INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The NF-B/Rel family of inducible transcription factors is involved in the expression of numerous genes involved in important cellular and physiological processes such as growth, development, apoptosis, and inflammatory and immune responses (15, 45). Members of the Rel family include p65 (RelA), p105/p50, p100/p52, RelB, and c-Rel. These transcription factors can form homo- or heterodimers with each other to make transcriptionally competent or repressive complexes, loosely referred to as the nuclear factor kappa B (NF-B). The biological activity of NF-B is tightly regulated by its inhibitor protein, IB. Members of the IB family include IB, IB, IB, Bcl-3, and the best-characterized member, IB (15). In most cells, IB and IB are found associated with the p50-p65 heterodimer, the most ubiquitous NF-B, to form a stable trimeric complex inside the cell.

The subcellular localization of NF-B-IB complexes dictates the ability of NF-B to be activated by extracellular stimuli such as tumor necrosis factor alpha (TNF-). We and others have previously shown that cytoplasmic localization of preinduced NF-B-IB complexes is important for efficient cytokine-dependent phosphorylation-ubiquitination and subsequent degradation of IB proteins, which cause the release of NF-B to the nucleus to alter gene expression (17, 38). Nuclear NF-B-IB complexes, however, are generally refractory to cytokine-induced IB degradation. These observations suggest that cytoplasmic localization of NF-B-IB complexes plays an important role during the pre- and postinduced stages of NF-B activation.

Localization of preinduced NF-B population is partly controlled by an N-terminal nuclear export signal (N-NES) on IB (17, 23, 38, 43). NF-B complexes formed with IB have a tendency to shuttle rapidly between the cytoplasm and nucleus, likely due to leaky exposure of p50 nuclear localization signal (NLS) coupled to a more dominant nuclear export by IB (17, 22). However, it is unknown whether IB or other IB members bound to NF-B can also shuttle nucleocytoplasmically in the absence of stimuli.

The localization of postinduced nuclear NF-B population is also carefully controlled, presumably by IB (1, 47). Postinduction repression refers to the condition in which activated nuclear NF-B upregulates the expression of IB due to NF-B consensus binding sites within the IB promoter (7, 8, 21, 27, 42), followed by nuclear accumulation of free IB, which dissociates NF-B from NF-B-bound DNA complexes to repress NF-B function (2). These newly formed nuclear NF-B-IB complexes are then exported out to the cytoplasm, thereby reestablishing the cytoplasmic pool of inactive NF-B complexes that are primed for another round of activation to take place (2). Recent reports have shown intrinsic nuclear export functions in both IB (2, 17, 23, 38, 39, 43) and the p65 subunit of NF-B complexes (16). However, it remains to be determined which of these newly characterized NESs can facilitate postinduction export of nuclear NF-B complexes.

The leucine-rich NES motif is an evolutionarily conserved sequence used by a variety of proteins to facilitate their delivery from the nucleus to the cytoplasm and is also used as an important point of control by the cell to regulate protein function through subcellular localization (24, 28). Leptomycin B (LMB) is an extremely useful tool used to selectively inhibit Crm1 (exportin-1)-dependent nuclear export of NES-containing proteins (10, 11, 26, 35, 41). LMB appears to covalently modify Crm1 export receptor at a conserved cysteine residue that renders the receptor incapable of forming the exporting trimeric complex between cargo protein and RanGTP (25).

In the present study, we provide evidence that IB may be the only NF-B inhibitor protein to possess an intrinsic nuclear export function. We employed LMB, knockout cells, chimeric constructs, and transient and stable transfection studies to monitor subcellular localization of NF-B-IB complexes, degradation of IB proteins, and NF-B DNA-binding activities during pre-and postinduction states. We found that N-NES of IB is primarily responsible for export of NF-B during pre- and postinduction states. However, NF-B-IB complexes are incapable of shuttling during both of these states, suggesting that, unlike IB, IB is capable of completely masking nuclear localization sequences of the NF-B dimer and incapable of exporting the complex out of the nucleus. Swapping the N-terminal region of IB for IB sequence allows NF-B bound to the chimeric protein to be exported out of the nucleus in a manner identical to IB. These results provide deeper insights into the fundamental differences between regulatory mechanisms governing function of IB and IB and possibly other IB members and implicate biological relevance of IB-specific nuclear export function in regulation of such processes as apoptosis and inflammatory responses.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. Mouse embryo fibroblasts (MEFs), 293 human embryonic kidney cells (HEK), and Cos-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Cellgro) supplemented with 10% fetal bovine serum (HyClone Laboratory, Inc.), 1,000 U of penicillin G (Sigma Chem. Co., St. Louis, Mo.), and 0.5 mg of streptomycin sulfate (Sigma) per ml in a humidified 10% CO₂ incubator (Forma Scientific). The p65^/ 3T3 cells were maintained in DMEM with 10% bovine calf serum and supplemented with antibiotics as above. Cells (embryonic fibroblasts and Cos-7) were transiently transfected using the Effectene transfection reagent (Qiagen) and the standard calcium phosphate precipitation method (5).

Reagents. Dimethyl sulfoxide (DMSO) and cycloheximide were purchased from Sigma. MG132 was purchased from Peptides International. Human recombinant TNF- was from CalBiochem and resuspended in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (fraction V; Sigma). In each experiment, all samples received the same amounts of DMSO to control for potential DMSO effects. Immunoglobulin G (IgG) antibodies against IB (C21), IB (C20), and p65 (C20) were purchased from Santa Cruz Biotechnology. The 5432 rabbit polyclonal antibody was raised against the N-terminal 56 amino acids of murine IB conjugated to glutathione S-transferase as described previously (30). A monoclonal anti-HA.11 (influenza virus hemagglutinin epitope) antibody was purchased from BabCO, and horseradish peroxidase-conjugated anti-rabbit and anti-mouse Ig antibodies were obtained from Amersham. Fluorescein isothiocyanate (FITC)- and tetramethyl rhodamine isocyanate (TRITC)-conjugated anti-rabbit and anti-mouse Ig antibodies were purchased from Sigma. Hoechst dye 33342 and rhodamine phalloidin (R-415) were purchased from Molecular Probes.

Western blot and EMSA procedures. Cell preparation and Western immunoblots were performed as described (31) and developed using the enhanced chemiluminescent (ECL) procedure according to the manufacturer (Amersham). Blots were then exposed to X-ray film (Kodak). The Ig-B oligonucleotide probe and conditions for electrophoretic mobility shift assay (EMSA) were previously described (31). The total cell extract buffer (20 mM HEPES [pH 7.9], 350 mM NaCl, 20% glycerol, 1% NP-40, 1 mM MgCl₂, 0.5 mM EDTA, 0.1 mM EGTA, 0.5 mM dithiothreitol, phenylmethylsulfonyl fluoride, and aprotinin) was used for both Western blot analysis and EMSA. Nuclear and cytoplasmic biochemical fractionations were accomplished using buffer C and buffer A, respectively, as described (31).

Construction of IB, IB and p65 chimeras and fusion proteins. N-terminally fused green fluorescent protein (GFP)-IB was generated by subcloning PCR-amplified human wild-type IB (MAD3) into the HindIII and BamHI sites of the pEGFP vector (Clontech). The human p65 cDNA was kindly provided by D. Ballard (Vanderbilt, Tenn.). To make the GFP N-terminally tagged p65 fusion protein, the p65 cDNA with HindIII and BamHI flanking sites was created by PCR and ligated in-frame into the pEGFP vector (Clontech). Truncated p65s (amino acids [aa] 1 to 420 and 1 to 450) were made with 3' primers specific to the C-terminal truncated portion of p65 and subcloned into the pEGFP vector as above. The murine p65 cDNA was blunt-end ligated into BamHI-NheI Klenow blunt-ended pCMX vector (provided by K. Umesono, Kyoto University). Similarly, CMX-p50 was generated by cleaving murine p105 cDNA at NcoI sites, Klenow filled, and blunt-ligated into BamHI-NheI Klenow blunt-ended pCMX vector. IB expression vector was constructed by inserting murine IB cDNA into pcDNA vector (Clontech). The amount of each plasmid vector was titrated to ensure the formation of IB and NF-B complexes in Cos-7 cells and MEFs. The IB construct (pLHL-CA-IB) was made by using specific primer sets with either 3' IB or 5' IB overhangs to PCR amplify pBS-IB (1 to 66) and pBS-IB (45 to 359). The two PCR products were joined by PCR using 5' IB and 3' IB-specific primers, subcloned into pBS at 5' BamHI and 3' XhoI, and subcloned into pLHL-CA vector at 5' BamHI-BglII and 3' XhoI sites. The N-NES mutant IB construct was constructed similarly to the IB construct except that the pBS-IB N-NES mutant construct was used as a template. The pBS-IB N-NES mutant template was generated using two-step PCR mutagenesis as described (17). The hemagglutinin (HA)-tagged murine IB gene (pLHL-CAHA-mIB) was constructed as described previously (31).

Visual analysis of NF-B and IB proteins. MEFs and Cos-7 cell lines were cultured in two-or four-welled chamber slides (Lab-Tek) and transiently transfected with various expression plasmid constructs. After drug treatment, cells were washed twice with ice-cold PBS and fixed with 3.7% formaldehyde in PBS for several hours at 4°C. Fixed cells were then rinsed with PBS and permeabilized with 0.2% Triton X-100 in PBS. Cells were blocked with 2% normal goat serum for 1 h and subsequently incubated with appropriate primary antibodies in PBS-0.2% Tween 20-2% goat serum at 37°C overnight. Staining was detected with either FITC- or TRITC-conjugated anti-mouse or -rabbit Ig secondary antibodies. Cells were mounted with Prolong Antifade (Molecular Probes) and visualized and photographed using a Zeiss Axioplan epifluorescence microscope with the aid of fluorescein- or rhodamine-specific filters. GFP fusion protein was visualized directly in living cells or under fixed conditions.

Generation of stable pools of MEFs. MEFs deficient for IB were reconstituted with either HA-tagged murine IB or the chimeric IB gene via retroviral infection as described previously (31). Briefly, pLHL-CAHA-mIB or pLHL-CA-IB and pLHL-CA-IB N-NESmut constructs were contransfected with pCLeco (32) helper virus in HEK 293 cells. Viruses were then harvested and transferred onto MEFs for infection. Stably infected pools of MEFs were selected in the presence of hygromycin. The stable pools of MEFs were then grown in the absence of hygromycin for 2 days before experiments were initiated.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

LMB cannot inhibit activation of NF-B associated with IB. It is known that nuclear export of NF-B can be mediated by the IB protein. It is yet unclear whether the capacity to export NF-B out of the nucleus is a general function of all IB family members. Scanning the primary amino acid sequences of other IB members, such as IB (44), p105/IB (14, 20, 37), p100 (29, 33), IB (46), and Bcl-3 (34), or cactus (13), the Drosophila homolog of IB, revealed no conserved NES motifs N-terminal of the first ankyrin repeat compared to IB (Fig. 1A; others not shown). Since IB is widely expressed and responds to stimulus-dependent degradation in a manner similar to IB, we chose to compare and contrast the mechanistic differences in the ability of IB and IB to modulate NF-B localization and activation. Consistent with earlier reports (17, 38), TNF--induced NF-B DNA-binding activity was inhibited by LMB (Fig. 1B, lower panel, lanes 2 and 3). However, surprisingly, LMB had no observable effect against TNF--induced NF-B activation in IB knockout cells (Fig. 1B, lower panel, lanes 6 and 7). Although TNF- induced the degradation of both IB and IB in the wild-type cells, LMB selectively retarded the degradation of IB, not IB (Fig. 1B, upper panel, lanes 2 and 3). Likewise, IB was degraded by TNF- stimulation but was not inhibited upon pretreatment with LMB in the IB-deficient cells (Fig. 1B, upper panel, lanes 6 and 7). These results show that LMB inhibits NF-B activation only through the IB protein. Together with a recent finding of the presence of a functional NES on p65 but not on the p50 or c-Rel subunit of NF-B (16), our data suggest that the p65 NES does not dominantly affect IB subcellular localization in an LMB-sensitive fashion. Alternatively, unlike IB, IB completely masks both nuclear localization sequences present on dimeric NF-B complexes, thereby preventing import of NF-B in the uninduced state. In the latter case, NF-B-IB complexes do not shuttle in an NES-dependent manner, making their subcellular localization LMB insensitive.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   IB is refractory to LMB inhibition of TNF--induced degradation. (A) N-terminal primary amino acid sequence alignment of IB and IB. Boxed region signifies the highly conserved IB kinase phosphorylation sites on the dual serines on IB and IB. Highlighted sequences (shaded box) are the recently identified "leucine-rich" N-terminal NES of IB that is not conserved in other IB proteins such as IB, IB, and cactus. (B) Wild-type and IB-deficient MEFs were untreated or treated with TNF- (10 ng/ml for 15 min; lanes 2 to 4 and 6 to 8) in the presence or absence of LMB (20 ng/ml; lanes 3 and 7) or MG132 (30 µM; lanes 4 and 8) pretreatment for 30 min. Total cell extracts were analyzed by EMSA by using an Ig-B probe (lower panel) and by Western blotting with IB (C-21) and IB (C-20) antibodies (upper panel). (C) p65-deficient 3T3 cells were treated with TNF- (10 ng/ml for 15 min; lanes 2 to 5) and pretreated in the presence or absence of LMB (20 ng/ml) for 15, 30, or 60 min (lanes 3, 4, and 5, respectively). Total cell extracts were analyzed by EMSA as described above (lower panel) and by Western blotting with IB- and IB-specific antibodies (upper panel).

IB is insensitive to LMB-induced nuclear accumulation in preinduction state. The above data suggest that NF-B-IB complexes do not shuttle in an LMB-sensitive fashion, unlike those containing IB. To directly determine whether the subcellular localization of preinduced IB is not affected by LMB, endogenous IB was stained with IB-specific antibodies and visualized under a fluorescent microscope. As an internal control, IB subcellular localization was similarly monitored. In untreated MEFs, both IB and IB were localized predominantly in the cytoplasm (Fig. 2A, upper left panels). Upon treatment of the cells with LMB, IB accumulated in the nucleus, as expected (Fig. 2A, upper right panels). In contrast, the localization of IB remained cytoplasmic. To ensure that selective nuclear accumulation of IB but not IB was achieved upon LMB treatment within the same cell, Cos cells were cotransfected with HA-tagged IB, IB, and p50-p65 NF-B subunits. Transfected cells were costained with antibodies specific to HA or IB, and localizations of IB and IB proteins were determined within the same cell. Exogenous IB was sensitive to LMB-induced nuclear accumulation, while IB did not migrate substantially into the nucleus (Fig. 2B, upper and middle panels). These results demonstrate that cytoplasmic localization of inactive NF-B is controlled by at least two different mechanisms. When NF-B is complexed with IB, the mechanism of localization is determined by the kinetics of nucleocytoplasmic shuttling, with nuclear export being dominant during the preinduced state. However, cytoplasmic localization of NF-B-IB is not regulated by an LMB-sensitive export mechanism.

View larger version (31K): [in this window] [in a new window]   FIG. 2.   NF-B-IB complexes do not shuttle between cytoplasm and nucleus in an LMB-sensitive manner. (A) Wild-type MEFs were untreated or treated with LMB (20 ng/ml for 30 min), fixed, and stained with either IB (C-20) or IB (C-21) antibodies. Cells were also costained with rhodamine-conjugated phalloidin to visualize actin for cell cytoskeleton integrity. For control, cells were also treated with LMB or TNF- as above and localization of p65 (C-20) was assessed. (B) Cos-7 cells were cotransfected with HA-tagged IB (1.0 µg), IB (1.0 µg), p50 (1.0 µg), and p65 (1.0 µg) constructs, untreated or treated with LMB as described for panel A, and stained with HA (red) and IB (green) antibodies. Cells were also stained with Hoechst DNA dye (blue).

NF-B cannot be sequentially activated without IB. It was recently shown that when the IB gene was replaced into the IB locus under control of the IB promoter, it could repress nuclear NF-B DNA-binding activity following pulse induction with TNF- (6). Thus, it was concluded that IB is functionally equivalent to IB. However, we hypothesize that the export of NF-B following the repression of NF-B DNA-binding activity allows the system to become permissive for reactivation or postrepression activation of NF-B. If NF-B-IB complexes are nuclear, reactivation by a second round of stimulation would not occur, since the activation event takes place primarily in the cytoplasmic compartment. To directly test this hypothesis, we first established the condition in which postinduction repression of NF-B DNA-binding activity can be reproducibly detected using both wild-type and IB-deficient MEFs (Fig. 3). As expected, resynthesis of IB directly correlated with the reduction of NF-B DNA-binding activity in wild-type cells (Fig. 3A and B, lanes 3 to 6). Similarly, NF-B DNA-binding activity decreased, in good agreement with increased IB protein levels in IB-deficient cells, but this effect required longer durations due to the lack of NF-B-dependent transcription of the IB gene (Fig. 3A and B, lanes 6 and 14). Once these conditions were established, MEFs were pulse stimulated with TNF- and then chased with medium without TNF- for 2 h before reactivation studies were performed (diagram in Fig. 4). Secondary stimulation of these cells with TNF- reactivated NF-B DNA-binding activity in wild-type cells (Fig. 4A, compare lanes 3 and 4). Importantly, while degradation of IB was observed during the reactivation phase, IB was mostly refractory to this process (Fig. 4A and B, lower panels, lanes 4). Consistent with the above observation and also with our hypothesis, TNF- could not efficiently activate NF-B for the second time in the absence of IB (Fig. 4B, lower panel, lanes 2 to 4). These results for the first time demonstrate that the postrepression activation of NF-B requires the NF-B-dependent resynthesis of IB, which cannot be compensated for by other IB family members.

View larger version (43K): [in this window] [in a new window]   FIG. 3.   Kinetic analyses of postinduction repression of NF-B DNA-binding activity in the presence and absence of IB protein. (A) Wild-type and IB-deficient MEFs were pulse treated for 10 min with TNF- (10 ng/ml; lanes 2 to 6 and 8 to 14) and subsequently chased with fresh medium without TNF- for the indicated amounts of time (0, 15, 30, 45, 60, 90, and 120 min). Lanes 1 and 7, untreated control samples. Total cell extracts from the samples were analyzed by Western blotting using IB- and IB-specific antibodies (A) and EMSA as described above (B).

View larger version (24K): [in this window] [in a new window]   FIG. 4.   NF-B cannot be sequentially activated in the absence of IB. (A) Wild-type MEFs were treated with TNF- (10 ng/ml for 10 min; lanes 2 to 4), then chased with fresh medium for 120 min (lanes 3 and 4), and finally restimulated with TNF- as above for 10 min (lane 4). Total cell extracts were analyzed by both EMSA (upper panel) and Western blotting (lower panel). The experimental setup is diagrammed above. Numbers in parentheses signify the lanes of the sample that were treated. (B) IB-deficient MEFs were treated as described above. IB- and IB-specific antibodies were used in both A and B.

Postinduction export of nuclear NF-B is inefficient in the absence of IB. Our EMSA analyses indicate that postinduction export of inactive NF-B complexes out of the nucleus requires the presence of IB. To directly determine the localization of NF-B-IB complexes during the postinduction phase, we examined the localization of the endogenous p65 by immunofluorescence in wild-type and IB^/ MEFs. The outline of the postinduction export experiment is diagrammed in Fig. 5. Consistent with NF-B-dependent synthesis of IB within 30 to 60 min post-TNF- stimulation (Fig. 3), nuclear p65 in the wild-type cells was efficiently exported out to the cytoplasm within 60 min of chase without TNF- (Fig. 5, panels C and D). Inclusion of cycloheximide or LMB during the chase period blocked the export of p65 (Fig. 5, panels G and H), demonstrating that nuclear export of NF-B during the postinduction phase indeed requires IB resynthesis and an NES-dependent process. However, in the IB-deficient MEFs, the majority of p65 remained in the nucleus after 60 min during the chase period (Fig. 5, panel K). Even after 3 h of chase, a large pool of p65 was nuclear (panels L and M), correlating with the resistance of NF-B to reactivation by TNF- (Fig. 4). Interestingly, adding LMB during the chase period caused a further increase in the amount of p65 in the nucleus (panel O). This observation suggests that an NES-dependent process was responsible for some p65 export during the prolonged chase period in IB-deficient cells. These observations together demonstrate that while efficient and rapid export of NF-B during postinduction phase requires the resynthesis of IB, it can be exported less effectively in the absence of IB, possibly via the recently identified NES on the p65 subunit.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Postinduction export of NF-B is sensitive to LMB and inefficient in the absence of IB. Wild-type and IB-deficient MEFs were treated with TNF- (10 ng/ml for 10 min; panels B to H and J to O) and chased with fresh medium for the indicated amounts of time (60, 120, or 180 min) in the presence or absence of cycloheximide (CX; 10 µg/ml) or LMB (20 ng/ml). The experimental setup is diagrammed above. The treated cells were then fixed and the endogenous p65 was stained with p65 (C-20)-specific antibody.

IB can enter the nucleus but is not efficiently exported during postinduction phase. The observations thus far suggest that NF-B associated with IB is exported efficiently but NF-B associated with IB is not during the postinduction phase. Direct examination of subcellular localization of IB during postinduction phase by both immunofluorescence and biochemical subfractionation analyses confirmed that indeed newly synthesized IB can enter the nucleus (Fig. 6A, vector only, panel 3, and 6B, compare IB, lane 7). Interestingly, the signal detected by C20 anti-IB antibody in immunofluorescence showed a similar extent of degradation as with IB after 30 min of TNF- stimulation (Fig. 6A, compare panels 2), but Western blot analyses consistently showed incomplete IB degradation compared to IB within the same time frame (Fig. 6B, IB and IB, lanes 2). The significance of this observation is discussed below (see Discussion). Importantly, while the presence of nuclear IB correlated well with the repression of NF-B DNA-binding activity in IB-deficient cells, IB was not exported out of the nucleus even up to 3 h during the chase period. LMB showed very little effect during the chase period (Fig. 6A, panel 4), suggesting that subcellular localization of IB is not affected by an NES-dependent export process. The lack of nuclear export of IB was not due to some defect of the IB-deficient cell system used because when IB tagged with the HA epitope (HA-IB) was stably expressed, HA-IB was efficiently exported during the postinduction phase in an LMB-sensitive fashion in a manner identical to endogenous IB in wild-type cells (Fig. 6A, HA-IB, panels 3 and 4). These results demonstrate that while newly synthesized IB can enter the nucleus, it fails to export efficiently out of the nucleus during the postinduction phase.

View larger version (48K): [in this window] [in a new window]   FIG. 6.   N-terminal region containing the N-NES of IB is sufficient to mediate the postinduction export function. (A) IB-deficient MEFs were stably transfected with either vector (vector only), HA-tagged IB (HA-IB), IB chimera (IB), or N-terminal NES mutant IB chimera contructs (IB N-NESmut). Pools of stable transfectants were isolated, and subcellular localization of IBs was visualized using antibodies against IB (C-20, vector only), IB (C-21, HA-IB), N-terminal 56 amino acids of IB (5432, IB, and IB N-NESmut). The cells were also treated with TNF- for 60 min (panels 2 to 4), chased with fresh medium for 180 min (panels 3), or in the presence of LMB (panels 4). (B) Stably transfected MEFs as above were either untreated (lanes 1 and 5), pulsed with TNF- for 30 min (lanes 2 and 6), chased with fresh growth medium for 120 min (lanes 3 and 7), and rechallenged with TNF- for another 30 min (lanes 4 and 8). Samples from each pool of cells were fractionated as described in Materials and Methods and analyzed by Western blotting using antibodies against IB (C-20), IB (C-21), and amino-terminal IB (5432) for the detection of the chimera protein. Samples from vector only, HA-IB, IB, or IB N-NESmut are shown as panels with arrows indicating IB, HA-IB, IB or IB N-NESmut, respectively.

IB N-NES is sufficient to mediate postinduction export function. There are a total of three independent NES motifs reported on IB-NF-B complexes, the N- and C- terminal NES motifs on IB and a p65 NES (2, 16, 17, 23, 40, 43). There is a controversy as to which of these putative NES sequences provide(s) the dominant nuclear export function for the complexes during the postinduction phase. To conclusively determine whether N-NES is sufficient to dominantly export the NF-B-IB complexes in the absence of C-NES, we asked whether IB N-terminal to the first ankyrin repeat could support dominant nuclear export function in the context of IB protein. We therefore generated an IB chimera expression construct in which the IB N-terminal region upstream of the first ankyrin repeat was swapped with the N-terminal aa 1 to 66 of IB (IB) and stably expressed it in IB-deficient MEFs. Similar to HA-IB, the chimeric protein was largely expressed in the cytoplasm (Fig. 6A, panel 1) and associated with NF-B as demonstrated by a coimmunoprecipitation assay (not shown). Degradation of IB after pulse stimulation with TNF- was complete like IB but unlike IB (Fig. 6A, IB, panel 2; Fig. 6B, IB, lane 2). Importantly, resynthesized IB was predominantly cytoplasmic, but LMB during the chase period trapped it in the nucleus (Fig. 6A, IB, panels 3 and 4). These results suggest that the N-terminal sequence of IB was able to dominantly export the chimeric complexes out of the nucleus.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Recently, several groups, including ours, discovered that NF-B-IB inactive complexes shuttle continuously between the nuclear and cytoplasmic compartments to achieve a predominant cytoplasmic localization during the absence of NF-B-activating signals (17, 23, 38, 43). Maintaining the cytoplasmic localization of the inactive complexes is essential for NF-B function, since signals derived from either the plasma membrane or the nucleus to target the degradation of IB are blocked if the inactive complexes are sequestered in the nuclear compartment (17, 18, 38). We therefore asked whether regulation of NF-B by nucleocytoplasmic shuttling was a conserved mechanism of all IB family members that negatively influence the important NF-B family of transcriptional regulators.

By scanning the primary amino acid sequences, we failed to identify a similar N-NES motif present in IB in the N-terminal regions of mammalian IB, p105/IB, p100, IB, Bcl-3, and the Drosophila IB homolog cactus (Fig. 1A). A classical NES sequence was also undetected in other regions of the IB family members. These observations suggested that IB might be the only IB family member that contained a novel nuclear export capacity to mediate rapid export of nuclear NF-B. These sequence analyses further implied that IB might be a more recently evolved family member. Coincidentally, IB is one of the major NF-B target genes in mammalian cells, which forms the autoinhibitory feedback loop to perform the postinduction repression of NF-B function. Even without the presence of a conserved N-terminal NES on IB, we could not rule out the possibility that the NF-B-IB complexes still shuttle in an LMB-sensitive manner, since a recent report has shown that the p65 subunit of NF-B possesses a bona fide NES motif (16).

In the present study, we found that LMB selectively blocked the signal-induced degradation of IB but not IB. IB knockout cells confirmed that TNF-induced NF-B activation cannot be inhibited by LMB in the absence of IB, since NF-B-IB, and possibly other NF-B-IB complexes, is insensitive to the LMB effect in these cells. Moreover, unlike IB-containing NF-B complexes, NF-B-IB complexes that contained the p65 subunit were still largely refractory to LMB-induced nuclear accumulation. These surprising results enabled us to modify the NF-B cytoplasmic sequestration model (3, 12), in which IB proteins, excluding IB, are likely proficient in masking the dual NF-B NLS motifs. Studies with IB suggest that the N-terminal sequence of IB is required for efficient masking of NF-B NLS sequences, since swapping of this region with that of IB permitted NF-B associated with the chimera protein to shuttle between the cytoplasm and nucleus (Fig. 6 and unpublished observation).

The NF-B-IB complexes, on the other hand, behave differently from other IB proteins in that cytoplasmic localization of the complexes is a result of a dynamic balance between nuclear import and active nuclear export forces. This notion is further supported by the observation of incomplete p50 NLS masking by IB ankyrin repeats in NF-B-IB cocrystals (19, 22) and the finding that the p65 NLS but not the p50 NLS motif is involved in direct association with IB (36). Even though at least three NES motifs have been identified in the NF-B-IB trimeric complex (2, 16, 17, 23, 40, 43), our data demonstrate that their nucleocytoplasmic shuttling is dominantly controlled by the IB N-NES motif, since (i) these complexes do not efficiently shuttle when the IB N-NES is mutated or deleted (17), (ii) an IB C-NES mutation has little effect on shuttling of the complexes (17, 23, 43), (iii) they also efficiently shuttle without p65 protein or when p65-NES is deleted (unpublished observation), and (iv) the IB N-NES is sufficient to confer dominant shuttling function when grafted onto the IB protein.

What is unclear is whether there are any physiological advantages to maintaining an energy-consuming and apparently futile shuttling process in order to preserve the asymmetric distribution of NF-B-IB complexes in the preinduced state. Perhaps it is more efficient to regulate the localization of shuttling proteins by simply adjusting the rate of nuclear entry versus export in order to drastically alter the steady-state subcellular distribution of the complexes. It is conceivable that as yet undiscovered physiological control of NF-B activity may exist which involves the altered regulation of nucleocytoplasmic shuttling of the inactive complexes. Adjusting the kinetics of shuttling by having the rate of import exceed the rate of export may be a novel mechanism to attenuate NF-B function. However, our findings suggest that, depending on the ratio of NF-B associated with IB or IB or other IB family members, only a subset of inactive NF-B pools may be subjected to this type of regulation. This may explain why certain investigators fail to observe large effects of LMB on NF-B activation in different experimental settings (23).

In contrast, what is clear from our present study is that the nuclear export function of IB N-NES is essential for rapid and efficient export of inactive NF-B complexes out of the nucleus during the postinduction phase. Our findings from IB knockout cells demonstrate that other IB family members could not efficiently perform this function. We found, in accordance with studies using cells isolated from IB knockin mice (6), newly synthesized IB protein was able to enter the nucleus and repress NF-B DNA-binding activity. However, IB, which does not possess a functional NES, could not efficiently carry NF-B out of the nucleus. Thus, only IB protein was able to prime the NF-B system for a subsequent reactivation event or provide a permissive condition for postrepression activation to take place.

Wild-type cells, which can properly export postinduced nuclear NF-B out to the cytoplasm, could efficiently respond to secondary NF-B stimuli. IB-deficient cells, however, were refractory to this NF-B reactivation process. Thus, it is logical to consider that the initial postinduction repression of NF-B DNA-binding activity is an important mechanism to rapidly shut off NF-B-dependent gene transcription following a short exposure to cytokine stimulation in a biological setting. This IB-mediated repression and removal of NF-B from its DNA-binding sites may be necessary, since NF-B has been shown to bind its cognate sites with very slow off rates in vitro (36). In addition, efficient removal of NF-B from cognate DNA-binding sites may also be critical for allowing the NF-B-IB complexes to interact with the soluble transport machinery, Crm1 and RanGTP (25, 28). Subsequently, the nuclear export of inactive NF-B-IB complexes provides another important function to allow cells to become quickly permissive for a second challenge with either cytokine, bacterial, or viral insults. Without efficient export of nuclear NF-B, the secondary activation process is defective. Thus, the nuclear export function of IB may contribute to the ability of an organism to respond rapidly to multiple cellular infections.

It is interesting to consider whether the N-NES of IB may have evolved to permit sequential NF-B activation. Because the N-NES of IB needs to interact with Crm1 for export function, i.e., exposed on the surface of NF-B-IB complexes, this domain may have lost the capacity to efficiently mask the p50 NLS. These events may have caused the NF-B-IB complexes to nucleocytoplasmically shuttle in the preinduced conditions by default. Important goals of future investigation include determining how nucleocytoplasmic shuttling of NF-B-IB complexes is regulated and whether disruption of these regulatory mechanisms has drastic biological and/or pathological consequences.

It is well known that IB is less responsive to stimulus-induced degradation than IB, as determined by Western blot analyses. However, the mechanism for this resistance is unclear. Unexpectedly, we found that when IB degradation was assessed by immunostaining using C20 IB antibody, TNF- stimulation caused almost complete loss of immunoreactivity, similar to that found with IB antibodies (Fig. 6). However, the same IB antibody still showed only partial degradation on Western blot analyses, consistent with the weaker responsiveness of IB to TNF--induced degradation. The pool of IB that was resistant to initial TNF- stimulation remained refractory to secondary stimulation even though it remained cytoplasmic, indicating that this pool of IB is not accessible to the stimulus-induced degradation pathway. A recent study by Ghosh and colleagues found that B-Ras might be critical for retarding IB degradation during the NF-B activation process (9). Interestingly, the interaction between B-Ras and IB is mediated by the C-terminal region of IB, which contains the epitope recognized by C20 antibody. It is possible that when B-Ras is bound to IB, C20 may not recognize IB due to epitope masking. Thus, it is tempting to speculate that there are at least two pools of IB, one free of B-Ras and the other associated with it. The IB pool that is free of B-Ras may be accessible for immunostaining with C20 and efficiently degraded by TNF- stimulation. In contrast, IB bound to B-Ras might be inaccessible for immunostaining using C20 antibody and resistant to degradation. Thus, our study suggests that the C20 antibody may provide a useful tool to help elucidate the mechanism of partial degradation of IB during NF-B activation processes.