Beyond antioxidant genes in the ancient Nrf2 regulatory network

Abstract

Nrf2, a basic leucine zipper transcription factor encoded by the gene NFE2L2, is a master regulator of the transcriptional response to oxidative stress. Nrf2 is structurally and functionally conserved from insects to humans, and it heterodimerizes with the small MAF transcription factors to bind a consensus DNA sequence (the antioxidant response element, or ARE) and regulate gene expression. We have used genome-wide chromatin immunoprecipitation and gene expression data to identify direct Nrf2 target genes in Drosophila and humans. These data have allowed us to construct the deeply conserved ancient Nrf2 regulatory network-target genes that are conserved from Drosophila to human. The ancient network consists of canonical antioxidant genes, as well as genes related to proteasomal pathways and metabolism and a number of less expected genes. We have also used enhancer reporter assays and electrophoretic mobility-shift assays to confirm Nrf2-mediated regulation of ARE activity at a number of these novel target genes. Interestingly, the ancient network also highlights a prominent negative feedback loop; this, combined with the finding that Nrf2-mediated regulatory output is tightly linked to the quality of the ARE it is targeting, suggests that precise regulation of nuclear Nrf2 concentration is necessary to achieve proper quantitative regulation of distinct gene sets. Together, these findings highlight the importance of balance in the Nrf2-ARE pathway and indicate that Nrf2-mediated regulation of xenobiotic metabolism, glucose metabolism, and proteostasis has been central to this pathway since its inception.

Keywords: ARE; Drosophila; Enhancer; Free radicals; Genomics; Human; Keap1; Nrf2; Transcriptional regulation.

Figure 1. Genes targeted by Cnc/Nrf2 in Drosophila

(A) Boxplot representing the range of gene expression changes in response to CncC overexpression for genes not bound by Cnc and the Class I, II, and III Cnc/Nrf2 targets. Gene expression profiling data are from [45] and and can be found under GEO Accession Number GSE30087; fold change represents expression in adult flies transiently overexpressing CncC relative to wild-type adult flies. (B) Schematic representation of the three Cnc isoforms. The Keap1 interaction region (ETGE) and DNA binding region (bZIP) are highlighted. (C) The NRF2 consensus motif (ARE), which is the top enriched motif in the Class III peaks as determined by i-cisTarget. (D) Graph representing the fraction of Class I, II, or III peaks containing a match to the general ARE consensus GCnnnnTCA (E) Graph representing the percent of CncC-induced genes as defined in [45] that are targeted by Cnc/Nrf2 via a peak that contains an ARE versus those targeted by Cnc/Nrf2 via a peak that does not contain an ARE.

Figure 2. Deeply conserved human NRF2 targets are upregulated by sulforaphane in human cells

(A) Percent of NRF2 target genes overlapping human orthologs of Drosophila Class I, II, or III genes. Orthologs were identified using either the top scoring ortholog only (best ortholog), or all orthologs scoring >2 as described in the text. *P<0.0005, **P<0.001, #P<0.05, based on hypergeometric test. (B) Gene Set Enrichment Analysis (GSEA) comparing conserved human NRF2 target genes (human orthologs of Drosophila Class I genes) and gene expression changes after treatment of LCL cells with sulforaphane (SFN).

Figure 3. Enhancers at deeply conserved human target genes are regulated by NRF2 in human cells

(A) Human NRF2 ChIP-seq signal from LCL cells treated with DMSO or sulforaphane (SFN) as indicated. Select ancient NRF2 target genes with highly significant binding are represented (ChIP y-axis scale = 0-5000). (B) ChIP-seq signal as in (A) at select ancient NRF2 target genes with moderate binding (ChIP y-axis scale = 0-500). (C) Heatmap representing the response to sulforaphane (SFN), tert-butylhydroquinone (tBHQ), overexpression of NRF2, or overexpression of a dominant negative version of NRF2 (NRF2^DN) for reporter constructs driven by the enhancer regions highlighted in panels (A) and (B). NQO1 is a positive control for human NRF2, but is not a conserved target because insects do not have an orthologous gene; the remaining nine are enhancers at deeply conserved NRF2 target genes.

Figure 4. Direct NRF2 regulation of enhancers at PSMA3, VCP, and PC in human cells

(A) NRF2 ChIP-seq data from human LCL cells treated with suforaphane (SFN). Top: NRF2 ChIP-seq peak and ARE sequence at the PSMA3 locus. Middle: NRF2 ChIP-seq peak and ARE sequence at the VCP locus. Bottom: NRF2 ChIP-seq peak and ARE sequence at the PC locus. Gene models are in red, with tall boxes representing coding regions, short boxes representing untranslated regions, and dashed lines representing introns. Enhancer regions tested in reporter assays are represented by the blue box at the bottom of each panel, with location of the tested ARE highlighted in red. ARE sequences are provided, with orientation of the ARE sequence matching orientation of the target gene. (B) Electrophoretic mobility shift assays in which, for both gels, lane 1 contains a labeled NQO1 ARE probe with no protein, lane 2 contains the NQO1 probe with purified MAFG only, lane 3 contains the NQO1 probe with purified NRF2 only, and lane for contains the NQO1 probe with NRF2 and MAFG. Binding is only seen when both proteins are present in the reaction. Wild-type (WT) and mutant (mut) versions of the AREs from PSMA3, VCP, and PC were used as cold competitors as indicated. Although competition strength varies depending on the ARE, all WT probes compete with the labeled NQO1 probe, and this competition is lost when the AREs are mutated, indicating that the AREs from PSMA3, VCP, and PC are bound by NRF2-MAFG. (C) Reporter assays in which the region highlighed in the (A) was cloned upstream of the luciferase reporter gene. Control represents human IMR32 cells that were transfected with a control vector (pEF) and treated with DMSO, and NRF2⁺ represents cells that were transfected with an NRF2 expression plasmid (pEF-NRF2) and treated with SFN. The PSMA3 enhancer is upreguated in NRF2⁺, and this induction is ARE-dependent (lost in a construct with a mutated ARE). (D) Same as (C) for the VCP locus. This enhancer is also upreguated in NRF2⁺ in an ARE-dependent manner. (E) Same as (C) for the PC locus. This enhancer is also upreguated in NRF2⁺ in an ARE-dependent manner.

Figure 5. Direct NRF2 regulation of enhancers at KEAP1 and SQSTM1 in human cells

(A) Human NRF2 ChIP-seq peak and ARE sequence at the KEAP1 locus; gene models and tested enhancer regions are indicated as described in Figure 4A. (B) Human NRF2 ChIP-seq peak and ARE sequence at the SQSTM1 locus; gene models and tested enhancer regions are indicated as described in Figure 4A. (C) Electrophoretic mobility shift assays as described in Figure 4B, with AREs from KEAP1 and SQSTM1 enhancers used as cold competitors; both are able to compete with labeled NQO1 probe in an ARE-dependent manner (lost with mutation of ARE), though the KEAP1 ARE is a weaker competitor than the SQSTM1 ARE. (D) Luciferease reporter assay as described in Figure 4C, only with enhancer from the KEAP1 locus. (E) Same as (D) with enhancer from the SQSTM1 locus. Both the KEAP1 (D) and SQSTM1 (E) enhancers are upreguated in NRF2⁺ in an ARE-dependent manner.

Figure 6. Relationship between ARE sequence, NRF2 DNA binding, and gene expression in human cells

(A) The top enriched motif in the human NRF2 ChIP-seq data (see Methods). This sequence matches the NRF2/ARE consensus. (B) Breakdown of the percent of NRF2 bound regions containing various iterations of the ARE consensus sequence. (C) Dot plot representing relationship between ChIP-seq signal and the ARE motif score in each NRF2 bound region. ARE motif score is determined by similarity to the sequence in (A). ChIP peaks were sorted into bins based on the similarity of their strongest ARE to the ARE position weight matrix from JASPAR (ID: MA0150.1) using a first-order transcription factor flexible model [36]. (D) Same as C, only representing the relationship between motif score and SFN-induced expression changes for the target genes associated with NRF2 binding regions. (E) Dose-response experiment using reporter constructs driven by NRF2-targeted enhancers from the NQO1 and KEAP1 loci. Human IMR32 cells were treated with a range of SFN concentrations (0 to 20 uM), and fold change was calculated relative to DMSO control (0 uM SFN). Lines best fitting the data (nonlinear for NQO1, linear for KEAP1) are represented.