DNA-dependent protein kinase catalytic subunit binds to the transactivation domain 1 of NF-κB p65

Abstract

Nuclear factor-kappa B (NF-κB) is a transcriptional factor that binds to the ∼10-base-pair κB motif on target genes and acts as an inflammatory regulator. Since dysregulation of NF-κB is thought to be related to various diseases, it would be very important to elucidate its post-translational modifications and binding partners in detail and to deeply understand mechanisms of the NF-κB dysregulation. NF-κB p65 is known to interact with the basic transcription factor TFIID subunit hTAF_II31/TAF9 through the ФXXФФ (Ф, hydrophobic amino acid; X, any amino acid) motif in a similar fashion to p53. MDM2 is known to inhibit p53 from binding to hTAFII31/TAF9 by masking p53's ФXXФФ motif. Here, as can be rationalized from this observation, we searched for novel nuclear proteins that interact with the transactivation domain 1 (TA1) of NF-κB p65 containing a ФXXФФ motif. We prepared a GST-tagged polypeptide, GST-p65_532–550, from Phe532–Ser550 of the TA1 domain and found various U937 cell nuclear proteins that bound to GST-p65_532–550. The largest bound protein the size of ∼400 kDa was subjected to mass spectrometric analysis and found to be DNA-dependent protein kinase catalytic subunit (DNA-PKcs). An immunoprecipitation experiment with an antibody against p65 and nuclear extracts from TNF-α-treated A549 cells suggested that NF-κB p65 indeed binds to DNA-PKcs in human cells. Furthermore, binding assays with a series of His-tagged DNA-PKcs fragments suggested that DNA-PKcs can bind to NF-κB p65 through the interaction of the TA1 domain with the region 541−750 in the N-HEAT domain or the region 2485−2576 in the M-HEAT domain.

Highlights

• •NF-κB p50/p65 is a transcriptional factor and acts as an inflammatory regulator.
• •NF-κB p65 has the transactivation domain 1 (TA1) containing a hydrophobic ФXXФФ motif.
• •DNA-PKcs may bind to p65 via interaction of the TA1 with either of two HEAT domains.

1. Introduction

Nuclear factor-kappa B (NF-κB) is a transcriptional factor that binds to the ∼10-base-pair κB motif on target genes and acts as an inflammatory regulator [1]. The NF-κB signaling pathway is activated by stress, tumor necrosis factor-α (TNF-α), UV light, and other stimuli [[1], [2], [3], [4], [5], [6]]. NF-κB is composed of two proteins such as p50/p65 from the five member NF-κB family and normally located in the cytosol as an inactive form [[7], [8], [9]]. Various extracellular stimulations induce NF-κB to translocate into the nucleus to activate proinflammatory cytokine-associated genes through binding to their κB motif [[10], [11], [12], [13]]. Transcription factors such as p53 and ESX are known to have regulators in the nucleus that either inhibit or activate their activity [[14], [15], [16]]. On the other hand, no regulators of NF-kB activation in the nucleus have been found. Since dysregulation of NF-κB is thought to be related to various diseases such as cancer and autoimmune diseases [1,2,5], it would be very important to elucidate its post-translational modifications and binding partners in detail and to deeply understand mechanisms of the NF-κB dysregulation.

NF-κB p65 consists of primarily two domains, a Rel homology domain, which is involved in dimerization and DNA binding, and a transactivation domain, which interacts with other transcription factors (Fig. 1A) [17,18]. The latter domain can be separated to two sub-domains, transactivation domain 1 (TA1) and transactivation domain 2 (TA2), and the TA1 the size of 31 amino acids that contains the ФXXФФ (Ф, hydrophobic amino acid; X, any amino acid) motif can form a stable α-helix structure [[17], [18], [19], [20]]. NF-κB p65 is known to interact with the basic transcription factor TFIID subunit hTAF_II31/TAF9 through the ФXXФФ motif in a similar fashion to p53 and ESX, which also contain the ФXXФФ motif (Fig. 1B) [14].

Fig. 1.

Structure of a probe from NF-κB p65 to search for its binding proteins. (A) Structure of NF-κB p65. RHD, Rel homology domain; NLS, nuclear localization signal; TAD, transactivation domain; TA1, transactivation domain 1; TA2, transactivation domain 2. (B) Transactivation domain sequences of p65, p53, and ESX with the ФXXФФ motif in red. (C) Structures of GST-tagged p65_532–550 and GST-tagged p65_{532–550, F542A + L546A}. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

NF-κB p50 becomes phosphorylated by DNA-PKcs in response to TNF-α stimulation, and promotes the expression of inflammation markers such as VCAM-1 [21]. DNA-PKcs, which comprises 4128 amino acids, plays a leading role in double-strand break repair as a ternary complex with Ku70 and Ku80 [22].

MDM2 is known to inhibit p53 from binding to hTAF_II31/TAF9 by masking p53's ФXXФФ motif [14,15]. Here, as can be rationalized from this observation, we searched for novel nuclear proteins that interact with the ФXXФФ motif of NF-κB p65 and found that at least two regions of DNA-PKcs can bind to the p65's TA1 domain.

2. Materials and methods

2.1. Cell culture

Human monocyte-like U937 cells were maintained in RPMI-1640 medium (189–02025, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) with 5% heat-inactivated fetal bovine serum (FBS) (S1810-500, Biowest, Nuaillé, France), 100 U/mL penicillin, and 100 μg/mL streptomycin. Human lung cancer A549 cells were maintained in Dulbecco's Modified Eagle's Medium (DMEM) with 10% heat-inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. Both cell lines were cultured in a 100 mm dish at 37 °C in a 5% CO₂ humidified incubator.

2.2. Expression and purification of the polypeptide p65_532–550 tagged with glutathione S-transferase (GST)

A cDNA corresponding to 532–550 amino acid sequence of Homo sapiens p65 and a cDNA corresponding to the sequence with two Ala substitutions at Phe542 and Leu546 were PCR-amplified using synthetic DNA primer pairs (Table S1) and inserted into the gateway entry vector, pDONR221 (Thermo Fisher Scientific Inc., Waltham, MA, USA). Expression plasmids for GST-tagged p65_532–550 (GST-p65_532–550) and GST-tagged p65_{532–550, F542A + L546A} (GST-p65_{532–550, F542A + L546A}) were constructed by LR reaction between the entry clone and pDEST15 vector (Thermo Fisher Scientific Inc.), and Escherichia coli BL21-AI cells (Thermo Fisher Scientific Inc.) transformed with each plasmid were treated with 0.2% l-arabinose for 3 h at 37 °C for induction. Cell lysates were prepared by sonication of the cells in a lysis buffer (50 mM Tris-HCl(pH8.0), 150 mM NaCl, 0.1% Triron-x-100 and protease inhibitor cocktail (Nacalai Tesque, Kyoto, Japan) using Astrason XL2020. The expressed GST-p65_532–550 and GST-p65_{532–550, F542A + L546A} in the lysates were purified with Glutathione Sepharose 4B beads (Cytiva, Tokyo, Japan). the E.Coli extracts were excessively added to the Glutathione-Sepharose beads to obtain saturated protein bound beads by checking the flow-through portion to contain over-dose GST-fusion protein. Moreover, to eliminate the non-specific proteins or the dimerized GST-p65_532-550 bound onto Glutathione Sepharose 4B beads, The beads were washed with a wash buffer (400 mM NaCl, 0.5% NP-40, 10 mM MgCl₂, 10% glycerol, and 20 mM Tris-HCl (pH 7.4)) at three times [23].

2.3. U937 nuclear extract preparation

U937 nuclear extracts were prepared as previously described [16]. Cell pellets obtained from a one L culture were suspended in buffer A (10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)-NaOH (pH 7.9), 1.5 mM MgCl₂, 10 mM NaCl, and 1% protease inhibitor cocktail (Nacalai Tesque)) and crushed with a Dounce homogenizer (tight pestle). The resulting lysate was centrifuged at 19,000×g for 20 min, and the cytosolic fraction was removed. The nuclear pellet was resuspended in buffer C (20 mM HEPES-NaOH (pH 7.9), 25% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM ethylenediaminetetraacetic acid (EDTA), and 1% protease inhibitor cocktail) and crushed with the Dounce homogenizer (tight pestle). The nuclear extracts were centrifuged at 20,000×g for 30 min, and the supernatant (nuclear soluble fraction) was diluted with buffer D (20 mM HEPES-NaOH (pH 7.9), 25% glycerol, 1.5 mM MgCl₂, 0.2 mM EDTA, and 1% protease inhibitor cocktail). To check quality of the nuclear soluble fraction, this fraction together with the other fractions was boiled for 5 min in an SDS sample loading buffer containing 100 mM dithiothreitol (DTT). The denatured samples were separated on an SDS/15% polyacrylamide gel and transferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked with 2.5% bovine serum albumin (BSA) and 2.5% skim milk/tris (hydroxymethyl) aminomethane buffered saline (TBS) at 25 °C for 1 h. The primary antibody used was a Histone H3 rabbit polyclonal antibody (1/1000 dilution) (ab18521, Abcam, Cambridge, UK). The membrane was washed with TBS-Tween (TBS-T) and incubated with alkaline phosphatase (AP)-conjugated goat anti-rabbit IgG (1/3000 dilution) (1706520, Bio-Rad, Hercules, CA, USA). Histone H3 was detected using an AP conjugate substrate kit (1706432, Bio-Rad, Hercules, CA, USA).

2.4. Detection of nuclear soluble proteins that bind to GST-p65_532–550

GST-p65_532–550 bound to Glutathione Sepharose 4B beads (500 μL) was reacted with 10 mL of the U937 nuclear soluble fraction (12 mg/mL) at 4 °C for 2 h. The beads were washed with a wash buffer (400 mM NaCl, 0.5% NP-40, 10 mM MgCl₂, 10% glycerol, and 20 mM Tris-HCl (pH 7.4)), and proteins bound to GST-p65_532–550 were eluted in an elution buffer (100 mM glutathione (reduced form), 150 mM NaCl, 300 mM Tris-HCl (pH 8.8)) and concentrated by the trichloroacetic acid precipitation method. An experiment with GST-p65_{532–550, F542A + L546A} was parallelly carried out. The recovered proteins were separated on an SDS/8% polyacrylamide gel and visualized with CBB R-250.

2.5. Mass spectrometry (MS)

A CBB-stained band corresponding to a protein the size of ∼400 kDa was excised from the polyacrylamide gel and decolorized with 50% acetonitrile (MeCN) containing 50 mM ammonium bicarbonate. The excised gel was minced and dehydrated with MeCN. The protein in the gel pieces was digested with 20 ng/mL modified porcine trypsin in 50 mM ammonium bicarbonate at 37 °C for 16 h. The digested peptides were extracted from the gel pieces twice with 50% MeCN containing 5% TFA and concentrated by lyophilization. The dried digested peptides were dissolved in 50% MeCN containing 0.1% TFA, mixed with saturated α-cyano-4-hydroxycinnamic acid (CHCA) solution (10 mg/mL) on a polished steel target plate, and analyzed using Autoflex III TOF/TOF (Bruker, Billerica, MA, USA). Protein identification was performed by MASCOT search of the SwissProt H. sapiens database with a reporter ion tolerance of 0.5 Da, trypsin enzyme with one missed cleavage, and variable modifications including oxidation on methionine of the digested peptide.

2.6. Immunoprecipitation

A549 cells cultured in a 100 mm dish were treated with 10 ng/mL TNF-α for 0.5 h, washed with cold PBS, and suspended in 4 mL of buffer A containing 1% protease inhibitor cocktail, 10 mM sodium fluoride, 20 mM imidazole, and 10 μM Cantharidin. The cells were homogenized with a Dounce homogenizer and centrifuged at 1000×g for 10 min. The pellet was resuspended in 4 mL of the lysis buffer (50 mM Tris-HCl (pH 7.4), 137 mM NaCl, 1% Triton x-100, 5 mM EDTA, 5 mM EGTA, 1% protease inhibitor cocktail, 10 mM sodium fluoride, 20 mM imidazole, and 10 μM Cantharidin), and incubated at 4 °C for 30 min. After centrifugation at 16,000×g for 10 min at 4 °C, the supernatants were used as nuclear extracts for immunoprecipitation experiments. The nuclear extracts (1 mg/mL) were incubated with 4 μg of rabbit IgG isotype control (sc-2027, Santa Cruz Biotechnology, Dallas, TX, USA) or rabbit p65 IgG pAb (GTX102090, GeneTex, Irvine, CA, USA) at 4 °C for 16 h, and then further incubated with 20 μL of Protein G-Sepharose beads (17061801, Cytiva) at 4 °C for 1 h. After washing the beads with the lysis buffer, bound nuclear proteins were detached from the beads by boiling them for 5 min in an SDS sample loading buffer containing 100 mM DTT. The detached proteins were separated on a 3−14% gradient gel (UH-T/R314, ATTO, Tokyo, Japan) and transferred onto a Nitrocellulose membrane using Transblot turbo system (1704150J8, Bio-rad). The membrane was blocked with 3% BSA/TBS-T at room temperature for 1 h and treated with a primary antibody DNA-PKcs (1/200 dilution) (sc-5282, Santa Cruz Biotechnology), Ku70 (1/200 dilution) (sc-17789, Santa Cruz Biotechnology), Ku80 (1/200 dilution) (sc-5280, Santa Cruz Biotechnology), or p65 (1/1000 dilution) (GTX102090, GeneTex) at 4 °C for 16 h. After washing out, the membrane was incubated with a HRP-conjugated goat anti-mouse IgG antibody (1/5000 dilution) (SA00001-1, Proteintech, Tokyo, Japan) at 25 °C for 1 h. Proteins were visualized with a LAS 3000 mini (Fujifilm, Tokyo, Japan).

2.7. Expression of his-tagged DNA-PKcs fragments and pull-down assay

cDNAs corresponding to H. sapiens DNA-PKcs fragments were obtained by PCR from H. sapiens PRKDC/pCMV6 mammalian expression vector (#83317, Addgene, Watertown, MA, USA) and inserted into pDONR221 or pCR8/GW/TOPO (Thermo Fisher Scientific Inc.) (Table S1). A plasmid for expression of each His-tagged DNA-PKcs fragment was constructed using the LR reaction of each entry clone with the pDEST17 vector (Thermo Fisher Scientific Inc.) (Table S1). Each His-tagged DNA-PKcs fragment was expressed in E. coli BL21-AI cells (Thermo Fisher Scientific Inc.), and cell lysates were prepared as above.

GST-p65_532–550-bound glutathione beads (40 μL) were incubated with the cell lysates expressing each His-tagged DNA-PKcs fragment at 4 °C for 2 h. After washing the beads with a wash buffer (150 mM NaCl, 0.1% Triton X-100, 10% glycerol, and 20 mM Tris-HCl (pH 7.4)), bound polypeptides were detached by boiling the beads for 5 min in the SDS sample loading buffer containing 100 mM DTT. The eluate was analyzed by western blotting using a monoclonal antibody against 6 × His (1/500 dilution) (Biodynamics Laboratory Inc., Tokyo, Japan) and an AP-conjugated-goat anti-mouse IgG (1/3000 dilution) (1706520, Bio-Rad.) or HRP-conjugated-goat anti-mouse IgG (1/5000 dilution) (SA00001-1, Proteintech). The band intensity was analyzed with an image processing program, ImageJ v1.52a (the National Institutes of Health, Bethesda, Maryland, USA).

3. Results

3.1. Preparation of a GST-tagged polypeptide from the TA1 domain

In order to search for novel nuclear proteins that interact with the TA1 domain of NF-κB p65, we prepared a GST-tagged polypeptide, GST-p65_532–550, from Phe532–Ser550 of the TA1 domain (Fig. 1C). The 19 amino-acid peptide has been shown to be sufficient for binding to EP300/CBP [19]. This polypeptide was expressed from the pDEST15-based plasmid in E. coli BL21-AI cells, and affinity-purified with glutathione-beads. Another polypeptide, GST-p65_{532–550, F542A + L546A}, which contains two Ala substitutions at Phe542 and Leu546 in the ФXXФФ motif of the TA1 domain, was parallelly prepared as a negative control.

3.2. Identification of a large protein that binds to GST-p65_532–550 as DNA-PKcs

Prior to searching for nuclear proteins that bind to GST-p65_532–550, we examined U937 cell nuclear extracts for quality. Whole cell lysate was fractionated to obtain soluble nuclear extracts, and each fraction was subjected to western analysis with anti-histone-H3 antibody (Fig. 2A and B). Judging from the result that histone H3 was detected only in the whole lysate and insoluble nuclear extracts, the quality of the obtained soluble nuclear extracts seemed to be good enough for further experiments.

Fig. 2.

Identification of an ∼400 kDa protein that binds to GST-p65_532–550 as DNA-PKcs. (A) Fractionation procedure to obtain a nuclear soluble fraction from U937 cells. (B) Western blotting analysis with an anti-histone H3 antibody to check quality of the nuclear soluble fraction. (C) U937 cell nuclear soluble proteins that bound to GST-p65_532–550 were separated on an SDS/8% polyacrylamide gel and visualized with CBB R-250. An ∼400 kDa protein was identified as Homo sapiens DNA-PKcs by mass spectrometry. M, protein standards. (D) Western blotting analysis for proteins in nuclear extracts from TNF-α treated A549 cells that were immunoprecipitated with an anti-p65 antibody. Antibodies against DNA-PKcs, Ku80, Ku70, and p65 were used for detection. Ab-, nuclear proteins precipitated without antibodies.

We reacted glutathione-bead-bound GST-p65_532–550 with the U937 cell soluble nuclear extracts and analyzed proteins that bound to GST-p65_532–550 by polyacrylamide gel electrophoresis. We found various nuclear proteins that bound to GST-p65_532–550 but not to GST-p65_{532–550, F542A + L546A} (Fig. 2C). The largest protein the size of ∼400 kDa was subjected to mass spectrometric analysis and found to be DNA-PKcs (Table 1 and Fig. S1).

Table 1.

Mass spectrometric analysis of the GST-p65_532–550 binding protein.

Observed mass (m/z)	Sequence	Protein (Acc.no.)	Score	Coveragea (%)
Monoisotopic
1065.7038	DFGLLVFVR	DNA-PKcs (P78527)	50	2.88
1588.0620	KEVYAAAAEVLGLILR		196
1708.015	LQYFMEQFYGIIR		93
1960.263	GPVLRNCISTVVHQGLIR		72
2119.4134	KDVLIQGLIDENPGLQLIIR		134
2268.769	DFSAFINLVEFCREILPEK		119
2746.7607	KAILELHYSQELSLLYLLQDDVDR		201

^aRatio of the numbers of matched peptide residues to the total length of the protein.

To clarify if full-length cellular NF-κB p65 binds to DNA-PKcs, we performed an immunoprecipitation experiment with an antibody against p65 and nuclear extracts prepared from TNF-α-treated A549 cells. We detected a specific band corresponding to DNA-PKcs in immunoprecipitants (Fig. 2D), suggesting that NF-κB p65 indeed binds to DNA-PKcs in cells. Ku70 and Ku80, which should have bound to DNA-PKcs, were not detected probably due to below detection limit.

3.3. Analysis for GST-p65_532–550 binding sites on DNA-PKcs

Next, we investigated which regions of DNA-PKcs GST-p65_532–550 binds to by preparing a series of His-tagged DNA-PKcs fragments of 175–789 amino acids (Fig. 3A). Each of 7 fragments was expressed in E. coli and cell extracts were prepared. The extracts were incubated with glutathione-bead-bound GST-p65_532–550, and proteins that bound to GST-p65_532–550 were analyzed by western blotting with an anti-6x His antibody. Two fragments, His-DNA-PKcs_541-988 and His-DNA-PKcs_2408-2819, were found to bind to GST-p65_532–550, although their binding efficiencies defined by bound fragment intensity/input fragment intensity were 17% and 99%, respectively, indicating that the latter fragment bound to GST-p65_532–550 much more stably (Fig. 3B and C).

Fig. 3.

Analysis for GST-p65_532–550 binding sites on DNA-PKcs. (A) Structure of DNA-PKcs and its regions corresponding to seven His-tagged DNA-PKcs fragments. (B–C and E–H) E. coli extracts expressing His-tagged DNA-PKcs fragments were incubated with glutathione-bead-bound GST-p65_532–550, and proteins that bound to GST-p65_532–550 were analyzed by western blotting with an anti-6x His antibody. (D) DNA-PKcs regions corresponding to smaller His-tagged DNA-PKcs fragments. GST-p65_532–550-bound DNA-PKcs fragments were denoted by fragment name in red and by red asterisk. A blue asterisk in (C) denotes a degradation product of the DNA-PKcs fragment E. NC, sample without extracts. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

To narrow down GST-p65_532–550 binding regions on DNA-PKcs, we prepared two sets of smaller His-tagged DNA-PKcs fragments, one from the region 541−998 and the other from the region 2408−2819 (Fig. 3D). Western analysis showed that GST-p65_532–550 binds to His-DNA-PKcs_541-750, His-DNA-PKcs_2485-2576, and His-DNA-PKcs_{2485-2576(Δ2530–2543)} (Fig. 3E−H). Since an expression level of several DNA-PKcs fragments was very low, the possibility that GST-p65_532–550 also binds to other fragments would not be excluded. These results suggest that the TA1 domain of NF-κB p65 can interact at least with the region 541−750 and the region 2485−2576, where the subregion 2530–2543 is dispensable, of DNA-PKcs.

4. Discussion

In this study, we searched for novel nuclear proteins that interact with NF-κB p65 through its TA1 domain, and the current data suggested that DNA-PKcs can bind to p65 through the interaction of the TA1 domain with the region 541−750 in the N-HEAT domain or the region 2485−2576 in the M-HEAT domain (Fig. 3). It has been shown that NF-κB p50 becomes phosphorylated at Ser20 by DNA-PKcs and that the phosphorylation efficiency increases in the presence of p65 [21]. These observations together imply that, in addition to a direct interaction between p50 and DNA-PKcs, the interaction of p65's TA1 domain with DNA-PKcs may be also important for the Ser20 phosphorylation. The formation of stable NF-κB p50/p60 may be physiologically prerequisite for the phosphorylation at Ser20 by DNA-PKcs.

Cryo-electron microscopic analysis of DNA-PKcs has shown that the region 541−750 and the region 2485−2576 are located at quite a distance (Fig. 4) [24]. This implies that there may be at least two modes of interaction between DNA-PKcs and NF-κB p50/p65, although the former interaction mode may be much weaker judging from the results shown in Fig. 3B and C. Transition between the two modes may occur during the phosphorylation of p50 at Ser20.

Fig. 4.

Structure of DNA-PKcs corresponding to GST-p65_532-550 -bound DNA-PKcs regions (PDB: 5w1r). Two potential binding regions for the TA1 domain of NF-κB p65 are shown in red. N-HEAT domain, blue; M-HEAT domain, green, FAT region, purple; kinase domain, yellow; FATC domain, orange. The diagram was generated by RasMol version 2.7.5.2. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Since, in the current study, we analyzed the interaction between p65 and DNA-PKcs using a crude system, effects of other proteins on its interaction cannot be excluded. In future study, the interaction analysis with purified components and determination of dissociation constants would be needed. And our final goal is to determine the whole structure of the NF-κB p50/p65 and DNA-PKcs complex.

CRediT authorship contribution statement

Yuta Hasegawa: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing - original draft; Writing - review & editing. Shinichi Asada: Conceptualization, Project administration, Resources, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the supplementary data to this article.