Phosphorylation of Serine 256 by Protein Kinase B Disrupts Transactivation by FKHR and Mediates Effects of Insulin on Insulin-like Growth Factor-binding Protein-1 Promoter Activity through a Conserved Insulin Response Sequence*

Insulin inhibits the expression of multiple genes in the liver containing an insulin response sequence (IRS) (CAAAA(C/T)AA), and we have reported that protein kinase B (PKB) mediates this effect of insulin. Genetic studies inCaenorhabditis elegans indicate that daf-16, aforkhead/winged-helix transcription factor, is a major target of the insulin receptor-PKB signaling pathway. FKHR, a human homologue of daf-16, contains three PKB sites and is expressed in the liver. Reporter gene studies in HepG2 hepatoma cells show that FKHR stimulates insulin-like growth factor-binding protein-1 promoter activity through an IRS, and introduction of IRSs confers this effect on a heterologous promoter. Insulin disrupts IRS-dependent transactivation by FKHR, and phosphorylation of Ser-256 by PKB is necessary and sufficient to mediate this effect. Antisense studies indicate that FKHR contributes to basal promoter function and is required to mediate effects of insulin and PKB on promoter activity via an IRS. To our knowledge, these results provide the first report that FKHR stimulates promoter activity through an IRS and that phosphorylation of FKHR by PKB mediates effects of insulin on gene expression. Signaling to FKHR-related forkheadproteins via PKB may provide an evolutionarily conserved mechanism by which insulin and related factors regulate gene expression.

Insulin exerts important effects on gene expression in multiple tissues (1). In the liver, insulin suppresses the expression of a number of genes that contain a conserved insulin response sequence (IRS) 1 (CAAAA(C/T)AA), including insulin-like growth factor-binding protein-1 (IGFBP-1), apolipoprotein CIII (apoCIII), phosphoenolpyruvate carboxykinase (PEPCK), and glucose-6-phosphatase (2)(3)(4)(5)(6). This observation suggests that insulin may regulate the expression of multiple hepatic genes through a common mechanism. Insulin rapidly suppresses the expression of IGFBP-1 and PEPCK at the transcriptional level, and this effect is not disrupted by pretreatment with cycloheximide (7,8), indicating that it is mediated by post-translational modification of pre-existing factors, perhaps by their phosphorylation. Specific factors that mediate the inhibitory effects of insulin on hepatic gene expression through a conserved IRS remain to be identified.
Recent studies indicate that protein kinase B (PKB) functions downstream from phosphatidylinositol 3Ј-kinase (PI3K) in the insulin signaling pathway (9,10) and that it plays an important role in mediating effects of insulin and related growth factors on glucose and amino acid transport, glycogen and protein synthesis, and cell survival (11)(12)(13)(14)(15)(16)(17)(18)(19). Following its activation, PKB is translocated to the nucleus where it may exert effects on gene expression (20,21). Activated PKB increases the expression of leptin and fatty acid synthase in adipocytes (22,23) and suppresses PEPCK mRNA levels in liver-derived cells stimulated by cAMP and glucocorticoids (24), mimicking the effects of insulin. Based on studies using pharmacological inhibitors and dominant negative and constitutively active forms of signaling peptides, we recently reported that PKB is both necessary and sufficient to mediate sequencespecific effects of insulin on basal promoter activity through an IRS (25). We now have sought to identify downstream mechanisms that mediate effects of insulin on gene expression through an IRS downstream from PKB.
Recent genetic studies in Caenorhabditis elegans indicate that specific members of the forkhead/winged-helix family of transcription factors may be major targets of insulin receptor signaling downstream from PKB. Mutation of the insulin/IGF-I receptor homologue (daf-2), the catalytic subunit of PI3K (age-1), or PKB (akt-1 and akt-2) in C. elegans results in increased longevity and constitutive dauer formation (26 -28), a stage of developmental arrest and reduced metabolic activity that enhances survival during periods of food deprivation and other environmental stresses (29). In each case, mutation of daf-16, which codes for a forkhead/winged-helix transcription factor, restores a normal life span and prevents entry into the dauer stage (28,30,31). These observations have suggested that daf-16 promotes entry into the dauer phase and enhanced longevity and that signaling via the insulin/IGF-I receptor-PI3K-PKB pathway may disrupt these effects of daf-16 (28,30,31).
Interestingly, analysis based on a consensus sequence for phosphorylation by PKB (Arg-Xaa-Arg-Xaa-Xaa-(Ser/Thr)-Hyd) (32) indicates that daf-16 contains four PKB phosphorylation sites (28), suggesting that it may be a direct target for signaling by PKB. Also, Ruvkun and co-workers (28) have noted that daf-16 can interact directly with an IRS in vitro. Several of these PKB sites are conserved in a group of closely related human forkhead proteins, including FKHR, FKHRL1, AFX1, and AF6q21 (33)(34)(35)(36). FKHR is expressed in the liver (34) and has three of these conserved PKB sites, including Thr-24, Ser-256, and Ser-319. In separate studies, we have found that each of these sites in FKHR is phosphorylated by PKB in vitro and in cells (37). The DNA binding domains of FKHR and daf-16 also are highly conserved, suggesting that FKHR may interact with and regulate promoter activity through an IRS in vivo. As a first step in evaluating the role that FKHR and/or closely related forkhead proteins may play in mediating effects of insulin on hepatic gene expression, we asked whether FKHR may regulate the activity of the IGFBP-1 promoter through a conserved IRS and mediate effects of insulin on promoter activity downstream from PKB.
All mutations were verified by dideoxy sequencing. The BamHI-NotI fragment containing 511 bp of the FKHR cDNA including 224 bp of 5Ј-untranslated sequence also was excised from pFB-12A2 and cloned into pcDNA3.1(Ϫ) (Invitrogen) in the antisense orientation (FKHR.AS) for antisense studies.
Cell Culture and Transient Transfection Studies-HepG2 cells plated in 60-mm dishes were transfected in triplicate with calcium phosphate precipitates containing equal amounts of DNA including reporter gene and expression vectors together with appropriate amounts of empty vector (pcDNA3), as previously reported (25). Transfected cells were refed with Dulbecco's modified Eagle's medium plus 1 g/liter fatty acid-free bovine serum albumin (Sigma) with/without 100 nM recombinant human insulin (Sigma) and/or 50 M PD98059 (Calbiochem), 50 M LY294002 (Calbiochem), or 200 nM rapamycin (Sigma) 18 h prior to the preparation of lysates and analysis of luciferase and ␤-galactosidase activity as reported (25).
Northern Blotting and RT-PCR-HepG2 cells were grown to ϳ70% confluency and then stabilized in serum-free medium for 24 h before total cellular RNA was prepared with TRI-Reagent-LS (Molecular Research). Total RNA (20 g) was loaded for gel electrophoresis, visualized with ethidium bromide, and transferred to a nitrocellulose mem-brane. The PstI-XhoI fragment containing 506 bp of the FKHR cDNA, including 133 bp of 3Ј-untranslated sequence, was excised from pFB.12A2 and labeled by random priming for use as a probe. Transcripts were identified by autoradiography.

RESULTS
To determine whether FKHR may influence the expression of genes that are regulated by insulin through an IRS, we first performed transient transfection studies in HepG2 cells with a luciferase reporter gene construct containing the IGFBP-1 promoter (BP1.Luc) and a ␤-galactosidase expression vector driven by the CMV promoter. As shown in Fig. 1, co-transfection with an FKHR expression vector stimulates the IGFBP-1 promoter and luciferase activity in a dose-dependent fashion without altering levels of ␤-galactosidase expression, indicating that this effect is promoter-specific.
We next examined whether this effect of FKHR is mediated through an IRS. The IGFBP-1 promoter contains 2 IRSs located ϳ100 bp 5Ј to the RNA cap site (IRSA (CAAAACAA) and IRSB (TTATTTTG)), and each is sufficient to mediate negative effects of insulin on promoter activity (3,25). As shown in Table  IA, mutation of IRSA and IRSB (BP1.mut) disrupts both the ability of insulin to inhibit and the ability of FKHR to stimulate promoter activity, whereas the presence of either IRSA (⌬IRS.A) or IRSB (⌬IRS.B) alone is sufficient to mediate effects of insulin and FKHR on promoter function. Mutation of a single base pair within either IRSA (⌬IRS.Amut) or IRSB (⌬IRS.Bmut) disrupts the effects of both insulin and FKHR on promoter activity, suggesting that both insulin and FKHR exert their effects in a sequence-specific fashion. Placing IRSA 3 bp 5Ј from its native location to disrupt potential interactions with flanking sequences also mediates effects of insulin and FKHR on promoter activity, whereas placing a mutated sequence at this location (⌬IRS.1M) is not effective. As shown in Table IB, introducing an array of IRSs (TK81.IRS3) immediately upstream from the thymidine kinase promoter also is sufficient to confer effects of both insulin and FKHR on promoter activity. In contrast, introducing another sequence of similar length containing two Gal4 binding sites at this location is not effective. Taken together, these results indicate that FKHR, like insulin, exerts effects on promoter activity through an IRS.
To determine whether the effects of insulin and FKHR on promoter activity are mediated through an IRS with similar sequence specificity, we created a series of reporter gene constructs where individual residues within a single IRS are mutated one at a time. As shown in Table II, point mutations which disrupt the ability of insulin to inhibit promoter activity also disrupt the ability of FKHR to stimulate promoter function. Fig. 2 shows the relationship between the ability of insulin to inhibit and the ability of FKHR to stimulate promoter activity in this series of reporter gene constructs, which is statistically significant (r ϭ 0.865, p Ͻ 0.01). Of note, constructs that contain the IRS found in the PEPCK or apoCIII gene (⌬IRS.1 m2 and ⌬IRS.1 m8) are responsive to both insulin and FKHR, suggesting that FKHR also may stimulate the activity of the PEPCK and apoCIII promoters through related IRSs. Together, these results indicate that FKHR and the endogenous factor(s) responsible for mediating the effects of insulin and PKB on promoter activity interact with the IRS with similar sequence specificity and suggest the possibility that endogenous FKHR or closely related proteins may contribute to the regulation of IGFBP-1 promoter activity through an IRS.
To address this question, we first determined whether FKHR is expressed in HepG2 cells. As shown in Fig. 3A, Northern blotting with a 32 P-labeled FKHR cDNA probe reveals the presence of an ϳ6.5-kb transcript in total cellular RNA pre-pared from HepG2 cells, similar to results obtained in human liver (34). As shown in Fig. 3B, RT-PCR with primers for the DNA binding domain (FKHR.DBD), another fragment of FKHR (FKHR), or GAPDH generated cDNA fragments of appropriate size (366, 437, and 189 bp, respectively). The FKHR.DBD and FKHR PCR products were cloned and sequenced completely. The sequences of these cDNAs (not shown) agreed entirely with the reported sequence for the FKHR cDNA (34), confirming that FKHR is expressed in HepG2 cells.
To test whether endogenous FKHR may contribute to the regulation of IGFBP-1 promoter activity in HepG2 cells, we next performed studies with a CMV-driven vector that expresses 244 bp of the 5Ј-untranslated region and the first 287 bp of the FKHR coding region in an antisense orientation. As shown in Fig. 4A, this construct disrupts the ability of FKHR to stimulate promoter activity through an IRS but does not interfere with the ability of HNF-3␤ to stimulate promoter function through a consensus HNF-3 binding sequence placed at the same location. This result indicates that this antisense construct disrupts IRS-dependent transactivation by FKHR selectively.
Subsequent studies with this antisense construct revealed that it reduces the activity of the IGFBP-1 promoter activity in a dose-dependent fashion without decreasing the level of ␤-galactosidase expressed by a CMV-driven vector, indicating that this effect is promoter-specific (data not shown). As shown in Fig. 4B, the ability of FKHR antisense RNA to inhibit activity of the IGFBP-1 promoter (BP1.Luc) is abolished when both IRSA and IRSB are mutated (BP1.mut), whereas the presence of either IRSA (⌬IRS.A and ⌬IRS.1) or IRSB (⌬IRS.B) alone is sufficient to render the IGFBP-1 promoter responsive to inhibition by FKHR antisense RNA. Similarly, introducing an array of IRSs 5Ј to the TK promoter is sufficient to confer a negative effect of FKHR antisense RNA on the activity of this heterologous promoter. These results support the concept that endogenous FKHR and/or closely related proteins contribute to basal promoter activity through an IRS-dependent mechanism.
We next used this antisense construct to determine whether endogenous FKHR may be required for the ability of insulin and PKB to inhibit basal promoter activity through an IRS. Previous studies have shown that insulin inhibits IGFBP-1 promoter activity through an IRS (2, 3) and that PKB is nec-TABLE I IRS-dependent effects of insulin and FKHR HepG2 cells were transfected with luciferase reporter gene constructs containing the native IGFBP-1 promoter (BP1.Luc) or constructs where IRSA (CAAAACAA) and/or IRSB (TTATTTTG) have been altered as described under "Materials and Methods." To examine effects of insulin on promoter activity, cells were transfected with 10 g of DNA/dish including 3 g of reporter gene construct plus empty vector, then refed with serum-free medium with/without 100 nM insulin 18 h prior to lysis and analysis of luciferase activity. To examine the effects of FKHR on promoter activity, cells were transfected with 10 g of DNA/dish including 3 g of reporter gene construct with/without 3 g of a CMV-driven FKHR expression vector and appropriate amounts of empty vector and then stabilized in serum-free medium prior to analysis of luciferase activity. The effects of insulin and FKHR on promoter activity are expressed as the percent inhibition or stimulation relative to control and reported as the mean Ϯ S.E. essary and sufficient to mediate this effect of insulin (25). As shown in Fig. 4C, expression of FKHR antisense RNA reduces the activity of the native IGFBP-1 promoter (BP1.Luc), and there is no additional reduction in promoter function by insulin or PKB in the presence of FKHR antisense RNA. Studies with reporter gene constructs containing either IRSA (⌬IRS.A) or IRSB (⌬IRS.B) alone or an array of IRSs introduced upstream from the thymidine kinase promoter (TK.IRS3) yielded similar results (Fig. 4C). In each case, FKHR antisense RNA reduces promoter activity and there is no additional inhibitory effect of insulin or PKB on promoter function in combination with FKHR antisense RNA. We next performed studies to determine whether FKHR antisense RNA disrupts the ability of insulin and PKB to inhibit promoter activity through an IRS selectively or also disrupts other effects of insulin and PKB in HepG2 cells. As shown in Fig. 4D, FKHR antisense RNA, insulin, and PKB each reduce promoter activity to a similar extent in a reporter gene construct containing a single IRS (⌬IRS.1), and there is no additional effect of insulin or PKB on promoter activity in combination with FKHR antisense RNA, consistent with results obtained with other constructs containing an IRS (Fig.  4C). In contrast, we observed that insulin and PKB both stimulate promoter activity when the IRS is replaced by an HNF-3 binding site (⌬IRS.HNF-3) and that FKHR antisense RNA does not disrupt this effect of insulin or PKB (Fig. 4D, right panel). This result confirms that FKHR antisense RNA disrupts the ability of insulin and PKB to suppress promoter activity through an IRS without disrupting other effects of insulin and PKB on promoter activity in HepG2 cells.

Construct
Taken together, these results indicate that FKHR, or a closely related factor, is required for the ability of insulin and PKB to suppress promoter activity through an IRS and suggest the possibility that insulin and PKB may inhibit promoter   Table II). The effects of insulin (percent inhibition) and FKHR (percent stimulation) on promoter activity for each construct and the linear regression line are shown. The relationship between the effect of insulin and FKHR on promoter activity was examined for significance by Pearson correlation analysis.

FIG. 3. Expression of FKHR mRNA in HepG2 cells.
A, Northern blotting. Total cellular RNA was prepared from HepG2 cells, and 20 g was loaded together for gel electrophoresis and transfer onto a nitrocellulose membrane. The membrane was probed with a 32 P-labeled FKHR cDNA probe, and transcripts were identified by autoradiography. B, RT-PCR of FKHR. RT-PCR was performed with 2 g of total RNA from HepG2 cells using 2 sets of primers for FKHR (FKHR.DBD and FKHR) and primers for GAPDH. PCR products were loaded for gel electrophoresis with a DNA sizing ladder and stained with ethidium bromide. A control reaction was performed in the absence of reverse transcriptase prior to PCR with the FKHR.DBD primers to exclude the possibility of contamination by genomic DNA (DBD.Bkg). activity largely by disrupting IRS-dependent transactivation by FKHR.
Based on these findings and previous studies indicating that PKB is necessary and sufficient for insulin to inhibit promoter activity through an IRS (25), we next examined whether insulin inhibits FKHR-stimulated promoter activity and whether PKB mediates this effect of insulin. As previously demonstrated, co-transfection with 1 g of a CMV-driven FKHR expression vector together with a reporter gene construct con-taining a single IRS (⌬IRS.1) results in a 7-fold stimulation of promoter activity (Table II). As shown in Fig. 5A, insulin inhibits FKHR-stimulated promoter activity by ϳ50%. This effect of insulin is not blocked by treatment with PD98059, a specific inhibitor of the activation of MAP kinase kinase 1 (MAPKK1) (43), or co-transfection with a dominant negative form of Raf (C4bRaf) (Fig. 5A), indicating that it is not mediated through the Ras-Raf-MAPKK1-MAPK pathway. In contrast, both treatment with LY294002, a highly specific inhibitor of PI3K (44), and co-transfection with a dominant negative form of the 85-kDa regulatory subunit of PI3K (⌬p85) block this effect of insulin completely, indicating that insulin suppresses FKHRstimulated promoter activity through a PI3K-dependent mechanism. Rapamycin, which prevents the activation of p70 S6 kinase downstream from PI3K (45), does not disrupt this effect of insulin. In contrast, expression of a kinase-deficient, dominant negative form of PKB (Lys-179-PKB) blocks the effect of insulin, whereas constitutively active PKB (Myr-PKB) inhibits FKHR-stimulated promoter activity, mimicking the effect of insulin. Together, these results indicate that PKB is necessary and sufficient to mediate the effect of insulin on FKHR-stimulated promoter activity.
FKHR contains three consensus PKB phosphorylation sites (Thr-24, Ser-256, and Ser-319), and we have found that PKB phosphorylates each of these sites in vitro and in cells (37). To determine whether the phosphorylation of these sites is required for insulin or PKB to disrupt transactivation by FKHR, we first mutated Thr-24, Ser-256, and Ser-319 to alanine individually (Thr-24 -Ala, Ser-256 -Ala, and Ser-319 -Ala FKHR) and together ((Thr/Ser/Ser)-Ala FKHR). Mutation of these residues to alanine does not disrupt the ability of FKHR to stimulate promoter activity (Fig. 5B). However, overexpression of (Thr/Ser/Ser)-Ala and Ser-256 -Ala (but not Thr-24 -Ala or Ser-319 -Ala) FKHR completely abolishes the ability of insulin and PKB to inhibit FKHR-stimulated promoter activity (Fig. 5B). Similar studies with the TK.IRS3 construct confirm that overexpression of (Thr/Ser/Ser)-Ala or Ser-256 -Ala (but not Thr-24 -Ala or Ser-319 -Ala) abolishes the ability of insulin and PKB to inhibit promoter function (data not shown). These findings indicate that phosphorylation of Ser-256 is required for the ability of insulin and PKB to disrupt IRS-dependent transactivation by FKHR.
To determine whether the introduction of a negative charge at these sites is sufficient to disrupt the ability of FKHR to stimulate promoter activity, we next mutated Thr-24, Ser-256, or Ser-319 to aspartate. As shown in Fig. 5C, mutation of Ser-256 to aspartate (Ser-256 -Asp) disrupts IRS-dependent transactivation by FKHR. In contrast, Thr-24 -Asp and Ser-319 -Asp mutations do not disrupt the ability of FKHR to stimulate promoter activity through an IRS-dependent mechanism (⌬IRS.1 versus ⌬IRS.1M). Similar studies with the TK.IRS3 construct confirm that mutation of Ser-256 (but not Thr-24 or Ser-319) to aspartate disrupts transactivation by FKHR (not shown). Taken together, these results indicate that phosphorylation of Ser-256 by PKB and the introduction of a negative charge at this site is necessary and sufficient to disrupt IRS-dependent transactivation by FKHR.
expressing dominant negative forms of Raf (C4bRaf) (5 g), the regulatory subunit of PI3K (⌬p85) (5 g), or PKB (Lys-179 -PKB) (10 g) or constitutively active PKB (Myr-PKB) (1 g) as indicated. Cells were refed with serum-free medium with/without 100 nM insulin and/or 50 M LY294002, 50 M PD98059, or 200 nM rapamycin for 18 h before lysates were prepared for luciferase assay. B, effect of alanine mutations on the ability of insulin and PKB to disrupt transactivation by FKHR. Cells were transfected with the ⌬IRS.1 reporter gene construct and (Thr/Ser/Ser)-Ala, Thr-24 -Ala, Ser-256 -Ala, or Ser-319 -Ala FKHR expression vector with/without 1 g of vector expressing constitutively active PKB (Myr-PKB) plus appropriate amounts of empty vector. Cells were refed with serum-free medium with/without insulin 18 h before lysis and analysis of luciferase activity. C, effect of aspartate mutations on transactivation by FKHR. Thr-24, Ser-256, and Ser-319 were mutated individually to aspartate. Cells were transfected with 10 g of DNA/dish including 3 g of ⌬IRS.1 or ⌬IRS.M. reporter gene construct with/without 1 g of wild type or Thr-24 -Asp, Ser-256 -Asp, or Ser-319 -Asp FKHR and stabilized in serum-free medium prior to lysis and analysis of luciferase activity. In the present study, we sought to determine whether FKHR, a member of the forkhead/winged-helix family of transcription factors, may contribute to the regulation of gene expression by insulin and provide a target for mediating effects of insulin and PKB on gene expression through a conserved IRS. To date, 80 members of the forkhead family have been identified, and many have been found to play an important role in development and in the determination of tissue-specific gene expression (46,47). Genetic studies in C. elegans have suggested that daf-16 is a major target for signaling by the insulin/ IGF-I receptor-PI3K-PKB pathway, based on the effects of mutations on the development and survival of intact organisms (28,30,31). A preliminary report by Paradis and Ruvkun, together with Nasrin and Alexander-Bridges (28) also indicates that daf-16 may interact directly with an IRS in vitro. In the present study, we utilized reporter gene constructs in a mammalian cell culture model to demonstrate that FKHR, a human homologue of daf-16, may contribute to the regulation of promoter activity through an IRS in vivo. The results of these studies demonstrate that FKHR stimulates promoter activity in a highly sequence-specific fashion through an IRS and that phosphorylation of Ser-256 by PKB is necessary and sufficient for insulin to disrupt IRS-dependent transactivation by FKHR. To our knowledge, these findings provide the first direct evidence that FKHR-like forkhead proteins stimulate promoter activity through a conserved IRS and that phosphorylation of FKHR or closely related proteins may mediate sequencespecific effects of insulin on gene expression downstream from PKB.
Analysis based on a known consensus sequence for phosphorylation by PKB (32) suggested that FKHR contains three PKB phosphorylation sites (Thr-24, Ser-256, and Ser-319) and we have found in separate studies that PKB phosphorylates each of these sites in vitro and in cells (37). In the present study, we used pharmacological inhibitors and expression vectors for dominant negative and constitutively active forms of signaling peptides to determine that insulin inhibits FKHR-stimulated promoter activity through a mechanism mediated by PKB. Based on studies where Thr-24, Ser-256, and Ser-319 were mutated to alanine, a neutral amino acid that is not susceptible to phosphorylation by kinases, or to aspartate, which has a negative charge, we showed that phosphorylation of Ser-256 by PKB is necessary and sufficient for insulin to disrupt IRS-dependent transactivation by FKHR. It remains to be determined whether the phosphorylation of Thr-24 or Ser-319 affects other functions of FKHR.
As shown in Fig. 6, Ser-256 is located in the basic region of the DNA binding domain of FKHR. X-ray crystallography performed with HNF-3␥ indicates that this region of the forkhead/ winged-helix DNA binding motif forms a random coil within the minor groove of target sites where it may interact with phosphate residues and stabilize DNA/protein interactions (48). It is interesting to speculate that the phosphorylation of Ser-256 and the introduction of a negative charge at this site might reduce the stability of this interaction and disrupt binding. However, in vivo footprinting and gel shift studies with nuclear extracts have so far failed to detect a nucleoprotein complex involving an IRS whose formation is disrupted by insulin treatment (2,3,49), suggesting that other mechanisms also must be considered. Another possibility is that phosphorylation of Ser-256 might disrupt interactions between FKHR and a co-activating factor required for transactivation or induce the recruitment of a co-repressor. Preliminary immunocytochemical studies in this laboratory indicate that phosphorylation by PKB also might result in the redistribution of FKHR from the nucleus to the cytoplasm within cells. 2 Additional studies are in progress to examine the specific mechanism(s) by which the phosphorylation of Ser-256 might disrupt IRS-dependent transactivation by FKHR.
It is important to note that this PKB phosphorylation site is conserved in several closely related forkhead family members, including FKHRL1, AFX1, AF6q21, and daf-16 where the serine residue is replaced by a threonine (Fig. 6). Because FKHRL1, AFX1, and AF6q21, like FKHR, are expressed in the liver (34 -36), it is possible that they also may be expressed in HepG2 cells and contribute to the regulation of IGFBP-1 promoter activity. In this study, we used an antisense construct to examine the role that endogenous FKHR may play in the regulation of promoter activity. Although our FKHR antisense construct did not disrupt the ability of HNF-3␤ to stimulate promoter activity through an HNF-3 binding site, it is possible that it might disrupt the expression of more closely related proteins, including FKHRL1, AFX1, or AF6q21. The observation that this PKB phosphorylation site is conserved in these proteins suggests the possibility that they also may contribute to the regulation of hepatic gene expression by insulin down-2 J. Wu, K. Colley, and T. Unterman, unpublished observations. FIG . 6. Comparison of forkhead DNA binding domains for FKHR, FKHRL1, AFX1, AF6q21, daf-16, HNF-3␤, and HNF-3␥. The positions of ␣ helices (Helix 1, Helix 2, and Helix 3), ␤ strands (S1, S2, and S3), and wing domains (W1 and W2) and the basic region within the forkhead/winged-helix DNA binding domain are indicated. Residues with greater than 50% homology among FKHR-related proteins (FKHR, FKHRL1, AFX1, AF6q21, and daf-16) are shaded. Serine 256 in FKHR is identified by a star. The percentage of amino acids that are homologous relative to FKHR is shown. stream from PKB, together with FKHR. Because FKHR, FKHRL1, AFX1, and AF6q21 are expressed in many tissues (34 -36), it is reasonable to speculate that they also may play an important role in mediating the effects of insulin and other growth factors on gene expression in other settings.
The observation that this PKB phosphorylation site is present in daf-16 (Fig. 6) supports the concept that signaling through the insulin/IGF-I receptor/PI3K/PKB pathway may disrupt transactivation by daf-16 in C. elegans, consistent with results of genetic studies indicating that signaling via PKB disrupts effects of daf-16 on longevity and dauer formation (28,30,31). At the same time, the fact that this phosphorylation site is absent in HNF-3 proteins (Fig. 6) is consistent with results in the present study indicating that PKB no longer inhibits promoter activity when an IRS is replaced by a consensus HNF-3 binding site (Fig. 4D). Several other studies also have indicated that interactions with HNF-3 proteins are insufficient to mediate inhibitory effects of insulin on promoter activity through an IRS (50 -52). If insulin and/or PKB do alter the transcriptional activity of HNF-3 members of the forkhead family, it is likely that they do so by a different mechanism.
Another feature, which distinguishes FKHR-related proteins from other members of the forkhead family, is the insertion of five additional amino acids at the N-terminal region of helix 3 within the DNA binding domain (Fig. 6). Crystallographic studies indicate that helix 3 of the forkhead/winged-helix DNA binding motif is presented to the major groove of target sequences where it is thought to play a critical role in sequencespecific interactions (48). Sequences flanking helix 3 also have been found to be important in determining sequence specificity for forkhead proteins (53). It is interesting to speculate that the insertion of these additional amino acids at the N-terminal end of helix 3 may cause FKHR-related members of the forkhead family to interact with a distinct set of related target sequences.
In this context, it is interesting to note similarities in the roles played by daf-16 in C. elegans and genes known to be regulated through an IRS in mammals. In C. elegans, daf-16 function is required for wild-type organisms to enter the dauer phase and enhance survival in response to nutrient deprivation (29). Similarly, several genes that are regulated by insulin in the liver through an IRS are important in the adaptation to nutritional restriction in mammals. The abundance of hepatic IGFBP-1 mRNA and circulating levels of IGFBP-1 are increased 10-fold in short-term fasting (54 -56) where high IGFBP-1 levels are thought to limit the anabolic effects of IGFs, sparing amino acid substrates for functions critical for survival, including gluconeogenesis. Hepatic PEPCK and glucose-6-phosphatase mRNA levels also are increased in fasting where PEPCK and glucose-6-phosphatase play a critical role in increasing the production and secretion of glucose by the liver (57)(58)(59)(60). It is interesting to speculate that FKHR-related forkhead proteins and their phosphorylation via the insulin/IGF-I receptor-PI3K-PKB pathway may play an important and evolutionarily conserved role in regulating metabolism at the genetic level in response to changes in nutrient availability.
The identification of a signaling pathway that may mediate effects of insulin on the expression of genes known to play an important role in the regulation of hepatic glucose production also has significant clinical implications. Unrestrained gluconeogenesis contributes to the pathogenesis of fasting hyperglycemia, the hallmark of diabetes mellitus (61). IGFBP-1 is produced largely by the liver, and hepatic production of IGFBP-1 is potently suppressed by insulin at the level of gene transcription (62,63). Recent studies indicate that circulating levels of IGFBP-1 are elevated in patients with Type 2 diabetes mellitus despite high levels of insulin (64). Taken together, these observations suggest that the ability of insulin to suppress hepatic production of IGFBP-1 may be impaired in patients with Type 2 diabetes mellitus. Based on the results of the present study, we speculate that defects in the ability of insulin to disrupt transactivation by FKHR-like forkhead proteins and suppress hepatic gene expression through a conserved IRS may contribute to both increased hepatic production of IGFBP-1 and unrestrained gluconeogenesis in diabetes. The development of interventions which restore the ability of insulin to regulate hepatic gene expression through this novel pathway may provide an effective therapeutic approach to the treatment of some forms of diabetes mellitus.