Development of Neh2-luciferase reporter and its application for high throughput screening and real-time monitoring of Nrf2 activators

Abstract

The NF-E2-related factor 2 (Nrf2) is a key transcriptional regulator of antioxidant defense and detoxification. To directly monitor stabilization of Nrf2, we fused its Neh2 domain, responsible for the interaction with its nucleocytoplasmic regulator, Keap1, to firefly luciferase (Neh2-luciferase). We show that Neh2 domain is sufficient for recognition, ubiquitination, and proteasomal degradation of Neh2-luciferase fusion protein. The Neh2-luc reporter system allows direct monitoring of the adaptive response to redox stress and classification of drugs based on the time course of reporter activation. The reporter was used to screen the Spectrum library of 2000 biologically active compounds to identify activators of Nrf2. The most robust and yet nontoxic Nrf2 activators found--nordihydroguaiaretic acid, fisetin, and gedunin--induced astrocyte-dependent neuroprotection from oxidative stress via an Nrf2-dependent mechanism.

Figure 1. Development of Neh2-luc Reporter

(A) Schematic presentation of reporter functioning. (B) Time course of the Neh2-luc reporter response to TBHQ compared to that for the commonly used ARE-luc reporter. (C) Time course of Neh2-luc and HIF ODD-luc reporter responses to lactacystin showing the lag period shortening with rising concentrations of the proteasomal inhibitor and, thus, confirming the switch of the rate-limiting step from specific recognition to proteasomal degradation. All values are presented as mean ± SEM. Increased expression of Nrf2-regulated genes in the Neh2-luc reporter line as a result of rescue of endogenous Nrf2 in the presence of the overexpressed Neh2-luciferase fusion, and the reporter response to canonical Nrf2 activators—PGJ2, TBHQ, and sulforaphane—in comparison to the absence of any response for HIF ODD-luc reporter (to confirm the specificity of each reporter) are shown in Figure S1.

Figure 2. Neh2-Luc Reporter Response to Upregulation and Downregulation of Keap1 Levels

(A) Keap1 overexpression results in a decreased luminescence in Neh2-luc cells transduced by Keap1 adenovirus. The efficiency of transfection of Neh2-luc cell line with FLAG-labeled Keap1-overexpressing adenovirus was 45%–70% as judged by immunostaining with anti-FLAG antibodies (see Figure S2A). (B) siRNA Keap1 knockdown results in an increased level of luminescence only in Neh2-luc cell line, but not in WT-luc line. (C) siRNA Keap1 knockdown has no effect on luminescent signal in WT-luc line. All values are presented as mean ± SEM. The siRNA Keap1 knockdown was confirmed by RT-PCR: it decreased levels of Keap1 mRNA and increased levels of mRNA of Nrf2-regulated genes in both Neh2-luc and WT-luc cell lines (see Figure S2B).

Figure 3. Structural Formulas of HTS Hits

Activation effects are shown in percentage (%) for 16 μM of the representative hits. See also Figure S3 for other novel scaffolds found.

Figure 4. Confirmation of Neh2-luc Fusion Accumulation Using Anti-Luciferase Antibodies

Concentration dependence of luciferase signal for “switch”-type hits (A and B) and confirmation of fusion protein accumulation by western blot (C). F, 3 hr treatment with 5 μM fisetin; N, 5 μM NDGA; Q, 4 μM quercetin. The control cell lines WT-luc and HIF ODD-luc (Smirnova et al.,2010) did not accumulate luciferase fusion under the same exposure conditions (see Figure S4).

Figure 5. Classification of HTS Best Hits Based on Kinetics of Reporter Activation

Switch-type activators (NDGA and fisetin); immediate alkylators (TBHQ, quercetin, sulforaphane, pyrithione); redox-cycling compounds undergoing prior oxidation and showing lag period (catechol, o-phenylene diamine); heavy metals (cadmium) working via inhibition of thiol-disulfide exchange and corresponding enzymes; Hsp90 inhibitors/destabilizers showing prolonged lag period (geldanamycin, TSA); and gedunin. Protein concentration 5.1 ± 0.2 μg per well. All values are presented as mean ± SEM. Keap1-labeling experiments in the presence of selected hits show that only sulforaphane behaves as a potent alkyating agent (see Figure S5).

Figure 6. Nrf2 Mediates the Astrocyte-Dependent Protective Effect of the Neh2-Luc Reporter Activators

(A–C) Cultured primary astrocytes were treated for approximately 24 hr with NDGA (A), fisetin (B), or gedunin (C). Immediately following complete wash off of the treatments, primary immature neurons were plated in the presence or absence of HCA. Forty-eight hours later, neuronal viability was determined. (D and E) Astrocytes were treated for 24 hr with 5 μM sulforaphane, 20 μM tBHQ, 10 μM NDGA, 24 μM gedunin, or 20 μM fisetin followed by RNA isolation. mRNA for NADPH quinone oxidoreductase 1 (NQO1) (D) or heme oxygenase-1 (HO-1) (E) was quantified with real-time PCR. (F) Astrocytes were treated for 24 hr with 5 μM sulforaphane, 20 μM fisetin, 24 μM gedunin, 5 μM NDGA (NDGA (5)), or 10 μM NDGA (NDGA (10)). Immunoblots show HO-1 and β-actin, used as a loading control, immunoreactivity. The last lane is recombinant HO-1. (G) Astrocytes were transfected with transfection reagent alone (Tsx-Ctl), a scrambled siRNA sequence (siScrml), or siRNA targeted against Nrf2 (siNrf2-1, siNrf2-2, siNrf2-3) and treated with 5 μM sulforaphane (SF) for 24 hr (total siRNA treatment 48 hr). Immunoblots show HO-1 and β-actin immunoreactivity. The last lane is recombinant HO-1 protein. (H–J) Astrocytes were treated with siRNAs for 24 hr (Ctl, nontreated; Tsx, transfection reagent alone; siScr, scrambled siRNA, or siRNA targeted against Nrf2, siN1, siN2, or siN3), followed by treatment with NDGA (H), fisetin (I), or gedunin (J) for 24 hr (total siRNA = 48 hr). Immediately following complete wash off of the treatments, primary immature neurons were plated with or without HCA. Statistical significance was determined via one-way ANOVA, followed by post hoc Dunnett’s test (A–C and H–J) or Student’s t test with Bonferroni correction (D and E). (A–C) *p < 0.05, comparisons are made within 5 mM HCA treatment groups and are versus 5 mM HCA alone. (D and E) **p < 0.01 all groups versus control. (H–J) *p < 0.05, versus Ctl NDGA plus HCA, Ctl fisetin plus HCA, or Ctl gedunin plus HCA. See Figure S6 for control experiments.

Figure 7. Schematic Representation of Different Mechanisms of Nrf2 Level Regulation and Plausible Mechanism of Gedunin Action

(A) Docking mode of gedunin in comparison with the binding mode of Neh2 portion into Keap1. (B) Overlap between Neh2 peptide and gedunin, from perpendicular views. (C) Hypothetic modes of Nrf2 level regulation (see text for details).