Biochemical Journal

Research article

p65 controls NF-κB activity by regulating cellular localization of IκBβ

Taras Valovka, Michael O. Hottiger

Abstract

NF-κB (nuclear factor κB) controls diverse cellular processes and is frequently misregulated in chronic immune diseases or cancer. The activity of NF-κB is regulated by IκB (inhibitory κB) proteins which control nuclear–cytoplasmic shuttling and DNA binding of NF-κB. In the present paper, we describe a novel role for p65 as a critical regulator of the cellular localization and functions of NF-κB and its inhibitor IκBβ. In genetically modified p65−/− cells, the localization of ectopic p65 is not solely regulated by IκBα, but is largely dependent on the NLS (nuclear localization signal) and the NES (nuclear export signal) of p65. Furthermore, unlike IκBα, IκBβ does not contribute to the nuclear export of p65. In fact, the cellular localization and degradation of IκBβ is controlled by the p65-specific NLS and NES. The results of our present study also reveal that, in addition to stimulus-induced redistribution of NF-κB, changes in the constitutive localization of p65 and IκBβ specifically modulate activation of inflammatory genes. This is a consequence of differences in the DNA-binding activity and signal responsiveness between the nuclear and cytoplasmic NF-κB–IκBβ complexes. Taken together, the findings of the present study indicate that the p65 subunit controls transcriptional competence of NF-κB by regulating the NF-κB/IκBβ pathway.

  • cellular localization
  • gene expression
  • inhibitory κB β (IκBβ)
  • nuclear factor κB (NF-κB)
  • tumour necrosis factor α (TNFα)

INTRODUCTION

NF-κB (nuclear factor κB) is a homo- or hetero-dimeric inducible transcription factor composed of several mammalian Rel proteins: p65 (RelA), p50 (NFKB1), p52 (NFKB2), c-Rel and RelB [1]. It controls various genes implicated in immune and stress responses, apoptosis, cell proliferation and differentiation [2]. To prevent aberrant activation of NF-κB, as it is often seen in inflammatory diseases or cancer [3,4], specific regulation of its transcriptional activity is essential. Controlled cellular localization of NF-κB is a critical mechanism of this regulation [5,6]. In most cells, NF-κB is retained in the cytoplasm by physical association with one of several IκB (inhibitory κB) proteins. The family of IκBs includes canonical IκBα, IκBβ and IκBϵ proteins, p50/p105 and p52/p100 precursors, as well as nuclear forms IκBζ andBcl-3 [5]. IκBα and IκBβ predominantly interact with ubiquitous forms of NF-κB containing p65, and modulate p65 functions in virtually all tissues [7,8]. The general role of IκBs in the regulation of NF-κB signalling is well established. However, the molecular mechanisms that determine localization and activity of NF-κB–IκB complexes are not well understood.

It has been established that NF-κB and IκBα continuously shuttle between the nucleus and the cytoplasm. A steady-state cytoplasmic localization of inactive NF-κB–IκBα is achieved due to the dominant nuclear export function of IκBα which shifts dynamic equilibrium of the shuttling [911]. Stimuli-induced degradation of IκBα releases NF-κB allowing its rapid accumulation in the nucleus mediated by the NLS (nuclear localization signal) activities of NF-κB subunits [1215]. Nuclear NF-κB binds to κB DNA consensus sequences and activates a specific subset of genes. An important consequence of NF-κB activation is an induction of the NFKBIA gene by NF-κB and re-synthesis of the IκBα protein. Newly synthesized IκBα dissociates NF-κB from DNA and promotes its nuclear export, thereby providing a negative regulatory feedback mechanism that critically influences the duration of the NF-κB response [1618].

The regulatory interplay between NF-κB and IκBβ, a homologue of IκBα, is less clear and a potential feedback mechanism has not been described. In contrast with IκBα, IκBβ shows stimuli-induced degradation that is characterized by a slow kinetics [19]. IκBβ is also not rapidly re-synthesized in a NF-κB-dependent manner as described for IκBα. Depending on the cell type or stimulus, IκBβ may instead undergo persistent degradation, leading to prolonged NF-κB activation [20]. Persistent NF-κB activity is also mediated by a hypophosphorylated IκBβ that forms a stable complex with NF-κB in the nucleus [21]. Basal phosphorylation of the C-terminal PEST (proline, glutamate, serine, threonine) domain in IκBβ inhibits NF-κB DNA binding and is thought to be primarily responsible for the formation of inactive NF-κB–IκBβ complexes. The main question here is how these latent NF-κB–IκBβ complexes are transported back to the cytoplasm when the NF-κB response is completed. Unlike IκBα, which contains a non-conventional nuclear import sequence [22,23] and a classical NES (nuclear export signal), IκBβ lacks both of these elements. This strongly argues that IκBβ cannot, on its own, provide nuclear import or export functions. It has been demonstrated that cytoplasmic κB-Ras protein interacts with the latent NF-κB–IκBβ complexes and prevents their nuclear import in resting cells [24]. However, this finding cannot explain how cytoplasmic localization of the NF-κB–IκBβ complexes is achieved in a post-induction period. Furthermore, it has been demonstrated that κB-Ras binds to only a fraction of the NF-κB–IκBβ complexes [25]. Thus it is necessary to clarify the mechanisms that regulate cellular distribution of the NF-κB–IκBβ complexes which do not bind to κB-Ras. The contribution of other proteins to the control of the IκBβ localization may provide an explanation.

In the present paper, we report that cellular localization of NF-κB and IκBβ is controlled by the NLS and NES of p65. Furthermore, we show that the constitutive nuclear localization of NF-κB and IκBβ mediated by p65 modulates the basal and cytokine-induced activation of inflammation-associated IP-10 [IFNγ (interferon γ)-induced protein 10 kDa] and ICAM1 (intercellular adhesion molecule 1) genes. The changes in gene expression correlated with differences in the DNA-binding activity of NF-κB and the stimulus-induced degradation of IκBβ observed for cytoplasmic and nuclear NF-κB–IκBβ. Taken together, the findings of the present study reveal a novel role for p65 in the regulation of the NF-κB/IκBβ pathway.

EXPERIMENTAL

Antibodies

Anti-p65(A) (sc-109), anti-p65(C20) (sc-372), anti-IκBα(H-4) (sc-1643) and anti-IκBβ(C20) (sc-945) (Santa Cruz Biotechnology), anti-HA(HA.11) (Covance) (HA is haemagluttinin), anti-α-tubulin (Sigma), HRP (horseradish peroxidase)-conjugated anti-rabbit and anti-mouse (Sigma), Cy3 (indocar-bocyanine)-conjugated anti-rabbit (Jackson Immunology) and FITC-conjugated anti-mouse (Covance) antibodies were used for immunological applications.

Plasmids

A Myc-tagged p65WT was cloned into retroviral pRLL vector [26] by substituting the GFP (green fluorescent protein) gene. A CITE (cap-independent translation enhancer) sequence fused to an antibiotic resistance bsd gene was introduced into the SalI site immediately downstream of the p65 coding sequence. To obtain a final pRRL-Myc-p65WT plasmid, a WPRE (woodchuck hepatitis virus post-transcriptional regulatory element) sequence [27] was introduced into an EcoRI site downstream of bsd. pRRL-Myc-p65NLSmut, pRRL-Myc-p65NESmut and pRRL-Myc-p65NLS/NESmut were generated by site-directed mutagenesis according to the QuikChange® protocol (Stratagene). Deletion mutants RHDL+NES and RHDL−NES were generated by PCR using the primers 5′-GCGGATCCACCATGGAGCAGAAGCTGATCAGCGAGGAGGACCTGATGGACGAACTGTTCCCCCTC-3′, 5′-TCGCTCGAGTTACAGGTCTTCATCATCAAACTGC-3′ and 5′-TCGCTCGAGTTACAGGGCCTCTGACAGCG-3′. The control plasmid contained the CITE/bsd insert instead of the GFP gene. The pRRL-HA-IκBβ plasmid was generated by cloning a PCR fragment coding for the human HA-tagged IκBβ1 isoform [28] into the BamHI and SalI sites of pRRL containing CITE/pac and WPRE sequences downstream of the insert. To generate the pLPCX-HA-IκBα plasmid, human IκBα cDNA was amplified by PCR as an HA-tagged version and cloned into the BglII and SalI sites of the pLPCX vector (Clontech). The envelope plasmid pMD.G and the packaging plasmid pCMV-ΔR8.91 have been described previously [29].

Cells and retroviral transduction

p65−/− MEFs (mouse embryonic fibroblasts) have been described previously [30]. Cells were cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% FBS (foetal bovine serum), 100 units/ml penicillin/streptomycin and non-essential amino acids at 37 °C and 5% CO2. Virus production and transduction of cells were performed as described previously [29]. Infected cells were selected using blasticidin S (2.5 μg/ml) or puromycin (3 μg/ml).

Immunoblot assay

For immunoblotting, cells were washed with PBS and lysed in ice-cold Nonidet P40 lysis buffer [20 mM Tris/HCl (pH 7.5), 200 mM NaCl, 0.5% Nonidet P40, 1 mM PMSF, 1 μg/ml pepstatin, 1 μg/ml aprotinin and 1 μg/ml leupeptin]. The lysates were resolved by SDS/PAGE, transferred on to a nitrocellulose membrane (Whatman), and immunoprobed with the antibodies indicated according to manufacturers' recommendations. Enhanced chemiluminescence detection (Pierce) was used in immunoblot assays.

Immunofluorescence

MEFs were plated at a density of 4.5×104 cells per chamber on poly-L-lysine-coated chamber slides (Lab-Tek) and incubated overnight at 37 °C and 5% CO2. The next day, cells were starved in serum-free medium for 14 h and stimulated with TNFα (tumour necrosis factor α) (20 ng/ml) for the times indicated. To inhibit nuclear export, leptomycin B (40 ng/ml) was added to the cell medium for the last 4 h of starvation. For immunofluorescent detection of proteins, cells were fixed in 4% paraformaldehyde for 10 min. After permeabilization in PBS containing 0.2% Triton X-100 for 5 min and blocking in 0.5% BSA for 20 min, slides were incubated with primary (1:300) antibodies followed by Cy3- or FITC-conjugated secondary antibodies (1:250). DNA was stained with a solution of PBS containing 300 nM DAPI (4′,6-diamidino-2-phenylindole) for 5 min. The chamber slides were rinsed with PBS, air-dried and mounted with Vectashield medium (Vector Lab). Immunofluorescence was analysed using an Olympus BX51 fluorescence microscope and images were captured using an Olympus DP71 CCD (charge-coupled device) camera. The representative images were used to determine the ratio of nucleus-to-cytoplasm fluorescence using the Olympus Soft Imaging System (Cell B).

Quantitative real-time PCR

Total RNA was isolated with a RiboPure kit (Ambion) and reverse-transcribed using a high-capacity cDNA reverse transcription kit (Applied Biosystems) according to manufacturers' recommendations. Quantitative real-time PCR was performed using mouse NFKBIA-, IP-10-, ICAM1- and β-actin-specific primers and SYBR Green PCR Core Reagents (Applied Biosystems) with subsequent detection of the melting curves using the StepOne Real-Time PCR System (Applied Biosystems). Results were normalized to β-actin expression and quantified using StepOne Software (v2.0) (Applied Biosystems).

EMSA (electrophoretic mobility-shift assay)

To prepare nuclear extracts for the EMSA, cells were starved for 14 h and treated with TNFα (10 ng/ml) for 20 min or left untreated. Cell fractionation was performed as described previously [31]. Nuclear extracts (5 μg) were incubated with a 32P-labelled DNA probe [31] in 20 μl of buffer [10 mM Tris/HCl (pH 7.5), 65 mM KCl, 5 mM MgCl2, 0.5 mM EDTA, 1 mM DTT (dithiothreitol), 0.1 μg/μl BSA, 10% glycerol and 0.05 μg/μl poly dI-dC] for 45 min at room temperature (21 °C). A supershift was produced by adding 1 μg of anti-p65(C-20) antibody to the mixture 5 min prior to loading. Samples were subjected to PAGE (4% gels) and analysed by autoradiography. Densitometric analysis was performed on digitized images using ImageQuant 5.2 software.

RESULTS

Constitutive and post-induction localization of p65 is regulated by its intrinsic NLS and NES activities

To study the role of p65-specific NLS and NES, p65−/− MEFs were genetically complemented with the wild-type or mutated p65 proteins by retroviral transduction. The mutants included full-length p65 mutated in the NLS and/or NES (p65NLSmut, p65NESmut and p65NLS/NESmut), as well as C-terminal deletion variants (RHDL+NES and RHDL−NES) lacking the transactivation domain (Figure 1A). The use of genetically modified p65−/− cells excludes dimerization of the mutants with endogenous p65 that can mask or reduce any possible effects of the mutations. The expression level of the recombinant full-length p65 proteins in complemented p65−/− MEFs was comparable with that of the endogenous p65 in normal MEFs (Figure 1B and Supplementary Figure S1A at http://www.BiochemJ.org/bj/434/bj4340253add.htm). The ectopic expression of p65WT or mutants restored the level of IκBα and IκBβ in the cells (Figure 1B). This suggested that the ectopic p65 proteins were able to activate transcription of the NFKBIA gene [17,18] and/or to enhance stability of IκBα and IκBβ [32,33], the functions previously assigned to endogenous p65. The relevance of the complementation has further been confirmed by revealing comparable TNFα-induced degradation and re-synthesis of IκBα, nuclear–cytoplasmic re-distribution of the endogenous and ectopic p65, and activation of the NFKBIA and IP-10 genes in the p65+/+ and p65−/− MEFs expressing ectopic p65WT (Supplementary Figures S1 and S2 at http://www.BiochemJ.org/bj/434/bj4340253add.htm).

Figure 1 Genetic complementation of p65−/− MEFs

(A) Schematic representation of retroviral constructs coding for the wild-type or mutated p65 proteins. The critical residues and inactivating mutations in the NLS and NES of p65 are underlined. (B) Mock-transduced and complemented p65−/− MEFs were analysed for the expression of ectopic p65 and endogenous IκBα and IκBβ proteins by immunoblot analysis. The molecular mass in kDa is indicated on the left-hand side.

Immunofluorescence microscopy of the quiescent and TNFα-stimulated p65−/− cells transduced with virally encoded p65 proteins revealed that mutation of the NLS and/or NES of p65 substantially changed steady-state and/or induced cellular distribution of the ectopic proteins (Figures 2A and 2C). We observed elevated nuclear staining of p65NESmut in unstimulated cells, whereas the wild-type, NLS and NLS/NES mutants of p65 were predominantly localized in the cytoplasm under similar conditions. The estimated nucleus-to-cytoplasm fluorescence ratio for p65NESmut was approx. 2.3-fold higher than that observed for p65WT despite comparable expression of the IκB proteins (Figures 2D and 2C). When cells were treated with TNFα, p65WT and p65NESmut efficiently accumulated in the nucleus displaying similar nucleus-to-cytoplasm ratios after 20 min of TNFα stimulation. This suggested that only a fraction of p65NESmut or its specific complexes was constitutively nuclear, and that the cytoplasmic portion of the protein could be induced to accumulate in the nucleus. Further analysis of the TNFα-stimulated cells revealed that, in contrast with p65WT, a post-induction nuclear export of p65NESmut was less efficient and delayed in time. After 60 min of TNFα stimulation, the ratios of nuclear-to-cytoplasm fluorescence specific for p65WT and p65NESmut were 0.35 and 0.86 respectively. Since degradation and re-synthesis of IκBα were similar in both cell lines, we concluded that nuclear export of NF-κB in the post-induction phase was dependent on the NES of p65 and could not be attributed to the NES function of IκBα alone. In agreement with others [12,14], we found that the induced nuclear accumulation of p65 required the presence of an intact NLS in the protein. Mutation of the NLS in p65 substantially reduced its nuclear import despite the normal kinetics of IκBα and IκBβ degradation. Moreover, inhibition of the nuclear import of p65 was accompanied by a reduction of the post-induction re-synthesis of IκBα. Simultaneous mutation of the NLS and NES elements in p65 partially restored the re-synthesis of IκBα, suggesting that the NF-κB response may depend on a ratio of its nuclear–cytoplasmic shuttling determined by the intrinsic NLS and NES activities of the p65 subunit (Figures 2A and 2D).

Figure 2 Cellular distribution of p65 is regulated by its NLS and NES elements

(A) and (B) Nuclear–cytoplasmic re-distribution of the wild-type and mutated forms of p65 in response to TNFα. The complemented p65−/− MEFs were starved for 14 h and then stimulated with TNFα (20 ng/ml) for the times indicated. The ectopic proteins were visualized by indirect immunofluorescence. Scale bar=20 μm. (C) The graphs summarize three experiments and present the nucleus-to-cytoplasm (Nuc/Cyt) ratio of p65 fluorescence in the cells (means±S.D.). (D and E) Degradation and re-synthesis of the endogenous IκBα and IκBβ in response to TNFα. The cells were treated as described in (A and B). Whole-cell lysates were immunoblotted with the antibodies indicated.

Expression of RHDL+NES and RHDL−NES in the p65−/− cells restored a steady-state level of IκBα, but prevented the p65-dependent re-synthesis of the inhibitor during a course of TNFα stimulation (Figure 2E). This allowed us to investigate TNFα-induced cellular distribution of the ectopic proteins without IκBα-mediated nuclear import/export. Similar to p65NESmut, RHDL−NES displayed increased nuclear staining in unstimulated cells (Figures 2B and 2C). Its NES-proficient counterpart RHDL+NES exhibited a cellular distribution comparable with that of p65WT, e.g. was found to localize predominantly in the cytoplasm. In response to TNFα, RHDL−NES translocated into the nucleus reaching a nuclear-to-cytoplasm ratio of 1.3. By contrast, RHDL+NES could not efficiently accumulate in the nucleus, with a large portion of the protein localizing in the cytoplasm at all time points of TNFα stimulation. The maximal ratio of the nuclear-to-cytoplasm fluorescence detected for this mutant was only 0.6. These results suggest that the NES of p65 operates to maintain nuclear export of the protein independently of IκBα. Despite the presence of NES in RHDL+NES, the protein was detected in both cytoplasmic and nuclear compartments upon TNFα treatment, indicating that the NES cannot fully compensate for the NLS function in this mutant. We therefore conclude that predominantly cytoplasmic localization of the p65 protein requires the intact NES of p65 and presence of IκBα. This is in agreement with previous findings suggesting the role of IκBα-specific NES in the regulation of nuclear–cytoplasmic shuttling of NF-κB−IκBα [911]. When the NES function of p65 and re-synthesis ofIκBα are compromised, as in the case of RHDL−NES, the protein remains predominantly nuclear over the entire period of TNFα stimulation (Figures 2B and 2C). Remarkably, the expression of IκBβ was not abrogated in the cells complemented with the deletion mutants of p65. However, the presence of IκBβ could not compensate for a loss of IκBα and prevent nuclear accumulation of RHDL−NES in the TNFα-treated cells. Thus we hypothesized that the expression of IκBβ is not sufficient for the nuclear export of NES-deficient p65 mutants.

p65 controls cellular localization of IκBβ

Next we investigated the effects of the p65 mutations on the localization of IκBα and IκBβ and their complexes. The cells were infected with viruses encoding HA-tagged IκBα or IκBβ. This enhanced expression of the individual IκB isoforms and their specific NF-κB–IκB complexes in the cells (Figure 3A and Supplementary Figure S3 at http://www.BiochemJ.org/bj/434/bj4340253add.htm). Immunofluorescent staining for p65 and the two IκBs revealed that both p65 and IκBα or IκBβ proteins were mainly cytoplasmic when p65WT was used (Figure 3B). TNFα-induced degradation of IκBα and IκBβ allowed p65WT to accumulate in the nucleus. At the later time points of the treatment, ectopic IκBα and IκBβ were re-synthesized from a constitutive viral promoter and co-stained with the p65WT predominantly in the cytoplasm. Similar cytoplasmic localization of IκBα was observed in unstimulated or TNFα-treated cells complemented with either p65NESmut, p65NLSmut or p65NLS/NESmut, suggesting that the deficiency in NLS and/or NES of p65 did not change the cellular distribution of IκBα. Interestingly, inactivation of the NES in p65 resulted in a constitutive nuclear staining of both p65 and IκBβ (Figure 3B and Supplementary Figure S3). An enriched nuclear presence of p65NESmut and IκBβ was also observed in TNFα-treated cells, suggesting that at least a fraction of the NF-κB–IκBβ complexes is regulated by the NES of p65. The use of the truncated RHDL+NES and RHDL−NES mutants (Supplementary Figure S3) revealed that the effect of the p65 NES dysfunction on IκBβ was not dependent on IκBα. Thus our results indicate that the NES of p65 is required to maintain predominantly cytoplasmic localization of IκBβ. By contrast, p65NLSmut or p65NLS/NESmut were found to co-localize with IκBβ mainly in the cytoplasm under all conditions tested. This indicates that inactivation of the NLS in p65 prevents constitutive nuclear accumulation of p65 and IκBβ, the phenotype observed for the p65NESmut-expressing cells. On the basis of this, we assume that the NLS of p65 is also involved in the regulation of the IκBβ compartmentalization. The mutual effect of the p65 mutants and IκBβ on their cellular localization suggests that formation of a complex between p65 and IκBβ is required for this phenomenon. This was confirmed by co-immunoprecipitation experiments demonstrating that p65WT and all mutants were able to form IκBα- or IκBβ-specific complexes (Supplementary Figure S4 at http://www.BiochemJ.org/bj/434/bj4340253add.htm).

Figure 3 The p65-specific NLS and NES control cellular localization of IκBβ

(A) Ectopic expression of the p65, HA–IκBα and HA–IκBβ proteins in the complemented p65−/− MEFs. The cells were infected with viruses encoding HA–IκBα or HA–IκBβ and were analysed for the expression of ectopic proteins by immunoblot analysis. The molecular mass in kDa is indicated on the left-hand side. (B) Co-localization of the wild-type and mutated forms of p65 (red) with IκBα or IκBβ (green) was analysed by immunofluorescence in unstimulated and TNFα-treated cells. Scale bar=20 μm.

Nuclear–cytoplasmic shuttling of p65 and IκBβ

Since mutation of the NLS and/or NES of p65 influenced the localization of IκBβ, we assumed that the NF-κB–IκBβ complex may shuttle between the nucleus and the cytoplasm. To explore this possibility, the unstimulated cells were treated with LMB (leptomycin B), an inhibitor of CRM1 (chromosome region maintenance 1)-mediated nuclear export. Within 4 h of LMB treatment, both p65WT and IκBα or IκBβ accumulated in the nucleus (Figure 4). This indicates that IκBα and IκBβ, and their complexes, shuttle between the nucleus and the cytoplasm. Mutation of the NLS or NES of p65 did not substantially change the ability of p65 and IκBα to be trapped by LMB in the nucleus. However, inactivation of the NLS in p65 reduced LMB-mediated nuclear accumulation of p65 and IκBβ, suggesting that NF-κB–IκBα and NF-κB–IκBβ complexes may possess a different rate of nuclear import. It is also evident from our results that the NLS from other proteins may contribute to nuclear import of NF-κB–IκBα and NF-κB–IκBβ when the p65-specific NLS is mutated. Those NLS may originate from other Rel proteins present in NF-κB–IκBα and NF-κB–IκBβ complexes (Supplementary Figure S4).

Figure 4 Nuclear–cytoplasmic shuttling of p65 and IκBβ

The cells were starved and treated with LMB (40 ng/ml) for the last 4 h of starvation. The p65 and HA–IκB proteins were visualized as described in Figure 3(B). DAPI (blue) was used to localize nuclei. Scale bar=20 μm.

Constitutive nuclear localization of p65 and IκBβ modulates activity of NF-κB

To explore the role of NF-κB and IκBβ shuttling between the nucleus and the cytoplasm, we analysed possible effects of changes in a steady-state localization of p65 and IκBβ on gene expression. Since mutation of the NES in p65 shifted both p65 and IκBβ to the nucleus, we compared basal and TNFα-induced expression of p65-dependent genes in cells ectopically expressing IκBβ and p65WT or p65NESmut (Figure 5). Quantitative real-time PCR of the IP-10 and ICAM1 gene products revealed that their basal expression was approx. 1.6-fold higher in the cells expressing constitutive nuclear p65NESmut–IκBβ complex when compared with control cells. Although nuclear retention of p65NESmut and IκBβ elevated basal NF-κB activity, it was not sufficient to trigger the full transcriptional activation by NF-κB. As shown in Figure 5, the expression of IP-10 and ICAM1 was stimulated by TNFα. Interestingly, the induced expression of these genes was reduced by up to 50% in cells expressing nuclear p65NESmut–IκBβ when compared with control cells. This suggests that the nuclear localization of p65 and IκBβ renders their complex less responsive to TNFα stimulation. At the same time, the expression of the NFKBIA gene did not differ significantly in those cells (Figure 5), indicating that only a subset of genes is dependent on a steady-state localization of p65 and IκBβ. Involvement of other complexes, e.g. NF-κB–IκBα, in the regulation of NFKBIA may provide an explanation. Similar results have been observed in the p65WT- or p65NESmut-complemented cells without ectopic expression of IκBβ (Supplementary Figure S5 at http://www.BiochemJ.org/bj/434/bj4340253add.htm).

Figure 5 Nuclear localization of p65 and IκBβ elevates basal but reduces TNFα-induced expression of inflammatory genes

Expression of the IP-10, ICAM1 and NFKBIA genes in response to TNFα. The gene-specific mRNAs were quantified by quantitative real-time PCR at the indicated time points of TNFα treatment (10 ng/ml) using a relative standard curve method. Values are means±S.D. of three experiments. RQ, relative quantification value.

Analysis of the NF-κB DNA-binding activity in the nuclear extracts of both cell lines revealed that the basal level of DNA binding of p65NESmut was elevated when compared with p65WT (Figure 6A). By contrast, the TNFα-induced DNA binding was higher in cells expressing p65WT. This correlated well with differences observed for the basal and induced expressionof IP-10 and ICAM1 in those cell lines. The DNA binding of NF-κB is regulated by the IκB component of the NF-κB–IκB complexes [34]. Since degradation of IκB is required for the NF-κB activation in response to various stimuli [13,19], we studied TNFα-induced degradation of IκBβ in cells expressing predominantly cytoplasmic or nuclear NF-κB–IκBβ complexes (Figure 6B). Degradation of ectopic (Figure 6B) as well as endogenous (Figure 2D) IκBβ proteins was reduced in cells expressing p65NESmut compared with control cells. Hence, nuclear retention of IκBβ protected it from TNFα-induced degradation that correlated with attenuation of IP-10 and ICAM1 expression. A similar resistance to signal-induced degradation has previously been reported for the nuclear fraction of IκBα [10,35].

Figure 6 Nuclear retention of p65 and IκBβ affects the DNA-binding activity of NF-κB and the stability of IκBβ

(A) EMSA of nuclear extracts from the cells expressing cytoplasmic p65WT-IκBβ and nuclear p65NESmut-IκBβ using a radioactively labelled NF-κB consensus oligonucleotide. Distinct lines of the same EMSA gel are separated by boxes. Quantification of the bound probe was performed by densitometric analysis of the representative gel. (B) Degradation of HA–IκBβ in cells expressing p65WT or p65NESmut in response to TNFα. Cells were starved and treated with TNFα (10 ng/ml) for 20 min. Whole-cell lysates were subjected to immunoblot analysis. The molecular mass in kDa is indicated on the left-hand side.

DISCUSSION

Changes in localization and processing of key NF-κB–IκB complexes are important regulatory steps in NF-κB signalling. Previous studies have implicated a critical role of IκBα in the control of nuclear–cytoplasmic distribution of the NF-κB–IκBα complexes. The importance of the CRM1-dependent nuclear export function of IκBα in maintaining predominantly cytoplasmic localization of inactive NF-κB has been demonstrated [911]. Although the proposed mechanism clarifies the function of IκBα, it does not explain all aspects of NF-κB regulation. For instance, it cannot be applied to the regulation of NF-κB–IκBβ complexes, as none of the known NES sequences are found in IκBβ. Recent studies have demonstrated that the regulation of NF-κB–IκBs is much more complex and may require a functional interplay betweenNF-κB and IκB proteins [32]. The crystal structure of the NF-κB–IκBα and NF-κB–IκBβ complexes reveals that the NLS of one NF-κB subunit does not contact IκB and is mostly solvent-exposed, suggesting its possible contribution to the nuclear import of the complexes [3638]. Furthermore, a putative NES has been identified in p65, but any functional relevance of this element in NF-κB–IκB complexes remains to be investigated [39]. Hence, it is possible that NF-κB itself is critically involved in the cellular localization of NF-κB–IκB complexes.

In the present study we have addressed an essential role of the p65 subunit of NF-κB in the regulation of localization and functions of NF-κB and its inhibitor IκBβ. We genetically reconstituted expression of p65 or its mutants in p65−/− MEFs and demonstrated that both the NLS and NES of p65 are essential for the regulation of its nuclear–cytoplasmic distribution before and after TNFα stimulation. Mutation of the p65 NES enriched constitutively the nuclear fraction of NF-κB which probably represented nuclear NF-κB–IκBβ complexes, as shown in our later experiments. However, the main fraction of p65NESmut remained cytoplasmic and co-localized with IκBα. Although IκBα retained a large portion of NES-deficient p65 in the cytoplasm of unstimulated cells, it failed to rapidly export this mutant from the nucleus after TNFα treatment, suggesting a contribution of the p65 NES to the termination of the NF-κB response. Interestingly, the activity of the p65 NES appears to be essential, but not sufficient, for the termination of the nuclear localization of NF-κB in a post-induction period. Despite the presence of NES in RHDL+NES, it was detected in both the nuclear and cytoplasmic compartments when IκBα re-synthesis was compromised in TNFα-treated cells. Thus we propose that p65 and IκBα operate together to rapidly terminate NF-κB activation and to restore the cytoplasmic pool of inactive NF-κB–IκBα. This statement is in good agreement with studies indicating that the NES of p65 is located outside of the region involved in the binding to IκBα [40]. It is likely that the NES of p65 is exposed and accessible for the nuclear export protein CRM1 when the NF-κB–IκBα complex is formed. Inactivation of the NLS in p65 not only reduced nuclear import of NF-κB in response to TNFα, but also prevented the phenotype observed for p65NESmut. This suggests that the NF-κB response is determined by the ratio of its intrinsic NLS and NES activities.

IκBβ has been shown to rescue the lethal phenotype observed for IκBα-deficient mice when it was expressed under the control of the promoter and regulatory sequence of the NFKBIA gene [41]. These results indicate that IκBα and IκBβ share significant similarities in their biochemical activities and may possess overlapping functions. Surprisingly, we found that IκBβ could not substitute for IκBα when the NES of p65 was mutated. In the absence of IκBα, newly re-synthesized IκBβ did not promote nuclear export of RHDL−NES at the late time points of TNFα stimulation. Furthermore, in contrast with IκBα which co-localized with p65NESmut in the cytoplasm, IκBβ was predominantly nuclear when co-expressed with p65NESmut. Our results therefore suggest that, unlike IκBα, IκBβ is not directly involved in the nuclear export of p65. In fact, we have shown that the cytoplasmic localization of IκBβ is mediated by the NES of p65. Another important observation is that the NLS of p65 participates in the nuclear import of IκBβ, supporting our hypothesis that p65 controls NF-κB–IκBβ compartmentalization.

A previous report has indicated that the NES of p65 is sensitive to LMB, which is a specific inhibitor of the nuclear export receptor CRM1 [39]. Given the role of the NLS and NES of p65 in the cellular localization of IκBβ it is logical to assume that complexes containing p65 and IκBβ may shuttle between the nucleus and the cytoplasm in a CRM1-dependent manner. Indeed, incubation of cells in the presence of LMB led to a nuclear accumulation of a substantial portion of the p65 and IκBβ proteins, indicating their nuclear–cytoplasmic shuttling. This observation contrasts with previous reports suggesting that IκBβ, but not IκBα, functions as a static cytoplasmic inhibitor of NF-κB which does not undergo dynamic shuttling [42,43]. The discrepancies could be explained by the differences in experimental settings, e.g. by the distinct time of LMB treatment. Similar discrepancies caused by different conditions of LMB treatment have been reported for NIK (NF-κB-inducing kinase) and IκBϵ. Both proteins were initially identified as non-shuttling proteins [10,43], but later they were shown to shuttle between the nucleus and the cytoplasm using a prolonged LMB treatment [44,45]. It is likely that NF-κB–IκBβ is characterized by a slower rate of nuclear import when compared with the NF-κB–IκBα complex. Hence, a short treatment with LMB used in the previous studies would be sufficient to trap NF-κB–IκBα in the nucleus, whereas NF-κB–IκBβ would still remain predominantly cytoplasmic. We have demonstrated that inactivation of the NLS of p65 reduced LMB-mediated nuclear staining of p65 and IκBβ, but not IκBα, suggesting that NF-κB–IκBα and NF-κB–IκBβ possess different rates of nuclear import. It has been shown previously that signalling molecules of the NF-κB pathway constitutively shuttle between the nucleus and the cytoplasm exhibiting distinct kinetics of nuclear import [44]. For example, IκBα showed a rapid nuclear accumulation and was detected in the nucleus after 30 min of LMB treatment. In contrast, nuclear accumulation of IKK1 (IκB kinase 1) required exposure to LMB for 2–4 h. The kinetics of accumulation in the nucleus will not only depend on the biochemical properties of NLS and NES, but also on complex formation with other proteins that influence nuclear import or export and dissociation parameters of these complexes. Interestingly, κB-Ras has been reported to bind to the ankyrin-repeat domain of IκBβ and to retain NF-κB–IκBβ in the cytoplasm. Similarly, a cytoplasmic retention of NF-κB–IκBα mediated by interaction with the G3BP2 protein has been described previously [46]. It is possible that κB-Ras and G3BP2 affect the dynamics of nuclear–cytoplasmic shuttling of NF-κB–IκBs by modulating the specific ratio of their NLS and NES activities.

The cellular distribution of individual NF-κB–IκBs appears to be very dynamic and specifically regulated suggesting that their compartmentalization may play a more complex regulatory role than simply shutting down NF-κB-induced gene expression. In the present study, we attempted to address how cellular localization of NF-κB–IκBβ affects basal and stimuli-induced expression of the NF-κB-responsive genes. For this purpose, we specifically modulated a steady-state cellular distribution of NF-κB–IκBβ by mutating the p65-specific NES. As discussed above, this led to constitutive accumulation of NF-κB–IκBβ in the nucleus, whereas NF-κB–IκBα remained predominantly cytoplasmic. The constitutive presence of NF-κB–IκBβ in the nucleus increased basal NF-κB activity which resulted in elevated levels of ICAM1 and IP-10 expression. The dynamic dissociation/association of the nuclear NF-κB–IκBβ could explain such a basal activity of NF-κB. A similar increase in basal expression of those genes was observed in various autoimmune and chronic inflammatory disorders, or in tumours which are often characterized by constitutive nuclear NF-κB activity [4,47,48]. Further studies are needed to investigate the possible role of nuclear NF-κB–IκBβ complexes in the up-regulation of inflammatory genes linked to the pathogenesis of these diseases. Interestingly we found that, similar to IκBα [10,35], nuclear retention of IκBβ reduced its TNFα-induced degradation. Consistent with this, p65 has been shown previously to regulate the stability of IκBβ, but the mechanism was not discussed [32]. In our studies, we observed that the nuclear retention of IκBβ correlated with a decrease in activation of the ICAM1 and IP-10, but not NFKBIA expression, in response to TNFα. Thus we propose that changes in steady-state localization of NF-κB–IκBβ may specifically regulate a subset of NF-κB-dependent genes. The differences in NF-κB response observed for the cytoplasmic and the nuclear NF-κB–IκBβ complexes may be dictated by a distinct repertoire and mode of function of various NF-κB regulators in those cellular compartments. For example, NF-κB–IκBβ can undergo different post-translational modifications or resides in distinct complexes depending on its actual compartmentalization.

The exact cellular role of nuclear NF-κB–IκBβ complexes currently remains unclear. Based on the results of our studies, it is tempting to speculate that localization of NF-κB–IκBβ may be affected by changes of its nuclear import/export rates under some physiological conditions to modulate constitutive and/or stimuli-induced NF-κB activity. This is supported by studies demonstrating that nuclear expression of IκBβ is associated with specific stages of B-cell differentiation [49]. In contrast with pre-B-cells, mature WEHI-231 B-lymphoma cells were characterized by the expression of a hypophosphorylated form of IκBβ highly abundant in the nucleus. It was hypothesized by the authors that nuclear IκBβ is responsible for the constitutive NF-κB activity generally observed for this cell line. There is also evidence that nuclear localization of IκBβ is associated with the pathology of HIV-1 infection. IκBβ was present in the nucleus of HIV-1-infected myeloid cells and participated in ternary complex formation with NF-κB and DNA maintaining persistent NF-κB activity [50].

In summary, we propose that the nature of NF-κB responses is determined by the specific compartmentalization of the NF-κB and IκBβ proteins which largely depends on the NLS and NES functions of p65. Further definition of the mechanisms regulating cellular localization of IκBβ by p65 may help to develop strategies for specific targeting of genes downstream of the NF-κB/IκBβ pathway.

AUTHOR CONTRIBUTION

Taras Valovka performed research and contributed to experimental design, data analysis and manuscript writing. Michael O. Hottiger contributed to data analysis and manuscript writing.

FUNDING

M.O.H. is supported by the Swiss National Foundation [grant number SNF 31–122421], and the Kanton of Zurich.

Acknowledgments

We thank Didier Trono (School of Life Sciences, Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland) for retroviral constructs and Anja Reintjes for technical assistance. T. V. is grateful to Klaus Bister for continuous support. We thank Klaus Bister and Markus Hartl for their critical reading of this manuscript prior to submission.

Abbreviations: CITE, cap-independent translation enhancer; CRM1, chromosome region maintenance 1; Cy3, indocarbocyanine; DAPI, 4′,6-diamidino-2-phenylindole; EMSA, electrophoretic mobility-shift assay; GFP, green fluorescent protein; HA, haemagluttinin; IκB, inhibitory κB; ICAM1, intercellular adhesion molecule 1; IP-10, IFNγ (interferon γ)-induced protein 10 kDa; LMB, leptomycin B; MEF, mouse embryonic fibroblast; NES, nuclear export signal; NF-κB, nuclear factor κB; NLS, nuclear localization signal; TNFα, tumour necrosis factor α; WPRE, woodchuck hepatitis virus post-transcriptional regulatory element

References

View Abstract