Logo of elifeeLifeRecent contentAbout eLifeFor authorsSign up for alerts
eLife. 2014; 3: e01856.
Published online Jan 21, 2014. doi:  10.7554/eLife.01856
PMCID: PMC3896944

p97-dependent retrotranslocation and proteolytic processing govern formation of active Nrf1 upon proteasome inhibition

Michael S Brown, Reviewing editor
Michael S Brown, UT Southwestern Medical School, United States;

Abstract

Proteasome inhibition elicits an evolutionarily conserved response wherein proteasome subunit mRNAs are upregulated, resulting in recovery (i.e., ‘bounce-back’) of proteasome activity. We previously demonstrated that the transcription factor Nrf1/NFE2L1 mediates this homeostatic response in mammalian cells. We show here that Nrf1 is initially translocated into the lumen of the ER, but is rapidly and efficiently retrotranslocated to the cytosolic side of the membrane in a manner that depends on p97/VCP. Normally, retrotranslocated Nrf1 is degraded promptly by the proteasome and active species do not accumulate. However, in cells with compromised proteasomes, retrotranslocated Nrf1 escapes degradation and is cleaved N-terminal to Leu-104 to yield a fragment that is no longer tethered to the ER membrane. Importantly, this cleavage event is essential for Nrf1-dependent activation of proteasome gene expression upon proteasome inhibition. Our data uncover an unexpected role for p97 in activation of a transcription factor by relocalizing it from the ER lumen to the cytosol.

DOI: http://dx.doi.org/10.7554/eLife.01856.001

Research organism: human, mouse

eLife digest

Cells exposed to high temperatures, infections and other forms of stress often produce oxygen ions and peroxide molecules that can cause damage to proteins and DNA. Cells therefore rely on molecular machines called proteasomes to eliminate damaged proteins, before they cause too much harm. Two related transcription factors—proteins that interact with DNA to ‘switch on’ the expression of genes—are involved in a cell’s responses to stress, but in different ways. Nrf2 switches on genes that limit the damage caused by oxygen ions and peroxide molecules, while Nrf1 switches on the genes that encode the components of the proteasome. As such, Nrf1 helps to restart proteasome activity if it has been shut off—a phenomenon known as ‘bounce-back’.

Within a cell, Nrf1 is known to start off embedded within the membranes of a structure called the endoplasmic reticulum. However, it is not clear how activated Nrf1 leaves this membrane and enters the nucleus to interact with the cell’s DNA. Now, Radhakrishnan et al. show that when Nrf1 is produced, most of its length is found inside the endoplasmic reticulum, with only a small piece being anchored in the surrounding membrane. This is unlike previously described transcription factors that associate with the endoplasmic reticulum, which are stuck to the outside of this structure.

Radhakrishnan et al. also discovered that the activation of Nrf1 depends on an enzyme called p97 or VCP. This enzyme helps to flip Nrf1 from the inside of the endoplasmic reticulum to its outside surface. In most cells, the proteasome then breaks down this part of Nrf1. However, if the proteasome is inhibited, an unknown enzyme cuts Nrf1 free from the endoplasmic reticulum, allowing it to migrate to the nucleus and promote the production of more proteasome components to counteract the inhibition.

Interestingly, drugs that inhibit the proteasome are used to combat cancer because the build-up of damaged proteins is toxic to the cancer cells. By showing that p97 promotes the ‘bounce-back’ of the proteasome, the work of Radhakrishnan et al. suggests that combining existing proteasome inhibitors with drugs that inhibit p97 could eventually lead to new, more effective, therapies for cancer or other diseases.

DOI: http://dx.doi.org/10.7554/eLife.01856.002

Introduction

Transcription factor Nuclear factor erythroid derived 2-related factor 1 (Nrf1) belongs to the cap ‘n’ collar basic leucine zipper (CNC-bZIP) family of proteins that are known to be activated in response to cellular stress (Sykiotis and Bohmann, 2010). Other members of this family include p45 NF-E2, Nrf2, Nrf3, Bach1, and Bach2 (Andrews et al., 1993; Moi et al., 1994; Oyake et al., 1996; Kobayashi et al., 1999). The CNC-bZIP transcriptions factors heterodimerize with small Maf proteins (Maf F, Maf G, and Maf K) and preferentially bind to anti-oxidant response elements (AREs) present in the promoter region of their target genes (Motohashi et al., 2002). The most studied member of the CNC-bZIP family is Nrf2, which in response to oxidative stress directs a transcriptional program that helps maintain cellular redox homeostasis (Kensler et al., 2007). Owing to their sequence similarity, Nrf1 was initially thought to have an overlapping function with Nrf2 in regulating antioxidant gene expression, but a growing body of emerging evidence contradicts this notion.

Recently, we demonstrated that in mouse embryonic fibroblasts, Nrf1 but not Nrf2 is necessary for elevated expression of proteasome subunit mRNAs observed in cells treated with proteasome inhibitor, leading to a recovery or ‘bounce-back’ of proteasome activity (Radhakrishnan et al., 2010). Consistent with our observation, TCF11 (a longer isoform of Nrf1 found only in humans) was subsequently reported to be a mediator of proteasome bounce-back response after proteasome inhibition in human cells (Steffen et al., 2010). Thus it appears that Nrf1 functions to combat proteotoxic stress caused by proteasome inhibition in mammals akin to transcription factors RPN4 in yeast (Xie and Varshavsky, 2001) and Cnc-C in Drosophila (Grimberg et al., 2011). From the standpoint of cancer treatment, blockade of this bounce-back response may be a viable strategy to enhance the efficacy of proteasome inhibition therapy, given that Nrf1 depletion slows the rate of recovery of proteasome activity following transient application of a covalent proteasome inhibitor and results in enhanced proteasome inhibitor-mediated apoptosis in cancer cells (Radhakrishnan et al., 2010). However, since the molecular requirements for Nrf1 activation have not been described, there remains no known mechanism to exploit this possibility.

In addition to its role in the induced synthesis of proteasome subunit (PSM) genes, Nrf1 has also been found to regulate their basal expression in certain scenarios. For instance, in Nrf1−/− mouse neurons, PSM gene expression is diminished resulting in impaired proteasome function and neurodegeneration (Lee et al., 2011). Also, a liver-specific knockout of Nrf1 in mice caused a similar attenuation of PSM expression in hepatocytes (Lee et al., 2013).

Despite the importance of Nrf1 in proteasome biology and its potential as an anti-cancer target, a thorough understanding of the molecular mechanism behind its activation is currently lacking. What is known is that Nrf1 exists in two forms, p120 and p110, both of which are unstable and accumulate when the proteasome is inhibited (Radhakrishnan et al., 2010). Whereas p120 is embedded in the ER membrane, p110 is soluble and can enter the nucleus (Biswas and Chan, 2009). However, the mechanism leading to the formation of these species is controversial, with one study pointing to proteolytic cleavage (Wang and Chan, 2006) and another to differential glycosylation (Zhang et al., 2007). We show here that, in striking contrast to other ER membrane-tethered transcription factors, the bulk of the Nrf1 polypeptide is inserted into the ER lumen. Activation of Nrf1 depends on p97/VCP-dependent transfer of the luminal segments of Nrf1 to the cytosolic side of the ER membrane, followed by a novel proteolytic processing step.

Results

Transcriptionally active Nrf1 p110 is derived from p120 by proteolytic processing

To discriminate between the possibilities that p110 is derived by cleavage or deglycosylation of p120, we first overexpressed 3×FlagNrf1HA in human HEK-293T and mouse NIH-3T3 cells. This construct contains three consecutive copies of the Flag tag at the N-terminus of Nrf1 and an HA tag at the C-terminus. In the presence of MG132, regardless of the cell type used, we observed that although the anti-HA antibody was able to detect two different forms of Nrf1 (~120 and ~110 kDa), the anti-Flag antibody was able to detect only the ~120 kDa form (Figure 1A). The simplest explanation for this observation is that Nrf1 p120 was cleaved somewhere close to the N-terminus to yield p110. To test this hypothesis and to identify the cleavage site, we overexpressed Nrf13×Flag (Nrf1 with C-terminal triple Flag tag) in HEK-293T cells and immunopurified the protein and subjected the p120 and p110 forms to Edman degradation-based N-terminal sequencing. Although the ~120 kDa band confirmed the intact N-terminus of the full-length Nrf1, sequence from the ~110 kDa band was consistent with a new N-terminus starting with Leu-104 (Figure 1B; putative cleavage site indicated with a scissor).

Figure 1.
Nrf1 p110 is derived from p120 by proteolytic processing.

To further confirm this observation, we introduced several mutations flanking the predicted cleavage site (m1 through m6 in Figure 1B) and compared these mutants to wild-type Nrf13×Flag for their ability to be processed into Nrf1 p110. We found that all mutants were defective to varying degrees in their ability to be processed to the p110 form, although the defects of m4 and m6 were relatively modest (Figure 1C). Interestingly, m4 and m6 are the only mutants that retain Trp-103 at the P1 position (with respect to the cleavage site), implying a crucial role of this Trp residue for protease recognition.

Although our data suggested that p110 was derived from p120, there is no decisive evidence that proves that this is the case. To determine if there is a post-translational precursor–product relationship between Nrf1 p120 and p110 forms, we performed a pulse-chase experiment. We used L-azidohomoalanine (analog of the amino acid L-methionine) to pulse-label newly synthesized proteins and followed the resultant pool of Nrf1 during a chase period in which protein synthesis and proteasome-dependent degradation were both inhibited by the addition of cycloheximide and MG132. We observed that over time, wild-type p120 was processed to p110 but the non-cleavable Nrf1(m1)3×Flag was not processed (Figure 1D). Taken together, our data are consistent with a model in which newly synthesized Nrf1 p120 undergoes proteolytic processing between Trp-103 and Leu-104 to generate the p110 form.

Next, we asked if this proteolytic processing of Nrf1 is biologically relevant. To this end, we compared the ability of wild-type and non-cleavable mutant Nrf1 to reconstitute the bounce-back response in Nrf1−/− mouse embryonic fibroblasts (Figure 2A). As expected, wild-type Nrf13×Flag was able to reinstate the induction of proteasome subunit (PSM) genes in response to MG132 treatment in these cells (Figure 2B). In contrast, non-cleavable Nrf1(m1)3×Flag was defective in inducing PSM genes under the same conditions, implicating a necessary role for proteolytic processing in Nrf1’s transcriptional program.

Figure 2.
Non-cleavable Nrf1 mutant is functionally defective.

p97 is required for Nrf1 activity and its proteolytic processing

p97 is a homo-hexameric AAA ATPase that is implicated in numerous cellular processes ranging from cell-cycle regulation to membrane fusion and protein degradation (Ye, 2006). p97 has also been implicated in the turnover of the TCF11 isoform of human Nrf1 via the endoplasmic reticulum-associated degradation (ERAD) pathway (Steffen et al., 2010). Similar to inhibition of the proteasome, depletion of p97 leads to the strong accumulation of ubiquitin-conjugated substrates (Wojcik et al., 2004), underscoring its role in enabling degradation of a subset of the ubiquitinated proteome by the proteasome (Kolawa et al., 2013). Given that cells depleted of p97 or treated with proteasome inhibitors accumulate both ubiquitinated proteins and Nrf1, we asked if p97 depletion was also sufficient to trigger the Nrf1-mediated bounce-back response. To this end, we made use of NIH-3T3 mouse fibroblasts that have been engineered to stably express a doxycycline-inducible form of an shRNA targeting p97. Unlike proteasome inhibition, depletion of p97 upon addition of doxycycline did not induce the bounce-back response (Figure 3A; compare bars 1 and 3 for the PSM genes). By contrast, we observed that MG132-mediated bounce-back response was severely impaired when p97 was knocked-down (Figure 3A; compare bars 3 and 4 for the PSM genes).

Figure 3.
p97 is required for processing and activation of Nrf1.

To investigate the molecular basis for the lack of bounce-back response in p97-depleted cells, we examined the fate of Nrf1. Consistent with Steffen et al., 2010, Nrf1 accumulated in cells depleted of p97 (Figure 3B, compare lanes 1 and 3). To address the key question of whether p97 contributes to processing of Nrf1, which is essential for its activation, we evaluated the forms of Nrf1 that accumulated in control vs p97-depleted cells treated with MG132. Whereas the control cells accumulated p110 (indicative of Nrf1 activation), this form was markedly absent in p97-depleted cells (Figure 3B, compare lanes 2 and 4). This failure to process p120 is sufficient to account for the lack of bounce-back response in these cells. However, one concern was that due to the 3–4 days required for p97 knock-down, the p120 form we observed in p97-depleted cells could represent an aberrant dead-end species that only accumulated upon sustained deprivation of p97 activity, and was not a true intermediate in the processing pathway. To address this possibility, we made use of the recently described compound NMS-873, a reversible inhibitor of p97 (Magnaghi et al., 2013; Polucci et al., 2013). Treatment of NIH-3T3 mouse fibroblasts or HEK-293-Nrf13×Flag cells with NMS-873 led to a robust accumulation of p120 Nrf1 (Figure 3C; lane 1), which was then efficiently converted to the p110 form with a half-life of ~30 min upon washout of NMS-873 in the presence of cycloheximide and MG132 (Figure 3C, lanes 2–6). The overproduced protein behaved similarly in HEK-293 cells, although the conversion of p120 to p110 was not as efficient as for the endogenous protein (Figure 3C, compare left and right panels). Thus, our results are consistent with a model in which p97 controls not only the degradation of Nrf1, but also its processing to the active p110 form.

p97 re-positions the C-terminal domain of Nrf1 into the cytosol

It has been proposed that during synthesis, Nrf1 becomes anchored in the ER membrane via an N-terminal transmembrane region (residues 7–24) (Zhang and Hayes, 2010). Given the role of p97 in Nrf1 processing and activation, we hypothesized that either p97 promotes processing of cytosolically-exposed Nrf1 and possibly its subsequent extraction from binding partners (as is the case for Spt23 and Mga2 in yeast, Hoppe et al., 2000; Rape et al., 2001), or p97 could be involved in re-positioning the TAD and DBD of Nrf1 from the luminal to the cytosolic side of the ER membrane to facilitate its processing. To distinguish between these possibilities, we performed a series of protease protection experiments to assess the topology of Nrf1 under various conditions (Figure 4A–D). Regardless of p97 status (normal in the absence of doxycycline, or depleted in the presence of doxycycline, Figure 4A), 3×FlagNrf1 was susceptible to Proteinase K (PK) digestion in the absence of detergent, implying that the N-terminus was oriented towards the cytosol at all times. Interestingly, however, the orientation of the C-terminus of Nrf13×Flag showed a strong dependence on p97 status (Figure 4B). In mock-depleted cells, Nrf13×Flag was largely susceptible to digestion by PK (Figure 4B, compare lanes 4 and 5). In stark contrast, when p97 was depleted by the addition of doxycycline, Nrf13×Flag was largely resistant to PK (Figure 4B, compare lanes 7 and 8, or lanes 10 and 11). The different behaviors of 3×FlagNrf1 and Nrf13×Flag and the failure to observe detectable trimming of Nrf13×Flag are consistent with a topology wherein only a very small amino-terminal segment of newly-synthesized Nrf1 is exposed to the cytosol (amino acids 1–6), whereas the bulk of Nrf1 (amino acids 25–742) remains in the lumen of the ER when the retrotranslocation activity of p97 is blocked. Interestingly, when we overexposed our blots, a small, protease-resistant pool of p120 was detectable at steady-state (Figure 4C, lanes 1 and 2). This pool quickly disappeared upon chasing with cycloheximide plus MG132 (lanes 4–11). We propose that retrotranslocation of p120 constitutively occurs at such a fast rate that, at steady-state, only a tiny fraction of luminal p120 can be detected unless retrotranslocation is blocked by depletion of p97 as in Figure 4B.

Figure 4.
The C-terminus of Nrf1 is re-positioned from the lumen to the cytosol in a p97-dependent manner.

If the type II transmembrane orientation (Ncytosol/Clumen) for Nrf1 indicated by the protease protection experiments represents a true intermediate in its biogenesis, the accumulation of this species should be reversed upon restoration of p97 activity. To test if this is the case, we utilized the p97 inhibitor NMS-873. Whereas Nrf13×Flag was largely resistant to digestion by PK in HEK293 cells treated with NMS-873 (Figure 4D, lanes 1 and 2), it became sensitive to PK upon washout of NMS-873 in the presence of MG132 plus cycloheximide (Figure 4D, lanes 4 and 5). In this case, the Nrf1 was detected as a p120 species, because the tagged protein was chased to p110 less efficiently than the endogenous protein (Figure 3C), and the p110 that did form during the chase fractionated with soluble proteins upon preparation of the membrane fraction. Nrf1(m1)3×Flag mirrored the behavior of the wild-type protein (Figure 4D), indicating that the lack of processing observed for non-cleavable mutant was not due to defective retrotranslocation.

To corroborate the results from the protease protection experiments, we also monitored the glycosylation status of Nrf1. We reasoned that when the C-terminal domain (CTD) is oriented towards the lumen, Nrf1 should be glycosylated on a set of Asn-X-Ser/Thr sites C-terminal to the transmembrane domain (Zhang et al., 2007), whereas when the CTD is retrotranslocated to the cytosol, it should become degylcosylated. To evaluate the glycosylation status of Nrf1, we monitored its sensitivity to the degylcosylating enzyme Endoglycosidase H (Endo H). Nrf1 p120 that accumulated in cells treated with NMS-873 was sensitive to Endo H (Figure 4E, compare lanes 1 and 2 or 5 and 6), whereas it became largely resistant to Endo H upon being chased in the presence of MG132 plus cycloheximide (Figure 4E, compare lanes 3 and 4 or 7 and 8). This behavior is consistent with the protease protection data that indicated that the CTD of p120 flips from a luminal to a cytosolic orientation upon restoration of p97 activity. Notably, the cytosolically-oriented Nrf1 that accumulated upon restoration of p97 activity migrated more slowly than expected for Nrf1 that has been deglycosylated and proteolytically processed. This was particularly evident when the migration of the primary translation product for amino acids 104–742 was compared to the species of Nrf1 that accumulated in cells subjected to various perturbations (Figure 4—figure supplement 1). Our data suggest that once it is retrotranslocated to the cytosol, Nrf1 is not only deglycosylated and proteolytically cleaved after Trp103, but it also gains one or more additional modifications, the nature of which remains unknown.

Discussion

Mouse Nrf1 and its human ortholog TCF11 are transcription factors that enhance the accumulation of mRNAs that encode proteasome subunits upon inhibition of proteasome activity (Radhakrishnan et al., 2010; Steffen et al., 2010). Prior work established that Nrf1 is initially integrated into the ER membrane but is turned over rapidly (Biswas and Chan, 2009; Radhakrishnan et al., 2010). Upon proteasome inhibition, it accumulates, migrates to the nucleus, and activates gene expression. In this work, we make several critical new observations about the biogenesis and processing of Nrf1 that allow us to formulate a model that explains how it becomes activated in response to proteasome inhibition. Specifically, we show that the domains of Nrf1 that are implicated in transcriptional activation are initially translocated into the lumen of the ER, but subsequently are retrotranslocated to the cytosolic side of the membrane in a manner that depends on the AAA ATPase p97/VCP. In unperturbed cells, the retrotranslocated Nrf1 is promptly degraded by the proteasome, resulting in a futile cycle of continuous synthesis, membrane insertion, retrotranslocation, and degradation (Figure 5, left-hand side). However, in cells with compromised proteasome activity, some of the retrotranslocated Nrf1 escapes degradation and is cleaved on the N-terminal side of Leu-104 to yield a p110 fragment that is no longer tethered to the ER membrane and can migrate to the nucleus (Figure 5, right-hand side). Importantly, this cleavage event is essential for Nrf1-dependent activation of proteasome gene expression upon inhibition of the proteasome.

Figure 5.
A model to explain Nrf1 activation.

This model raises two important questions. First, what is the identity of the protease that clips p120 to yield p110? This remains unclear, but the available evidence rules out the most likely candidates. Cleavage by the site-1 and site-2 proteases involved in mobilization of SREBP appears unlikely, because CHO cell lines deficient in these activities (Rawson et al., 1997, 1998) were nonetheless able to accumulate p110 upon proteasome inhibition (R Rawson, personal communication). In addition, the proteasome, which cleaves Spt23 and Mga2 from the ER membrane in yeast (Hoppe et al., 2000), appears to be dispensable for cleavage because we observed robust formation of p110 in two different cell lines with three different proteasome inhibitors across a wide range of concentrations (Figure 1, and data not shown). Finally, an inhibitor of ER-localized rhomboid protease(s) (Pierrat et al., 2011) failed to impede the accumulation of p110 in cells treated with MG132 (data not shown). The second question is why does p110 not normally accumulate in unperturbed cells? Although it will be easier to address this question when the protease is in hand, a reasonable working hypothesis is that in unperturbed cells, p97-dependent retrotranslocation and proteasome-dependent degradation are so intimately coupled that there is no opportunity for the cleavage event to occur as Nrf1 is threaded from the membrane into the maw of the proteasome. However, when the proteasome is inhibited, Nrf1 accumulates on the cytosolic side of the membrane where it is susceptible to cleavage. Even if the protease occasionally intercedes before the proteasome can act and small amounts of p110 are generated, p110 is intrinsically unstable (Radhakrishnan et al., 2010) and thus does not accumulate to substantial levels.

It should be immediately evident from the foregoing discussion how the activation of Nrf1 differs markedly from the activation of all other ER membrane-tethered transcription factors that have been studied to date, including SREBP, ATF6, OASIS, CREB-H, Spt23, and Mga2 (Brown and Goldstein, 1997; Hoppe et al., 2000; Ye et al., 2000; Kondo et al., 2005; Zhang et al., 2006). In all of those cases, the precursor is inserted into the ER membrane such that the functional domains involved in transcription are displayed on the cytosolic side of the membrane. Cleavages in the cytosolic or intramembrane domains can therefore liberate soluble fragments that translocate to the nucleus. By contrast, since Nrf1 begins its life with its functional domains in the ER lumen, it must be retrotranslocated to the cytosol before processing can yield a functional fragment that can relocalize to the nucleus. However, because the p97-dependent retrotranslocation pathway traversed by Nrf1 is normally employed to clear misfolded proteins from the ER (Wolf and Stolz, 2012), the inevitable consequence of Nrf1 piggybacking on this mechanism is that it is rapidly degraded by the proteasome as soon as it gains access to the cytosol. This precludes the accumulation of Nrf1 precursor on the cytosolic side of the ER membrane, except under circumstances where there is insufficient proteasome activity to keep pace with the appearance of substrates.

Recently, Zhang and Hayes (Zhang and Hayes, 2013) put forth a model for Nrf1 activation that shares some features with the model proposed here but is very different in its key aspects, leading to a fundamentally different view regarding the biogenesis and maturation of p120 into an active transcription factor. Specifically, Zhang and Hayes proposed that p120 spans the ER membrane at least three times. They argued that turnover of this species is promoted by calpain, not the proteasome. p120 that escapes degradation was envisioned to remain stably integrated as a multispanning transmembrane protein until it is mobilized by an unknown signal, which triggers retrotranslocation of the luminal domains followed by their deglycosylation to yield active Nrf1. It was proposed that Nrf1 is later cleaved by the proteasome to yield lower molecular weight soluble species, but appearance of active Nrf1 precedes this cleavage event. However, the causal relationship between cleavage and activation could not be evaluated because the cleavage site(s) was not identified. Another key distinction is that Zhang and Hayes do not perform true chase experiments, and thus their data do not distinguish whether there is a precursor–product relationship between the various Nrf1 species, or they arise from newly-synthesized Nrf1 following different biogenetic pathways. We do not understand the basis for the multiple discrepancies between our data and those of Zhang and Hayes.

Our observation that a functional transcription factor can begin its life in the lumen of the ER is unexpected but not without precedent. Ricin, a plant toxin, and some bacterial toxins including cholera toxin, and shiga toxin likewise are retrotranslocated from the ER lumen to the cytosol, where they exert their biological functions (Inoue et al., 2011). Another example is the Hepatitis E Virus capsid protein ORF2, which co-opts the ERAD machinery to gain access to the cytosol (Surjit et al., 2007). To our knowledge, the only host protein that undergoes retrotranslocation from the ER but escapes degradation by the proteasome is calreticulin (Afshar et al., 2005; Gold et al., 2010). Upon inhibition of the proteasome, presumably the entire complement of ERAD substrates mimics Nrf1 in that these proteins are retrotranslocated from the luminal to the cytosolic side of the ER membrane but are not degraded. Our observations provoke the question of whether some of these proteins, like Nrf1, exert novel functions upon their accumulation in the cytosol. An obvious question that emerges from our findings is why is Nrf1 regulated by such a baroque mechanism? Targeting Nrf1 to the ER may ensure that the basal level of Nrf1 activity is essentially zero in the absence of proteasome stress, due to the normally tight coupling between retrotranslocation and degradation (Wolf and Stolz, 2012). In cells that experience insufficient proteasome activity—brought about either through environmental insults including proteasome or chaperone inhibitors produced by microbes, or through changes in gene expression—accumulation of active Nrf1 can restore an adequate level of proteasome function.

Our finding that p97 is required for the Nrf1-mediated proteasome bounce-back response potentially has therapeutic implications. Previously, we demonstrated that in cancer cells, depletion of Nrf1 slows down recovery of proteasome activity upon transient inhibition of the proteasome and enhances apoptosis caused by a covalent proteasome inhibitor (Radhakrishnan et al., 2010). Although inhibition of transcription factors with small molecules has proven difficult (Berg, 2008), p97 is a druggable target (Chou et al., 2011, 2013; Magnaghi et al., 2013). Indeed, recent papers have reported enhanced killing of cancer cells when proteasome and p97 inhibitors are used in combination (Auner et al., 2013; Chou et al., 2013). It will be interesting to see whether blockade of the Nrf1-mediated bounce-back response through inhibition of p97 enhances the efficacy of proteasome inhibitor therapy in multiple myeloma or enables expansion of proteasome inhibitor therapy into new indications.

Materials and methods

Constructs

3×FlagNrf1 (RDB-2411) and 3×FlagNrf1HA (RDB-2412) constructs have been described previously (Radhakrishnan et al., 2010). Human Nrf1 coding region along with a C-terminal 3×Flag sequence was cloned into pcDNA3.1+ (Invitrogen, Carlsbad, CA) to establish the wild-type construct Nrf13×Flag (RDB-2867). The following mutants were derived from Nrf13×Flag via site-directed mutagenesis using the indicated forward primers together with their corresponding reverse complement primers: Nrf1(m1)3×Flag (RDB-2868; forward primer: 5′-ACA GGT TCC AGG TGC CAA CCA CTG AGG TAG CTG CCG CGG CGG CTG CCC GAG ACC CAG AGG G-3′), Nrf1(m2)3×Flag (RDB-2869; forward primer: 5′-GGT GCC AAC CAC TGA GGT AGC TGC CGC GCT GGT TCA CCG AGA-3′), Nrf1(m3)3×Flag (RDB-2870; forward primer: 5′-CAC TGA GGT AAA TGC CGC GCT GGT TCA CCG AGA C-3′), Nrf1(m4)3×Flag (RDB-2871; forward primer: 5′-GAG GTA AAT GCC TGG GCG GTT CAC CGA GAC CC-3′), Nrf1(m5)3×Flag (RDB-2872; forward primer: 5′-CCA CTG AGG TAA ATG CCG CGG CGG TTC ACC GAG ACC CAG-3′), and Nrf1(m6)3×Flag (RDB-2873; forward primer: 5′-ACT GAG GTA AAT GCC TGG GCG GCT GCC CGA GAC CCA GAG GGG TC-3′).

The doxycycline-inducible shRNA expression construct pTRIPZ-p97sh (RDB-2874) targeting p97 was based on a 21-mer sequence (AAC AGC CAT TCT CAA ACA GAA) present in the coding region of both human and mouse genes and was cloned into the pTRIPZ vector (Open Biosystems, Huntsville, AL).

Cell culture

HEK-293T, HEK-293, and NIH-3T3 cell lines were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (Atlanta Biologicals, Norcross, GA), penicillin, and streptomycin (Invitrogen) at 37°C in 5% CO2. Immortalized Nrf1−/− mouse embryonic fibroblasts (MEFs) (Radhakrishnan et al., 2010) were grown as above except that the medium was additionally supplemented with β-mercaptoethanol and non-essential amino acids (Invitrogen).

Transient transfection and retroviral transduction

Transfection reagents Lipofectamine 2000 (for HEK-293T, HEK-293 cells) or Lipofectamine LTX (for mouse embryonic fibroblasts) were used for transient transfections as per manufacturer’s recommendations (Invitrogen).

For retroviral production, HEK-293T cells were transfected with the required retroviral construct along with helper plasmids. 48 hr after transfection, media supernatant containing the retrovirus was collected every 4–5 hr for 2 days. This retrovirus-containing medium, supplemented with polybrene (10 µg/ml), was used to transduce the target cells.

Stable cell lines

NIH-3T3-p97sh cell line (DTC-138) expressing doxycyline-inducible p97-specific shRNA was generated by transducing the mouse NIH-3T3 cells with the pTRIPZ-p97sh retroviral construct and selecting them in the presence of 7.5 µg/ml of puromycin. The HEK-293-p97sh cell line (DTC-139) was generated similarly except that 2 µg/ml puromycin was used for selection.

Stable cell lines HEK-293-Nrf13×Flag (DTC-140) and HEK-293-Nrf1(m1)3×Flag (DTC-141) were established by transfecting HEK-293 cells with the constructs Nrf13×Flag and Nrf1(m1)3×Flag respectively and subjecting the cells to 500 µg/ml Geneticin selection.

Immunoblot analysis

Cells were lysed in RIPA buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% NP40, 1% Na.Deoxycholate, 0.1% SDS, 1 mM EDTA) supplemented with a protease and phosphatase inhibitor cocktail (Pierce, Rockford, IL). Immunoblots were performed with antibodies specific for Nrf1 (8052S; Cell Signaling, Danvers, MA), Flag tag (A8592; Sigma–Aldrich, St. Louis, MO), HA tag (2013819; Roche Diagnostics, Indianapolis, IN), p97 (sc-20799; SantaCruz Biotechnology, Dallas, TX), Calnexin (sc-11397; SantaCruz Biotechnology), and β-actin (A5441; Sigma–Aldrich).

N-terminal Edman degradation sequencing

HEK-293T cell line was transiently transfected with the construct Nrf13×Flag and 48 hr later was treated with 5 µM MG132 for 5 hr. Cell lysate was prepared in lysis buffer (50 mM Tris pH 7.4, 0.5 M NaCl, 1 mM EDTA, 1% Triton X-100) supplemented with protease and phosphatase inhibitor cocktail (Pierce). The lysate was then subjected to immunoprecipitation with anti-Flag beads (Sigma-Aldrich) and eluted with Flag peptide (Sigma-Aldrich). The eluate was resolved by SDS-PAGE and transferred on to a PVDF membrane in the presence of 10 mM CAPS (3-cyclohexylamino-1-propane sulfonic acid), pH 11.0 and 10% Methanol. The membrane was stained with Ponceau S solution, and the Nrf1-specific bands (p120 and p110) were excised and sequenced using a protein micro-sequencer (Procise cLC 492A sequencer; Applied Biosystems, Foster City, CA).

Whereas the p120 sample confirmed the intact N-terminus, the p110 sample indicated a mix of amino acid sequences LVHRD and VHRD. The latter sequence is likely to be a frayed-end version of the former that was generated during sample processing (F Rusnak, personal communication), thus revealing the N-terminus of p110 to be Leu-104.

Quantitative reverse transcription PCR

RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA). cDNA was prepared using the Superscript III first strand synthesis kit (Invitrogen) according to the manufacturer’s recommendations. Quantitative PCR (qPCR) was performed using the SYBR GreenER supermix (Invitrogen). The primers used to quantify murine proteasome subunits (PSM) and GAPDH mRNA levels have been described previously (Radhakrishnan et al., 2010).

Pulse-chase assay

The pulse-chase experiment to establish the precursor–product relationship between Nrf1 p120 and p110 was carried out using the Click-iT Metabolic labeling and detection kits (Invitrogen) as per the manufacturer’s recommendations. Briefly, HEK-293 cells stably expressing either wild-type Nrf13×Flag or Nrf1(m1)3×Flag were starved for an hour in methionine-free medium and pulsed for an additional hour with 50 µM L-azidohomoalanine. Then the cells were washed with PBS and subjected to a chase with the addition of 5 µM MG132, 50 µg/ml cycloheximide, and 2 mM L-methionine. The cells were harvested at different time points and the cell lysates were prepared in lysis buffer (50 mM Tris pH 7.4, 0.5 M NaCl, 1 mM EDTA, 1% Triton X-100) supplemented with protease and phosphatase inhibitor cocktail (Pierce). The lysates were then used for immunoprecipitation with anti-Flag beads (Sigma-Aldrich). The immunoprecipitates were labeled with Tetramethylrhodamine (TAMRA)-alkyne and resolved by SDS-PAGE. The labeled bands were visualized in a Typhoon laser scan imaging system (GE Healthcare, Pittsburgh, PA).

Protease protection assay

The cells were washed once in PBS, resuspended in 10 mM Hepes-KOH pH 7.5 buffer and incubated on ice for 10 min. The swollen cells were then sedimented, resuspended in homogenization buffer (10 mM Hepes-KOH pH 7.5, 10 mM KCl, 1.5 mM MgCl2, 5 mM EGTA, and 250 mM sucrose) and passed through a 27G syringe needle several times. The homogenate was then subjected to serial centrifugations at 600×g (10 min), 3000×g (10 min) and 100,000×g (60 min). The microsomes collected at the end of the 100,000×g ultracentrifugation step were resuspended in membrane buffer (10 mM Hepes-KOH pH 7.5, 50 mM KOAc, 2 mM Mg(OAc)2, 1 mM DTT, and 250 mM sucrose).

Microsomes were mock-treated or subjected to Proteinase K (0.5 µg/µl) treatment either in the absence or presence of 1% Triton X-100 for 60 min on ice. The samples were then precipitated by trichloroacetic acid (TCA), resuspended in boiling sample buffer, resolved by SDS-PAGE and subsequently analyzed by immunoblotting with appropriate antibodies.

Deglycosylation assay

Cells were lysed in IP buffer (50 mM HEPES/KOH pH 7.5, 5 mM Mg(OAc)2, 70 mM KOAc, 0.2% Triton X-100; 10% glycerol, and 0.2 mM EDTA) supplemented with a protease and phosphatase inhibitor cocktail (Pierce). These cell lysates were used for deglycosylation reactions with the enzyme Endoglycosidase H for 1 hr at 37°C as per the manufacturer’s recommendations (New England Biolabs, Ipswich, MA).

Acknowledgements

We are grateful to the members of the Deshaies’ lab for helpful advice and critical reading of the manuscript. We thank R Rawson (UT Southwestern Medical Center at Dallas) for testing Nrf1 processing in S1P/S2P deficient CHO cells. We gratefully acknowledge the help of J Zhou and F Rusnak (Protein/Peptide Micro Analytical Laboratory, California Institute of Technology) for N-terminal sequencing of Nrf1. We thank Han-Jie Zhou and BioDuro, LLC for the synthesis of Rhomboid inhibitor. We thank Lev G Lis and Michael A Walters (University of Minnesota) for the synthesis of NMS-873. RJD is a Howard Hughes Medical Institute (HHMI) Investigator, and this work was supported in part by HHMI.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

This paper was supported by the following grant(s):

Howard Hughes Medical Institute to Raymond J Deshaies.
National Cancer Institute’s Howard Temin Pathway to Independence Award K99/R00 K99CA154884 to Senthil K Radhakrishnan.

Funding Information

This paper was supported by the following grants:

  • Howard Hughes Medical Institute to Raymond J Deshaies.
  • National Cancer Institute’s Howard Temin Pathway to Independence Award K99/R00 K99CA154884 to Senthil K Radhakrishnan.

Additional information

Competing interests

RJD: RJD is a founder and shareholder of Cleave Biosciences, and an eLife reviewing editor.

The other authors declare that no competing interests exist.

Author contributions

SKR, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article.

WB, Analysis and interpretation of data, Contributed unpublished essential data or reagents.

RJD, Conception and design, Analysis and interpretation of data, Drafting or revising the article.

References

  • Afshar N, Black BE, Paschal BM. 2005. Retrotranslocation of the chaperone calreticulin from the endoplasmic reticulum lumen to the cytosol. Molecular and Cellular Biology 25:8844–8853. 10.1128/MCB.25.20.8844-8853.2005 [PMC free article] [PubMed] [Cross Ref]
  • Andrews NC, Erdjument-Bromage H, Davidson MB, Tempst P, Orkin SH. 1993. Erythroid transcription factor NF-E2 is a haematopoietic-specific basic-leucine zipper protein. Nature 362:722–728. 10.1038/362722a0 [PubMed] [Cross Ref]
  • Auner HW, Moody AM, Ward TH, Kraus M, Milan E, May P, Chaidos A, Driessen C, Cenci S, Dazzi F, Rahemtulla A, Apperley JF, Karadimitris A, Dillon N. 2013. Combined inhibition of p97 and the proteasome causes lethal disruption of the secretory apparatus in multiple myeloma cells. PLOS ONE 8:e74415. 10.1371/journal.pone.0074415 [PMC free article] [PubMed] [Cross Ref]
  • Berg T. 2008. Inhibition of transcription factors with small organic molecules. Current Opinion in Chemical Biology 12:464–471. 10.1016/j.cbpa.2008.07.023 [PubMed] [Cross Ref]
  • Biswas M, Chan JY. 2009. Role of Nrf1 in antioxidant response element-mediated gene expression and beyond. Toxicology and Applied Pharmacology 244:16–20. 10.1016/j.taap.2009.07.034 [PMC free article] [PubMed] [Cross Ref]
  • Brown MS, Goldstein JL. 1997. The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89:331–340. 10.1016/S0092-8674(00)80213-5 [PubMed] [Cross Ref]
  • Chou TF, Brown SJ, Minond D, Nordin BE, Li K, Jones AC, Chase P, Porubsky PR, Stoltz BM, Schoenen FJ, Patricelli MP, Hodder P, Rosen H, Deshaies RJ. 2011. Reversible inhibitor of p97, DBeQ, impairs both ubiquitin-dependent and autophagic protein clearance pathways. Proceedings of the National Academy of Sciences of the United States of America 108:4834–4839. 10.1073/pnas.1015312108 [PMC free article] [PubMed] [Cross Ref]
  • Chou TF, Li K, Frankowski KJ, Schoenen FJ, Deshaies RJ. 2013. Structure-activity relationship study reveals ML240 and ML241 as potent and selective inhibitors of p97 ATPase. ChemMedChem 8:297–312. 10.1002/cmdc.201200520 [PMC free article] [PubMed] [Cross Ref]
  • Garrison JL, Kunkel EJ, Hegde RS, Taunton J. 2005. A substrate-specific inhibitor of protein translocation into the endoplasmic reticulum. Nature 436:285–289. 10.1038/nature03821 [PubMed] [Cross Ref]
  • Gold LI, Eggleton P, Sweetwyne MT, Van Duyn LB, Greives MR, Naylor SM, Michalak M, Murphy-Ullrich JE. 2010. Calreticulin: non-endoplasmic reticulum functions in physiology and disease. FASEB Journal 24:665–683. 10.1096/fj.09-145482 [PMC free article] [PubMed] [Cross Ref]
  • Grimberg KB, Beskow A, Lundin D, Davis MM, Young P. 2011. Basic leucine zipper protein Cnc-C is a substrate and transcriptional regulator of the Drosophila 26S proteasome. Molecular and Cellular Biology 31:897–909. 10.1128/MCB.00799-10 [PMC free article] [PubMed] [Cross Ref]
  • Hoppe T, Matuschewski K, Rape M, Schlenker S, Ulrich HD, Jentsch S. 2000. Activation of a membrane-bound transcription factor by regulated ubiquitin/proteasome-dependent processing. Cell 102:577–586. 10.1016/S0092-8674(00)00080-5 [PubMed] [Cross Ref]
  • Inoue T, Moore P, Tsai B. 2011. How viruses and toxins disassemble to enter host cells. Annual Review of Microbiology 65:287–305. 10.1146/annurev-micro-090110-102855 [PubMed] [Cross Ref]
  • Kensler TW, Wakabayashi N, Biswal S. 2007. Cell survival responses to environmental stresses via the Keap1-Nrf2-ARE pathway. Annual Review of Pharmacology and Toxicology 47:89–116. 10.1146/annurev.pharmtox.46.120604.141046 [PubMed] [Cross Ref]
  • Kobayashi A, Ito E, Toki T, Kogame K, Takahashi S, Igarashi K, Hayashi N, Yamamoto M. 1999. Molecular cloning and functional characterization of a new Cap’n’ collar family transcription factor Nrf3. The Journal of Biological Chemistry 274:6443–6452. 10.1074/jbc.274.10.6443 [PubMed] [Cross Ref]
  • Kolawa N, Sweredoski MJ, Graham RL, Oania R, Hess S, Deshaies RJ. 2013. Perturbations to the ubiquitin conjugate proteome in yeast deltaubx mutants identify ubx2 as a regulator of membrane lipid composition. Molecular and Cellular Proteomics 12:2791–2803. 10.1074/mcp.M113.030163 [PMC free article] [PubMed] [Cross Ref]
  • Kondo S, Murakami T, Tatsumi K, Ogata M, Kanemoto S, Otori K, Iseki K, Wanaka A, Imaizumi K. 2005. OASIS, a CREB/ATF-family member, modulates UPR signalling in astrocytes. Nature Cell Biology 7:186–194. 10.1038/ncb1213 [PubMed] [Cross Ref]
  • Lee CS, Ho DV, Chan JY. 2013. Nuclear factor-erythroid 2-related factor 1 regulates expression of proteasome genes in hepatocytes and protects against endoplasmic reticulum stress and steatosis in mice. The FEBS Journal 280:3609–3620. 10.1111/febs.12350 [PMC free article] [PubMed] [Cross Ref]
  • Lee CS, Lee C, Hu T, Nguyen JM, Zhang J, Martin MV, Vawter MP, Huang EJ, Chan JY. 2011. Loss of nuclear factor E2-related factor 1 in the brain leads to dysregulation of proteasome gene expression and neurodegeneration. Proceedings of the National Academy of Sciences of the United States of America 108:8408–8413. 10.1073/pnas.1019209108 [PMC free article] [PubMed] [Cross Ref]
  • Magnaghi P, D’Alessio R, Valsasina B, Avanzi N, Rizzi S, Asa D, Gasparri F, Cozzi L, Cucchi U, Orrenius C, Polucci P, Ballinari D, Perrera C, Leone A, Cervi G, Casale E, Xiao Y, Wong C, Anderson DJ, Galvani A, Donati D, O’Brien T, Jackson PK, Isacchi A. 2013. Covalent and allosteric inhibitors of the ATPase VCP/p97 induce cancer cell death. Nature Chemical Biology 9:548–556. 10.1038/nchembio.1313 [PubMed] [Cross Ref]
  • Moi P, Chan K, Asunis I, Cao A, Kan YW. 1994. Isolation of NF-E2-related factor 2 (Nrf2), a NF-E2-like basic leucine zipper transcriptional activator that binds to the tandem NF-E2/AP1 repeat of the beta-globin locus control region. Proceedings of the National Academy of Sciences of the United States of America 91:9926–9930. 10.1073/pnas.91.21.9926 [PMC free article] [PubMed] [Cross Ref]
  • Motohashi H, O’Connor T, Katsuoka F, Engel JD, Yamamoto M. 2002. Integration and diversity of the regulatory network composed of Maf and CNC families of transcription factors. Gene 294:1–12. 10.1016/S0378-1119(02)00788-6 [PubMed] [Cross Ref]
  • Oyake T, Itoh K, Motohashi H, Hayashi N, Hoshino H, Nishizawa M, Yamamoto M, Igarashi K. 1996. Bach proteins belong to a novel family of BTB-basic leucine zipper transcription factors that interact with MafK and regulate transcription through the NF-E2 site. Molecular and Cellular Biology 16:6083–6095 [PMC free article] [PubMed]
  • Pierrat OA, Strisovsky K, Christova Y, Large J, Ansell K, Bouloc N, Smiljanic E, Freeman M. 2011. Monocyclic beta-lactams are selective, mechanism-based inhibitors of rhomboid intramembrane proteases. ACS Chemical Biology 6:325–335. 10.1021/cb100314y [PMC free article] [PubMed] [Cross Ref]
  • Polucci P, Magnaghi P, Angiolini M, Asa D, Avanzi N, Badari A, Bertrand J, Casale E, Cauteruccio S, Cirla A, et al. 2013. Alkylsulfanyl-1,2,4-triazoles, a new class of allosteric valosine containing protein inhibitors. Synthesis and structure-activity relationships. Journal of Medicinal Chemistry 56:437–450. 10.1021/jm3013213 [PubMed] [Cross Ref]
  • Radhakrishnan SK, Lee CS, Young P, Beskow A, Chan JY, Deshaies RJ. 2010. Transcription factor Nrf1 mediates the proteasome recovery pathway after proteasome inhibition in mammalian cells. Molecular Cell 38:17–28. 10.1016/j.molcel.2010.02.029 [PMC free article] [PubMed] [Cross Ref]
  • Rape M, Hoppe T, Gorr I, Kalocay M, Richly H, Jentsch S. 2001. Mobilization of processed, membrane-tethered SPT23 transcription factor by CDC48(UFD1/NPL4), a ubiquitin-selective chaperone. Cell 107:667–677. 10.1016/S0092-8674(01)00595-5 [PubMed] [Cross Ref]
  • Rawson RB, Cheng D, Brown MS, Goldstein JL. 1998. Isolation of cholesterol-requiring mutant Chinese hamster ovary cells with defects in cleavage of sterol regulatory element-binding proteins at site 1. The Journal of Biological Chemistry 273:28261–28269. 10.1074/jbc.273.43.28261 [PubMed] [Cross Ref]
  • Rawson RB, Zelenski NG, Nijhawan D, Ye J, Sakai J, Hasan MT, Chang TY, Brown MS, Goldstein JL. 1997. Complementation cloning of S2P, a gene encoding a putative metalloprotease required for intramembrane cleavage of SREBPs. Molecular Cell 1:47–57. 10.1016/S1097-2765(00)80006-4 [PubMed] [Cross Ref]
  • Steffen J, Seeger M, Koch A, Kruger E. 2010. Proteasomal degradation is transcriptionally controlled by TCF11 via an ERAD-dependent feedback loop. Molecular Cell 40:147–158. 10.1016/j.molcel.2010.09.012 [PubMed] [Cross Ref]
  • Surjit M, Jameel S, Lal SK. 2007. Cytoplasmic localization of the ORF2 protein of hepatitis E virus is dependent on its ability to undergo retrotranslocation from the endoplasmic reticulum. Journal of Virology 81:3339–3345. 10.1128/JVI.02039-06 [PMC free article] [PubMed] [Cross Ref]
  • Sykiotis GP, Bohmann D. 2010. Stress-activated cap’n’collar transcription factors in aging and human disease. Science Signaling 3:re3. 10.1126/scisignal.3112re3 [PMC free article] [PubMed] [Cross Ref]
  • Wang W, Chan JY. 2006. Nrf1 is targeted to the endoplasmic reticulum membrane by an N-terminal transmembrane domain. Inhibition of nuclear translocation and transacting function. The Journal of Biological Chemistry 281:19676–19687. 10.1074/jbc.M602802200 [PubMed] [Cross Ref]
  • Wojcik C, Yano M, DeMartino GN. 2004. RNA interference of valosin-containing protein (VCP/p97) reveals multiple cellular roles linked to ubiquitin/proteasome-dependent proteolysis. Journal of Cell Science 117:281–292. 10.1242/jcs.00841 [PubMed] [Cross Ref]
  • Wolf DH, Stolz A. 2012. The Cdc48 machine in endoplasmic reticulum associated protein degradation. Biochimica et Biophysica Acta 1823:117–124. 10.1016/j.bbamcr.2011.09.002 [PubMed] [Cross Ref]
  • Xie Y, Varshavsky A. 2001. RPN4 is a ligand, substrate, and transcriptional regulator of the 26S proteasome: a negative feedback circuit. Proceedings of the National Academy of Sciences of the United States of America 98:3056–3061. 10.1073/pnas.071022298 [PMC free article] [PubMed] [Cross Ref]
  • Ye J, Rawson RB, Komuro R, Chen X, Dave UP, Prywes R, Brown MS, Goldstein JL. 2000. ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs. Molecular Cell 6:1355–1364. 10.1016/S1097-2765(00)00133-7 [PubMed] [Cross Ref]
  • Ye Y. 2006. Diverse functions with a common regulator: ubiquitin takes command of an AAA ATPase. Journal of Structural Biology 156:29–40. 10.1016/j.jsb.2006.01.005 [PubMed] [Cross Ref]
  • Zhang K, Shen X, Wu J, Sakaki K, Saunders T, Rutkowski DT, Back SH, Kaufman RJ. 2006. Endoplasmic reticulum stress activates cleavage of CREBH to induce a systemic inflammatory response. Cell 124:587–599. 10.1016/j.cell.2005.11.040 [PubMed] [Cross Ref]
  • Zhang Y, Hayes JD. 2010. Identification of topological determinants in the N-terminal domain of transcription factor Nrf1 that control its orientation in the endoplasmic reticulum membrane. Biochemical Journal 430:497–510. 10.1042/BJ20100471 [PubMed] [Cross Ref]
  • Zhang Y, Hayes JD. 2013. The membrane-topogenic vectorial behaviour of Nrf1 controls its post-translational modification and transactivation activity. Scientific Reports 3:2006. 10.1038/srep02006 [PMC free article] [PubMed] [Cross Ref]
  • Zhang Y, Lucocq JM, Yamamoto M, Hayes JD. 2007. The NHB1 (N-terminal homology box 1) sequence in transcription factor Nrf1 is required to anchor it to the endoplasmic reticulum and also to enable its asparagine-glycosylation. Biochemical Journal 408:161–172. 10.1042/BJ20070761 [PMC free article] [PubMed] [Cross Ref]
2014; 3: e01856.
Published online Jan 21, 2014. doi:  10.7554/eLife.01856.009

Decision letter

Michael S Brown, Reviewing editor
Michael S Brown, UT Southwestern Medical School, United States;

eLife posts the editorial decision letter and author response on a selection of the published articles (subject to the approval of the authors). An edited version of the letter sent to the authors after peer review is shown, indicating the substantive concerns or comments; minor concerns are not usually shown. Reviewers have the opportunity to discuss the decision before the letter is sent (see review process). Similarly, the author response typically shows only responses to the major concerns raised by the reviewers.

[Editors’ note: although it is not typical of the review process at eLife, in this case the editors decided to include the reviews in their entirety for the authors’ consideration as they prepared their revised submission.]

Thank you for sending your work entitled “p97-dependent retrotranslocation and proteolytic processing govern formation of active Nrf1 upon proteasome inhibition” for consideration at eLife. Your article has been favorably evaluated by a Senior editor and three reviewers, two of whom are members of our Board of Reviewing Editors.

The reviewers have made some suggestions that you should consider prior to submitting a final manuscript. The most important question is whether you wish to revise Figure 5 to indicate that the proposed model for C-terminal retrotranslocation led by the C-terminus is only provisional and other models are also possible. We are including the reviews below for your information.

Reviewer #1:

This important paper documents a novel mechanism for control of proteasome biosynthesis through the p57-mediated retrotranslocation of Nrf1 from ER lumen to cytosol and eventually to the nucleus. The experiments are comprehensive and well performed.

I see only one major problem. The diagram in Figure 5 implies that Nrf1 is retrotranslocated with the COOH-terminus leading the way. This requires a new membrane-spanning segment that the authors do not document. It seems equally likely that the NH2-terminal transmembrane helix is actually an uncleaved signal sequence that may never dissociate from the Sec61 translocation machinery. In this case the retrotranslocation may proceed by simply reversing the original membrane insertion mechanism so that the NH2- terminus would lead the way. If the authors agree they should consider modifying Figure 5 to indicate this alternate mechanism for retrotranslocation.

Reviewer #2:

This manuscript reports on a mechanism for the regulated activation of the transcription factor Nrf1 in mammalian cells with diminished proteasome activity. Using a panel of suitably tagged mutant versions of Nrf1 and exploiting inhibitors of the AAA protein p97 and the proteasome, Radhakrishnan and colleagues provide compelling evidence that the bulk of Nrf1 (its C-terminus) is normally translocated into the lumen of the ER and then subject to proteasome-dependent ER-associated degradation (a process i.e. dependent on the activity of p97). Under conditions of diminished proteasome activity, the C-terminal portion of Nrf1 is untethered from its membrane anchor by a yet-to-be identified protease (that cleaves after Trp103) and finds its way to the nucleus where it serves to activate the transcription of genes encoding proteasome subunits.

This study is remarkable both in terms of its physiological significance (clarifying a mechanism relevant to the regulated expression of a key component of the cell's protein degradation apparatus) and in terms of the level of detail it provides on the molecular aspects. The cleavage site of Nrf1 is precisely mapped and, crucially, the proposed unusual itinerary of Nrf1's C-terminal portion is confirmed by tracing the glycosylation status of the various trapped intermediates. Thus, this study will be of interest to a wide range of investigators.

Two specific comments follow (the first critical and the second speculative):

The critical comment:

The supplement to Figure 4 is a bit cryptic. The experiment exploits cotransin - presumably the compound described by Taunton's group in 2005 (PMID: 16015336) to inhibit Sec61-mediated translocation of something. But neither the logic of the experiment nor the provenance of cotransin (Taunton's paper is not cited), nor the significance of the observations are discussed in any detail. eLife's readers would benefit from rectification of these shortcomings.

The speculative comment:

Comparing the mobility of the P110 fragment synthesized in vitro, with that isolated from cells, the researchers conclude that the latter is subject to a novel post-translational modification. This is likely true, but it is worth perhaps considering the possibility that the modification in question may be a direct consequence of cytosolic deglycosylation of the single (predicted) N-linked glycan in Nrf1, which would convert Asn 347 to an aspartic acid (EndoH results in a different remnant, a GlcNAc attached to Asn 347 accounting perhaps for the different mobility of bands c and b in Figure 4). If the deglycosylated residue were important to P110's activity, this could provide a satisfying explanation for Nrf1's unusual itinerary: its purpose is to convert Asn347 to Asp347. This idea would be supported (though by no means proven) by judicious mutagenesis of Asn347.

Reviewer #3:

This manuscript demonstrates a novel transcriptional regulatory mechanism. The authors show that the transcriptional activation domain of Nrf1 is localized in the ER lumen after its initial synthesis, and then subject to constitutive retrotranslocation followed by rapid proteasomal degradation under normal circumstance. Upon inhibition of proteasomes, a proteolytic event-generated C-terminal fragment of the protein (p110) activates transcription of genes encoding proteasome subunits. These results are potentially interesting, but they are not enough to support the model shown in Figure 5, primarily because the authors fail to show how the precursor of Nrf1 (p120) associates with membranes.

1) The authors cited a previous study showing that Nrf1 (p120) is an intrinsic membrane protein. However, after reading the cited reference, this reviewer found that the particular study did not address whether the protein is really an intrinsic membrane protein or a soluble ER luminal protein. Indeed, the reviewer suspects that Nrf1 (p120) may be a soluble ER luminal protein as the so-called N-terminal transmembrane segment is not hydrophobic enough to be a transmembrane helix but appears to be a good candidate for a signal sequence. This hypothesis can be tested by alkaline treatment or other methods to break microsome membranes to see whether such treatment releases the protein from membrane pellets in cells in which p97 is depleted. An intrinsic ER membrane protein such as calnexin and a soluble ER luminal protein such as calreticulin should be used as controls in the experiment.

2) If NRF1 (p120) is a soluble ER protein, then after retrotranslocation it should be a soluble cytosolic protein. Under this condition, the proteolytic cleavage may allow p110 to enter the nucleus where it activates transcription of its target genes. This hypothesis should be tested by cell fractionation experiments to determine the subcellular localization of p120 and p110 in cells treated with a proteasomal inhibitor but with p97 present.

3) Even if NRF1 (p120) is an intrinsic membrane protein, the model shown in Figure 5 still does not make sense as it suggests the presence of a new transmembrane domain during the retrotranslocation. Is there any evidence supporting this hypothesis? A more likely model is that during retrotranslocation the transmembrane helix horizontally dips into membranes so that both the N and C terminal segment of the proteins can face the cytosol. If this is the case, the authors need to demonstrate that whereas p120 is an intrinsic membrane protein, p110 is a soluble protein that can enter the nucleus.

2014; 3: e01856.
Published online Jan 21, 2014. doi:  10.7554/eLife.01856.010

Author response

Reviewer #1:

This important paper documents a novel mechanism for control of proteasome biosynthesis through the p57-mediated retrotranslocation of Nrf1 from ER lumen to cytosol and eventually to the nucleus. The experiments are comprehensive and well performed.

I see only one major problem. The diagram in Figure 5 implies that Nrf1 is retrotranslocated with the COOH-terminus leading the way. This requires a new membrane-spanning segment that the authors do not document. It seems equally likely that the NH2-terminal transmembrane helix is actually an uncleaved signal sequence that may never dissociate from the Sec61 translocation machinery. In this case the retrotranslocation may proceed by simply reversing the original membrane insertion mechanism so that the NH2-terminus would lead the way. If the authors agree they should consider modifying Figure 5 to indicate this alternate mechanism for retrotranslocation.

We were attempting to imply as little as we possibly could because we do not really know in detail how the retrotranslocation process works. We originally drew the “new” transmembrane domain as being within the retrotranslocon, to imply that what is shown is an intermediate in the retrotranslocation/degradation process. However, we felt that this made the drawing excessively complicated. Another possibility in addition to the one mentioned by the reviewer is that the N-terminal membrane anchor domain switches polarity during the retrotranslocation process. Given that our efforts at simplification appear to have backfired, what we have done is to modify the drawing to include a putative retrotranslocon, and then state explicitly in the text of the figure legend that the drawing reflects one possibility among several.

Reviewer #2:

[…] This study is remarkable both in terms of its physiological significance (clarifying a mechanism relevant to the regulated expression of a key component of the cell's protein degradation apparatus) and in terms of the level of detail it provides on the molecular aspects. The cleavage site of Nrf1 is precisely mapped and, crucially, the proposed unusual itinerary of Nrf1's C-terminal portion is confirmed by tracing the glycosylation status of the various trapped intermediates. Thus, this study will be of interest to a wide range of investigators.

Two specific comments follow (the first critical and the second speculative):

The critical comment:

The supplement to Figure 4 is a bit cryptic. The experiment exploits cotransin - presumably the compound described by Taunton's group in 2005 (PMID: 16015336) – to inhibit Sec61-mediated translocation of something. But neither the logic of the experiment nor the provenance of cotransin (Taunton's paper is not cited), nor the significance of the observations are discussed in any detail. eLife's readers would benefit from rectification of these shortcomings.

The speculative comment:

Comparing the mobility of the P110 fragment synthesized in vitro, with that isolated from cells, the researchers conclude that the latter is subject to a novel post-translational modification. This is likely true, but it is worth perhaps considering the possibility that the modification in question may be a direct consequence of cytosolic deglycosylation of the single (predicted) N-linked glycan in Nrf1, which would convert Asn 347 to an aspartic acid (EndoH results in a different remnant, a GlcNAc attached to Asn 347 accounting perhaps for the different mobility of bands c and b in Figure 4). If the deglycosylated residue were important to P110's activity, this could provide a satisfying explanation for Nrf1's unusual itinerary: its purpose is to convert Asn347 to Asp347. This idea would be supported (though by no means proven) by judicious mutagenesis of Asn347.

Nrf1 actually contains 9 potential glycosylation sites but it is not known how many of them are modified. We presume from the context of his or her comment that the reviewer is referring to the bands b and e in Figure 4–figure supplement 1, not bands b and c. Although it is possible that the difference in mobility between these bands is due to conversion of a set of Asn residues to Asp, this seems unlikely given that the shift in MW is on the order of ~10 kD. Nevertheless, we now mention this possibility in the revised figure legend. It is possible that conversion of Asn to Asp results in activation of Nrf1, and in fact it has been suggested that this is the case (Zhang et al., 2009 Biochem J. 418, 293-310). In the Zhang study, they created a mutant in which 7 putative glycosylation sites were changed to aspartate. They reported enhanced activity for the mutant construct on a synthetic reporter, but only two-fold (Figure 3C of their paper). But, there are a number of problems with their experiment. First of all, it was done by transient transfection and there is no information regarding the degree of overexpression. Second, the authors did not quantify their Western blot of the wild type and mutant constructs and it actually appears that there might be a bit more of the mutant protein, so it is unclear if increased transcription of the reporter gene is due to increased expression or increased specific activity. Finally, it was not established that each of the sites that is mutated is actually glycosylated. Even if the Zhang result is correct, a two-fold effect is of questionable significance. This experiment would need to be done under more physiological conditions to test properly the interesting hypothesis raised by the reviewer.

Reviewer #3:

This manuscript demonstrates a novel transcriptional regulatory mechanism. The authors show that the transcriptional activation domain of Nrf1 is localized in the ER lumen after its initial synthesis, and then subject to constitutive retrotranslocation followed by rapid proteasomal degradation under normal circumstance. Upon inhibition of proteasomes, a proteolytic event-generated C-terminal fragment of the protein (p110) activates transcription of genes encoding proteasome subunits. These results are potentially interesting, but they are not enough to support the model shown in Figure 5, primarily because the authors fail to show how the precursor of Nrf1 (p120) associates with membranes.

1) The authors cited a previous study showing that Nrf1 (p120) is an intrinsic membrane protein. However, after reading the cited reference, this reviewer found that the particular study did not address whether the protein is really an intrinsic membrane protein or a soluble ER luminal protein. Indeed, the reviewer suspects that Nrf1 (p120) may be a soluble ER luminal protein as the so-called N-terminal transmembrane segment is not hydrophobic enough to be a transmembrane helix but appears to be a good candidate for a signal sequence. This hypothesis can be tested by alkaline treatment or other methods to break microsome membranes to see whether such treatment releases the protein from membrane pellets in cells in which p97 is depleted. An intrinsic ER membrane protein such as calnexin and a soluble ER luminal protein such as calreticulin should be used as controls in the experiment.

The reviewer’s argument is contradicted by the data we show in Figure 4A,B. In cells depleted of p97, essentially 100% of the accumulated p120 molecules have an N-terminal Flag tag removed in a protease protection experiment conducted in the absence of detergent (panel A), whereas a C-terminal Flag tag is completely protected (panel B). The most logical conclusion from this experiment is that the N-terminus faces the cytosol and the C-terminus faces the lumen (i.e., Nrf1 cannot be completely luminal). Regarding the point about the N-terminal transmembrane domain, we analyzed the Nrf1 sequence using the transmembrane domain prediction program TMpred (http://www.ch.embnet.org/software/TMPRED_form.html), which indicated a strong transmembrane domain (amino acids 7 to 26).

For the sake of thoroughness, we have also done membrane extraction experiments to address this issue, but the results have not been straightforward. About half of the Nrf1 is carbonate extractable in cells depleted of p97 or treated with p97 inhibitor (by contrast, 100% of calnexin resists carbonate extraction in both cases). By contrast most of Nrf1 is carbonate extractable in cells treated with MG132, consistent with the idea that it has been dislocated from the ER membrane. In cells treated with both p97 and proteasome inhibitors, the p97-deficient phenotype is epistatic. Thus, while the protease protection suggests that essentially all Nrf1 molecules at a p97 block are transmembrane, the carbonate extraction suggests that roughly 50% of these transmembrane species remain in an aqueous compartment. The simplest resolution of this is to hypothesize that a significant fraction of full-length Nrf1 molecules remain associated with the translocon, reminiscent of what was recently suggested for Deg1-Sec62 chimeras (Rubenstein et al., 2012 J. Cell Biol. 197:761-773). Nailing this point down would require considerable more experimentation but would not alter our fundamental conclusion that the C-terminus of Nrf1 starts out its life in the ER lumen and p97 is required for its retrotranslocation, processing, and activity. Hence we decided to not include the extraction data.

2) If NRF1 (p120) is a soluble ER protein, then after retrotranslocation it should be a soluble cytosolic protein. Under this condition, the proteolytic cleavage may allow p110 to enter the nucleus where it activates transcription of its target genes. This hypothesis should be tested by cell fractionation experiments to determine the subcellular localization of p120 and p110 in cells treated with a proteasomal inhibitor but with p97 present.

It has been previously established that p110 is nuclear (e.g., Wang and Chan, 2006. J Biol Chem, 281:19,676-19,687), so we did not feel it was necessary to repeat these observations. As part of the extraction experiments described in the response to point #1 we observed that p110 largely sediments but is readily extractable from the pellet with either carbonate or non-ionic detergent, consistent with the hypothesis that it is no longer anchored by a transmembrane domain.

3) Even if NRF1 (p120) is an intrinsic membrane protein, the model shown in Figure 5 still does not make sense as it suggests the presence of a new transmembrane domain during the retrotranslocation. Is there any evidence supporting this hypothesis? A more likely model is that during retrotranslocation the transmembrane helix horizontally dips into membranes so that both the N and C terminal segment of the proteins can face the cytosol. If this is the case, the authors need to demonstrate that whereas p120 is an intrinsic membrane protein, p110 is a soluble protein that can enter the nucleus.

What the reviewer proposes is very similar to our suggestion, described in the response to Reviewer #1, that the transmembrane domain may switch polarity during retrotranslocation. As described in more detail in the response to Reviewer #1, we have redrawn the figure and modified the legend to make more clear what we originally intended to show.


Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...