HIPK2 restricts SIRT1 activity upon severe DNA damage by a phosphorylation-controlled mechanism

Abstract

Upon severe DNA damage a cellular signalling network initiates a cell death response through activating tumour suppressor p53 in association with promyelocytic leukaemia (PML) nuclear bodies. The deacetylase Sirtuin 1 (SIRT1) suppresses cell death after DNA damage by antagonizing p53 acetylation. To facilitate efficient p53 acetylation, SIRT1 function needs to be restricted. How SIRT1 activity is regulated under these conditions remains largely unclear. Here we provide evidence that SIRT1 activity is limited upon severe DNA damage through phosphorylation by the DNA damage-responsive kinase HIPK2. We found that DNA damage provokes interaction of SIRT1 and HIPK2, which phosphorylates SIRT1 at Serine 682 upon lethal damage. Furthermore, upon DNA damage SIRT1 and HIPK2 colocalize at PML nuclear bodies, and PML depletion abrogates DNA damage-induced SIRT1 Ser682 phosphorylation. We show that Ser682 phosphorylation inhibits SIRT1 activity and impacts on p53 acetylation, apoptotic p53 target gene expression and cell death. Mechanistically, we found that DNA damage-induced SIRT1 Ser682 phosphorylation provokes disruption of the complex between SIRT1 and its activator AROS. Our findings indicate that phosphorylation-dependent restriction of SIRT1 activity by HIPK2 shapes the p53 response.

Figure 1

SIRT1 and HIPK2 interact in response to DNA damage. (a) Interaction of endogenous SIRT1 and HIPK2 upon DNA damage (0.75 μg/ml Adriamycin for 24 h). HIPK2 was precipitated from U2OS cell lysates as indicated and coprecipitated SIRT1 was analysed by immunoblotting. The input control is 2% of the total cell lysate. (b) Interaction of ectopically expressed SIRT1 and HIPK2 after DNA damage. Flag-SIRT1 and GFP-HIPK2 were expressed in 293T cells. Twenty-four hours after transfection cells were treated with adriamycin (0.75 μg/ml for 24 h) or left untreated. GFP-HIPK2 was precipitated from the lysates and the binding of Flag-SIRT1 to GFP-HIPK2 was measured by immunoblot analysis. The input control is 10% of total cell lysates. (c) In vitro interaction between SIRT1 and HIPK2. GST-SIRT1 and GST were incubated with in vitro translated ³⁵S-HIPK2, GST pull-downs were performed and analysed by SDS–PAGE and autoradiography. Two percent input was loaded as input control. Total amounts of proteins were analysed by Coomassie Brilliant Blue staining. (d) GST pulldown assays were performed with recombinant GST-SIRT1 truncations and ³⁵S-labelled HIPK2. Two percent input was loaded as control. (e) Flag-HIPK2 and the truncation mutants GFP-SIRT1 (aa 261–447) and (aa 448–747) were expressed in 293T cells. Flag-HIPK2 protein was precipitated from the lysates and co-immunoprecipitation of GFP-SIRT1 was analysed by immunoblot using the indicated antibodies. As input controls 5% of the total cell lysates were analysed. (f) GST-pulldown with recombinant GST-HIPK2 truncations and ³⁵S-labelled SIRT1. In all, 2% input were loaded as input control. (g) GFP-SIRT1 and the truncation mutants Flag-HIPK2 (aa 1–553) and (aa 551–1191) were expressed in 293T cells. Twenty-four hours after transfection cells were incubated with Adriamycin (0.75 μg/ml for 24 h) or left untreated. Flag-HIPK2 protein was precipitated from the lysates and co-immunoprecipitation of GFP-SIRT1 was analysed by immunoblot using the indicated antibodies. As input controls 10% of the total cell lysates were used. (h) Schematic representation of the SIRT1–HIPK2 interaction

Figure 2

HIPK2 phosphorylates SIRT1 in vitro and in cellular context. (a) HIPK2 directly phosphorylates SIRT1. Recombinant 6xHis-HIPK2, kinase-deficient 6xHis-HIPK2^K221A and GST-SIRT1 were purified from E. coli. In vitro kinase assays were performed and SIRT1 phosphorylation as well as HIPK2 autophosphorylation was examined by SDS–PAGE and autoradiography. Protein levels were analysed by either Coomassie Brilliant Blue staining or immunoblotting. (b) HIPK2 phosphorylates SIRT1 at Ser27 and Ser682 in vitro. GST-SIRT1 wild type, SIRT1 S27A, SIRT1 S682A and the double mutant SIRT1 S27A, S682A were incubated with a truncated, catalytically active GST-HIPK2^1–553. In vitro kinase assay was analysed by SDS–PAGE and autoradiography. Protein levels were analysed by Coomassie Brilliant Blue staining. The ratio (amount pSIRT1)/(amount pHIPK2) was quantified by densitometry using the ImageJ software. (c) Recombinant GST-SIRT1 and 6xHis-HIPK2 purified from E.coli were subjected to an in vitro kinase reaction. GST-SIRT1 phosphorylation was analysed by immunoblotting using the indicated antibodies. (d) U2OS cells were retrovirally transduced with a SIRT1-specific shRNA or a control shRNA. Depletion of SIRT1 was examined by immunoblot analysis using the indicated antibodies. The expression of GFP protein was used as a control for the transduction efficiency. (e) HIPK2 phosphorylates SIRT1 in cells. HA-HIPK2 wild type or kinase-deficient point mutant HA-HIPK2^K221A constructs were expressed along with Flag-SIRT1 in 293T cells. Twenty-four hours after transfection cell lysates were analysed by immunoblotting using the indicated antibodies

Figure 3

Phosphorylation of SIRT1 at Ser682 by HIPK2 upon DNA damage. (a) SIRT1 Ser682 phosphorylation is linked to DNA damage. U2OS cells were treated with 0.75 μg/ml of Adriamycin for the indicated time points. Total cell lysates were analysed by immunoblotting using the indicated antibodies. (b) SIRT1 Ser682 phosphorylation depends on HIPK2. U2OS cells were transfected with HIPK2-targeted siRNA. Twenty-four hours post-transfection cells were treated with Adriamycin (0.75 μg/ml) for 24 hours or left untreated. Total cell lysates were analysed by immunoblotting using the indicated antibodies. (c) U2OS cells were treated with sublethal (0.1 μg/ml) and lethal (0.75 μg/ml) doses of Adriamycin for the indicated time points. Total cell lysates were analysed by immunoblotting using the indicated antibodies. (d) SIRT1 Ser682 phosphorylation is linked to lethal DNA damage. U2OS cells were treated with sublethal (0.1 μg/ml) and lethal (0.75 μg/ml) doses of Adriamycin for 24 hours. Total cell lysates were analysed by immunoblotting using the indicated antibodies

Figure 4

Regulation of SIRT1 Ser682 phosphorylation at PML-NBs. (a) PML provokes colocalization of SIRT1 and HIPK2 at PML-NBs. GFP-SIRT1 (green), mCherry-HIPK2 (red) (upper panels) and PML IV (lower panels) were co-expressed in U2OS cells and analysed by immunofluorescence staining and confocal microscopy. DNA is stained by Hoechst33342 (blue). Scale bar, 10 μm. (b) Endogenous SIRT1 and HIPK2 co-localize at PML-NBs upon DNA damage. U2OS cells were treated with adriamyin (0.75 μg/ml) for 24 h. Endogenous PML, HIPK2 and SIRT1 were stained using indirect immunofluorescence. Representative confocal images are shown. Scale bar, 5 μm. (c) Phosphorylated SIRT1 localizes to PML-NBs. Flag-PML IV, SIRT1 and HA-HIPK2 were expressed in U2OS cells. pSer682 SIRT1 (anti-pSer682) and PML (anti-Flag) were examined by immunofluorescence staining and confocal microscopy. Scale bar, 10 μm. DNA is visualized with Hoechst33342 (blue). Representative cells are shown. (d) Endogenous SIRT1 phosphorylated at Ser682 localizes upon DNA damage to PML-NBs. U2OS cells were treated with adriamyin (0.75 μg/ml) for the indicated timepoints and analysed by indirect immunofluorescence staining. Representative confocal images are shown. Scale bar, 5 μm. (e) Localization of phosphorylated SIRT1 to PML-NBs is dependent on HIPK2 expression. U2OS cells were transfected with HIPK2-targeted siRNA. Twenty-four hours post-transfection cells were treated with adriamycin (0.75 μg/ml) for 6 and 24 h or left untreated. Total cell lysates were analysed by immunoblotting using the indicated antibodies. In parallel, cells were analysed by indirect immunofluorescence staining. Representative confocal images are shown. Scale bar, 5 μm. (f) Ectopic expression of PML potentiates SIRT1 phosphorylation at Ser682. U2OS cells were transfected with the indicated constructs and cell lysates were analysed by immunoblotting using the indicated antibodies. (g) Depletion of endogenous PML expression by RNA interference results in diminished SIRT1 Ser682 phosphorylation after DNA damage. U2OS cell were treated with the indicated siRNA and treated with Adriamycin (0.75 μg/ml) for 24 h. Total cell lysates were analysed by immunoblotting using the indicated antibodies

Figure 5

Ser682 phosphorylation regulates SIRT1 deacetylase activity, p53 acetylation and cell death. (a) HIPK2 is important for p53 Lys382 acetylation after DNA damage. HIPK2 was depleted in U2OS cells by RNA interference and cells were treated with Adriamycin (0.75 μg/ml) as indicated. Total cell lysates were analysed by immunoblotting using the indicated antibodies. (b) HIPK2 antagonizes SIRT1-mediated deacetylation of p53. Flag-SIRT1, HA-HIPK2, p53 and HA-CBP were expressed in H1299 cells as indicated and total cell lysates were analysed by immunoblotting using the indicated antibodies. GFP expression was used to control the transfection efficiency. Total p53 protein expression levels were adjusted to equal levels to be able to compare p53 acetylation under different conditions. The ratio (amount p53 acK)/(amount p53) was quantified by densitometry using the ImageJ software. (c) SIRT1 Ser682 is required for HIPK2-mediated SIRT1 inhibition. H1299 cells were transfected with the indicated constructs and total cell lysates were analysed by immunoblotting using the indicated antibodies. GFP expression was used to control the transfection efficiency. Total p53 protein expression levels were adjusted to equal levels to be able to compare p53 acetylation under different conditions. The ratio (amount p53 acK)/(amount p53) was quantified by densitometry using the ImageJ software. (d) SIRT1 Ser682 is important for DNA damage-stimulated p53 acetylation. H1299 cells were transfected with the indicated expression constructs. Twenty-four hours after transfection cells were treated with Adriamycin (0.75 μg/ml) or left untreated. Total cell lysates were analysed by immunoblotting. GFP expression was used to control the transfection efficiency. Total p53 protein expression levels were adjusted to equal levels to be able to compare p53 acetylation under different conditions. The ratio (amount p53 acK)/(amount p53) was quantified by densitometry using the ImageJ software. (e) SIRT1 Ser682 phosphorylation regulates p53-dependent transcription. U2OS cells were transfected with the indicated expression constructs together with a luciferase PUMA-reporter. Firefly reporter activity was normalized to Renilla reporter activity. Data are shown as means±S.D.; n=3; P<0.05, Student's t-test. (f) U2OS cells were transfected with the indicated expression constructs and expression of p53 target genes: p21, PUMA, BAX, NOXA and p53AIP1 were analysed by qRT-PCR. mRNA expression levels were normalized to the empty vector control. Data are shown as means±S.D.; n=3; P<0.05, Student's t-test. (g) SIRT1 Ser682 phorphorylation potentiates apoptosis. U2OS cells were transfected with the indicated expression constructs. Twenty-four hours after transfection cells were treated with Adriamycin (1 μg/ml, 48 h) and subsequently analysed by FACS using Annexin V-FITC

Figure 6

Regulation of the SIRT1–AROS complex by SIRT1 Ser682 phosphorylation and DNA damage. (a) SIRT1 Ser682 phoshorylation regulated SIRT1–AROS binding. 293T cells were transfected with the indicated expression constructs. Twenty-four hours after transfection Flag-SIRT1 was precipitated from the lysates and co-immunoprecipitation of AROS was analysed by immunoblotting using the indicated antibodies. Ten percent of the total cell lysate are shown as input control. (b) No effect of SIRT1 Ser682 phosphorylation on SIRT1–DBC1 interaction. 293T cells were transfected with the indicated expression constructs. Flag-SIRT1 was precipitated from the lysates and co-immunoprecipitation of AROS was analysed by immunoblotting using the indicated antibodies. Ten percent of the total cell lysate are shown as input control. (c) Disruption of the endogenous SIRT1–AROS complex after DNA damage. U2OS cells were treated with Adriamycin (0.75 μg/ml for 4 h) and MG-132 (20 μM) or left untreated. Subsequently, endogenous SIRT1 was precipitated from the lysates and co-immunoprecipitation of endogenous AROS was detected by immunoblotting. As control 10% of the total cell lysates were analysed. (d) Reduced interaction of ectopically expressed SIRT1 and AROS after DNA damage. HA-AROS and Flag-SIRT1 were expressed in 293T cells. Twenty-four hours after transfection cells were treated with Adriamycin (0.75 μg/ml for 4 h) and MG-132 (20 μM) or left untreated. Flag-SIRT1 was precipitated from the lysates and co-immunoprecipitation of HA-AROS was examined by immunoblotting. Ten percent of the total cell lysates are shown as input control. (e) Ser682 is required for dissociation of the SIRT1–AROS complex after DNA damage. HA-AROS and the phosphorylation-deficient mutant Flag-SIRT1^S682A were expressed in 293T cells. Cells were treated with Adriamycin (0.75 μg/ml for 4 h) and MG-132 (20 μM) or left untreated and Flag-SIRT1^S682A was precipitated from the lysates and analysed by immunoblotting using the indicated antibodies. Controls correspond to 10% of the total cell lysates

Figure 7

Proposed model for the interplay between HIPK2 and SIRT1 at PML-NBs upon severe genotoxic stress. Our data propose that in response to severe DNA damage SIRT1 and HIPK2 are recruited to PML-NBs by the PML isoform IV. Co-recruitment of both p53 regulatory enzymes facilitates crosstalk between SIRT1 and HIPK2 at the PML-NB. HIPK2 phosphorylates SIRT1, which in turn inhibits SIRT1 activity through dissociation of AROS. Reduced SIRT1 activity enables efficient p53 acetylation, expression of pro-apoptotic p53 target genes and potentiation of the DNA damage-induced cell death response