Figure 1.
The NHB1-CNC subfamily of membrane-bound transcription factors.
(A) The structural domains of NF-E2 p45-related CNC-bZIP transcription factors have been identified by bioinformatic analyses of their amino acid sequences. The Neh4 and Neh5 domains, which act as transactivation domains (TADs) in Nrf2 [49], [75], are represented by Neh4L and Neh5L in other family proteins. In Nrf1, AD1 is an essential TAD, containing the PEST1, Neh2L, CPD and Neh5L subdomains (see Text). Neh2L contains the DIDLID/DLG element and the ETGE motif; both are present in CncC and Nrf2 where they regulate protein stability. In addition to AD1, the AD2 region also functions as a TAD in Nrf1 [6] and is conserved amongst all other CNC family members, where it has been labeled AD2L. The ER-targeting NHB1 peptide of Nrf1/TCF11 and its NST glycodomain [7] are represented in Nrf3, CncC and Skn-1. We propose that Nrf1, Nrf3, CncC and Skn-1 constitute a subfamily of CNC transcription factors, called NHB1-CNC, which are membrane-bound proteins that are glycosylated in the lumen of the ER. For definition of the major acronyms, see Box S1. (B) The conserved topological structure of NHB1-CNC factors within and around membranes is predicted by bioinfomatics. Their ER-targeting mechanism has been confirmed in Nrf1, Nrf3 and CncC [5], [11], to occur via the conserved TM1 motif. The ability of NHB1-CNC factors (except Skn-1) to bind ARE sequences in target gene promoter regions is mediated through their CNC/bZIP domains that are retained on the cyto/nucleoplasmic side of membranes. The DNA-binding activity of Skn-1 is attributed to its CNC domain [76]. The TADs of the membrane-bound factors are transiently translocated into the luminal side of the ER during the initial co-translactional topogenesis. When these factors are required to activate their target genes, the luminal TADs are repartitioned and dislocated/retrotranslocated out of the luminal side across membranes into the cytoplasmic and/or nucleoplasmic compartments, where they are presented to the general transcriptional machinery before transactivating target gene expression. In addition, the asterisk* indicates the presence of putative GSK-3 phosphorylation sites in Nrf1, TCF11 and Skn-1.
Figure 2.
Regulation of Nrf1 by glycosylation and deglycosylation of its NST domain.
(A) The left schematic illustrates structural domains of Nrf1 and its N/D-scanning mutants in the NST glycodomain. The right panel shows reporter gene activity measured after COS-1 cells had been cotransfected with each of expression constructs (1.2 μg), together with PSV40GSTA2-6×ARE-Luc (0.6 μg) and β-gal plasmid (0.2 μg), and allowed to recover in fresh media for an additional 24 h before lysis. The data were calculated as a fold change (mean ± S.D) of transactivation by N/D mutants of Nrf1, as described elsewhere [34]. Significant increases in activity, relative to wild-type Nrf1, are indicated: $, p<0.05 and $$, p<0.001, n = 9). (B and C) PNGase F-catalyzed deglycosylation was performed on total lysates of cells that expressed wild-type Nrf1, its N/D mutants (B, lanes 2 to 9) or N/Q mutants (C, lanes 11 to 19). The digest products were resolved by 4–12% LDS/NuPAGE and visualized by western blotting with V5 antibodies. (D) The left schematic depicts the N/Q-scanning mutants, and locations of the TM1 and TMi sequences (). The right panel shows the reporter gene activities produced by Nrf1 and its N/Q mutants. Significant decreases in activity are indicated: *, p<0.05 and **, p<0.001 (n = 9). (E and F) Inhibition of Nrf1 deglycosylation by C19, C24, C45 and Z-VAD-FMK (zVF) causes significant increases in the amount of the 120-kDa Nrf1 glycoprotein. COS-1 cells were cotransfected with an expression construct for wild-type Nrf1 or an empty vector (as a control), along with PSV40GSTA2-6×ARE-Luc and the β-gal plasmid. The cells were allowed to recover in fresh medium containing 5.5 mM glucose and 10% FBS for 8 h, before being treated for 18 h with the above chemicals in fresh medium with 10% dialyzed FBS that contained no added-glucose (i.e. ‘no-glucose’). Repression of Nrf1 activity by the PNGase inhibitors was analyzed by luciferase reporter assay (E), showing a significant difference (*p<0.05; n = 9) between the indicated inhibitors and DMSO. Expression of Nrf1 proteins was visualized by immunoblotting with V5 antibodies (F). β-actin was employed as an internal control for protein loading.
Figure 3.
Glucose deprivation activates Nrf1 through TADs other than the NST domain.
(A) Cells expressing wild-type Nrf1 were allowed to recover from transfection in fresh 5.5 mM-glucose-containing-medium for 8 h, and were thereafter cultured for a further 18 h in media containing 0, 1.1 or 25 mM glucose. The cell lysates were resolved by 4-12% LDS/NuPAGE, followed by immunoblotting with V5 antibodies to detect ectopic Nrf1 protein. (B) Increased activity of ectopic wild-type Nrf1 resulting from exposure to glucose deprivation (i.e. ‘no-glucose’) conditions ($$, p<0.001, n = 9) was determined by reporter gene assays, in which the transfected cells were allowed to recover for 8 h in medium containing 5.5 mM glucose before they were subjected to an additional 18-h culture in either glucose-free or 25-mM glucose medium. (C) Transactivation of an ARE-driven luciferase gene by Nrf1 or mutants, following 18-h no-glucose starvation, was calculated from three independent reporter gene assays. Significant increases in transactivation activity ($, p<0.05; $$, p<0.001, n = 9) and significant decreases (*, p<0.05; **, p<0.001, n = 9) are shown.
Figure 4.
AD1 contributes to Nrf1-mediated transactivation of ARE-driven reporter genes.
(A) The left schematic illustrates the relative positions of PEST1, Neh2L, CPD and Neh5L within AD1. The DIDLID/DLG element and the ETGE motif are situated in Neh2L, which overlaps PEST1. The right panel shows that discrete regions of AD2 make different contributions to Nrf1 activity. Cells were transfected with the indicated expression plasmids, along with that for GSTA2-6×ARE-Luc reporter construct. After recovery in 5.5 mM-glucose medium, the cells were cultured for a further 18 h in glucose-free or 25 mM-glucose-containing medium, before luciferase activity was measured. Significant decreases (**p<0.001, n = 9) relative to wild-type Nrf1 activity are indicated. (B and C) These samples were also subjected to western blotting and cross-reacting polypeptides were visualized by ECL.
Figure 5.
TADs are transiently translocated in the lumen of ER before transactivating Nrf1-target genes.
(A) AD1 was mapped by the introduction of eN glycosylation sites into Nrf1(1-7)xN/Q (Figure S3, A to C). Following treatment of cell lysates that expressed Nrf1 eN mutants with Endo H or PNGase F to deglycosylate proteins, the products were analyzed by LDS/NuPAGE containing 7% Tris-Acetate gel (a1) or 4–12% Bis-Tris gel (a2), before immunoblotting. (B) The activity of Nrf1(1-7)xN/Q and its eN mutants was determined using the GSTA2-6×ARE-Luc reporter. Significant increases ($, p<0.05, n = 9) and significant decreases (*, p<0.05; **, p<0.001, n = 9) in the transactivation activity are shown. (C) Total lysates of cells expressing Nrf1 eN mutants within AD2 and SR-PEST2 (Figure S3D) were deglycosylated by digestion with Endo H (c1) or PNGase F (c2). The electrophoretic mobilities of Nrf1 proteins were monitored by immunoblotting. (D) The activity of Nrf1(1-7)xN/Q and its eN mutants was determined using a GSTA2-6×ARE-Luc reporter assay. The statistical significance of data was calculated ($, p<0.05 and $$, p<0.001, n = 9).
Figure 6.
Partial repartitioning of the NST-adjoining TADs across membranes into the cyto/nucleoplasm.
(A) Schematic of a series of Nrf1 deletion mutants lacking discrete portions of AD1 (including Neh5L and DIDLID/DLG), TMi-containing NST, AD2, SR, TMp-containing Neh6L, and bZIP. In addition, the locations of the eN mutants are also indicated across the AD2, SR and Neh6L domains. (B and C) Cells expressing wild-type Nrf1 (b1), its mutant Nrf1Δ280-298, Nrf1Δ171-186 (b2), or others indicated (C) were subjected to subcellular fractionation, followed immediately by an intact ER membrane protection assay to measure the sensitivity of the ectopic proteins to digestion by PK (50 μg/ml); proteolysis was allowed to proceed in the presence or absence of 1% TX in reaction mixtures placed on ice. The products were examined by immunoblotting with polyclonal antibodies against Nrf1β before being re-probed with antibodies against calreticulin (CRT) as a marker for luminal proteins. The intensity of these blots was estimated by dividing the value for Nrf1 with that for CRT, and the relative percentage (%) amount of Nrf1 that remained after PK digestion was normalized to the total amount of Nrf1 in reactions without PK digestion. The results are shown graphically (c, mean ± S.D, n = 4), allowing the stability of different Nrf1 mutants in membrane PK protection reactions to be compared (see Figure S4). (D and E) Membrane PK protection reactions using intact ER-enriched fractions purified from cells expressing Nrf1Δ374-393, Nrf1Δ409-428, Nrf1Δ466-488, Nrf1Δ508-513 or Nrf1Δ519-537 proteins (D) or other mutants indicated (E). The relative percentage of protein remaining after PK digestion was calculated as described above. The results are shown graphically (e, mean ± S.D, n = 4), allowing the stability of different Nrf1 mutants in membrane PK protection reactions to be compared (also see Figure S6). (F) The left schematic shows Nrf1 mutants lacking various portions of the protein. Their contributions to changes in Nrf1 activity in response to glucose starvation, when compared with activity observed under 25 mM-glucose conditions (control), were examined using the reporter assay. Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.05, **p<0.001, n = 9) are indicated, relatively to the wild-type Nrf1 activity obtained from the 25 mM-glucose conditions.
Figure 7.
Live-cell imaging of Nrf1/GFP to determine its dynamic movement out of the ER into the cytoplasm.
COS-1 cells were co-transfected with expression constructs for Nrf1/GFP fusion protein and the ER/DsRed marker, and were then subjected to live-cell imaging combined with the in vivo membrane protease protection assay. (A) The cells were permeabilized by digitonin 20 mg/ml) for 10 min, (B) before being co-incubated with PK (50 mg/ml) for 35 min prior to addition of 1% Triton X-100. Over this time interval, real-time images were acquired using the Leica DMI-6000 microscopy system. The merged images of Nrf1/GFP with ER/DsRed are presented (on the third row of panels), whereas changes in the intensity of their signals are shown graphically (bottom). The characteristic features of the arrow-indicated cells are described in the main text. Overall, the images shown herein are a representative of at least three independent experiments undertaken on separate occasions that were each performed in triplicate (n = 9).
Figure 8.
A proposed model to explain the molecular mechanisms controlling Nrf1.
Since Nrf1 is a mobile membrane-associated protein that engages in dynamic topologies [29], we propose a model to explain the molecular mechanisms controlling both its post-translational processing and its activity. The model involves seven stages. I) After being targeted to the ER, Nrf1 is anchored in the membrane through TM1. II) The NST-adjoining TADs in Nrf1 are transiently translocated into the lumen, where they are glycosylated to yield a 120-kDa glycoprotein. III) During topogenesis, the TMi-adjacent amphipathic regions in Nrf1 are tethered to the luminal leaflet of the membrane, whilst TMp dynamically associates within membranes, and its flanking PEST2 and Neh6L may be partitioned into distinct compartments. During this stage, the basic CNC-bZIP domain is retained in the cyto/nucloplasm, and its connecting TMc region is likely to be either left in the cytoplasm or integrated into membranes. IV) Once the TMi region in Nrf1 is liberated from the restraint of its flanking glycopeptides, it is reintegrated into membranes. This process should enable repartitioning of AD2 and SR out of membranes enabling it to function as a TAD. V) When required, the luminal NST and AD1 are repartitioned across the membrane into the cyto/nucleoplasm, thereby enabling deglycosylation of Nrf1 to produce the 95-kDa active transcription factor that up-regulates genes through its TADs. VI) An 85-kDa cleaved isoform of Nrf1 is generated upon removal of the NTD, allowing it to be released into the nucleus where it transactivate ARE-driven genes. VII) Distinct degrons can trigger proteolysis of Nrf1 to yield 55-kDa Nrf1β/LCR-F1 (acting as a weak activator), and/or the dominant-negative 36-kDa Nrf1γ and 25-kDa Nrf1δ isoforms. Abbreviations: GTM, general transcriptional machineries; ‘Retro?’, an unidentified retrotranslocon complex.