A Homeostatic Shift Facilitates Endoplasmic Reticulum Proteostasis through Transcriptional Integration of Proteostatic Stress Response Pathways

Abstract

Eukaryotic cells maintain protein homeostasis through the activity of multiple basal and inducible systems, which function in concert to allow cells to adapt to a wide range of environmental conditions. Although the transcriptional programs regulating individual pathways have been studied in detail, it is not known how the different pathways are transcriptionally integrated such that a deficiency in one pathway can be compensated by a change in an auxiliary response. One such pathway that plays an essential role in many proteostasis responses is the ubiquitin-proteasome system, which functions to degrade damaged, unfolded, or short half-life proteins. Transcriptional regulation of the proteasome is mediated by the transcription factor Nrf1. Using a conditional knockout mouse model, we found that Nrf1 regulates protein homeostasis in the endoplasmic reticulum (ER) through transcriptional regulation of the ER stress sensor ATF6. In Nrf1 conditional-knockout mice, a reduction in proteasome activity is accompanied by an ATF6-dependent downregulation of the endoplasmic reticulum-associated degradation machinery, which reduces the substrate burden on the proteasome. This indicates that Nrf1 regulates a homeostatic shift through which proteostasis in the endoplasmic reticulum and cytoplasm are coregulated based on a cell's ability to degrade proteins.

Keywords: ER stress; ERAD; Nrf1; UPR; proteasome; proteostasis.

FIG 1

The ER stress sensor gene Atf6 is significantly downregulated in Nrf1 CKO mice. (A) RT-qPCR analysis of wild-type and Nrf1 CKO liver tissue shows that a broad range of proteasome subunits are significantly downregulated in Nrf1 CKO mice. (B) Schematic representation of the experimental approach. mRNA was extracted from the livers and used for RT-qPCR analysis of gene expression. (C) RT-qPCR analysis of the expression of the ER stress sensor genes Perk, Ire-1, and Atf6 clearly shows that while the expression of Perk and Ire-1 are unaltered in Nrf1 CKO liver, Atf6 is significantly downregulated in the absence of Nrf1. In both panels A and C, gene expression in Nrf1 CKO liver tissue is shown relative to the wild type, with wild-type expression fixed at 1. Error bars display the SEM (n = 4; *, P < 0.05; **, P < 0.01).

FIG 2

ChIP-Seq analysis of Nrf1. (A) Wild-type MEFs were used for the ChIP-Seq experiments. In response to the proteasome inhibitor MG132, the Nrf1 protein significantly accumulates within cells. Two isoforms of Nrf1 are present in MEFs: a high-molecular-weight form (open triangle) and a cleaved low-molecular-weight nuclear form (closed triangle). (B) Manual ChIP-qPCR of the Nrf1 target genes Psmb6, Psmc4, and Psmd12 showed significant enrichment during pulldown with the TFC11 antibody compared to the control locus Txs. Error bars display the SEM (n = 3). (C) Summary of the three Nrf1 ChIP-Seq data sets. Library 3 contains the most highly enriched peaks from libraries 1 and 2, suggesting that it was created using more stringent conditions. (D) KEGG pathway analysis of the genes associated with the Nrf1 binding peaks identified from the ChIP-Seq data. (E) The relative distance of the identified Nrf1 binding sites relative to the transcriptional start site (TSS) shows clear clustering around the TSS. (F) Genomic distribution of the Nrf1 binding peaks identified by ChIP-Seq. (G) The consensus Nrf1 binding site identified using MEME-ChIP. (H) Comparison of binding site locations between Nrf1 and a previously published Nrf2-sMaf ChIP-Seq analysis.

FIG 3

Nrf1 binds to an enhancer in the Atf6 locus. (A) Nrf1 ChIP-Seq binding site in the Atf6 locus, coupled with binding sites for MafK, p300, H3K4m1, and H3K4m3 obtained from the USC genome browser (https://genome.ucsc.edu). (B) Manual ChIP-qPCR showing enrichment of Nrf1 at the Psma3 and Atf6 loci relative to the negative control, Txs. Error bars display the SEM (n = 3; *, P < 0.001).

FIG 4

Microarray analysis shows a significant change in ER homeostasis gene expression in Nrf1 CKO mice. A microarray performed comparing gene expression in Nrf1 CKO liver and WT liver tissue revealed downregulation of proteasome subunit genes (A), the ATF6-dependent transcription program (B), and genes involved in protein processing in the ER (C).

FIG 5

The ATF6 and ERAD transcriptional programs are downregulated in Nrf1 CKO mice. (A) RT-qPCR of ATF6 target genes revealed that they are significantly downregulated in Nrf1 CKO liver in comparison to wild-type tissue. (B) Immunoblot analysis of three wild-type and Nrf1 CKO mice clearly shows that, at the protein level, the cellular abundance of ATF6 is reduced in Nrf1 CKO cells. (C) RT-qPCR analysis of additional ERAD components shows that they are significantly downregulated in Nrf1 CKO liver tissue. (D) Diagram showing the proposed mechanism by which Nrf1 regulates ER homeostasis. In the absence of Nrf1, the transcription factor ATF6 is downregulated, leading to a decrease in ERAD. In both panels A and C, the error bars display the SEM (n = 4; *, P < 0.05; **, P < 0.01).

FIG 6

Nrf1 directly regulates transcription of the ERAD machinery. Nrf1 ChIP-Seq binding sites in the promoters of the Ufd1l (A), Nploc4 (B), and VCP (C) genes are shown. (D) A manual ChIP-qPCR shows enrichment of Nrf1 at the Ufd1l, Nploc4, and VCP loci relative to the negative control, Txs. Error bars display the SEM (n = 3; *, P < 0.005).

FIG 7

The IRE1 and PERK pathways are unaffected by Nrf1 deletion. (A) RT-PCR analysis of wild-type and Nrf1 CKO mice both in the basal state and treated with tunicamycin for 24 h to induce ER stress shows that the cleavage pattern of the mRNA for XBP1 is identical across genotypes. This indicates that IRE1 activity is not changed upon Nrf1 deletion. (B) Immunoblot analysis shows that at the protein level the accumulation of ATF4, under both basal and tunicamycin (TM)-induced ER stress conditions, is unaltered in Nrf1 CKO mice relative to wild-type mice. This indicates that PERK activity is not changed upon Nrf1 deletion. (C) RT-qPCR analysis shows that XBP1 target gene expression is unaltered in Nrf1 CKO mice. (D) RT-qPCR analysis shows that Atf4 target gene expression is unaltered in Nrf1 CKO mice. In both panels C and D, gene expression in Nrf1 CKO liver tissue is shown relative to the wild type, with wild-type expression fixed at 1. Error bars display the SEM (n = 4).

FIG 8

A CHOP-dependent gene signature in upregulated in Nrf1 CKO mice. (A) A microarray comparing gene expression in Nrf1 CKO liver and WT liver tissue revealed a significant increase in expression of Chop, in addition to the CHOP target genes Bcl2l11, Atf3, and Osgin2. (B) RT-qPCR analysis of CHOP target genes confirmed the microarray data, showing that a CHOP-dependent gene signature is significantly upregulated in Nrf1 CKO liver in comparison to wild-type tissue. (C) Immunoblot analysis of three wild-type and Nrf1 CKO mice clearly shows that at the protein level the cellular abundance of CHOP is increased in Nrf1 CKO cells. Gene expression in Nrf1 CKO liver tissue is shown relative to the wild type, with wild-type expression fixed at 1. Error bars display the SEM (n = 4; *, P < 0.05; **, P < 0.01).

FIG 9

Cellular stress significantly increases the CHOP-dependent gene expression in Nrf1 CKO mice and Nrf1 KO fibroblasts. (A) Tunicamycin treatment leads to a significant increase in Chop expression in Nrf1 CKO mice. Both wild-type and Nrf1 CKO mice exhibit increased Chop expression in response to tunicamycin-induced ER stress; however, the increase in Chop expression in Nrf1 CKO mice is significantly enhanced relative to wild-type mice. (B) Tunicamycin treatment leads to a significant increase in a CHOP-dependent gene expression signature, represented by the CHOP target genes Gadd34, Bax, Puma, Bcl2l11, and Atf3 in Nrf1 CKO mice relative to wild-type mice. (C) As an alternative stress model, MG132-mediated proteasome inhibition leads to a significant enhancement in the transcriptional induction of Chop, and the CHOP target genes Bcl2l11 and Gadd34 in Nrf1 KO fibroblasts compared to wild-type fibroblasts. The data are presented as the fold induction of gene expression, with the basal expression level set at 1. Error bars display the SEM (n = 3). (D) Treatment of wild-type and Nrf1 KO MEFs with the stressing agents MG132 and tunicamycin for 24 h shows a significant increase in cell death in Nrf1 KO cells relative to wild-type cells. In both panels A and B, gene expression in Nrf1 CKO liver tissue is shown relative to the wild type, with wild-type expression in response to PBS fixed at 1. Error bars display the SEM (n = 6; *, P < 0.05; **, P < 0.01).

FIG 10

Nrf1-dependent regulation of ATF6 integrates the ERAD and proteasome transcriptional programs. Nrf1 regulates the transcription of the proteasome subunit genes, and therefore the loss of Nrf1 leads to a reduction in proteasome activity. Through the direct regulation of ATF6, downregulation of Nrf1 results in a decrease in ERAD, which reduces the flow of protein substrates to the proteasome, resulting in a homeostatic shift which allows cells to survive in response to decreased proteasome activity. Since the downregulation of ERAD may make cells vulnerable to the deleterious effects of unfolded proteins, reduced proteasome activity, reduced proteasome activity is coupled with increased CHOP activity, which primes cells for apoptosis in response to further stress.