A novel mechanism of nuclear factor-κB regulation by adenoviral protein 14.7K

Abstract

Viruses have evolved many different ways to evade immune attacks. The adenoviral E3 protein 14·7K effectively inhibits antiviral immunity and inflammation. However, the underlying mechanism for this effect is unclear. Here we show that 14·7K is a potent inhibitor of nuclear factor (NF)-κB transcriptional activity following Toll-like receptor (TLR) or tumour necrosis factor (TNF) receptor signalling. The inhibition of the NF-κB activity occurs downstream of IκBα degradation and NF-κB translocation into the nucleus. Analysis of NF-κB DNA binding reveals that 14·7K specifically inhibits p50 homodimer DNA binding and that this inhibition is mediated through the interaction of 14·7K with p50. We propose that 14·7K inhibits NF-κB activity through directly blocking p50 binding to DNA and that this is the basis for its anti-inflammatory properties. Our data also indicate a role for p50 homodimer-dependent transcription in inflammation.

Introduction

Viruses have developed many different ways to evade immune attacks. The E3 transcriptional unit of the human adenovirus type 5 contains several genes that function to counteract the host antiviral immunity. Deletion of E3 genes enhances antiviral immune responses and inflammation^1,2 whereas enforced expression of the E3 14·7K gene inhibits inflammation in both lung³ and liver.⁴ The E3 14·7K gene, which encodes a polypeptide of molecular weight 14 700·(hence 14·7K), has also been shown to inhibit tumour necrosis factor-α (TNF-α)-induced⁵ and TNF-related apoptosis-inducing ligand (TRAIL) -induced⁶ apoptosis in certain cell types. However, it is unclear if such an anti-apoptotic activity is responsible for its immunosuppressive effect. 14·7K has also been demonstrated to interact with IκB kinase/NF-κB essential modulator (IKKγ/NEMO) although no effect on the NF-κB activating function of IKKγ was observed.⁷ Understanding the mechanisms of 14·7K-mediated suppression of immune responses is crucial not only for the treatment of viral infections but also for the application of adenovirus-based vectors for vaccination and gene therapy because most E3 genes are deleted from the viral vectors used for the latter applications.

The inducible expression of inflammatory cytokines such as interleukin-6 (IL-6) and IL-12 by Toll-like receptors (TLR) is a key component of the innate immune response to infection and is critical for the development of an adequate adaptive immune response. The expression of many inflammatory cytokines including IL-6 and IL-12 is under the control of the NF-κB family of transcription factors, which include p50, p52, p65, c-Rel and RelB. The NF-κB family of proteins is present in the cytoplasm as an inactive homo- or hetero-dimer in association with an inhibitory IκB protein, which acts by masking the nuclear localization signal of the NF-κB, preventing its nuclear translocation. A wide variety of stimuli including TLR ligands and cytokines can activate NF-κB. Activation of NF-κB involves phosphorylation and proteolytic degradation of the IκB by specific IκB kinases. The free NF-κB in turn enters the nucleus, and binds to the κB sites of gene promoters. Members of the NF-κB family play crucial roles in both innate and adaptive immune responses. Disruption of NF-κB expression or transcriptional activity significantly compromises the ability of the immune system to control infections. It is therefore not surprising that viruses have evolved means to disrupt the NF-κB pathway to evade immune attacks.⁸

In this study, we seek to elucidate the molecular mechanisms whereby 14·7K inhibits innate immunity. We report that in response to TLR and TNF receptor signalling, 14·7K inhibits NF-κB transcriptional activity at a point distal to the initiating signal. We demonstrate that 14·7K prevents DNA binding of NF-κB complexes by directly binding to p50. We propose a novel mechanism of viral protein-mediated regulation of the NF-κB pathway.

Materials and methods

Mice

Normal C57BL/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME). All animal procedures were preapproved by the University of Pennsylvania Animal Care and Use Committee.

Recombinant adenoviruses and plasmids

The AdE1°E3° and Ad14·7K adenoviruses were kindly provided by Dr R. Rooke (Transgene, Strasbourg, France).⁴ Both viruses are based on human adenovirus type 5 with the E1 and E3 genes deleted. Ad14·7K carries the 14·7K gene driven by the CMV promoter in the E1-deleted region but is otherwise identical to AdE1°E3° (Fig. 1a). Recombinant adenoviruses were propagated in 293HEK cells and their titres were determined using the Adeno-X Rapid Titer kit acording to the manufacturer's instructions (Clontech, Palo Alto, CA). Gene transfer was performed by intravenous injection of 2 × 10⁹ infectious units (IU)/mouse of the viruses in 100 μl phosphate-buffered saline. The pRK5 FLAG-tagged 14·7K expression vector was generated by inserting the 14·7K polymerase chain reaction product amplified from wild-type adenovirus type 5 genomic DNA.

Figure 1.

14·7K inhibits inflammatory cytokine expression in vivo and in vitro. (a) Schematic representation of the genomic organization of the adenoviral vector deleted in the E1 and E3 regions (AdE1°E3°) and the recombinant adenovirus carrying the 14·7K expression cassette under the control of the CMV promoter inserted into the E1-deleted region (Ad14·7K). (b) IL-6 and IL-12p40 production in the liver. C57BL/6 mice (n = 4) were injected intravenously with 100 μl PBS (uninfected), or 2 × 10⁹ IU of AdE1°E3° or Ad14·7K in 100 μl PBS. Twenty-four hours later, liver extracts were prepared and cytokine concentrations were determined by ELISA as described in Materials and methods. Values shown are means and SEM of cytokines detected per mg of liver protein. (c) IL-6 and IL-12p40 production by macrophages. Bone marrow-derived macrophages from C57BL/6 mice, 10⁶ cells/culture, were infected with 2 × 10⁸ IU AdE1°E3° (white bars) or Ad14·7K (black bars). Forty-eight hours later, macrophages were either left untreated (UNT) or activated with 100 ng/ml of LPS. Cells were cultured for an additional 24 hr and cytokine concentrations in the supernatant were determined by ELISA. Representative results from three independent experiments are presented in (b) and (c).

Luciferase assay

Promoters containing genomic sequences of IL-6 (−1232 to +3) and IL-12p40 (−700 to +54) were cloned into the pGL3-basic vector (Promega, Madison, WI). Cells were transfected with expression and reporter constructs using Fugene-6 reagent (Roche, Indianapolis, IN). Luciferase activities of whole cell lysates were analysed using the Dual-luciferase reporter assay system (Promega). Co-transfection of the Renilla-luciferase expression vector pRL-TK (Promega) was used as an internal control. For all samples, the reporter data were normalized for transfection efficiency by dividing firefly luciferase activity by that of the Renilla luciferase.

Cell isolation, culture and treatment

RAW 264·7 macrophage cells and 293HEK cells were maintained in Dulbecco's modified essential Eagle's medium containing 10% fetal calf serum. Transfection using equal quantities of plasmid DNA was carried out using Fugene-6 (Roche), according to the manufacturer's recommendations. Transfection of cells using a green fluorescent protein (GFP) expression vector as control was used to establish a transfection efficiency of at least 90% in 293T cells. Bone marrow-derived macrophages were isolated and cultured in 30% L929 conditioned medium. Bone marrow-derived macrophages were infected with AdE1°E3° or Ad14·7K using a multiplicity of infection (MOI) of 200 IU/cell for 90 min. RAW macrophages and bone marrow-derived macrophages were stimulated with 100 ng/ml lipopolysaccharide (LPS) isolated from Escherichia coli 055:B5 (Sigma-Aldrich, St Louis, MI) for 24 hr before supernatants were collected for cytokine analysis.

Cytokine assays

Liver homogenates were prepared as previously described.⁹ Quantification of cytokines in liver homogenates and cell culture supernatants was carried out by sandwich enzyme-linked immunosorbent assay (ELISA) as reported.¹⁰

Immunoprecipitation and Western blotting

To prepare total protein extracts, cells were lysed in RIPA buffer containing 50 mm Tris–HCl pH 7·4, 1% nonidet P-40, 0·25% deoxycholate, 1·50 mm NaCl, 1 mm ethylenediamineteraacetic acid (EDTA), 1 mm sodium orthovanadate and 1 × Complete protease inhibitors (Roche, Indianapolis, IN). Cytoplasmic and nuclear extracts were obtained using a nuclear extract kit (Active Motif, Carlsbad, CA) following the manufacturer's instructions. For immunoprecipitation, equal amounts of cytoplasmic or nuclear extracts were adjusted to 1 mg/ml in RIPA buffer and incubated with primary antibody overnight at 4°, followed by incubation with protein A/G-agarose (Invitrogen, Carlsbad, CA) for 2 hr. Pellets were washed three times in RIPA buffer and resuspended in 2× sample buffer (100 mm Tris–HCl pH 6·8, 20% (w/v) sodium dodecyl sulphate, 20 mg/ml bromophenol blue and 20% glycerol). Equal volumes of resuspended immunoprecipitates were resolved by denaturing sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to a nitrocellulose membrane. Following immunoblotting, the membrane was developed using Pierce Supersignal reagent (Pierce, Rockford, IL). Anti-FLAG antibody was obtained from Sigma-Aldrich, and anti-p65, anti-p50, anti-p52, anti-c-Rel and anti-RelB antibodies were purchased from Santa Cruz Biotech (Santa Cruz, CA). Anti-IκBα and anti-histone H1 antibodies were obtained from Upstate Biotech (Charlottesville, VA).

Electrophoretic mobolity shift assay (EMSA) and DNA affinity purification assay (DAPA)

Five micrograms of nuclear protein was used for each sample. NF-κB consensus (5′-AGTTGAGGGGACTTTCCCAGG-3′) double-stranded oligonucleotides were purchased from Santa Cruz Biotechnology. Oligonucleotides were end-labelled with [γ³²P]ATP (Amersham Biosciences, Piscataway, NJ) using the T4 polynucleotide kinase (Promega). Binding reactions were prepared using 5 μg of nuclear extract with 50 000 counts/min of oligonucleotide in a 25 μl reaction volume containing 10 mm HEPES-KOH (pH 7·9), 50 mm KCl, 2·5 mm MgCl₂, 1 mm dithiothreitol, 10% glycerol, 1 μg DNAse free bovine serum albumin and 2·5 μg poly[d(I-C)] at room temperature for 30 min. For supershift analysis, antibodies were added to the reaction mixture on ice for 20 min prior to the addition of radiolabelled probes. Binding reactions were resolved on a 4% non-denaturing polyacrylamide gel at 22 mA for 3 hr at 4° in 1 × TBE (0·089 Tris–borate, 0·089 m boric acid and 0·002 m EDTA). Gels were subsequently dried and exposed to a phosphor screen and visualized on a phosphoimager (Amersham Biosciences). Densitometry was performed using imagequant 5·2 software (Amersham Biosciences). For measurement of NF-κB DNA-binding activity by DNA affinity purification, DNA-binding proteins were precipitated from nuclear extracts using agarose-conjugated consensus oligonucleotides (Santa Cruz Biotech) and analysed by Western blotting.

Results

14·7K inhibits inflammatory cytokine production both in vivo and in vitro

Although it has been reported that the adenoviral protein 14·7K inhibits innate immune responses to LPS,⁴ which activates NF-κB through TLR-dependent and TLR-independent means, the effect of 14·7K on the innate immunity to adenoviruses has not been studied. One of the earliest innate immune responses to adenoviruses in vivo is the expression of pro-inflammatory cytokines such as IL-6 and IL-12. To determine whether 14·7K regulates this response during adenoviral infection, we infected intravenously C57BL/6 mice with 2 × 10⁹ IU of human E1/E3-deleted adenovirus serotype 5 that did or did not carry the 14·7K gene (Fig. 1a). As reported, only the adenovirus that carried the 14·7K (Ad14·7K) conferred 14·7K expression in this system.⁴ Twenty-four hours after the infection, liver homogenates were prepared and tested for IL-12p40 and IL-6 by ELISA. Livers from uninfected mice were used as controls. As shown in Fig. 1(b), uninfected mice displayed no detectable IL-12 or IL-6 in the liver while mice infected with E1/E3-deleted adenovirus (AdE1°E3°) showed significant levels of IL-6 and IL-12. By contrast, Ad14·7K-infected mice showed significantly reduced levels of IL-6 and IL-12 in their liver homogenates compared to AdE1°E3°-infected animals (Fig. 1b). These data demonstrate that the expression of 14·7K is capable of inhibiting the innate immune response to adenovirus in the liver.

To determine if 14·7K-mediated inhibition of cytokine expression in the liver can be generalized to a better-defined TLR system in vitro, we examined the expression of IL-6 and IL-12 in bone marrow-derived macrophages following LPS stimulation. Consistent with the results from the liver, macrophages infected with Ad14·7K expressed significantly less IL-6 and IL-12 following stimulation with LPS relative to the control AdE1°E3°-infected cells (Fig. 1c). Infection of macrophage with a control GFP-expressing adenovirus demonstrated an infection efficiency of at least 70% under the conditions used (see Materials and methods). For both viruses, infected cells that were not treated with LPS did not express detectable levels of either IL-6 or IL-12. Taken together, these data demonstrate that 14·7K inhibits IL-6 and IL-12 production both in vivo and in vitro, a finding that has not been reported previously.

14·7K inhibits the promoter activities of inflammatory cytokines and impedes TLR-induced NF-κB activation

To determine whether the inhibitory effect of 14·7K on IL-6 and IL-12 expression occurs at or upstream of the transcriptional level, we utilized luciferase-based reporter assays to test the cytokine gene promoter activities. RAW macrophages were transiently transfected with empty vector or a 14·7K expression construct, along with IL-6 and IL-12p40 promoter reporter plasmids. Twenty-four hours after the transfection, cells were stimulated with LPS for 8 hr and their luciferase activities were measured. All luciferase activities were normalized to a cotransfected Renilla luciferase plasmid control. As shown in Fig. 2(a), expression of 14·7K in cells significantly inhibited the reporter activities of both the IL-6 and IL-12p40 promoters following stimulation with LPS. The inhibitory effect of 14·7K on the transactivation of these promoters suggests interference at or upstream of the transcriptional level.

Figure 2.

14·7K inhibits IL-6 and IL-12 promoters and impedes NF-κB-dependent gene transcription following Toll-like receptor ligation. (a) RAW macrophages were transiently cotransfected with the IL-6 or IL-12 promoter-luciferase (firefly) constructs (100 ng), a constitutive Renilla luciferase expression vector (10 ng) and 400 ng of empty expression vector (control) or 14·7K expression vector (14·7K). 24 hr later, cells were treated with or without 100 ng/ml LPS for 8 hr followed by measurement of luciferase activities. Promoter luciferase activity for each sample was normalized to the Renilla luciferase activity for the same sample. (b) RAW macrophages were transiently cotransfected with a NF-κB luciferase reporter construct containing the NF-κB consensus sequence (100 ng), a constitutive Renilla luciferase expression vector (10 ng) and 400 ng of empty expression vector (control) or 14·7K expression vector (14·7K). Twenty-four hours later, cells were treated with or without LPS (100 ng/ml), CpG DNA (1 μm), poly(I:C) (50 μg/ml), peptidoglycan (PGN, 10 μg/ml) and loxobrine (100 μm) for 8 hr and luciferase activity assayed. Reporter luciferase activity for each sample was normalized to the Renilla luciferase activity for the same sample. Data presented in both (a) and (b) are mean ± SEM of triplicate cultures and are representative of three independent experiments. Inhibition of reporter activity by 14·7K is statistically significant for all the ligands tested (P < 0·05 as determined by Student's t-test). UNT, untreated cultures.

The NF-κB transcription factor is a key mediator of TLR signalling and activator of cytokine genes such as IL-6 and IL-12. To investigate whether 14·7K interferes with NF-κB-mediated signalling of TLR, we examined the effect of 14·7K on a NF-κB reporter, which consists of four repeats of the NF-κB consensus binding site upstream of the firefly luciferase gene. RAW macrophages were transfected with the NF-κB reporter construct as well as an empty vector or 14·7K-containing expression vector. Twenty-four hours later, cells were stimulated with five different TLR ligands (LPS, CpG DNA, poly(I:C), peptidoglycan and loxobrine), all of which are known to activate the NF-κB pathway in these cells. Remarkably, 14·7K inhibited the NF-κB reporter activities induced by all these ligands, with the most significant effect on LPS-induced activity (Fig. 2b). A NF-κB reporter plasmid employing the NF-κB binding site of the interferon-β promoter yielded similar results (data not shown). These data indicate that 14·7K may directly target the NF-κB pathway.

14·7K acts downstream of IκB degradation and NF-κB translocation into the nucleus

The NF-κB transcription factor is composed of a family of five related proteins which may form homodimers or heterodimers. Activation of NF-κB occurs as a result of the phosphorylation and subsequent degradation of the inhibitor protein, IκBα. The liberated NF-κB then moves from the cytoplasm to the nucleus where it regulates gene transcription. Although previous studies showed that 14·7K did not inhibit degradation of the IκBα¹¹ the effect of 14·7K on NF-κB function downstream of IκBα has not been explored. To address this issue, we transfected 293HEK cells with either empty vector or 14·7K-containing expression vectors and the NF-κB consensus site reporter plasmid. As shown in Fig. 3(a), expression of 14·7K significantly inhibited NF-κB activity following TNF-α stimulation. Similar inhibition of NF-κB activity by 14·7K was observed following IL-1β stimulation (data not shown). Analyses of IκBα protein levels in cells after TNF-α stimulation demonstrated that 14·7K expression does not inhibit IκBα degradation, suggesting that the inhibition of NF-κB activity occurs downstream of IκBα degradation (Fig. 3b). To address this possibility, we next monitored the nuclear translocation of NF-κB. Cells were transfected with either an empty expression vector or a 14·7K-FLAG expression vector and stimulated with TNF-α for 15 min. Cells were then harvested and their nuclear and cytoplasmic extracts were prepared and analysed by Western blot using antibodies against p65 and p50. Interestingly, no significant differences were detected in NF-κB p65 and p50 nuclear translocation between control and 14·7K-transfected groups (Fig. 3c). No significant expression or translocation of other NF-κB family members was detected in these cells (data not shown). These results indicate that 14·7K-mediated inhibition of NF-κB transcriptional activity occurs downstream of NF-κB activation (as defined by IκBα degradation) and NF-κB translocation to the nucleus.

Figure 3.

14·7K inhibits NF-κB transcriptional activity downstream of IκB degradation and nuclear translocation. (a) 293HEK cells were cotransfected with a NF-κB luciferase reporter construct containing the NF-κB consensus sequence (100 ng), a constitutive Renilla luciferase expression plasmid (10 ng) and 400 ng of empty expression vector (control, white bars) or 14·7K expression plasmid (14·7K, black bars). After 24 hr, cells were treated with or without TNF-α (20 ng/ml) for 8 hr and assayed for luciferase activities. Reporter luciferase activity for each sample was normalized to the Renilla luciferase activity for the same sample. Inhibition of reporter activity by 14·7K is P < 0·05 when compared to control. (b) 293HEK cells were treated as in (a). Fifteen minutes after the TNF treatment, cells were harvested and whole cell protein lysates were prepared. Lysates were resolved by SDS–PAGE prior to immunoblotting using antibodies against IκBα and FLAG (which is tagged to the 14·7K plasmid). Immunoblotting with an antibody against β-actin was used to show equal loading of protein across samples. (c) 293HEK cells were treated as in (a). Fifteen minutes after the TNF-α treatment, cells were harvested, and nuclear and cytoplasmic were extracts prepared. Equal amounts of proteins from nuclear (N) and cytoplasmic (C) extracts were resolved by SDS–PAGE and immunoblotted using antibodies against p65, p50 and FLAG. Immunoblotting with an antibody against Histone H1 was performed to demonstrate efficient nuclear fractionation. Data presented in all panels are representative of three independent experiments.

14·7K blocks p50 binding to DNA in the nucleus

The normal distribution of NF-κB to the nucleus following TNF-α stimulation of 14·7K-expressing cells indicates that the inhibition of NF-κB transcriptional activity in these cells occurs in the nucleus. Interestingly, significant amounts of 14·7K protein are present in the nucleus of transfected cells suggesting a nuclear role for 14·7K (Fig. 3c). To test this possibility, we next examined the NF-κB-binding activities in 14·7K-expressing cells. 293HEK cells were transfected with 14·7K-FLAG-containing expression vector and stimulated with TNF-α for 15 min. Nuclear lysates were then prepared and the DNA-binding assay was performed using a double-stranded oligonucleotide containing the NF-κB consensus binding site. Immunoblotting with antibodies against p65 and p50 revealed a significantly reduced p50-binding activity in 14·7K-expressing cells relative to the control, while the p65 binding remained unchanged (Fig. 4a). Densitometric analysis of the blots showed more than 60% reduction in p50 binding but not p65 binding in 14·7K-expressing cells (Fig. 4b). These data indicate that 14·7K inhibits NF-κB activity through inhibition of p50 DNA binding but not of p65 DNA binding.

Figure 4.

14·7K inhibits p50 binding to DNA. (a) 293HEK cells were transiently transfected with either empty expression vector (Control) or a 14·7K expression construct (14·7K) as in Fig. 3. Twenty-four hours after the transfection, cells were treated with or without TNF-α for 15 min and harvested for the extraction of the nuclear lysate. Equal amounts of nuclear protein lysates were used in a DAPA employing the NF-κB consensus oligonucleotide. The purified DNA binding activity was then resolved by SDS–PAGE and immunoblotted using antibodies against p65 and p50. (b) The relative levels of p65 and p50 obtained in the DAPA were determined by densitometric analysis of the immunoblots. (c) RAW macrophages were infected with either AdE1°E3° control vector or the 14·7K-expressing Ad14·7K virus at a MOI of 200. Forty-eight hours after the infection, cells were treated with or without LPS (100 ng/ml) for 15 min and then harvested for preparing nuclear lysate extracts. Equal amounts of nuclear proteins were added to an EMSA binding reaction and incubated with a radiolabelled double-stranded oligonucleotide containing the NF-κB consensus binding site. p50 complexes were identified by the addition of an antibody against p50. Complexes were resolved by non-denaturing PAGE and visualized by exposure to autoradiographic film. * Indicates higher mobility p50 homodimer complexes.

To further study this novel mechanism of NF-κB regulation, an EMSA was performed using nuclear extracts obtained from cells infected with control AdE1°E3° virus or the 14·7K-expressing Ad14·7K virus. The identity of the p50-containing binding complexes was determined using the NF-κB consensus probe and the anti-p50 antibody. In agreement with the results from the DAPA, EMSA analysis demonstrated a marked inhibition of a higher mobility p50 complex corresponding to p50 homodimers in stimulated 14·7K-expressing RAW macrophages (Fig. 4c). Similar results were obtained from TNF-α-treated 293HEK cells (data not shown). These data reveal that the p50 homodimer complex is specifically targeted by 14·7K to inhibit NF-κB transcriptional activity.

14·7K binds to p50

To elucidate the mechanism underlying 14·7K-mediated inhibition of p50 DNA binding, we examined the possibility that 14·7K might interact with DNA, thereby preventing p50 homodimer binding. However, gel shift and super shift assays failed to show a 14·7K–oligonucleotide complex (data not shown). We next investigated the possibility that 14·7K might interact with the p50 directly thereby preventing it from binding to DNA. 293HEK cells were transfected with 14·7K-FLAG expression vector and stimulated with TNF-α for 15 min. Immunoprecipitation was performed employing an anti-FLAG antibody to pull down 14·7K followed by Western blot with an antibody to p50. We found that 14·7K interacts with p50 in both TNF-stimulated and unstimulated cells (Fig. 5). Additionally, the levels of coimmunoprecipitated p50 in nuclear and cytosolic fractions correlated with the translocation of p50 following TNF-α treatment. Similar interactions were obtained when immunoprecipitations were performed using an anti-p50 antibody (data not shown). By contrast, no interaction between 14·7K and other NF-κB members was detected under the same conditions (data not shown) indicating that 14·7K specifically targets p50.

Figure 5.

14·7K interacts with p50. 293HEK cells were transiently transfected with a 14·7K-FLAG expression construct as in Fig. 3. After 24 hr, cells were treated with or without TNF-α (20 ng/ml) for 15 min. Nuclear and cytoplasmic protein extracts were then prepared and used in an immunoprecipitation assay with an anti-FLAG antibody. Immunoprecipitation of TNF-α-stimulated nuclear lystate with preimmune mouse serum (PI) was used as a control for the specificity of the immunoprecipitation. Equal amounts of lysates and immunoprecipitates were resolved by SDS–PAGE and immunoblotted using antibodies against p50 and FLAG.

Discussion

Mutagenesis and deletional analysis of the adenoviral genome has identified a critical role for the E3 proteins in the attenuation of the host antiviral immune responses. Similar approaches have also highlighted an immune modulating role for these proteins in several experimental conditions including diabetes,¹² transplantation¹³ and hepatitis.⁴ The E3 protein 14·7K possesses potent anti-inflammatory and antiapoptotic properties that appear to be independent of other E3 proteins. However, the host target of 14·7K is not clear, although its interaction with caspase-8 was reported by one group¹⁴ but not by others.^15–18 Equally uncertain is whether the antiapoptotic properties of 14·7K contribute to its anti-inflammatory effects in animal models of inflammation.

In this study, we found that 14·7K negatively regulates inflammatory cytokine expression in macrophages and liver, a key component of the innate immune response. We further identified NF-κB as a novel host target of 14·7K. 14·7K modulates NF-κB transcriptional activity through interaction with p50 and inhibition of its DNA binding. This inhibition results in loss of NF-κB reporter activity in vitro, which may impair the expression of the NF-κB-dependent cytokines such as IL-12 and IL-6. Indeed, mice infected with adenoviruses constitutively expressing 14·7K display significantly reduced levels of these cytokines, indicating that 14·7K may play an anti-inflammatory role in vivo independent of its effect on apoptosis. However, it is possible that the 14·7K-mediated inhibition of the NF-κB activity is also related to its antiapoptotic effects. While the roles of NF-κB in apoptosis are complex, evidence exists that it may both promote and inhibit apoptosis.^19,20 Interestingly, 14·7K selectively inhibits apoptosis induced by TNF-α and TRAIL, but not Fas ligand, which may relate to the fact that Fas ligand is a considerably weaker inducer of NF-κB activity than TNF-α or TRAIL.²¹ However, it remains to be formally tested whether the inhibition of NF-κB by 14·7K, in addition to preventing IL-12 and IL-6 expression, also determines the sensitivity of cells to apoptosis.

The selective targeting of NF-κB by 14·7K underlines its anti-inflammatory effects in vivo. NF-κB is an essential component of the host immune and inflammatory responses and modulation or inhibition of NF-κB activity can have profound effects on these processes.²² NF-κB is critical for the expression of IL-6 and IL-12 following macrophage activation which, in turn, are crucial for initiating adequate adaptive immune responses. Indeed, mice deficient in IL-12 are resistant to LPS-induced hepatitis, a phenotype strikingly similar to Ad14·7K-infected mice. Therefore, inhibition of IL-12 expression by 14·7K may be responsible for its protection against LPS-induced hepatitis.

The selective targeting of NF-κB p50 by 14·7K and the subsequent inhibition of NF-κB transcriptional activity also point to an essential role of p50 homodimers in NF-κB-mediated transcription. Although the p50 homodimer is often regarded as a repressor of NF-κB-dependent gene transcription, accumulating evidence suggests that it also possesses positive transcriptional activation potentials. Transgenic mice lacking the inhibitory C-terminal domain of the p105 precursor of p50 display an accumulation of p50 homodimers accompanied by cell-specific positive and negative gene expression abnormalities.²³ In vitro studies have demonstrated that p50 homodimers act as potent transcriptional activators for several κB sites.²⁴ The expression of a number of genes, including Bcl-2,²⁵ p-selectin,²⁶ HIV1²⁷ and COX-2,²⁸ are positively regulated by p50 homodimers. The regulation of these genes by p50 homodimers involves the association of accessory proteins such as the transcriptional initiation factor TFII-I²⁷ and the IκB family member Bcl-3,^26,29 as well as other transcription factors such as AP-1³⁰ and members of the C/EBP family.²⁸ These findings indicate that p50 homodimer-driven gene expression may be regulated by the binding of alternative preinitiating complexes in a cell-specific manner. Alternatively, p50 homodimers may play a role in promoter remodelling, which is different from the intrinsic transactivating activity. Indeed, the recent identification of a number of mammalian regulatory proteins, which enhance^31,32 or inhibit³³ p65-dependent transcriptional activation through direct interaction, point to an additional level of regulation for the NF-κB family of transcription factors. In support of this view, it has recently been shown that mice deficient in IκBζ, a modulator of p50 homodimer DNA binding, are also deficient in LPS induced IL-6 and IL-12 expression.³⁴

In the context of adenoviral infection, the inhibition of p50 homodimer DNA binding and modulation of NF-κB transcriptional activity probably function to prevent elimination of the infected cells by suppressing the host antiviral defence. Interestingly, NF-κB can induce transcription of the E3 genes through the activity of p65-containing complexes at two κB sites located on the E3 promoter³⁵ which may not be affected by 14·7K because of its selective targeting to p50 homodimers. 14·7K has been previously used to inhibit inflammatory responses in gene therapy studies.⁴ Elucidation of the molecular basis for its anti-inflammatory properties as reported here will allow for further rational design of vectors incorporating 14·7K, for conditions under which modulation of the NF-κB response may be desirable.

Abbreviations

Ad

Adenovirus

DAPA

DNA affinity purification assay

EDTA

ethylenediaminetetraacetic acid

ELISA

enzyme-linked immunosorbent assay

EMSA

electrophoretic mobolity shift assay

IL-6

interleukin-6

IU

infectious unit

LPS

lipopolysaccharide

MOI

multiplicity of infection

NF-κB

nuclear factor-κB

SDS–PAGE

sodium dodecyl sulphate–polyacrylamide gel electrophoresis

TLR

Toll-like receptor

TNF-α

tumour necrosis factor-α