Defining the Functional Targets of Cap'n'collar Transcription Factors NRF1, NRF2, and NRF3

Abstract

The NRF transcription factors NRF1, NRF2, and NRF3, are a subset of Cap'n'collar transcriptional regulators which modulate the expression of genes harboring antioxidant-response element (ARE) sequences within their genomic loci. Despite the emerging physiological importance of NRF family members, the repertoire of their genetic targets remains incompletely defined. Here we use RNA-sequencing-based transcriptional profiling and quantitative proteomics to delineate the overlapping and differential genetic programs effected by the three NRF transcription factors. We then create consensus target gene sets regulated by NRF1, NRF2, and NRF3 and define the integrity of these gene sets for probing NRF activity in mammalian cell culture and human tissues. Together, our data provide a quantitative assessment of how NRF family members sculpt proteomes and transcriptomes, providing a framework to understand the critical physiological importance of NRF transcription factors and to establish pharmacologic approaches for therapeutically activating these transcriptional programs in disease.

Keywords: NRF transcription factors; cellular pathway analysis; oxidative stress response; proteomics; transcriptomics.

Figure 1

An overexpression system for evaluating the activities of NRF transcription factors in HEK293T cells. (A) Schematic depicting the FLAG-NRF constructs used in this work; numbers above each transgene indicate the amino acid positions of the reference ORF. (B) Western blotting analysis for FLAG protein content from HEK293T cells 24 h after transfection with the indicated constructs. (C) Transcript levels of HMOX1 and NQO1 as measured by qRT-PCR from HEK293T cells 24 h after transfection with the indicated constructs (n = 3; mean and s.d.). (D) Relative ARE reporter activity from HEK293T cells overexpressing the indicated FLAG-NRF construct (n = 8; mean and s.d.).

Figure 2

Schematic depicting the strategy used in this work to define the functional targets of the NRF transcription factors.

Figure 3

Venn diagrams depicting the number of transcripts upregulated (top) or downregulated (bottom) in response to overexpression of the indicated NRF transgenes. Boxes depict the results of pathway analyses of the indicated regions of Venn overlap.

Figure 4

A consensus set of NRF2 target genes associated with oxidative stress resistance. (A) Venn diagram depicting the overlapping and differentially expressed transcripts induced by NRF2 activation identified in this work (Ibrahim, p < 0.05 and absolute value log₂ fold change > 0.075) in comparison to those obtained with reported inducible U2OS-based system (Liu et al.). (B) Plot depicting the linear relationship of differentially expressed transcripts derived from the common targets in this work and from Liu et al. (C) Linear regression plot of log₂ fold change values of protein and transcript levels of significantly changed genes in response to NRF2 overexpression. (D) Summary of GSEA analysis of RNA-seq profiling in response to NRF2 overexpression. (E) Heatmap depicting log₂ fold change of transcript levels associated with a response to ROS that is upregulated by NRF2 overexpression by DAVID analysis.

Figure 5

A consensus set of NRF1 target genes associated with proteostasis. (A) Venn diagram depicting the overlapping and differentially expressed transcripts induced by NRF1 activation identified in this work (Ibrahim, p < 0.05 and absolute value of log₂ fold change > 0.075) in comparison to those obtained with reported inducible U2OS-based system (Liu et al.). (B) Plot depicting the linear relationship of differentially expressed transcripts derived from the common NRF1 targets in this work and from Liu et al. (C) Linear regression plot of log₂ fold change values of protein and transcript levels of significantly altered genes in response to NRF1 overexpression. (D) Summary of GSEA analysis of RNA-seq profiling in response to NRF1 overexpression (E) Heatmap depicting log₂ fold change of transcript levels associated with proteostasis that are upregulated by NRF2 overexpression by DAVID analysis.

Figure 6

Defining common NRF3 targets within transcriptomes and proteomes. (A) Plot depicting the linear relationship of differentially expressed transcripts derived from the common NRF3 targets in this work and from Liu et al. (B) Linear regression plot of log₂ fold change values of protein and transcript levels of significantly modulated genes in response to NRF1 overexpression.

Figure 7

Co-expression patterns of NRF2 target transcripts in human tissues. Genes are separated into three groups by hierarchical clustering. First row: Correlation matrices of brain, liver, and heart clustered by the similarity of correlation patterns in the brain. Second row: Correlation matrices of brain, liver, and heart clustered by the similarity of correlation patterns in the liver. Third row: Correlation matrices of brain, liver, and heart clustered by the similarity of correlation patterns in the heart.