Differential Yet Integral Contributions of Nrf1 and Nrf2 in the Human HepG2 Cells on Antioxidant Cytoprotective Response against Tert-Butylhydroquinone as a Pro-Oxidative Stressor

In the past 25 years, Nrf2 (nuclear factor erythroid 2-related factor 2, also called NFE2L2) had been preferentially parsed as a master hub of regulating antioxidant, detoxification, and cytoprotective genes; albeit as a matter of fact that Nrf1 (nuclear factor erythroid 2-related factor 1, also called NFE2L1)—rather than Nrf2—is indispensable for cell homeostasis and organ integrity during normal growth and development. Herein, distinct genotypic cell lines (i.e., Nrf1α−/−, Nrf2−/−ΔTA, and caNrf2ΔN) are employed to determine differential yet integral roles of Nrf1 and Nrf2 in mediating antioxidant responsive genes to tert-butylhydroquinone (tBHQ) serving as a pro-oxidative stressor. In Nrf1α−/− cells, Nrf2 was highly accumulated but also could not fully compensate specific loss of Nrf1α’s function in its basal cytoprotective response against endogenous oxidative stress, though it exerted partially inducible antioxidant response, as the hormetic effect of tBHQ, against apoptotic damages. By contrast, Nrf2−/−ΔTA cells gave rise to a substantial reduction of Nrf1 in both basal and tBHQ-stimulated expression levels and hence resulted in obvious oxidative stress, but it can still be allowed to mediate a potent antioxidant response, as accompanied by a significantly decreased ratio of GSSG (oxidized glutathione) to GSH (reduced glutathione). Conversely, a remarkable increase of Nrf1 expression resulted from the constitutive active caNrf2ΔN cells, which were not manifested with oxidative stress, whether or not it was intervened with tBHQ. Such inter-regulatory effects of Nrf1 and Nrf2 on the antioxidant and detoxification genes (encoding HO-1, NQO1, GCLC, GCLM, GSR, GPX1, TALDO, MT1E, and MT2), as well on the ROS (reactive oxygen species)-scavenging activities of SOD (superoxide dismutase) and CAT (catalase), were further investigated. The collective results unraveled that Nrf1 and Nrf2 make distinctive yet cooperative contributions to finely tuning basal constitutive and/or tBHQ-inducible expression levels of antioxidant cytoprotective genes in the inter-regulatory networks. Overall, Nrf1 acts as a brake control for Nrf2’s functionality to be confined within a certain extent, whilst its transcription is regulated by Nrf2.


Introduction
With the development of science and technology, much more effective compounds with antioxidant properties have been discovered or synthesized insofar as to prevent lipid and protein oxidation [1,2]. Amongst them, tert-butylhydroquinone (tBHQ) is a well-known small molecule phenolic antioxidant, which is a main metabolite of 3-tertbutyl-hydroxyanisole (BHA) in vivo in humans, dogs, and rats, since it is widely used as a preservative in oils and processed foods [3,4]. However, tBHQ (and its precursor BHA) is de facto identified as a double-faced compound with both effects to be exerted as an antioxidant and also a pro-oxidant in biological systems [5,6]. Of great note, such a double-bladed sword impact of tBHQ is further unraveled by chemoprotective and carcinogenic effects Collectively, these demonstrate that mouse Nrf1 (and its isoforms) fulfills an indispensable function in regulating critical genes for maintaining robust redox homeostasis and organ integrity, so that the normal physiological development and growth are perpetuated in life process. However, it is regrettable that these achievements are made mostly from mouse models. Such being the case, the underlying mechanism(s) by which human Nrf1 (or its derived isoforms) also contributes to similar pathophysiological cytoprotective responses remains elusive.
For this reason, we have established three specific-knockout cell lines by gene-editing of human Nrf1 or Nrf2 on the base of HepG2 cells (named Nrf1α −/− , Nrf2 −/−∆TA , and caNrf2 ∆N , respectively) [31,40]. Here, these three distinct genotypic cell lines together with wild-type cells were stimulated by tBHQ and subjected to a series of experimental interrogation of both basal and inducible expression levels of certain antioxidant, detoxification, and cytoprotective genes. The resulting evidence has been presented by us, revealing that human Nrf1 and Nrf2 can make differential, yet integral, contributions to synergistic regulation of antioxidant and detoxification genes induced by tBHQ as a pro-oxidative stressor. Of great note, it is plausible that the presence of Nrf1 determines the basal redox steady-state and normal antioxidant cytoprotective responses against endogenous oxidative damages and apoptosis, albeit Nrf2 is involved in this homeostatic function [12,14,21,31]. This study also provides a better understanding of the inter-regulatory roles of Nrf1 and Nrf2 within the redox control system, except that both factors can exert their specific yet combinational functions in the process. Thereby, such cautions should also be severely taken into account for us to develop new drugs targeting Nrf1 or Nrf2 alone or both in the biomedical translational study.

Cell Lines and Regents
The human hepatomacellular carcinoma (HepG2) cells (wide type, WT; i.e., Nrf1/2 +/+ ) were purchased from American Type Culture Collection (ATCC, Manassas, VA, USA). Three HepG2-derived cell lines with distinct knockout types of Nrf1α −/− (with a specific deletion mutant of full-length Nrf1/TCF11 and its derived isoforms), Nrf2 −/−∆TA (lacking its longer transactivation domain-containing fragment), or caNrf2 ∆N (i.e., a constitutive active mutant of Nrf2 that lacks its N-terminal Keap1-binding Neh2 domain) had been established in our laboratory, as described in detail by Qiu et al. [31]. It is also worth mentioning that the authenticity of HepG2 cell line had been confirmed by its authentication analysis and STR (short tandem repeat) typing map (which was carried out by Shanghai Biowing Applied Biotechnology Co., Ltd., Shanghai, China). All these cell lines were, separately, cultured in DMEM (Dulbecco's Modified Eagle's Medium) supplemented with 10% (v/v) FBS (fetal bovine serum) and 100 units/L double-antibiotic (penicillin and streptomycin, Solarbio, Beijing, China) and incubated at 37 • C in a humidified 5% CO 2 atmosphere.

Cell Viability with the MTT Assay
All the indicated experimental cells were digested by trypsin and diluted into a suspension of 5 × 10 4 cell/mL, before being seeded into 96-well plates (5 × 10 3 cells/well). After the cells were completely adherent to the plates, they were treated with tBHQ at Antioxidants 2021, 10, 1610 4 of 25 different concentrations (i.e., 0-100 µM) for 24 h, or with 50 µM of tBHQ for distinct time periods (i.e., 0-24 h). The cell viability was evaluated by assaying 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide (MTT, ST1537, Beyotime, Shanghai, China) to form an insoluble product formazan in all living cells. The resulting data were calculated by dividing the experimental absorbance by relevant control values. The results were shown as a percentage of mean ± SD (n = 5 × 3), which are representative of at least three independent experiments, each of which was performed in quintuplicates.

Quantitative RT-PCR Analysis of mRNA Expression
All experimental cell lines growing in logarithmic phases were digested with trypsin and diluted by a complete medium into the suspension of 3.5 × 10 5 cell/mL. Then equal amounts of cells were inoculated in 6-well plates (3.5 × 10 5 cells/well) and cultured until being completely adherent. Thereafter, they were treated for different time periods (i.e., 0, 4, 8, 12, 16, 20, or 24 h) with 50 µM of tBHQ. Subsequently, total RNA was isolated using a RNA extraction kit (TIANGEN, Beijing, China), 500 ng of which was then subjected to the reaction with reverse transcriptase (Promega, Madison, WI, USA) to synthesize the single strand cDNAs, that served as PCR templates. Lastly, both basal and tBHQ-induced mRNA expression levels of those indicated genes were detected by quantitative real-time PCR (RT-qPCR) with each pair of their primers (Table 1). The RT-qPCR reaction was carried out with GoTaq ® qPCR Master Mix (Promega, Madison, WI, USA) on a CFX96 instrument (Bio-Rad, Hercules, CA, USA). The specific reaction procedure was followed by all relevant experimental groups, which were first inactivated at 95 • C or 3 min, and then amplified by 40 reaction cycles of 15 s at 95 • C and 30 s at 60 • C. The resulting data were analyzed by the Bio-Rad CFX 96 Manager 3.0 software (Hercules, CA, USA), whilst β-actin expression level served an internal reference control. All the experimental values were further calculated by normalization to the basal values obtained from WT cells that had been treated with tBHQ for 0 h (this value of 1 is set). The results of all the examined genes were shown as fold changes (Mean ± SD, n = 3 × 3), which are representative of at least three independent experiments being each performed in triplicates.

Western Blotting Analysis of Protein Expression
All those indicated experimental cells were allowed for preparation of a suspension of 3.5 × 10 5 cell/mL, and seeded into 6-well plates at a final density of 3.5 × 10 5 cells/well. After the cells were completely adherent to the plates, they were treated with 50 µM tBHQ for distinct lengths of time (i.e., 0, 1, 2, 4, 8, 12, 16, 20, and 24 h). After they were collected in a lysis buffer, the proteins were extracted for total lysates, which were further diluted with a 3× loading buffer and denatured by boiling at 100 • C for 10 min. The resulting total proteins from each of experimental groups were subjected to separation by SDS Antioxidants 2021, 10, 1610 5 of 25 (sodium dodecylsulfate)-PAGE (polyacrylamide gel electrophoresis) gels containing 8% polyacrylamide (to resolve Nrf1 and Nrf2) or 10% polyacrylamide (to resolve GSR, GCLC, GCLM, GPX1, HO-1, NQO1, and TALDO), which were allowed for running at 50 v for 30 min and then changed into 100 v to continue running for 2 h, before being transferred on the PVDF (polyvinylidene fluoride) membrane (Millipore Co., Tullagreen, Ireland) at 200 A for 2 h. After the protein-botted membranes were blocked by 5% skimmed milk for 1 h, they were incubated with each of the indicated primary antibodies at 4 • C overnight, and then re-incubated with the secondary antibody at room temperature for 2 h. The protein blots were developed by the enhanced chemiluminescence as described previously [31]. The intensity of relevant immunoblots was calculated by using the Quantity One software (Bio-Rad Laboratories, Hercules, CA, USA) and also normalized to the value of β-actin as a loading control. The resulting data were also shown as fold changes (Mean ± SD, n = 3) relative to the respective controls.

Detection of Cellular ROS and Apoptosis by Flow Cytometry
After a suspension of each of experimental cell lines were prepared in a complete medium, equal amounts of cells were seeded in 6-well plates (4 × 10 5 cells/well) and then allowed for growth until being completely adherent. Subsequently, they were treated for different time periods (i.e., 0, 4, 16 h) with 50 µM of tBHQ. After collecting the cells in each group, they were resuspended in pre-cooled PBS (phosphate buffered saline). In order to determine intracellular ROS levels, all the experimental groups were exposed to 100 µM dichlofluorescein diacetate (DCFH-DA included in a detection kit, S0033S, Beyotime, Shanghai, China) for 30 min at the incubator (37 • C, 5% CO 2 ). After being washed twice with PBS, all they were centrifuged and resuspended in serum-free media. The resulting 2 7 -dichlofluorescein (DCFH) was detected at the excitation wavelength of 488 nm and the emission wavelength of 525 nm by a flow cytometry (FlowJo, Ashland, OR, USA). The final results were expressed by the fluorescence intensity of DCFH detected in distinct cell lines. Furthermore, all the experimental cells were treated as abovementioned method and collected by centrifuging at 1000× g for 5 min, and stained with a binding buffer containing of both Annexin V-FITC and propidium iodide (PI) for 15 min. After the cells were washed twice to remove the excess staining reagent, they were subjected to detection of cell apoptosis by flow cytometry. The resulting data were shown by different fluorescence intensity in distinct states of cells.
2.6. The Assays for Total, Reduced, and Oxidized Glutathione Levels All experimental cell lines were suspended in a complete medium, and then equal amounts of cells were allowed for growth in 6-well plates (4 × 10 5 cells/well) until being completely adherent. Thereafter, they were treated for different time periods (i.e., 0, 4, 16 h) with 50 µM of tBHQ. All the cells were collected in PBS and then subjected to the measurement of total glutathione, reduced glutathione (GSH) and oxidized glutathione (GSSG) by using a glutathione assay kit (A061-1, Nanjing Jiancheng, Nanjing, China) according to the manufacturer's instruction. Of note, two standards of GSH and GSSG were also prepared in the same assays. The assay was designed by employing an Ellman's reagent (5,5'-disulfidebis-2-nitrobenzoic acid, DNTB), which can react with GSH to form 2-nitro-5-thiobenzoic acid, a yellow product with an absorbance at a wavelength of 405 nm. In addition, the protein concentrations in all experiment cells were determined by the bicinchoninic acid assay (BCA, P1511, ApplyGene Co., Beijing, China) and used as an internal control for the normalization, along with relevant standard curves, in order to calculate amounts of total glutathione, GSSG, and GSH by the formula provided by this manufacturer. The final resulting data are shown by a ratio of GSSG to GSH levels.

Assays for ROS-Scavenging Activities of Superoxide Dismutase and Catalase
Equal amounts of each of experimental cell lines suspended in a complete medium were seeded in 6-well plates (4 × 10 5 cells/well) and then allowed for growth until being completely adherent. Thereafter, they were treated for different time periods (i.e., 0, 4, 16 h) with 50 µM of tBHQ. All groups of experimental cells had been collected and subjected to assays for ROS-scavenging activities of superoxide dismutase (SOD), that were determined according to the instruction of enhanced SOD assay kit (A001-3, Nanjing Jiancheng, Nanjing, China). Besides, another ROS-scavenging enzyme catalase (CAT) activity was detected through the instructions of CAT kit (BC0205, Solarbio, Beijing, China).

ARE-Luciferase Reporter Assays
All experimental cells were, separately, seeded into 12-well plates (1.5 × 10 5 cells/well), and allowed for growth to reach 80% of confluence, before the cells were co-transfected using a lipofectamine 3000 mixture with each of ARE (antioxidant response elements)-driven luciferase plasmids (which were made by inserting each of the indicated ARE sequences into the pGL3-Promoter vector) or non-ARE reporter plasmids (as an internal background control), together with an experiment construct for Nrf1, Nrf2, or an empty pcDNA3.1 vector. In this test, the Renilla expression by pRL-TK (a plasmid encoding renilla luciferase driven by the thymidine kinase promoter) served as an internal quality control for transfection efficiency. Thereafter, the luciferase activity was measured by the dual-luciferase reporter system (Beyotime, Shanghai, China). The resulting data were calculated as fold changes (mean ± SD, n = 3 × 3) relative to the controls, which are representative of at least three independent experiments being each performed in triplicates.

Statistical Analysis
All the relevant results are presented as mean ± SD (n = 3 × 3 or 5 × 3) relative to the indicated controls. The comparison of the various experimental groups and their corresponding controls was carried out by one-way ANOVA, and analyzed by the post-hoc test with Fisher's least significant difference (LSD). The differences in between distinct treatments were considered to be statistically significant at p < 0.05.

Different Effects of Nrf1 and Nrf2 on Cell Growth during tBHQ Intervention
To gain an insight into endogenous Nrf1-and Nrf2-mediated antioxidant responses to tBHQ, we here confirmed that four distinct genotypes of cell lines are true ( Figure 1A), as described previously [31,40]. Of striking note, human Nrf1α is manifested with four major isoforms, as identified by Xiang et al. [41], of which its A and B isoforms represent the fulllength glycoprotein and deglycoprotein of Nrf1, respectively, whilst its C and D isoforms denote two distinct lengths of Nrf1 s N-terminally-truncated isoforms. Specific knockout of Nrf1α (by its gene-editing to delete a very short segment adjoining its translational start codons) led to a complete loss of all four Nrf1α-derived isoforms in Nrf1α −/− cells, albeit with a retention of other two minor proteins Nrf1 ∆N and Nrf1β ( Figure 1A). By contrast, all four Nrf1α-derived isoforms A to D were also substantially diminished by Nrf2 −/−∆TA , but its B to D isoforms (with distinct potentials of its trans-activity) were significantly augmented by caNrf2 ∆N . Intriguingly, Nrf1β abundances were also markedly suppressed by Nrf2 −/−∆TA or caNrf2 ∆N . These imply that Nrf1 expression and processing may be monitored by Nrf2, besides itself. Conversely, only a major Nrf2 isoform-A, but not its isoforms B or C, was incremented in Nrf1α −/− cells ( Figure 1A). However, all three isoforms A to C of Nrf2 were completely abolished by specific deletion of its transactivation Neh4-Neh5 domains (to yield an inactive Nrf2 −/−∆TA mutant), but their disappearance seemed to be replaced by other three smaller isoforms with a faster electrophoretic mobility (which were generated from Nrf2 ∆TAD , a dominant-negative mutant retaining its prototypic DNA-binding activity competitively against relevant wild-type factors), when compared with those equivalents examined in WT cells. By contrast, three slightly shorter isoforms of caNrf2 ∆N (closely to wild-type isoforms A to C, respectively) were retained and enhanced by this constitutive active mutant factor, because the N-terminal Keap1-binding Neh2 domain of Nrf2 was removed from its genomic locus. Collectively, these indicate that Nrf2 expression and processing may also be monitored by itself, as well as by Nrf1, within an inter-regulatory feedback cycle.
be replaced by other three smaller isoforms with a faster electrophoretic mobility (which were generated from Nrf2 ΔTAD , a dominant-negative mutant retaining its prototypic DNAbinding activity competitively against relevant wild-type factors), when compared with those equivalents examined in WT cells. By contrast, three slightly shorter isoforms of caNrf2 ΔN (closely to wild-type isoforms A to C, respectively) were retained and enhanced by this constitutive active mutant factor, because the N-terminal Keap1-binding Neh2 domain of Nrf2 was removed from its genomic locus. Collectively, these indicate that Nrf2 expression and processing may also be monitored by itself, as well as by Nrf1, within an inter-regulatory feedback cycle. The MTT assay was repeated three times, each performed in at least quintuplicates. The expanded graphic areas (at 0 to 20 μM between 0 and 4 h) were placed in the interior windows. (D,E) Four distinct cell lines were or were not treated with 50 μM tBHQ for a short time (from 0 to 2 h), followed by western blotting of Nrf1 (d1), Nrf2 (d2) and other ARE-driven target gene products as indicated (d3, d4 and e1-e5). The intensity of immunoblots was calculated at the gray values listed under the corresponding protein bands with statistical analysis of significant increases (^, p < 0.01) and significant decreases (*, p < 0.01) in their expression levels. These data shown herein are representative of at least three independent experiments.
Next, the cytotoxic effect of tBHQ on the aforementioned four different cell lines was evaluated by a MTT assay for the formation of formazan precipitates with succinate dehydrogenase in the mitochondria of all living cells only, and changes in the absorbance The MTT assay was repeated three times, each performed in at least quintuplicates. The expanded graphic areas (at 0 to 20 µM between 0 and 4 h) were placed in the interior windows. (D,E) Four distinct cell lines were or were not treated with 50 µM tBHQ for a short time (from 0 to 2 h), followed by western blotting of Nrf1 (d1), Nrf2 (d2) and other ARE-driven target gene products as indicated (d3, d4 and e1-e5). The intensity of immunoblots was calculated at the gray values listed under the corresponding protein bands with statistical analysis of significant increases (ˆ, p < 0.01) and significant decreases (*, p < 0.01) in their expression levels. These data shown herein are representative of at least three independent experiments.
Next, the cytotoxic effect of tBHQ on the aforementioned four different cell lines was evaluated by a MTT assay for the formation of formazan precipitates with succinate dehydrogenase in the mitochondria of all living cells only, and changes in the absorbance were measured to reflect the cell viability. As shown in Figure 1B, the viability of three examined cell lines except wild-type (WT) cells was modestly decreased by intervention with 5 µM tBHQ, but 10 µM of this chemical enabled these cell viability to return closely to their basal levels (obtained from treatment of cells with the vehicle of 0.1% DMSO). Then, a relatively stable viability of Nrf1α −/− cells was maintained between 10-60 µM of tBHQ, followed by a gradual decrease to 80% of its viability until its concentration increased to 100 µM ( Figure 1B). By contrast, a narrow window of stable WT cell viability was defined by 5-20 µM tBHQ, followed by a fairly sloping downhill to 70% viability of the cells treated with 100 µM tBHQ, while the other two close smoothly growth curves emerged from 10 to 80 µM tBHQ treatments of either Nrf2 −/−∆TA or caNrf2 ∆N cell lines, before their viability decreased to 80% and 70%, respectively, upon treatment of 100 µM tBHQ ( Figure 1B).
Based on the dose-dependent effects, 50 µM tBHQ was selected for intervention of the above-described four cell lines to assess distinct time-dependent growth courses ( Figure 1C). The results showed that the viability of all four cell lines decreased to different extents of between 90% and 75% by tBHQ intervention for 1 h. Of note, the continuous treatment enabled the viability of WT and Nrf2 −/−∆TA cell lines to smoothly decrease to 85-80% or 75-75% from 2 h or 4 h to 24 h, respectively ( Figure 1C). By sharp contrast, the viability of Nrf1α −/− and caNrf2 ∆N cell lines appeared to elevate respectively to 100% or 90% in a modest 'bounce-back' response to tBHQ-continued treatment from 2 h to 4 h, and then both declined to 75% at 12 h of treatment. Thereafter, the viability of Nrf1α −/− cells continued to gradually reduce to 70% until 24 h of tBHQ treatment, whilst the viability of caNrf2 ∆N cells was maintained to 75% from 12 h to 24 h treatments. As such, all cell viability reached a relatively stable level of them after 16 h of t-BHQ intervention. Therefore, the optimal concentration of tBHQ and its optimal time course were selected in the follow-up experiments to assess the cytoprotective roles of Nrf1 and Nrf2 against this chemical. For this end, we mainly investigated their expressional differences between these four cell lines in responses to 50 µM tBHQ intervention for different time periods as indicated.

Short-Term Intervening Effects of tBHQ on Nrf1, Nrf2, and AREs-Driven Genes in Distinct Genotypic Cells
Herein, short-term effects of tBHQ intervention for 1-2 h on Nrf1 and Nrf2 were first examined by western blotting ( Figure 1D,E). The results showed that tBHQ treatment of WT cells caused modest increases in Nrf1-processed isoforms C/D, as well as Nrf1 ∆N ( Figure 1D (d1)). Such altered Nrf1 ∆N also emerged in Nrf1α −/− cells, albeit it lacked A to D isoforms, implying it is not originated from the full-length Nrf1α processing. By contrast, a slight enhancement in the remnant Nrf1α-derived isoforms and Nrf1 ∆N expression in tBHQ-treated Nrf2 −/−∆TA cells, but both their basal and tBHQ-stimulated levels were increased in caNrf2 ∆N cells ( Figure 1D (d1)). For Nrf2, its protein expression was more sensitive to tBHQ stimulation in WT cells, and also increased significantly after 1 h treatment ( Figure 1D (d2)), when compared with those in the other three cell lines. They appeared to be largely insensitive to tBHQ, even although altered Nrf2 expression levels were evidently enhanced in Nrf1α −/− and caNrf2 ∆N , except Nrf2 −/−∆TA , cells lines, but all three with no obvious changes after treatment with tBHQ.
For further insights into differential expression patterns of these antioxidant cytoprotective genes among different genotypic cell lines, the following experiments were performed by long-term stimulation of cells with tBHQ for 4 h to 24 h.

Long-Term Stimulating Effects of tBHQ on Nrf1, Nrf2, and Downstream Targets in Distinct Genotypic Cells
To give a proper understanding of long-term tBHQ-stimulated effects on Nrf1, Nrf2, and downstream genes, their mRNA expression levels were determined by quantitative real-time PCR (Figure 2). The results revealed an obvious increase of Nrf1 mRNA expression after 12-h tBHQ stimulation of WT cells; this increase was maintained to 24 h of treatment of this chemical (Figure 2A and Figure S1A). By contrast, basal mRNA expression level of Nrf1 (measured at 0 h) was substantially abolished or diminished by knockout of Nrf1α −/− or Nrf2 −/−∆TA , respectively. Therefore, although Nrf2 was rather highly expressed in Nrf1α −/− cells ( Figures 1D and 2B), the remnant Nrf1 shorter isoforms in Nrf1α −/− cells were insensitive to tBHQ (Figure 2A and Figure S1A), whereas the residual Nrf1 in Nrf2 −/−∆TA cells could also hardly trigger a marginal response to this chemical. Conversely, caNrf2 ∆N cells had given rise to a remarkable increase in basal Nrf1 mRNA levels, but only a modest tBHQ-inducible increase of Nrf1 expression was detected after 20-24 h stimulation of this cell line (Figure 2A and Figure S1A). Together, these results indicate that transcriptional expression of human Nrf1 gene is monitored by Nrf2, as well by Nrf1 itself, even in the response to tBHQ.
Treatment of WT cells with tBHQ caused a gradual modest induction of Nrf2 mRNA expression levels from 8 h to16 h, which was maintained to 20 h, followed by a marked peak of its induction at 24 h, of this chemical stimulation ( Figure 2B and Figure S1B). Both basal and tBHQ-stimulated Nrf2 expression levels were completely abolished by Nrf2 −/−∆TA . Rather, it is interesting that, even though basal Nrf2 mRNA expression was significantly augmented by Nrf1α −/− or caNrf2 ∆N , its tBHQ-stimulated expression levels were roughly unaffected or even partially reduced in such two distinct genotypic cell lines ( Figure 2B and Figure S1B). Collectively, these demonstrate that transcriptional expression of human Nrf2 gene is bidirectionally regulated by itself and Nrf1, but upon stimulation by tBHQ, itself regulation by Nrf2 per se appears to be attributable to its N-terminal Keap1-binding Neh2 domain of the latter CNC-bZIP (Cap'n'Collar basic region-leucine zipper) factor.
Besides Nrf1 and Nrf2, downstream target genes HO-1 ( Figure 2C and Figure S1C) and NQO1 ( Figure 2D and Figure S1D) were also induced by tBHQ treatment of WT cells in a time-dependent manner. Upon loss of Nrf1α-derived isoforms, significant increments in basal and tBHQ-stimulated mRNA expression levels of HO-1 and NQO1 were determined in Nrf1α −/− cells. The first sharp maximum peak of HO-1 occurred at 4 h of induction by tBHQ, followed by a gradual decline to 16 h and then the second peak at 20 h treatment of Nrf1α −/− cells ( Figure 2C and Figure S1C). By contrast, only a smooth increase of NQO1 was obtained from 4 h to 12 h of tBHQ induction of Nrf1α −/− cells to a higher level, which was then maintained at such high level until 24 h ( Figure 2D and Figure S1D). However, loss of Nrf2 −/−∆TA led to an evident diminishment or even abolishment in basal and tBHQ-stimulated expression levels of HO-1 and NQO1 ( Figure 2C,D), except for a marginal induction of NQO1 by tBHQ at 24 h. Conversely, constitutive active caNrf2 ∆N appeared to have no significant effects on both basal and tBHQ-stimulated expression of HO-1 and NQO1 ( Figure 2C,D), albeit with a weak induction of NQO1 by tBHQ at 24 h and another similar lower stimulation of HO-1 at 12 h to 20 h ( Figure S1C,D). Altogether, these data indicate that HO-1 and NQO1 serve as two representative targets of Nrf2, and both genes regulated by Nrf2 may also be monitored positively by its N-terminal Neh2 domain, aside from the potential negative regulation of Nrf2 and its targets by Nrf1.
of NQO1 by tBHQ at 24 h. Conversely, constitutive active caNrf2 ΔN appeared to have no significant effects on both basal and tBHQ-stimulated expression of HO-1 and NQO1 (Figure 2C,D), albeit with a weak induction of NQO1 by tBHQ at 24 h and another similar lower stimulation of HO-1 at 12 h to 20 h ( Figure S1C,D). Altogether, these data indicate that HO-1 and NQO1 serve as two representative targets of Nrf2, and both genes regulated by Nrf2 may also be monitored positively by its N-terminal Neh2 domain, aside from the potential negative regulation of Nrf2 and its targets by Nrf1. (K). Then, mRNA expression was calculated as described in "Section 2". The resulting data were shown as fold changes (mean ± SD, n = 3 × 3), which are representative of at least three independent experiments being each performed in triplicates. Significant increases ($, p < 0.05; $$, p < 0.01) and significant decreases (* p < 0.05; ** p < 0.01), in addition to the nonsignificance (#), were statistically analyzed when compared with the corresponding cell line controls (measured at 0 h), respectively. (L) A schematic representation of several major ROS-scavenging enzymes (e.g., SOD and CAT) and relevant redox signaling (e.g., GPX and GSR) to defend against oxidative stress over-stimulated by ROS along with oxygen free radicals. Of note, as two important antioxidant players, GSH and NADPH are yielded from cell metabolism through key . Then, mRNA expression was calculated as described in "Section 2". The resulting data were shown as fold changes (mean ± SD, n = 3 × 3), which are representative of at least three independent experiments being each performed in triplicates. Significant increases ($, p < 0.05; $$, p < 0.01) and significant decreases (* p < 0.05; ** p < 0.01), in addition to the non-significance (#), were statistically analyzed when compared with the corresponding cell line controls (measured at 0 h), respectively. (L) A schematic representation of several major ROS-scavenging enzymes (e.g., SOD and CAT) and relevant redox signaling (e.g., GPX and GSR) to defend against oxidative stress over-stimulated by ROS along with oxygen free radicals. Of note, as two important antioxidant players, GSH and NADPH are yielded from cell metabolism through key enzymes GCLC/M and TALDO, respectively. Thereby, these key gene products exert their vital redox-regulatory functions in antioxidant, detoxification, and cytoprotective processes.

Long-Term Stimulation of Human Antioxidant and Detoxification Genes by tBHQ in Distinct Genotypic Cells
It is of crucial antioxidant and detoxification to be merited by glutathione (GSH)conjugates in the redox signaling cycles. The intracellular biosynthesis of GSH, as an important cellular antioxidant, is controlled by a key rate-limiting enzyme consisting of both GCLC and GCLM subunits. As shown in Figure 2E,F, a time-dependent increment in the mRNA expression of GCLC and GCLM induced by tBHQ from 4 h to 24 h was determined in WT cells (also see Figure S1E,F). By contrast, Nrf1α −/− cells gave rise to a rapid induction of GCLC mRNA expression by 4-h of tBHQ stimulation, to a maximum peak similar to that of WT cells, which was then maintained at a higher level until 24 h ( Figure 2E and Figure S1E). However, no striking changes in both basal and tBHQ-stimulated GCLC expression were observed in Nrf2 −/−∆TA or caNrf2 ∆N cell lines, albeit with a few exceptions of marginal alternations ( Figure S1E). Interestingly, further examinations revealed that basal and tBHQ-stimulated GCLM expression levels were substantially augmented in Nrf1α −/− cells as the time was extended to 24 h treatment ( Figure 2F and Figure S1F), whilst caNrf2 ∆N cells only gave rise to a relatively lower induction of GCLM by tBHQ, although its basal expression was also certainly elevated at a similar level to that obtained from Nrf1α −/− cells. Of note, Nrf2 −/−∆TA cells could still retain a considerably lower induction of GCLM by tBHQ to the constructive basal level of caNrf2 ∆N ( Figure 2F and Figure S1F). Altogether, these demonstrate differential contributions of Nrf1 and Nrf2 to basal and inducible regulation of both GCLC and GCLM genes controlling GSH biosynthesis in distinct genotypic cells.
As a central enzyme of the intracellular antioxidant defense, GSR can reduce the oxidized glutathione disulfide (GSSG) to the sulfhydryl form (GSH). In such a thiol-based redox cycle, another key enzyme GPX1, belonging to the glutathione peroxidase family, can catalyze the glutathione to reduce hydrogen peroxide (H 2 O 2 ) and other organic hydroperoxides, in order to detoxify the oxidants and hence protect the cells from oxidative damages. Thereby, we examine the intervening effects of tBHQ on induction of GSR and GPX1 mRNA expression mediated by Nrf1 and/or Nrf2 in distinct genotypic cell lines. As anticipated, RT-qPCR results revealed that GSR mRNA levels were strikingly gradually upregulated by tBHQ stimulation of WT cells from 8 h to 24 h ( Figure 2G and Figure S1G), while GPX1 expression was unaffected by this chemical treatment ( Figure 2H and Figure  S1H). Upon knockout of Nrf1α −/− , basal GSR and GPX1 mRNA levels were markedly enhanced, but only modest induction of GSR, rather than GPX1, by tBHQ occurred from 4 h to 8 h and thereafter maintained at a maximum level that was yet lower than equivalent values measured from tBHQ-treated WT cells ( Figure 2G,H). Such tBHQ-trigged induction of GSR, as well as its basal expression levels, was substantially attenuated or abolished by knockout of Nrf2 −/−∆TA ( Figure 2G and Figure S1G). However, GPX1 was smoothly downregulated by tBHQ stimulation of Nrf1α −/− cells ( Figure 2H and Figure S1H), though hyper-expression of Nrf2 was preserved in this knockout cell line ( Figure 2B), but this effect appeared to be completely prevented by Nrf2 −/−∆TA ( Figure 2H and Figure S1H). Conversely, constitutive active caNrf2 ∆N only gave rise to a remarkable increase in basal expression levels of GPX1, but not GSR, except that both genes were almost insensitive to stimulation by tBHQ ( Figure 2G,H).
Furthermore, TALDO is a key enzyme of the non-oxidative pentose phosphate pathway (PPP) providing ribose-5-phosphate for nucleic acid synthesis and NADPH for lipid biosynthesis [42]. Notably, NADPH arising from this pathway enables glutathione to be maintained at a reduced state, thereby protecting sulfhydryl groups and cellular integrity from oxygen radicals. Here, our results unraveled a stepwise inducible increase of TALDO mRNA expression levels from 4 h to 20 h of its maximum stimulation by tBHQ of WT cells, before being maintained until 24 h of this chemical stimulation ( Figure 2I and Figure S1I). By contrast, Nrf1α −/− caused significant increments in basal and tBHQ-stimulated expression levels of TALDO from 4 h to 12 h of its rapidly inducible peak that was much higher than the values obtained from WT control cells, which was then largely retained to 24 h. Such induction of TALDO was almost abolished by Nrf2 −/−∆TA (except for a marginal induction of it by 24-h stimulation of tBHQ), and also suppressed by caNrf2 ∆N (albeit its basal levels were augmented) ( Figure 2I and Figure S1I). Collectively, these indicate that Nrf1 and Nrf2 exert differential yet integral roles in mediating the aforementioned antioxidant cytoprotective genes against tBHQ.
In addition, it is worth mentioning that metallothioneins (MT) cannot only maintain the metal homeostasis in vivo, but also serve as a redox buffer for ROS and other free radicals to play an essential role in the cytoprotective process [43]. However, our examinations of MT1E and MT2 unraveled that both genes were not merely insensitive to tBHQ, but were modestly downregulated by this chemical intervention of WT cells ( Figure 2J,K, and see Figure S1J,K). Of note, basal mRNA expression of MT1E, along with its inhibitory effect of tBHQ, was markedly diminished or completely abolished in caNrf2 ∆N or Nrf1α −/− cells, respectively ( Figure 2J and Figure S1J), but both cell lines gave rise to a remarkable increase of basal MT2 expression, aside from that tBHQ-stimulated expression of MT2 was elevated in caNrf2 ∆N , rather than Nrf1α −/− , cells ( Figure 2K and Figure S1K). By striking contrast, Nrf2 −/−∆TA cells could give rise to significant increases in basal and tBHQ-inducible mRNA expression profiles of MT1E from 4 h to 12 h of this chemical stimulation, prior to being maintained at a considerably higher levels until 24 h ( Figure 2J and Figure S1J), whereas basal MT2 expression level was modestly downregulated by Nrf2 −/−∆TA , but with a marginal induction by tBHQ treatment from 12 h to 24 h ( Figure 2K and Figure S1K). Together, these suggest a remarkable distinction in contributions of Nrf1 and Nrf2 to transcriptional regulation of MT1E and MT2, respectively.

Distinct Time-Dependent Effects of tBHQ on Nrf1, Nrf2, and Target Gene Expression in Different Cell Lines
As shown in Figure 3A (a1), Nrf1α-derived isoforms A to D were obviously enhanced after 4 h of tBHQ treatment in WT cells and then maintained to their considerably higher extents between 8 h and 24 h, as illustrated graphically ( Figure 3A (a6). Similarly, the abundance of Nrf2 proteins was rapidly significantly augmented by tBHQ stimulation of WT cells from 1 h to 24 h ( Figure 1D (d2) and Figure 3A (a2)). Although hyper-expressed Nrf2 was retained in Nrf1α −/− cells, its protein abundances were unaffected by tBHQ intervention of Nrf1α-specific knockout cells ( Figure 3B (b2, b6)), in which Nrf1α-derived isoforms were constitutively lacked, but its N-terminal portion-truncated Nrf1 ∆N abundances were rather promoted by tBHQ ( Figure 3B (b1)). By contrast, Nrf2 −/−∆TA cells could only yield considerably weaker abundances of Nrf1α-derived isoforms A to D, but they still were enabled to respond to tBHQ in a biphasic manner, with the first peak at 4 h of this stimulation and the recurring second peak at 20 h of stimulation ( Figure 3C (c1, c6)), whilst the remnant Nrf2 ∆TAD mutant proteins were evidently time-dependently inhibited by tBHQ intervention of Nrf2 −/−∆TA cells ( Figure 3C (c2)), as compared with the positive reference control of 4-h tBHQ-treated WT cells (as indicated in the same gels). Conversely, caNrf2 ∆N cells led to strikingly increased abundances of basal Nrf2 ∆N and Nrf1α-derived isoforms ( Figure 1D (d1, d2)), and also their time-dependent inducible expression changes were here determined in tBHQ-stimulated caNrf2 ∆N cells ( Figure 3D (d1, d2, d6)).
only a marginal induction of caNrf2 ΔN by tBHQ was observed, but this was accompanied by significant induction of HO-1 and NQO1 (but still were lower than the positive control level of 4-h tBHQ-treated WT cells), as well as Nrf1α-derived proteins (to a greater extent than the tBHQ-treated WT controls), which occurred in their distinct time-dependent courses ( Figure 3D (d1-d6)). Taken together, these results demonstrate that such two interregulatory factors of Nrf1 and Nrf2-together with distinct targets-could mediate differential yet integral responses to tBHQ intervention of different genotypic cell lines. c Figure 3. Long-term effects of tBHQ on protein abundances of Nrf1, Nrf2 and co-target genes. Experimental cells, WT (A), Nrf1α −/− (B), Nrf2 −/−∆TA (C), and caNrf2 ∆N (D), were or were not treated with 50 µM tBHQ for 0 to 24 h, before basal and tBHQ-inducible protein changes of Nrf1 (a1, b1, c1, d1), Nrf2 (a2, b2, c2, d2), HO-1 (a3, b3, c3, d3), and NQO1 (a4, b4, c4, d4) were determined by Western blotting with their indicated antibodies, whilst β-actin served as a loading control. The intensity of those immunoblots, representing different protein expression levels, was also quantified by the Quantity One 4.5.2 software (Bio-Rad, Hercules, CA, USA). The data were representative of at least three independent experiments, as shown graphically (in right panels), after being calculated by a formula of Ln([A] t /[A] 0 , in which [A] t indicated a fold change (mean ± SD) in each of those examined protein expression levels at different times relative to the corresponding controls measured at 0 h (i.e., [A] 0 ). In addition, it should be noted that two big red crosses represent the constitutive losses of Nrf1 or Nrf2 (b6, c6), respectively.
Next, further examinations revealed that tBHQ-inducible expression levels of HO-1 in WT cells were gradually incremented, as its intervening time extended from 4 h to 20 h, to a considerably higher level, before being slightly declined ( Figure 3A (a3, a6)), whilst NQO1-induced expression levels were rapidly triggered by 4-h of this stimulation to a certain extent, and then maintained until 24 h ( Figure 3A (a4, a6)). Of great note, although basal abundance of HO-1 was substantially diminished by Nrf2 −/−∆TA (Figure 1D (d3)), it remained to be significantly induced by tBHQ from 8 h to 24 h of stimulation to a maximum extent ( Figure 3C (c3, c6)), whilst another rapid modest induction of NQO1 by tBHQ were detected in Nrf2 −/−∆TA cells ( Figure 3C (c4, c6)), when they compared to a positive reference control obtained from 4-h tBHQ-treated WT cells (as shown in the same gels). In Nrf1α −/− cells, even though the hyper-expressed Nrf2 was insensitive to tBHQ, both HO-1 and NOQ1 were still rapidly induced by this chemical from 4 h to 8 h of stimulation and then maintained to their respective higher extents until 24 h ( Figure 3B  (b3, b4,b6)). Rather, only a marginal induction of caNrf2 ∆N by tBHQ was observed, but this was accompanied by significant induction of HO-1 and NQO1 (but still were lower than the positive control level of 4-h tBHQ-treated WT cells), as well as Nrf1α-derived proteins (to a greater extent than the tBHQ-treated WT controls), which occurred in their distinct time-dependent courses ( Figure 3D (d1-d6)). Taken together, these results demonstrate that such two inter-regulatory factors of Nrf1 and Nrf2-together with distinct targets-could mediate differential yet integral responses to tBHQ intervention of different genotypic cell lines.

Different Time-Dependent Effects of tBHQ on Antioxidant Cytoprotective Gene Expression in Distinct Cell Lines
As shown in Figure 4A, tBHQ stimulation of WT cells caused significant time-dependent induction of GCLC, GCLM, GSR, GPX1, and TALDO (a1 to a7). Amongst them, GCLC was relatively slowly induced after 8 h of tBHQ stimulation and then gradually incremented to a maximum inducible extent at 20 h of this chemical treatment, before being slightly declined (a1, a7). In contrast, GCLM, GSR, and TALDO was rapidly induced within 4 h of stimulation by tBHQ and then presented within stepwise ascending trends from 8 h to 24 h of this treatment (a2, a3, a5 and a7), whilst GPX1 induction by tBHQ appeared to rise and fall within a biphasic waving mode (a4, a7). Such differences in these examined enzymes may be attributable to distinct involvement of their upstream factors (e.g., Nrf1 and Nrf2) in the cellular response to tBHQ.
Next, we determine distinct contributions of Nrf1 and Nrf2 to alterations in tBHQstimulated abundances of antioxidant and detoxification enzymes. As revealed in Nrf1α −/− cells, GCLC and GCLM were successively induced by tBHQ from 4 h to 24 h of their maximum stimulation ( Figure 4B, (b1, b2, b7)), while a lag induction of GSR occurred at 8 h of tBHQ stimulation, which was maintained to 16 h and then declined gradually to its basal levels at 24 h of this treatment (b3 and b7). As such, TALDO only displayed modest induction by tBHQ in a biphasic stepwise, with the first induction at 8 h and the second higher induction at 20 h before be declined nearly to its basal level (b5 and b7), except largely no induction of GPX1 in Nrf1α −/− cells (b4 and b7), although hyper-active Nrf2 was retained. However, the inactive Nrf2 −/−∆TA could still give rise to gradual enhancements in inducible GCLC, GCLM, GPX1, and TALDO abundances from 4 h to 24 h of tBHQ stimulation ( Figure 4C (c1, c2, c4-c7)), albeit all occurred to lower extents than the positive control levels of 4-h tBHQ-treated WT cells (as in the same gels), whereas GSR was slightly downregulated by this chemical in Nrf2 −/−∆TA cells (c3 and c7). By sharp contrast, caNrf2 ∆N could also only lead to a significant increment in induction of GCLM by tBHQ from 4 h to 24 h ( Figure 4D (d2, d7)), in addition to only modest induction of GCLC, GPX1,TALDO, but not GSR, by this chemical stimulation, which was maintained from 8 h to 12 h and then declined to relatively lower levels (d1, d3-d7).  GCLC (a1, b1, c1, d1), GCLM (a2, b2, c2, d2), GSR (a3, b3, c3, d3), GPX1 (a4, b4, c4, d4), and TALDO (a5, b5, c5, d5), were determined by western blotting with the indicated antibodies. The intensity of relevant immunoblots representing different protein expression levels was also quantified by the Quantity One 4.5.2 software. The resulting data were then shown graphically (in right panels), after being calculated by a formula of Ln([A]t/[A]0), in which [A]t indicated a fold change (mean ± SD) in each of those examined protein expression levels at different times relative to the corresponding controls measured at 0 h (i.e., [A]0), which were representative of at least three independent experiments.
Next, we determine distinct contributions of Nrf1 and Nrf2 to alterations in tBHQstimulated abundances of antioxidant and detoxification enzymes. As revealed in Nrf1α −/− cells, GCLC and GCLM were successively induced by tBHQ from 4 h to 24 h of their maximum stimulation ( Figure 4B, (b1,b2,b7)), while a lag induction of GSR occurred at 8 h of tBHQ stimulation, which was maintained to 16 h and then declined gradually to its basal levels at 24 h of this treatment (b3 and b7). As such, TALDO only displayed modest induction by tBHQ in a biphasic stepwise, with the first induction at 8 h and the second higher induction at 20 h before be declined nearly to its basal level (b5 and b7), except largely no  GCLC (a1, b1, c1, d1), GCLM (a2, b2, c2, d2), GSR (a3, b3, c3, d3), GPX1 (a4, b4, c4, d4), and TALDO (a5, b5, c5, d5), were determined by western blotting with the indicated antibodies. The intensity of relevant immunoblots representing different protein expression levels was also quantified by the Quantity One 4.5.2 software. The resulting data were then shown graphically (in right panels), after being calculated by a formula of Ln([A] t /[A] 0 ), in which [A] t indicated a fold change (mean ± SD) in each of those examined protein expression levels at different times relative to the corresponding controls measured at 0 h (i.e., [A] 0 ), which were representative of at least three independent experiments.

Different Antioxidant Responses of Four Distinct Genotypic Cell Lines to tBHQ as a Pro-Oxidative Stressor
The above experiments revealed there exists a synergistic effect of those antioxidant and detoxification genes regulated by Nrf1 and/or Nrf2. Just such synergistic effects can fully ensure the stable and effective function of this antioxidant cytoprotective system (to yield GSH and NADPH) to remove the excessive ROS produced from oxidative stressor, so that a certain redox homeostasis is being maintained to ensure the proper physiological operation of a heathy body. As a general term, ROS represents a set of all oxygen-containing reactive substances, including superoxide anion, hydrogen peroxide, and relevant free radicals. To date, they remain to be hardly detected, owing to their characteristics of strong oxidative activity with such a short life to be rapidly scavenged and detoxified by antioxidants (i.e., GSH). As such, the intracellular redox state was herein measured directly by DCFH-DA, one of the most widely-used assays to evaluate the resulting oxidative damages, because it can react directly with ROS to give rise to an extremely sensitive, but impermeable, dichlofluorescein probe, as detected by flow cytometry [44]. As shown in Figure 5A, a left shift of the dichlofluorescein image resulted from 16-h tBHQ intervention of WT cells (a1, also see Figure S2), implying a relative decrease of intracellular ROS levels, when compared with the control image obtained from the untreatment with tBHQ (i.e., 0.1% DMSO vehicle at 0 h). Further examinations revealed that Nrf1α −/− or Nrf2 −/−∆TA gave rise to a significant increase in basal ROS levels under the vehicle-treated conditions, and also an evident left-shift of their tBHQ-intervening images to varying extents ( Figure 5A,B (a2,  a3) and Figure S2), when compared with those measured from the WT cells. These indicate that, despite loss of Nrf1 or Nrf2 alone, the remaining portions of both factors still enabled either Nrf1α −/− or Nrf2 −/−∆TA cell lines to be stimulated by tBHQ to trigger antioxidant cytoprotective responses against their endogenic oxidative stress. However, it is rather intriguing that almost no changes in both basal and tBHQ-stimulated dichlofluorescein images were determined in caNrf2 ∆N cells, when compared to WT cells ( Figure 5A,B (a4) and Figure S2). This implies that the N-terminal Keap1-binding domain of Nrf2 is required for mediating tBHQ-triggered antioxidant response.
Further glutathione assays unraveled that the ratio of GSSG to GSH was marginally reduced by tBHQ stimulation of WT cells ( Figure 5C). By sharp contrast, Nrf1α −/− led to a remarkable increase in its basal GSSG to GSH ratio, but significant decreases of this ratio occurred after tBHQ stimulation. This indicates putative endogenous oxidative stress to yield the excessive GSSG, more than GSH levels, in this Nrf2-hyperexpressed Nrf1α −/− cells, but the remaining antioxidant response in this knockout cell line may be still triggered by tBHQ. Contrarily, Nrf2 −/−∆TA and caNrf2 ∆N further caused substantial decreases in their basal GSSG to GSH ratio, although their stimulated ratios were less or not promoted by tBHQ, respectively ( Figure 5C). This implicates such two distinctive mutants can still enable to yield a certain amount of GSH in Nrf2 −/−∆TA and caNrf2 ∆N cell lines, but could not be enough to allow for effective conversion of GSH into GSSG, even under tBHQ-stimulated conditions.
To gain insights into the initial scavengers of ROS, the activity of two key enzymessuperoxide dismutase (SOD) and catalase (CAT)-was examined herein. As shown in Figure 5D, significant increases in the basal activity of SOD were determined in Nrf1α −/− , Nrf2 −/−∆TA , or caNrf2 ∆N cell lines, and tBHQ-stimulated activity of SOD was further promoted only in Nrf1α −/− , Nrf2 −/−∆TA , but not caNrf2 ∆N , cell lines. Of note, a longer term (of 16 h) treatment of Nrf2 −/−∆TA cells with tBHQ caused a substantial reduction of SOD activity to its basal level, which was, though, still higher than that measured from WT cells ( Figure 5D). However, caNrf2 ∆N cells displayed no significant changes in tBHQ-inducible SOD activity, albeit its basal activity was much more than that of WT cells. Additionally, no obvious changes in the SOD activity were detected in WT cells that had or had not been treated by tBHQ. However, further examinations revealed that CAT activity was evidently stimulated by tBHQ in WT cells ( Figure 5E). By contrast, basal CAT activity was increased in both cell lines of Nrf1α −/− and Nrf2 −/−∆TA , but its tBHQ-stimulated activity was markedly elevated only in Nrf1α −/− , rather than Nrf2 −/−∆TA , cells ( Figure 5E). Furthermore, tBHQ stimulation of caNrf2 ∆N cells caused a striking suppression or even complete abolishment of its inducible CAT activity at 4 h or 16 h of this treatment, respectively, although its basal activity was greatly substantially augmented. Such discrepant activities of SOD and CAT in between these cell lines, together with their differential expression results as published previously [45], demonstrate to be attributable to distinctive yet cooperative contributions of Nrf1 and Nrf2 at regulating different target genes. Further glutathione assays unraveled that the ratio of GSSG to GSH was marginally reduced by tBHQ stimulation of WT cells ( Figure 5C). By sharp contrast, Nrf1α −/− led to a remarkable increase in its basal GSSG to GSH ratio, but significant decreases of this ratio occurred after tBHQ stimulation. This indicates putative endogenous oxidative stress to yield the excessive GSSG, more than GSH levels, in this Nrf2-hyperexpressed Nrf1α −/− cells, but the remaining antioxidant response in this knockout cell line may be still triggered by tBHQ. Contrarily, Nrf2 −/−ΔTA and caNrf2 ΔN further caused substantial decreases in their basal GSSG to GSH ratio, although their stimulated ratios were less or not promoted by tBHQ, respectively ( Figure 5C). This implicates such two distinctive mutants can still enable to yield a certain amount of GSH in Nrf2 −/−ΔTA and caNrf2 ΔN cell lines, but could not

Distinct Roles of Nrf1 and Nrf2 in Different Cell Apoptosis Induced by tBHQ as a Pro-Oxidative Stressor
Further analysis by flow cytometry unraveled that only a few number of apoptotic cells were indeed examined in WT cells that had been intervened with tBHQ for 16 h ( Figure 6A,B). By contrast, a considerable augment in basal apoptosis of Nrf1α −/− cells reached to a much higher rate than that of the other cell lines, but its tBHQ-stimulated apoptosis was significantly decreased after intervention of Nrf1α −/− cells by this chemical for 4 h to 16 h ( Figure 6A,B), to a similar level to that of tBHQ-treated WT cells. This phenomenon appeared to be almost consistent with the results of changing ROS levels as detected above ( Figure 5A,B). These indicate that tBHQ can induce antioxidant cytoprotective response against endogenous oxidative stress arising from Nrf1α −/− cells, in which putative hyper-expressed Nrf2 may be allowed for a certain extent to ameliorate potential oxidative damage and apoptosis caused by loss of Nrf1α, but could not fully compensate the constitutive loss of Nrf1 s function. Contrarily, no significant differences in basal apoptosis of either Nrf2 −/−∆TA or caNrf2 ∆N cell lines were observed when compared with that of WT cells ( Figure 6A,B), but both mutants led to a modest or less increase in tBHQ-triggered apoptosis after intervention of Nrf2 −/−∆TA or caNrf2 ∆N cells, respectively. This indicates that the sensitivity to tBHQ cytotoxicity may be weakened by Nrf2-deficient mutants, allowing for the resistance of these two cell lines to a considerable extent of pro-oxidative stress. Overall, it could be concluded that both Nrf1 and Nrf2 play distinctive roles in mediating differential antioxidant cytoprotective responses against oxidative stress-induced apoptosis.
This concluding notion is supported by further luciferase reporter assays ( Figure 6C), in which the reporter gene was driven by two different ARE-battery sequences existing in the promoter region of human MT1E (i.e., MT1E-ARE1 and MT1E-ARE2). The results revealed that the transactivation activity of MT1E-2×ARE1-luc was mediated by Nrf1 rather than Nrf2, but no changes in transcriptional expression of MT1E-2×ARE2-luc were examined ( Figure 6D). However, a significant amplified activity of MT1E-6×ARE2-luc was mediated by Nrf2 rather than Nrf1, even although the MT1E-6×ARE1-luc was still modestly induced by Nrf1, but not Nrf2 ( Figure 6E). Such differences in the gene activity of between MT1E-2×ARE2-luc and MT1E-6×ARE2-luc may be relevant to their contexts in the reporter gene constructs, but the detailed mechanism remains elusive.  copies of those two ARE-sequences from the MT1E gene were allowed for driving relevant luciferase reporter genes. Each of the ARE-driven reporters-together with expression constructs for Nrf1, Nrf2, or an empty plasmid-were co-transfected into WT cells for the indicated times. Then the cells were allowed for recovery from transfection, before the luciferase activity was measured, as shown graphically as fold changes (mean ± SD, n = 3 × 3) relative to the background controls (* p < 0.05 and ** p < 0.01 indicate significant differences with controls). The data were representative of at least three independent experiments, each of which was performed in triplicates. (F) A proposed model is provided to give a better explanation of distinctive yet cooperative roles of Nrf1 and Nrf2 in synergistically regulating antioxidant cytoprotective genes against the pro-oxidative stressor tBHQ. Of note, the ER-associated Nrf1 was subject to selective topovectorial processing of this protein to yield an active N-terminally-truncated CNC-bZIP factor and then translocate the nucleus before transcriptionally regulating its target genes. Importantly, Nrf1 can act as a brake control for Nrf2's functionality to be confined within a certain extent, albeit its transcriptional expression is also positively regulated by Nrf2. This is based on the fact that specific loss of Nrf1α −/− enables Nrf2 to be accumulated as a hyper-active factor. In addition, tBHQ can also trigger a certain yield of ROS in different cell lines, before its stimulation of antioxidant cytoprotective responses mediated by cooperation of Nrf1, Nrf2, and their target genes within a complex hierarchical regulatory network.

Discussion
Since oxidative stress was initially formulated by Helmut Sies in 1985 and later redefined by Dean P. Jones in 2006 [46][47][48], an overwhelming number of publications by this conceptual term had been collected within at least 331,795 entries of the PubMed (https://pubmed.ncbi.nlm.nih.gov, accessed on 5 June 2021). Such a perennially vital topic as oxidative stress (and redox signaling) is open to arouse great concerns from researchers in distinct fields, but also is one of the most persistently-existing intractable problems to be addressed for health and disease, particularly in changing environmental conditions. Amongst its merits elicited by evoking biological stress responses, a steadystate redox balance is maintained within certain threshold ranges by cell respiration, aerobic metabolism, and redox switches governing oxidative stress responses [46,47]. However, the pitfalls of oxidative stress can also lead to indiscriminate use of this term as a global concept, but without a clear relation to redox chemistry, in each of the particular cases. For the underlying molecular details, the major role in antioxidant defense is fulfilled by antioxidant enzymes, but not by small-molecule antioxidant compounds (e.g., tBHQ), in the cellular biochemical processes.
In all life forms, distinct types of cells can constantly generate a certain amount of ROS (and free radicals) during aerobic metabolism, such that its hormetic effects could be triggered in order to establish normal physiological cytoprotective mechanisms against oxidative damages. Of note, oxidative stress occurs in cells when ROS production overwhelms the natural antioxidant defenses and/or redox controls are disrupted [46,47]. If oxidative stress is over-stimulated for a long period, the resulting damages lead to various pathophysiological conditions which can result in many human chronic diseases-including cancer, diabetes, atherosclerosis, and neurodegenerative diseases [28,49]. Thereby, to combat the excessive production of ROS, all the cells have been evolutionarily armed with a series of innate powerful antioxidant defense systems. Amongst them is a set of essential antioxidant, detoxification, and cytoprotective mechanisms governed by the CNC-bZIP family of transcription factors [12,50,51]. In mammalian cells, Nrf1 and Nrf2 are two principal CNC-bZIP factors to regulate target genes by specific ARE-binding sequences in the promoter regions. To date, a large number of studies on Nrf2 had revealed it functions as a master regulator of antioxidant response and relevant redox signaling [25]. Rather, such versatile Nrf2 acts de facto as a promiscuous, but not essential, player for the optimal ARE-binding to most of its target genes [52], supporting the concluding notion that Nrf2 is dispensable for normal growth and development [26], with no any pathological phenotypes being manifested in its global knockout mice. As a matter of fact, Nrf1, rather than Nrf2, is a living fossil with its ancestral properties, because it shares a highly evolutionary conservativity with SKN-1, Cnc, and Nach factors [51]. Like its ancient homologues [53,54], Nrf1 is topologically dislocated across ER membranes and then processed to give rise to an N-terminally-truncated active factor, similar to Nrf2, before regulating its cognate target genes [55][56][57]. Thus, it is inferable that a unique conserved, indispensable role is fulfilled by Nrf1, but not by Nrf2, in maintaining the steady-state threshold of robust redox homeostasis during healthy life process.
The evidence has been provided in the present study, unraveling differential yet integral contributions of Nrf1 and Nrf2 to synergistic regulation of antioxidant cytoprotective genes at basal and tBHQ-inducible expression levels in wild-type (WT) cells. Specific knockout of Nrf1α −/− leads to severe endogenous oxidative stress as elicited by increased basal ROS levels; this is accompanied by increased ratios of GSSG to GSH and apoptosis. In Nrf1α −/− cells, Nrf2 was highly accumulated, but also cannot fully compensate loss of Nrf1α's function in its basal cytoprotective response against endogenous oxidative stress, even though it had exerted partially inducible antioxidant response as the hormetic effect of tBHQ against apoptotic damages. By striking contrast, Nrf2 −/−∆TA cells were also manifested by obvious oxidative stress, partially resulting from a substantial reduction of Nrf1 in basal and tBHQ-stimulated expression levels. However, the inactive Nrf2 −/−∆TA cells can be still triggered to mediate a potent antioxidant response to tBHQ, as deciphered by a significantly decreased ration of GSSG to GSH. Conversely, a remarkable increase of the Nrf1 expression was obtained from the constitutive active caNrf2 ∆N cells, in which neither oxidative stress nor apoptotic damages had occurred, no matter if it was intervened with tBHQ. Thereby, distinct yet joint functions of Nrf1 and Nrf2 may be executed through their inter-regulatory effects on cognate genes against oxidative stress ( Figure 6F).
Differences in ROS-scavenging activities of SOD and CAT were determined in distinct genotypic cell lines. Basal activities of SOD and CAT were significantly increased by Nrf1α −/− , Nrf2 −/−∆TA , or caNrf2 ∆N , when compared to those of WT cells. tBHQ-inducible SOD activity were marginally elevated in Nrf1α −/− , Nrf2 −/−∆TA , but not caNrf2 ∆N or WT, cell lines, as accompanied by an exceptional decrease of its activity by 16 h of this stimulation. The modest changes suggest that SOD activity may be monitored by other factors beyond Nrf1 and Nrf2. This notion is also supported by the previous RT-qPCR data [45]. Further evidence also revealed that tBHQ-stimulated CAT activity was markedly augmented in Nrf1α −/− cells (with hyper-active Nrf2 accumulation), and thereby completely abolished in Nrf2 −/−∆TA cells. However, basal increased CAT activity was substantially reduced by tBHQ intervention of caNrf2 ∆N cells (also with enhanced expression of Nrf1), although this constitutive activator per se was unaffected by this chemical. These imply that CAT activity is regulated positively by Nrf2, and also monitored negatively by Nrf1, particularly during tBHQ-stimulated conditions. As such, it cannot also be ruled out that loss of the N-terminal Keap1-binding Neh2 domain from Nrf2 to yield caNrf2 ∆N may cause a negative effect on the tBHQ-stimulated CAT activity.
As a widely used Nrf2-activator, tBHQ can also trigger a certain activating effect on the expression of Nrf1 in WT cells, as well in caNrf2 ∆N cells, but this effect is almost totally abolished by Nrf1α −/− or Nrf2 −/−∆TA . Conversely, induction of Nrf2 expression by tBHQ only occurred in WT cells alone, but not in Nrf1α −/− , Nrf2 −/−∆TA , or caNrf2 ∆N cell lines, although its basal expression levels are significantly augmented in either Nrf1α −/− or caNrf2 ∆N cell lines. Together, with our previous data [31,40], these indicate that Nrf1 has an ability to confine Nrf2 within a certain extent, albeit its transcriptional expression is positively regulated by Nrf2, which functions as a limited chameleon activator. Such interregulatory effects of both Nrf1 and Nrf2 on antioxidant, detoxification and cytoprotective genes, such as HO-1, NQO1, GCLC, GCLM, GSR, GPX1, TALDO, MT1E and MT2, were further determined in distinct genotypic cell lines. As anticipated, the comprehensive experimental evidence has been provided herein, unraveling that HO-1, NQO1, GCLC, GCLM, GSR, and TALDO were induced by tBHQ stimulation of WT cells, but GPX1, MT1E, and MT2 were not stimulated or even slightly suppressed by this chemical. By contrast, Nrf1α −/− cells were still allowed for tBHQ-increased expression of HO-1, NQO1, GCLM, and TALDO, but with an exceptional decrease of GPX1, even though hyper-expressed Nrf2 was roughly unaffected by tBHQ. This implies an additional involvement of other transcriptional factors beyond Nrf1 and Nrf2 in mediating these gene response to tBHQ as a pro-oxidative stressor. More intriguingly, both basal and tBHQ-stimulated expression levels of MT1E were strikingly augmented in Nrf2 −/−∆TA , but not Nrf1α −/− or caNrf2 ∆N , cell lines, whereas MT2 was marginally induced by tBHQ in caNrf2 ∆N cells. This finding implies that MT1E, but not MT2, may serve as an Nrf1-specific target gene, as further evidenced by its relevant reporter assays. Moreover, these was also accompanied by a modest inducible enhancement of GLCM and Nrf1, whereas all other examined genes were, to lesser or no extents, stimulated by tBHQ in either Nrf2 −/−∆TA or caNrf2 ∆N cell lines. These indicated that most of all other examined genes except MT1E are regulated primarily by Nrf2, but induction of its transactivation activity by tBHQ is also limited by its constitutive loss of the Keap1-binding Neh2 domain in the mutant caNrf2 ∆N factor.

Concluding Remarks
Dramatic research advances of the past 25 years, since a fascinating discovery by Itoh et al. [23], have witnessed an overwhelmingly preferential option for a sole Nrf2 focus in all relevant fields, whereas Nrf1 was almost totally ignored by nearly all others except for a handful of groups. Such disproportionately biased consequence has resulted in a general misunderstanding of Nrf2 as an only master hub of predominantly regulating antioxidant, detoxification, and cytoprotective genes; regardless of the exciting fact that Nrf1, rather than Nrf2, is highly conserved with those more ancient SKN-1, Cnc, and Nach factors [51], and that it can also fulfill unique indispensable roles for cell homeostasis and organ integrity during the life process. In the present study, together with our previous publications [31,40], Nrf1 and Nrf2 are experimentally evidenced to elicit differential yet integral roles in mediating antioxidant cytoprotective responsive genes against pro-oxidative stress induced by tBHQ. The inter-regulatory effects of Nrf1 and Nrf2 on differential expression levels of antioxidant cytoprotective genes-e.g., HO-1, NQO1, GCLC, GCLM, GSR, GPX1, TALDO, MT1E, and MT2-as well on the ROS-scavenging activities of SOD and CAT, were determined in depth. The collective results demonstrate that both Nrf1 and Nrf2 can make distinctive yet cooperative contributions to finely tuning basal and/or tBHQ-stimulated expression of target genes within their inter-regulatory networks. Overall, Nrf1 can be allowed to act as a brake control for confining Nrf2 s functionality within a certain extent, albeit its transcriptional expression is positively regulated by Nrf2.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/antiox10101610/s1, Figure S1: Time-dependent changes in the mRNA expression of distinctive responsive genes to tBHQ; and Figure S2: Different time-dependent effects of tBHQ on ROS level in distinct cell lines.
Author Contributions: Both R.W. and Z.F. performed the experiments with help of K.L., collected all the relevant data, and wrote a draft of this manuscript with most figures and supplemental information. Y.Z. designed and supervised this study, analyzed all the data, helped to prepare all figures with cartoons, and wrote and revised the paper. All authors have read and agreed to the published version of the manuscript.
Funding: This study was funded by the National Natural Science Foundation of China (NSFC, with a key program 91429305 and other two projects 81872336 and 82073079) awarded to Yiguo Zhang (at Chongqing University, Chongqing, China).

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable. Data Availability Statement: All data needed to evaluate the conclusions in the paper are present in this publication along with the supplementary information documents that can be found online. The additional data related to this paper may also be requested from the corresponding author on reasonable request.