SCF/{beta}-TrCP promotes glycogen synthase kinase 3-dependent degradation of the Nrf2 transcription factor in a Keap1-independent manner

Abstract

Regulation of transcription factor Nrf2 (NF-E2-related factor 2) involves redox-sensitive proteasomal degradation via the E3 ubiquitin ligase Keap1/Cul3. However, Nrf2 is controlled by other mechanisms that have not yet been elucidated. We now show that glycogen synthase kinase 3 (GSK-3) phosphorylates a group of Ser residues in the Neh6 domain of mouse Nrf2 that overlap with an SCF/β-TrCP destruction motif (DSGIS, residues 334 to 338) and promotes its degradation in a Keap1-independent manner. Nrf2 was stabilized by GSK-3 inhibitors in Keap1-null mouse embryo fibroblasts. Similarly, an Nrf2(ΔETGE) mutant, which cannot be degraded via Keap1, accumulated when GSK-3 activity was blocked. Phosphorylation of a Ser cluster in the Neh6 domain of Nrf2 stimulated its degradation because a mutant Nrf2(ΔETGE 6S/6A) protein, lacking these Ser residues, exhibited a longer half-life than Nrf2(ΔETGE). Moreover, Nrf2(ΔETGE 6S/6A) was insensitive to β-TrCP regulation and exhibited lower levels of ubiquitination than Nrf2(ΔETGE). GSK-3β enhanced ubiquitination of Nrf2(ΔETGE) but not that of Nrf2(ΔETGE 6S/6A). The Nrf2(ΔETGE) protein but not Nrf2(ΔETGE 6S/6A) coimmunoprecipitated with β-TrCP, and this association was enhanced by GSK-3β. Our results show for the first time that Nrf2 is targeted by GSK-3 for SCF/β-TrCP-dependent degradation. We propose a "dual degradation" model to describe the regulation of Nrf2 under different pathophysiological conditions.

FIG. 1.

GSK-3 modulates Nrf2 levels. (A) HEK293T cells were maintained in low-serum medium (0.5% FBS) for 16 h before finally being treated with 15 μM tBHQ, 10 μM SFN, or 20 μM SB216763 for 6 h. Upper blot, Nrf2 immunodetection in cell lysates; middle blot, β-catenin levels as a control for GSK-3 inhibition; lower blot, β-actin levels showing similar protein loads per lane. (B) Quantitative reverse transcriptase PCR (RT-PCR) determination of mRNA of Nrf2 normalized by β-actin from HEK293T cells treated as in panel A (expressed in arbitrary units). Variations are not statistically significant. (C) HEK293T cells were transfected with the 3×ARE-LUC and renilla control vectors, and after transfection the cells were treated with 3 μM tBHQ, 3 μM SFN, or 20 μM SB216763 for 16 h before luciferase activity was measured. Asterisks denote statistically significant differences between the untreated and SB216763-treated groups according to a Student t test.

FIG. 2.

GSK-3 inhibition promotes Nrf2 protein accumulation in a Keap1-independent manner. (A) MEFs from Keap1-deficient (Keap1^−/−) or wild-type (Keap1^+/+) littermates were maintained in low-serum medium for 16 h and then treated with 20 μM SB216763 for the times indicated. Upper blot, total Nrf2 protein levels; lower blot, β-actin levels showing that similar amounts of protein were loaded per lane. (B) MEFs were maintained in low-serum medium for 16 h and then treated with 20 μM SB216763 for 2 h prior to inhibition of protein synthesis with 40 μg/ml cycloheximide (CHX). Whole-cell lysates were prepared at the indicated times after addition of CHX. Upper blots, Nrf2 protein levels in Keap1^+/+ and Keap1^−/− fibroblasts. Lower blots, β-actin levels showing similar protein loaded per lane from Keap1^+/+ and Keap1^−/− fibroblasts. (C and D) Both graphs depict the natural logarithm of the relative levels of the Nrf2 protein as a function of CHX chase time in Keap1^+/+ (C) or Keap1^−/− (D) cells. The protein half-life has been determined in the linear range of the degradation curve.

FIG. 3.

Modulation of Nrf2 protein levels by GSK-3 is independent of Keap1. (A and C) HEK293T cells were transfected with either V5-tagged Nrf2, the wild type, or the Keap1-insensitive Nrf2 version (Nrf2^ΔETGE-V5), maintained in low-serum medium for 16 h, and then treated with 20 μM SB216763 for the indicated times. Upper blots, Nrf2-V5 (A) or Nrf2^ΔETGE-V5 (C) protein levels; middle blots, β-catenin levels in the same cell lysates as a control for GSK-3 inhibition; lower blots, β-actin levels showing similar protein loads per lane. (B and D) HEK293T cells were transfected with either Nrf2-V5 (B) or Nrf2^ΔETGE-V5 (D). After 24 h they were further transfected with siRNAs for GSK-3α, GSK-3β, or both or with a control scrambled siRNA as explained in Materials and Methods. Cells were lysed 24 h after siRNA transfection. Upper blots, Nrf2-V5 (B) or Nrf2^ΔETGE-V5 (D) protein levels; middle blots, GSK-3α and -β protein levels; lower blots, β-actin levels showing that similar amounts of protein were loaded in each lane.

FIG. 4.

Nrf2 is regulated by the E3 ligase β-TrCP complex through a destruction motif within its Neh6 domain. (A) Upper panel of sequences, primary structure of Nrf2 between residues 317 and 359 in the murine protein. The bold/underlined residues correspond to the putative site of phosphorylation by GSK-3, and the boxed ones correspond to the β-TrCP consensus motif. Lower panel of sequences, alignment of different well-known β-TrCP substrates showing the consensus sequence of the degradation motif. (B) Keap1^−/− MEFs were transfected with siRNA against β-TrCP1 and/or β-TrCP2 as described in Materials and Methods. Upper blot, Nrf2 total protein levels; lower blot, glyceraldehyde-3-phosphate dehydrogenase (GADPH) levels showing similar protein loads per lane. (C) Quantitative RT-PCR determination of mRNA for β-TrCP1 and β-TrCP2 normalized by β-actin from MEFs transfected as in panel B. Asterisks denote statistically significant differences with P < 0.05. (D) HEK293T cells were cotransfected with the Nrf2-V5 or Nrf2^ΔETGE-V5 expression vector and the indicated amounts of the β-TrCP dominant-negative (β-TrCP^ΔFbox-HA) mutant and then maintained in low-serum medium for 16 h. Whole-cell lysates were immunoblotted against anti-V5 antibody (upper blot) or anti-HA antibody (middle blot) or with anti-β-actin antibody showing similar amounts of protein per sample (lower blot). (E) In vitro ubiquinitation of Nrf2 by the β-TrCP complex. Bacterially expressed His-tagged Nrf2 was submitted to an in vitro kinase assay in the absence (Nrf2) or presence (phospho-Nrf2) of recombinant GSK-3β. Nrf2 and phospho-Nrf2 (20 ng) were incubated at 25°C for 1 h with purified ubiquitin, E1/cdc34b, β-TrCP/Skp1, and Cul1/Rbx1 as indicated. Polyubiquitinated Nrf2 was detected by immunoblotting with antiubiquitin antibody. Upper blot, antiubiquitin. E1-Ub, monoubiquitinated E1. Lower blot, anti-Nrf2 showing similar amounts of substrate loaded per lane. (F) p100 dishes of HEK293T cells were transfected with the indicated plasmids. After transfection (24 h), a whole-cell lysate (input) and an affinity-purified His-tagged fraction (His-pull-down) were blotted with anti-V5 antibody. Upper blot, anti-V5 input; middle blot, ectopically expressed HA-tagged GSK-3β; lower blot, anti-V5 detection in the His pulldown fraction. The bracket indicates the mobility of polyubiquitinated Nrf2-V5 (Nrf2-Ub) forms.

FIG. 5.

Mutation of 6S to A in the Neh6 domain of Nrf2 reduces its phosphorylation by GSK-3β. (A) Representative in vitro kinase assay on recombinant Nrf2 or Nrf2^6S/6A proteins with HA-tagged GSK-3β immunoprecipitated from HEK293T cells transfected with either the inactive HA-GSK-3β^Y216F mutant or the constitutively active HA-GSK-3β^Δ9 or HA-GSK-3β^S9A mutant. Upper blot, immunocomplex kinase assay; middle blot, immunodetection of Nrf2 to ensure similar quantity per reaction; lower blot, immunoblot with anti-HA antibody showing similar amount of GSK-3β per assay. (B) Basal levels of Nrf2^ΔETGE-V5 or Nrf2^{ΔETGE 6S/6A}-V5 in HEK293T-transfected cells. Upper blot, anti-V5; lower blot, β-actin. (C) Densitometric quantification of representative blots from panel B. (D) Nrf2 mRNA levels analyzed by quantitative RT-PCR from cells transfected as in panel B. (E) Graph depicts the ratio between Nrf2 protein and mRNA levels as shown in panels C and D.

FIG. 6.

Mutation of the 6S cluster within the Neh6 domain increases Nrf2 protein stability. (A) HEK293T cells were transfected with Nrf2-^ΔETGE-V5 or Nrf2^{ΔETGE 6S/6A}-V5, serum starved for 16 h, and finally subjected to protein synthesis inhibition with 100 μg/ml CHX. Whole-cell lysates were prepared at the indicated times after addition of CHX. Upper blot, Nrf2-V5 protein levels. Lower blot, β-actin levels showing similar protein loads per lane. (B) The graph depicts the natural logarithm of the relative levels of the Nrf2-V5 protein as a function of CHX chase time. The protein half-life was determined using the linear part of the degradation curve. (C) HEK293T cells were transfected with Nrf2-^ΔETGE-V5 or Nrf2-^{ΔETGE 6S/6A}-V5, serum starved for 16 h, and then subjected to [³⁵S]methionine/cysteine labeling for 1 h. Then, cells were incubated in high-methionine- and cysteine-containing medium and collected at the indicated times. For the 120-min time point, the cells were treated with MG132 as an internal control. Upper blot, ³⁵S autoradiography; lower blot, anti-V5 antibody. (D) The graph shows the natural logarithm of the relative levels of Nrf2 as a function of ³⁵S chase time. The half-life has been determined in the linear range of the degradation curve.

FIG. 7.

Mutation of the 6S cluster within the Neh6 domain renders the Nrf2 protein insensitive to GSK-3. (A) HEK293T cells were transfected with Nrf2-^ΔETGE-V5 or Nrf2^{ΔETGE 6S/6A}-V5, maintained in low-serum medium for 16 h, and then incubated with 20 μM SB216763 for the indicated times. Upper blot, anti-V5 immunoblot; middle blot, β-catenin levels as a control of GSK-3 inhibition; lower blot, β-actin levels showing the similar protein loads per lane. (B) HEK293T cells were transfected with either Nrf2^ΔETGE-V5 or Nrf2^{ΔETGE 6S/6A}-V5. After 24 h they were further transfected with siRNAs for GSK-3α, GSK-3β, or both or with a control scramble siRNA as described in Materials and Methods. Cells were lysed 24 h after siRNA transfection. Upper blot, anti-V5 antibody; middle blot, GSK-3α and -β protein levels; lower blot, β-actin levels showing similar protein loads per lane.

FIG. 8.

The 6S cluster within the Neh6 domain is a target of β-TrCP in a GSK-3-dependent manner. (A) p100 dishes of HEK293T cells were transfected with the indicated plasmids. After transfection (24 h), a whole-cell lysate (input) and an affinity-purified His-tagged fraction (His-Pull-down) were blotted with anti-V5 antibody. Upper panel, anti-V5 input detection; middle panel, ectopically expressed HA-tagged GSK-3β; lower panel, anti-V5 detection in His pulldown fractions. The bracket indicates the mobility of polyubiquitinated Nrf2-V5 (Nrf2-Ub) forms. (B) Keap1^−/− or Nrf2^−/− MEF cells were serum starved for 6 h before they were treated for a further 2 h with vehicle control, 10 μM LY294002, or 5 μM CT99021. Thereafter, the fibroblasts that had been subjected to different treatments were harvested separately, and each was lysed in 0.45 ml of buffer. A 50-μl portion of the lysate was retained as “input” (labeled on the left), and the remainder was immunoprecipitated with antibodies against either Nrf2 (labeled IP Nrf2) or β-TrCP (labeled IP β-TrCP) using Sepharose-protein G beads. The immunoprecipitated material was analyzed by Western blotting using antibodies against Nrf2 or β-TrCP as indicated to the left of the blots. (C) HEK293T cells were cotransfected with the indicated plasmids or with an empty vector (Mock). One-fifth of whole-protein lysate was used to control for protein expression as shown in the three upper panels. The rest of the protein lysates were immunoprecipitated with anti-Flag or anti-V5 antibodies and immunoblotted as indicated in the four lower panels. The arrow points the specific β-TrCP immunoreactive band. (D) After cotransfection of HEK293T cells with either Nrf2^ΔETGE-V5 or Nrf2^{ΔETGE 6S/6A}-V5 and β-TrCP^ΔFbox-HA, cells were maintained in low-serum medium for 16 h. Upper panel, anti-V5 antibody; middle panel, anti-HA antibody; lower panel, anti-β-actin antibody showing a similar protein load.

FIG. 9.

Scheme showing the main participants in the “dual degradation” model. See Discussion for details.