SENP2 associates with NEMO and negatively regulates NF-κB activation by genotoxic stress. (A) Flag-SENP1, 2, 3, 5, 6, and 7 were co-transfected with Myc-NEMO in HEK293 cells. Transfected cells were treated with 10 μM VP16 for 1 h, lysed and NEMO was then precipitated with Myc antibody and analyzed by immunoblot using Flag or NEMO antibody. (B) HEK293 cells were transfected with vector (−) or SENP constructs and treated as above. Lysates were used for EMSA using a radio-labeled Igκ-κB or Oct-1 probe. (C) Phosphorimage quantified results from four experiments as in (B) were plotted (mean+SEM). *, p<0.03. (D) HEK293 cells were transfected with varying amounts of Flag-SENP2 wild type (wt) or a catalytically inactive mutant (c/s) and were analyzed by EMSA as in (A). (E) HEK293 cells were co-transfected with SENP2-wt, -c/s, or -siRNA (SMARTpool, Dharmacon) and 3x-κB-luciferase and β-galactosidase constructs. Twenty-four hours following transfection, cells were treated with VP16 or DMSO for 8 h and extracts were analyzed by immunoblotting using SENP2 or tubulin antibody. (F) Cell samples as in (E) were analyzed for luciferase and β-galactosidase activities and relative luciferase activity was plotted (mean+SEM). *, p<0.03; **, p<0.001. See also Figure S1.

SENP2 binds to NEMO and promotes NEMO deSUMOylation. (A) HEK293 cells stably expressing 6xMyc-NEMO were transfected with Flag-SENP2-wt or -c/s constructs, treated with VP16 (1 h) at 24 h after transfection, and then lysed in 1% SDS lysis buffer. Cell lysates were immediately boiled and then SDS diluted to 0.1% final with IP buffer before NEMO was precipitated using Myc antibody for subsequent immunoblot analysis using Myc or SUMO1 antibody. SUMOylated NEMO is indicated by S~NEMO. (*) indicates the IgG heavy chain. (B) HEK293 cells were treated as indicated and used for NEMO IP followed by immunoblot using SENP2, IKKα/β, or NEMO antibody. (C) HEK293 cells were cotransfected with Flag-SENP2-cs and Myc-NEMO truncation constructs (N1–4) and used for Myc IP followed by Flag or Myc immunoblot. Binding data (Bd) data are summarized with diagrams of mutants. (D) HEK293 cells were transfected with Myc-NEMO and Flag-SENP2 truncation constructs (S1–S5) and used for Flag IP followed by Myc or Flag immunoblot. Binding (Bd) data are summarized as above with S (*) indicating SUMO-interaction motif (SIM) and C (#) the catalytic site. (E) HEK293 cells were co-transfected with Myc-NEMO or SUMO-1 (minus the C-terminal di-glycine motif) fused NEMO (S1-NEMO) and Flag-SENP2 and used for Myc IP followed by immunoblotting using indicated antibodies. (F) HEK293 cells were transfected with varying amounts of Flag-SENP1-wt or Flag-SENP2-wt constructs, treated with VP16 for 1 h and used for Flag IP. SUMOylated NEMO (S~NEMO) was prepared using in vitro translation and SUMOylation assay (Mabb et al. 2006). SUMOylated NEMO was incubated with Flag-purified SENP1 or SENP2 for 2 h at 37°C and analyzed by NEMO or Flag immunoblotting. Lower graph: the percentages of SUMOylated NEMO amounts remaining with increasing amounts of transfected SENP constructs are indicated (mean+SD) from two experiments. See also Figure S2.

SENP2 and SENP1 genes are NF-κB-regulated genes selectively induced in response to genotoxic stress. (A) HEK293 cells treated as indicated were analyzed for SENP2 mRNA expression by qRT-PCR and graphed (mean+SEM). (B, C) Indicated parental and those stably expressing HA-IκBα-S32/36A were treated with VP16 for indicated times and analyzed as in (A) for SENP2 (B) or SENP1 mRNA (C) expression. The difference compared to untreated samples in parental cells; *, p<0.05; **, p<0.02; #, p<0.001. (D, E) CEM cells and those stably expressing HA-IκBα-S32/36A were treated as indicated and analyzed by EMSA (D) or immunobloting using indicated antibodies (E). See also Figure S3.

SENP2 is a direct NF-κB target gene. (A) A diagram showing putative κB binding sites in the 5′ region of the SENP2 locus. (B) CEM cells were incubated with VP16 for 3 h and then subjected to ChIP analysis using IgG, p65, or CBP antibodies and qPCR using primers (Supplement Information) spanning each of the putative κB binding sites (indicated with numbers). % input over total qPCR signals for each antibody is shown (mean+SEM). (C) Similar analyses as in (B) were done except for treating cells with TNFα for 3 h. (D) HEK293 cells were transfected with an empty luciferase reporter construct (CTL), SENP2 5′ sequence-luc (−2520), or SENP2 5′ sequence-luc with the κB sites mutated (indicated by “X”), along with β-galactosidase control vector, and treated with VP16 for 10 h. Luciferase activity corrected for β-galactosidase activity was plotted as fold luciferase activity (mean+SEM). *, p<0.001. (E) Similar analyses as in (D) were performed except for exposure of cells to TNFα for times indicated. See also Figure S4.

ATM-regulated histone methylation of SENP2 promoter in response to DNA damage. (A) HEK293 cells were incubated with TNFα (T) or VP16 (V) for 3 h. ChIP was performed using H3K4me2 or IgG control antibody and analyzed by qPCR with specific primers around the indicated SENP2 κB sites. The data are displayed as in Fig. 4B (mean+SEM). *, p<0.01. **, p<0.001. (B) HEK293 cells stably expressing HA-IκBα-S32/36A were analyzed as in (A). **, p<0.001. n.s., not significant. (C) HEK293 cells pretreated with KU55933 for 1 h followed by TNFα (T) or VP16 (V) for 3 h were analyzed as in (A). (D) HEK293 cells were transfected with control or ATM-targeting siRNA and then treated with VP16 (V) for 3 h prior to ChIP analysis using antibodies indicated. The efficiency of ATM knockdown was also evaluated by immunoblotting with ATM antibody. See also Figure S5.

SENP2 is required to attenuate a biphasic NEMO SUMOylation and NF-κB activation in response to DNA damage. (A) Different MEFs treated as indicated were analyzed by EMSA. (B) Phosphorimage quantification of NF-κB/AP1 binding from three independent experiments as in (A) was plotted (mean+SD). *p<0.001. **, p<0.01. (C) Different MEFs treated as indicated were processed as in Fig. 2A and SUMOylated NEMO (S~NEMO) was detected by immunoblotting with SUMO1 antibody (upper). (D) S~NEMO signals from (C) from two experiments were quantified by ImageJ analysis of scanned blots and plotted with the mean+SD for each time point indicated. (E) IKK immune-complex kinase assay was performed using GST-IκBα(1–56) as substrate. The amounts of the substrate and NEMO used for IKK immunoprecipitation are shown by immunoblotting. (F) NF-κB activation in Wt and Nfkbia^−/− MEFs were analyzed as in (A). (G) IKK kinase assay was performed using Nfkbia^−/− MEF as in (E). See also Figure S6.

SENP2 is required to limit NF-κB-dependent cell survival in response to DNA damage. (A) Wt and Senp2^−/− MEFs treated as indicated were analyzed by immunoblotting with indicated antibodies. (B) Senp2^−/− MEFs stably reconstituted with SENP2-wt, SENP2-c/s or vector control (Vec) were analyzed as in (A). (C) Wt, Senp2^−/− and Senp2^−/− stably expressing HA-IκBα-S32/36A (Senp2^−/−S32/36A) were exposed to different doses of CPT and cell viability was analyzed as in ‘Experimental Procedures’ and plotted (mean+SEM). (D) MEF cells as in (B) were analyzed as in (C). *, p<0.01. **, p<0.001. (E) A model depicting negative feedback regulation of NF-κB signaling induced by TNFα or DNA damage. See Discussion for details. See also Figure S7.