The interferon-inducible p202a protein modulates NF-kappaB activity by inhibiting the binding to DNA of p50/p65 heterodimers and p65 homodimers while enhancing the binding of p50 homodimers.

p202a is a member of the interferon-inducible murine p200 family of proteins. These proteins share 1 or 2 partially conserved 200 amino acid segments of the a or the b type. The known biological activities of p202a include among others the regulation of muscle differentiation, cell proliferation, and apoptosis. These biological activities of p202a can be correlated with the inhibition of the activity of several transcription factors. Thus, the binding of p202a results in the inhibition of the sequence-specific binding to DNA of the c-Fos, c-Jun, E2F1, E2F4, MyoD, myogenin, and c-Myc transcription factors. This study concerns the mechanisms by which p202a inhibits the activity of NF-kappaB, a transcription factor involved among others in host defense, inflammation, immunity, and the apoptotic response. NF-kappaB consists of p50 and p65 subunits. We demonstrate that p202a can inhibit in vitro and in vivo the binding to DNA of p65 homodimers and p50/65 heterodimers, whereas it increases the binding of p50 homodimers. Thus p202a can impair NF-kappaB activity both by inhibiting the binding to DNA of the transcriptionally active p65 homodimers and p50/p65 heterodimers and by boosting the binding of the repressive p50 homodimers. p202a can bind p50 and p65 in vitro and in vivo, and p202a can be part of the p50 homodimer complex bound to DNA. p50 binds in p202a to the a type segment, whereas p65 binds to the b type segment. Transfected ectopic p202a increases the apoptotic effect of tumor necrosis factor (at least in part) by inhibiting NF-kappaB and its antiapoptotic activity.

Functionally, the best characterized murine 200 family proteins are p204 and p202a. Both of these proteins include two partially conserved 200-amino acid segments: one a type segment and one b type segment. Overexpression of p204 in transfected mammalian cells leads to inhibition of growth (6,7). p204 is required for the differentiation of cultured myoblasts to myotubes, and the level of p204 increases during differentiation as a consequence of transactivation of the gene by MyoD (8). p204 enables the differentiation in part by overcoming the inhibition of the activities of MyoD, E12/E47, and other myogenic transcription factors by the Id proteins, Id1, Id2, and Id3 (9) (see also Ref. 10). p204 binds and sequesters the Id proteins and also elicits a decrease in their level (9). p204 also inhibits ribosomal RNA synthesis, by inhibiting the sequence-specific binding to DNA of the transcription factor UBF (6). p204 is also induced by infection with cytomegalovirus and is required for the replication of this virus (11).
Similarly to p204, overexpression of p202a in transfected mammalian cells also inhibits cell proliferation (12). This inhibitory effect can be correlated with the binding and inhibition of the activity of several transcription factors by p202a. These include c-Fos, c-Jun, AP2, E2F1, E2F4, MyoD, myogenin, c-Myc, and NF-B (13)(14)(15)(16)(17). p202a also binds pRb (12) and inhibits the activity of p53 (18). In turn, the transcription of the Ifi202a gene (encoding p202a) is inhibited by p53 (19). The inhibitory activity of p202a can be overcome by the binding of p202a to the p53 binding protein 1 (18). p202a binds the human adenovirus E1A oncoprotein and reduces E1A-mediated apoptosis (20). E1A in turn can partially overcome the inhibition of cell proliferation by p202a. Lowering the concentration of serum in the medium results in the increase in the level of p202a in fibroblasts (21). p202a is also induced during skeletal muscle differentiation in consequence of transactivation of the Ifi202a gene by the muscle-specific MyoD transcription factor (16,22). p202a can modulate during this process the activity of the muscle-specific transcription factors MyoD and myogenin, and it also inhibits apoptosis (16,17). Furthermore, the levels of 202a RNA and p202a are increased in NIH 3T3 cells by the H-ras oncogene (23). This is apparently in consequence of the activation by the H-Ras oncoprotein of the transcription factor c-Jun, which binds to two AP-1 DNA binding sites in the Ifi202a gene encoding p202a. It was proposed that the increased expression of p202a in the cells transfected with the activated H-ras expression plasmid increases the survival of the cells in low serum conditions (23). p202a has a sister protein, p202b, which is encoded by the Ifi202b gene and differs from p202a only in 7 of 445 amino acids (1). The disruption of the Ifi202a gene in mice has no obvious phenotype, apparently in consequence of the increase of the p202b level, compensating for the loss of p202a. Overexpression of p202 (which of the two types remains to be identified) was linked in mice to susceptibility to the autoimmune disease lupus erythematosus (24). The binding of p202a to transcription factors blocks transactivation in most cases (e.g. c-Fos, c-Jun, E2F1, E2F4, MyoD, and myogenin) by inhibiting their sequence-specific binding to DNA (13)(14)(15)(16). However, c-Myc-dependent transactivation is blocked by p202a in consequence of the inhibition of the binding of c-Myc to Max (17).
We have reported earlier that p202a also binds and inhibits the activity of the transcription factor NF-B (13). NF-B activates numerous genes involved in host defense, inflammation, immunity, apoptosis, and tumor development (25)(26)(27)(28)(29)(30)(31). NF-B is a member of the NF-B/Rel transcription factor family. In mammals, this consists of p65 (RelA), RelB, c-Rel, p105/ p50 (NF-B1), and p105/p52 (NF-B2). All family members share a partially conserved domain of 300 amino acids, the Rel homology domain. This is required for dimerization, and the members form various homo-and heterodimers and exert their activities in the form of dimers. p65, RelB, and c-Rel contain potent transactivation domains. In contrast, the mature p50 and p52 proteins lack strong activation domains, and the p50 and p52 homodimers were shown to act as repressors (27,(32)(33)(34)(35)(36)(37). However, there are several reports suggesting that in certain cases p50 homodimers can transactivate gene expression in vitro (38) and in vivo, especially after interaction with particular nuclear proteins (39 -42). The p50/p65 complex (usually designated as NF-B) is the most abundant of the NF-B/Rel family heterodimers (27). It is present in essentially all cells.
The NF-B proteins can be retained in the cytoplasm in an inactive form as a consequence of binding to proteins of the IB family (IB␣, IB␤, and IB⑀) (27). (For a discussion of the possibility that NF-B⅐IB complexes might shuttle between the cytoplasm and the nucleus, see Ref. 31.) The activation of NF-B and its movement to the nucleus occurs following the proteolytic degradation of the IB proteins. The degradation of these is triggered by phosphorylation, mediated by the IKK (inhibitor of IB kinase) complex together with the regulatory protein NEMO (43). Phosphorylation leads to polyubiquitination and degradation of IB by the proteasomes. The phosphorylation also activates PKAc, the catalytic subunit of protein kinase A, the cyclic AMP activatable protein kinase, which can phosphorylate and activate the NF-B p65 subunit (44). Activators of the NF-B family of transcription factors include a large variety of agents (e.g. double-stranded RNA, bacterial lipopolysaccharides, phorbol esters, cytokines (e.g. TNF 1 and interleukin-1), UV light, and viruses (27,31,45). Activated NF-B can associate in the nucleus with a DNA binding motif with the consensus sequence 5Ј-GGGRNNYYCC-3Ј (27,28).
The interferon-inducible p202a protein was found earlier to bind to the p50 and p65 subunits of the NF-B transcription factor and to inhibit NF-B-dependent transcription of reporter constructs and endogenous genes in various cultured cell lines (13). Here we report observations that help to elucidate the mechanism of this inhibition: (a) p202a inhibits the rate and extent of the sequence-specific binding to DNA of in vitro synthesized murine p65 homodimers and p50/p65 heterodimers, whereas it increases the binding of p50 homodimers; (b) p202a added to nuclear extracts also decreases the binding to DNA of native p50/p65 heterodimers, whereas it increases the binding of p50 homodimers; and (c) overexpression of ectopic p202a in transfected AKR-2B cells also inhibits the binding to DNA of endogenous p50/p65 heterodimers while increasing the binding of p50/p50 homodimers. Consequently, p202a might diminish the transcriptional activity of NF-B in two ways: (a) by inhibiting the binding to DNA of the transcriptionally active p50/p65 heterodimers and p65 homodimers and (b) by boosting the binding of the transcriptionally inactive p50 homodimers. Antisera to p50 or p65 coimmunoprecipitated p202a, and antiserum to p202a coimmunoprecipitated p50 and p65 from the cell extracts, revealing an interaction between these proteins in vivo. An antiserum to p202a supershifted the p50 complex in an electrophoretic mobility shift assay, but only when p202a was present, suggesting that p202a can be part of the p50 complex bound to DNA. This conclusion was in agreement with the finding that p202a was retained by p50 bound to an NF-B oligodeoxynucleotide coupled to beads. The NF-B subunits bound to distinct segments in p202a: p50 to the a segment and p65 to the b segment. Whereas, in the binding to p202a only one short region of p65 close to the C terminus was required, several regions of p50 were involved. In accord with an earlier report (46), transfected ectopic p202a increased the apoptotic effect of tumor necrosis factor on NIH 3T3 cells, presumably by suppressing the antiapoptotic effect of NF-B.

Construction of the ov1, ov2, and ov3 Cell Lines
Overexpressing p202 AKR-2B cells grown to 60% confluence were transfected with a p202a expression plasmid (pcDNA3-p202a) or with the vector (pcDNA3 serving as control) using LipofectAMINE according to the instructions of the manufacturer (Invitrogen). 48 h after transfection the cells were split 1:5 into selection medium containing G418 (500 g/ml). After a 10 -14day incubation, G418-resistant colonies were picked. Three of these colonies (ov1, ov2, and ov3) were amplified, and their p202 levels were compared with that of the control AKR-2B line by immunoblotting. The p202a levels in the overexpressing lines were also compared with the levels of p202a induced by treatment of control AKR-2B cells with 1000 units/ml interferon for various lengths of time. The levels of p202a as well as of NF-B p50 and p65 were determined in 10-g protein samples from the culture lysates. As an internal control, the levels of ␤-actin were also determined.

Fusion Proteins and Deletion Mutants
The preparation of GST-p50 was reported (49). His-tagged p65 was expressed in Escherichia coli and purified by affinity chromatography (50). The construction of expression plasmids encoding GST-p202a and its truncated fusion proteins were reported earlier (17). The proteins expressed in E. coli were purified using glutathione-agarose beads (Sigma) and Sephadex G50 columns. Protein concentrations were determined using the Bio-Rad reagent.

Electrophoretic Mobility Shift and Supershift Assays
Double-stranded oligodeoxynucleotide probes were labeled using [␥-32 P]dATP and T4 polynucleotide kinase and purified on Micro-SpinG-25 columns (Amersham Biosciences). The reaction mixtures (20 l) included 40 fmol of oligodeoxynucleotide probe, 200 ng of protein expressed in vitro, or 10 g of nuclear extract protein in 10 mM Hepes (pH 7.9), 10 mM KCl, 0.5 mM DTT, 0.5 mM EDTA, 0.2-1 g of poly(dI-dC), and 5% glycerol. Incubations were at room temperature for the times indicated. Unlabeled double-stranded oligodeoxynucleotide (in 10-fold excess over the probe) was added, if so indicated, 10 min prior to adding the labeled probe. Supershift assays involved the incubation of reaction mixtures including nuclear extract with 2 l of antibodies, at 4°C overnight, before adding the labeled probe. Samples were loaded onto a 5% nondenaturing polyacrylamide gel in 0.5ϫ TBE buffer (0.045 M Tris borate, 1 mM EDTA) and electrophoresed at 300 V for 2 h. The separated DNA-protein complexes were visualized by autoradiography. If so indicated, the signal strength was determined by densitometry. The average signal intensity and S.D. values were determined based on the results of three independent experiments. Nuclear extracts were prepared as described (51) with slight modifications. Briefly, cells were washed three times with phosphate-buffered saline (PBS), scraped into 500 l of PBS buffer, and spun down, and the pellet was resuspended in 3 packed cell volumes of Triton lysis buffer (9 mM Tris-HCl (pH 8.0), 60 mM EDTA, 1 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride, and Triton X-100 (the concentration of Triton X-100 was 0.3% for HeLa and AKR2B cells and 0.2% for EL4 and 293 cells)). After 5 min on ice, the lysates were sedimented by centrifugation. The pelleted nuclei were washed, and nuclear proteins were extracted with 2 packed cell volumes of nuclear extract buffer (20 mM Hepes (pH 8.0), 420 mM NaCl, 1.5 mM MgCl 2 , 0.5 mM DTT, 0.2 mM EDTA, and 25% glycerol) at 4°C for 45 min. The soluble material was pelleted, and the supernatant fraction was dialyzed at 4°C for 1 h against Shapiro's buffer D (20 mM Hepes (pH 7.9), 20% glycerol, 100 mM KCl, 2 mM DTT, 0.2 mM EDTA, 0.2 mM EGTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.7 mg pepstatin A per liter and 0.5 mg/liter leupeptin) (52). The precipitate was removed by centrifugation, and the supernatant fraction was stored at Ϫ80°C (nuclear extract).

Immunoprecipitation and Immunoblotting
Immunoprecipitation and immunoblotting were performed as described (53). The antibody ␣-p202, ␣-p65, ␣-p50, or preserum was coupled to Gel10 beads (Sigma) by mixing antibody and Gel10 beads in a 1:1 ratio, incubated at 4°C with rotation for 4 h, sedimented, washed with 0.1ϫ MOPS buffer (20 mM MOPS, 5 mM sodium acetate, and 1 mM EDTA) four times, and resuspended in 1ϫ immunoprecipitation buffer (150 mM NaCl, 10 mM Tris-HCl (pH 7.4), 1 mM phenylmethylsulfonyl fluoride, 0.1% Nonidet P-40, 0.5 mg/liter leupeptin). The beads were then added to cell extracts, incubated with rotation at 4°C and 1 h, and washed five times with 1ϫ immunoprecipitation buffer. The proteins were eluted from the beads with 1ϫ SDS sample buffer, analyzed by SDS-PAGE and immunoblotting, and visualized using the ECL system (Amersham Biosciences).
It should be noted that extracts from various murine cells and tissues were reported earlier to contain mRNAs for both p202a and p202b. The ratios of these two mRNAs varied with their sources. Furthermore, treatment with interferons increased the levels of both 202a and 202b mRNAs (1). An antiserum raised against p202a recognized and immuno-precipitated both p202a and p202b, which differ only in 7 of 445 amino acids (1). Thus, we could not determine the proportions of p202a and p202b immunocoprecipitated from cell extracts. Therefore, we have continued to designate as p202a coprecipitates that contained both p202a and p202b. This is the case, for example, in Fig. 4, A and B.

GST Pull-down Assay
Covalent Coupling of Purified GST Fusion Proteins to CNBr-Sepharose Beads-Purified GST, GST-p50, GST-p202a, and truncated fusion proteins were coupled to CNBr-activated Sepharose beads (54) using 0.5 mg of protein/ml of resin in 500 mM NaCl, 100 mM NaHCO 3 (pH 8.3), at room temperature, for 3 h. The beads were washed four times with PBS, and the remaining active groups on the beads were blocked in 200 mM glycine (pH 8.0) at room temperature for 3 h. The beads were stored at 4°C in PBS and 0.001% sodium azide.
The various truncated forms of p65 protein were generated by cleavage of a cDNA plasmid encoding p65 (550 aa) (54) with Tth111 I (504 aa), EcoNI (477 aa), PstI (441 aa), BspHI (311 aa), and ScaI (288 aa) (where the numbers of aa in the truncated proteins encoded by the various products of the cleavages are indicated in parentheses) (55). Truncated forms of p50 were generated by cleavage of p50 cDNA (429 aa) with HindIII (363 aa), EarI (281 aa), and SauIII AI (191 aa) (49). The plasmids were linearized and expressed in vitro using 35 S methionine and the TNT Quick Coupled Transcription/Translation System (Promega). The plasmids (based on pGEX3X) encoding p202a and its truncated versions (17) were expressed in E. coli. 2-l aliquots of the proteins were analyzed by SDS-PAGE and staining with Coomassie Blue. In the pull-down assay, equal amounts of GST fusion proteins (0.5-1.0 g) were immobilized on CNBr-activated Sepharose beads and incubated at 4°C for 1 h with 2.5 l of the labeled proteins translated in vitro in TNT buffer (20 mM Tris HCl (pH 8.0), 150 mM NaCl, 1% Triton-X, 0.5 mM phenylmethylsulfonyl fluoride). After extensive washing with TNT buffer, the bound proteins were eluted in SDS sample buffer and analyzed by SDS-PAGE and autoradiography.
Assay for the Interaction of Immobilized NF-B-specific Oligodeoxynucleotide with p50 and Labeled p202a Using Magnetic Beads, an Approach Based on Ref. 56 -[ 35 S]Methionine-labeled p202a was generated by expressing a pcDNA3-p202a plasmid in a reticulocyte-coupled transcription-translation system (TNT) according to the instructions of the manufacturer. Labeling of the NF-B-specific oligodeoxynucleotide with biotin was performed according to the BioNick labeling method following the instructions of Invitrogen. 20-l aliquots of Streptavidincoupled magnetic beads (50%, v/v) were distributed into five Eppendorf tubes, washed twice with 20 l of 1ϫ binding buffer (20 mM Hepes (pH 7.6), 30 mM KCl, 10 mM (NH 4 ) 2 SO 4 , 1 mM EDTA, 1 mM DTT, 1% Tween 20), and supplemented with 10 l of [ 35 S]p202a. The volume of the reaction mixture in each tube was adjusted to 50 l by adding water and 2ϫ binding buffer (i.e. the final concentration of binding buffer was 1ϫ). Tube 1 was mixed, incubated with rotation at room temperature for 1 h, and processed by washing the beads three times with 1ϫ binding buffer and eluted with 50 l of 1ϫ elution buffer (i.e. binding buffer supplemented with 1 or 1.5 M NaCl). Tube 2 was also supplemented with 7 l (0.56 pmol) biotin-labeled oligodeoxynucleotide and incubated and processed as tube 1. Tube 3 was also supplemented with 3 g of GST-p50 and incubated and processed as tube 1. Tube 4 was supplemented with both 7 l of biotin-labeled oligodeoxynucleotide and 3 g of GST-p50 and was incubated and processed as Tube 1. Tube 5 was supplemented with 7 l of biotin-labeled oligodeoxynucleotide and 3 g of His-p65 and was incubated and processed as Tube 1. The eluates from the washed beads from each tube were analyzed by SDS-PAGE and autoradiography.

Assays of Apoptosis
FACS Analysis-Mouse 3T3 cultures derived from embryonic fibroblasts of wild type Swiss Webster mice and from p65 Ϫ/Ϫ Swiss Webster mice (57) were cultured in complete Dulbecco's modified Eagle's medium. If so indicated, the cultures were transfected at 50 -60% confluence with pcDNA3 control vector or pCMV 202a expression plasmid using 2 g of DNA/60-mm culture dish following the Lipofect-AMINE procedure (Promega). After 6 h, the cultures were washed with PBS and incubated in fresh complete Dulbecco's modified Eagle's medium for 48 h. Subsequently, 10 ng/ml murine TNF-␣ was added to the medium (58) for 0-, 10-, or 24-h incubations as indicated. Thereafter, for flow cytometry, both floating and attached cells were collected, fixed, stained with propidium iodide, and analyzed by FACS (using a Becton-Dickinson FACS Vantage flow cytometer).
TUNEL Assay-3T3 cultures were derived from p65 Ϫ/Ϫ or wild type p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 Swiss Webster mouse embryonic fibroblasts and were cultured as above. p65 Ϫ/Ϫ cultures were used without transfection. 3T3 cultures from wild type fibroblasts were transfected as indicated above with pcDNA3 control vector or pCMV 202a expression plasmid. All three types of cultures were incubated with 10 ng/ml murine TNF-␣ for 24 h. The cultures were fixed with 4% paraformaldehyde and apoptotic cells were detected using the in situ cell death detection kit AP (Roche Applied Science) following the manufacturer's suggestions (59).

RESULTS
p202a Inhibited the Sequence-specific Binding to DNA of p65 Homodimers and p50/p65 Heterodimers but Increased the Binding of p50 Homodimers-We started to explore the mechanism of inhibition of NF-B activity by p202a by performing electrophoretic mobility shift assays. These assays revealed that: (a) p202a decreased the rate and the extent of the sequence-specific binding to DNA of in vitro-synthesized murine p65 homodimers (Fig. 1, A1 and A2) and p50/p65 heterodimers (C1 and C2), whereas it increased the binding of p50 homodimers (B1 and B2); (b) it is in line with the above finding that p202a added to nuclear extracts from 293 human embryonic kidney cells, HeLa human cervical carcinoma cells, and EL4 murine T cells also decreased the binding to DNA of native p50/p65 heterodimers, whereas it increased the binding of p50 homodimers (Fig. 2).
To establish whether this effect of p202a is also manifested in vivo, we transfected AKR-2B cultures with a p202a expression plasmid and generated (by G418 selection) three lines (ov1, ov2, and ov3*) overexpressing p202a. The extent of overexpression varied among these lines from 2.7-fold in ov1 to 5.5-fold in ov3* (Fig. 3A, top panel). The overexpression diminished the level of p65 (but not of p50) by about 50% in the three cell lines. The overexpressed level of p202 (even in ov3*) was in a physiological range (i.e. it was much lower than that obtained FIG. 1. p202a inhibited the sequence-specific binding to DNA of synthetic, purified NF-B p65 homodimers and p50/p65 heterodimers but increased the binding of p50 homodimers. A1, p202a inhibited the sequence-specific binding to DNA of p65 homodimers. His-p65 and poly(dI-dC) were incubated with a 32 P-labeled NF-B-specific RRBE oligodeoxynucleotide, without or with GST-p202a, for the times indicated. The specificity of the binding was established in 20-min incubations. These revealed no binding to GST-p202a in the absence of His-p65, no inhibition of the binding to His-p65 by GST (a component of GST-p202a), and no binding to His-p65 in the presence of excess unlabeled oligonucleotide (Cold oligo). The incubated reaction mixtures were analyzed by EMSA. A2, time courses of the binding without and with GST-p202a. Shown is a densitometric analysis of the data in A1. The extent of the binding to His-p65 without GST-p202a at 30 min was taken as 1. The S.D. values based on three experiments are shown. B1 and B2, p202a increased the sequence-specific binding to DNA of p50 homodimers. The experiments were as in A1 and A2, except that GST-p50 was substituted for His-65. In B2, the extent of binding to GST-p50 in the absence of GST-p202a at 30 min was taken as 1. C1 and C2, p202a inhibited the sequence-specific binding to DNA of p50/p65 heterodimers. The experiments were as in A1 and A2, except that GST-p50/His-p65 was substituted for His-p65. In C2, the extent of binding to p50/p65 in the absence of GST-p202a at 30 min was taken as 1. For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 after exposing wild type AKR-2B cells to interferon (1000 unit/ ml) for 24 h (Fig. 3A, bottom panel)).
Electrophoretic mobility shift assays (Fig. 3B) revealed that when testing cultures in which the NF-B system was not activated by PMA in the extract of the p202a overexpressing AKR/ov3* line, there was a strong decrease in the level of the sequence-specific binding to DNA of p50/p65 heterodimers and a strong increase in the binding of p50/p50 homodimers compared with the case of the extract from the AKR control cells (compare lanes 9 and 12). In experiments with cultures, in which the NF-B system was activated by PMA, in the extract from the p202a-overexpressing AKR/ov3* line there was a huge increase in the level of the sequence-specific binding of p50/p50 homodimers and a smaller decrease in the binding of p50/p65 heterodimers, compared with the case of the extract from the control AKR cells (compare lanes 10 and 13). As expected, the addition of an excess of unlabeled DNA probe eliminated the binding of the labeled probe to NF-B homo and heterodimers (compare lanes 9 and 11 and lanes 12 and 14).
In vivo the p50/p65 heterodimers and the p65 homodimers serve as transcription factors, whereas the p50 homodimers are repressors of transcription (37). Consequently, p202a decreased the transcriptional activity of NF-B (13) in vitro and in vivo in two ways: (a) by inhibiting the binding to DNA of the FIG. 2. p202a inhibited the binding to DNA of p50/p65 heterodimers but increased the binding of p50 homodimers in nuclear extracts from 293, HeLa, and EL4 cells. Aliquots of nuclear extracts from 293, HeLa, and EL4 cell cultures were incubated with 32 P-labeled RRBE oligodeoxynucleotide in 20 l of EMSA reaction mixtures in the absence of GST-202a (lanes 1, 5, and 9), or in the presence of 100 ng of GST-202 (lanes 2, 6, and 10), 200 ng of GST-202a (lanes 3, 7,  and 11), or 400 ng of GST-202a (lanes 4, 8,  and 12). After incubation, the reaction mixtures were analyzed by EMSA. The positions of the p50/p65 heterodimer and p50 homodimer bands as well as of the free probe are indicated. For further details, see "Experimental Procedures." FIG. 3. The overexpression of p202a in ov3, an AKR-2B line (which was generated by transfection with a p202a expression plasmid and selection), inhibits in the nuclear extract the sequence-specific binding to DNA of p50/p65 heterodimers but increases the binding of p50/p50 homodimers. A, top panel, comparison of the levels by immunoblotting of p202a, NF-B p50 and p65, and ␤-actin (serving as an internal control) in three cell lines, ov1, ov2, and ov3* (generated by the transfection of a p202a expression plasmid into AKR-2B cells and selection), with the levels of the same proteins in the control AKR-2B line (AKR). The p202a, p50, p65, and ␤-actin levels are indicated. The ov3 line used in the experiment in B is indicated by an asterisk. Bottom panel, level of p202a in a control AKR-2B line, without and after exposure to 1000 units/ml interferon for various lengths of time, as indicated. As an internal control, ␤-actin was used. The p202a and ␤-actin bands as revealed by immunoblotting are indicated by arrows. B, overexpression of p202a in the ov3 line (AKR/ov3*) inhibits in the nuclear extract the sequencespecific binding to DNA of NF-B p50/p65 heterodimers but increases the binding of p50/p50 homodimers. The p202a level in the ov3 line was about 5.5-fold higher than in the control (AKR) line. Nuclear extracts (NE) were prepared from cells of the AKR/ov3* and the control AKR lines. The sequence-specific binding to a labeled double-stranded oligodeoxynucleotide probe of NF-B p50/p65 heterodimers and p50/p50 homodimers in the nuclear extracts from the AKR/ov3* and the AKR lines were compared in an EMSA. Lanes 9 -11, nuclear extracts from AKR; lanes 12-14, nuclear extracts from AKR/ov3*. The cells from which the nuclear extracts were prepared were, if so indicated (lanes 10 and 13), incubated with 50 ng/ml PMA (Sigma) for 12 h to activate the NF-B system. The reaction mixtures (in lanes 11 and 14) were supplemented with a 20-fold excess of cold competitor oligodeoxynucleotide. The p50/p65 and p50/p50 bands as well as the positions of the free probe are indicated. For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 transcriptionally active p50/p65 heterodimers and p65 homodimers and (b) by boosting the binding of the repressive p50 homodimers. A further decrease in the NF-B activity by p202 may be due to the decrease in the level of p65 protein.
Coimmunoprecipitation of p202a with p50 and p65-Antisera to p65 or p50 (but not preimmune serum) coimmunoprecipitated p202a from extracts of AKR-2B cells (Fig. 4, A and B). The amount of p202a coimmunoprecipitated by an antiserum to p65 was greater in an extract from cells that had been exposed to interferon, compared with extracts from control cells (Fig. 4A). This is as expected, since the level of p202a is low in control cells and is increased severalfold in cells exposed to interferon (Fig. 3A). Furthermore, as expected, the amount of p202a coimmunoprecipitated by an antiserum to p50 was greater from an extract of cells that had been transfected with an expression plasmid encoding p202a, compared with an extract of cells transfected with only the empty vector (B). The amounts of p65 (or p50) coimmunoprecipitated by an antiserum to p202a were much greater from extracts of AKR-2B or 293 cells that had been exposed to interferon than from extracts of control cells (Fig. 3, C and D). This is a consequence of the increase in p202a levels by interferon. These coimmunoprecipitation experiments reveal an interaction between p202a and p50 as well as between p202a and p65.
Identification of the Regions of Interaction between p202a and p50 and between p202a and p65-We performed GST pull-down assays for identifying the regions of interaction between p202a and NF-B p50 and p65 subunits. We generated a series of fusion proteins consisting of a GST moiety at their N termini, to which different segments of p202a were fused (for a schematic drawing of the fusion proteins and the designation of the p202a segments, see Fig. 5D). The fusion proteins were expressed in E. coli and purified (Fig. 5A). a and b stand for the two 200-amino acid segments whose sequence is partially conserved in all 200 family proteins. a1 and a2 are two subsegments of a, and b1 and b2 are two subsegments of b. Each of the a1, a2, b1, and b2 subsegments is encoded by a distinct exon.
The pull-down assays revealed that p65 bound to the complete GST-p202a, and the binding could be localized to the a1 subsegment (Fig. 5B). p50, as expected, also bound to the complete GST-202a; however, it bound poorly to the a segment but bound strongly to the b segment. Within the b segment, binding to b2 was more efficient than to b1 (Fig. 5C).
These results indicate that p65 and p50 can bind to distinct segments in p202a: p65 to the a1 segment, and p50 to the b segment (see the table in Fig. 5D).
For identifying the regions in p65 and p50 that bind to p202a, GST-p202a fusion protein and a series of C-terminally truncated p65 and p50 proteins were used (schematic drawings of these are shown in the lower panels of Fig. 6, A and B). The p65 and p50 segments added to the reaction mixtures (in Input) and retained on GST-202a in (GST-p202a pull-down) are shown in the upper parts of Fig. 6, A and B. The data reveal that, as expected, complete p65 was bound to GST-p202a, and the lack of a 46-amino acid segment from the C terminus of p65 did not decrease the binding. However, the deletion of a further 27-amino acid segment eliminated the binding to p202a (Fig.  6A). Complete p50, as expected, bound to GST-202a (Fig. 6B).
Deletion of a C-terminal segment of 66 amino acids strengthened the binding. The deletion of a further 82-amino acid segment eliminated the binding. Unexpectedly, an even further truncation by 90 amino acids restored a weak binding. These results suggest that within p65 a short (27-amino acid) segment close, but not directly adjacent to, the C terminus is required for the binding to p202a (Fig. 6A). However, the case of p50 is more complex. There is a binding site within the N-terminal 191-amino acid segment. This site can apparently be masked in the longer, 281-amino acid N-terminal segment. The even larger 363 amino acid N-terminal segment binds more strongly than the complete p50 protein. This suggests that the segment contains a further binding site. This further site is, however, weakened by the 66-amino acid C-terminal segment.
p202a Can Form a Complex with p50 Bound to Its Recognition Sequence in DNA: Supershift and Bead Assays-The results presented so far indicate that p202a interacts with p50 and p65 in vivo and increases the sequence-specific binding of p50 to DNA. One mechanism by which p202a might increase this binding is by becoming part of a complex with p50 and its recognition sequence in DNA. We performed two types of tests FIG. 4. Antibodies to p65 or p50 coimmunoprecipitated p202a and an antiserum to p202a coimmunoprecipitated p65 and p50 from cell extracts. A and B, immunoprecipitation (IP). Preimmune serum (Pre-S) or antisera to p65 (␣-p65) or p50 (␣-p50) were coupled to Gel 10 beads, and the beads were incubated with extracts from control AKR-2B cells (Ϫ), cells treated with interferon (IFN), cells transfected with pCMV vector (V), or cells with pCMV-p202a expression plasmid (p202a). IB, immunoblotting. The beads were washed and eluted with SDS, and the eluted proteins were analyzed by immunoblotting with ␣-p202a. The position of the p202a band is indicated. It should be noted that ␣-p202a recognized in Western blotting and immunoprecipitated both p202a and p202b. C and D, immunoprecipitation. Preimmune serum or antiserum to p202a (␣-p202a), as indicated, was coupled to Gel 10 beads, and the beads were incubated with extracts from control (Ϫ) or interferon-treated (IFN) AKR-2B or 293 cells. The beads were washed and eluted with SDS, and the eluted proteins were analyzed by immunoblotting with ␣-p65 or ␣-p50. The positions of the p65 and p50 bands are indicated. For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 to explore whether such a hypothetical complex can be formed. The first of these tests was based on examining by EMSAs whether antibodies to p202a can supershift a complex formed in a reaction mixture including p202a, p50, and the appropriate DNA probe. The results of this test suggested that the hypothetical complex was in fact formed (Fig. 7). Control experiments revealed that, as expected, p50 bound to the labeled DNA probe, and GST-p202a (which alone did not bind) strengthened the binding of p50 (Fig. 7A). Excess cold oligodeoxynucleotide eliminated the binding, preimmune serum did not affect it, and ␣-p50 supershifted the p50 DNA complex. Most importantly, as noted, the complex was supershifted by ␣-p202a or ␣-GST (the latter was active due to the presence of GST in the GST-p202a used). The mobility of the band supershifted by ␣-p50 was decreased by the presence of GST-p202a in the reaction mixture. All of these findings suggested that GST-p202a did become part of a complex including p50 and the DNA probe. An important control in Fig. 7B revealed that, as expected, ␣-p202a did not supershift the p50-DNA complex in the absence of GST-p202a.
EMSAs with p65 (in the form of His-p65) and the same DNA probe as used with p50 are shown in Fig. 7C. The complex, including His-p65 was, as expected, supershifted by ␣-p65, but not by ␣-p202a, even if the reaction mixture included GST-p202a. This lack of supershifting was expected, since p202a (which increased the sequence-specific binding of p50 to DNA) strongly inhibited the binding of p65 to DNA (Fig. 1). This suggests that little or no p202a-p65 complex could bind to DNA.
The data in Fig. 7, D and E, were obtained in experiments with nuclear extracts from AKR-2B cells. These data confirmed the results in Figs. 1 and 2 by showing that GST-p202a, (but not GST), decreased the binding of p50-p65 heterodimers to DNA, whereas it increased the binding of p50 homodimers to DNA. Moreover, the EMSAs in Fig. 7D revealed that, as expected, ␣-p65 supershifted the p50/p65 heterodimers but not the p50 homodimers. Significantly, ␣-p202 supershifted the p50 homodimers but only when GST-p202a was also present. However, ␣-p202a did not supershift in the absence of p202a, even when tested over a wide range of concentrations (Fig. 7E). These results (in Fig. 7, A-E) clearly established that ␣-p202a could supershift in an EMSA a complex including p202a attached to p50 bound to its recognition site in DNA.
To verify the existence of such a p202a-p50-DNA complex, we performed a second type of test (Fig. 8). This involved the use of 35 S-labeled p202a, biotin-labeled NF-B-specific oligodeoxynucleotide immobilized on streptavidin-beads, and purified GST-p50 or His-p65. The results of the test revealed that the binding of p202a to the streptavidin beads depended on both FIG . 5. In p202a, primarily the a segment bound p65, and primarily the b segment bound p50. GST pull-down assays. A, the fusion proteins, consisting of GST linked to different segments from p202a (for a schematic drawing with the designations of the segments, see D), were expressed in bacteria and purified. The various GST-p202a fusion proteins used were examined by SDS-PAGE and Coomassie Blue staining. The p202a segments present in the various fusion proteins are indicated. The segments specified are indicated by arrows. B and C, binding of 35 S-labeled p65 and p50 to GST-p202a and its various segments. GST-pull-down assays, SDS-PAGE, and autoradiography were performed. D, schematic drawings of the GST-p202a segments and their strengths of binding of p50 and p65. For further details see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 the biotin-labeled, NF-B-specific oligodeoxynucleotide and GST-p50. GST (substituting for GST-p50) did not allow the binding of p202a to the streptavidin beads (not shown). His-65 (substituting for GST-p50) allowed the binding of at most onetwentieth as much p202a as GST-p50 did.
These results confirm those of the EMSAs (in Fig. 7) in establishing the existence of a complex including p202a attached to p50 bound to its DNA recognition site. The finding that His-p65 was not able to substitute for GST-p50 in this test is in line with the inability of ␣-p202 to supershift a p65-DNA complex formed in a reaction mixture including GST-p202a (Fig. 7C).
The Inhibition of NF-B Activity by p202a Increased the Susceptibility to Apoptosis Triggered by TNF-␣-Treatment with TNF-␣ transmits signaling cascades, triggering apoptotic cell death, and other signals, eliciting the protection of the cells from TNF-␣-induced apoptosis (58, 60 -64). This protection is apparently the consequence of the transactivation of the expression of several antiapoptotic proteins by NF-B. Thus, an inhibition of NF-B activity should increase cell killing by TNF-␣. It is in line with this expectation that proteolysis of IKK␤, a protein required for the activation of NF-B, in response to TNF-␣ increased the TNF-␣-induced apoptosis (65). Moreover, the inhibition of NF-B activity in two human breast cancer lines by transfection with a murine p202a expression plasmid was also reported to increase the TNF-␣-mediated apoptosis (46). Since the human counterpart of the murine p202a protein has not been identified, we proceeded to confirm and extend this report (46) using a completely murine system (i.e. 3T3 cells derived from wild type Swiss Webster mouse embryonic fibroblasts and a pCMV-202a expression plasmid). As a control, we also included 3T3 cells derived from p65 Ϫ/Ϫ Swiss Webster mouse embryonic fibroblasts that are lacking FIG. 6. In p65, a short segment (amino acids 477-504) close to the C terminus was essential for the binding to p202a, and in p50 several regions were involved in the binding. GST pull-down assays are shown. Aliquots of a cDNA expression plasmid for p65 (A) and p50 (B) were truncated by digestion with different restriction enzymes. The resulting cDNA fragments were transcribed and translated in vitro into 35 S-labeled segments from p65 (A) and p50 (B). The segments were quantified by SDS-PAGE and autoradiography (Input). Aliquots of the labeled segments were incubated with GST-p202a; the bound segments were coprecipitated by binding to glutathione-agarose beads and thereafter released and analyzed by SDS-PAGE and autoradiography (GST-p202a pull-down assay). The autoradiograms indicate the restriction enzymes used for generating the p65 and p50 fragments, the amounts of input fragments, and the amounts of the fragments bound to GST-p202a. Schematic drawings of the p65 and p50 fragments are provided, and the strengths of their binding to GST-p202a are indicated (ϩϩϩ, strong binding; ϩϩ, weaker binding; Ϫ, no binding). FL, full-length. For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 NF-B activity and consequently are highly sensitive to TNFinduced apoptosis (57,58).
As shown in Fig. 9A, an exposure to TNF-␣ for 24 h resulted in the apoptosis of ϳ53% of the p65 Ϫ/Ϫ cells, 13% of the untransfected wild type cells, 20% of the wild type cells transfected with the pCMV control vector, and 36% of the wild type cells transfected with the pCMV-p202a expression plasmid. Thus, an increase in p202a level (and the resulting decrease in the NF-B activity) in wild type cells transfected with the pCMV-p202a plasmid greatly increased the level of apoptosis (from 20 to 36%). This assay was based on the determination of the percentage of cells with sub-G 1 amount of DNA by FACS. A TUNEL assay (Fig. 9B) also revealed that the (low) level of apoptosis in the culture of wild type murine 3T3 cells (transfected with the pcDNA3 control vector (b)), was strongly increased in the culture transfected with the pCMV-202a expression plasmid (c), reaching a high level of apoptosis similar to that in a culture of p65 Ϫ/Ϫ 3T3 cells (a), which are known to be especially prone to TNF-␣-induced apoptosis.
These results confirm and extend the conclusions of Ref. 46 in establishing that an increase in the p202a level increases the susceptibility of cultured fibroblasts to apoptosis induced by TNF-␣. It remains to be established whether the proapoptotic effect of p202a in the case of cells treated with TNF-␣ is only the consequence of the inhibition of NF-B activity by p202a or if other activities of p202a contribute to the effect. DISCUSSION p202a, the interferon-inducible p200 family protein, modulates various cellular responses. It can retard cell proliferation (12). During skeletal muscle differentiation, it is induced by myogenic transcription factors, and it controls the expression and activity of such factors (16). It can inhibit or increase apoptosis, depending on the triggering agent (17, 46) (this study). Expression of the H-ras oncogene also results in an increase in the level of p202a (23), whereas an increase in the FIG. 7. An antiserum to p202a supershifted the position of p50, but not of p65, in the presence of p202a in an EMSA. A and B, p50 was incubated with 32 P-labeled RRBE oligodeoxynucleotide in EMSA mixtures, as indicated, without or with GST-p202a, GST, cold RRBE oligodeoxynucleotide (Cold oligo), and one of the following antisera: preimmune serum (Pre-S), antiserum to p50 (␣-p50), antiserum to GST (␣-GST), or antiserum to p202a (␣-p202a). The reaction mixtures were analyzed by EMSA. The positions of the p50/p50 band and of the supershifted p50 bands are indicated. C, His-p65 was incubated with 32 P-labeled RRBE oligodeoxynucleotide without or with GST-p202a and one of the following antisera: preimmune serum, ␣-p65, or ␣-p202a. The reaction mixtures were analyzed by EMSA. The positions of the p65/p65 band and of the supershifted p65 band are indicated. D, nuclear extract from AKR-2B cells (AKR NE) was incubated with 32 P-labeled RRBE oligodeoxynucleotide in an EMSA reaction mixture as indicated, without or with GST, GST-p202a, and one of the following antisera: preimmune serum, ␣-p65, ␣-p50, or ␣-p202a. The reaction mixtures were analyzed by EMSA. The positions of the p50/p65 and p50/p50 bands and of the supershifted bands are indicated. E, AKR NE in the amounts of protein indicated were incubated in 20 l of EMSA reaction mixtures with the amounts of ␣-p202a indicated. The EMSA revealed that ␣-p202a did not supershift either p50/p65 or p50/p50 (unless GST-p202a was added as in D). For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 level of p53 results in a decrease of the p202a level (19). Overexpression of p202a in lymphoid cells was correlated with the autoimmune disease lupus erythematosus, conceivably in con-sequence of the antiapoptotic activity of p202a (24,66). The above listed biological effects of p202a were related to the modulation of activity of numerous transcription factors by p202a (13). In the case of most of the transcription factors modulated (e.g. c-Fos, c-Jun, AP2, E2F1, E2F4, MyoD, and myogenin), p202a bound to the transcription factor and inhibited its sequence-specific binding to DNA (13)(14)(15)(16). In the case of c-Myc, p202a inhibited binding to Max (17). By modulating the activities of several transcription factors, each of which is known to effect the expression of many genes (e.g. c-Myc, E2F1, E2F4, MyoD, myogenin, p53, pRb, c-Fos, c-Jun, and NF-B), p202a modulates indirectly the expression possibly of hundreds of genes.
In this study, we have explored the mechanisms by which p202a inhibits the activity of NF-B (13). Experiments in vitro involving the use of synthetic NF-B p50 and p65 subunits ( Fig. 1) and nuclear extracts from cultured mammalian cells (Fig. 2) revealed that added p202a inhibited the sequence-specific binding to DNA of p65 homodimers and p50/ p65 heterodimers, whereas it enhanced the binding of p50 homodimers.
The effects of increasing the endogenous level of p202a on the sequence-specific binding to DNA of NF-B p50/p65, and p50/ p50 dimers were also tested. The experiments involved a comparison of extracts from AKR cell cultures transfected with a p202a expression plasmid and thus overexpressing p202a with extracts from appropriate control cultures. The results obtained (Fig. 3) were in accord with the results with experiments in vitro involving the addition of p202a to purified NF-B p50 and/or p65 dimers (Fig. 1) or to nuclear extracts from various cultured cells (Fig. 2). An increase in the endogenous p202a level decreased the binding to DNA of p50/p65 heterodimers, whereas it increased the binding of p50 homodimers. This was the case with extracts from cultures in which the NF-B system was activated by PMA and with extracts from control cultures not exposed to PMA.
Interactions in vivo between p202a and p50 as well as between p202a and p65 were revealed by coimmunoprecipitation (Fig. 4). The binding of p202a to p50 subunits attached to NF-B-specific oligodeoxynucleotides was established in an assay involving supershifting elicited by an antiserum to p202a in an EMSA (Fig. 7) and was confirmed by demonstrating that p202a was retained by p50 subunits bound to an immobilized NF-B-specific oligodeoxynucleotide (Fig. 8). GST pull-down assays established that the two NF-B subunits bound to distinct segments in p202a: p50 to the b segment and p65 to the a segment (and within this primarily to the a1 subsegment) (Fig.  5). In the interaction with p202a, several regions of p50 were involved, whereas only one short p65 region close to the C terminus was responsible for binding to p202a (Fig. 6).
We reported earlier that p202a inhibited the transactivation of several NF-B-dependent reporter genes (e.g. PRDII-CAT, including a segment from the interferon-␤ gene regulatory region that contained a B site, and HIV-Luc, including a human immunodeficiency virus 5Ј long terminal repeat segment with two B sites) (13).
The experiments presented in this report suggest that p202a might influence the transcription of NF-B-regulated genes in two ways. First, by inhibiting the binding to DNA of p50/p65 heterodimers and p65 homodimers, which represent transcriptionally active forms of NF-B, p202a can directly interfere with the transcriptional activity of NF-B. Second, by enhancing the binding of p50 homodimers that are known to act as repressors of NF-B-dependent genes (37, 67), p202a can suppress the activity of NF-B.
The inhibitory role of p50 homodimers in the expression of A, detection of apoptosis by FACS. 3T3 cells derived from wild-type Swiss Webster mouse embryonic fibroblasts (CON), and p65 Ϫ/Ϫ Swiss Webster embryonic fibroblasts (p65 Ϫ/Ϫ) were used, together with wild-type 3T3 cells transfected with the pcDNA3 control vector (pcDNA3) or with the pCMV-p202a cDNA expression plasmid (pCMV-p202a) encoding p202a. The cultures were treated with murine TNF-␣ for 0, 10, or 24 h, as indicated. Thereafter, both the floating and the attached cells were collected, stained with propidium iodide, and analyzed by FACS. Cells with sub-G 1 amounts of DNA were taken as apoptotic; these are shown as percentage of control cells not treated with TNF-␣. S.D. values were calculated from three independent experiments. B, detection of apoptosis by TUNEL assay. 3T3 cells derived from p65 Ϫ/Ϫ Swiss Webster mouse embryo fibroblasts were used without further transfection (a), whereas 3T3 cells from wild type Swiss Webster mouse embryo fibroblasts were transfected with pcDNA3 control vector (b) or with pCMV-p202a expression plasmid (c). After a 24-h incubation with murine TNF-␣, the three cultures were fixed with paraformaldehyde, and the apoptotic cells were visualized using an in situ cell death detection kit. The arrows point to apoptotic cells. For further details, see "Experimental Procedures." p202a Inhibits Binding to DNA of NF-B p50/p65, Boosts That of p50 NF-B-regulated genes is well established (e.g. (a) in unstimulated cells p50 homodimers form a complex with histone deacetylase 1, and this complex binds to DNA and suppresses NF-B-dependent gene expression (37), and (b) studies with p50 knockout animals revealed a severalfold increase in the expression of the interferon ␤ gene (68)).
In enhancing the binding of p50 homodimers to B sites in DNA, p202a resembles the Bcl-3 protein (40,69). The induction of this protein by immunological adjuvants was shown to promote the survival of activated T cells (70). Bcl-3 was reported on the one hand to antagonize the p50-mediated inhibition of NF-B target genes by sequestering p50 homodimers and thereby increasing access by NF-B RelA-50 (71). On the other hand, Bcl-3 was shown to bind several proteins including Tip60, a histone acetylase, and the resulting Tip60-Bcl-3-p50 (or p52) complex was reported to activate gene expression (40). This was established in the case of a Drosophila gene (P-selectin) with an NF-B DNA binding site that is preferentially bound by p50 (or p52) (72).
It remains to be seen whether conditions can be found in which p202a (which was shown so far to inhibit transactivation by at least 10 distinct transcription factors) (13)(14)(15)(16)(17)(18) can enable transactivation of any gene by p50 homodimers.
In addition to triggering a proapoptotic signal, TNF-␣ can also activate NF-B and thereby protect cells from apoptosis (58, 60 -64). This suggested that by inhibiting NF-B activity p202a may increase the susceptibility of cells to apoptosis triggered by TNF-␣. Our results (shown in Fig. 9) are in accord with this expectation. The results confirm and extend an earlier report that was based on a heterologous (human-mouse) test system (46).
The identification of the various protein regions among p202a and the NF-B p50 and p65 subunits involved in the inhibitory interactions (Figs. 5 and 6) may therefore facilitate the development of antitumor agents that inhibit the activity of NF-B and thereby increase the proapoptotic activity of TNF-␣ and some other agents.
It should be noted that c-Myc also sensitizes cells to TNFmediated apoptosis by inhibiting NF-B transactivation. This was reported to be due to an interference by c-Myc with p65 transactivation (73). In contrast to the proapoptotic activity of p202a in the case of apoptosis triggered by TNF-␣, p202a is antiapoptotic in the case of apoptosis elicited by low serum concentration or increased c-Myc activity (17) (see also Ref. 74). This antiapoptotic activity of p202a might be a consequence, at least in part, of the inhibition by p202a of transactivation by the proapoptotic agents p53, E2F1, and c-Myc (14,17,18).
Finally, it is of interest that interferon can not only inhibit NF-B activity (by inducing the formation of p202a) (13) (this study) but can also activate NF-B. Thus, interferon was shown to activate NF-B in various cell lines by triggering the phosphorylation and degradation of IB␣, the protein that can retain NF-B in the cytoplasm in an inactive form (75). The inhibition of NF-B activity in consequence of the induction of p202a takes several hours of exposure of the cells to interferon and can persist for at least 48 h (13). The activation of NF-B, however, can be detected within 5 min after beginning the exposure of the cells to interferon, and it disappears within 4 h (75).