Figure 1.
The Neh3L-containing CTD of Nrf1 is conserved in the CNC-bZIP family.
(A) Schematic representation of discrete domains of Nrf1 and Nrf2. Locations of the ER signal, transactivation domains (TADs, including AD1, NST and AD2), DNA-binding domain (DBD, including CNC and bZIP) are indicated within Nrf1. The Neh3L region is situated within the C-terminal domain (CTD) of Nrf1. The positive regulation of Nrf2 by its Neh3 domain occurs through direct interaction with CHD6 [37], but it is not identified as one of Nrf1-interacting proteins [71]. (B) An alignment of amino acids covering CTD in Nrf1 and other CNC-bZIP factors with ER-resident proteins. The CNC family comprises both water-soluble members (i.e. NF-E2p45 and Nrf2) and membrane-bound NHB1-CNC members (including Nrf1, TCF11, Nrf3, CncC, and Skn-1, albeit the latter lacks both the corresponding ZIP and Neh3L regions). The distinction between Neh3L and Neh3 from Nrf1 and Nrf2 is attributable to different positioning relatively to membranes. Amongst the NHB1-CNC proteins, the core Neh3L is conserved with an ER-resident protein, omeg-3 fatty acid desaturase (O3FADS). Its N-terminally flanking CRAC motif is present in Nrf1 (numbered as CRAC5), TCF11 and Nrf3, but is absent from other members. The C-terminal basic cluster is predicted to possess an ER-retention signal (K/RxK/R), which ensembles to those in calnexin (CNX), O3FADS and Rit (Ras-like protein in all tissues). The conversed hydrophobic pentapeptide is boxed due to the representative in Nrf2 that is essential for its interaction with CHD6 [37]. (C) Bioinformatic prediction of three discrete regions within CTD of Nrf1. It is proposed that both CRAC5 and TMc sequences could be wheeled into two relative stable amphipathic helices only upon interaction with amphipathic membranes, whilst a positively-charged helix folded by the basic C-terminal peptide could interact electrically with the putative negatively-charged head group of membrane lipids. Three physico-chemical parameters related with the helical folding (i.e. aliphaticity, hydropathicity and amphipathicity) were calculated using the ProParam tool (http://web.expasy.org/protparam/).
Figure 2.
Nrf1 is negatively regulated by its CTD.
(A) Diagrammatic representation of various lengths of CTD in Nrf1 and its mutants. The putative secondary structure of discrete regions within CTD is shown (upper cartoon). (B) Luciferase activity was measured from COS-1 cells had been transfected with each of expression constructs for Nrf1 or its mutants (1.2 µg), together with PSV40Nqo1-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 Nrf1 or its mutants. Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (**, p<0.001, n = 9) in activity relatively to wild-type Nrf1 are indicated. (C) The above-prepared cell lysates (30 µg of protein) were resolved by gradient LDS/NuPAGE containing 4–12% polyacrylamide in a Bis-Tris buffer system and visualized by western blotting with antibody against the V5 epitope. The amount of proteins loaded into each electrophoresis sample well was adjusted to ensure equal loading of β-gal activity. An arrow indicates Nrf2 with a molecular mass of ∼80-kDa estimated (upper panel), whereas another arrow points to the brightly-contrasted band of ∼55-kDa Nrf1β (middle panel) that was cropped from the same gel as shown in the upper panel. GAPDH served as a protein-loading control (lower panel). It is notable that the same protein exhibits distinct mobility on different electrophoretic gels in different running buffer systems (cf. Figs. 2C with 5C). (D) COS-1 cells were co-transfected with 1.3 µg DNA of each of the above-described expression constructs and 0.2 µg of the ER/DsRed plasmid, and then allowed to recover from transfection for 24 h before being fixed. Subcellular location of proteins was examined by immunocytochemistry followed by confocal imaging. FITC-labelled second antibody was used to locate V5-tagged proteins. Nuclear DNA was stained by DAPI. The ER/DsRed gave a red image in the ER. The merge signal represents the results obtained when the three images were superimposed. (E) The quantitative data of imaging (corresponding to those shown in panel D) were calculated by determining the percentage of cells (at least 200 cells counted) in which the extra-nuclear stain, i.e. cytoplasmic plus ER (called simply C) was greater than or equal to the nuclear stain (called N), as opposed to the percentage of cells in which the extra-nuclear stain was less than the nuclear stain. Bar = 20 µm.
Figure 3.
Imaging of fixed and live cells expressing GFP fusion protein with CTD of Nrf1 or its mutants.
(A) Schematic of Six expression constructs for the GFP-CTD fusion protein and its mutants; these fusion proteins have been created by attachment of various lengths of CTD of Nrf1 to the C-terminus of GFP. (B) These indicated expression constructs each were transfected into COS-1 cells for 6 h. The cells were then allowed to recover from transfection in fresh medium for 18 h before being fixed by 4% paraformaldehyde and stained for the nuclear DNA by DAPI. The green signals from GFP were observed under confocal microscope and merged with the DNA-staining images. (C and D) Live-cell imaging of GFP-CTD and its mutant GFP-CTDΔ731–741(lacking its basic c-tail). COS-1 cells had been transfected with expression constructs for either GFP-CTD (C) or GFP-CTDΔ731–741 (D), together with the ER/DsRed marker, before being subjected to real-time live-cell imaging combined with the in vivo membrane protease protection assay. The cells were permeabilized by digitonin 20 µg/ml) for 10 min, before being co-incubated with PK (50 µg/ml) for 30 min. In the time course, real-time images were acquired using the Leica DMI-6000 microscopy system. The merged images of GFP with ER/DsRed are placed (on the third raw of panels), whereas changes in the intensity of their signals are shown graphically (bottom). 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). The arrow indicates a ‘hernia-like’ vesicle protruded from the cytoplasm.
Figure 4.
Live-cell imaging of both mutants GFP-CTDΔ723 −741 and GFP-CTDΔ714–722.
COS-1 cells co-expressing either GFP-CTDΔ723 −741 (A) or GFP-CTDΔ714–722 (B), along with the ER/DsRed marker, were subjected to live-cell imaging combined with the in vivo membrane protease protection assay, as described above in Figure 3. 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 5.
Opposing regulation of ARE-driven reporter genes by distinct Nrf1 isoforms.
(A) Schematic shows structural domains of five different isoforms of Nrf1. Locations of ER-targeting signal, AD1 and PEST2 are also indicated within distinct domains. (B) Shows luciferase reporter gene activity measured from COS-1 cells that had been co-transfected with 1.2 µg of each expression construct for Nrf1 isoforms, together with 0.6 µg of P-1061/nqo1-Luc (that is driven by the 1061-bp promoter of Nqo1) and 0.2 µg of β-gal plasmid. The data were calculated as a fold change (mean ± S.D) of transactivation by distinct Nrf1 isoforms. Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (**, p<0.001, n = 9) in activity were calculated relatively to the background activity (obtained from transfection of cells with an empty pcDNA3 with reporter plasmids). (C) Total lysates of COS-1 cells expressing each of Nrf1 isoforms or Nrf2 were resolved by 12% SDS-PAGE in a Bis-Tris buffer system and visualized by immunoblotting with the V5 antibody. The position of migration of the V5-tagged polypeptide was estimated to be 120, 95, 55, 46, 38, 36 and 25 kDa, and GAPDH was used as an internal control to verify amounts of proteins loaded into each electrophoretic well. (D) Nrf1γ inhibits transactivation of ARE-driven genes by Nrf1 or Nrf2. COS-1 cells were co-transfected with indicated amounts of expression constructs for Nrf1, Nrf1γ and/or Nrf2, together with 0.6 µg of P-1061/nqo1-Luc (D1) or PSV40Nqo1-ARE-Luc (D2) and 0.2 µg of β-gal plasmid. Thereafter, luciferase activity was measured and is shown as a fold change (mean ± S.D). Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (**, p<0.001, n = 9) in activity relatively to the background activity are indicated. (E) Total lysates of COS-1 cells co-transfected with expression constructs for Nrf1, Nrf1γ and/or Nrf2 alone or in combination (as indicated corresponding to those in panel D) was subject to separation by 4–12% LDS/NuPAGE in a Bis-Tris buffer system. The upper two panels represent similar images from different independent gels, on which location of Nrf2 migration is arrowed, whilst a non-specific protein band is starred (*). The position of the V5-tagged Nrf1 polypeptides of 120, 95, 85, 55, and 36 kDa is indicated. It is notable that the same proteins exhibit distinct mobilities on different electrophoric gels in different running buffer systems (cf. C with E).
Figure 6.
The weak activator Nrf1β/LCR-F1 is negatively regulated by its CTD.
(A) The middle schematic representation of Nrf1β, Nrf1γ and their deletion mutants lacking various lengths of aa 297–741 of Nrf1 (a3). The contributions of the deleted regions to changes in the activity of Nrf1β and Nrf1γ, when compared with the background value, were examined using the PSV40Nqo1-ARE-Luc reporter assay as described above. The right panel shows ARE-driven luciferase activity (a4) that was measured from COS-1 cells that had been co-transfected with each of numbered expression constructs and reporter plasmids. The data are shown as a fold change (mean ± S.D), and significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.05, **p<0.001, n = 9) are indicated, relatively to the background value from transfection with an empty pcDNA3 control vector alone (C). The left two panels show western blotting of some of the above-transfected cell lysates with antibodies against either V5 (a1) or Xpress (a2). In addition, a non-specific protein-band is indicated (by arrow). The amount of protein applied to each polyacrylamide gel sample well was adjusted to ensure equal loading of β-gal activity. (B) COS-1 cells were co-transfected with each of the above-numbered expression constructs for Nrf1β, Nrf1γ and their mutants, together with an expression vector for wild-type Nrf1 (N1) or Nrf2 (N2), PSV40Nqo1-ARE-Luc and β-gal plasmids. The cells were allowed to recover from transfection for 24 h before luciferase activity was measured. The data are shown as a fold change (mean ± S.D) of ARE-driven gene activity when compared with the background (value of 1.0). Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.05, **p<0.001, n = 9) are indicated. (C) The above-prepared cell lysates (b1 and b2) were resolved using 4–12% LDS/NuPAGE and visualized by western blotting with V5 antibody (c1 and c2). The electrophoresis band representing Nrf2 is indicated ((by arrow). The amount of protein loaded to each electrophoretic well was adjusted to ensure equal loading of β-gal activity.
Figure 7.
Blockage of Nrf1γ results an increase in the transactivation activity of Nrf1 and Nrf1β/LCR-F1.
(A) Confocal imaging of COS-1 cells that had been transfected with 1.3 µg DNA of each expression construct for Nrf1, Nrf1β and Nrf1γ or mutants, before their subcellular locations were then examined by immunocytochemistry with FITC-labelled second antibody in order to locate V5-tagged proteins. Nuclear DNA was stained by DAPI. The merge signal represents the results obtained when the two images were superimposed with DIC from normal light microscopy. Bar = 20 µm. The quantitative data (bottom) were calculated as described in Figure 2D. (B) Western blotting of COS-1 cells that had been transfected with the indicated expression constructs for V5-tagged Nrf1, Nrf1β and Nrf1γ and their point mutants (Met into Leu, below). The right panel shows that the same gel as the left panel was exposed to X-ray for a little longer time. Two bands representing the 36-kDa Nrf1γ and a 38-kDa polypeptide are indicated (arrows). GAPDH served as an internal control to verify the amount of proteins applied to each electrophoresis well. (C) Schematic representation of Nrf1, Nrf1β, and their Met-to-Leu mutants with various deletions. The upper left panel shows amino acids adjoining five numbered Met residues; their mRNA codons can be recognized by ribosome for the internal initiation to translate Nrf1β or Nrf1γ. The first four or all five Met-to-Leu mutants were made respectively to yield Nrf14xM/L and Nrf15xM/L, whilst Nrf1β M/L contains the fifth Met-to-Leu mutant. Additional deletion mutants were created on the base of Nrf15xM/L and Nrf1βM/L. The right panel shows luciferase reporter activity of COS-1 cells that had been transfected with 1.2 µg of each of indicated expression constructs, together with 0.6 µg of PSV40nqo1-ARE-Luc and 0.2 µg of β-gal plasmids. The data are shown graphically as fold changes (mean ± S.D.) of transactivation by indicated factors. Significant increases ($, p<0.05 and $$, p<0.001, n = 9) in the activity are compared to the activity of the intact Nrf1 or Nrf1β.
Figure 8.
Both AD2 and NST domains positively regulates chimaeric Gal4-Nrf1 and Gal4-Nrf1β factors.
(A) Schematic representation of expression constructs for Gal4D (Gal4 DNA-binding domain) fusion proteins containing various portions of Nrf1 or Nrf1β (left panel). They were created by ligation of their encoding cDNA fragments into the BamHI/EcoRI sites of the pcDNA3/Gal4-V5 vector. The left panel shows Gal4D-directed reporter activity that was measured from COS-1 cells had been cotransfected with each of indicated expression constructs for the various Gal4D/Nrf1 fusion proteins (1.2 µg), together with PTKUAS×4-Luc (0.6 µg) and b-gal (0.2 µg) plasmids. The data are shown graphically as fold changes (mean ± S.D.) of transactivation by indicated Gal4-fusion factors when compared with the background (value of 1.0). Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.05, **p<0.001, n = 9) in activity relatively to the referenced activity are indicated (arrows). (B) The above-prepared cell lysates were resolved using 4–12% LDS/NuPAGE and examined by western blotting with V5 antibody. The electrophoretic bands representing free Gal4D and Gal4-Nrf1 fusion proteins are indicated. Samples loaded on each well were calculated to contain equal amounts of β-gal activity.
Figure 9.
Nrf2-target gene expression is up-regulated by AD2 and NST domains of Nrf1 within chimaeras N607:C270Nrf2 and Nrf1βN607:C270Nrf2, but is also down-regulated by CTD of Nrf1 within additional chimaeras Nrf2:C112 Nrf1.
(A) The cartoon shows structural domains of Nrf1 and Nrf2. The C-terminal residues 629–741 of Nrf1 (i.e. C112Nrf1) cover both its bZIP and CTD regions (upper). In Nrf2, the C270Nrf2 represents its C-terminal 270 aa between positions 328–597 that cover its Neh6, CNC, bZIP and Neh3 domains (lower). (B) Diagrammatic representation of chimaeras that were composed of various portions Nrf1 and Nrf2. The cDNA fragments encoding different portions of the N-terminal aa 1–607 of Nrf1 (e.g. N607Nrf1) and various lengths of the central aa 292–607 (i.e. Nrf1βN607) were ligated into the BamHI/EcoR1 sites of the Nrf2/pcDNA4His/Max B construct. Thus a series of chimaeric proteins were created by fusing different regions of either N607Nrf1 or Nrf1βN607 to the N-terminus of C270Nrf2. (C) The transactivation activity of the above-described chimaeric factors as well as wild-type Nrf1 and Nrf2. This was determined by using PSV40Nqo1-ARE-Luc and β-gal reporters that had been co-transfected with each of indicated expression constructs into COS-1 cells. The data are shown as fold changes (mean ± S.D.) of the transactivation activity when compared with the background (value of 1.0) that was measured from the blank co-transfection of cells with an empty pcDNA4 vector and the above two reporters. Thereafter, significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.05, **p<0.001, n = 9) in activity relatively to the referenced activity are indicated (arrows). (D) Additional three chimaeras are schematically shown (left panel), which were created by fusing the full-length Nrf2 to the N-terminus of either C112Nrf1 or its mutants. These expression constructs for Nrf2 and its chimaeric proteins, together with PSV40Nqo1-ARE-Luc and β-gal reporters, were co-transfected into either COS-1 cells (right panel of D) or RL-34 cells (E), before luciferase activity was assayed and the data are presented as fold changes (mean ± S.D). Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (**p<0.001, n = 9) in activity were calculated relatively to the activity arrowed. (F) Total lysates of COS-1 cells that had been co-transfected with expression constructs for Nrf1, Nrf2 and its three chimaeric proteins (shown in D) were resolved using 4–12% LDS/NuPAGE and then visualized by western blotting with antibodies against either Nrf2 or the V5 epitope (left and right panels, both blotting in the same gel-transferred nitrocellulose membranes). Amounts of protein loaded to each electrophoretic well were adjusted to ensure equal loading of β-gal activity. In addition, a non-specific protein band recognized by anti-Nrf2 antibody is arrowed (right panel).
Figure 10.
Endogenous genes are up-regulated by Nrf1 and Nrf1β/LCR-F1 but also down-regulated by Nrf1γ and Nrf1δ.
(A) Knockdown of Nrf1 by its targeting siRNA, which, along with a scramble siRNA (as an internal control), was transfected into HEK 293T cells as described previously [12] (and maintained in our laboratory). Subsequently, changes in the mRNA expression of both the endogenous Nrf1 per se and Nrf1-target genes were analyzed by real-time qPCR. The data are shown as fold changes (mean ± S.D) in gene knockdown by Nrf1-siRNA relatively compared to the scramble value (1.0 set). Significant decreases (*p<0.005, **p<0.001, n = 9) in gene expression relatively to the basal level are indicated. (B to D) Expression constructs for Nrf1 (B), Nrf1β (C), Nrf1γ and Nrf1δ (D) (2 µg of cDNA each, along with an empty pcDNA3 control vector) were transfected into HEK 293T cells. Thereafter, alterations in the expression of Nrf1-target genes were determined by real-time qPCR, and were calculated as fold changes (mean ± S.D) in gene regulation by distinct Nrf1 isoforms when compared to the background (value of 1.0). Significant increases ($, p<0.05 and $$, p<0.001, n = 9) and decreases (*p<0.005, **p<0.001, n = 9) in gene expression relatively to the basal level are indicated. (E and F) Nrf1 and Nrf1β, Nrf1γ and Nrf1δ were restored into Nrf1−/− MEFs, in which Nrf1 has been lost (see Fig. S3) before being transfected with expression constructs for distinct isoforms alone or in combination, which were indicated (+, 1 µg of cDNA; ++, 2 µg of cDNA). Subsequently, real-time qPCR was performed to determine changes in the expression of GCLM (E) and PSMB6 (F). The data are presented as folds (mean ± S.D) relatively to the blank transfection with pcDNA3 alone (value of 1.0). Significant decreases (*p<0.005, **p<0.001, n = 9) in gene expression were calculated when compared to the level of genes regulated by Nrf1 or Nrf1β (arrows).