INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The NF-B/Rel family of transcription factors participates in the regulation of the immunomodulatory genes and activates numerous cellular genes as well as viral genes including the human immunodeficiency virus type 1 (HIV-1) long terminal repeat (LTR) (6, 7, 52, 60). The NF-B/Rel family members can be subdivided into two subgroups according to their structure and function: the DNA binding proteins NF-B1(p50), NF-B2(p52), RelA(p65), c-Rel, and RelB and the NF-B1(p105) and NF-B2(p100) precursors which are proteolytically cleaved to generate DNA binding proteins (p50 and p52, respectively). All members of the family share an N-terminal 300-amino-acid domain known as the NF-B/Rel/dorsal homology region which is responsible for binding to DNA (consensus sequence GGGRNNYYCC [6, 7, 52, 60]), dimerization, and nuclear translocation of NF-B (5). The dimer composition of different NF-B subunits and the sequence context of NF-B sites in different promoters contribute to the differential specificity of gene activation (22, 34, 38, 46).

The members of the IB family include IB (26), IB (58), IB (61), IB (24), and Bcl-3 (27), as well as the NF-B proteins p105 (39) and p100 (41), which contain ankyrin repeats in the C-terminal portion of the molecule and bind to NF-B, masking the nuclear localization sequence (10, 11). The most extensively characterized of the IB proteins is IB. Upon stimulation by many activating agents, including tumor necrosis factor (TNF) and phorbol 12-myristate 13-acetate (PMA), IB is rapidly phosphorylated by the recently identified 700- to 900-kDa complex containing IB kinase (IKK) (19, 48, 62). Phosphorylation targets IB for ubiquitination and degradation by the 26S proteasome, resulting in the release of NF-B (16). The substitution of alanine for Ser-32 and Ser-36 within the N-terminal signal response domain abolished the signal-induced IB phosphorylation and degradation, resulting in a blockage of NF-B activation (13, 14, 59). These mutations also abrogated in vitro ubiquitination of the IB protein (16, 50, 55). The amino terminus of IB is necessary for signal-induced degradation, but degradation of IB also requires the C-terminal domain of the protein (9, 14, 21, 30, 37, 38, 49) which is constitutively phosphorylated by casein kinase II (8, 37, 40). Once released, NF-B is able to activate target genes until new IB is synthesized.

Since the IB gene contains NF-B binding sites in its promoter, NF-B is able to autoregulate the transcription of its own inhibitor (15, 17, 29, 35, 57). This autoregulatory control of IB expression is in part responsible for the transient nature of the NF-B activation of gene expression; newly synthesized IB can localize to the nucleus and directly interfere with gene expression by dissociating protein-DNA complexes (3, 53). The IB gene has also been shown to be regulated by RelA(p65) at both the mRNA and protein levels: RelA(p65)-IB protein interactions increased the half-life of the inhibitory protein, and IB mRNA was induced by RelA(p65) as a consequence of increased IB gene transcription (12, 56). Stimulation of Jurkat T cells by TNF- or PMA induced degradation of IB protein concomitant with NF-B release and activation (36, 42). Activation was followed by de novo IB synthesis in an NF-B-dependent manner, and cycloheximide treatment prior to induction resulted in the inhibition of IB resynthesis, as well as prolonged NF-B DNA binding (57).

Previous analysis of the human IB promoter (29, 35) identified three NF-B sites (63 to 53, 225 to 216, and 319 to 310) and two NF-B-like sites (159 to 150 and 34 to 24) (see Fig. 2C). By deletion and point mutagenesis only the B1 site from 63 to 53 was shown to be functionally important for inducibility, since disruption of the B1 site completely abolished IB promoter activity, whereas deletion of the B2 site from 319 to 310 and B3 site from 225 to 216 had no effect. Another study demonstrated that in addition to the B1 site, the B-like site located between 34 to 24 is essential for IB gene expression since activation of the IB promoter by TNF- was abolished when B1 and B-like sites were mutated (29). Both sites bound p50 homodimers in resting HeLa cells and p50, p65, and c-Rel complexes in TNF--induced cells.

In Jurkat T cells that express transdominant repressors of IB (TD-IB) under the control of a tetracycline (Tet)-inducible system, endogenous IB protein expression was blocked by TD-IB induction (31). We now demonstrate that inducer-dependent induction of IB gene transcription was blocked by the transdominant repressor expression at the transcriptional level. To further analyze the autoregulatory control of IB expression, dimethyl sulfate (DMS) genomic footprinting was used to determine the pattern of protein-DNA interactions at the IB locus in stimulated Jurkat T cells and in TD-IB-expressing cells. These studies permit the first in vivo characterization of IB transcriptional autoregulation by NF-B and identify the promoter-proximal NF-B/Sp1 transcriptional switch as an essential component in the regulation of the IB promoter.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell lines and reagents. Jurkat cells and Jurkat cells stably expressing TD-IB were described previously (31). IB(2N4) contains alanine substitutions at the Ser-32 and Ser-36 inducible phosphorylation sites as well as a 22-amino-acid deletion of the C-terminal domain of IB, a region of the PEST domain that is dispensable with regard to binding of NF-B subunits but is important for IB degradation (9, 37). All cells were grown in RPMI 1640 containing 10% fetal bovine serum (FBS), 2 mM glutamine, and 10 µg of gentamicin per ml. Cells were stimulated by PMA (50 ng/ml; ICN, Costa Mesa, Calif.) plus phytohemagglutinin (PHA; 1 µg/ml; Sigma, St. Louis, Mo.) or TNF- (10 µg/ml; R&D System, Minneapolis, Minn.).

Immunoblot analysis. To characterize kinetics of expression, Jurkat T cells transfected with rtTA-Neo (rtTA-Neo Jurkat cells) and rtTA-IB(2N4) Jurkat cells were cultured in the presence of doxycycline (Dox; 1 µg/ml; Sigma) for various times. Cells were then washed with phosphate-buffered saline (PBS) and lysed in a mixture containing 10 mM Tris-HCl (pH 8.0), 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 0.5% Nonidet P-40 (NP-40), 0.5 mM phenylmethysulfonyl fluoride, leupeptin (10 µg/ml), pepstatin (10 µg/ml), and aprotinin (10 µg/ml). Equivalent amounts of whole-cell extract (20 µg) were subject to sodium dodecyl sulfate-polyacrylamide gel electrophoresis in a 10% polyacrylamide gel. After transfer, the Hybond membrane (Amersham, Cleveland, Ohio) was incubated overnight with N-terminal IB monoclonal antibody MAD10B (30) at 4°C. After four 10-min washes with PBS, membranes were reacted with a peroxidase-conjugated secondary goat anti-mouse antibody (Kierkegaard & Perry Laboratories, Gaithersburg, Md.) at a dilution of 1:1,000. The reaction was then visualized with an enhanced chemiluminescence detection system as recommended by the manufacturer (NEN Life Science, Boston, Mass.).

RNase protection assay. A 221-bp XbaI-PstI fragment was obtained by PCR amplification with an IB cDNA clone (cloned into pSVK3) by using specific primers containing restriction enzyme sites corresponding to positions 824 to 839 (5'ATCATCTAGAAACAGAGTTACCTACC3') and 1030 to 1045 (5'ATCACTGCAGTAACGTCAGACGCTGG3'); the XbaI-PstI fragment of the PCR product was cloned into the XbaI-PstI site of the pDP18-T7/T3 transcription vector (Ambion, Inc., Austin, Tex.) to generate pDP18CU-/CTERM. ³²P-labeled antisense RNA probe was transcribed by using an in vitro transcription kit (Pharmingen, San Diego, Calif.), and RNase protection was carried out with an RNase protection kit (Pharmingen). A -actin antisense probe (pTRI--actin; Ambion) was synthesized by the same protocol and used in the same reaction with an IB probe; 5 to 10 µg of total RNA extracted by using an RNeasy mini kit (Qiagen, Valencia, Calif.) from unstimulated or stimulated rtTA-Neo or rtTA-IB(2N4) Jurkat cells was used. The resulting protected RNAs were resolved on a 5% denaturing gel and exposed to X-ray film.

EMSA. Following the addition of 1 µg of Dox per ml to the culture medium for 24 h, nuclear extracts were prepared from rtTA-Neo or rtTA-IB(2N4) Jurkat T cells or Jurkat T cells after induction with TNF- or PMA-PHA for times ranging from 10 min to 24 h. Nuclear extracts were prepared as previously described (31) and subjected to electrophoretic mobility shift assay (EMSA) by using ³²P-labeled probes corresponding to the IB promoter either in NF-B DNA binding buffer (20 mM HEPES [pH 7.9], 5% glycerol, 0.1 M KCl, 0.2 mM EDTA [pH 8.0], 0.2 mM EGTA [pH 8.0]) or in NF-B/Sp1 DNA binding buffer (20 mM HEPES [pH 7.9], 100 mM KCl, 20% glycerol, 0.1 mM EDTA, 0.25 mM ZnSO₄, 0.05% NP-40), together with 5 or 0.5 µg of poly(dI-dC), respectively. Oligonucleotides used are as follows: B1, 5'-GATCTTGGAAATTCCCCGA-3'; Sp1, 5'-TCGAGACCCCGCCCCAG-3'; consensus Sp1, 5'-ATTCGATCGGGGCGGGGCGAGC-3'; mutated Sp1 probe, 5'-ATTCGATCGGTTCGGGGCGAGC-3'; B1/Sp1, 5'-TCGATTGGAAATTCCCCGAGCCTGACCCCGCCCCAG-3'; mutB1/Sp1, 5'-TCGATTGTCAATTCCCCGAGCCTGACCCCGCCCCAG-3'; B1/mutSp1, 5'-TCGATTGGAAATTCCCCGAGCCTGACCAAGCCCCAG-3'; and +5 B1/Sp1, 5'-TCGATTGGAAATTCCCCGAGCTGCAGCTGACCCCGCCCCAG-3'. Underlining delineates the Sp1 site, and boldface letters indicate mutations in either B1 or Sp1 sites. Recombinant proteins (glutathione S-transferase [GST]-NF-B fusion proteins [38, 47] and Sp1 [Promega Inc.]) were also incubated with the probes in a different DNA binding buffer [10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM dithiothreitol, 0.1 mg bovine serum albumin per ml, 50 µM MgCl₂, 1 mM ATP, 5 µg of poly(dI-dC) per ml]. The resulting protein-DNA complexes were resolved on 5 to 6% polyacrylamide-1× Tris-borate-EDTA gels and exposed to X-ray film. To demonstrate the specificity of protein-DNA complex formation, 125-fold molar excess of unlabeled oligonucleotide was added to the nuclear extract before addition of labeled probe. Supershift analysis was performed with anti-p65, anti-p50, anti-c-Rel, and anti-Sp1 antibodies (Santa Cruz Biotechnology Inc.).

In vivo genomic footprinting. Jurkat cells (10⁸) were harvested and resuspended in 1 ml of RPMI 1640-10% FBS containing 20 mM HEPES (pH 7.3). The methylation reaction was performed in presence of 10 µl of concentrated DMS (Aldrich Chemical Company, Milwaukee, Wis.) for 1 min. The reaction was then quenched by two washes in cold PBS containing 2% -mercaptoethanol. Genomic DNA extraction was performed as previously described (1). Briefly, cells were lysed in 2 ml of Tris buffer (pH 7.5) containing 10 mM NaCl and 10 mM EDTA and supplemented with 100 µl of proteinase K (20 mg/ml), 100 µl sodium dodecyl sulfate (20%), and 100 µl of NP-40 (10%) and then incubated at 50°C overnight. Proteins were precipitated by adding 1.2 ml of 5 M NaCl and centrifuged for 40 min at 7,500 × g (SS34 rotor, Sorval RS5 Superspeed). Cleared supernatant was ethanol precipitated to obtain genomic DNA; the pellet was resuspended in 200 µl H₂O with 20 µl of piperidine (Aldrich) and incubated 30 min at 90°C to provide cleavage of methylated G (or A) residues. The DNA control (naked DNA) was first extracted from cells and then submitted to DMS treatment and subsequently to piperidine cleavage to allow methylation and cleavage to all G residues of the sequence. For each sample, 2 µg of DNA were submitted to ligation-mediated PCR (LM-PCR) using Vent DNA polymerase (New England Biolabs, Mississauga, Ontario, Canada) as described elsewhere (23, 43). PCR amplification was for 2 min for the first cycle and was progressively increased to 10 min in the last cycle. A total of 18 cycles were performed for DNA amplification. The third primer was radiolabeled by end labeling using T4 polynucleotide kinase (Pharmacia Biotech) and [-³²P]ATP (ICN). Two more PCR cycles were performed to radiolabeled elongated DNA. The final labeled PCR product was analyzed on a 5% Hydrolink Long Ranger sequencing gel (Baker, Phillipsburg, N.J.) in 1.0× Tris-borate-EDTA at 65 W and exposed for 12 to 36 h with BioMax sensitive film (Eastman Kodak, Rochester, N.Y.). For the LM-PCR, two sets of oligonucleotides were used: for the noncoding strand, primers 1 (5'-CTCATCGCAGGGAGTTTCT-3'; melting temperature [T_m], 55°C; 2 (5'-CCCAGCTCAGGGTTTAGGCTTCTTT-3'; T_m, 63°C); and 3 (5'-GGGTTTAGGCTTCTTTTTCCCCCTAGCAG-3'; T_m, 66°C); for the coding strand, primers 1B (5'-ACTGCTGTGGGCTCTGCA-3'; T_m, 63°C); 2B (5'-TAAACGCTGGCTGGGGATTTCTCTG-3'; T_m, 63°C); and 3B (5'-TGGGGATTTCTCTGGGGCGGGGTCAGGCT-3'; T_m, 71°C) (see Fig. 2C).

Plasmid construction and mutagenesis. 0.4SK-pGL3 Luc was obtained by subcloning the 0.4-kb fragment of the IB promoter (a kind gift from A. Israël) into KpnI/SacI-digested pGL3. Plasmids carrying point mutations in the B1 or Sp1 site or in both sites were obtained by the subcloning of PCR-amplified fragments into KpnI/SacI-digested pGL3. Briefly, these constructs were obtained in two steps by a procedure previously described (33). The first round of amplification used 0.4SK-pGL3 Luc as template, 5'-CACGCGTAAGAGCTCCACCG-3' (SacI primer) as 3' primer, and 5'-GGAAATTCaaCGAGCCTGAC-3' (nucleotides in lowercase indicate point mutations) as 5' primer for B1 site mutagenesis or 5'-CCTGACCaaGCCCCAGAGAA-3' as 5' primer for Sp1 site mutagenesis. The amplified fragments were used as the 3' primer in a second PCR using 0.4SK-pGL3 Luc as template and 5'-CTATCGATAGGTACCGGGCC-3' (KpnI primer) as 5' primer. In each case, the final products were purified, digested by KpnI and SacI, and inserted between these sites in the pGL3 polylinker. The construct carrying both B1 and Sp1 mutations was similarly obtained by using Sp1-mutated IB promoter as template. The promoter constructs carrying internal deletions or insertions were obtained by the ligation of two separately amplified fragments, one digested by SacI and the other digested by KpnI, to KpnI/SacI-digested pGL3. Thus, plasmid 8-IB was constructed by ligation of two fragments amplified with 5'-CCCCGCCCCAGAGAAATC-3'/SacI primers and KpnI/5'-GGGGAATTTCCAAGCCAGT-3' primers, in the presence of 0.4SK-pGL3 Luc as template. The +5- and +9-IB plasmids were similarly constructed with 5'-TGCAGCTGACCCCGCCCCAGAGAAA-3'/SacI (inserted nucleotides are underlined) and KpnI/5'-GCTCGGGGAATTTCCAAGCCA-3' primers and with 5'-GCAGCATCGCTGACCCCGCCCCAGAGAAA-3'/SacI and KpnI/5'-GCTCGGGGAATTTCCAAGCCA-3' primers, respectively. The correct sequences of the constructs presented were confirmed by DNA sequence analysis. Information concerning PCR is available upon request.

Transfection and luciferase assay. Jurkat T cells were transiently transfected by the DEAE-dextran method (32). One microgram of 0.4SK-pGL3 (wild-type IB promoter) Luc reporter plasmids or mutant 0.4SK-pGL3 (mutB1, mutSp1, and mutB1/Sp1) along with pRL-TK (for transfection normalization; Promega) was resuspended in TS solution (8 mg of NaCl per ml, 0.38 mg of KCl per ml, 0.1 mg of Na₂HPO₄ · 7H₂O per ml, 3.0 ml of Tris, 0.1 mg of MgCl₂ per ml, 0.1 mg of CaCl₂ per ml [pH 7.4]) with 25 µg of DEAE-dextran (Pharmacia). For each transfection, 5 × 10⁵ cells in exponential phase were washed once in 100 µl of TS, resuspended with the DNA solution, and incubated at room temperature for 20 min. Cells were then incubated at 37°C for 30 min in 0.5 ml with medium containing 10% FBS and 0.1 mM chloroquine (Sigma), after which they were centrifuged and resuspended in fresh medium and serum. At 30 h after transfection, cells were induced with TNF- or PMA-PHA. At 16 h after induction, cells were harvested and lysed by 1× passive lysis buffer, and then luciferase activity was analyzed by the Dual-Luciferase reporter assay system (Promega) as specified by the manufacturer. The background obtained from mock-transfected cells was subtracted from each experimental value. The experiments were performed in triplicate in 24-well plates, and the average fold induction was calculated.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Tet-induced TD-IB blocks expression of the IB gene. As shown previously (31), endogenous IB protein expression was abolished after TD-IB induction for 24 h (Fig. 1A, lane 7 to 11). To determine whether IB gene transcription was downregulated by TD-IB, rtTA-IB(2N4) Jurkat cells were treated with TNF- or PMA in the presence or absence of Dox, and endogenous IB and TD-IB mRNAs were analyzed by RNase protection analysis with a 27-nucleotide (nt) 3' cDNA probe that specifically recognized the C terminus of endogenous IB mRNA as well as the truncated IB(2N4) mRNA (Fig. 1B). After 10 min of TNF- or PMA-PHA induction, we detected 20- and 40-fold increases in IB mRNA, respectively (Fig. 1B, lanes 2 and 5); subsequently the level of IB mRNA declined with time in rtTA-IB(2N4) Jurkat cells (Fig. 1B, lanes 4 and 7). In TD-IB-expressing cells, endogenous IB mRNA induction was decreased fivefold by Dox addition, whereas high-level expression of the IB(2N4) transgene was easily detected by RNase protection (Fig. 1B, lanes 8 to 14). These results indicate that Dox-induced TD-IB expression interfered with the induced but not the basal level of endogenous IB mRNA transcription.

View larger version (55K): [in this window] [in a new window]   FIG. 1.   Tet-induced TD-IB inhibits endogenous IB expression. (A) rtTA-Neo (lanes 1 to 6) and rtTA-IB(2N4) (lanes 7 to 12) Jurkat cells were incubated with Dox (1 µg/ml) for 0, 3, 6, 14, 24, and 48 h. Endogenous IB (top arrow) and TD-IB (bottom arrow) were detected by immunoblotting with antibody MAD10B. (B) Schematic representation of C-terminal IB probe used in RNase protection analysis. rtTA-IB(2N4) Jurkat cells were treated with TNF- (10 ng/ml; lanes 2 to 4 and 9 to 11) or PMA (50 ng/ml) plus PHA (1 µg/ml) (lanes 5 to 7 and 12 to 14) for 0, 10 min, 4 h, or 24 h in the absence (lanes 1 to 7) or presence (lanes 8 to 14) of Dox (1 µg/ml, 24 h). Endogenous (wild-type [wt]) IB and TD-IB mRNAs were detected with using the 276-nt 3' cDNA probe by RNase protection assay. Arrows indicate -actin, IB (221-nt band), and IB(2N4) (155-nt band). The results shown are representative of at least three independent experiments.

In vivo genomic footprinting of the IB promoter. To analyze the in vivo occupancy of the IB promoter, genomic footprinting analysis was performed in Jurkat and IB(2N4)-expressing Jurkat cells. Following stimulation by either TNF- or PMA-PHA, living cells were submitted to DMS treatment, which methylates G residues and to a lesser extent A residues; genomic DNA was then extracted and submitted to piperidine treatment to cleave methylated residues. Then, piperidine-cleaved DNA was amplified by LM-PCR using specific primers for IB promoter as detailed in Fig. 2C. A G-specific sequence ladder was also generated as reference and analyzed by sequencing.

View larger version (64K): [in this window] [in a new window]   FIG. 2.   In vivo footprinting of the proximal IB gene promoter in Jurkat T cells. (A) Noncoding strand analysis; (B) coding strand analysis. Naked DNA was treated in vitro with DMS (lane 1). Cells were either nonstimulated (lane 2) or treated with PMA-PHA (lane 3) or by TNF- (lane 4) for 40 min and then were treated with DMS. Genomic DNA was extracted and treated with piperidine. All DNA samples were amplified by LM-PCR and visualized on a Long-Ranger sequencing gel. (C) Sequence of the IB promoter. Major consensus sites for protein binding such as the NF-B sites and sites of Ets-1, Sp1, and AP2 are enclosed in boxes. The mRNA start site and TATA box are also shown. Arrows correspond to primers used in genomic footprinting. Primers 1, 2, and 3 were designed to characterize the noncoding strand; primers 1B, 2B, and 3B were used for the coding strand.

Scanning analysis of the methylation pattern of the 20 to 70 IB promoter region. Since the modifications at the Sp1 site were more discrete than those detected at the B1 site, changes in the methylation pattern of the 20 to 70 region of IB promoter which includes the AP2, Sp1, and B1 sites were analyzed by densitometry scanning (Fig. 3). A representative autoradiograph is presented in Fig. 3A. Comparison of nonstimulated and naked DNA patterns revealed increased methylation of the 42G residue and a slight decrease in 41G methylation at the Sp1 site, whereas no significant differences in methylation were observed for the bordering 20G and 66G residues, indicating an even methylation pattern in the scanned region (Fig. 3B). These data indicate that the Sp1 site was constitutively occupied in resting Jurkat T cells. Similarly, hypermethylation of 24 and 25G residues in nonstimulated Jurkat cells suggests a constitutive binding at the AP2 site. In contrast, no binding was detected on the B1 site in unstimulated conditions. Following stimulation by either TNF-, as shown in Fig. 3C, or PMA-PHA (data not shown), protection of the Sp1 site was modified, as observed by a strong decrease in methylation at residues 36G, 37G, 39G, 41G, 43G, and 44G; only 42G remained methylated. As clearly seen in Fig. 2A, inducible binding at the B1 site is characterized by decreased methylation of 54G, 55G, and 56G as well as a very strong increase in 53G methylation, detected as a broad peak by densitometric scanning (Fig. 3C). Interestingly, methylation of 48G and 49G, which are located between Sp1 and B1 sites, was also significantly decreased after stimulation, indicating that the inducible changes at the B1 and Sp1 sites affect the whole region delimited by these two sites (Fig. 3C). In contrast, no significant modifications of 66G and 21G were observed after stimulation, thus restricting the inducible region to the Sp1 and B1 sites.

View larger version (30K): [in this window] [in a new window]   FIG. 3.   Scanning analyses of In vivo footprinting of the proximal IB gene promoter in Jurkat T cells. Methylation patterns observed on the 20 to 70 region of noncoding strand of IB promoter with naked DNA or from nonstimulated or TNF stimulated cells for 4 h (A) were analyzed by densitometry scanning using a Hewlett-Packard Scan Jet 4c scanner and NIH Image 1.60 software. Comparison of profiles obtained with naked DNA versus DNA from resting Jurkat T cells (B) or with DNA from resting cells versus TNF--stimulated cells (C) corresponds to profiles obtained in at least three independent experiments. Open arrows represent constitutive modifications; filled arrows correspond to inducible changes. Arrows pointing up or down represent increased or decreased methylation on G residues. The sequence of the scanned region, where the methylated G residues are in capital letters, is indicated below each graph.

Footprinting analysis of the IB upstream promoter. The potential role of other upstream NF-B sites that may play a role in IB inducibility (29, 35), including B2 and B3 sites was also evaluated. No significant modification of the patterns was observed in the region corresponding to the B2 site at 319 to 310 or in the putative B site at 34 to 24, located downstream of B1. However, the A residue at position 222 in the B3 site appeared methylated in both unstimulated and TNF--stimulated cells, suggesting constitutive protein binding to this site (data not shown). These data indicate that although other B sites in the IB promoter are recognized in vitro by NF-B complexes (29, 35) and in vivo by constitutive binding complexes, they are not modified in vivo after stimulation of Jurkat T cells. Only the B1 site appears to be targeted by inducible NF-B binding proteins as detected by in vivo genomic footprinting (Fig. 2).

NF-B protection of the B1 site is blocked in TD-IB-expressing cells. Next, control (Neo) Jurkat and TD-IB(2N4)-expressing cells were treated with TNF- or PMA-PHA for 40 min following 24 h of Dox induction (Fig. 4) and then subjected to genomic footprint analysis using noncoding specific primers 1, 2, and 3. In control cells, the 4G ladder was easily identified (Fig. 4, lanes 1 to 3), and following TNF- or PMA-PHA addition, the characteristic hypersensitive cleavage of 53G was detected (lanes 4 to 7). In IB(2N4)-expressing cells, TD-IB induction resulted in complete inhibition of inducible binding complexes to the B1 site in a Dox-dependent manner; TNF- or PMA-PHA stimulation in the absence of Dox-induced TD-IB resulted in a footprint at the B1 site that was indistinguishable from that of stimulated Jurkat cells (lanes 4 to 7, 10, 12, and 14), whereas in the presence of Dox-induced TD-IB, the observed footprint pattern resembled that for unstimulated control Jurkat cells (lanes 2, 3, 8, 9, 11, and 13). Although less clearly resolved, the pattern of methylation and cleavage of the adjacent Sp1 site was also sensitive to Dox induction. For example, in TD-IB-inducible cells, PMA-PHA treatment resulted in protection of the Sp1 site with the exception of 42G (Fig. 4, lane 12), whereas Dox induction resulted in a pattern of protection at the Sp1 site that was characteristic of unstimulated control Jurkat cells (Fig. 4, lane 13; compare with lanes 2, 3, and 8). These results further indicate that binding of complexes to the Sp1 site may be coordinately regulated by the adjacent B1 site in a TD-IB-inducible manner.

View larger version (76K): [in this window] [in a new window]   FIG. 4.   In vivo footprint of the IB gene promoter in rtTA-Neo and Tet-inducible TD-IB-expressing cells. Noncoding strand analysis was performed with rtTA-Neo (lanes 2 to 7) or rtTA-IB(2N4) (lanes 8 to 14) Jurkat cells, either unstimulated (lanes 2, 3, 8, 9, and 14) or stimulated by TNF- (10 ng/ml; lanes 4, 5, 9, and 10) or PMA-PHA (50 ng/ml and 1 µg/ml, respectively; lanes 6, 7, 12, and 13) for 40 min. Naked DNA was methylated in vitro (lanes 1). Where indicated (+), cells were pretreated with Dox (1 µg/ml, 24 h).

Prolonged NF-B binding and temporal switch in the composition of NF-B complexes at the B1 site. To determine the kinetics of the in vivo occupancy of B1 and Sp1 sites of the IB promoter, control Neo (Fig. 5A) and TD-IB-expressing (Fig. 5B) cells were analyzed at different times after stimulation by TNF- or PMA-PHA. Surprisingly, the same protection of the B1 site was observed from 10 min to 24 h following TNF- or PMA-PHA treatment in both cell types (Fig. 5A, lanes 4 to 15; Fig. 5B, lanes 4, 6, 8, 10, 12, and 14). In TD-IB-expressing cells pretreated with Dox, the pattern of methylation and cleavage of the B1 site remained characteristic of unstimulated cells (Fig. 5B, lanes 5, 7, 9, 11, 13, and 15), regardless of the time of TNF- or PMA-PHA stimulation. Prolonged protection of the adjacent Sp1 site was also observed from 10 min to 24 h in stimulated control or TD-IB-expressing cells (Fig. 5A, lanes 4, 5, 7, 10, 12, 14, and 15; Fig. 5B, lanes 4, 8, and 12). Again, Dox induction of TD-IB expression resulted in a methylation pattern at the Sp1 site that was characteristic of unstimulated cells (Fig. 5B, lanes 5, 7, 9, and 13).

View larger version (96K): [in this window] [in a new window]   FIG. 5.   Kinetics of protein-DNA interactions on the proximal IB promoter. For noncoding strand analysis of rtTA-Neo (A) and rtTA-IB(2N4) (B) Jurkat T cells, naked DNA was methylated in vitro (lane 1). Cells were either unstimulated (lanes 2 and 3) or stimulated with TNF- (10 ng/ml; lanes 4 to 9) or PMA-PHA (50 ng/ml and 1 µg/ml, respectively; lanes 10 to 15). The time of cell harvesting following TNF- or PMA-PHA stimulation is indicated above the lanes. Where indicated (+), cells were pretreated with Dox (1 µg/ml, 24 h).

View larger version (57K): [in this window] [in a new window]   FIG. 6.   Switch in the composition of NF-B complexes bound to the B1 site. Mobility shift and supershift analyses were performed with nuclear extracts from rtTA-Neo Jurkat cells treated with PMA-PHA for the times indicated (lanes 1 to 4), or analyzed by using anti-p50 (lanes 5 to 7), anti-p65 (lanes 8 to 10), and anti-c-Rel (lanes 11 to 13) antibodies with a [-32P]ATP labeled B1 probe. The NF-B and supershifted (S.S.) complexes are indicated.

p65 and Sp1 bind together to the B1/Sp1 site of IB promoter. To assess Sp1 binding to the NF-B/Sp1 region, EMSA analysis was performed with probes encompassing the 44 to 36 Sp1 site, the 63 to 36 B1 site, or both B1 and Sp1 sites (B1/Sp1 probe from the 65 to 34 region of the IB promoter). PMA treatment of Jurkat cells resulted in a 10-fold increase in the intensity of the Sp1 binding complex (Fig. 7A, top panel, lanes 1 and 2); complex formation was blocked in competition reactions by both Sp1 and B1/Sp1 binding sites (Fig. 7A, top panel, lanes 3 and 7) but not by the B1 site alone (Fig. 7A, lane 6). This complex was further identified as an Sp1 binding activity by its competition in the presence of the consensus Sp1 sequence, whereas only partial inhibition was observed when a mutated Sp1 sequence was used as competitor (Fig. 7A, top panel, lanes 4 and 5). As a control, no inhibition of NF-B binding to B1 was detected in the presence of the Sp1 site of the IB promoter, consensus Sp1 or mutated Sp1 sequences (Fig. 7A, middle panel, lanes 2 to 5). Identification of the complexes detected by B1/Sp1 probe was further determined by supershift analysis (Fig. 7B). As expected, anti-p50 and anti-p65 antibodies abolished p65/p50 binding to B1/Sp1 probe, whereas anti-Sp1 antibodies reacted against the Sp1-containing complex (Fig. 7B, lanes 3 to 5). The faster-migrating band observed with the B1/Sp1 probe (Fig. 7A, bottom panel, lanes 1, 2, 5, and 6) is likely to be a degradation product of Sp1-containing complex generated during nuclear protein extraction; this band is competed by Sp1-related oligonucleotides (Fig. 7A, lanes 3, 4, and 7) but is not affected by anti-Sp1 antibodies (Fig. 7B, lane 5). This EMSA analysis failed to reveal a complex formed by both endogenous Sp1 and NF-B bound to the B1/Sp1 probe.

View larger version (43K): [in this window] [in a new window]   FIG. 7.   NF-B and Sp1 bind to the 63 to 36 region of the IB promoter. (A) EMSA analysis was performed with radiolabeled oligonucleotide probes specific to Sp1 (top panel), B1 (middle panel), and B1/Sp1 (bottom panel). Nuclear extracts prepared from Jurkat cells were either unstimulated (lane 1) or treated with PMA (50 ng/ml) for 2 h (lanes 2 to 7). Competition was performed in the presence of a 125-fold excess of unlabeled oligonucleotide: Sp1 site of IB promoter (lane 3), Sp1 consensus (cons. Sp1; lane 4), mutant Sp1 (mut. Sp1; lane 5), B1 (lane 6), or B1/Sp1 (lane 7). To facilitate detection of simultaneous binding of NF-B and Sp1, EMSA buffer conditions were modified as described in Materials and Methods and the amount of extract used in the binding reactions was varied between 150 ng and 3 µg; for the binding reactions shown, 150 ng of Jurkat nuclear extract and 500 ng of poly(dI-dC) were used. (B) Complex composition was analyzed by supershift analysis. PMA-induced nuclear extracts were incubated with anti-p65 (lane 3), anti-p50 (lane 4), and anti-Sp1 (lane 5) antibodies.

View larger version (61K): [in this window] [in a new window]   FIG. 8.   NF-B and Sp1 can co-occupy the B1/Sp1 site of the IB promoter. (A) EMSA was performed with recombinant p65 (GST-Np65), p50 (GST-p50), and Sp1 proteins, using radiolabeled B1/Sp1 site [-32P]ATP labeled. Each recombinant p65 (2 ng), p50 (2.73 ng), and Sp1 (3 ng) was used alone in lanes 2 to 4. Increasing amounts of p65 (0.5, 1, and 2 ng) combined with increasing amounts of p50 (0.65, 1.3, and 2.73 ng) were incubated in the absence (lanes 5 to 7) or presence (lanes 8 to 10) of Sp1 (3 ng). Increasing amounts of Sp1 (1, 3, and 5 ng) were also tested with the same amount of p65 (1 ng) and p50 (1.3 ng) in lanes 11 to 13. Shifted complexes are indicated by arrows. (B) Combinations of recombinant p65 (2 ng), p50 (2.73 ng), and Sp1 (3 ng) proteins were incubated with radiolabeled B1/Sp1 oligonucleotides in the presence or absence of specific antibodies for p65, p50, and Sp1 proteins. The composition of the complexes was analyzed by supershift analysis using anti-p65, anti-p50, or anti-Sp1 antibodies (lanes 2 to 4).

IB gene expression is dependent on both NF-B and Sp1 binding. We next examined the functional role of the NF-B and Sp1 sites in IB gene transcription in Jurkat cells by transient cotransfection with luciferase reporter constructs driven by the wild-type IB promoter (0.4SK) (35) or by mutated versions of the IB promoter (Fig. 9). Treatment of transfected Jurkat cells with TNF or PMA-PHA resulted in 4- and 7-fold stimulation of gene activity, respectively. Deletion or point mutation of the B1 site of the IB promoter (B and mutB1) abrogated TNF- and PMA-PHA-induced gene activation relative to the wild-type promoter (Fig. 9A). Strikingly, point mutation of the Sp1 site (mutSp1) also dramatically decreased induction of IB gene expression and also slightly decreased basal-level promoter activity. As expected, mutation of both B1 and Sp1 sites also completely inhibited gene activity. As shown in Fig. 9B, EMSA analysis demonstrated that impairment of transactivation was due to lack of NF-B binding (Fig. 9B, lanes 4 to 6) or Sp1 binding (Fig. 9B, lanes 7 to 9) to B1/Sp1 sites. From these results, we conclude that both B1 and Sp1 sites are required for full induction of IB promoter. To further analyze whether activation requires direct contact between NF-B and Sp1, mutant luciferase reporter plasmids containing deletions or insertions between B1 and Sp1 sites were tested (Fig. 9C). Interestingly, the introduction of 5 or 9 nt between the B1 and Sp1 sites which alters the helical relationship between the two sites decreased but did not eliminate IB inducibility (Fig. 9C). When 8 nt (8) between B1 and Sp1 sites were deleted, transcriptional inducibility was completely abolished. Together, these results indicate that both B1 and Sp1 sites are required for full induction of IB promoter.

View larger version (46K): [in this window] [in a new window]   FIG. 9.   Both B and Sp1 are required for full TNF-- or PMA-induced IB promoter activity. (A) Jurkat cells were transfected with 1 µg of luciferase reporter plasmid containing wild-type (0.4SK), mutant B1 (B or mutB1), mutant Sp1 (mutSp1), or mutant B1/Sp1 (mutB1/Sp1) IB promoter. Twenty four hours after transfection, cells were treated with TNF- (10 ng/ml) or PMA (50 ng/ml) or left untreated for an additional 16 h. Transfection efficiency was normalized to that of Renilla luciferase (see Materials and Methods). The experiments were performed in triplicate, and the average fold induction was calculated. (B) EMSA was performed with different radiolabeled oligonucleotide probes (B1/Sp1, mutated B1/Sp1 [mutB1], and B1/mutated Sp1 [mutSp1]), using uninduced or PMA (50 ng/ml for 2 h)-induced Jurkat nuclear extract. (C) Luciferase assays were performed as described for panel A, in using 1 µg of reporter plasmids containing an 8-nt deletion (8), 5-nt addition (+5), and 9-nt addition (+9) between B1 and Sp1 sites of the IB promoter. Transfection efficiency was normalized to that of Renilla luciferase (see Materials and Methods). The experiments were performed in triplicate, and the average fold induction was calculated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This report presents, for the first time, an in vivo genomic footprinting analysis of the IB promoter and characterizes the autoregulatory control of IB transcription by an NF-B/Sp1 transcriptional switch. In previous studies, TD-IB expression was shown to inhibit endogenous IB at the protein level as well as to interfere with NF-B binding and HIV-1 multiplication in Jurkat cells (31). We now demonstrate that induction of endogenous IB after TNF- or PMA-PHA treatment is suppressed by TD-IB at the transcriptional level. Tet-induced TD-IB expression blocked NF-B binding activity at the proximal 63 to 53 B1 site of the IB promoter. In vivo genomic footprinting revealed multiple protein-DNA interactions in the region of the IB promoter between 250 to +100 bp in Jurkat T cells; protection of Sp1, AP2, Ets-1, and B3 sites in unstimulated cells indicates that these sites participate in basal-level IB transcription. In response to stimulation of Jurkat T cells by PMA-PHA or TNF-, changes in methylation of the B1 site and the adjacent Sp1 site were observed, whereas no inducible changes were detected at B3 or other sites (data not shown). The protection observed at B1 and Sp1 sites was sustained from 10 min to 24 h and together with the EMSA results demonstrated that binding of p50-p65 heterodimer correlated with early transcriptional induction of the IB gene; at later times, the switch in composition of the NF-B complexes to predominantly p50-c-Rel heterodimers correlated with transcriptional downregulation. Deletion and point mutagenesis demonstrated that both B1 and Sp1 sites were absolutely required for IB promoter induction, whereas only Sp1 was involved in basal transcription of this promoter; a strict spacing requirement between B1 and Sp1 sites was also essential for full activation of the IB promoter. Together, these studies suggest a model for IB transcriptional regulation in Jurkat T cells, as summarized in Fig. 10. Early activation of IB promoter activity is accompanied by p65-p50 binding to the B1 site of the promoter, as well as by modulation of Sp1 binding at the adjacent Sp1 site. Downregulation of IB transcription, occurring at later times after induction, is associated with a switch in the composition of NF-B complexes, from p50-p65 to p50-c-Rel heterodimers. This mechanism is in agreement with inhibition of p65-mediated transcription of HIV-1 LTR and interleukin-2 receptor alpha-chain promoters by c-Rel (20). Furthermore, c-Rel is induced with delayed kinetics compared to p65 (20) and may inhibit IB transcription by competition with p50 and p65 for occupancy of the B1 binding site (Fig. 10).

View larger version (35K): [in this window] [in a new window]   FIG. 10.   Schematic representation of the protein-DNA interactions regulating the IB promoter. IB promoter organization, including B1 to B3 as well as Sp1, Ets-1, and AP2 binding sites, is shown at the top. In resting Jurkat T cells, protection is observed at the B3, Ets-1, Sp1, and AP2 sites, which likely contribute to basal transcription. Early after induction by PMA-PHA or TNF-, the B1 site (63 to 53) is occupied by p65-p50 heterodimers. At later times, a switch in complex composition to p50-c-Rel heterodimers correlates with downregulation of IB transcription. Protection is also observed at the adjacent Sp1 site (44 to 36) and is also modulated by inducer-mediated stimulation or by activation of Tet-induced TD-IB expression which sequesters p65 in the cytoplasm. Binding to the Sp1 site may be related to the occupancy of the B1 site by inducible NF-B complexes. No changes were observed on B3, Ets-1, and AP2 sites during cell activation.

The EMSA and genomic footprinting data are consistent with inhibition of IB promoter activity identified by deletion of the B1 site (35); the B2 and B3 sites appear to play no role in the inducibility of the IB promoter at least in Jurkat T cells stimulated by PMA-PHA or TNF-. Our results are also in agreement with the mutagenesis analysis performed by Ito et al. (29), showing a predominant role for the B1 site. This analysis had also suggested that full activation of the IB promoter also required another B-like site located downstream of B1 between nt 34 and 24, as well as the upstream B2 site. In the present study, no inducible in vivo protein-DNA interactions were observed at either of these sites. Although a role for B2 and B3 sites cannot be excluded, our in vivo data clearly demonstrate that B1 and Sp1 sites play the major role in the inducibility of the IB promoter in Jurkat T cells. Furthermore, in vivo genomic footprinting experiments performed with the U937 promonocytic cell line revealed the same in vivo protection pattern as observed with Jurkat T cells: only the Sp1 site was protected before stimulation and both Sp1 and B1 sites were targeted by inducible complexes after induction (data not shown). Thus, a common mechanism of IB regulation involving the NF-B/Sp1 transcriptional switch is likely active in multiple cell types, including T cells and monocytes/macrophages.

Many genes regulated by NF-B also contain adjacent Sp1 sites, and direct interaction between NF-B proteins and Sp1 has been demonstrated (45); a recent study also identified in vitro binding of Sp1 to the B sites located on promoters such as the interleukin-6 and P-selectin (28). We have demonstrated binding of p50-c-Rel and p50-p65 heterodimers to the B1 site in response to cell induction, as well as binding of Sp1 to its own site in IB promoter. Furthermore, the Sp1 protection observed was reversed with TD-IB activation (Fig. 10). This coordinate change suggests that the binding of NF-B inducible complexes to the B1 site may also facilitate an increased Sp1 binding affinity to the adjacent Sp1 site. Interestingly, the residues at positions 48 and 49, located between the NF-B and Sp1 sites, became more sensitive to methylation and cleavage, suggesting that the protein-DNA conformation of the entire NF-B/Sp1 region is modified after stimulation (Fig. 3).

Sp1 binding to its consensus site prior to stimulation indicates that it may contribute to basal transcription of the IB gene. Interestingly, a different methylation and cleavage pattern was observed at the Sp1 site after stimulation. Direct Sp1 conformational changes and/or alterations in Sp1 binding affinity may be induced after stimulation via Sp1 posttranslational modification. Sp1 has been described as a zinc finger phosphoprotein which upon cell activation undergoes specific phosphorylations and dephosphorylations that regulate its DNA binding activity and its interactions with other proteins (4, 51, 54). Following cell stimulation, Sp1 can be phosphorylated by casein kinase II, by protein kinase A, and by a recently described 60-kDa kinase, activated in response to Neu differentiation factors (2). The inducible change at the Sp1 binding site in response to Jurkat T-cell stimulation may reflect such Sp1 modifications leading to increased binding on DNA.

The critical role of B1 site and the adjacent Sp1 site in the inducibility of the IB promoter is further supported by the conservation of these two sites in the murine and porcine homologs of the IB promoter (17, 18). Moreover, not only are the exact sequences conserved, but also the 10-bp spacing between both sites is maintained between species. This distance, corresponding to one helical turn of DNA, may permit a physical interaction between proteins bound to the Sp1 site and the p65-p50 complex on the same face of chromatin in vivo, as shown for the HIV-1 LTR promoter (45). Although no cooperativity in the binding of NF-B and Sp1 was observed by EMSA, transfection studies using hybrid IB promoters in which nucleotides between B1 and Sp1 sites were inserted or deleted revealed a strict spacing requirement for maximal inducibility of IB promoter. Deletion of the 8 nt between both sites or insertion of 5 or 9 nt significantly lowered IB promoter inducibility. The fact that addition of an half helical turn or a complete helical turn led to a similar decrease in IB gene inducibility argues against a direct physical interaction between NF-B and Sp1 and suggests instead a requirement for the interaction of NF-B and/or Sp1 with basal transcription factors such as TATA binding protein (TBP)-associated factors or with the transcription machinery for maximal induction of IB promoter. Sp1 has been found to associate with CBP/p300 in a multiprotein complex (44), and the NF-B p65 subunit is able to interact with the amino-terminal region of the coactivator p300, resulting in gene activation of E-selectin and VCAM-1 (25). Further studies are required to characterize the association of NF-B and Sp1 with TBP-associated factors, TBP, or CBP/p300 and their role in IB regulation.