Phosphatidylinositol 3-Kinase in Interleukin 1 Signaling

The signaling mechanisms utilized by the proinflammatory cytokine interleukin-1 (IL-1) to activate the transcription factors NFκB and activator protein-1 (AP-1) are poorly defined. We present evidence here that IL-1 not only stimulates a dramatic increase in phosphatidylinositol 3-kinase (PI 3-kinase) activity but also induces the physical interaction of its type I receptor with the p85 regulatory subunit of PI 3-kinase. Furthermore, two PI 3-kinase-specific inhibitors, wortmannin and a dominant-negative mutant of the p85 subunit, inhibited IL-1-induced activation of both NFκB and AP-1. Transient transfection experiments indicated that whereas overexpression of PI 3-kinase may be sufficient to induce AP-1 and increase nuclear c-Fos protein levels, PI 3-kinase may need to cooperate with other IL-1-inducible signals to fully activate NFκB-dependent gene expression. In this regard, cotransfection studies suggested that PI 3-kinase may functionally interact with the recently-identified IL-1-receptor-associated kinase to activate NFκB. Our results thus indicate that PI 3-kinase is a novel signal transducer in IL-1 signaling and that it may differentially mediate the activation of NFκB and AP-1.

The biological processes of growth, differentiation, and immunity are dependent on the highly regulated action of transcription factor families such as NFB and activator protein-1 (AP-1). 1 These two transcription factors participate not only in normal physiology but in diseased conditions as well. Extracellular stimuli such as interleukin-1 (IL-1), tumor necrosis factor (TNF), viruses, and UV light are among the known potent inducers of NFB (1). Of these, IL-1, which is a major proinflammatory cytokine, is responsible for mediating numerous host responses, including fever, activation of lymphocytes, and the induction of acute-phase proteins (2). Elevated levels of IL-1 have been associated with various pathological conditions, including rheumatoid arthritis (2). Although several biological activities of IL-1 have been characterized, the molecular mechanisms by which its signals are transduced from the plasma membrane to affect gene transcription in the nucleus remain to be elucidated.
One of the most prominent IL-1-inducible signals, one which requires the type I IL-1 receptor (IL-1RI), involves the rapid and dramatic activation of NFB and AP-1 and results in the induction of discrete sets of genes (2). Since IL-1RI shares no significant homology with conserved protein kinase domains, it is unlikely to have any intrinsic protein kinase activity (3) and may need to recruit specific cytoplasmic proteins to transmit its signals. One such protein, recruited to the receptor in response to IL-1 stimulation, is the IL-1 receptor-associated protein kinase (IRAK), which bears significant homology to the Drosophila protein Pelle (4). Although there is no evidence linking IRAK directly to NFB activation, studies on the function of Pelle have demonstrated its importance in the activation of Dorsal, the mammalian equivalent of NFB. Recently, IRAK has been shown to physically associate with a protein belonging to the TNF receptor-associated factor family called TRAF6 (5). In contrast to TNF, which recruits TRAF2 to activate NFB, IL-1-induced activation of NFB is mediated by TRAF6 (5). However, the mechanisms by which the recruitment of TRAFs to the IL-1 and TNF receptors leads to the activation of NFB are not understood.
NFB is normally inactive and kept sequestered in the cytoplasm by its interaction with the inhibitory subunit IB (1). Upon stimulation, IB is rapidly phosphorylated, ubiquitinated, and then degraded, resulting in the release and subsequent nuclear translocation of active NFB (1). In addition to the TRAFs, several other factors appear to play a role in NFB activation. Although some interact with the TRAFs (6), others such as Raf kinase, tyrosine kinases, reactive oxygen intermediates, and sphingomyelinases have also been reported (1). AP-1 is predominantly a heterodimeric complex of c-Fos and c-Jun proteins, and its activation is mainly due to induction of c-Fos synthesis and the phosphorylation of both c-Jun and c-Fos (7). Recent studies have indicated that phosphatidylinositol 3-kinase (PI 3-kinase) may up-regulate c-Fos synthesis (8) and stimulate the Jun N-terminal kinase pathway (9), which might then lead to the phosphorylation and activation of c-Jun. Other studies have implicated PI 3-kinase in epidermal growth factor-induced AP-1 activation (10). PI 3-kinase consists of a catalytic subunit (p110) associated with a regulatory polypeptide (p85) (11). Ligand-dependent interactions between the SH2 domains of the p85 subunit and the phosphotyrosinecontaining YXXM motif present on several cytokine/growth factor receptors have been reported (11). The phosphorylated lipid products generated by this enzyme may act as second messengers to activate protein kinases such as the Akt gene product (12) or certain isoforms of protein kinase C (13). Its reported interactions with Ras (14) and its potential ability to activate several signaling pathways (9) underscore the importance of PI 3-kinase in various cellular functions.
We have demonstrated that the acute-phase gene serum amyloid A 1 and 3 (15)(16)(17) are highly inducible by cytokines such as IL-1 and shown that NFB is critical for their expression (16 -18). In this study, we investigated the signaling events that lead to NFB and AP-1 activation in IL-1-stimulated cells. Our results indicate a prominent role for PI 3kinase in the activation of both NFB and AP-1.

EXPERIMENTAL PROCEDURES
Cell Lines and Reagents-HepG2 and KB cells were obtained from the American Type Culture Collection and maintained at 37°C in modified Eagle's medium containing 10% fetal bovine serum and antibiotics. Wortmannin and phosphatidylinositol were from Sigma. PD98059 was obtained from Biomol Research Laboratories Inc., and staurosporine was obtained from Life Technologies, Inc.
Electrophoretic Mobility Shift Assays-Cells were serum-starved for 4 h before treatment with various inhibitors and IL-1. Nuclear extracts were prepared and electrophoretic mobility shift assays reactions performed as described earlier (19). Nuclear extracts of 2.5 g were incubated with radiolabeled probe. Protein-DNA complexes were resolved on 5% polyacrylamide gels and visualized by autoradiography. The results presented here are representative of at least three independent experiments.
PI 3-Kinase Assays and Coimmunoprecipitation-KB or HepG2 cells were seeded in 60-mm culture dishes overnight. Cells were serumstarved for 4 h before treatment for the indicated times with or without IL-1. Whole cell lysates were prepared in ice-cold lysis buffer containing 50 mM HEPES, pH 7.5, 10% glycerol, 150 mM sodium chloride, 1% Nonidet P-40, 1.5 mM magnesium chloride, 1 mM EGTA, 50 mM sodium fluoride, 1 mM sodium orthovanadate, 14 mM ␤-mercaptoethanol, 0.2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 3 mM benzamidine. After 20 min on ice, the lysates were clarified by centrifugation at 14,000 ϫ g. The supernatant was then utilized for PI 3-kinase assays as described earlier (20), and the labeled [ 32 P]phosphatidylinositol 3-phosphate was resolved by thin layer chromatography (21). For coimmunoprecipitation experiments, whole cell lysates were prepared as described above and incubated with anti-p85 PI 3-kinase antibodies (Santa Cruz) or anti-IL-1R type I antibodies (Santa Cruz) for 1 h before incubation with protein A-Sepharose beads (1 h). The immune complexes were boiled in Laemmli buffer and resolved on 10% denaturing polyacrylamide gels. After transfer to nitrocellulose membranes, proteins were probed with anti-IL-1R type antibodies or anti-p85 PI 3-kinase antibodies in accordance with the manufacturer's specifications. Proteins were visualized by ECL (Amersham) using goat anti-rabbit horseradish peroxidase-conjugated secondary antibodies.
Reporter Assays-HepG2 cells were plated in 100-mm dishes and transfected the next day with the (NFB) 3 /CAT or (AP-1) 2 /CAT report-ers and various expression vectors. Transfections were done using the Polybrene method. Sixteen hours after transfection, cells were treated with cytokines or left untreated for 20 h. All other procedures were performed as described elsewhere (22).
Construction of IRAK Expression Vector-A forward primer 5Ј-GCG-GCGGCAACCATGGCCGGG-3Ј and a reverse primer 5Ј-TCTGCCGG-TACCAACACATCAGCTCTGGAATTC3-Ј containing a KpnI restriction site were used to amplify IRAK cDNA from a HeLa cDNA library by the polymerase chain reaction. A 2.1-kilobase polymerase chain reaction product was cloned into the p85DN vector SR␣ (23) from which the p85DN cDNA had been excised after digestion with EcoRI and KpnI.
Immunostaining-Twelve hours after transfection, cells were washed with phosphate-buffered saline and fixed with 10% paraformaldehyde for 10 min before being permeabilized with 0.2% Nonidet P-40 for 2 min. Cells were washed five times with phosphate-buffered saline and then incubated with c-Fos polyclonal antibody (Santa Cruz) for 60 min. After washing, these cells were incubated with tetramethylrhodamine B isothiocyanate-conjugated goat anti-rabbit IgG (Jackson Im-munoResearch Laboratories) for 45 min. The coverslips were washed several times and then inverted on glass slides containing 90% glycerol and sealed. Cells were then observed under a Nikon Diaphot TMD-EF inverted microscope with epi-fluorescence attachment. All other procedures were performed as described in the Santa Cruz Research Applications (Santa Cruz).

Wortmannin Inhibits IL-1-induced Activation of NFB and
AP-1-To identify potential mediators of IL-1 signaling, we used gel mobility shift assays to examine the ability of various inhibitors to block the activation of NFB and AP-1 by IL-1. Epidermoid carcinoma KB or hepatoma HepG2 cells were stimulated with IL-1 after pretreatment with either wortmannin, a potent PI 3-kinase inhibitor (24), PD98059, an inhibitor of the mitogen-activated protein kinase pathway (25), or staurosporine, a general protein kinase inhibitor. Whereas DNA binding activities of both NFB and AP-1 were potently induced by IL-1, their activation was inhibited completely in cells pretreated with wortmannin ( Fig. 1, A and B). Inhibition by wortmannin was observed in both KB and HepG2 cells, indicating that the wortmannin-sensitive mechanism in NFB activation may be a general mechanism. PD98059, which inhibits the mitogen-activated protein kinase pathway, also had an inhibitory effect but was less effective than wortmannin (Fig. 1). Consistent with earlier reports (26), staurosporine had little or no effect on the activation of NFB or AP-1 by IL-1.
Rapid and Transient Activation of PI 3-Kinase Activity by IL-1-The concentration of wortmannin that we used has been shown to selectively inhibit PI-3 kinase activity (24). To determine if PI 3-kinase activity is indeed regulated by IL-1, cytokine-treated cell extracts were assayed for PI 3-kinase activity with phosphatidylinositol as substrate. Insulin, a known potent activator of PI 3-kinase activity (11), was included as positive control. Treatment of KB cells with IL-1 or insulin resulted in elevation of PI-3 kinase activity in a time-dependent manner (Fig. 2). Within 7 min, insulin stimulated PI 3-kinase activity by about 27-fold ( Fig. 2A), which is comparable to the foldactivation reported by others (27). The insulin-stimulated increase in PI 3-kinase activity declined rapidly thereafter. In contrast, IL-1 stimulated PI 3-kinase with much faster induction kinetics (3 min) and to a greater magnitude (108-fold). Like insulin, the IL-1-stimulated increase in PI 3-kinase activity was also transient. Similar results were obtained in HepG2 cells (data not shown). IL-1-inducible PI 3-kinase activity was dependent on the presence of phosphatidylinositol in the assays and was inhibited completely by 0.2% Nonidet P-40 (data not shown). The presence of detergent in these assays has little effect on PI 4-kinase activity and thus differentiates the activity of PI 3-kinase from that of any contaminating PI 4-kinase. Pretreatment of cells with wortmannin prevented the stimulation of PI 3-kinase activity by IL-1 (Fig. 2B), further linking PI 3-kinase functionally to NFB and AP-1 activation. The kinetics of PI 3-kinase activation preceded that of NFB (5-10 min) and AP-1 (30 min), consistent with a role for PI 3-kinase in the activation of these transcription factors.
Physical Interaction between PI 3-Kinase and the Type I IL-1 Receptor-The stimulation of PI 3-kinase by platelet-derived growth factor and other growth factors involves the physical association of PI 3-kinase with their receptors (11). The SH2 domains of the p85 subunit of this enzyme recognize and associate with the phosphotyrosine-containing YXXM (11) or YVXV motif (28) present on those receptors. Analysis of the human IL-1RI protein sequence revealed one tyrosine residue located at amino acid 496 of the cytoplasmic domain that matched the YXXM motif (YEKM). Similar sequence motifs are also found in the mouse IL-1RI and the Drosophila Toll protein. This raises the possibility that in response to IL-1 stimulation, the SH2 domain of p85 might also physically associate with the human IL-1RI through the phosphorylated Y 496 EKM sequence. To test whether this interaction between p85 and IL-1RI exists, polyclonal antibodies against the p85 PI 3-kinase and IL-1RI were used to coimmunoprecipitate their associated proteins. The immunoprecipitates were then Western-blotted with anti-IL-1RI (for anti-p85 precipitate) and anti-p85 antibodies (for anti-IL-1RI precipitate). In two different cell lines, KB and HepG2, IL-1RI could be detected in the anti-p85 PI 3-kinase immune complexes (Fig. 3A). The association of IL-1RI with p85 PI 3-kinase was dependent entirely on IL-1 stimulation. Furthermore, their association was time-dependent, with maximal association achieved at 0.5 min after IL-1 addition, and little or no interaction was detected after 3 min (Fig. 3A). The physical coupling of these two proteins was further confirmed in converse experiments wherein anti-IL-1RI immune complexes coprecipitated p85 PI 3-kinase (Fig. 3B).
PI 3-Kinase Is Necessary for NFB Activation, but Overexpression of the p110 Catalytic Subunit Alone Is Insufficent to Activate NFB-dependent Gene Expression-Since wortmannin blocked the activation of both NFB and AP-1, the role of PI-3 kinase in the activation of these two transcription factors was explored further by transient transfection assays. The functional importance of PI 3-kinase was evaluated by cotransfection of an (NFB) 3 /CAT reporter plasmid, with either the dominant-negative mutant of the p85 regulatory subunit (p85DN) or p110 catalytic subunit of PI 3-kinase. To demonstrate that PI-3 kinase functions as a mediator in NFB activation, we first employed a dominant-negative mutant of PI 3-kinase, p85DN (23) in these cotransfection assays. When p85DN was overexpressed, it inhibited IL-1-induced (NFB) 3 /CAT expression by approximately 70% (Fig. 4A), implicating PI 3-kinase as an important mediator of NFB activation. To ascertain the specificity of this inhibition, we tested its effect on the induction of a T1 kininogen/CAT reporter gene. We have previously shown that this reporter can be induced by a combination of IL-6 and Dex (22) and that the induction does not involve NFB. In contrast to the (NFB) 3 /CAT reporter, the induction of the T1 kininogen/CAT gene was unaffected by overexpression of p85DN, suggesting that its inhibition on NFB-dependent gene expression was specific. Overexpression of p110 in- duced CAT activity approximately 3-fold and in a dosedependent manner (Fig. 4B). Interestingly, when p110transfected cells were further stimulated with IL-1, synergistic activation of the (NFB) 3 /CAT reporter was observed (Fig. 4C). Maximal synergism was observed at suboptimal concentrations of IL-1 (data not shown). In addition, this synergism between IL-1 and p110 was dependent on the dose of p110 used for transfection, the best effect being achieved with 5-10 g of transfected p110 DNA. Induction of the (NFB) 3 /CAT reporter by IL-1 and its synergistic activation by IL-1 and p110 can be blocked by IB (Fig. 4C), the inhibitory subunit of NFB. This indicates that induction of the reporter gene by these stimuli specifically involved NFB activation. As a control, the mutant (NFB) 3 /CAT reporter construct that lacks a functional NFB binding site was nonresponsive regardless of the treatment.

PI 3-Kinase Is Necessary and Sufficient for AP-1-dependent Gene Expression-
The effect of p110 overexpression on an (AP-1) 2 /CAT reporter was also examined by cotransfection studies. In contrast to its effects on the (NFB) 3 /CAT reporter, overexpression of p110 alone induced the (AP-1) 2 /CAT reporter to the same extent as IL-1 treatment (Fig. 5E), and no additional increase was observed upon inclusion of IL-1, indicating that the reporter could be fully activated by p110 alone. Furthermore, p85DN blocked IL-1-induced AP-1 activation completely (Fig. 5E). Our data are therefore consistent with recent reports for a role of PI-3 kinase in epidermal growth factor-induced AP-1 activation (10) and further demonstrate the involvement of PI-3 kinase in IL-1-induced AP-1 activity. Activation of AP-1 involves a dramatic elevation of c-Fos protein that can then dimerize with Jun proteins to produce AP-1 complexes. There is evidence that PI 3-kinase is involved in the activation of c-Fos synthesis (8). Since IL-1 up-regulates c-Fos mRNA (2) and there is a dramatic induction of c-Fos protein in IL-1treated cells (Fig. 5B), we examined the ability of overexpressed PI 3-kinase to induce c-Fos protein levels. Overexpres- FIG. 4. PI 3-kinase is necessary but not sufficient for maximal activation of NFB by IL-1. A, PI 3-kinase is required for maximal NFB activation. The p85DN (dominant negative mutant of the p85 subunit of PI 3-kinase) or empty vector was cotransfected with either a CAT reporter gene containing three copies of wild-type NFB binding sites or with a CAT reporter under control of the T1 kininogen promoter (22). B, effect of p110 overexpression on (NFB) 3 /CAT reporter gene activity. Wild-type (B) or mutant (mB) CAT reporter genes were cotransfected into HepG2 cells with various doses of a p110 PI 3-kinase expression vector. C, synergistic activation of NFB by p110 and IL-1. Cells were cotransfected with either 5 g of p110 expression vector and the NFB reporter or vector control. Wherever indicated, DNA mixtures also contained 1 g of IB-expression vector or empty vector. The averages and standard deviations from three independent experiments are shown. sion of p110 PI 3-kinase induced c-Fos, as evidenced by the intense staining of nuclei in about 3-5% of the cells (Fig. 5D), which correlates with the percentage of our transfection efficiency in these cells. No cells stained above background levels in control cells (Fig. 5A) or cells transfected with the vector alone (Fig. 5C). These results suggest that overexpression of PI 3-kinase is sufficient to induce c-Fos protein levels and further emphasize its role in the activation of AP-1 by IL-1.
PI 3-Kinase Synergizes with IRAK to Activate NFB-dependent Reporter Gene Expression-Our results showed that the p85DN mutant of PI-3 kinase inhibited the activation of NFB by IL-1, and yet overexpression of the p110 catalytic subunit was unable to fully activate the (NFB) 3 /CAT reporter gene. We reasoned that since p110 can synergize with IL-1 to induce CAT activity, PI 3-kinase may need to cooperate with other IL-1-inducible signals or pathways to activate NFB. We cotransfected IRAK expression plasmid (4) with the (NFB) 3 / CAT gene to test for its ability to activate NFB. A dose-dependent increase in CAT activity was observed with a maximal induction of 13-fold at 15 g of IRAK ( Fig. 6 and data not shown), implicating IRAK as a potential mediator of IL-1induced NFB activation. When p110 and IRAK were coexpressed with the reporter gene, a synergistic activation of the reporter was observed (Fig. 6A), suggesting that these molecules may cooperate in vivo to activate NFB. We next evaluated the possibility that IRAK would synergize with IL-1. IRAK, like PI 3-kinase, synergized with IL-1 (Fig. 6B) to a comparable extent (compare Figs. 4C with 6B), and interestingly, this effect was blocked by coexpression of the p85DN mutant of PI-3 kinase (Fig. 6B). DISCUSSION We have presented evidence to indicate that the interaction of PI 3-kinase with IL-1RI is one of the early events in IL-1stimulated cells and that PI 3-kinase may be indispensable for the activation of NFB and AP-1 by IL-1. Our results show that wortmannin blocked the activation of both transcription factors by IL-1. Interestingly, wortmannin, which is a fungal metabolite, has been shown in experimental animals to exert antiinflammatory or immunosuppressive effects (24,29). It remains to be determined whether these effects can be accounted for, at least in part, by its inhibitory effects on NFB activation.
Since wortmannin blocked the activation of NFB and AP-1 by IL-1, it was predicted that IL-1 would stimulate PI 3-kinase activity. Indeed, IL-1 stimulated PI 3-kinase activity very potently and with rapid kinetics of activation. The activation kinetics of NFB and AP-1 have been well established and are slower in comparison to those for PI 3-kinase in IL-1-treated cells, consistent with a role for this enzyme in their activation. PI 3-kinase has been attributed with both lipid kinase and protein kinase activities, and since wortmannin inhibits both, it is difficult to assess at this time the relative importance of each activity in IL-1 signaling.
Wortmannin has recently been shown to inhibit phospholipase A 2 at concentrations that earlier were thought to be selective for PI 3-kinase (30). We therefore used a dominant negative mutant, p85DN, of the regulatory subunit of PI 3kinase to verify the importance of this enzyme in the activation of NFB and AP-1. p85DN has been used previously to confirm a role for PI 3-kinase in various cellular functions (23,31). Consistent with our wortmannin experiments, overexpression of p85DN strongly inhibited the activation of NFB-and AP-1dependent gene expression by IL-1. PI 3-kinase therefore appears to be essential for their activation by IL-1.
The cytoplasmic domain of IL-1RI protein contains a sequence (Y 496 EKM) which fits the YXXM motif that has been demonstrated to mediate direct physical interaction between receptor proteins and PI 3-kinase. Our coimmunoprecipitation studies demonstrate that the IL-1RI and PI 3-kinase do interact and that this association is induced very rapidly by IL-1. Although these data strongly suggest that the interaction between p85 PI 3-kinase and IL-1RI is direct, the presence of an accessory factor that links these two proteins cannot be ruled out. Interestingly, this Y 496 EKM motif is within the 50-amino acid (477-527) region in the IL-1RI cytoplasmic domain found to be essential for IL-1 signal transduction (32). Similar sequence motifs are also found in the mouse IL-1RI and the Drosophila Toll protein, which is equivalent to human IL-1RI.
Various studies have indicated that for the YXXM motifs to be capable of binding to p85 PI 3-kinase, the tyrosine residues must be phosphorylated (Ref. 33, see references in Ref. 34). Several receptor and nonreceptor tyrosine kinases that could potentially phosphorylate these sites and lie upstream of PI 3-kinase in various signaling pathways have been implicated (11). In one study, a sequence in the insulin receptor substrate-1 containing a PI 3-kinase binding motif and several flanking amino acid residues was evaluated as a substrate for different nonreceptor tyrosine kinases (35). Using synthetic peptides in protein kinase assays, it was determined that the tyrosine kinases differed in their tolerance for various amino acid substitutions, and in particular, an aspartate immediately N-terminal to the tyrosine was indispensable. In this context, it may be noted that in addition to other acidic amino acids, there is an aspartic acid (Asp 495 ) immediately adjacent to the Y 496 EKM sequence on the IL-1RI. This suggests that the Y 496 EKM sequence on the IL-1RI may be a good substrate for tyrosine kinases and may facilitate direct interaction between IL-1RI and PI 3-kinase. An IL-1-inducible tyrosine kinase activity has been reported but not identified (36). The involvement of tyrosine kinase activities in the steps leading to NFB activation has been reported for various inducers (see references in Ref. 1).
Overexpression of p110 was sufficient to induce the AP-1/ CAT reporter and to increase nuclear c-Fos protein levels as potently as IL-1. These results are consistent with recent reports that epidermal growth factor-induced activation of AP-1 requires PI 3-kinase (10). In sharp contrast to the AP-1 re-FIG. 6. IRAK and PI 3-kinase can synergistically activate NFB. 5 (A) or 10 g (B) of IRAK expression vector or the empty vector were cotransfected into HepG2 cells with the wild-type NFB reporter. The transfection mixture also contained 5 g of p110 PI-3 kinase or p85DN expression vector. Total DNA used for each transfection was maintained at 20 g by the addition of empty vector. Cells were then stimulated with IL-1 or left untreated, and CAT activities were determined. * indicates that 10 g of IRAK were used in the experiment. This data is representative of two independent experiments. porter, we observed that although overexpression of p110 activated the NFB/CAT gene in a dose-dependent manner, the activation was much less than what we normally observe with IL-1. Addition of IL-1 to p110-overexpressing cells, however, resulted in a synergistic activation of the NFB/CAT reporter. It appears therefore that whereas PI 3-kinase is necessary for IL-1-induced activation of NFB, its overexpression is not sufficient to fully activate this transcription factor. To identify signaling molecules that can cooperate with PI 3-kinase, we expressed the IRAK in various cotransfection experiments. We show evidence here that overexpression of IRAK is sufficient to activate the NFB reporter in a dose-dependent manner. To address the possibility that IRAK and PI 3-kinase cooperate to activate NFB, we followed two approaches. We cotransfected various doses of IRAK with the p110-expression vector and the NFB/CAT reporter. A clear synergism with p110 could be observed at lower doses of IRAK. In the second approach, since IRAK synergized with IL-1, we used the dominant negative mutant p85DN to test whether PI 3-kinase may be involved. Cotransfection of p85DN inhibited the synergism between IRAK and IL-1, lowering the CAT activity levels to those activated by IRAK in the absence of IL-1. In addition to providing further evidence for the involvement of PI 3-kinase in IL-1induced activation of NFB, these data support the idea that to mediate NFB activation, PI 3-kinase cooperates with molecules such as IRAK that possibly lie on other signaling pathways (Fig. 7). Similar observations have been made for TNFand CD40L-induced activation of NFB (37). While TRAF2 overexpression is sufficient to activate NFB, it was recently shown that TRAF2 may need a coactivator protein TRAF family member-associated NFB activator. In contrast to TRAF2, overexpression of TRAF family member-associated NFB activator alone was unable to activate NFB. The function of PI 3-kinase in transcription factor activation may not be limited to NFB and AP-1, for it has been implicated recently in the activation of STAT3 (38). Whereas the tyrosine phosphorylation of STAT3 plays a prominent role in its activation, phosphorylation of this factor on a specific serine residue is important and may require PI 3-kinase.
The mechanisms by which activated PI 3-kinase may ulti-mately result in the activation of NFB and AP-1 are unclear. However, several signaling molecules have been shown to affect directly or indirectly the pathway leading to the activation of these transcription factors. For example, two atypical forms of protein kinase C, aPKC (39) and aPKC (10), have been shown to be involved in NFB and AP-1 activation, respectively. Since the phosphorylated lipid products of PI 3-kinase activate various isoforms of PKC in vitro, including aPKC (40), PKC may lie downstream of PI 3-kinase. aPKC activity was recently reported to be stimulated by IL-1 in rat renal mesangial cells (41). Small G-proteins and the protein kinase mitogenactivated protein kinase-extracellular signal-regulated kinase kinase kinase 1 (MEKK1) are among the other signaling molecules that have been implicated in the activation of NFB. Physical interaction between the small GTPase Cdc42 and PI 3-kinase may result in the stimulation of PI 3-kinase activity (42,43). Cdc42 appears to be required for the activation of NFB by TNF (44), and our results suggest that the mechanism may involve interactions with PI 3-kinase. Finally, MEKK1, a key mediator of the Jun N-terminal kinase pathway, has recently been implicated as an upstream regulator of the 700-kDa IB kinase (45). Since overexpression of PI 3kinase activates the Jun N-terminal kinase pathway (9), it would be of interest to determine if it is involved in the upregulation of MEKK1 activity in IL-1-stimulated cells. It is not known whether any of these proteins, Cdc42, PKCs, or MEKK1, interact with IRAK and if they do, whether those interactions would account for the synergism between PI 3kinase and IRAK in NFB activation. Since our results suggested that PI 3-kinase may functionally interact with IRAK to activate NFB, we are currently investigating the molecular basis of such interactions. It is hoped that in addition to providing greater insights into the mechanisms employed by IL-1 for NFB activation, these studies would also reveal the identity of the downstream targets of both IRAK and PI 3-kinase.