High Mobility Group-I(Y) Protein Facilitates Nuclear Factor-κB Binding and Transactivation of the Inducible Nitric-oxide Synthase Promoter/Enhancer*

Nitric oxide (NO), a free radical gas whose production is catalyzed by the enzyme NO synthase, participates in the regulation of multiple organ systems. The inducible isoform of NO synthase (iNOS) is transcriptionally up-regulated by inflammatory stimuli; a critical mediator of this process is nuclear factor (NF)-κB. Our objective was to determine which regulatory elements other than NF-κB binding sites are important for activation of the iNOS promoter/enhancer. We also wanted to identify transcription factors that may be functioning in conjunction with NF-κB (subunits p50 and p65) to drive iNOS transcription. Deletion analysis of the iNOS promoter/enhancer revealed that an AT-rich sequence (−61 to −54) downstream of the NF-κB site (−85 to −76) in the 5′-flanking sequence was important for iNOS induction by interleukin-1β and endotoxin in vascular smooth muscle cells. This AT-rich sequence, corresponding to an octamer (Oct) binding site, bound the architectural transcription factor high mobility group (HMG)-I(Y) protein. Electrophoretic mobility shift assays showed that HMG-I(Y) and NF-κB subunit p50 bound to the iNOS promoter/enhancer to form a ternary complex. The formation of this complex required HMG-I(Y) binding at the Oct site. The location of an HMG-I(Y) binding site typically overlaps that of a recruited transcription factor. In the iNOS promoter/enhancer, however, HMG-I(Y) formed a complex with p50 while binding downstream of the NF-κB site. Furthermore, overexpression of HMG-I(Y) potentiated iNOS promoter/enhancer activity by p50 and p65 in transfection experiments, suggesting that HMG-I(Y) contributes to the transactivation of iNOS by NF-κB.

Nitric oxide (NO) is a free radical gas that participates in the physiologic or pathophysiologic regulation of multiple organ systems (1)(2)(3). Production of NO from its substrate, L-arginine, is catalyzed by NO synthase (NOS) 1 (4,5). Three isoforms of NOS (NOS 1-3) are present in mammalian cells, each encoded by a unique gene. The high output pathway of NO production is catalyzed by NOS2, a transcriptionally regulated gene that is induced after immunologic or inflammatory stimuli. We refer to this inducible NOS2 isoform as iNOS. Although iNOS was originally identified and characterized in macrophages (6 -8), it is present in numerous cell types including vascular smooth muscle cells (9 -11).
Much of the initial evaluation of the mouse iNOS gene focused on its transcriptional regulation in macrophages after lipopolysaccharide (LPS) and interferon (IFN)-␥ stimulation (12,13). Dissection of the iNOS promoter/enhancer revealed that a downstream nuclear factor (NF)-B site (Ϫ85 to Ϫ76, NF-Bd) is critical for activation of iNOS by LPS in macrophages. In the presence of LPS, c-Rel or Rel A (p65) binds with p50 to form heterodimers on the iNOS promoter/enhancer, in conjunction with additional unidentified proteins (14). Further studies revealed that a more upstream region of the iNOS promoter/enhancer (Ϫ951 to Ϫ911) is responsible for the synergistic induction of iNOS by IFN-␥ and LPS. IFN regulatory factor (IRF)-1 binding to an IRF binding site (IRF-E) (15) and Stat1␣ binding to an IFN-␥-activated site (16) contribute to optimal induction of iNOS by IFN-␥ and LPS.
Interleukin (IL)-1␤ and tumor necrosis factor-␣ (important pro-inflammatory cytokines generated after LPS stimulation in vivo; Refs. 17 and 18) are potent activators of iNOS transcription in vascular smooth muscle cells (10,11,19). We have shown that the downstream NF-Bd site (Ϫ85 to Ϫ76) is important for proinflammatory cytokine induction of the iNOS promoter/enhancer in vascular smooth muscle cells (20). However, elements other than this NF-Bd site in the downstream portion of the iNOS 5Ј-flanking sequence (Ϫ234 to ϩ31) also appear to be crucial for full activation of iNOS (20). We designed the present study to further elucidate the important regulatory elements in region Ϫ234 to ϩ31 responsible for iNOS induction in vascular smooth muscle cells by the proinflammatory cytokine IL-1␤. Goldring and colleagues (21) demonstrated by in vivo foot-print analysis in macrophages that, in addition to the NF-B sites, nuclear protein binding occurred after LPS stimulation at NF-IL6 (Ϫ150 to Ϫ142) and octamer (Oct) (Ϫ61 to Ϫ54) sites of the iNOS promoter/enhancer (21). Their report revealed potential binding sites in region Ϫ234 to ϩ31 of the iNOS promoter/ enhancer but lacked a detailed functional analysis of these sites. Until our present study, the functional importance of these sites in vascular smooth muscle cells had not been elucidated. Furthermore, there had been no identification of nuclear proteins (binding in the downstream region of the iNOS promoter/enhancer) that facilitate iNOS transactivation by NF-B. Nuclear proteins that interact with members of the NF-B family include the nonhistone chromosomal proteins of the high mobility group (HMG)-I(Y) family (22). HMG-I(Y) proteins play a role in the transcriptional regulation of certain mammalian genes whose promoter/enhancer regions contain AT-rich sequences (22)(23)(24).
HMG-I(Y) refers to two proteins, HMG-I and HMG-Y, that are alternatively spliced products of the same gene (25). HMG-I(Y) binds to AT-rich regions in the minor groove of DNA (26), and it is known to bind Oct sequence motifs (27). HMG-I(Y) facilitates the assembly of functional nucleoprotein complexes (enhanceosomes) by modifying DNA conformation and by recruiting nuclear proteins to an enhancer (28,29). The role of HMG-I(Y) in enhanceosome assembly has been studied extensively in the IFN-␤ gene after viral stimulation (28 -33). HMG-I(Y) has been shown to enhance the binding of transcription factors, such as NF-B and activating transcription factor-2, to their binding sites by DNA-protein and protein-protein interactions (28 -33). Because of the aforementioned properties of HMG-I(Y), we determined whether this protein bound to the iNOS promoter/enhancer and interacted with NF-B subunits to regulate iNOS gene transcription. We also determined whether the site of HMG-I(Y) binding overlapped a binding site for NF-B (as occurs with IFN-␤), or if HMG-I(Y) bound at a site different from NF-B.

EXPERIMENTAL PROCEDURES
Materials-Salmonella typhosa LPS (Sigma) was dissolved in 0.9% saline and stored at Ϫ20°C. Recombinant human IL-1␤ (Collaborative Biomedical, Bedford, MA) was stored at Ϫ80°C until use. Recombinant human NF-B subunit p50 (Promega Corp., Madison, WI) was also stored at Ϫ80°C until use. A goat polyclonal antibody to p50 (Santa Cruz Biotechnology, Santa Cruz, CA) was used for supershift experiments.
Plasmids-Plasmid pGL2-Basic contained the firefly luciferase gene without any promoter (Promega, Madison, WI). Reporter constructs containing fragments of the mouse iNOS 5Ј-flanking sequence were named according to the location of the fragment from the transcription start site in the 5Ј and 3Ј directions. A 1516-base pair (bp) fragment amplified from mouse genomic DNA, containing 1485 bp of the 5Јflanking region and the first 31 bp after the transcription start site, was named iNOS(Ϫ1485/ϩ31), as described (20). A shorter fragment, iNOS(Ϫ234/ϩ31), was generated by polymerase chain reaction from iNOS(Ϫ1485/ϩ31) as described (20). The downstream NF-Bd site was mutated (Ϫ85 to Ϫ83, GGG to CTC) in the Ϫ1485 to ϩ31 fragment by using a site-directed mutagenesis technique (20), and this construct was named iNOS(Ϫ1485/ϩ31 NF-Bm).
Transfections-RASMC were transfected by a DEAE-dextran method (20). In brief, 500,000 cells were plated onto 100-mm tissue culture dishes and allowed to grow for 48 -72 h (until 80 -90% confluent). Then iNOS luciferase constructs and pOPRSVI-CAT (to correct for differences in transfection efficiency) were added (5 g each) to the RASMC in a solution containing 500 g/ml DEAE-dextran. RASMC were subsequently shocked with 5% dimethyl sulfoxide solution for 1 min and then allowed to recover in medium containing 10% heatinactivated FBS. Twelve hours after transfection, RASMC were placed in 2% FBS. RASMC were then stimulated with vehicle, human recombinant IL-1␤ (10 ng/ml), or LPS (1 g/ml) for 48 h. The doses of IL-1␤ and LPS, and the duration of stimulation, were chosen on the basis of pilot experiments (data not shown).
Rat alveolar macrophages (NR8383) were also transfected by the same DEAE-dextran method (20), with the exception that they were treated in a floating suspension (the RASMC were attached to culture dishes). Twelve hours after transfection, the macrophages were stimulated with LPS (1 g/ml) for 48 h. In both the RASMC and macrophage transfection experiments, cell extracts were prepared by detergent lysis (Promega), and luciferase activity was measured with an AutoLumat LB953 luminometer (EG&G, Gaithersburg, MD) and the Promega luciferase assay system. To evaluate the efficiency of transfection, we performed a CAT assay by a modified two-phase fluor diffusion method as described (40,41). The ratio of luciferase to CAT activity in each sample served as a measure of normalized luciferase activity. SL2 cells were transfected by the calcium-phosphate method according to Di Nocera and Dawid (42). In brief, SL2 cells were plated in six-well tissue culture dishes (Costar Corp, Cambridge, MA) 24 h before transfection. Transfection was then performed in six separate wells for each condition. iNOS plasmids were added at 1 g/well. Plasmids p50-pPAC, p65-pPAC, and phsp82LacZ were added at 100 ng/well. Plasmid pPACHMGI was added at 1 g/well, alone or in combination with p50-pPAC or p65-pPAC (or both). The expression plasmid doses were chosen on the basis of pilot experiments (data not shown). Fortyeight hours after the initial transfection, extracts from the SL2 cells were prepared and luciferase activity was measured as described for RASMC. ␤-Galactosidase assays were performed as described elsewhere (43). The ratio of luciferase activity to ␤-galactosidase activity in each sample served as a measure of normalized luciferase activity.
Protein Expression and Purification-The prokaryotic expression plasmid for mouse HMG-I(Y), pRSETHMG-I(Y), was prepared as described (44). Plasmid pRSETHMG-I(Y) (containing the HMG-I(Y) protein, a vector-derived polyhistidine tag, and an enterokinase cleavage site) was transferred into Escherichia coli strain BL21(DE3)pLysS. Expression was induced in culture (mid-log phase) by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside. Three hours after induction of expression, the bacteria were lysed and HMG-I(Y) was purified by cobalt affinity chromatography under denaturing conditions according to the instructions of the manufacturer (CLONTECH). Proteins were renatured by dialysis overnight against 20 mM HEPES (pH 7.8), 20 mM KCl, and 0.2% Tween 20 at 4°C. The dialysate was stored at Ϫ80°C until use.
Probes were also generated with mutations in the Ϫ85 to Ϫ76 NF-Bd site (TGCTCCAGAGGGCTTTGGGAACAGTTATGCAAAATA) and the Ϫ61 to Ϫ54 Oct site (TGGGGACTCTCCCTTTGGGAACAGTTCG-TACCCCTA). Prior to annealing, polynucleotide kinase (Boehringer Mannheim) was used to label the oligonucleotides with [␥-32 P]ATP. A typical binding reaction contained 20,000 cpm DNA probe, 10 mM Tris-HCl (pH 7.5), 50 mM KCl, 0.1 mM EDTA, 250 g/ml acetylated bovine serum albumin, 1 mM dithiothreitol, 5% glycerol, 500 ng of poly(dG-dC⅐dG-dC), and recombinant HMG-I(Y) and/or p50 protein in a final volume of 25 l. The reaction mixture was allowed to incubate for 20 min at room temperature. The DNA-protein complexes were then fractionated on a 5% native polyacrylamide electrophoretic gel in a 0.25ϫ Tris borate-EDTA recirculating buffer system at 4°C.
Statistics-Data from the SL2 cell transfection experiments were subjected to analysis of variance followed by Scheffe's test. Significance was assumed at p Ͻ 0.05.

A Downstream Oct Binding Site Is Pivotal to Induction of the iNOS Promoter/Enhancer by IL-1␤ and LPS in Vascular
Smooth Muscle Cells-We have demonstrated elsewhere that the IL-1␤-responsive elements in iNOS reside between bp Ϫ234 and ϩ31 of the 5Ј-flanking sequence (20). Our previous data also suggested that elements other than NF-Bd in this downstream region contributed to activation of the iNOS promoter/enhancer by IL-1␤ in vascular smooth muscle cells. To reveal these other IL-1␤-responsive elements in the iNOS 5Јflanking sequence, we generated deletion constructs containing a mutated NF-Bd site. This approach ensured that any induction of iNOS promoter/enhancer activity after transfection of the constructs into vascular smooth muscle cells would not be the result of nuclear protein binding to the NF-Bd site. Beyond a decrease in iNOS promoter/enhancer activity after mutation of the NF-Bd site (Fig. 1), no further reduction in IL-1␤ responsiveness occurred, even after the iNOS 5Ј-flanking sequence had been reduced to a construct containing bases Ϫ69 to ϩ31 of the downstream promoter. Other than the TATA box, an AT-rich Oct site (Ϫ61 to Ϫ54) remained in this portion of the iNOS 5Ј-flanking sequence.
Using site-directed mutagenesis, we generated constructs of the downstream iNOS 5Ј-flanking sequence that contained no mutations (iNOS(Ϫ234/ϩ31)), a mutation at the NF-Bd site (iNOS(Ϫ331/ϩ31 NF-Bm)), a mutation at the Oct site (iNOS(Ϫ331/ϩ31 OCTm)), or mutations at both sites (iNOS(Ϫ331/ϩ31 NF/OCTm)). These constructs were transfected into RASMC and stimulated with vehicle or IL-1␤. Mutation of the NF-Bd site produced a 69% reduction in iNOS promoter/enhancer activity after stimulation with IL-1␤ (Fig.  2). Furthermore, mutation of the Oct site led to an even greater reduction (90%) in iNOS activity after IL-1␤ stimulation. Mutating both sites did not cause iNOS promoter/enhancer activity to fall below the level obtained by mutating the Oct site alone. Taken together, the data in Fig. 2 suggest that the Oct binding site (Ϫ61 to Ϫ54) is critical for induction of iNOS promoter/enhancer activity by IL-1␤ in vascular smooth muscle cells. The absence of a further reduction in iNOS promoter/ enhancer activity after mutation of both sites suggests that there may be an interaction between nuclear proteins that bind at the Oct site and nuclear proteins that bind at the NF-Bd site. There were no significant differences in promoter/enhancer activity among iNOS constructs that received vehicle alone.
To determine if this Oct site was important for iNOS induction by inflammatory stimuli other than IL-1␤, we transfected constructs iNOS(Ϫ234/ϩ31), iNOS(Ϫ331/ϩ31 NF-Bm), iNOS(Ϫ331/ϩ31 OCTm), and iNOS(Ϫ331/ϩ31 NF/OCTm) into RASMC and stimulated the cells with LPS. LPS and IL-1␤ had an almost identical effect on iNOS promoter/enhancer activity. Mutation of the NF-Bd site caused a significant reduction (72%) in iNOS promoter/enhancer activity after LPS stimulation; activity was reduced even further (87%) in the construct containing a mutated Oct site (Fig. 3A). Again, activity after mutation of both sites was not different from activity after mutation of the Oct site alone. We also transfected these constructs into alveolar macrophages and stimulated the cells with LPS. In comparison with iNOS promoter/enhancer activity in vascular smooth muscle cells, mutation of the NF-Bd site produced a more dramatic reduction in macrophages (85%) after LPS stimulation (Fig. 3B). Mutation of the Oct site again caused a dramatic reduction (92%) in iNOS promoter/enhancer activity after LPS stimulation. These experiments demonstrate that binding of nuclear proteins at the downstream Oct site is important for activation of the iNOS promoter/enhancer in both vascular smooth muscle cells and macrophages, and that this site is important for iNOS activation after stimulation with two mediators of inflammation, IL-1␤ and LPS.

HMG-I(Y) Binds to the iNOS 5Ј-Flanking Sequence at the Downstream Oct Binding Site-To determine if HMG-I(Y)
could bind to the iNOS promoter/enhancer region containing the NF-Bd and Oct sites, we performed EMSA with a radiolabeled probe encoding region Ϫ87 to Ϫ52 of the iNOS 5Јflanking sequence. Incubation of recombinant HMG-I(Y) with the probe containing intact NF-Bd and Oct sites resulted in a DNA-protein complex (Fig. 4A). This complex was specific because a 500-fold molar excess of unlabeled identical oligonucleotide, but not unrelated oligonucleotide, competed for HMG-I(Y) binding and abolished the DNA-protein complex. To localize the site of HMG-I(Y) binding, we incubated recombinant HMG-I(Y) with a radiolabeled probe containing intact NF-Bd and Oct sites, a mutated NF-Bd site, or a mutated Oct site. Like the wild-type probe containing intact binding sites, the probe containing a mutated NF-Bd site was able to bind to HMG-I(Y) (Fig. 4B). Mutation of the AT-rich Oct site, however, resulted in a marked reduction in HMG-I(Y) binding. In addition, a probe containing a mutated region between the NF-Bd and Oct sites did not disrupt HMG-I(Y) binding (data not shown). These data suggest that binding of HMG-I(Y) within region Ϫ87 to Ϫ52 of the iNOS promoter/enhancer occurs at the Oct binding site.

HMG-I(Y) and NF-B Subunit p50 Bind to the iNOS Promoter/Enhancer and Form a Ternary Complex-EMSA
were performed with a radiolabeled probe encoding region Ϫ87 to Ϫ52 of the iNOS 5Ј-flanking sequence (containing intact NF-Bd and Oct sites) and recombinant p50 (an important DNA binding subunit of NF-B; Refs. 45 and 46), in the absence or presence of recombinant HMG-I(Y). Incubating the probe with p50 resulted in a slowly migrating DNA-protein complex whose intensity increased with increasing concentrations of p50 (Fig.  5). A polyclonal antibody to p50 completely supershifted this DNA-protein complex. Incubation with HMG-I(Y) caused the probe to form a more rapidly migrating DNA-protein complex than did incubation with p50. When the probe and increasing concentrations of p50 were incubated in the presence of HMG-I(Y), the DNA-protein complex was more intense and migrated more slowly than did the complex formed with p50 alone (Fig.  5). As this upper DNA-protein complex formed, the intensity of the HMG-I(Y) band decreased, suggesting that HMG-I(Y) was being incorporated into a ternary complex. The addition of a polyclonal antibody to p50 supershifted all components.
We have shown recently that IL-1␤ and LPS are able to induce HMG-I(Y) in vitro and in vivo, respectively (47). Thus, to determine if HMG-I(Y) had a dose-dependent effect on formation of this ternary complex, we incubated the same radiolabeled probe with p50 in the presence of increasing concentrations of HMG-I(Y). The intensity of the ternary complex increased as the concentration of HMG-I(Y) increased (Fig. 6). Addition of the p50 antibody resulted in a complete supershift of this upper, ternary complex. These data suggest that HMG-I(Y) and NF-B subunit p50 bind to the iNOS promoter/enhancer and that HMG-I(Y) assists in the formation of a ternary complex.
For HMG-I(Y) to recruit transcription factors to their DNA binding sites, it must usually bind to DNA. However, recent studies have revealed that HMG-I(Y) mutants incapable of binding to DNA were still able to enhance serum response factor binding to DNA (44). Therefore, we performed further studies to determine if HMG-I(Y) binding to DNA was necessary for recruitment of NF-B subunit p50 to the iNOS promoter/enhancer and formation of a ternary complex. A radiolabeled probe encoding region Ϫ87 to Ϫ52 of the iNOS 5Јflanking sequence and containing a mutated Oct site was incubated with p50 in the absence or presence of HMG-I(Y). Incubation of p50 alone with the probe containing the Oct site mutation resulted in a slowly migrating DNA-protein complex, and this complex was supershifted by a polyclonal antibody to p50 (Fig. 7). Incubation of this probe with HMG-I(Y) resulted in no DNA-protein complex formation and thus no HMG-I(Y) binding. Also, incubation of this probe with p50 in the presence of HMG-I(Y) resulted in no ternary complex formation (Fig. 7). These data suggest that HMG-I(Y) must bind to DNA at the Oct site to facilitate the formation of a ternary complex between HMG-I(Y), p50, and the iNOS promoter/enhancer. Subunits p50 and p65-Recently, we have shown that, in the presence of p50 and p65, HMG-I(Y) was able to increase iNOS promoter/enhancer activity in a dose-dependent manner (47). To extend these studies, we cotransfected the iNOS(Ϫ1485/ϩ31) promoter/enhancer construct with expression plasmids for p50, p65, or a combination of p50 and p65 in the absence or presence of an expression plasmid for HMG-I(Y). The transfections were performed in Drosophila SL2 because they contain far less endogenous HMG-I(Y) than do mammalian cells (29). This experiment allowed us to determine which members of a nucleoprotein com-plex (containing HMG-I(Y) and NF-B subunits) would be necessary to drive iNOS transcription most efficiently. p65 alone or in combination with p50 increased iNOS reporter activity,

FIG. 4. HMG-I(Y) binding to the iNOS 5-flanking sequence by EMSA.
A, oligonucleotides containing bp Ϫ87 to Ϫ52 of the iNOS promoter/enhancer were radiolabeled and incubated with recombinant HMG-I(Y) (100 ng). Unlabeled competitors were added at a 500-fold molar excess as indicated. I, identical competitor; NI, non-identical competitor. B, EMSA were performed with radiolabeled oligonucleotides encoding bp Ϫ87 to Ϫ52 of the iNOS promoter/enhancer. The oligonucleotides contained the wild-type sequence (iNOS WT), a mutated NF-B site (NF-B mut), or a mutated Oct site (Oct mut). The radiolabeled probes were incubated in the presence (ϩ) or absence (Ϫ) of recombinant HMG-I(Y) (100 ng).

FIG. 5. HMG-I(Y) and NF-B sub-
unit p50 bind to the iNOS promoter/ enhancer and form a ternary DNAprotein complex. EMSA were performed with a radiolabeled oligonucleotide encoding bp Ϫ87 to Ϫ52 of the iNOS promoter/ enhancer. Increasing concentrations of p50 were incubated with the radiolabeled probe in the presence (ϩ) or absence (Ϫ) of recombinant HMG-I(Y) (100 ng). Antibody to p50 (p50 Ab, 1 g) was used to verify the presence of p50 in the slowly migrating DNA-protein complexes (supershift).
FIG. 6. Increasing concentrations of HMG-I(Y) facilitate the formation of a ternary DNA-protein complex with NF-B subunit p50. EMSA were performed with a radiolabeled oligonucleotide encoding bp Ϫ87 to Ϫ52 of the iNOS promoter/enhancer. Increasing concentrations of HMG-I(Y) were incubated with the radiolabeled probe in the presence (ϩ) or absence (Ϫ) of recombinant p50 (50 ng). Antibody to p50 (p50 Ab, 1 g) was used to verify the presence of p50 in the slowly migrating DNA-protein complexes (supershift).
but the most potent effect was in the presence of both plasmids (Fig. 8). HMG-I(Y) alone did not significantly increase iNOS promoter/enhancer activity. When HMG-I(Y) was transfected in conjunction with p50 alone or p65 alone, however, iNOS reporter activity increased further in comparison with transfections lacking HMG-I(Y). Transfection of HMG-I(Y) in conjunction with both p50 and p65 resulted in a potentiated increase in iNOS promoter/enhancer activity in comparison with transfections lacking HMG-I(Y) (Fig. 8). These data suggest that HMG-I(Y) can use either p50 or p65 to drive iNOS transcription; however, the most dramatic effect on transactivation of the iNOS promoter/enhancer occurs when HMG-I(Y) is coexpressed with both NF-B subunits.
Because the most important IL-1␤-responsive elements are located within the Ϫ234 to ϩ31 region of the iNOS 5Ј-flanking sequence, we transfected construct iNOS(Ϫ234/ϩ31) into the SL2 cells. As in the transfections with iNOS(Ϫ1485/ϩ31), HMG-I(Y) potentiated iNOS(Ϫ234/ϩ31) transactivation by p50 and p65 (Fig. 9A). In addition to iNOS(Ϫ234/ϩ31), we also transfected construct iNOS(Ϫ331/ϩ31 NF-Bm) into the SL2 cells. The construct was cotransfected with expression plasmids p50, p65, and HMG-I(Y). Mutation of the NF-Bd site disrupted the ability of HMG-I(Y) to potentiate iNOS transactivation. This same disruption also occurred when the NF-Bd site was mutated in the construct containing region Ϫ1485 to ϩ31 of the iNOS promoter/enhancer (data not shown). These data suggest that enhanced transactivation of the iNOS promoter/enhancer by HMG-I(Y) requires an intact NF-Bd site in the downstream region of the iNOS 5Ј-flanking sequence. DISCUSSION The high output pathway of NO production is catalyzed by the inducible isoform of NOS. iNOS induction can be regulated in some cells by posttranscriptional mechanisms (48) and by the availability of cofactors such as heme and tetrahydrobiopterin (49 -51). In most cells, however, the induction of iNOS by inflammatory stimuli is tightly regulated at the level of gene transcription (5,(11)(12)(13)20). Activation of NF-B is a critical component of the transcriptional induction of iNOS by inflammatory stimuli (14). We have observed elsewhere that elements other than an NF-Bd binding site, in the downstream portion of the iNOS promoter/enhancer, may be involved in iNOS induction after IL-1␤ stimulation in vascular smooth muscle cells (20).
The present analysis of the iNOS promoter/enhancer revealed that an AT-rich region (bases Ϫ61 to Ϫ54) downstream of the NF-Bd site is essential for full induction of iNOS by IL-1␤ in vascular smooth muscle cells (Fig. 2). The importance of this AT-rich region was not limited to IL-1␤ stimulation, as induction of iNOS by LPS in vascular smooth muscle cells (Fig.  3A) and by LPS in macrophages (Fig. 3B) required an intact Ϫ61 to Ϫ54 site. By in vivo footprint analysis in macrophages, Goldring et al. showed that nuclear proteins bind to this ATrich region, corresponding to an Oct site, after LPS stimulation (21).
Two recent studies in macrophages have suggested that the downstream Oct site may be important for iNOS activation by LPS (52) and IL-6 (53). Xie (52) demonstrated that an LPSresponsive element, termed LRE AA , resides within the downstream Oct site. Xie suggested that Oct-1-like proteins, but not other POU domain proteins (Oct-2 or Pit 1) (54), are contained in a complex that binds at the LRE AA site. Sawada and colleagues investigated the importance of the downstream Oct site in the activation of iNOS by IL-6 (53). They demonstrated that Oct-2-related proteins bind at this site during IL-6-induced macrophage differentiation, and that this Oct site is essential for induction of iNOS by IL-6. The binding activity of Oct-1-related proteins appeared to decrease after IL-6 stimulation. Although these studies showed that Oct-1-like proteins and Oct-2 proteins are components of the DNA-protein com- plexes that bind at the downstream Oct site in macrophages, their specific roles in iNOS activation and their ability to work in conjunction with, or independently of, NF-B were not investigated.
Because the downstream Oct site is rich in A and T residues, we became interested in the potential role of HMG-I(Y) as a mediator of iNOS activation. HMG-I(Y) is an architectural transcription factor that binds to AT-rich regions in the minor groove of DNA (26). HMG-I(Y) by itself is not a transcriptional activator; instead, HMG-I(Y) facilitates the assembly and stability of stereospecific DNA-protein complexes that drive efficient gene transcription (22,31). HMG-I(Y) performs this task by modifying DNA conformation and by recruiting nuclear proteins (such as subunits of NF-B) to an enhanceosome complex (28,29). Our data revealed that HMG-I(Y) bound specifically in region Ϫ87 to Ϫ52 of the iNOS promoter/enhancer (Fig.  4A), and that this binding occurred at the AT-rich Oct site, not at the NF-Bd site (Fig. 4B). HMG-I(Y) assisted in the formation of a ternary complex containing itself, p50, and the iNOS promoter/enhancer (Figs. 5 and 6). The formation of this complex required the binding of HMG-I(Y) to the downstream Oct site (Fig. 7). Although HMG-I(Y) alone had no significant effect on iNOS reporter activity, HMG-I(Y) did increase iNOS activity in the presence of p50 or p65. Moreover, the most dramatic increase in iNOS transactivation occurred when HMG-I(Y) was coexpressed with both p50 and p65 (Fig. 8). These data suggest that HMG-I(Y) works in conjunction with subunits of NF-B to drive transcription of the iNOS promoter/enhancer. Deletion of an upstream NF-B site (Ϫ971 to Ϫ962) had no effect on HMG-I(Y)'s ability to potentiate iNOS transactivation by p50 and p65 (Fig. 9A); however, mutation of the downstream NF-Bd site (Ϫ85 to Ϫ76) disrupted this HMG-I(Y) response (Fig.  9B). Thus, binding of HMG-I(Y) at the Oct site (Ϫ61 to Ϫ54) in conjunction with binding of NF-B subunits at the NF-Bd site (Ϫ85 to Ϫ76) appears to be essential for the most potent activation of the iNOS promoter/enhancer.
Previous studies have shown that binding sites for HMG-I(Y) partially or fully overlap binding sites for transcription factors that are incorporated into an enhanceosome complex (31). By binding to DNA in the minor groove, HMG-I(Y) is able to recruit transcription factors to the major groove. This scenario does not appear to apply to the iNOS promoter/enhancer. The AT-rich Oct site, the location of HMG-I(Y) binding in the iNOS promoter/enhancer, resides 15 bp downstream of the NF-Bd site. Our experiments show that NF-B subunits p50 and p65 drive iNOS transcription most efficiently in the presence of HMG-I(Y), even though the binding sites for these factors do not overlap. Future studies will focus on additional transcription factors that may interact with NF-B and HMG-I(Y) to form an enhanceosome complex and drive iNOS transcription. Because HMG-I(Y) binds at the AT-rich Oct site in the iNOS promoter/enhancer and HMG-I(Y) is known to interact with POU domain proteins (55), we will initially concentrate on transcription factors that preferentially bind to AT-rich sequences and POU domain proteins.