Requirement of Phosphoinositide 3-Kinase and Akt for Interferon- (cid:1) mediated Induction of the (cid:1) - R1 ( SCYB11 ) Gene*

We previously reported and Chem. that IFN- (cid:1) induction of (cid:1) -R1 in fibrosarcoma cells required transcription factors ISGF-3 and NF- (cid:2) B. IFN- (cid:1) treatment did not augment the abundance of NF- (cid:2) B, but led to phosphorylation of the NF- (cid:2) B subunit p65 and induced a nuclear activity capa-ble of phosphorylating a p65-GST fusion construct in the carboxy-terminal transactivation domain (TAD), residues 354–551. We now present evidence for the involve-ment of phosphoinositide 3-kinase (PI3K) in this pathway. Pretreatment of HT1080-derived fibrosarcoma cells with pharmacological inhibitors of PI3K (wortman-nin or LY294002) selectively inhibited IFN- (cid:1) -induced (cid:1) -R1 mRNA accumulation in a dose-dependent manner. In stably transfected cell lines, bovine p85,

The IFN-␣/␤ receptor consists of two transmembrane chains (IFNAR-1 and IFNAR-2) that specifically pre-associate with intracellular signaling kinases, including TYK2 and JAK1, and latent non-phosphorylated transcription factors called signal transducers and activators of transcription (STAT). 1 Following engagement of ligand with the receptor, the kinases become activated and phosphorylate STATs. Activated STAT complexes translocate to the nucleus and bind with high affinity to target sequences and enhance transcription initiation of interferon-stimulated genes (ISGs) (1). The JAK-STAT signaling pathway is clearly necessary but not sufficient for the universe of IFN-mediated responses. Cells expressing kinase-deficient JAK1 were competent for IFN-␥-inducible JAK-STAT signaling and gene induction, yet lacked antiviral response (2). Nonreceptor components were also required selectively by type I IFN to mediate either antiviral effects (3) or high-affinity binding and anti-proliferative responses (4).
We previously reported the selective induction of the ␤-R1 (SCYB11) gene by IFN-␤, but not IFN-␣, in human fibrosarcoma and astrocytoma cells (5). ISGF-3 was essential but not sufficient for IFN-␤-dependent induction of ␤-R1 (5,6). The transcription factor nuclear factor-B (NF-B) was required for the induction of this gene by IFN-␤. IFN-␤ treatment did not induce NF-B DNA binding activity or nuclear translocation, but instead led to phosphorylation of the p65 subunit of NF-B (7).
In recent years insulin receptor substrate (IRS) proteins (8), CrkL adaptor protein (9), vav proto-oncogene (10,11), p38 MAP kinase (12), p42/p44 MAP kinases (13), and phosphoinositide 3-kinase (PI3K) (8,14) have been reported to be activated and participate in the generation of IFN signals. The role of PI3K in the transcriptional response to IFN has been enigmatic, although it has been demonstrated repeatedly that PI3K is activated by type I IFNs. The mechanisms by which PI3K is induced by type I IFNs in various cell types appear to differ (8,(15)(16)(17)(18)(19). We demonstrated that IFN-␤ treatment of fibrosarcoma cells induced a phosphatidylinositol (PI) lipid kinase activity, which co-precipitated with IFNAR-1, the ␣-subunit of the receptor (19) and found the association of p85 with IF-NAR-1 to be ligand independent (19).
PI3K is a dual-specificity cytoplasmic enzyme, possessing both serine and lipid kinase activity (20) that phosphorylates the D3 position of the inositol ring of PI. PI3K is a heterodimer consisting of a 110-kDa catalytic subunit and a 85-kDa regulatory subunit that provides SH-2-dependent adapter function, recruiting the catalytic subunit to phosphotyrosine residues on activated cellular components. Roles for PI3K and the serinethreonine kinase Akt have been defined recently in cytokine signaling. In particular, these components are required for generating phosphorylated p65 and for enhancing the transactivation function of NF-B complexes in response to IL-1 or TNF-␣ (21)(22)(23)(24).
Since we had found that p65 TAD was phosphorylated in IFN-␤-treated cells and that PI3K activity was induced by IFN-␤, we addressed the requirement of PI3K for induction of the ␤-R1 gene by IFN-␤. The results demonstrated that PI3K activity is required for IFN-␤-mediated phosphorylation of the p65 subunit of NF-B and transcriptional induction of the ␤-R1 gene.
Stable Transfections-The plasmid SR␣-wt-p85 coding for the entire sequence of bovine p85␣ and the SR␣-⌬p85, a deletion mutant of bovine p85␣ that lacks the binding site for the p110 catalytic subunit of P13-kinase because of a deletion of 35 amino acids (residues 479 -513), were obtained from Dr. K. Yonezawa (Kobe University) (26). U1.wt cells were grown to 60 -70% confluency in 150-mm plates and co-transfected with 10 g of wt-p85 plasmid or ⌬p85 plasmid and 1 g of pCMVpuromycin DNA. Cells were selected in puromycin, and single-cell clones were isolated. A reverse-transcriptase polymerase chain reaction (RT-PCR) assay was used to screen the clones (27). The sequence of the forward primer was 5Ј-GGCTTCTCTCTGACCCATTAACCTTC-3Ј, and the reverse primer was 5Ј-CCTCTTCCAATCTCCTTCTGCTG-3Ј. The reverse primer contained a 3-base mismatch compared with the human sequence and was selective for bovine p85. The size of the expected product was 472 bp for wtp85 expressing cells and 370 bp for the cells expressing ⌬p85. The amount of RNA used in the assay was normalized to tubulin expression using tubulin-specific primers that have been reported (27). The cell lines that contained PCR amplification products of expected sizes were further tested for protein expression by Western blot assay.
Western Blot-Cells were treated with IFN-␤ (5000 units/ml) at 37°C for varying times as indicated. Cells were washed twice with ice-cold phosphate buffered saline. Cells were lysed on ice for 20 min in lysis buffer containing 50 mM Tris, pH 7.4, 10% glycerol, 1% Triton X-100, 0.1 mM EDTA, 150 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, and 1 mM ␤-glycerophosphate and lysates clarified by centrifugation in a microcentrifuge at high speed for 20 mins at 4°C. Proteins were separated by SDS-PAGE and adsorbed to polyvinylidene difluoride membranes. Incubation with primary antibody was at 4°C overnight, and secondary antibody was for 1 h at 37°C. Immunoreactivity was visualized with the ECL Western blotting system (Amersham Biosciences). For immunoprecipitation 500 g of whole cell extracts were incubated with the indicated antibodies overnight at 4°C on a rotator followed by incubation with Protein A-Sepharose for 2 h at 4°C. At the end of the incubation the extracts were centrifuged, and the pellets were washed four times in lysis buffer and resolved by SDS-PAGE gel.
In Vitro Phosphorylation of p65 Transactivation Domain-The carboxyl terminus of p65 fused to GST (28) was expressed in Escherichia coli, and cells were lysed in 0.5% Nonidet P-40 lysis buffer as described (29) and bound to glutathione-sepharose beads (Amersham Biosciences). In vitro phosphorylation was performed using the GST-fused p65 substrate and nuclear extracts from U1.wt cells as described (7).

Activation of Akt by IFN-␤-
We previously showed that PI3K was activated in response to IFN-␤. As PI3K responses are commonly associated with activation of Akt, a downstream effector, we determined if Akt was activated in IFN-␤-treated cells. IFN-␤ induced phosphorylation of Akt. It was first detected at 5 min with a peak at 30 min, decaying after one hour (Fig. 1a, top panel). The activation of Akt in response to IFN-␤ was inhibited by PI3K inhibitor LY294002 (Fig. 1b).
Requirement of PI3K for Phosphorylation of p65 Subunit of NF-B by IFN-␤-Phosphorylation of p65 was increased in cells treated with IFN-␤ (7). To address the role of IFN-␤induced PI3K in the phosphorylation of p65 by IFN-␤, cells were pretreated for 30 min with LY294002 followed by addition of IFN-␤. Then, phosphorylation of GST-p65 was examined. IFN-␤ induced phosphorylation of p65 (Fig. 1c), whereas pretreatment with LY294002 blocked the IFN-␤-induced phosphorylation of p65 (Fig. 1c). This result suggested that IFN-␤induced phosphorylation of p65 required PI3K activity.
PI3K Inhibitors Selectively Block Induction of ␤-R1 by IFN-␤-The role of PI3K activity in endogenous gene induction by IFNs has not been elucidated. To address the requirement of PI3K for transcriptional induction of ␤-R1, we treated cells with IFN-␤ in the presence of chemical inhibitors of PI3K, wortmannin, or LY294002. Total cellular RNA was isolated and analyzed by RNase protection assay. Treatment with either inhibitor resulted in a selective inhibition of ␤-R1 gene induction (Fig. 1d). Induction of 6 -16, a type I IFN-induced gene that has an ISRE but lacks a B site in its promoter, was unaffected by treatment with either inhibitor. Wortmannin caused significant ␤-R1 inhibition at a concentration of 10 nM and inhibited ␤-R1 gene induction in a dose-dependent fashion with 99% suppression at 100 nM. Accumulation of 6 -16 mRNA was inhibited only modestly, and this effect was not concentration dependent. LY294002 also selectively abrogated ␤-R1 induction by IFN-␤ in a dose-dependent fashion with 95% inhibition at a dose of 25 M.
Overexpression of Bovine p85 Suppresses IFN Signaling to PI3K-Chemical inhibitors such as wortmannin can exert nonspecific effects. To further address the role of PI3K in IFN-␤ signaling, we constructed stable cell lines overexpressing wildtype bovine p85␣ (wt-p85␣) or a mutant bovine p85␣ that lacks the binding site for the p110 catalytic subunit of P13K because of a deletion of 35 amino acids (residues 479 -513) (Fig. 2, a and  b). Based on prior reports that addressed signaling from insulin receptors to PI3K in rat adipocytes (30), it was anticipated that mutant p85 would act as a dominant negative toward PI3K signaling. In representative clonal isolates of stably transfected cell lines, mRNAs encoding bovine p85 were present (Fig. 2a), and antibodies to bovine anti-p85 detected either wild-type (in cell line w14) or mutant forms (in cell line ⌬7) of bovine p85 but not human p85 (in untransfected cell line U1.wt) (Fig. 2b, top). Bovine p85 in ⌬7 cells migrated at a lower position on SDS-PAGE because of the 35 amino acid deletion. The abundance of human p85 was unaffected by overexpression of bovine p85 (Fig. 2b, bottom).
Unexpectedly, wild-type but not mutant bovine p85 exerted a dominant-negative function toward activation of PI3K by either epidermal growth factor (EGF) or type I IFN (Fig. 2, c and  d). In particular, Akt was activated in response to EGF in ⌬7 cells after 2 min, with peak activation at 8 min and decaying after 15 min ( Fig. 2c and Table I). Further, the time course of Akt activation by IFN-␤ was equivalent in U1.wt ⌬7 cells (compare Figs. 1a and 2d, Table I). However, IFN-␤-mediated Akt activation was not detected in w14 cells (overexpressing wild-type bovine p85) (Fig. 2d, Table I).
This unexpected dominant-negative effect of wild-type bovine p85 was characterized further. Initial experiments confirmed that bovine p85 and human p110␣ were found to be associated only in w14 cells (Fig. 2e). Taken together with the failure of EGF-or IFN-␤-induced Akt activation (Fig. 2, c and d ,  Table I), this result suggested that wild-type but not mutant bovine p85 might sequester human p110␣ in nonfunctional complexes. A corollary of this prediction was that nonfunctional bovine p85/human p110 complexes might exert dominant-negative effects toward IFN signaling to PI3K by displacing human p85/p110 complexes from association with human IFNAR-1. To address this issue, we immunoprecipitated IF-NAR-1 and examined its association with human p85 by West-ern analysis. Anti-IFNAR-1 immunoprecipitates contained human p85 only in ⌬7 and wild-type (untransfected) cells, but not w14 cells (Fig. 2f). These results suggested that bovine p85/ human p110␣ complexes were not functional for signaling to Akt. Interestingly, it appeared that human p85 did not associate with IFNAR-1 unless complexed with p110. These results (Fig. 2, a-f) indicated that w14 cells could be used as a genetic model for disruption of signaling from the type I IFN receptor to PI3K.
RNase protection assays revealed induction of ␤-R1 in ⌬7 but not in w14 cells (Fig. 3, a). Expression of the 6 -16 gene was FIG. 2. Wild-type but not mutant bovine p85 exerts a dominant-negative function toward cytokine and growth factor signaling to PI3K in human cells. Panel a, reverse-transcriptase polymerase chain reaction (RT-PCR) analysis was performed using RNA isolated from cells stably transfected with wild-type p85 (w14) or mutant p85 (⌬7) cDNA. The PCR products were separated on 1% agarose gels by electrophoresis and DNA visualized by ethidium bromide staining. PCR products in w14 and ⌬7 migrated at the expected sizes of 472 bp and 370 bp, respectively. Panel b, whole cell extracts were prepared from parental U1.wt cells and w14 and ⌬7 cells. Cell lysates were resolved by SDS-PAGE, transferred to polyvinylidene difluoride membranes, and probed with anti-bovine p85 (top panel), which detects both forms of the bovine protein but does not detect the human protein in U1.wt. Blots were stripped and reprobed with anti-human p85 (bottom panel), demonstrating that levels of the endogenous p85 were equivalent in all three cell lines. Panel c, w14 and ⌬7 cells were treated with recombinant EGF (10 ng/ml) for various times as indicated or reserved as untreated controls. Western analysis was performed to detect phosphorylated Akt using anti-phospho S473 Akt (top panel) and anti-Akt to determine total Akt protein (bottom panel). Panel d, w14 or ⌬7 cells were treated with recombinant IFN-␤-1b (5000 units/ml) for various times as indicated or reserved as untreated controls. Western analysis was performed to detect phosphorylated Akt using anti-phospho S473 Akt (top panel) and anti-Akt to determine total Akt protein (bottom panel). Panel e, whole cell extracts from w14, ⌬7, and wt cells were immunoprecipitated with antibody to human p110␣. Proteins were separated by SDS-PAGE, blotted onto polyvinylidene difluoride membranes and probed with antibodies to bovine p85␣. Bovine p85 was associated with human p110␣ in w14 cells overexpressing wild-type bovine p85 but not in ⌬7 cells expressing mutant p85. Panel f, whole cell extracts from w14, ⌬7, and wt cells were immunoprecipitated with human IFNAR-1 antibody. Proteins were seperated by SDS-PAGE, blotted onto polyvinylidene difluoride membranes, and probed with antibody to human p85␣. Lack of association of bovine p85 in w14 cells with the human IFN receptor was observed.
induced in both cell lines, with a 2-fold higher induction in ⌬7 cells. These results were representative of those obtained in two clones expressing wild-type bovine p85 and two clones expressing mutant p85. Also, IFN-␤-mediated induction of the ␤-R1 gene but not the 6 -16 gene in ⌬7 cells was sensitive to inhibition by LY294002 (results not shown). Taken with observations using chemical inhibitors of PI3K (Fig. 1d), these results supported the hypothesis that IFN-␤-mediated signaling through PI3K to Akt was selectively required for the induction of ␤-R1. These stable transfectants also provided an opportunity to address the role of IFN-␤-induced PI3K activity in the phosphorylation of the p65 TAD. To examine this question we prepared nuclear extracts from w14, ⌬7, and wild-type cells. IFN-␤ induced a nuclear activity that increased p65 TAD phosphorylation in ⌬7 and wild-type cells, but not in w14 cells (Fig.  3, b and c).
Overexpression of PTEN or Dominant-negative Akt Blocked Induction of a ␤-R1 Promoter Reporter by IFN-␤-Additional genetic approaches were taken to disrupt either signaling by PI3K to Akt or events downstream of Akt. These experiments permitted us to evaluate the effects on IFN-␤-induced expression of a ␤-R1 promoter reporter. To block PI3K signaling to Akt, we transfected cells with an expression construct encoding PTEN, a dual-specificity phosphatase that dephosphorylates phosphatidylinositol(3,4,5)P 3 at the 3 position in vitro and in vivo (31)(32). This activity precludes signaling downstream of PI3K to Akt. Overexpression of PTEN blocked IFN-␤-mediated stimulation of luciferase activity directed by the ␤-R1 promoter-reporter construct (Fig. 4a). PTEN also encodes a protein phosphatase activity. None of three PTEN mutants that lacked lipid phosphatase activity suppressed induction of ␤-R1 (Fig.  4a), although two mutants, G129E and D92A, retained protein phosphatase activity. Therefore, the ability of PTEN to suppress IFN-␤-dependent transcription of ␤-R1 depended on its lipid phosphatase activity.
This result suggested that signaling by PI3K to Akt was specifically involved in the pathway by which IFN-␤ induced ␤-R1 transcription. This hypothesis was addressed by overexpression of a dominant-negative catalytically inactive Akt mutant. Overexpression of catalytically inactive Akt selectively suppressed induction of the wt-␤-R1 promoter reporter by IFN-␤ (Fig. 4b), indicating a dominant-negative effect in this system. As a control for nonspecific effects of Akt inhibition, we analyzed IFN-␤-mediated induction of a promoter-reporter con-  struct derived from the p561 gene (a type I IFN-induced gene). Induction of the 561 promoter reporter was unaffected by overexpression of catalytically inactive Akt (Fig. 4b). Similar results were obtained by overexpression of a separate dominantnegative Akt mutant (results not shown).

DISCUSSION
Recent published reports delineated a role for PI3K and Akt in phosphorylating p65 in cytokine signaling pathways (21)(22)(23)(24). Our previous work showed that the p85 subunit of PI3K was associated with IFNAR-1 and that an inducible PI kinase activity could be co-precipitated with IFNAR-1 after IFN-␤ treatment. We also showed that ␤-R1 induction by IFN-␤ required NF-B and was associated with phosphorylation of the p65 TAD. Taken together, these observations raised the question of whether PI3K played a similar role in type I IFN signaling, an issue that is addressed in this report.
In this regard PI3K activity is induced by IFN-␣/␤ by a mechanism that appears to differ in various cell lines (8,(15)(16)(17)(18)(19). This conservation of PI3K activation despite varied mechanisms suggests an important function in IFN signaling. In Daudi cells, STAT3 serves as an adapter and recruits PI3K to the interferon-alpha receptor (IFNAR) complex in an SH2-dependent interaction (15)(16). In other cells of hematopoietic origin, association of PI3K with activated STAT3 could not be demonstrated (17), and a role for IRS-1 in recruiting p85/p110 PI3K was demonstrated in these cells (18,8). In HT1080derived fibrosarcoma cells, IFN-inducible PI kinase activity was detected in anti-IFNAR-1 immunoprecipitates (19).
Activation of PI3-kinase increases intracellular phosphatidylinositol-3,4-P2 and phosphatidylinositol-3,4,5-P3. The phosphorylated lipid products act as second messengers. Various intracellular downstream targets of PI3K have been identified (33)(34). Among them, the serine-threonine kinase Akt has received special attention because of its central role in regulation of cell survival or death (35). Akt is encoded by the Akt proto-oncogene and has an amino-terminal pleckstrin homology (PH) domain, which is essential for binding of phosphatidylinositol 3,4-P2. Activation of Akt by PI3K is dependent on the integrity of the PH domain (36) and is susceptible to wortmannin (36 -37). In vitro incubation of inactive Akt from serum-starved NIH3T3 cells with synthetic PPIs activated Akt in a dose-dependent manner, indicating Akt is a direct target of PI3K (36). In this report we demonstrate that IFN-␤-induced activation of Akt is sensitive to blockade by LY294002 (Fig. 1b), placing it downstream of PI3K. Functional significance for this signaling pathway was documented by showing that overexpression of catalytically inactive Akt selectively abrogated induction of a ␤-R1 promoter-reporter construct (Fig. 4b).
Pretreatment of cells with either of two different pharmacological inhibitors of PI3K inhibited the induction of endogenous ␤-R1 mRNA by IFN-␤ (Fig. 1d). The effect was selective since a variety of other IFN-induced genes including 6 -16 (Fig. 1d), p56, ISG-54, and 9 -27 (results not shown) were unaffected by PI3K inhibition. To exclude nonspecific effects of chemical PI3K inhibitors and further address the mechanism by which PI3K promotes induction of ␤-R1, we constructed stable cell lines expressing the full-length cDNA of bovine p85 or a mutant cDNA of bovine p85 (26). The mutant ⌬p85 lacks the ability to bind p110 and functions as a dominant negative toward the rodent insulin receptor in adipocytes (30). We found that the ⌬7 cell line expressing ⌬p85 was fully permissive for induction of Akt activation and ␤-R1 by IFN-␤ (Figs. 2d and 3a) in contrast to its phenotype in other systems (30). Unexpectedly, the w14 cell line expressing wt-p85 failed to support the Akt activation or induction of ␤-R1 by IFN-␤ (Figs. 2d and 3a), indicating that wild-type bovine p85 acted as a dominant negative for IFN signaling to PI3K.
These results suggested that bovine p85/human p110 complexes were nonfunctional in the IFN signaling pathway. Further, PI3K was not activated in w14 cells after treatment with EGF (Fig. 2c), raising the possibility that bovine p85/human p110 complexes were non-responsive to various cytokines and growth factors. Additional analysis revealed that bovine p85 was co-precipitated with human p110 (Fig. 2e), supporting the possibility that the dominant-negative effect was mediated by sequestering endogenous p110 in inactive complexes.
It remained to be determined if human p85 could associate with IFNAR-1 alone or only in complex with p110. We found (Fig. 2f) that human p85 was not detected in IFNAR-1 immunoprecipitates from w14 cells, although levels of the protein in cytosolic extracts were not reduced (Fig. 2b, bottom panel). Our results indicate that bovine wild-type p85 protein associates in nonfunctional complexes with the human IFNAR-1 and p110 (Fig. 2, e and f) and exerts dominant-negative function for type I IFN signaling to PI3K and downstream mediators (Figs. 2d and 3b). In these experiments ⌬7 cells expressing bovine ⌬p85 protein provide a useful negative control, as the mutant protein, being unable to associate with p110 (Fig. 2e), leaves PI3K signaling by cytokines and growth factors (Figs. 2, c and d and  3, a and b) unaffected. We speculate that activation of the PI3K pathway in fibrosarcoma cells is unlikely to utilize IRS-1 as an adaptor, since the bovine deletion mutant ⌬p85 acts as a dominant negative in signaling through the insulin receptor in both rat adipose cells and Chinese hamster ovary cells (30, 38 -39).
Other genetic experiments supported this model of IFN signaling to PI3K through Akt. PTEN is a natural antagonist of PI3K activity. Lipid products of PI3K, critical for activation of Akt, are dephosphorylated at the 3-OH position by the lipid phosphatase activity of PTEN (40). Lipid products of PI3K activity target Akt to the plasma membrane where it is fully activated through phosphorylation of Ser-473 and Thr-308. PTEN possesses activity toward both phospholipid and protein.
Overexpression of wild-type PTEN abrogated the induction of a ␤-R1 promoter reporter in IFN-␤-treated cells (Fig. 4a). By comparing the suppressive effects of overexpressing wild-type PTEN, D92A mutant PTEN, G129E mutant PTEN, and C124A mutant PTEN (41) we addressed whether the phosphoprotein or phospholipid products of PI3K activity were essential for IFN-␤-mediated induction of ␤-R1 (Fig. 4a). The C124A mutant is catalytically inactive, retaining neither lipid nor protein phosphatase activity. Both the G129E and D92A mutations result in loss of lipid phosphatase activity with retention of protein phosphatase activity (41). Only wild-type PTEN suppressed IFN-␤ induction of ␤-R1, indicating that phospholipid products of PI3K were essential for the induction of ␤-R1. This result suggested that Akt activation would also prove to be required for ␤-R1 induction by IFN-␤. This hypothesis was tested by co-transfecting a plasmid encoding a catalytically inactive Akt mutant along with the ␤-R1 promoter reporter. Expression of the dominant-negative Akt also suppressed induction of the ␤-R1 promoter reporter by IFN-␤ (Fig. 4b). Furthermore, nuclear extracts prepared from IFN-␤-treated w14 cells did not mediate phosphorylation of the p65 TAD (Fig. 3, b  and c). These results delineated an accessory signaling pathway that supplements JAK-STAT signaling and show that this pathway is essential for the induction of one gene, ␤-R1, by IFN-␤. The accessory signaling pathway involves the activation of PI3K and one downstream effector, Akt, and leads to the phosphorylation of the p65 subunit of NF-B. This scheme is summarized in Fig. 5. As shown in Fig. 5, it remains unclear how PI3K/Akt mediated signals lead to cytokine-induced NF-B p65 phosphorylation. Akt mediates IKK␣ phosphorylation at Thr-23, and mutation of this Thr residue blocked both phosphorylation by Akt and liberation of NF-B by TNF-␣ (42). The interaction of Akt with IB kinase can be constitutive (42) or induced by platelet-derived growth factor (PDGF) (43). Interestingly, IFN-␤ treatment leads to phosphorylation but not degradation of IB␣ in HT1080-derived cells (7), suggesting the possibility that signaling through IKK␣ may occur in this pathway.
In summary, we provide evidence that IFN-␤ signals to PI3K with Akt as a downstream effector to enhance the transactivation competence of NF-B through phosphorylation of p65 within the TAD. Along with ISGF-3, this pathway is required for the transcriptional induction of the ␤-R1 gene by IFN-␤. The downstream effectors of PI3K and Akt required for the transactivation of p65 remain to be identified.