The transcription factor Nrf2 (nuclear factor-erythroid 2-related factor 2) co-ordinately regulates ARE (antioxidant-response element)-mediated induction of cytoprotective genes in response to electrophiles and oxidative stress; however, the molecular mechanism controlling Nrf2-dependent gene expression is not fully understood. To identify factors that regulate Nrf2-dependent transcription, we searched for proteins that interact with the Nrf2-NT (N-terminal Nrf2 transactivation domain) by affinity purification from HeLa nuclear extracts. In the present study, we identified KAP1 [KRAB (Krüppel-associated box)-associated protein 1] as a novel Nrf2-NT-interacting protein. Pull-down analysis confirmed the interaction between KAP1 and Nrf2 in cultured cells and demonstrated that the N-terminal region of KAP1 binds to Nrf2-NT in vitro. Reporter assays showed that KAP1 facilitates Nrf2 transactivation activity in a dose-dependent manner. Furthermore, the induction of the Nrf2-dependent expression of HO-1 (haem oxygenase-1) and NQO1 [NAD(P)H quinone oxidoreductase 1] by DEM (diethyl maleate) was attenuated by KAP1 knockdown in NIH 3T3 fibroblasts. This finding established that KAP1 acts as a positive regulator of Nrf2. Although Nrf2 nuclear accumulation was unaffected by KAP1 knockdown, the ability of Nrf2 to bind to the regulatory region of HO-1 and NQO1 was reduced. Moreover, KAP1 knockdown enhanced the sensitivity of NIH 3T3 cells to tert-butylhydroquinone, H2O2 and diamide. These results support our contention that KAP1 participates in the oxidative stress response by maximizing Nrf2-dependent transcription.
- haem oxygenase-1 (HO-1)
- Krüppel-associated box-associated protein 1 (KAP1)
- NAD(P)H quinone oxidoreductase 1 (NQO1)
- nuclear factor-erythroid 2-related factor 2 (Nrf2)
- oxidative stress
- transcription regulation
The transcription factor Nrf2 (nuclear factor-erythroid 2-related factor 2) is activated by electrophiles and oxidative stress. Consequently, it co-ordinately regulates a battery of genes that encode cytoprotective proteins including NQO1 [NAD(P)H quinone oxidoreductase 1], HO-1 (haem oxygenase-1) and glutamate-cysteine ligase subunits via the ARE (antioxidant-response element) . The Nrf2 protein possesses six evolutionarily conserved domains named Neh (Nrf2-ECH homology) 1–6 . Keap1 (Kelch-like ECH-associated protein 1) binds to the N-terminal Neh2 domain and negatively regulates Nrf2 activity . Under homoeostatic conditions, Keap1 acts as an ubiquitin E3 ligase adaptor molecule for the Cul3 ubiquitin ligase complex, and it induces Nrf2 degradation through the ubiquitin–proteasome pathway . Electrophiles or ROS (reactive oxygen species) oxidize several reactive cysteine residues on Keap1  and inactivate the Keap1-mediated degradation of Nrf2 by mechanisms that are still not completely understood. Subsequently, stabilized Nrf2 translocates into the nucleus, heterodimerizes with small Maf proteins to bind to AREs and activates ARE-mediated gene expression . Numerous studies using Nrf2-knockout mice have demonstrated the biological significance of Nrf2 in the oxidative stress response and in the pathogenesis of various environmental- and aging-associated diseases, such as acute oxidative injury , chemically induced carcinogenesis , acute inflammation  and autoimmune diseases .
KAP1 [KRAB (Krüppel-associated box)-associated protein 1], also known as TIF1β (transcriptional intermediary factor 1β) or Trim28 (tripartite motif 28), belongs to the Trim protein family [10,11]. KAP1 has multiple structural motifs in its N-terminal region, including a RING (really interesting new gene) finger, two B-boxes and coiled-coil domains, which are collectively called the RBCC domain. It also has PHD (plant homeodomain) and Bromo domains in its C-terminal region . The RBCC domain of KAP1 is responsible for binding to KRAB domain-containing zinc finger proteins (KRAB proteins), which constitute a large family of transcriptional silencers (over 400 genes in the human genome) . In addition, there is a consensus HP1 (heterochromatin protein 1) binding motif (PXVXL) in the central region of KAP1 . KAP1 was originally identified by its ability to interact with two chromosomal proteins, HP1α and MOD1 (HP1β) . KAP1 functions as a scaffold protein for KRAB proteins and mediates transcriptional repression through epigenetic mechanisms such as histone tail modifications. KAP1 recruits the histone methyltransferase SETDB1 and the NuRD repressor complex to the promoter region of KRAB-regulated genes [16–18]. It also mediates the deacetylation of histones, the methylation of histone H3 lysine 9 (H3 K9), and the deposition of HP1 on target promoter regions ; however, several studies have indicated that KAP1 also acts as a co-activator for transcription caused by activation of C/EBPβ (CCAAT-enhancer-binding protein β), the GR (glucocorticoid receptor)  and the Nur77 orphan receptor .
The Nrf2 transactivation activity mainly relies on its Neh4 and Neh5 domains. We previously demonstrated that CBP [CREB (cAMP-response-element-binding protein)-binding protein] directly and co-operatively interacts with the Neh4 and Neh5 domains to activate Nrf2-dependent transcription . In addition, BRG1 (Brahma-related gene 1), a core ATPase subunit of the SWI/SNF chromatin remodelling complex, interacts with Nrf2 in a Neh5-dependent manner and selectively induces the expression of HO-1 [23,24]. Nioi et al.  showed that the Nrf2 C-terminal Neh3 domain is important for Nrf2 transactivation through an interaction with the chromo-ATPase/helicase DNA-binding protein CHD6. These findings suggest the importance of orchestration by the multiple co-activators in the regulation of Nrf2-dependent transcription; however, the underlying molecular mechanisms of Nrf2-dependent transcription are not fully understood. In the present study, we identified KAP1 as a novel cofactor for Nrf2 and examined its role in Nrf2-dependent transcription.
MATERIALS AND METHODS
DEM (diethyl maleate) and tBHQ (tert-butylhydroquinone) were purchased from Wako. Diamide was purchased from Sigma–Aldrich.
HEK-293 cell (human embryonic kidney-293 cell) lines were maintained in DMEM (Dulbecco's modified Eagle's medium; Sigma–Aldrich) containing 10% (v/v) FBS (fetal bovine serum) and 100 units/ml penicillin–streptomycin (Invitrogen). HeLa cells were cultured in RPMI 1640 medium (Sigma–Aldrich) containing 10% (v/v) FBS and 100 units/ml penicillin–streptomycin. The Nrf2-KO-MEFs (mouse embryonic fibroblasts derived from Nrf2-knockout mice) were cultured in Iscove's modified Dulbecco's medium (Sigma–Aldrich) containing 10% (v/v) FBS and 100 units/ml penicillin–streptomycin. All cells were cultured at 37 °C with 5% CO2 and saturated humidity.
To construct the mouse KAP1 expression plasmid, the KAP1 cDNA (amino acids 1–825) was amplified as three DNA fragments by PCR using NIH 3T3 cDNA as a template with the following primers: fragment 1: 5′-GGAAGCTTATGGCGGCCTCGGCGGCA-3′ and 5′-TAGACAGGGCAGCAGCCG-3′; fragment 2: 5′-GCCTGCGGCCCGAGCGGG-3′ and 5′-CTGAAACTTCATCTCACC-3′; and fragment 3: 5′-TATTTCCAGCTGCATCGG-3′ and 5′-CTGAAACTTCATCTCACC-3′. These DNA fragments were subcloned into the pGEM-T Easy Vector (Promega). cDNA fragments were digested with HindIII–BamHI, BamHI and BamHI–XbaI. They were subsequently ligated into a single cDNA fragment and subcloned into the pcDNA3 vector (Invitrogen), thereby creating pcDNA3-mKAP1. The human KAP1 expression plasmid, pFN21A-human full-length KAP1 (Flexi HaloTag clone, FHC06144), was purchased from the Kazusa DNA Research Institute. To make mammalian expression plasmids with a HaloTag fused to the hKAP1-NT (N-terminal half of human KAP1) or the hKAP1-CT (C-terminal half of human KAP1), DNA fragments were amplified by PCR using pFN21A-human full-length KAP1 as a template with the following primers: hKAP1-NT: 5′-AGCGATAACGAGATCGCCATG-3′ and 5′-GAATTCGTTTAAACCACAATCATCTGGAGGGCCCG-3′; and hKAP1-CT: 5′-TAACGCGATCGCCGATCCCGTGGAGCCCATGGC-3′ and 5′-CGAGCCCGAATTCGTTTAAACA-3′. These DNA fragments were inserted into the SgfI and PmeI sites of pFN21A (Promega). These vectors were referred to as pFN21A-hKAP1-NT and pFN21A-hKAP1-CT.
To generate bacterial expression plasmids for GST–Nrf2-NT [GST (glutathione transferase) fused to the N-terminal domain of Nrf2;, amino acid residues 1–316], GST–Neh2–4 (GST fused to the Neh2–4 domains; amino acid residues 1–171), GST–Neh5 (GST fused to the Neh5 domain; amino acid residues 172–201), GST–Neh4-NTend (GST fused to the Neh4-NTend domain; amino acid residues 116–316), GST–Neh5-NTend (GST fused to the Neh5-NTend domain; amino acid residues 172–316) and GST–Neh4–5 (GST fused to the Neh4–5 domain; amino acid residues 116–201), DNA fragments were amplified by PCR using pcDNA3-mouse full-length Nrf2 plasmid as a template with the following primers: mNrf2-NT: 5′-CGGGTCGACAAATGATGGACTTGGAGTTGCCA-3′ and 5′-CCGCTCGAGACACAGTGACAGGTCACA-3′; mNrf2 Neh2–4: 5′-CGGGTCGACAAATGATGGACTTGGAGTTGCCA-3′ and 5′-CCGCTCGAGCTGCTTAAATCAGTCATGGCT-3′; mNrf2 Neh5: 5′-CGGGATCCAAATAGAGCAGGACATGGAGCAAGT-3′ and 5′-CCGCTCGAGATCAGCCAGCTGCTTGTTTTCGGT-3′; mNrf2 Neh4-NTend: 5′-CGGGATCCAATTTGAAGACTGTATGCAGCTT-3′ and 5′-CCGCTCGAGACACAGTGACAGGTCACA-3′; mNrf2 Neh5-NTend: 5′-CGGGATCCAAATAGAGCAGGACATGGAGCAAGT-3′ and 5′-CCGCTCGAGACACAGTGACAGGTCACA-3′; and mNrf2 Neh4–5: 5′-CGGGATCCAATTTGAAGACTGTATGCAGCTT-3′ and 5′-CCGCTCGAGATCAGCCAGCTGCTTGTTTTCGGT-3′. DNA fragments were inserted into the SalII and XhoI sites (Nrf2-NT and Neh2–4) or the BamHI and XhoI sites (Neh5, Neh4-NTend, Neh5-NTend and Neh4–5) of pGEX-5X-1 (GE Healthcare). These vectors will be referred to as pGEX-mNrf2-NT, pGEX-mNrf2 Neh2–4, pGEX-mNrf2 Neh5, pGEX-mNrf2 Neh4-NTend, pGEX-mNrf2 Neh5-NTend and pGEX-mNrf2 Neh4–5 respectively.
The pcDNA3-GBD (Gal4 DNA-binding domain)-mNrf2-NT vector was constructed by inserting PCR fragments of GBD–mNrf2-NT into the HindIII and XbaI sites of pcDNA3. To construct a mouse Nrf2 NT deletion expression plasmid, the mouse Nrf2 with the NT region deleted was amplified by PCR using mouse Nrf2 cDNA as a template with the following primers: 5′-GGGGTACCCCTCAGCATGATGGACAAAGCTTTCAACCCGAAGCAC-3′ and 5′-CCCAATTGACTAGTTTTTCTTTGTATCTGGCT-3′. The amplified DNA fragment was digested with KpnI and SpeI, cloned into the KpnI–XbaI sites of pcDNA3, and designated pcDNA3-mNrf2 ΔNT.
To obtain a human NQO1–luciferase DNA fragment, the human NQO1 promoter region (1040 bp) was amplified by PCR using human aortic endothelial cell genomic DNA as a template with the following primers: 5′-GGGGCTAGCTCCGGGTTCAAGCGATTC-3′ and 5′-CCCCCGGGGCTCTGGTGCAGTCCGG-3′. This DNA fragment was digested by NheI–SmaI and ligated into the NheI–SmaI sites of the pGL3-Basic vector (Promega). To generate a pCEP4-human NQO1 luciferase plasmid, the human NQO1–luciferase DNA fragment was cut out of the pGL3-Basic human NQO1 vector via NheI–SalI digestion, blunt-ended by filling in reaction with T4 DNA polymerase, and then ligated into the SalI-blunt-ended site of pCEP4.
Purification of Nrf2-interacting proteins
The preparation of nuclear extracts from HeLa cells has been described previously . For the purification of the Nrf2-NT-interaction protein, an incubation mixture (total volume 1.5 ml) was made that was incubated at 4 °C for overnight, and then applied to a column. This mixture contained 5 mg of HeLa cell nuclear extracts, 20 μg of GBD or GBD-NT recombinant protein and 500 μl of anti-FLAG antibody-conjugated agarose (Sigma–Aldrich) in IP buffer [20 mM Hepes (pH 7.9), 14% glycerol, 280 mM NaCl, 0.17 mM EDTA and 1 × protease inhibitor cocktail (Roche)]. After multiple washes with 5 ml of FLAG-wash buffer [20 mM Hepes (pH 7.9), 10% glycerol, 150 mM NaCl, 0.2 mM EDTA and 0.1% Nonidet P40], the bound proteins were eluted using 400 μl of FLAG-elution buffer [20 mM Hepes (pH 7.9), 10% glycerol, 100 mM NaCl, 0.2 mM EDTA, 1 × protease inhibitor cocktail and 200 μg/ml FLAG peptide (Sigma–Aldrich)]. The eluted fraction was concentrated to 100 μl using Amicon Ultra centrifugal filter units, and then the collected protein sample was subjected to SDS/PAGE. The SDS/PAGE was stained with SYPRO Ruby. Subsequently, we arbitrarily cut out five bands from the SDS/PAGE and in-gel digested with trypsin. The eluted peptides were then analysed with nano-ESI-MS/MS (nano-electronspray ionization tandem MS).
HaloLink resin pull-down assay
HEK-293 cells were transfected with either pcDNA3-full-length mouse Nrf2 or pcDNA3-mouse Nrf2 ΔNT and pFN21A-human KAP1 full-length plasmids using FuGENE HD (Roche). After a 36 h incubation, the whole-cell extracts of the transfected HEK-293 cells were prepared using lysis buffer [50 mM Tris/HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, 1 × protease inhibitor cocktail and 10 μM MG132]. To detect an interaction between Nrf2 and KAP1, the whole-cell extracts were incubated with HaloLink Resin (Promega) in lysis buffer overnight at 4 °C. After extensive washing with lysis buffer, the bound protein fractions were digested with ProTEV protease (Promega). The digested proteins were subjected to SDS/PAGE followed by immunoblot analysis using anti-HaloTag (Promega), anti-KAP1 (Abcam, ab3831) or anti-Nrf2 antibodies (Santa Cruz Biotechnology, sc-722).
In vitro transcription/translation, recombinant protein preparation and pull-down assays
Recombinant HaloTag-human KAP1 proteins were synthesized using the TNT® T7 Quick-Coupled Transcription/Translation System (Promega) with pFN21A-human KAP1, pFN21A-hKAP1-NT and pFN21A-hKAP1-CT as templates. GST-tagged proteins were expressed in Escherichia coli strain BL21 and purified with glutathione–Sepharose 4B beads (GE Healthcare). GST-tagged recombinant proteins (20 μg) were immobilized to glutathione–Sepharose 4B beads and mixed with in-vitro-translated proteins in pull-down buffer [50 mM Tris/HCl (pH 7.5), 10% glycerol, 100 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, 1 × protease inhibitor cocktail and 10 μM MG132; 300 μl total volume] for 4 h at 4 °C. After washing three times with pull-down buffer (0.5 ml), the precipitated proteins were stained with HaloTag Ligand TMR (Promega) for 30 min at room temperature (20–25 °C). The prepared proteins were resolved on SDS/PAGE, and HaloTag proteins and GST–mouse Nrf2 proteins were detected with the Typhoon FLA9000 (GE Healthcare) imaging and CBB (Coomassie Brilliant Blue) staining respectively.
HeLa cells were transfected with pcDNA3 or pcDNA3-full-length mouse Nrf2 plasmids using FuGENE HD (Roche) and cultured on glass coverslips. After washing with PBS, the cells were fixed with 4% (w/v) paraformaldehyde/PBS, permeabilized with 0.25% Triton X-100/PBS and blocked with 2% BSA/PBS. The cells were then incubated with 2% BSA/PBS containing 1 μg/ml anti-Nrf2 antibody (Santa Cruz Biotechnology, sc-13032) and 1 μg/ml anti-KAP1 antibody (Abcam, ab3831) overnight at 4 °C. To visualize the Nrf2 and KAP1 signals, the cells were stained with anti-rabbit IgG-Biotin (Jackson ImmunoResearch Laboratories) plus Streptavidin-Alexa Fluor® 488 and anti-goat-IgG Alexa Fluor® 594 (Invitrogen) respectively. The fluorescence images were observed with a C1si confocal laser-scanning microscope system (Nikon).
For Gal4-based reporter assays, the Nrf2-KO-MEFs were transfected with the reporter plasmids pCEP4-Gal4-UAS Luc  and pRLTK (Promega), and the effector plasmids pcDNA3-GBD-mNrf2-NT and pcDNA3-mKAP1 using Lipofectamine™ and Plus Reagent (Invitrogen). For human NQO1 reporter assays, Nrf2-KO-MEFs were transfected with the pCEP4-based human NQO1 reporter plasmid, pRLTK, pcDNA3-full-length mouse Nrf2 or pcDNA3-mouse Nrf2 ΔNT and pcDNA3-mKAP1. At 4 h after transfection, the medium was replaced with fresh medium containing 10% FBS, and the cells were incubated for 36 h at 37 °C with 5% CO2. Dual luciferase assays were performed according to the manufacturer's recommended protocol (Promega).
siRNA (small interfering RNA) knockdown analysis
An siRNA against mouse KAP1 was synthesized by Qiagen with the following sequence: 5′-CCAAAGACAUCGUGGAGAATT-3′ . This siRNA was transfected using the Lipofectamine™ 2000 reagent according to the manufacturer's recommended protocol (Invitrogen). After 24 h of transfection, the cells were treated with DEM for the time-course analyses.
Q-RT–PCR (quantitative reverse-transcription–PCR)
cDNAs were synthesized using PrimeScript® RTase (TaKaRa Bio) with total RNA as a template. Q-RT–PCR analyses were performed using SYBR® Premix Ex Taq™ II (Perfect Real Time) (TaKaRa Bio) with a CFX Real-Time PCR Detection System (Bio-Rad). The cyclophilin A gene was used as an internal control with the following primers: 5′-AAGACTGAATGGCTGGATGG-3′ and 5′-AGCTGTCCACAGTCGGAAAT-3′. The DEM-induced gene expression was measured by Q-RT–PCR with the following primers: HO-1, 5′-GTGATGGAGCGTCCACAGC-3′ and 5′-TTGGTGGCCTCCTTCAAGG-3′; NQO1, 5′-TTTAGGGTCGTCTTGGCAAC-3′ and 5′-AGTACAATCAGGGCTCTTCTCG-3′.
ChIP (chromatin immunoprecipitation) assay
The ChIP assays were performed as described previously . The chromatin immunoprecipitaed DNA was quantified by real-time PCR using the following primers: HO-1 E1 enhancer: 5′AGAGGGAACAGAGGGTGACTC-3′ and 5′-TGCTTTTATGCTGTGTCATGGT-3′, HO-1 E2 enhancer: 5′-GGGCTAGCATGCGAAGTGAG-3′ and 5′-AGACTCCGCCCTAAGGGTTC-3′, HO-1 exon 3: 5′-TTCTGTGCAATCTTCTTCAGGA-3′ and 5′-CACTCACCCTGAGCTGCTG-3′, and NQO1-promoter, 5′TCTTCCCAAGATGCCTCTGG-3′ and 5′-GGCTGGCTACAGGCTAGGCTA-3′.
Cell viability analyses
Control or KAP1 siRNAs were transfected into the NIH 3T3 cells as described above. After 24 h of transfection, the cells were plated into 96-well plates at a density of 2×103 cells/well and incubated for 12 h. The culture medium was replaced with DMEM containing 100 μM DEM, and the plate was incubated for another 12 h. Then, the medium was replaced with DMEM containing various stress reagents, and the plate was incubated for a further 6 h. The cell viability was measured using a Cell Counting Kit-8 (Dojin) according to the manufacturer's recommended protocol.
The results were expressed as means±S.E.M. Statistical significance was determined by a Student's t test for two-parameter comparisons or a one-way ANOVA followed by a Dunnett's post-hoc test for multiple parameter comparisons with the control.
Identification of a novel Nrf2-interacting factor, KAP1
The Nrf2-NT region, which includes Neh2, Neh4 and Neh5, functions as a potent transactivation domain when fused to a GBD [2,22]. To identify cofactors for Nrf2 transactivation, we performed an affinity-purification experiment. As bait, we used recombinant Nrf2-NT that was N-terminally linked to FLAG–GBD and C-terminally linked to a His6 tag (FLAG–GBD–Nrf2-NT–His) (Figures 1A and 1B). The FLAG–GBD–His that lacked Nrf2-NT was used as a negative control. HeLa cell nuclear extracts were incubated with the bait, and Nrf2-NT-interacting proteins were immunoprecipitated with anti-FLAG antibody-conjugated beads. Several proteins that co-precipitated with the bait, but not with the control protein, were detected on SDS/PAGE (Figure 1C). By mass spectrometric analysis, we identified proteins including BAF170, TIF1γ and KAP1/TIF1β (hereafter referred as KAP1) (Figure 1C) using the two criteria. More than two peptide signals had to be detected and the protein sizes of those identified must be appropriately compared with those of the previously reported ones. The peptide signals detected are summarized in Supplementary Table S1 (available at http://www.BiochemJ.org/bj/436/bj4360387add.htm). Two TRIM family proteins, TIF1γ and KAP1, are involved in the transcriptional regulation, and both proteins possess an RBCC domain that is responsible for the interactions with multiple proteins. Furthermore, our preliminary analysis demonstrated that KAP1 activated, but TIF1γ repressed Nrf2 transactivation activity in a transient transfection assay (results not shown). Therefore we focused on KAP1 in the present study and analysed the role of KAP1 in an Nrf2-dependent transcriptional regulation. To confirm the interaction between Nrf2 and KAP1, we performed a pull-down assay using HEK-293 cells that transiently expressed full-length Nrf2 or the N-terminal deletion mutant (ΔNT) with HaloTag–human KAP1. The pull-down assay demonstrated that full-length Nrf2, but not ΔNT, binds to KAP1 (Figure 1D).
KAP1 directly interacts with Nrf2 and localizes to the nucleus
To examine which Nrf2 domain binds to KAP1, we performed an in vitro pull-down assay using a series of recombinant proteins with GST fused to Nrf2-deletion constructs that included Nrf2-NT, Neh2–4 and Neh5. These proteins were incubated with in-vitro-translated HaloTag–KAP1 protein (Figure 2A). The results from the pull-down assays indicated that KAP1 binds to Nrf2-NT, but not to Neh2–4 and Neh5 (Figure 2B). To further dissect which Nrf2-NT regions are responsible for the interaction with KAP1, we generated recombinant proteins including GST–Neh4-NTend, GST–Neh5-NTend and GST–Neh4–5. We detected an interaction between Neh4-NTend and KAP1, although it was a weaker interaction than that of GST–Nrf2-NT. In addition, we detected faint, but significant, binding of GST–Neh5-NTend and GST–Neh4–5, but not of GST–NTend (which contains only amino acids 202–316 of Nrf2) with KAP1 (Figure 2C and results not shown). To further investigate which KAP1 domains are responsible for the interaction with Nrf2-NT, we performed a pull-down assay using GST–Nrf2-NT, and the KAP1-NT or KAP1-CT proteins translated in vitro. The results showed that KAP1-NT, but not KAP1-CT, binds to Nrf2-NT (Figure 2D). This finding demonstrates that Nrf2-NT directly binds to KAP1-NT, which contains an RBCC domain.
To investigate whether Nrf2 co-localizes with KAP1 in cells, we performed immunocytochemical analyses in HeLa cells using specific antibodies against Nrf2 and KAP1. Because endogenous Nrf2 protein undergoes rapid turnover by proteasomal degradation even under stressful conditions , it was barely detectable with our immunostaining conditions. Therefore we transiently overexpressed Nrf2 to visualize it more clearly. Images obtained by confocal microscopy revealed that Nrf2 was uniformly distributed in the nucleus of Nrf2-transfected cells, but not in control-transfected cells (Figure 3, top and middle panels respectively). KAP1 was also evenly distributed in the nucleoplasm and excluded from the nucleoli. A closer inspection demonstrated that substantial portions of KAP1 and Nrf2 exist in the small foci-like structures in the nucleus, and some of these structures are double-positive for KAP1 and Nrf2 (Figure 3, bottom panels). Thus these results support our contention that Nrf2 interacts with KAP1 in the nucleus.
KAP1 activates Nrf2-dependent transcription
To investigate the role of KAP1 in Nrf2-dependent transcription, we performed a reporter assay in embryonic fibroblasts derived from Nrf2-KO-mice. For this assay, we used a reporter plasmid that has multiple GBD-binding sites upstream of a luciferase gene . Our results showed that KAP1 up-regulates the transactivation activity of GBD–Nrf2-NT in a dose-dependent manner (Figure 4A). In the absence of GBD-Nrf2-NT, KAP1 did not significantly affect reporter activity (results not shown). Next, to examine the effects of KAP1 on the native Nrf2-responsive promoter, we utilized a luciferase reporter plasmid linked to the human NQO1 promoter. Nrf2 expression in Nrf2-KO-MEFs increased human NQO1 reporter activity, and KAP1 facilitated Nrf2-dependent transcription in a dose-dependent manner (Figure 4B). The maximum dose of KAP1 significantly activated Nrf2 transactivation activity 2.1-fold over the cells without KAP1 transfection. In contrast, KAP1 only slightly activated NQO1-reporter activity in the Nrf2 ΔNT-transfected cells, although the effect was statistically significant. The maximal dose of KAP1 did not affect the reporter activity significantly. These results indicate that KAP1 functions as a positive regulator of Nrf2-dependent transcription.
To analyse the effects of KAP1 on endogenous Nrf2-dependent gene expression, we knocked down endogenous KAP1 using mouse KAP1-specific siRNA in NIH 3T3 fibroblasts. Knockdown of KAP1 protein by siRNA was confirmed by immunoblot analysis using an anti-KAP1 antibody (Figure 5A). Moreover, the KAP1 knockdown did not affect the nuclear accumulation of Nrf2 in response to DEM. We then investigated the gene expression of two representative Nrf2 target genes, HO-1 and NQO1, in response to DEM by Q-RT–PCR. In KAP1 knockdown cells (siKAP1), inducible expression was significantly suppressed at the 6- and 12-h time points for HO-1 and at the 24-h time point for NQO1, as compared with siCont (control siRNA)-transfected cells (Figure 5B). To further investigate the extent to which KAP1 knockdown affects Nrf2-mediated gene expression, we performed immunoblot analyses of HO-1 and NQO1 (Figure 5C). Densitometric analysis demonstrated that HO-1 protein expression is significantly reduced at the 6- and 12-h time points and NQO1 protein expression is reduced at the 24-h time point in KAP1-knockdown fibroblasts compared with the control cells (Figure 5D). These results indicate that KAP1 activates the expression of the endogenous Nrf2 targets HO-1 and NQO1.
KAP1 participates in Nrf2 binding to the HO-1 E2 enhancer and the NQO1 promoter
To investigate the mechanism by which KAP1 enhances Nrf2-dependent gene expression, we examined the ability of Nrf2 to bind to AREs in KAP1-knockdown cells using ChIP. Because the recruitment of Nrf2 to AREs normally precedes gene induction, we performed the ChIP assays at 4- and 8-h time points after DEM administration. Primers were generated against the NQO1 promoter region that contains a single functional ARE  and against the E1 and E2 enhancer regions of the HO-1 gene, each containing multiple Nrf2-binding sites located approx. −4 and −10 kb upstream of the transcription start site respectively . Nrf2 binding to the three aforementioned regions was induced by DEM, and the maximum binding peaks were observed 4 h after DEM administration in siCont-transfected cells for each of the three regions (Figures 6A–6C). The DEM-induced Nrf2 binding to the NQO1 promoter was significantly decreased at the 4-h time point in the KAP1-knockdown cells compared with the siCont-transfected cells (Figure 6A). Nrf2 binding to the HO-1 E2 region was also significantly reduced in KAP1-knockdown cells at the 4-h time point (Figure 6B). In contrast, Nrf2 binding to the HO-1 E1 region was not significantly affected by KAP1 knockdown (Figure 6C). As a negative control, we examined Nrf2 binding to the exon 3 region of HO-1 and found no enrichment in this region (results not shown). These results indicate that KAP1 influences Nrf2 binding to AREs in a binding-site-specific manner.
KAP1 knockdown increases the susceptibility to oxidative stress
Nrf2 exerts a cytoprotective effect on cell damage induced by oxidative or chemical stress [6–9]. To explore the physiological significance of KAP1 on the Nrf2-dependent cytoprotective mechanism, we investigated whether NIH 3T3 cells were susceptible to oxidative stress when KAP1 was knocked down. The NIH 3T3 cells were transfected with either siKAP1 or siCont, and the Nrf2 pathway was activated for 12 h with 100 μM DEM. Subsequently, the cells were treated with various oxidative stress-inducing reagents. A cell viability assay demonstrated that the KAP1-knockdown cells were more sensitive to H2O2, tBHQ and diamide, but not to Cisplatin (results not shown), in comparison with the siCont cells (Figure 7). These results demonstrated that decreased expression of KAP1 increases susceptibility to oxidative stress.
We identified KAP1 as a novel Nrf2-NT-interacting protein (Figure 1C and Supplementary Table S1). Furthermore, we also identified TIF1γ among the Nrf2-interacting proteins (Figure 1C). Interestingly, both KAP1 and TIF1γ possess an RBCC domain that is responsible for interactions with multiple proteins . Our pull-down assay demonstrated that Nrf2-NT interacts with KAP1-NT, which contains an RBCC domain (Figure 2D). This finding suggests that TIF1γ also interacts with Nrf2-NT through an RBCC domain. Indeed, we have confirmed the interaction of Nrf2 and TIF1γ in a co-immunoprecipitation assay in HEK-293T cells [HEK-293 cells expressing the large T-antigen of SV40 (simian virus 40); results not shown].
Using pull-down assays and a co-immunoprecipitation assay, we demonstrated that KAP1 interacts with Nrf2 in cells (Figure 1D) and that Nrf2-NT directly interacts with KAP1-NT in vitro (Figures 2B–2D). Although we previously demonstrated that Neh4 and Neh5 act as independent transactivation domains , each Nrf2 portion alone was not sufficient for binding KAP1 (Figure 2B). Both Neh5-NTend and Neh4–5 fragments faintly, but significantly, interacted with KAP1 (Figure 2C), indicating that the Neh5 domain is essential, but the surrounding region is required for the interaction. In addition, the binding of Neh4-NTend with KAP1 was stronger than Neh5-NTend and Neh4–5, but weaker than the whole Nrf2-NT fragment. Thus we surmise that the whole structure of Nrf2-NT is important for stable binding with KAP1. These observations corroborate our previous findings that Nrf2-NT has much higher transactivation activity compared with those of Neh4 and Neh5 alone .
Immunocytochemical analyses demonstrated that KAP1 and Nrf2 exist in the small foci in the nucleus and partially co-localized (Figure 3). Interestingly, these KAP1 and Nrf2 double-positive foci were associated with weak DAPI (4′,6-diamidino-2-phenylindole) staining (Figure 3, bottom panel), which suggests that Nrf2 co-localizes with KAP1 in the euchromatic (transcriptionally active) regions in the nucleus.
Using the reporter assay in Nrf2-KO-MEFs, we showed that KAP1 enhances Nrf2-dependent transactivation in a concentration-dependent manner, which demonstrates that KAP1 acts as a positive regulator of Nrf2 (Figures 4A and 4B). Curiously, we also observed a minor but significant enhancement of NQO1 reporter activity by KAP1 in the Nrf2 ΔNT-transfected cells (Figure 4B). Furthermore, KAP1 increased NQO1 reporter activity in the absence of Nrf2 expression plasmids; however, the effect was not statistically significant. Since we performed the assay in Nrf2-KO-MEFs lacking functional Nrf2, we speculate that KAP1 activates NQO1 reporter activity by acting on transcription factors other than Nrf2.
We found that KAP1 knockdown decreased the expression of HO-1 and NQO1 at both the mRNA and protein levels in NIH 3T3 cells (Figure 5). HO-1 mRNA showed a sharp peak of induction 6 h after DEM treatment, and KAP1 knockdown significantly suppressed the peak induction. NQO1 mRNA gradually increased during the time course, and KAP1 knockdown significantly decreased NQO1 induction at the 24-h time point. Although KAP1 knockdown affected HO-1 and NQO1 expression at different time points, it significantly affected their maximal protein expression. Therefore we surmise that KAP1 itself is not necessarily required for the induction itself; instead, it acts to maximize the expression of Nrf2-target genes.
Hu et al.  recently demonstrated that KAP1 binds to the HO-1 E1 enhancer and the NQO1 promoter in mouse embryonic stem cells by ChIP sequencing analysis. This observation suggests that KAP1 is recruited to the regulatory region of HO-1 and NQO1 by trans-acting factors such as Nrf2. In the present study, we found that the binding of Nrf2 to the HO-1 E2 enhancer and the NQO1 promoter, but not to the HO-1 E1 enhancer, is significantly reduced in KAP1-knockdown cells (Figures 6A–6C). The Nrf2-binding site in the NQO1 promoter that we detected by ChIP analysis is reportedly essential for Nrf2-dependent gene induction . The HO-1 E1 and E2 enhancers possess multiple Nrf2 binding sites that are required for maximum activity of the enhancers in response to the various stressors, including heavy metals and electrophiles [30,32]. Reichard et al.  showed that arsenate-activated Nrf2 preferentially binds to the HO-1 E2 region rather than the HO-1 E1 region in HaCaT cells. Although it is not clear how KAP1 knockdown selectively affected Nrf2 binding to the HO-1 E2 region but not to the HO-1 E1 region, we speculate that the decrease in binding to the HO-1 E2 region and NQO1 promoter contributes to the decrease in Nrf2-mediated transcription in the KAP1-knockdown cells.
We observed a substantial difference between the impact of KAP1 on the expression of HO-1 and NQO1 (Figures 5B and 5C). This difference may indicate that KAP1 is somewhat selectively required for HO-1 transcription. Indeed, we previously demonstrated that both BRG1 and an actin-related motif of Neh5 are selectively required for HO-1 transcription, but not for NQO1 transcription [23,24]. However, the underlying mechanisms are not presently clear. Also, the regulatory regions of Nrf2 target genes such as HO-1 and rat GSTA2 have C/EBP-binding sites [30,34,35]. The HO-1 E1 and E2 enhancers possess multiple binding sites for Nrf2 and C/EBP, both of which are required for the maximal activity of the enhancers in response to the various electrophiles [30,34]. On the other hand, there is no report that the NQO1 regulatory region has any C/EBP-binding sites. Intriguingly, KAP1 also acts as a co-activator for C/EBPβ . Therefore KAP1 may mediate the co-operative enhancement of HO-1 transcription through both the Nrf2 and the C/EBP-binding site. These possibilities may cause the substantial difference between the impact of KAP1 on the expression of HO-1 and NQO1.
Nrf2 binding to AREs in chromatin is controlled by multiple factors. These include the interaction between small Maf proteins  and Nrf2 modifications such as acetylation . Sun et al.  recently proposed that the direct acetylation of Nrf2 by CBP/p300 promotes Nrf2 DNA binding. They demonstrated that the inhibition of Nrf2 acetylation by lysine-to-arginine mutations on the acetylation sites impairs Nrf2 binding to the regulatory regions of NQO1, GCLM and TXNRD1, but not to the human HO-1 E2 region. Thus we speculate that CBP enhances Nrf2 binding to AREs through mechanisms that are distinct from a KAP1-mediated mechanism. KAP1 is phosphorylated in response to irradiation-induced DNA double-stranded breaks and plays an important role in DNA repair via active nucleosome relaxation . Because ROS also induce the DNA damage response, KAP1 might be phosphorylated under oxidative stress conditions. We surmise that the Nrf2-recruited KAP1 may relax chromatin structure locally and in turn enhance Nrf2 binding to chromatin.
Retroviral propagation in mouse embryonic cells is suppressed by transcriptional repression of the provirus and primer-binding-site-mediated gene repression is a mechanism of such repression. Wolf et al. [39,40] has shown that KAP1 recruits the repressor complex to primer binding sites adjacent to 5′-LTRs (5′-long terminal repeats) of retroviruses, and it silences retrovirus transcription through histone H3 K9 dimethylation in mouse embryonic stem cells [39,40]. It was also reported that Nrf2 is involved in the inhibition of Tat-induced HIV-1 LTR transactivation . In addition, retroviral infections such as HIV-1 and MLV (murine leukaemia virus) cause oxidative stress and deplete cellular glutathione [42,43]. In the present study, we demonstrated for the first time that KAP1 is involved in cytoprotective mechanisms against oxidative stress (Figure 7). Therefore KAP1 may protect cells from retroviral infection via multiple mechanisms, including the Nrf2-mediated oxidative stress response. Alternatively, Jakobsson et al.  reported that the disruption of KAP1 in the adult forebrain results in increased anxiety-like behaviour and sensitivity to stress during spatial learning and memory. Because Nrf2 plays an important neuroprotective function , the down-regulation of Nrf2-mediated transcription might be involved in stress sensitivity in KAP1-knockout mice.
In summary, we identified KAP1 as a novel Nrf2-interacting protein. KAP1 directly interacts with Nrf2-NT and functions as a positive regulator of Nrf2-dependent transcription. Moreover, KAP1 knockdown in NIH 3T3 cells increased their susceptibility to oxidative stress. These results demonstrated that KAP1 plays a critical role in the oxidative stress response by promoting the Nrf2-mediated cytoprotective response. The clarification of the role of KAP1 in Nrf2-mediated transcription in vivo is an important issue that needs to be explored in the future.
Atsushi Maruyama, Keizo Nishikawa, Yukie Kawatani and Tomonori Hosoya performed the experiments and analysed the data; Atsushi Maruyama, Keizo Nishikawa, Masayuki Yamamato and Ken Itoh designed the research and co-wrote the manuscript; Junsei Mimura and Nobuhiko Harada analysed the data and co-wrote the manuscript.
This work was supported, in part, by grants from JST-ERATO and the Ministry of Education, Science, Sports and Technology, the Hirosaki University Institutional Research (to K.I.) and The Karoji Memorial Fund for Medical Research (to A.M.).
We thank the Biomedical Research Core of Tohoku University Graduate School of Medicine and the Research Laboratory for Radioisotopes at Hirosaki University for their technical support.
Abbreviations: ARE, antioxidant-response element; BRG1, Brahma-related gene 1; CBB, Coomassie Brilliant Blue; CBP, CREB (cAMP-response-element-binding protein)-binding protein; C/EBPβ, CCAAT-enhancer-binding protein β; ChIP, chromatin immunoprecipitation; DAPI, 4′,6-diamidino-2-phenylindole; DEM, diethyl maleate; DMEM, Dulbecco's modified Eagle's medium; FBS, fetal bovine serum; GBD, Gal4 DNA-binding domain; GST, glutathione transferase; GST–Neh2–4, GST fused to the Neh2–4 domains; GST–Neh4-NTend, GST fused to the Neh4-NTend domain; GST–Neh4–5, GST fused to the Neh4–5 domain; GST–Neh5-NTend, GST fused to the Neh5-NTend domain; GST–Nrf2-NT, GST fused to the N-terminal domain of Nrf2; HEK-293, cell, human embryonic kidney-293 cell; hKAP1-CT, C-terminal half of human KAP1; hKAP1-NT, N-terminal half of human KAP1; HO-1, haem oxygenase-1; HP1, heterochromatin protein 1; KRAB, Krüppel-associated box; KAP1, KRAB-associated protein 1; Keap1, Kelch-like ECH-associated protein 1; LTR, long terminal repeat; Neh, Nrf2-ECH homology; NQO1, NAD(P)H quinone oxidoreductase 1; Nrf2, nuclear factor-erythroid 2-related factor 2; Nrf2-KO-MEF, mouse embryonic fibroblast derived from Nrf2-knockout mice; Nrf2-NT, N-terminal Nrf2 transactivation domain; PHD, plant homeodomain; Q-RT–PCR, quantitative-reverse transcription–PCR; RING, really interesting new gene; ROS, reactive oxygen species; siCont, control siRNA; siRNA, small interfering RNA; tBHQ, tert-butylhydroquinone; TIF1β, transcription intermediary factor 1β; Trim28, tripartite motif 28
- © The Authors Journal compilation © 2011 Biochemical Society