J Biol Chem, Vol. 275, Issue 20, 15193-15199, May 19, 2000

Inducible NF-B Activation Is Permitted by Simultaneous Degradation of Nuclear IB^*

Patricia Renard^§, Yann Percherancier^§, Mathias Kroll^¶, Dominique Thomas, Jean-Louis Virelizier, Fernando Arenzana-Seisdedos, and Françoise Bachelerie

From the Unité d'Immunologie Virale, Institut Pasteur, 28 rue du Dr. Roux, 75015 Paris, France

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Signal-induced phosphorylation and ubiquitination of IB targets this inhibitor of NF-B for proteasome-mediated degradation, thus permitting the release of active NF-B. Upon cell stimulation, NF-B activation results in neotranscription and neosynthesis of its own inhibitor, IB. As reported earlier, the neosynthesized inhibitor is then accumulated in the nucleus, where it rapidly binds to and terminates the function of nuclear NF-B upon withdrawal of the stimulus. The present work was aimed at understanding how NF-B activity is preserved while stimuli persist, despite intense, simultaneous IB neosynthesis, which would be expected to end NF-B activity. We here show that incoming IB in the nucleus represents a target for resident nuclear proteasome complexes. Signal-induced, proteasome-dependent degradation of phosphorylated and ubiquitinated IB occurs in the nucleus, thus permitting the onset and persistence of NF-B activity as long as stimulation is maintained. Our results suggest that intranuclear proteolysis of IB is necessarily required to avoid self-termination of NF-B activity during cell activation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Inhibitory IB proteins tightly control the biological activity of Rel/NF-B transcription factors through their association with homo- or heterodimers of this family. Members of the family share a highly conserved NH₂-terminal sequence termed the Rel homology domain, which is required for DNA binding, dimerization, nuclear localization, and interaction with the IB molecules. In response to an inflammatory stimulus, cytokine, or viral infection, IB proteins are rapidly degraded by the 26 S multicatalytic proteasome. Degradation of IB, the most intensively characterized inhibitor, requires phosphorylation on serine residues 32-36 (1-6) by the activated IB kinase complex (reviewed in Ref. 7). This modification triggers recognition of IB by the F-box/WD -TrCP protein, the receptor of the SCF E3 ubiquitin ligase, which marks IB for ubiquitin-mediated proteolysis (8-13). As a consequence of IB degradation, the freed NF-B accumulates in the nucleus, where it activates gene transcription. NF-B acts on genes coding for cytokines, chemokines, immune receptors, and adhesion molecules, and its activation leads to a coordinated increase in the expression of inflammatory and immune response mediators (reviewed in Ref. 14).

Apart from the well characterized inhibitory function on NF-B in the cytoplasm, IB also participates in the inhibition of NF-B-dependent transcription in the cell nucleus. Once the stimulus is withdrawn, NF-B activity is rapidly shut down, ensuring that the B-dependent transcriptional activity is only transient (15, 16). This is accounted for by two mechanisms. First, free, non-NF-B-associated IB has the capacity to enter the nucleus when the protein is overexpressed from a heterologous promoter (17, 18). Such a property seems to rely on an active process mediated by a non-canonical nuclear import sequence located within the second ankyrin domain of IB protein (19, 20). Second, IB has the ability to both prevent NF-B binding to and to dissociate NF-B from specific DNA consensus sequences (18, 21, 22). Nuclear localization of IB is induced by stimuli activating NF-B and can be considered as part of a physiological mechanism regulating NF-B-dependent transcription. This assumption is supported by the fact that a massive accumulation of IB, which becomes detectable in the nucleus upon extinction of the cell signaling, occurs concomitantly with loss of NF-B-DNA binding activity and extinction of NF-B-dependent transcription (15). Compelling additional evidence for a role of nuclear IB in the regulation of NF-B activity in vivo came from a murine model of IB gene knockout (23, 24). Indeed, fibroblasts from IB-deficient mice showed an abnormally long lasting expression of nuclear NF-B upon removal of TNF (tumor necrosis factor ) (23). In addition to that mechanism, termination of NF-B-dependent transcription by IB could be completed by a newly described retrograde transport of NF-B·IB complex from the nucleus to the cytoplasm (25). A nuclear export sequence (NES)¹ (IQQQLGQLTLENL) located in the C-terminal (residues 265-277) region of IB, which resembles the prototypical human immunodeficiency virus type 1 Rev NES (LPPLERLTLD) (25, 26), would confer the protein with the capacity to interact with the Crm1-dependent export pathway (27). This pathway ensures retrograde transport toward the cytoplasm of numerous proteins containing a homologous NES (reviewed in Ref. 28). However, a new NES motif (MVKELQEIRLE) was recently identified within the N-terminal 45-55 residues of IB (29, 30). Both studies demonstrate that nuclear exclusion of the NF-B·IB complexes depends critically on this N-NES and is mediated by a Crm1-dependent pathway (29, 30).

Persistent activation of NF-B, required for full T-cell responses in vivo (31), takes place while IB is being resynthesized and localized to the nucleus. The question arises of how termination of NF-B-dependent transcriptional activity by IB is prevented if continuous nuclear import of the IB molecule occurs early following the onset of cell activation and is prolonged while cell signaling persists. To address this question, we have investigated whether IB expression could be regulated in the nucleus during the process of cell activation, which initiates and maintains NF-B-dependent transcription. Based on the susceptibility of post-translationally modified IB to the multicatalytic core activity of the 26 S proteasome and the well characterized existence of nuclear proteasome complex (32-35), we hypothesized that IB in the nuclei of activated cells could be exposed to proteolytic attack by the proteasome. Thus, in situ degradation of IB would prevent precocious and abrupt termination of ongoing NF-B-dependent gene expression.

Here we provide evidence that the proinflammatory cytokines TNF and IL-1 (interleukin-1) activate a permanent shuttling of IB from the cytoplasm to the nucleus. IB becomes detectable in the cell nucleus within minutes following cell induction. The amount of IB accumulating in the nucleus while cell signaling and NF-B activation persist is regulated in situ by an inducible proteasome-mediated degradation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Cell Culture, Transfection, and Luciferase Assay-- Human HeLa cells were maintained in Dulbecco's modified Eagle's medium containing 5% fetal calf serum (FCS). Cells were transiently transfected with CMV-IE-promoter-driven (pcDNA3) vectors (200 ng for 10⁴ cells) expressing either wild type (-TrCP) or mutant forms (-TrCPF) of -TrCP. HeLa 57A cells stably co-transfected with Rous sarcoma virus--galactosidase and 3EnhConA luciferase expression vectors were provided by R. Hay (University of St. Andrews, United Kingdom) and maintained in Dulbecco's modified Eagle's medium with 5% FCS and 200 µg/ml Geneticin (Life Technologies, Inc.). Luciferase activity contained in total extracts was measured using a luminometer (Lumat LB 9501, Berthold), and results were expressed as relative luciferase units per µg of protein, the background signal being subtracted from the values obtained from each sample. -Galactosidase measurement on total cellular extracts was performed using a kit according to the instructions provided by the manufacturer (Roche Molecular Biochemicals).

Immunofluorescence Microscopy-- HeLa cells (1 × 10⁵) were seeded on glass coverslips in 24-well plates and stained for indirect immunofluorescence 36-40 h later. They were fixed for 10 min with 4% paraformaldehyde phosphate-buffered saline and permeabilized with 0.1% Triton X-100 for 10 min. Anti-human IB or RelA rabbit polyclonal (C-21 and C-20, respectively (Santa Cruz Biotechnology, Inc., Santa Cruz, CA)) antibodies were applied for 1 h followed by a 1-h incubation with fluorescein isothiocyanate-conjugated goat anti-rabbit IgG (Southern Biotechnology). In some experiments, nuclei were counterstained by a 1-h incubation in propidium iodide and RNase A (Amersham Pharmacia Biotech). Coverslips were mounted in Mowiol (Hoechst). Confocal laser microscopy analysis was performed on a Leica TCS4D instrument with the 488-nm and 568 laser wave lines to excite fluorescein isothiocyanate and propidium iodide dyes, respectively, using a × 63 oil immersion PL APO objective. All of the data were recorded at the same laser and multiplier settings.

Western Blot, Electrophoretic Mobility Shift Assay, and Immunoprecipitation Analysis-- Cytoplasmic extracts were prepared as described (36) by lysis of cells in hypotonic buffer containing protease inhibitors and a mixture of phosphatase inhibitors. Nuclear extracts were prepared by plasma membrane permeabilization in hypotonic buffer containing 40 µg/ml digitonin, 25% glycerol, 1 mM phenylmethylsulfonyl fluoride, 7 mM -mercaptoethanol, 10 mM HEPES, pH 8, 50 mM NaCl, 1 mM EDTA. Nuclei were then lysed in hypertonic buffer (15) containing 1% Triton X-100. Cytoplasmic and nuclear extracts were analyzed by Western blotting performed as described (36) using anti-human IB monoclonal (MAD3 10B) (37) or anti-human IB (C-21 or C-15, Santa Cruz Biotechnology), anti-human RelA (C-20), anti-human Sp1 (sc-059; Santa Cruz Biotechnology), anti--tubulin (Amersham Pharmacia Biotech), and anti-human LDH (K90153S, Interchim) polyclonal antibodies. Immobilized antigen-antibody complexes were detected with secondary anti-IgG-horseradish peroxidase conjugates (Pierce) and an enhanced chemiluminescence detection system (Pierce). Lactate dehydrogenase activity was estimated in freshly collected cytosolic and nuclear fractions using a kit (DG 1340, Sigma) according to instructions provided by the manufacturer. Electrophoretic mobility shift assay was performed as described (15) with 4 µg of nuclear extracts incubated for 15 min at room temperature with a [-³²P]ATP-labeled, double-stranded oligonucleotide containing the human immunodeficiency virus type 1 LTR binding site for NF-B. For immunoprecipitation, magnetic beads (Dynal, M-280) were coupled with anti-human IB (C-15) or anti-human RelA (C-20) and were then incubated overnight at 4 °C in the presence of cytoplasmic or nuclear extracts. The immunoprecipitated proteins were analyzed by Western blotting performed as described above.

Materials-- The proteasome inhibitor carbobenzoxyl-leucin-leucin-leucinal (Z-LLL-H) was provided by F. Baleux (Unité de Chimie Organique, Institut Pasteur, Paris). Human recombinant TNF was provided by the Medical Research Council AIDS reagent program. Human recombinant IL-1 was a gift from Dr. A. Allison. Digitonin was purchased from Fluka. Leptomycin B was a kind gift of B. Wolff (Novartis, Vienna, Austria).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

IB Enters the Cell Nucleus While Cell Signaling Persists-- The kinetics of endogenous IB and NF-B proteins expression were analyzed in HeLa cells following stimulation with the proinflammatory cytokine TNF. Immunofluorescence microscopy revealed that within the first minutes following TNF treatment, a substantial reduction of the cytoplasmic amount of IB accompanied by the localization of cytoplasmic RelA to the cell nucleus is observed (Fig. 1A, TNF 5'). Concomitantly, and paralleling the accumulation of RelA, an increase in the nuclear content of p50 occurred, which suggests the presence of bona fide NF-B in the cell nucleus of TNF-activated cells (data not shown).

View larger version (48K): [in this window] [in a new window]   Fig. 1.   IB enters the cell nucleus while cell signaling persists. The kinetics and the fate of endogenous IB and RelA were determined by indirect immunofluorescence microscopy (A, B, and D) with anti-IB and RelA sera or by immunoblot analysis after cell fractionation (C). A, cells were left unstimulated or were stimulated with TNF (10 ng/ml) for the indicated times. A phase-contrast microscopy view of the corresponding field is shown for TNF 30'. B, cells were left unstimulated or were stimulated with TNF for 30 or 40 min. Following a 30-min induction with TNF, Z-LLL-H (100 µM) was added for 10 min (TNF 30' + Z-LLL-H 10'). In one case, nuclei were counterstained with propidium iodide (TNF 40'). C, cells were stimulated with TNF for 30 min followed by a 90-min chase (TNF (Chase)). In some cases (+), 10 min prior to subcellular fractionation, z-LLL-H or IL-1 (5 ng/ml) alone or together were added to the culture, and nuclear and cytosolic extracts (40 µg) were analyzed by Western blotting with antibodies specific for IB, RelA, Sp1, and lactate dehydrogenase. The positions of IB and phosphorylated IB (IBP) are indicated. IB detection in nuclear extracts corresponds to a longer exposure compared with cytosolic extracts. RelA detection in nuclear and cytosolic fractions from the aforementioned cells corresponds to the same exposure. The amount of nuclear protein load was assessed by detection of the transcription factor Sp1. To evaluate the presence of cytoplasmic proteins in the nuclear fractions, the content of cytoplasmic enzyme lactate dehydrogenase was assessed by immunoblot. Detection of lactate dehydrogenase in nuclear fractions was estimated by enzymatic assays and found to be routinely 2% of the cytoplasmic levels (data not shown). Lactate dehydrogenase and Sp1 detection in nuclear and cytosolic fractions correspond to the same exposure. D, cells were stimulated as described in C (panels B, C, D, and E), and immunofluorescence analysis was performed. Z-LLL-H or IL-1 alone (panels C and D, respectively) or both (panel E) were added to the culture 10 min prior to fixation of the cells.

Replenishment of the cytoplasmic pool of IB by newly synthesized protein becomes detectable 20 min after TNF induction (Fig. 1A, TNF 20'). Remarkably, by 30 min after the onset of the induction, a shift in the pattern of IB subcellular distribution was observed. At that time and while RelA was abundantly expressed in the cell nucleus, the neosynthesized IB showed apparently a preferential localization to the nucleus (Fig. 1A, TNF 30').

Neither induced IB phosphorylation or ubiquitination permit dissociation of the inhibitor integrated in NF-B·IB complex (14). Thus, it is conceivable that IB trafficking to the cell nucleus arises from a massively resynthesized, NF-B-unbound pool of the protein. This and the reported decreased activity of IB kinase complexes, which declines progressively in TNF time course experiments (38) along with the IB kinase complexes' preference for IB as a substrate when bound to NF-B (39), may explain that a pool of freed IB escapes degradation in the cytoplasm and enters the cell nucleus.

Degradation of IB Accumulating in the Cell Nucleus Shows Similar Characteristics to the Cytoplasmic Proteolysis of the Inhibitor-- The presence of IB in the cell nucleus 30 min after the addition of TNF should counteract the capacity of NF-B to bind specific DNA consensus sequences and, consequently, NF-B-dependent transcription. However, it is well established that a functional NF-B-DNA binding activity is generally reached by 30 min following TNF induction (see Fig. 4A, lane 3) by the time IB localization to the cell nucleus becomes apparent. How, under those circumstances, are NF-B-DNA binding and NF-B-dependent transcription protected from a precocious blockade by nuclear IB? Preservation of NF-B-DNA binding activity could be explained if only low amounts of IB enter and accumulate in the nucleus at that time. Alternatively, IB could translocate continuously and abundantly to the nucleus of activated cells, but the levels of the protein in this compartment could be actively down-regulated, thus preventing blockade of NF-B-dependent functions. Two mechanisms capable of reducing the content of IB must be considered: export from the nucleus to the cytoplasm and in situ nuclear proteolysis of IB.

A reduced stability of IB in the cell nucleus could be the consequence of the proteolytic activity of the proteasome. Indeed, the nuclear localization of the 26 S proteasome is well established, and its catalytic activity has been shown to participate in the regulation of nuclear transcriptional activators (40-43). We hypothesized that if the proteosome was involved in the degradation of nuclear IB, incubation with pseudopeptides that inhibit the multicatalytic activity of the proteasome complex (44) must stabilize the protein.

To this purpose, subcellular fractionating and biochemical analysis of the fate of IB both in the cytoplasm and nucleus were carried out. It should be stressed that in contrast with the intense nuclear IB staining observed in situ by immunofluorescence, the amount of nuclear IB detected upon subcellular fractionating is relatively low and is probably underestimated as a result of the loss of a significant fraction during preparation of nuclear and cytoplasmic fractions (15).

Cells were induced with TNF for 30 min and maintained for an additional 10 min in the presence or absence of the proteasome inhibitor Z-LLL-H (Fig. 1B, TNF 30', TNF 30' + Z-LLL-H, and TNF 40', respectively). The short treatment with Z-LLL-H led to a substantial increase of the cytoplasmic pool of IB and an even more marked effect on the accumulation of the protein in the cell nucleus (Fig. 1B). The accumulation of IB in the cell nucleus promoted by Z-LLL-H can be explained by the inhibition of an ongoing proteolytic mechanism activated by TNF, which is simultaneously operating in both the cytoplasmic and nuclear compartments. To further characterize the degradation mechanisms of IB in the nucleus of activated cells, we investigated the presence of phosphorylated forms of the inhibitor, which is a prerequisite for its subsequent degradation by the proteasome.

Furthermore, we sought to investigate whether a pool of nuclear IB may be phosphorylated in situ. To this aim, we used an experimental approach described previously, which permits induction and long lasting accumulation of IB in the cell nucleus leading to termination of NF-B-dependent transcription. Briefly, cells were treated for a short time by TNF (30-min pulse). Upon withdrawal of the stimulus, cells were cultured for an additional period of 90 min (chase), and thereafter nuclear proteins were obtained by subcellular fractionating and analyzed by immunoblotting. The fate of IB was analyzed in parallel by indirect immunofluorescence.

The mechanisms propitiating the constitution of an apparent resident pool of IB in the cell nucleus under those experimental conditions (Fig. 1, C, lane 5 and D, panel B) could be explained at least in part by the above discussed, time-dependent decline of the phosphorylation capacity of IB kinase complex in the course of TNF stimulation (38), a process that could be specially marked upon the extinction of cell signaling. Analysis of proteins from pulse-chase TNF-induced cells revealed, as expected, a massive accumulation of the nuclear content of IB, where the native form of the protein is predominant and co-exists with the slowly migrating form characteristic of the Ser³²-Ser³⁶-phosphorylated species (Fig. 1C, lane 5). Recognition of the slowly migrating form of IB by a polyclonal antibody that specifically recognizes IB phosphorylated on Ser³² further characterized the modification undergone by the protein (data not shown). The absence of detectable lactate dehydrogenase enzyme in the nuclear compartment excludes significant contamination of nuclear cell extracts by residual cytoplasmic proteins.

In the presence of Z-LLL-H, the amount of phosphorylated IB increased, thus reflecting the stabilization of the protein in the cell nucleus (Fig. 1C, compare lanes 5 and 6). The reinforced staining of the nuclear pool of IB parallels that finding and sustains that assumption (Fig. 1D, panel C). However, from this finding, it cannot be concluded whether the accumulation of the phosphorylated form of IB results from arrival of cytoplasmic-phosphorylated IB or from in situ modification of the nuclear pool. In this regard, induction with IL-1 of pulse-chase TNF-induced cells proved to be more informative. Indeed, IL-1 stimulation for 10 min efficiently promoted activation of NF-B as proved by a substantial enhancement of the nuclear content of RelA (Fig. 1C, -RelA, lane 7). Concomitantly, IL-1 induction lead to a profound reduction of the total content of IB (Fig. 1C, -IB, lane 7; Fig. 1D, panel D). If only IB phosphorylated in the cytoplasm was the target of degradation induced by IL-1 in the cell nucleus, the pool of unmodified nuclear IB should remain unaffected. However, it should be noted that IL-1 treatment induced in the nucleus a profound decrease of the native, nonphosphorylated IB that co-existed with the detection of phosphorylated IB (Fig. 1C, -IB, lane 7). These findings are compatible with an in situ modification of the pre-existing nuclear pool of IB that would permit ulterior degradation by the proteasome. Supporting this assumption, the simultaneous addition of Z-LLL-H and IL-1 preserved the nuclear pool of IB (Fig. 1C, lane 8; Fig. 1D, panel E), promoting the preferential accumulation of stable, phosphorylated forms of the protein (Fig. 1C, lane 8). As expected from an in situ phosphorylation of IB, the stabilization and accumulation of phosphorylated IB was paralleled by a significant reduction of the unmodified pool of the protein (Fig. 1C, lane 8).

The putative phosphorylation of IB in the cell nucleus is intriguing, since so far, the IB kinases  and  appear to be mainly cytoplasmic. However, the kinases have the capacity to localize in the nucleus. Indeed, it has been shown that when overexpressed from a heterologous promoter, both kinases localize in both cytoplasmic and nuclear compartments (38). The recent cloning of a new IB kinase complex induced by inflammatory cytokines (45) and interacting with TANK (a TRAF-binding protein (TNF receptor-associated factor) (46)) opens the possibility that, either constitutively or upon cell activation, this or another new kinase from that rapidly growing family could account for the phosphorylation of IB in the cell nucleus.

We and others showed that the F-box-containing -TrCP protein integrated in a SCF complex (for Skp1-cullin-F-box protein) is the ubiquitin-E3 ligase that ultimately marks Ser³²-Ser³⁶-phosphorylated IB for proteolysis by the proteasome machinery (8-12). When cells are stimulated with NF-B activators, phosphorylated IB is recruited to the SCF^-TrCP complex through its recognition by the first of the seven WD domains located at the C terminus of -TrCP. Deletion of the -TrCP F-box domain generates a mutant protein (-TrCPF) that is unable to bind to Skp1 and prevents ubiquitination of SCF^-TrCP substrates (reviewed in Ref. 47). However, -TrCPF retains intact its capacity to interact with the phosphorylated form of IB and therefore behave as a potent transdominant negative protein. Consequently, when overexpressed, -TrCPF can inhibit the metabolism of IB, thus permitting the stabilization of IB and the accumulation of the Ser³²-Ser³⁶-phosphorylated form. To further characterize the mechanism of IB degradation in the cell nucleus, -TrCPF was transiently expressed following cell transfection. This approach offers the invaluable advantage of permitting stabilization of IB while avoiding secondary effects on cell metabolism derived from the use of proteasome inhibitory compounds.

Fig. 2, A and B, shows data from two independent experiments. While the ectopically expressed wild-type -TrCP did not interfere with the TNF-induced degradation of IB, overexpression of -TrCPF reduced the induced proteolysis of IB and promoted a dramatic stabilization of the phosphorylated forms of the protein in both the cytoplasm and the nucleus of transfected cells (Fig. 2, A and B, lanes 7-9 and 10-12, respectively). However, simultaneous inhibition of the proteasome (Fig. 2, A and B, lanes 9 and 12) further stabilized IB, thus suggesting that some degree of NF-B activation, and probably of IB resynthesis and trafficking, escapes to the blockade imposed by -TrCPF upon induction by TNF. The phosphorylated form of IB could even be detected in the cytoplasm and the nucleus of unstimulated cells (Fig. 2, A and B, lanes 7 and 10, respectively), confirming that the -TrCPF mutant acts as a transdominant negative regulator of both constitutive and TNF-induced proteolysis of IB.

View larger version (38K): [in this window] [in a new window]   Fig. 2.   Degradation of IB accumulating in the cell nucleus shows similar characteristics to the cytoplasmic proteolysis of the inhibitor. HeLa cells maintained in 10% FCS were transiently transfected with -TrCP wild type (-TrCP) or its F-box-deleted counterpart (-TrCPF). After 24 h, cells were stimulated with TNF for 15 min in the presence or absence of Z-LLL-H. The amount of wild type and mutant -TrCP proteins was assessed by Western blot analysis to ensure that both proteins were expressed at comparable levels (data not shown). Shown are two representative experiments (A and B). Nuclear (60 and 30 µg for A and B, respectively) and cytosolic (30 and 20 µg for A and B, respectively) proteins were analyzed by Western blotting with anti-IB rabbit serum. IB detection in nuclear and cytosolic fractions correspond to the same exposure.

Whereas exogenous -TrCP has the capacity to localize to the nucleus (12),² the physiological cell distribution of the protein remains to be characterized. Interestingly, recent observations in mammalian cells have localized partly in the nucleus endogenous Skp1 and Cul1 (the human homologue of the yeast cullin Cdc53p), both being components of the ubiquitin ligase complex (48).

Degradation of IB in the Cell Nucleus Protects NF-B-dependent Transcription from Precocious Termination-- The fate of NF-B-dependent transcription under conditions where IB was massively accumulated in the cell nucleus was investigated. To this purpose, we developed an approach aimed at promoting trapping of IB in the cell nucleus after blocking the nuclear export of the protein. Thus, preventing the retrograde transport to the cytoplasm allowed us to investigate whether NF-B induced by an extracellular signaling could promote gene transcription despite the presence of nuclear IB to whom it could complex in an inactive form.

In unstimulated HeLa cells, IB shows an apparently constitutive cytoplasmic-nuclear trafficking. This bidirectional trafficking across the nuclear membrane reflects some degree of basal cell activation that largely depends on the particular conditions under which the experiments are carried out (i.e. FCS concentration in the culture medium). In agreement with this interpretation, the addition of leptomycin B (LMB), a Streptomyces metabolite that binds CRM1 and inhibits nuclear export of NES-carrying proteins at nanomolar concentrations (27, 49-52) permitted both the accumulation of IB and RelA in the cell nucleus (Fig. 3, A and B). The simultaneous presence of IB and RelA in the cell nucleus, promoted by a 40-min treatment with LMB, led to the formation of an IB-RelA complex that could be detected by immunoprecipitation (Fig. 3B, lane 8). This interaction was shown to be slightly lower compared with that observed in the cytosolic fraction (Fig. 3B, lane 6), a difference consistent with the relative levels of RelA in both compartments as assessed by direct immunoblot analysis (Fig. 3B, lanes 9-12). These results reveal that LMB induced the nuclear accumulation of RelA-bound IB and suggest a continuous shuttling of the complex between the cytoplasm and the nucleus.

View larger version (82K): [in this window] [in a new window]   Fig. 3.   Simultaneous presence of IB and RelA in nuclei of LMB-treated cells. HeLa cells maintained in 0.5% FCS were left untreated or exposed to LMB (20 nM) for 40 min prior to fixation of the cells for indirect immunofluorescence analysis (A) or cell fractionation for immunoprecipitation analysis (B). A, in LMB-treated cells, phase-contrast microscopy views of corresponding fields are shown. B, immunoblot analysis was used to detect RelA in the complexes precipitated with an anti-IB antibody (IP-IB) from cytosol and nuclei of HeLa cells (lanes 5-8) treated with LMB (+) or untreated. Direct immunoblot of cytoplasmic and nuclear fractions using anti-IB (lanes 1-4) and anti-RelA (lanes 9-12) antibodies was performed on the same cell extracts.

Upon induction with TNF the nuclear content of IB could be enhanced by the combined action of an intensified nuclear import of newly synthesized protein and the accumulation promoted by LMB. As a consequence, the NF-B-dependent transcription set in motion by TNF could be severely impaired or even abrogated. However, combined analysis of NF-B-DNA binding activity and NF-B-dependent gene expression demonstrated that the functional activity of NF-B is not affected by the simultaneous accumulation of IB promoted by LMB (Fig. 4, A and B). Indeed, exposure to TNF induced NF-B activation (Fig. 4A, lane 7) and NF-B-dependent expression of the luciferase reporter gene regardless of LMB pretreatment (Fig. 4B). Control expression of the -galactosidase gene placed under the control of an NF-B-independent promoter remained unchanged by those various treatments (data not shown). In conclusion, NF-B-dependent transcription is initiated and prolonged in cells despite nuclear import of IB.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   NF-B activation in LMB-treated cells. HeLa 57A cells stably transfected with a 3EnhConA luciferase vector were left untreated or were treated with LMB (20 nM) or Z-LLL-H (50 µM) for 40 min. A, half of the cultures were stimulated for 30 min with TNF before cell fractionation and electrophoretic mobility shift assay were performed on nuclear extracts. Specific complexes (NF-B) and free probe are indicated by arrows. The competition experiment indicated by a triangle was performed by adding a 40-fold molar excess of unlabeled B motif oligonucleotide (lane 6). B, half of each culture either untreated or treated with LMB was then stimulated with TNF. Luciferase activity measured at the indicated times of stimulation with TNF (hours) represents the mean of triplicates and is expressed as relative luciferase units (rlu)/µg of protein.

To explain this apparent paradox, we postulated that the pool of IB accumulated by exposure to LMB was degraded in the cell nucleus as a consequence of ongoing stimulation by TNF. Immunoblotting of subcellular protein fractions provided the clue to explain the fate of IB in the nuclear compartment. Indeed, we show that IB accumulated in nuclei from cells treated by LMB (concentrations ranging from 2 up to 200 nM) (Fig. 5A, lanes 9-12) is susceptible to proteolytic attack (Fig. 5A, lanes 13-16). We noticed that the susceptibility to degradation is only partial, since, unexpectedly, the presence of LMB affected the signaling initiated by either IL-1 (not shown) or TNF, leading to phosphorylation of the IB protein (Fig. 4A, compare lanes 3 and 7; Fig. 5C, compare lanes 4 and 5) and the activation of NF-B (Fig. 5 B, upper panel, lane 8). However, TNF-dependent NF-B-DNA binding and RelA accumulation were observed in the nuclei of activated cells (Fig. 5B, lanes 5-8, upper and lower panel, respectively). Thus, despite some interference of LMB in the TNF signaling pathway, the pool of IB accumulated in cells exposed to 20 nM LMB was severely reduced following induction with TNF, and a weak but still significant reduction could be observed at 200 nM LMB (Fig. 5A, short exposure, compare lanes 12 and 16). Thus, under experimental conditions ensuring specific blockade of IB export, the content of the nuclear pool of the protein was significantly down-regulated by cell signaling, which simultaneously induced degradation of cytoplasmic IB. Proteolysis of nuclear IB is likely to be a physiological mechanism preventing abrupt and precocious termination of NF-B-dependent transcription while cell signaling persists.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Signal-dependent degradation of IB in nuclei of LMB-treated cells. A, HeLa cells were pretreated with increasing amounts of LMB (0 nM (lanes 1, 5, 9, and 13), 2 nM (lanes 2, 6, 10, and 14), 20 nM (lanes 3, 7, 11, and 15), 200 nM (lanes 4, 8, 12, and 16)) for 40 min and then stimulated (+) or not () with TNF for 20 min prior to harvesting. Proteins (60 µg) from nuclear and cytoplasmic extracts were analyzed by Western blotting with anti-IB rabbit antibodies. IB detection in nuclear and cytosolic fractions corresponds to the same exposure. B, upper panel, electrophoretic mobility shift assay was performed on nuclear extracts from the aforementioned cells. Specific NF-B complexes are indicated by a bracket. The competition experiment is indicated by a triangle. Lower panel, proteins from nuclear extracts were also analyzed by Western blotting with anti-RelA antibody. C, HeLa cells left untreated or pretreated with LMB (20 nM) and/or Z-LLL-H (50 µM) for 40 min. Half of the cultures were stimulated with TNF for 30 min prior to harvesting. Proteins (40 µg) from cytosolic fractions were analyzed by immunoblotting with anti-IB antibody.

Our findings are opposed to those from a recent report concluding that IB accumulated upon LMB treatment in the cell nucleus represents a stable pool resistant to cell signaling-induced degradation (53). The apparent discrepancy is certainly linked to the narrow range of concentration offered by the drug to obtain specific inhibition of export without interfering in the signaling leading to IB phosphorylation. Indeed, a careful monitoring of LMB by side effects must be performed, especially when cells are exposed to LMB for long periods of time (53). Supporting our analysis, a recent study published during completion of this manuscript (29) reported that LMB used at low concentrations (5 nM) does not interfere in the TNF-inducible degradation of IB. Based on that evidence, these authors also conclude that the pool of nuclear IB can be destabilized by cell signaling (29).

Conclusion-- The main observation in this study is that nuclear IB, known to be detectable transiently upon extinction of the activation stimulus, is in fact entering the nucleus steadily in cells continuously exposed to stimulation but is constantly degraded in the nuclear compartment as long as stimulation persists. Our findings reveal the existence of two dynamically related mechanisms finely tuning the transcriptional activity of NF-B into the nucleus of mammalian cells. One permits nuclear NF-B to remain transcriptionally active as long as stimulation is ongoing, and it results from proteasome-mediated degradation of nuclear IB, thus suppressing the termination properties of this inhibitor. The second, intervening later when NF-B activity is no longer needed, results from retrograde transport of NF-B proteins to the cytoplasm by nuclear IB molecules whose destruction is stopped when stimulation is finished, thus liberating the nucleus from then unwanted NF-B molecules. These two mechanisms would thus successively act to optimize the efficiency and the timing of NF-B-dependent gene transcription, adapting the latter to cell activation or rest, death, or survival.

	ACKNOWLEDGEMENTS

We are indebted to E. Perret for confocal microscopy and to Dr. S. Michelson and Dr. J. Matthews for helpful and critical reading of the manuscript. We thank B. Wolff for the kind gift of LMB reagent and R. Hay for the gift of HeLa 57A cells.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

This work was supported in part by grants from the Agence Nationale de la Recherche sur le Sida (ANRS, France), the Association pour la Recherche sur le Cancer (France), and the European Communities Concerted Action BIOMED II (ROCIO Project).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Research assistant of the Fonds National de la Recherche Scientifique (Belgium). Present address: Unité de Biologie et Biochimie Cellulaire, Facultés Universitaires Notre-Dame de la Paix, 61 rue de Bruxelles. B-5000 Namur, Belgique.

^§ These authors contributed equally to this work.

^¶ Supported by a doctoral training grant from ANRS.

To whom correspondence should be addressed. Tel.: 33 1 40 61 34 67; Fax: 33 1 45 68 89 41; E-mail: fbachele@pasteur.fr.

² M. Kroll, unpublished results.

	ABBREVIATIONS

The abbreviations used are: NES, nuclear export sequence; FCS, fetal calf serum; Z-LLL-H, carbobenzoxyl-leucin-leucin-leucinal; IL, interleukin; TNF, tumor necrosis factor; LMB, leptomycin B.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES