Phosphorylation of the Transcription Factor Forkhead Family Member FKHR by Protein Kinase B*

Protein kinase B lies “downstream” of phosphatidylinositide (PtdIns) 3-kinase and is thought to mediate many of the intracellular actions of insulin and other growth factors. Here we show that FKHR, a human homologue of the DAF16 transcription factor in Caenorhabditis elegans, is rapidly phosphorylated by human protein kinase Bα (PKBα) at Thr-24, Ser-256, and Ser-319in vitro and at a much faster rate than BAD, which is thought to be a physiological substrate for PKB. The same three sites, which all lie in the canonical PKB consensus sequences (Arg-Xaa-Arg-Xaa-Xaa-(Ser/Thr)), became phosphorylated when FKHR was cotransfected with either PKB or PDK1 (an upstream activator of PKB). All three residues became phosphorylated when 293 cells were stimulated with insulin-like growth factor 1 (IGF-1). The IGF-1-induced phosphorylation was abolished by the PtdIns 3-kinase inhibitor wortmannin but not by PD 98059 (an inhibitor of the mitogen-activated protein kinase cascade) or by rapamycin. These results indicate that FKHR is a physiological substrate of PKB and that it may mediate some of the physiological effects of PKB on gene expression. DAF16 is known to be a component of a signaling pathway that has been partially dissected genetically and includes homologues of the insulin/IGF-1 receptor, PtdIns 3-kinase and PKB. The conservation of Thr-24, Ser-256, and Ser-319 and the sequences surrounding them in DAF16 therefore suggests that DAF16 is also a direct substrate for PKB in C. elegans.

In recent years evidence has accumulated that many of the metabolic actions of insulin may be mediated by a protein kinase cascade that lies "downstream" of phosphatidylinositide (PtdIns) 1 3-kinase and the second messengers PtdIns(3,4,5)P 3 and PtdIns(3,4)P 2 (reviewed in Refs. 1 and 2). A central player in this cascade is protein kinase B (PKB, also called c-Akt).
PKB mediates the metabolic actions of insulin by phosphorylating regulatory proteins at serine or threonine residues that lie in Arg-Xaa-Arg-Xaa-Xaa-(Ser/Thr) motifs (9), of which the best characterized are the cardiac isoform of 6-phosphofructo-2-kinase (PFK2) (2,10), the protein kinase glycogen synthase kinase 3 (GSK3) (11,12), and the mammalian target of rapamycin (mTOR) (13), as well as the proapoptotic protein BAD (reviewed in Ref. 2). Phosphorylation by PKB activates cardiac PFK2, and this is thought to underlie the insulininduced stimulation of glycolysis in the heart. Phosphorylation inhibits GSK3 and is thought to contribute to the stimulation of glycogen synthesis and global protein synthesis by insulin (1,2). Phosphorylation by PKB activates mTOR, allowing it to catalyze several phosphorylation events that enhance the translation of specific proteins. The overexpression of PKB has also been shown to mimic other metabolic actions of insulin, such as the stimulation of glucose (14) and amino acid transport (15).
When cells are stimulated with IGF-1, PKB is initially translocated to the plasma membrane where it becomes activated by PDK1 and PDK2, but it subsequently accumulates in the nucleus (16). This raises the question of whether PKB mediates some of the effects of insulin on specific gene transcription, and several pieces of evidence would appear to support this contention. For example, the overexpression of constitutively active mutants of PKB mimics the effects of insulin in stimulating the transcription of the obesity gene product leptin (17) and in inhibiting the transcription of IGF-binding protein 1 (IGFBP-1) (18). The insulin-induced suppression of phosphoenolpyruvate carboxykinase (19) and IGFBP-1 (18) is prevented by inhibitors of PtdIns 3-kinase (wortmannin, LY 294002) and unaffected by the drugs that inhibit mTOR (rapamycin) or the classical mitogen-activated protein (MAP) kinase cascade (18,19). Studies with constitutively active and dominant negative forms of PKB have shown that PKB may mediate transcriptional effects of insulin through a conserved insulin response sequence present in a number of genes known to be inhibited by insulin in the liver, such as IGFBP-1 and PEPCK (18). This suggests that PKB may indeed play an important role in mediating the effects of insulin on hepatic gene expression The insulin/IGF-1-stimulated PKB cascade has also been identified in Caenorhabditis elegans, where it is known to stimulate metabolism, to inhibit dauer arrest and to shorten the life span of this nematode (reviewed in Ref. 20). In this pathway, which has been partially dissected by genetic techniques, the DAF2 gene encodes a homologue of the IGF-1 receptor and lies "upstream" of the AGE1 gene that encodes a PtdIns 3-kinase homologue and the AKT1 and AKT2 genes that encode homologues of PKB. Downstream of PKB is the transcription factor DAF16; mutations in DAF16 return life span to normal that has been lengthened by inactivating mutations in AGE1 or AKT1/2 (21,22). Whether DAF16 is phosphorylated directly by AKT1/AKT2 is unknown, but we noticed that it possesses three consensus sequences for phosphorylation by PKB, all of which are highly conserved in several mammalian DAF16 homologues, namely the "Forkhead" family members FKHR, FKHRL1, and AFX (23). Two of the three sites are conserved in a further DAF16 homologue encoded by the AF6q21 gene (24). Here we establish that FKHR is phosphorylated at these three sites by PKB in vitro and in cotransfection experiments and that the same sites become phosphorylated in response to IGF-1 in 293 cells via a PtdIns 3-kinasedependent pathway that is independent of mTOR or the classical MAP kinase cascade. The accompanying paper (25) demonstrates that FKHR-stimulated reporter gene expression is dependent on an intact insulin response sequence (IRS) and that transactivation by FKHR is inhibited by insulin via the phosphorylation of Ser-256 (25). The conservation of Thr-24, Ser-256, and Ser-319 and the sequences surrounding them in DAF16 suggests that DAF16 is likely to be a direct substrate for PKB in C. elegans. Taken together, these results indicate that PKB regulates the ability of FKHR to stimulate transcription. Cell Culture, Transient Transfections, and Cell Lysis-293 cells were cultured at 37°C in an atmosphere of 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal calf serum. Transfection of cells was carried out using the calcium chloride precipitation method, using 10 g of DNA per 10-cm diameter dish. Prior to lysis, cells were serum-starved for 12 h. Cells were lysed in 1 ml ice-cold Buffer A (50 mM Tris acetate (pH 7.5), 1 mM EGTA, 1% (w/v) Triton X-100, 1 mM EDTA, 50 mM NaF, 10 mM sodium ␤-glycerophosphate, 5 mM sodium pyrophosphate, 1 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride, and 0.1% (v/v) ␤-mercaptoethanol). The lysates were centrifuged at 13,000 ϫ g, and the supernatants were removed, frozen immediately in liquid nitrogen, and stored at Ϫ80°C until use.

Materials-Tissue
Alignment, Cloning, Expression, and Purification of FKHR-Inspection of the public EST data bases identified 41 clones from 14 different tissues that encoded fragments of FKHR. Inspection of these sequences showed 100% identity with the nucleotide sequence of the second FKHR sequence to be reported (GenBank TM accession number AF032885 (26)). An epitope-tagged, full-length human FKHR construct was generated as follows. A fusion was generated in which the oligonucleotide 5Ј-GCGGG GATCC CGCCA CCATG GAGTT CATGC CCATG GAGCT GGTGC TGCCA T-3Ј encoding the epitope Glu-Phe-Met-Pro-Met-Glu (termed EE-tag) preceded the first 300 base pairs of FKHR (EST 1061191). This 350-base pair oligonucleotide produced the EE-FKHR fusion in a PCR reaction with a second oligonucleotide: 5Ј-GGCCG CGGCG GCCGC CGCCG CCACC GCCGC CGCCA CGGAG CC-3Ј corresponding to bases 259 -300 of the FKHR coding sequence. This PCR product was cloned into the TA cloning vector pCR2.1 TOPO. The remaining 1665 base pairs of FKHR (EST 166558) were incorporated into pCR2.1 by a non-PCR TA cloning method using the NotI and BspMI restriction sites (27). A vector containing a construct of full-length FKHR was then generated by ligating the two halves of FKHR in pCR2.1. This construct was then subcloned into pCMV-5 and pEBG2T to allow expression of recombinant FKHR in mammalian cells. Finally, the FKHR construct was subcloned into a pGEX4T-3 vector for expression of a GST-FKHR fusion protein in E. coli. Expressed GST-FKHR was purified by glutathione-Sepharose affinity chromatography and stored at Ϫ80°C in 50 mM Tris-HCl, pH 7.5 (20°C), 0.2 mM glutathione at a concentration of 2 mg/ml (4).
Generation of Phosphospecific Antibodies for FKHR-Phosphopeptides and dephosphopeptides were synthesized corresponding to residues 19 -31 (RPRSCTpWPLPRPE), 248 -262 (KSPRRRAASp-MDNNSK), and 311-324 (TTFRPRSSSpNASVS) of FKHR, where p indicates the site of phosphorylation (Thr-24, Ser-256, and Ser-319). The peptides were conjugated to both keyhole limpet hemocyanin and bovine serum albumin and injected into sheep at the Scottish Antibody Production Unit (Carluke, UK). Six weeks later, antisera were passed through a CH-Sepharose column to which a dephosphopeptide had been coupled, followed by affinity chromatography on CH-Sepharose to which the corresponding phosphopeptide antigen had been attached covalently. Phospho-specific antibodies were eluted with 0.1 M glycine, pH 2.4, immediately adjusted to pH 8 with Tris base and stored at 4°C. A further antibody was generated toward residues 636 -651 of FKHR (LPNQSFPHSVKTTTHS) that recognizes the dephosphorylated as well as the phosphorylated form of the protein.
Activation of Protein Kinase B and Phosphorylation of FKHR and BAD-PKB␣ (0.15-5.0 units/ml) that had been expressed in Sf9 cells and activated by phosphorylation with PDK1 (4) was incubated for 30 min at 30°C with FKHR (0.1 mg/ml) or BAD (0.05 mg/ml) in 50 mM Tris-HCl, pH 7.5), 0.1 mM EGTA, 2.5 M PKI, 10 mM magnesium acetate, 0.1 mM ATP (10 6 cpm/nmol). One unit of PKB activity was that amount that catalyzed the phosphorylation of 1 nmol of the peptide GRPRTSSFAEG in 1 min (11).
To determine the stoichiometry of phosphorylation of FKHR and BAD, aliquots of the reaction were denatured in SDS and subjected to polyacrylamide gel electrophoresis. After staining with Coomassie Blue and destaining, the gels were dried and the concentrations of GST-FKHR and GST-BAD were determined by densitometry using a Fujifilm LAS-1000 luminescent image analyzer calibrated with different concentrations of bovine serum albumin run in parallel on the same gel. The molecular masses of GST-FKHR and GST-BAD were taken as 97 kDa and 56 kDa, respectively.

Phosphorylation of FKHR by PKB in Vitro-
The observation that FKHR contains three residues, located at Thr-24, Ser-256, and Ser-319, that lie in consensus sequences for phosphorylation by PKB (see introduction) led us to study the phosphorylation of this protein by PKB in vitro. FKHR was phosphorylated by PKB to a stoichiometry of Ͼ1 mol of phosphate/mol of protein (Fig. 1A). The initial rate of phosphorylation was much faster than that of the pro-apoptotic protein BAD (Fig. 1B), which is thought to be an in vivo substrate for PKB (29).
To identify the sites of phosphorylation, we raised phosphospecific antibodies that only recognize FKHR if it is phosphorylated at Thr-24, Ser-256, or Ser-319. The specificity of these antibodies is demonstrated in Fig. 2. The antibody raised against a phosphopeptide corresponding to the sequence surrounding Thr-24 recognized FKHR only after phosphorylation by PKB, and its recognition of phospho-FKHR was prevented if the antibody was first preincubated with the phosphopeptide used to raise it but not by preincubation with either of the other two phosphopeptides. The specificities of the antibodies raised against phosphopeptides corresponding to the sequences surrounding Ser-256 and Ser-319 were established in an analogous manner (Fig. 2). These results demonstrate that PKB phosphorylates FKHR at Thr-24, Ser-256, and Ser-319. The phosphorylation of all three sites approached a plateau after 5-10 min when high concentrations of PKB were present in the assays (5 units/ml) (Fig. 3A). However, when the concentration of PKB was reduced to 0.15 unit/ml to measure initial rates of phosphorylation, Ser-256 was found to be phosphorylated more rapidly than the other two sites (Fig. 3B).
Phosphorylation of FKHR in 293 Cells-To investigate whether FKHR could be phosphorylated by PKB in a cellular context, we co-transfected into 293 cells DNA encoding FKHR with that encoding a constitutively active PKB and then examined FKHR phosphorylation by immunoblotting with the phospho-specific antibodies. These experiments demonstrated that Thr-24, Ser-256, and Ser-319 were phosphorylated to low levels when FKHR was transfected alone and that phosphorylation of each site increased greatly when FKHR was cotransfected with PKB (Fig. 4A). We also found that FKHR became phosphorylated at all three sites when cotransfected with PDK1 (Fig. 4B), one of the upstream activators of PKB (reviewed in Refs. 1  and 2).
In 293 cells PKB is maximally activated (50-fold) 5 min after stimulation with IGF-1 with a half-time for activation of 1 min (3). IGF-1 also stimulated the phosphorylation of FKHR at Thr-24, Ser-256, and Thr-319 (Fig. 5), maximal phosphorylation occurring within 10 min (data not shown). Consistent with the IGF-1-induced phosphorylation of FKHR being mediated by PKB, phosphorylation of Thr-24, Ser-256, and Ser-319 was prevented if the cells were incubated with the PtdIns 3-kinase inhibitor wortmannin (100 nM) prior to stimulation with IGF-1 (Fig. 5). The basal level of phosphorylation was also abolished by wortmannin. Residues 24,256, and 319 all lie in Arg-Xaa-Arg-Xaa-Xaa-(Ser/Thr) sequences that are not only consensus sequences for phosphorylation by PKB but also for phosphorylation by MAP kinase-activated protein kinase 1 (MAPKAP-K1, also called p90 rsk ) and p70 S6 kinase (9,30). MAPKAP-K1 (which lies immediately downstream of MAP kinase) and p70 S6 kinase (which lies downstream of mTOR) are both activated in response to insulin or IGF-1, and their activation is also inhibited by wortmannin (31,32). To investigate which of these protein kinases mediates the IGF-1-induced phosphorylation of FKHR in vivo, we therefore carried out additional experiments in which 293 cells were incubated with either rapamycin (which prevents the activation of p70 S6 kinase by inhibiting mTOR (33)) or PD 98059 (which prevents the activation of MAP kinase kinase-1 and hence the activation of MAPKAP-K1) (34). Neither of these drugs affected the basal or IGF-1-induced phosphorylation of FKHR at Thr-24, Ser-256, and Ser-332 (Fig. 5) although, at the same concentrations, they prevented the basal and IGF-1-induced phosphorylation of p70 S6 kinase and MAP-KAP-K1 (data not shown). DISCUSSION In this study we have demonstrated a rapid phosphorylation of FKHR by PKB at Thr-24, Ser-256, and Ser-319. This occurs both in vitro (Figs. 2 and 3) and in cotransfection experiments Bacterially expressed GST-FKHR was maximally phosphorylated with PKB (columns 2-5) or left unphosphorylated (column 1), and aliquots (1 g) were spotted onto a nitrocellulose membrane. They were then immunoblotted with affinity-purified antibodies raised against phosphopeptides corresponding to the sequences surrounding Thr-24, Ser-256, and Ser-319, as well as a dephosphopeptide corresponding to residues 636 -651 of FKHR. In columns 3, 4, and 5, antibodies were preincubated with 50 mM Thr-24, Ser-256, and Ser-319 phosphopeptide antigens prior to immunoblotting as indicated below. Columns 1 and 2, antibodies used without preincubation; column 3, antibodies incubated with the Thr-24 phosphopeptide antigen (50 mM); column 4, antibodies incubated with the Ser-256 phosphopeptide antigen (50 mM); column 5, antibodies incubated with the Ser-319 phosphopeptide antigen (50 mM).

FIG. 3. PKB phosphorylates FKHR at three residues.
A, bacterially expressed GST-FKHR was incubated with MgATP and PKB (5 units/ml). At the times indicated, aliquots were denatured in SDS, electrophoresed on 8% polyacrylamide gels, transferred to nitrocellulose, and immunoblotted with the antibodies in Fig. 2. B, same as A, except that the concentration of PKB was reduced to 0.15 units/ml, and after immunoblotting, the membranes were analyzed by densitometry to measure the extent of phosphorylation of each site.
using either PKB (Fig. 4A) or PDK1 (Fig. 4B). PDK1 is an upstream activator of PKB (reviewed in Refs. 1 and 2), and the phosphorylation of FKHR when cotransfected with PDK1 is presumably catalyzed by the endogenous PKB that becomes activated by the transfected PDK1. Similar observations were made with GSK3, another physiological substrate of PKB (35).
We have also shown that all three sites on FKHR become rapidly phosphorylated when 293 cells are stimulated with IGF-1 (Fig. 5). The IGF-1-induced phosphorylation is prevented by inhibitors of PtdIns 3-kinase but not by inhibitors of the activation of the MAP kinase cascade or p70 S6 kinase. These experiments, together with the in vitro studies and cotransfection experiments, indicate that the IGF-1-induced phosphorylation of FKHR is mediated by PKB or a closely related enzyme.
A pathway in C. elegans has been partially dissected by genetic techniques in which DAF16 (an FKHR homologue) is a downstream component of a signaling cascade that includes homologues of the insulin/IGF-1 receptor, PtdIns 3-kinase, and PKB (see introduction). PDK1 has not yet been identified as a component of this pathway, perhaps because its inactivation has other (lethal) consequences stemming from its role in activating additional protein kinases that lie in distinct protein kinase cascades. This may explain why disruption of the two PDK1 homologues in Saccharomyces cerevisiae is also lethal (28). We have noticed that, as expected, the genome of C. elegans does indeed encode a PDK1 homologue located on the X chromosome (genomic clone number H33H01).
The accompanying paper (25) shows that FKHR stimulates the transcription of a reporter gene, that this stimulation is dependent on an intact insulin response sequence (IRS), and that transcription is suppressed by 50% when HepG2 cells are exposed to insulin or transfected with constitutively active PKB. In this system, phosphorylation of Ser-256 appears to be required and sufficient to mediate the action of insulin because its mutation to alanine abolishes the effect, whereas mutation to aspartic acid mimics it. Interestingly Ser-256 is the residue in FKHR that is phosphorylated most rapidly by PKB in vitro (Fig. 3B). In contrast, mutation of Thr-24 or Ser-319 to alanine has no effect on the ability of insulin or constitutively active PKB to disrupt activation by FKHR, and the functional consequences of these other phosphorylation events remain to be determined. Nevertheless, the conservation of the sequences that surround Thr-24 and Ser-319 in two other mammalian FKHR homologues (23) as well as DAF16 of C. elegans (21,22) suggests that their phosphorylation is likely to play a role(s) in vivo.
In conclusion this study, together with that presented in the accompanying paper (25), provides the first direct biochemical evidence that the kinase cascade leading to the regulation of DAF16 is present in mammalian cells. The finding that phosphorylation of FKHR inhibits its ability to stimulate transcription is consistent with genetic evidence in C. elegans that has shown that mutations in the DAF16 gene have the opposite phenotype to mutations in either the AGE1 (PtdIns 3-kinase) or AKT1/2 (PKB) genes (21)(22)(23). These findings raise the possibility that phosphorylation of FKHR (or its homologues) underlies the regulation of at least some of the genes whose transcription is inhibited by insulin.
Acknowledgments-We thank our colleagues Dr. Andrew Paterson (expression and activation of PKB), Dr. Maria Deak (construction of vectors expressing PDK1 and constitutively active PKB), and Dr. Takayasu Kobayashi (purified BAD protein) for reagents.
Note Added in Proof-While this paper was under review, similar results to those presented here and in the following paper (25) have been found by three other laboratories. In analysis of Forkhead family members namely FKHRL1 (36), AFX (37), and FKHR1 (38), phosphorylation was accompanied by nuclear export (36,37) as a result of interaction of the phosphorylated transcription factor with 14-3-3 proteins (36). 293 cells were transiently transfected with a plasmid expressing GST-FKHR. 16 h post-transfection, the cells were serum-starved for a further 12 h and then stimulated for 10 min with or without 100 ng/ml IGF-1. Aliquots (10 g of lysate protein) were then denatured in SDS, electrophoresed on 8% polyacrylamide gels, and after transfer to nitrocellulose, immunoblotted with the four antibodies used in Fig. 2. Where indicated, the cells were incubated for 10 min with 100 nM wortmannin or for 60 min with 100 nM rapamycin plus 50 M PD 98059 prior to stimulation with IGF-1.