes. The immunoprecipitates were separated by SDS-PAGE and analysed by western blot. As shown, an increase in binding of both TNKS-1 and -2 to axin was observed after TNKSi treatments. In addition, an increase in β-catenin binding to axin was seen after 6 h treatment. In the presence of MG132, the induction of TNKS-axin complexes by G007-LK was blocked as shown. Quantification of axin-bound TNKS-1, TNKS-2 and β-catenin band intensities from two separate experiments are shown (mean ± SD). Values are corrected towards the total axin pulled down per sample. B. Western blot of total protein extract demonstrates that total levels of TNKSs were not modified by TNKSi treatment, but displayed a modest increase after 6 h MG132. The G007-LK treatment did modestly increase axin levels (and decrease β-catenin levels after 24 h treatment) as expected. This control blot indicates that the strong induction of TNKS-axin complexes above is not due to changes in protein expression. C. A diagram outlining how these novel effects of TNKSi on axin-TNKS complexes may contribute to axin/TNKS positive puncta formation in the cell.
Fig 4.
Induction of axin puncta by TNKSi is proteasome-dependent.
A. SW480 cells were treated with single or combined doses of tankyrase (5 μM G007LK) and proteasome (20 μM MG132) inhibitors. The MG132 was added for 6 h, either simultaneously with G007-LK for a 6 h treatment, or during the last 6 h of a 24 h G007-LK treatment. As shown in the immunofluorescence images, the addition of MG132 caused a reduction in the % of cells with visible TNKSi-induced axin puncta after 6 h treatment. The 6 h MG132 treatment also caused a modest increase in nuclear staining of axin. The later addition of MG132 (at the end of a 24 h G007-LK treatment) caused the induced axin puncta to relocate to the perinuclear region. B. The number of axin puncta per cell were scored by microscopy and categorized (<5, 5–15 or >15 puncta per cell). As shown, TNKSi induced a high number of puncta per cell (more than 80% of cells scored >15 puncta per cell) and this was blocked at 6 h or reduced after 24 h by MG132 treatment. C. A diagram summarizing the effect of G007-LK +/- MG132 on axin pattern and cellular distribution (shown in green).
Fig 5.
Tankyrase inhibitors promote inclusion of axin and TNKS into insoluble complexes.
A. NIH 3T3 cells were left untreated (- CSK) or washed for 5 min with 0.2% Triton X-100 containing MT-buffer (+CSK) to permeabilize the plasma membrane and remove soluble proteins, before immunolabeling with fluorescent antibodies for axin and tubulin. Axin puncta in untreated cells and after induction by TNKSi were still visible after the detergent extraction, despite a decrease in background staining. This implies that the axin puncta are part of an insoluble pool resistant to extraction. B. To determine if axin and TNKS accrued in an insoluble fraction in vitro after TNKSi treatment in SW480 cells, cells were extracted either with RIPA buffer (left panel) or enriched for the insoluble fraction by extracting cells on the plate with SDS containing sample buffer (right panel). The addition of TNKSi (G007-LK) increased the accumulation of the TNKS-1 and -2 in the insoluble fractions and this was at least partly reduced for TNKS2 and axin following MG132 treatment, indicating that TNKSi might reduce solubility and mobility of axin and TNKS. Quantification of band intensity was measured and values were corrected to the endogenous control β-actin (a second experiment is shown in S6 Fig).
Fig 6.
Proteasome inhibitor induces PARylation of TNKSs and TNKSi reduces basal PARylation levels.
A. SW480 and HEK293 cells were exposed to combinations of TNKSi (G007-LK) and MG132 treatments similar to that shown in legend to Fig 3. The cells were then lysed and extracts analysed by IP assay using a specific antibody against PAR. PARylated forms of TNKS1/2 were then detected by western blot. The IgG control was negative. Pull-down of PARylated proteins revealed an increase in PARylation of TNKS1 by MG132 treatment, and a decrease of PARylation induced by TNKSi. When combined, the MG132 dominated and caused some induction of PARylated TNKS1 in both SW480 and HEK293 cells. Levels of PARylated TNKS2 (and also axin, not shown) were too low for detection. Quantification of band intensities (mean ± SD) was from two separate experiments. B. Total extract western blot shows similar TNKS levels after different treatments.
Fig 7.
Model summarizing the action of TNKSi on formation of TNKS-axin complex formation and puncta assembly.
A. Table with diagrams. B. Hypothetical model summarizing the effects of TNKSi on TNKS/axin binding, oligomerization and puncta formation by blocking PARylation of TNKS1. This model is strengthened by a consistent negative impact of MG132 on each of these different steps including the ability of TNKSi to dePARylate TNKS, induce axin-TNKS binding complexes and to induce cytoplasmic axin-TNKS puncta. The mechanism by which MG132, through proteasome inactivation, is able to block puncta formation occurs at least in part through stabilizing the PARylation of TNKS (and possibly axin), but is yet to be fully defined. We propose the following: treatment with MG132 alone stabilizes PARylated TNKS, which causes it to dissociate from axin leading to increased nuclear axin. Treatment with TNKSi alone can stabililize unPARylated TNKS which then binds axin and retains it in the cytoplasm in large complexes and puncta. When MG132 is added after TNKSi treatment, it will induce some increase in the pool of PARylated TNKS, and causes a perinuclear shift in the TNKS/axin puncta through a mechanism yet to be defined.