Nuclear Retention of IκBα Protects It from Signal-induced Degradation and Inhibits Nuclear Factor κB Transcriptional Activation*

Transcriptional activation of nuclear factor κB (NF-κB) is mediated by signal-induced phosphorylation and degradation of its inhibitor, IκBα. However, NF-κB activation induces rapid resynthesis of IκBα, which is responsible for post-induction repression of transcription. Newly synthesized IκBα translocates to the nucleus, where it dissociates NF-κB from DNA and transports NF-κB from the nucleus to the cytoplasm in a nuclear export sequence-dependent process that is sensitive to leptomycin B (LMB). In the present study, LMB was used as a tool to inhibit nuclear export sequence-mediated nuclear protein export and evaluate the consequences for regulation of NF-κB-dependent transcriptional activity. Pretreatment of cells with LMB inhibits NF-κB-dependent transcriptional activation mediated by interleukin 1β or tumor necrosis factor α. This is a consequence of the inhibition of signal-induced degradation of IκBα. Although LMB treatment does not affect the signal transduction pathway leading to IκBα degradation, it blocks IκBα nuclear export. IκBα is thus accumulated in the nucleus, and in this compartment it is resistant to signal-induced degradation. These results indicate that the signal-induced degradation of IκBα is mainly, if not exclusively, a cytoplasmic process. An efficient nuclear export of IκBα is therefore essential for maintaining a low level of IκBα in the nucleus and allowing NF-κB to be transcriptionally active upon cell stimulation.

The NF-B 1 /Rel family of transcription factors is implicated in regulation of the expression of a number of cellular genes involved in immune responses, inflammation ,and apoptosis (for recent reviews, see Refs. [1][2][3]. In vertebrates, the NF-B family of proteins is composed of transcriptionally active p65/ Rel A (4,5), c-Rel (6), or Rel B (7) and transcriptionally silent p50/NF-B1 (8,9) or p52/NF-B2 (10 -12). All NF-B proteins share a conserved region known as the Rel homology domain that contains the nuclear localization signal as well as the dimerization and DNA binding functions. The NF-B form activated by extracellular signals is composed of p50 and p65. NF-B transcriptional activity is controlled by inhibitor IB proteins that contain ankyrin repeat domains. Association of p50/p65 with IB not only occludes the nuclear localization sequence of p50 and p65, leading to cytoplasmic sequestration, but also prevents NF-B DNA binding activity. Several IBs have been described including IB␣ (13), IB␤ (14), IB⑀ (15), and Bcl-3 (16). Additionally, the precursors of p50 (p105) and p52 (p100) possess inhibitory ankyrin repeat domains that in isolation are known as IB␥ (17)(18)(19) and IB␦ (20,21), respectively.
Newly synthesized IB␣ is accumulated in the cytoplasm but also in the nucleus, where it terminates NF-B-dependent transcription. This is accomplished by inhibition of the NF-B/ DNA interaction and export of NF-B back to the cytoplasm (38). The latter function of IB␣ is conferred by a leucine-rich nuclear export sequence (NES) present in its carboxyl-terminal region (aa 265-277) (39) and homologous to the NES found in many proteins including the human immunodeficiency virustype 1 Rev protein and the protein kinase A inhibitor (40 -42). Such NESs constitute transferable transport signals that are necessary and sufficient to mediate rapid and active export from the nucleus to the cytoplasm. The nuclear protein CRM1 (also known as exportin 1) has been recently identified as the NES receptor (43)(44)(45)(46). CRM1 belongs to the karyopherin ␤ family and, in particular, it shares sequence homology in the Ran-GTP binding domain with members from this family (47). The formation of CRM1/NES complex is facilitated by the presence of Ran in its GTP-bound form. It has been proposed that this ternary complex is transported through the nuclear pore complex and dissociates in the cytoplasm due to GTP hydrolysis by Ran-GAP (43). In addition, CRM1 has been shown to be the cellular target of the drug leptomycin B (LMB) that inhibits NES-mediated protein export both in vivo and in vitro (43)(44)(45)48).
In the present study, LMB was used as a tool to inhibit NES-mediated nuclear protein export and evaluate the consequences for regulation of NF-B-dependent transcriptional activity. Pretreatment of cells with LMB inhibits NF-B-dependent transcriptional activation mediated by IL-1␤ or TNF␣. This is a consequence of the inhibition of signal-induced degradation of IB␣. Although LMB treatment does not affect the signal transduction pathway leading to IB␣ degradation, it blocks IB␣ nuclear export. IB␣ is thus accumulated in the nucleus, and in this compartment it is resistant to signal-induced degradation. These results indicate that the signal-induced degradation of IB␣ is mainly, if not exclusively, a cytoplasmic process. An efficient nuclear export of IB␣ is therefore essential for maintaining a low level of IB␣ in the nucleus and allowing NF-B to be transcriptionally active upon cell stimulation.

EXPERIMENTAL PROCEDURES
Reagents-Leptomycin B was a gift from B. Wolff-Winiski (Novartis) and was used at 20 nM. TNF␣, which was obtained from the MRC Reagent Project, and IL-1␤ (Sigma) were used at 10 ng/ml for the indicated time. Cycloheximide (Sigma) was used at 100 g/ml. Z-LLL-H (MG132) was a gift from F. Baleux (Institut Pasteur) and was used at 20 M.
Plasmid Construction-The pSVB (CLONTECH) and pRC/RSV (Invitrogen) vectors were digested with NotI. Corresponding DNA fragments were ligated to obtain the pRC/RSV reporter plasmid in which ␤-galactosidase transcription is driven by the RSV promoter.
Cell Culture and Transfections-HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Before treatment with IL-1␤ or TNF␣, the medium was changed to Dulbecco's modified Eagle medium without serum. HeLa cells were transfected by electroporation as described previously (39). A total of 10 g of plasmid DNA encoding the chimeric protein 4NBC (24) was transfected in 5 ϫ 10 6 HeLa cells. After transfection, cells were seeded in four wells of 6-well plates, and incubation continued for 24 h.
The expression of both the 3Enhancer-kB-conA-Luc (49) and the pRC/RSV-␤-galalactosidase plasmids was stabilized in HeLa cells using neomycin selection. Single cell clones were obtained by limiting dilution of the neomycin-resistant cells. The 57A cell line was selected on the basis of TNF␣-induced luciferase activity and constitutive ␤-galactosidase activity.
Western Blot Analysis-Cells grown in 6-well plates were harvested in 150 l of lysis buffer (50) for Western blot analysis. 15 l of whole cell extracts were resolved in 10% SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidine difluoride membranes (Sigma) by electroblotting, and processed for Western blotting as described previously (38). When indicated, membranes were stripped and processed as reported previously (29). Primary polyclonal antibodies anti-IB␣ or anti-p65 (C-21 and C-20) were from Santa Cruz Biotechnology. The previously described anti-SV5 tag monoclonal antibody (51) to detect the chimeric 4NBC protein was obtained from Dr. R. E. Randall. Monoclonal anti-␣-tubulin and horseradish peroxidase-labeled anti-rabbit or anti-mouse antibodies were purchased from Amersham. An enhanced chemiluminescence detection system was used to detect specific antigen-antibody interactions (POD; Boehringer Mannheim).
Cell Fractionation and Immunoprecipitation-Cells were washed twice with Dulbecco's modified Eagle medium and once with Buffer S (115 mM potassium acetate, pH 7.3, 25 mM Hepes, pH 7.4, and 2.5 mM MgCl 2 ) at 37°C and then treated with 2 g/ml streptolysin O (SLO) (52) in Buffer S for 3 min at 37°C. SLO supernatant (cytosolic extracts) was kept on ice, cells were washed once with Buffer S, and the resulting washing volume was mixed with SLO supernatant (cytosolic extracts). Cells were then washed three times with Buffer S and solubilized in Buffer T (30 mM Tris, pH 8.6, 150 mM NaCl, 2 mM EDTA, and 2% Triton X-100). This extract corresponds to the nuclear fraction. The quality of fractionation was controlled by Western blotting using monoclonal antibodies against heterogeneous nuclear ribonucleoprotein C (nuclear protein) and ␥-adaptin (cytosolic protein). Nuclear and cytosolic extracts were incubated for 15 min at 4°C and centrifuged at 10,000 ϫ g for 10 min. Anti-p65 polyclonal antibodies and protein G-agarose beads were added to the supernatants and incubated for 4 h at 4°C. Beads were then washed, boiled for 5 min in Laemmli sample buffer, and analyzed by 10% SDS-polyacrylamide gel electrophoresis and Western blotting with anti-IB␣ polyclonal antibodies.
Indirect Immunofluorescence Analysis-For indirect immunofluores-cence analysis, HeLa cells grown on coverslips were fixed with 3% paraformaldehyde and permeabilized with 0.1% Triton X-100 for 10 min. Monoclonal antibodies to IB␣ (10B) and polyclonal antibodies to NF-B p65 (C-20; Santa Cruz Biotechnology) were applied for 30 min, followed by a 30-min incubation with fluorescein isothiocyanate or Texas Red (TR)-conjugated donkey anti-mouse or anti-rabbit IgG (Jackson). Coverslips were mounted in Mowiol (Hoechst, Frankfurt, Germany). Confocal laser scanning microscopy and immunofluorescence analysis were performed with a TCS4D confocal microscope based on a DM microscope interfaced with a mixed-gas argon-krypton laser (Leica Laser Technik). Fluorescence acquisitions were performed with the 488 nm and 568 nm laser lines to excite fluorescein isothiocyanate and Texas Red dyes, respectively, with a ϫ100 oil immersion PL APO objective. Data presented on the same figure were registered at the same laser and multipliers settings. Measurement of Luciferase and ␤-Galactosidase Activities-75 ϫ 10 3 HeLa 57A cells for each time point were stimulated with 10 ng/ml TNF␣ or 10 ng/ml IL-1␤ as described previously and incubated for an additional 7 h. Cells were lysed as reported previously (38). ␤-Galactosidase activity was measured using the Galacto-Light Plus kit (Tropix) according to the manufacturer's instructions. Both ␤-galactosidase and luciferase activities were measured in a bioluminometer (Berthold). Values for fold activation of the luciferase reporter are the average of four separate determinations and are compared with the uninduced value. In each case, ␤-galactosidase activity was used as an internal control.

Transcriptional Activation of NF-B Is Inhibited by Leptomycin B-
It has been previously reported that IB␣ that is newly synthesized in response to NF-B activation translocates to the nucleus, where it dissociates NF-B from DNA and transports NF-B from the nucleus to the cytoplasm in a NESdependent process that is sensitive to LMB. To explore the role of nuclear export in the regulation of NF-B-dependent transcriptional activity, LMB was used to inhibit NES-mediated nuclear protein export. To monitor transcriptional activity, a cloned HeLa cell line (57A) was derived that contains NF-Bdependent luciferase and NF-B-independent ␤-galactosidase reporters stably integrated into the genome. To activate NF-B-dependent transcription, HeLa 57A cells were stimulated with TNF␣ or IL-1␤ for 7 h ( Fig. 1; TNF or IL-1 conditions) or for 30 or 40 min, respectively, and further incubated for 7 h at 37°C in the absence of stimulus ( Fig. 1; TNFϩchase or IL-1ϩchase). Unstimulated cells were exposed to control medium lacking activators ( Fig. 1; NS). After the indicated time, cells were lysed, and reporter activity was measured. Because ␤-galactosidase reporter activity does not change in response to TNF␣ or IL-1␤, this was used to normalize the NF-B-dependent luciferase activity.
Continuous exposure of cells to TNF␣ and IL-1␤ for 7 h induced a 196-and 358-fold stimulation, respectively, compared with the basal noninduced activity. After 30 min of TNF␣ or IL-1␤ incubation followed by a 7-h incubation in the absence of stimulus, the transcriptional activity was 71-and 213-fold activation, respectively. This results in a 64% and 41% reduction in activity compared with the continuously activated cells. In contrast, when cells were treated with 20 nM LMB for 30 min before cell stimulation, only 51-and 102-fold activation was obtained by continuous stimulation with TNF␣ and IL-1␤, respectively. LMB treatment thus led to a 74% and 72% inhibition of the NF-B transcriptional activity induced by TNF␣ or IL-1␤, respectively. Moreover, remaining NF-B activities were not modified when LMB-treated cells were transiently exposed to TNF␣ or IL-1␤ and incubation continued in the absence of stimulus. These results indicate that an inhibition of NESmediated nuclear protein export not only prevents the postinduction repression of NF-B-dependent transcription but also strongly represses the initial activation of NF-B upon cell stimulation. Similar results were obtained with an independent lung-derived cell clone (A549) that contains NF-B-depend-ent luciferase and NF-B-independent ␤-galactosidase reporters stably integrated into the genome (data not shown).
Leptomycin B Inhibits Signal-induced Degradation of IB␣-Activation of NF-B transcriptional activity is mediated by signal-induced degradation of IB␣, which allows the released NF-B to translocate to the nucleus. The effect of LMB on signal-induced degradation of IB␣ expression was therefore examined. Thus, untreated HeLa or LMB-pretreated cells were exposed to TNF␣ for 30 min or to IL-1␤ for 40 min (Fig. 2,  A and B). Whole cell extracts were analyzed by Western blotting using an anti-IB␣ antibody. As expected, IB␣ was rapidly degraded after TNF␣ or IL-1␤ treatments, but pretreatment of the cells with LMB substantially inhibited signalinduced degradation of IB␣. This effect was not restricted to HeLa cells because TNF␣-induced degradation of IB␣ was also inhibited in 293 or COS7 cells pretreated with LMB (Fig.  2, C and D). These data suggest that LMB inhibits NF-B transcriptional activity by reducing the signal-induced degradation of IB␣.
Nuclear Accumulation of IB␣ Mediated by Leptomycin B Protects It from Signal-induced Degradation-To understand how LMB, a drug that blocks NES-mediated nuclear protein export, inhibits IB␣ degradation, the subcellular localization of both IB␣ and NF-B p65 was analyzed by cell fractionation and biochemical analysis as well as indirect immunofluorescence. HeLa cells were either untreated or pretreated with LMB 30 min before stimulation with TNF␣. After 30 min, TNF␣ was removed, and the cells were either analyzed directly or incubated for an additional 60 min in the absence of TNF␣ (chase). IB␣ levels were determined by Western blotting of whole cell extracts. In the absence of LMB, TNF␣ induced IB␣ degradation, followed by resynthesis during the chase period (Fig. 3A, lanes 1-3). LMB treatment inhibited the TNF␣-induced reduction in IB␣ level but had no effect on the amount of IB␣ present after the 60-min chase period (Fig. 3A, lanes  4 -6).
To establish the subcellular localization of IB␣ and the p65 subunit of NF-B under these experimental conditions, cells were fractionated using Streptolysin O, a bacterial toxin that permeabilizes cells without affecting the integrity of the nuclear envelope (52)(53)(54). Cytoplasmic and nuclear extracts were analyzed either by Western blotting with an anti-p65 antibody or by immunoprecipitation with an anti-p65 antibody and Western blotting with an anti-IB␣ antibody (Fig. 3B). In parallel, intact cells treated under the same experimental conditions were processed for immunofluorescence and analyzed using anti-p65 and anti-IB␣ antibodies (Fig. 3C).
In the absence of any cell stimulation, p65 was predominantly cytoplasmic and was found to interact with IB␣, whereas IB␣ was localized in both the cytoplasm and the nucleus (Fig. 3B, lane 1; Fig. 3C, left panel). Treatment with LMB led to the nuclear accumulation of IB␣ and p65, although to a lesser extent (Fig. 3B, lane 4; Fig. 3C, right panel).  3 and 6). Whole cell extracts were analyzed by Western blotting with a polyclonal antibody specific for IB␣ and a monoclonal antibody to ␣-tubulin. B, HeLa cells were treated as indicated in A and fractionated into cytoplasmic and nuclear fraction using SLO. Cytoplasmic and nuclear extracts were either analyzed directly by Western blotting with an anti-p65 antibody (upper panel) or immunoprecipitated with an anti-p65 polyclonal antibody before analysis by Western blotting with an IB␣ polyclonal antibody (lower panel). C, HeLa cells treated as indicated in A were processed for indirect immunofluorescence and double stained with a mouse monoclonal antibody to IB␣ and a rabbit polyclonal antibody to Rel A. Primary antibodies were detected with a fluorescein isothiocyanate-conjugated anti-mouse IgG and a Texas Red-conjugated anti-rabbit IgG. Cells were visualized by confocal laser scanning microscopy, and photographs correspond to the accumulation of four optical sections in one projection. This result suggests that IB␣ and a fraction of NF-B are continuously shuttling between the nucleus and the cytoplasm, even in the absence of cell stimulation. Upon TNF␣ stimulation of cells that were not treated with LMB, IB␣ was degraded in the cytoplasm with a small fraction still present in the nucleus, and p65 was partially translocated to the nucleus (Fig. 3B, lane  2; Fig. 3C, left panel). The addition of TNF␣ to LMB-treated cells led to the loss of IB␣ from the cytoplasm without affecting the nuclear content of IB␣. Thus, the remaining IB␣ after TNF␣ stimulation of LMB-treated cells was exclusively nuclear. Nuclear translocation of p65 was increased by LMB treatment but mainly resulted in the nuclear accumulation of a transcriptionally inactive IB␣-bound form of Rel A ( Fig. 1; Fig.  3B, lane 5; Fig. 3C, right panel). In cells that were not treated with LMB that had been exposed to TNF␣ but incubated for an additional 60 min in the absence of TNF␣, IB␣ returned to prestimulation levels as a result of the de novo synthesis of the protein. Binding of the newly synthesized IB␣ to p65 allowed the IB␣/p65 complexes to relocalize to the cytoplasm (Fig. 3B,  lane 3; Fig. 3C, left panel). Under identical conditions, IB␣ levels in LMB-pretreated cells returned to those observed before TNF␣ stimulation. However, LMB treatment inhibited the relocalization of both IB␣ and Rel A to the cytoplasm (Fig. 3B,  lane 6; Fig. 3C, right panel).
In conclusion, inhibition of NES-mediated nuclear protein export leads to the accumulation of IB␣ in the nucleus, where it is resistant to signal-induced degradation. Nuclear IB␣ can interact with NF-B and therefore prevent the DNA binding of the transcription factor. Moreover, LMB-mediated inhibition of nuclear export blocks the transport of NF-B/IB␣ complexes back to the cytoplasm. An efficient nuclear export of IB␣ is thus required to maintain a low level of IB␣ in the nucleus and allow NF-B to be efficiently activated upon cell stimulation.
Leptomycin B Does Not Increase IB␣ Synthesis-The data presented in Figs. 1-3 indicated that LMB inhibits signalinduced activation of NF-B by partitioning IB␣ in the nucleus, where it is resistant to degradation. However, it is a formal possibility that LMB has no effect on IB␣ degradation but rather stimulates the synthesis of IB␣. To distinguish between these possibilities, cells pretreated with LMB or untreated cells were exposed to TNF␣ in the presence of cycloheximide. The expression and subcellular localization of IB␣ and p65 were analyzed by Western blotting (Fig. 4A) and indirect immunofluorescence using anti-p65 and anti-IB␣ antibodies (Fig. 4B).
Simultaneous TNF␣ and CX treatments did not affect signalinduced IB␣ degradation and nuclear translocation of p65 (Fig. 4A, lane 2; Fig. 4B, left panel). However, CX abolished the de novo synthesis of IB␣ and relocalization of p65 to the cytoplasm that occurs after TNF␣ treatment and an additional 1-h incubation in the absence of TNF␣ (Fig. 4A lane 3; Fig. 4B,  left panel). These data confirm that the expression of newly synthesized IB␣ in the nucleus is responsible for the nuclear FIG. 4. LMB does not increases IB␣ synthesis. A, HeLa cells were treated as described in the Fig. 3 legend, but cycloheximide was added together with TNF␣ or for the last 30 min in unstimulated cells. Whole cell extracts were analyzed by Western blotting with a polyclonal antibody specific for IB␣ and a monoclonal antibody to ␣-tubulin. B, HeLa cells treated as indicated in A were processed for indirect immunofluorescence and double stained with a mouse monoclonal antibody to IB␣ and a rabbit polyclonal antibody specific for p65. Primary antibodies were detected with a fluorescein isothiocyanate-conjugated anti-mouse IgG and a Texas Red-conjugated anti-rabbit IgG. Cells were visualized by confocal laser scanning microscopy, and photographs correspond to the accumulation of four optical sections in one projection. export of NF-B to the cytoplasm. Treatment of cells with LMB and CX in the presence or absence of TNF␣ stimulation did not substantially modify the subcellular distribution of IB␣ and p65 observed previously in cells treated only with LMB (Fig.  4B, right panel). IB␣ was still detected in LMB-and CXtreated cells stimulated with TNF␣, and when the incubation was continued in the absence of TNF␣, the amount of IB␣ in LMB-and CX-treated cells was only slightly decreased (Fig.  4A, lanes 4 -6). Thus, LMB treatment does not induce IB␣ synthesis but rather protects IB␣ from signal-induced degradation as a consequence of its nuclear sequestration.
Nuclear IB␣ Is Not Accessible to Signal-induced Modification-To confirm that LMB treatment protects IB␣ from signal-induced degradation through its accumulation in the nucleus and not through an inhibition of the TNF␣ signal transduction cascade, a fusion protein containing the aminoand carboxyl-terminal regions of IB␣ fused to Escherichia coli ␤-galactosidase (24) was used. Because this protein contains the necessary sequence information from IB␣, it undergoes signal-induced degradation in response to TNF␣. However, because it does not contain the ankyrin repeats from IB␣, it does not translocate to the nucleus and is localized exclusively in the cytoplasm, even in the presence of LMB (data not shown). Thus, in LMB-treated cells transfected with a plasmid expressing 4NBC, the endogenous IB␣ will be located in the nucleus, whereas 4NBC will be located in the cytoplasm. HeLa cells were transiently transfected with the 4NBC-encoding plasmid, treated with LMB or control medium, and then exposed to TNF␣ and CX or control medium. After 60 min, the levels of 4NBC and endogenous IB␣ were determined by Western blotting. The combined action of TNF␣ and CX induced the degradation of both endogenous IB␣ and exogenous 4NBC, although, as expected, to varying degrees (Fig. 5A, IB␣, 90% degraded; 4NBC, 58% degraded). In the presence of LMB, the extent of TNF␣-and CX-induced degradation of 4NBC was unaltered (60% degraded), whereas IB␣ was inhibited (46% degraded, Fig. 5A), indicating that LMB does not directly affect the TNF␣ signal transduction cascade. Moreover, in vitro kinase assays using immunopurified IKK␣ and IKK␤ from HeLa cell extracts showed that the TNF␣-induced IKK activity was not affected by the LMB treatments (data not shown).
To determine the step of the signal-induced degradation of IB␣ that is prevented by the localization of IB␣ in the nucleus, untreated HeLa cells or HeLa cells treated with the proteasome inhibitor MG132 were exposed to TNF␣ or control medium for 15 min. This experimental condition allows the accumulation of a more slowly migrating phosphorylated form of IB␣ (Fig. 5B, lanes 1-4). To confirm that this form corresponded to phosphorylated IB␣, the same blot was stripped and reprobed with an antibody specifically recognizing IB␣ phosphorylated at Ser-32 (data not shown). In contrast, IB␣ degradation was not observed when cells were pretreated with LMB, and the TNF␣ plus MG132-mediated accumulation of the phosphorylated form of IB␣ was severely reduced (Fig. 5B,  lanes 5-8). These data indicate that nuclear IB␣ is not accessible to signal-induced phosphorylation.
However, this observation does not exclude the possibility that other downstream steps leading to signal-induced degradation of IB␣ could also be affected; thus, any effect of LMB on IB␣ ubiquitination was evaluated. HeLa cells were pretreated with the proteasome inhibitor MG132 and then treated with either LMB, TNF␣, or a combination of LMB and TNF␣. As expected, MG132 treatment prevented TNF␣-induced degradation and allowed the accumulation of slowly migrating multiubiquitinated forms of IB␣ (29) (Fig. 5C, lane 2). In the pres-ence of MG132, LMB, and TNF␣, the amount of ubiquitinated IB␣ was strongly reduced but not abolished (Fig. 5C, lane 4) and probably corresponds to modification of the remaining cytoplasmic IB␣. Thus, nuclear sequestration of IB␣ by LMB prevents its proper phosphorylation and subsequent ubiquitination after signal induction. The conditions described above did not alter the levels of the SUMO-1 modified form of IB␣ (50). DISCUSSION The experiments reported here indicate that inhibition of IB␣ nuclear export not only prevents the post-induction repression of NF-B-dependent transcription but also strongly represses the initial activation of NF-B upon cell stimulation.  3 and 4) for 60 min and further treated with TNF␣ plus CX (lanes 2 and 4) for 60 min. Total cell extracts were analyzed by Western blotting with the SV5 Pk tag monoclonal antibody and a polyclonal antibody specific for p65. The same membrane was stripped and reprobed with an anti-IB␣ antibody. Levels of IB␣ and 4NBC were determined by densitometry. B, HeLa cells were untreated (lanes 1 and 2) or pretreated for 60 min with MG132 (lanes 3 and 4), LMB (lanes 5 and 6), or both MG132 and LMB (lanes 7 and 8). After a 15-min incubation with TNF␣ (lanes 2, 4, 6, and 8), whole cell extracts were analyzed by Western blotting with a polyclonal antibody specific for IB␣ and a monoclonal antibody to ␣-tubulin. C, HeLa cells were pretreated with MG132 (lanes 1-4) and LMB (lanes 3 and 4) for 90 min and further incubated for 30 min with TNF␣ (lanes 2 and 4). Whole cell extracts were analyzed by Western blotting with a polyclonal antibody specific for IB␣ and a monoclonal antibody to ␣-tubulin. SUMO-1-and ubiquitin-conjugated forms of IB␣ are indicated (B). Indeed, nuclear IB␣ appears to be resistant to signal-induced phosphorylation and degradation, and this results in nuclear accumulation of transcriptionally inactive IB␣/NF-B complexes. Nuclear export of IB␣ and many other proteins that shuttle between the nucleus and the cytoplasm is mediated by a leucine-rich NES that is recognized by CRM1/exportin 1 (43)(44)(45)(46). In our experiments, nuclear export of IB␣ was inhibited by the drug LMB. LMB specifically targets CRM1 by blocking its interaction with the NES (45). Although the inhibition of CRM1 by LMB is highly specific, it was important to rule out the possibility that LMB might be interfering with the signal transduction pathways that lead to NF-B activation. To address this point, we used a lacZ fusion protein (4NBC) containing the amino and carboxyl termini of IB␣. When this protein is expressed in cells, it is unable to translocate to the nucleus but undergoes signal-induced degradation in response to agents such as TNF␣ and IL-1␤ (24). Signal-induced degradation of the 4NBC protein was unaffected by LMB (Fig. 5A), indicating that LMB does not inhibit the signal transduction pathway that leads to IB␣ degradation. It is also clear that LMB does not inhibit transcription in a nonspecific fashion because the activity of the integrated RSV-driven lacZ reporter was unaffected by the presence of LMB (Fig. 1). The conclusion from these experiments is that in HeLa cells, signal-induced phosphorylation and degradation of IB␣ occurs exclusively in the cytoplasm. One possibility to explain this restriction is that an essential component of the signal transduction pathway that leads to IB␣ phosphorylation cannot gain access to the nucleus. It appears that the IB␣ kinases (IKK␣ and IKK␤) are present in a large signaling complex that may contain upstream kinases such as NF-B inducing kinase (55,56) and scaffolding proteins such as NEMO (57). It has yet to be determined if this large complex can be imported into the nucleus. The alternative argument would be that the signal modification machinery has access to the nucleus, but that the nuclear IB␣ is in some way refractile to modification. Mechanisms to achieve this could include prior covalent modification of IB␣ to a form that is no longer recognized by the IKK signaling complex. It has recently been demonstrated that IB␣ is modified by the small ubiquitin-like protein SUMO-1 (50). Because this protein is linked to the same lysine residues that are used for ubiquitination of IB␣, this renders the SUMO-1 modified form of IB␣ resistant to signal-induced degradation. Although the known proteins that are substrates for SUMO-1 modification have been detected in the nucleus or are involved in nuclear transport (58), we have no evidence to support the notion that nuclear IB␣ detected in the presence of LMB is resistant to signal-induced phosphorylation because it is modified by SUMO-1. It is also possible that IB␣ could interact with a nuclear protein that occludes the region in IB␣ containing residues Ser-32 and Ser-36, thus protecting it from signalinduced phosphorylation.
In unstimulated cells, there is clearly a requirement for the transcription of essential NF-B-dependent genes. Low-level transcription of these genes does not take place simply as a consequence of NF-B-independent transcription, because IB␣ overexpression effectively abolishes the activity of a NF-B-dependent reporter in unstimulated cells. 2 Thus, it appears that the cell has evolved a highly dynamic system to provide for continued low-level transcription of NF-B-dependent genes. This homeostatic mechanism requires the continuous proteasome-mediated breakdown of IB␣, which generates a stream of free NF-B that can translocate to the nucleus. Once in the nucleus, NF-B activates NF-B-dependent genes, including that of IB␣. After transport to the cytoplasm, IB␣ mRNAs are translated, and the free IB␣ is directed to the nucleus, where it interacts with DNA-bound NF-B and dissociates the DNA-protein complex. By virtue of the presence of a NES in IB␣ (39), NF-B/IB␣ complexes are recognized by CRM1, which mediates nuclear export (45). At this point, equilibrium is reestablished. Thus, rather than having a simple on-off switch, the cell can delicately alter the NF-B transcriptional response by varying the rate at which IB␣ is turned over. The most extreme perturbation of this equilibrium comes after exposure of the cells to agents such as TNF␣ or IL-1␤. In this situation, cytoplasmic IB␣ is completely degraded, and a massive pulse of NF-B is released into the nucleus to initiate high-level transcription of NF-B-dependent genes. However, the same mechanism is used to bring the system back into homeostasis (38,39). A remarkably similar homeostatic mechanism seems to operate to control the level of p53 within the cell. In this case, the product of the hdm2 gene targets p53 for ubiquitin-mediated proteasomal degradation, and disruption of this interaction during the damage response leads to the accumulation of p53. Nuclear translocation of p53 activates transcription of the hdm2 gene, and the newly synthesized protein enters the nucleus, where it terminates p53-dependent transcriptional activation. hdm2 also contains a NES, and this is used to export the p53/hdm2 complex to the cytoplasm using the same pathway that is used for nuclear export of IB␣. Inhibition of hdm2-mediated export revealed that nuclear export of hdm2 is required to accelerate the degradation of p53 (59). In the case of both IB␣ and p53, ubiquitin-mediated proteasomal degradation occurs in the cytoplasm, even though proteasomes are found in both compartments. However, proteasomal components are distributed differentially between the nucleus and the cytoplasm (60,61), suggesting that nuclear and cytoplasmic proteasomes may have unique properties. The advantage to the cell of these homeostatic mechanisms to control NF-B-and p53-dependent transcription is that they are both highly sensitive to perturbation, and they can provide a finely tuned response to external signals.