Insulin Regulation of Insulin-like Growth Factor-binding Protein-1 Gene Expression Is Dependent on the Mammalian Target of Rapamycin, but Independent of Ribosomal S6 Kinase Activity*

Insulin inhibits the expression of the hepatic insulin-like growth factor-binding protein-1 (IGFBP-1) and glucose-6-phosphatase (G6Pase) genes. The signaling pathway that mediates these events requires the activation of phosphatidylinositol 3-kinase, whereas transfection studies have suggested an involvement of Akt (protein kinase B) and FKHR, a transcription factor regulated by Akt. We now demonstrate that insulin repression of endogenous IGFBP-1 gene transcription was blocked by rapamycin or by amino acid starvation. Rapamycin inhibited the mammalian target of rapamycin (mTOR) and the subsequent activation of p70/p85 S6 protein kinase-1 (S6K1) by insulin, whereas amino acid depletion prevented insulin induction of these signaling molecules. Importantly, we demonstrate that insulin regulation of the thymine-rich insulin response element of the IGFBP-1 promoter was also inhibited by rapamycin. However, sustained activation of S6K1 did not repress this promoter. In addition, rapamycin did not affect insulin regulation of G6Pase expression or Akt activation. We propose that these observations indicate that an mTOR-dependent, but S6K-independent mechanism regulates the suppression of IGFBP-1 (but not G6Pase) gene expression by insulin. Therefore, although the insulin-responsive sequence of the G6Pase gene promoter is related to that of the IGFBP-1 promoter, the signaling pathways that mediate suppression of these genes are distinct.

Insulin regulates hepatic metabolism by inducing a complex signaling network that ultimately mediates the ability of the liver to maintain glucose homeostasis (1)(2)(3)(4). Defects in this signaling network may underlie the development of insulin resistance. Upon binding to its cognate receptor at the cell membrane, insulin promotes a conformational change that results in autophosphorylation and activation of the intracellular domain of the receptor. This activation of the insulin receptor tyrosine kinase promotes phosphorylation of a number of intracellular substrates, including a group of related proteins known as insulin receptor substrates 1-4 (5). These proteins act as adaptors, recruiting and activating proteins that bind to the phosphotyrosine motifs targeted by the insulin receptor. One such molecule is the lipid kinase phosphatidylinositol 3-kinase (PI3K). 1 Activation of PI3K results in the production of 3-phosphorylated inositides, including phosphatidylinositol 3,4,5-trisphosphate. These molecules interact with multiple pleckstrin homology domain-containing proteins, including the protein kinases 3-phosphoinositide-dependent protein kinase-1 and Akt (also known as protein kinase B) (6,7). Accumulation of phosphatidylinositol 3,4,5-trisphosphate activates S6 protein kinase (S6K), Akt, and serum-and glucocorticoid-regulated protein kinase (SGK) through a 3-phosphoinositide-dependent protein kinase-1-dependent mechanism (for reviews, see Refs. 8 and 9). In addition, PI3K activation also regulates signaling through the protein kinase termed mammalian target of rapamycin (mTOR) by a less well defined mechanism that may require Akt activation (10,11). Activation of S6K (but not Akt or SGK) requires both PI3K and mTOR activity (11)(12)(13).
In yeast, TOR signaling is known to regulate the cellular localization of specific transcription factors (14), whereas in mammals, the ciliary neurotrophic factor regulation of STAT3 phosphorylation and activation is mediated via mTOR (15). Downstream of mTOR, S6K phosphorylates ribosomal S6 protein in vitro and in vivo (12,13) and may be involved in insulin regulation of protein translation (16,17). Mammals express two homologous S6K proteins from distinct genes, termed S6K1 and S6K2. Mice lacking S6K1 have a reduced pancreatic ␤-cell size and are thus hypoinsulinemic and glucose-intolerant (18).
The mTOR pathway has been linked to the regulation of expression of the insulin gene in pancreatic ␤-cells (19), the Na ϩ /P i cotransporter-1 gene in H4IIE cells (20), the hexokinase II gene in L6 myotubes (21), and the p85 ␣-regulatory subunit of PI3K in isolated muscle cells (22). The molecular connection between mTOR and these gene promoters is not clear. In contrast, Akt and SGK have been implicated in the regulation of multiple genes through the phosphorylation of FKHR/ FKHR-L1 and AFX (23)(24)(25)(26)(27)(28), related members of a subfamily of the FOX(o) transcription factor family (29). For example, overexpression studies and treatment of cells with selective PI3K inhibitors have implicated PI3K, Akt, and FKHR in the regulation of IGFBP-1 gene expression by insulin (25, 30 -32). A well characterized thymine-rich insulin response element (TIRE) is fundamental to the correct regulation of the IGFBP-1 promoter by insulin (30,33), and FKHR-related proteins can bind to this TIRE (T(G/A)TTT(T/G)(T/G)) in vitro (34). Homologous TIRE sequences are found in a number of other insulinrepressed gene promoters, including the glucose-6-phosphatase (G6Pase) and phosphoenolpyruvate carboxykinase (PEPCK) promoters. Therefore, a PI3K-, Akt-, and FKHR-dependent mechanism for regulation of G6Pase and PEPCK expression also has been proposed (35).
However, some recent studies have questioned this hypothesis. For example, the mTOR inhibitor rapamycin blocks the regulation of glucocorticoid-induced IGFBP-1 expression by insulin (36). In addition, detailed characterization of the IGFBP-1 and PEPCK TIRE sequences suggests that FKHR can bind to mutant TIRE sequences that do not respond to insulin (37).
To resolve this discrepancy, we have carried out a detailed analysis of the effect of rapamycin on the regulation of basal and glucocorticoid-induced IGFBP-1 and G6Pase gene transcription as well as its effect on insulin regulation of the IGFBP-1 TIRE, with and without FKHR overexpression. We found that mTOR activity is required for full insulin regulation of IGFBP-1 (but not G6Pase or PEPCK) gene transcription and that S6K activation is not sufficient to reproduce this action of insulin.
Cell Culture-The rat hepatoma cell line H4IIE was maintained in Dulbecco's modified Eagle's medium (DMEM) containing 1000 mg/liter glucose and 5% (v/v) fetal calf serum as described previously (38).
RNA Isolation and RNase Protection Assay-H4IIE cells were serum-starved overnight and treated with hormone/inhibitor for the times and at the concentrations indicated in the figure legends. Total cellular RNA (ϳ100 g/10 6 cells) was isolated using TriReagent TM (Sigma) following the manufacturer's instructions. An RNase protection assay was carried out to determine the relative amounts of IGFBP-1, G6Pase, ␤-actin, or cyclophilin mRNA in each sample. The synthesis of the rat IGFBP-1 and G6Pase cDNAs has been described previously (38,39). A similar strategy was used to construct mouse IGFBP-1 and G6Pase probes. Briefly, mouse cDNAs for IGFBP-1 (from nucleotide ϩ625 to ϩ830) or G6Pase (from nucleotide ϩ63 to ϩ237) were amplified from a mouse liver cDNA library (CLONTECH), and the resultant product was cloned into pCR4-TOPO (Invitrogen). The plasmid was cut with SpeI to produce a linear DNA template. pTRI-cyclophilin (rat) and pTRI-actin (mouse) linear plasmids (Ambion Inc.) were used as control DNA templates. All RNA probes were synthesized by in vitro transcription following the instructions of the MaxiScript kit (Ambion Inc.). 3000 cpm of each probe was hybridized with 10 g of total RNA following the instructions of the RPA II kit. Samples were analyzed on an 8 M urea and 5% polyacrylamide gel; band intensity was quantitated on a Phos-phorImager (Fuji); and data were calculated as the ratio of IGFBP-1 or G6Pase to cyclophilin or ␤-actin RNA.
Preparation of Cell Extract for Kinase Assays and Western Blotting-H4IIE cells were incubated in serum-free medium with hormones and inhibitors for the times and at the concentrations indicated in the figure legends. Cells were then scraped into ice-cold lysis buffer (25 mM Tris-HCl (pH 7.4), 50 mM NaF, 100 mM NaCl, 1 mM sodium vanadate, 5 mM EGTA, 1 mM EDTA, 1% (v/v) Triton X-100, 10 mM sodium pyrophos-phate, 1 mM benzamidine, 0.1 mM phenylmethylsulfonyl fluoride, 0.27 M sucrose, 2 M microcystin, and 0.1% (v/v) 2-mercaptoethanol). Cell debris was removed by centrifugation at 13,000 ϫ g for 5 min, and the protein concentration was determined by the method of Bradford (63) using bovine serum albumin as an internal standard.
Antibodies for Immunoprecipitation and Western Blot Analysis-The phospho-specific Ser-235 ribosomal S6 protein antibody was raised against a peptide derived from amino acids 229 -242 (AKRRRLpSSL-RASTS) and is commercially available from Upstate Biotechnology, Inc. The phospho-specific Thr-32 FKHR-L1, Thr-24 FKHR, and Ser-319 FKHR antibodies (25) and the control anti-FKHR, anti-S6K, and anti-GSK3␣ antibodies were purchased from Upstate Biotechnology, Inc., whereas the phospho-specific Ser-9/Ser-21 GSK3, Thr-202/Tyr-204 p42/ p44 MAPK, and Ser-256 FKHR antibodies and the control p42/p44 MAPK antibody were purchased from New England Biolabs, Inc. Immunoprecipitation of endogenous FKHR was performed as described above, and total or isolated protein was separated on Novex SDS-4 -12% polyacrylamide gel. Following transfer to nitrocellulose, blots were incubated with 5% (w/v) nonfat milk for 2 h, and primary antibodies were incubated at 4°C overnight prior to incubation for 1 h at room temperature with secondary antibody and development using the ECL kit (Amersham Biosciences, Inc.) following the manufacturer's instructions.
Plasmids and Transient Transfections-The plasmids BP-1WT and BP-1DM5 were a gift from Dr. Rob Hall and Professor Daryl K. Granner (Vanderbilt University, Nashville, TN). The BP-1WT plasmid represents a luciferase reporter construct under the control of a thymidine kinase promoter containing the IGFBP-1 wild-type TIRE sequence (5Ј-CAAAACAAACTTATTTTG). Two base pair mutations of the wild-type TIRE sequence at residues equivalent to position 5 in each of the A and B sites (5Ј-CAAAAGAAACTTCTTTTG) produces a mutant promoter (BP-1DM5) that is no longer responsive to insulin (37). Transfections were performed using the calcium phosphate procedure as described previously (42). H4IIE cells were transfected with 10 g of BP-1WT or BP-1 DM5 with or without 10 g of glutathione S-transferase-FKHR (25). Cells were then incubated for 20 h in serum-free medium with or without hormones as described in the figure legends. Cells were lysed in 900 l of 1ϫ cell lysis buffer (Promega, Southampton, UK) and centrifuged at 13,000 rpm for 2 min, and the supernatant was stored at Ϫ70°C. Luciferase assays were performed using the firefly luciferase assay system (Promega) following the manufacturer's instructions, with the luciferase activity being corrected for the protein concentration in the cell lysate as determined by the method of Bradford (63).
Isolation of Nuclei and Nuclear Run-on Assay-Nuclei from H4IIE cells were isolated by a modification of the method described previously (43). Briefly, cells were harvested in ice-cold phosphate-buffered saline, pelleted at 1200 rpm for 5 min, and resuspended in 4 ml of sucrose buffer I (0.32 M sucrose, 3 mM CaCl 2 , 2 mM magnesium acetate, 0.1 mM EDTA, 10 mM Tris-HCl (pH 8), 1 mM dithiothreitol, and 0.5% (v/v) Nonidet P-40). The mixture was then transferred to a Wheaton Dounce tissue grinder, and the cells were lysed by five strokes each with a type B pestle followed by a type A pestle. 4 ml of sucrose buffer II (2 M sucrose, 5 mM magnesium acetate, 0.1 M EDTA, 10 mM Tris-HCl (pH 8), and 1 mM dithiothreitol) was added to the homogenate. Nuclei were collected by overlaying 4.4 ml of sucrose buffer II with the sample and ultracentrifugation at 30,000 ϫ g for 45 min at 4°C. The nuclei pellet was resuspended in 300 l of nuclear freezing buffer (50 mM Tris-HCl (pH 8.3), 40% (v/v) glycerol, 5 mM MgCl 2 , and 0.1 mM EDTA) and stored at Ϫ80°C.
A cDNA for rat IGFBP-1 (see "RNA Isolation and RNase Protection Assay"), a positive control cDNA (rat ␤-actin), and a negative control cDNA (pCR2.1, Invitrogen) were linearized; denatured in 0.3 M NaOH; applied (5 g/slot) to 0.45-m nitrocellulose membrane; and fixed by UV cross-linking. The rat ␤-actin was amplified from total rat liver RNA by reverse transcription-PCR using primer sequences 5Ј-TCA TGA AGT GTG ACG TTG ACA TCC GT-3Ј and 5Ј-CCT AGA AGC ATT TGC GGT GCA CGA TG-3Ј (Promega). The resultant product was cloned into pCR2.1 and linearized with BamHI. The negative clone (pCR2.1) was created by digesting pCR2.1/rat ␤-actin with EcoRI to remove the ␤-actin insert and religating the vector to produce the empty vector. To examine transcription initiation, isolated nuclei (ϳ50 ϫ 10 6 ) were incubated in 5 mM Tris-HCl (pH 8), 2.5 mM MgCl 2 , 150 mM KCl, 0.25 mM ATP, 0.25 mM CTP, 0.25 mM GTP, and 150 Ci of [␣-32 P]UTP for 30 min at 37°C. Proteins and DNA were removed by phenol/chloroform extraction. The radiolabeled RNA was recovered by ethanol precipitation, and the total incorporation of radioactivity for each RNA sample was determined by scintillation counting (44). Equal amounts of radiolabeled RNA (6 ϫ 10 6 cpm) were incubated with the nitrocellulose-immobilized DNA slot blots in a final volume of 1 ml of TES/NaCl solution (10 mM TES (pH 7.4), 10 mM EDTA, 0.2% (w/v) SDS, and 0.6 M NaCl) for at least 36 h. After hybridization, the membranes were washed with 2ϫ SSC at 65°C for 1 h, followed by 2ϫ SSC containing 0.1% SDS at 65°C for 1 h. An additional wash was performed in the presence of 10 g/ml RNase A at 37°C for 30 min to digest any radiolabeled RNA that had not annealed to the membrane. The RNase was removed by a single wash with 2ϫ SSC for 1 h. The blots were visualized by autoradiography and PhosphorImager analysis.
Preparation of Adenoviruses and Treatment of H4IIE Cells with Adenoviruses-Adenoviral vectors were generated using the system described by Hardy et al. (45). Briefly, an XbaI-PstI fragment containing the S6K1 cDNA (either wild-type or kinase-dead due to mutation to Gln of the codon encoding Lys-100) with a Myc tag at the N terminus (17,46) was subcloned into pAdlox cut with XbaI and PstI. 10 g of the resulting plasmids was cotransfected together with 3 g of the donor viral DNA ⌿5 in 293 cells expressing the recombinase CRE (CRE8 cells). When cells rounded up, they were collected, and an extract was prepared by sonication. The extracts were used for further amplification of the adenoviral vectors in CRE8 cells. All additional viral amplifications were carried out in human embryonic kidney 293 cells. H4IIE cells were infected at a multiplicity of infection of 5 in serum-free DMEM. After 36 h, cells were treated with or without insulin and lysed (as described above) for Western blot analyses, or RNA was prepared for gene expression analysis.
Statistical Analyses-As a measure of statistical significance of differences in experimental groups, Student's t tests were performed, and 5% confidence limits were applied.

RESULTS
The regulation of basal or glucocorticoid/cAMP-induced G6Pase gene expression in H4IIE cells by insulin was blocked by inhibitors of PI3K, but was not affected by the presence of rapamycin (Fig. 1, A and B). In contrast, inhibition of mTOR by treatment of H4IIE cells with as little as 1 nM rapamycin strongly antagonized the ability of insulin to repress either basal (Fig. 1C) or glucocorticoid-induced (Fig. 1D) IGFBP-1 gene expression. However, this action of insulin was not completely inhibited by this treatment, even at rapamycin concentration as high as 200 nM (data not shown). Consistent with these findings, nuclear run-on analysis demonstrated that this mTOR-dependent pathway mediated about half of the insulin regulation of IGFBP-1 gene transcription (Fig. 2).
Transient transfection of H4IIE cells with a luciferase reporter gene (BP-1WT) under the control of a thymidine kinase promoter containing the IGFBP-1 TIRE (nucleotides Ϫ96 to Ϫ72 relative to the transcription start site) rendered luciferase expression sensitive to insulin (Fig. 3). The response of this TIRE to insulin was ablated by two base pair mutations (BP-1DM5) of residues equivalent to position 5 in each of the A and B sites of the IGFBP-1 TIRE (Fig. 3) (37). Meanwhile, the ability of insulin to regulate BP-1WT was completely lost in the presence of 10 nM rapamycin (Fig. 3A). Previously, this element was proposed to be regulated independently of mTOR activity, via Akt regulation of FKHR (25, 30 -32). We therefore checked whether FKHR is also regulated in a rapamycin-sensitive manner. When FKHR was coexpressed with BP-1WT in H4IIE cells, the expression of luciferase was induced by Ͼ5-fold, whereas insulin produced a 50% block of this induced expression (Fig. 3B). Interestingly, the ability of insulin to inhibit this FKHR-mediated luciferase expression in H4IIE cells was slightly reduced in the presence of 10 nM rapamycin (Fig. 3B); however, the effect of rapamycin on FKHR activity was not as significant as the rapamycin block of insulin regulation in the absence of FKHR overexpression (Fig. 3, A and B). Coexpression of FKHR with BP-1DM5 in H4IIE cells did not induce luciferase expression or impart insulin sensitivity to this promoter (Fig. 3B), consistent with its inability to bind to this mutant TIRE (37).
Importantly, treatment of H4IIE cells with as low as 1 nM rapamycin completely blocked the activation of S6K1 by insulin (Fig. 4A). However, the presence of 10 nM rapamycin did not affect the ability of insulin to activate MAPK (Fig. 4B) or to induce phosphorylation of the Akt substrates GSK3␣ and GSK3␤ at Ser-21 and Ser-9, respectively (Fig. 4B), or FKHR at Thr-24, Ser-256, and Ser-319 (Fig. 4C). Insulin did not activate SGK in H4IIE cells (data not shown).
Depriving cells of amino acids is an alternative mechanism for inactivating the mTOR/S6K pathway (47,48). The ability of insulin to induce phosphorylation of S6 in H4IIE cells deprived of amino acids and glucose for 1 h was almost completely lost (Fig. 5A). Meanwhile, Akt activation by insulin was not affected by amino acid deprivation, but basal Akt activity was reduced after 2 h of deprivation (as judged by phosphorylation of GSK3 and FKHR-L1) (Fig. 5B). Under similar conditions, the insulin repression of IGFBP-1 gene expression was severely affected (Fig. 5C). Indeed, amino acid deprivation had a very similar effect on IGFBP-1 gene expression compared with treatment of cells with rapamycin.
These data suggest that the mTOR/S6K pathway is required for full regulation of IGFBP-1 gene expression by insulin. To examine whether S6K activity alone is sufficient to reduce the expression of IGFBP-1, we expressed active S6K1 in H4IIE cells using adenovirus (Fig. 6). As a control, we expressed catalytically inactive S6K1 (kinase-dead S6K1) at a level similar to that of active S6K1 (Fig. 6A). At this level of protein expression, the phosphorylation of S6 by active S6K1 was similar to that seen with insulin treatment of these cells, yet not high enough to permit kinase-dead S6K1 to produce a dominant-negative effect on insulin-induced S6 phosphorylation (Fig. 6A). However, there was no reduction in IGFBP-1 gene expression following activation of S6K1 by this method (Fig.  6B). Therefore, either the pathway from mTOR to the IGFBP-1 promoter does not require S6K, or S6K activation is not sufficient (or present in the appropriate compartment) to mimic insulin regulation of this promoter.

The Insulin Signaling Pathway That Regulates IGFBP-1 Gene Expression Is Distinct from the Pathway That Regulates
G6Pase and PEPCK Gene Expression-The search for insulinresponsive DNA promoter elements and their binding proteins has led to the identification of a TIRE in the PEPCK promoter and subsequently to the identification of similar insulin-responsive sequences within a number of other gene promoters, including IGFBP-1 and G6Pase (35,49,50). Recently, FKHR and related transcription factors have been shown to bind to TIREs in vitro and to regulate their activity in intact cells (25, 30 -32, 34). The discovery that FKHR is regulated in a PI3Kand Akt-dependent manner suggests a potentially complete signaling pathway from the insulin receptor to a subgroup of insulin-regulated gene promoters. However, we now demonstrate that complete repression of the IGFBP-1 TIRE requires mTOR activity, whereas G6Pase or PEPCK (51) expression is insensitive to rapamycin. Thus, it seems that these related DNA sequences are actually regulated by distinct mechanisms. Indeed, although there is a great deal of circumstantial evidence linking Akt and FKHR to regulation of IGFBP-1, G6Pase, and PEPCK expression, there are also a number of reports suggesting that the regulation of G6Pase and PEPCK expression by insulin may require other signaling pathways. For example, Tavare and co-workers (52) were unable to repress the G6Pase promoter by cotransfection of an active form of Akt. Although these workers did not check FKHR activity in their experiments, others have demonstrated strong repression of FKHR when Akt is activated. Similarly, Granner and coworkers (37) have questioned the role of FKHR as a regulator of the PEPCK and IGFBP-1 promoters. A base-by-base examination of the PEPCK (and IGFBP-1) TIRE revealed that FKHR-L1 could bind to mutant TIRE sequences that had lost the ability to mediate a response to insulin. Also, overexpression of "dominant-negative" Akt reduces the regulation of FIG. 2. Insulin regulation of IGFBP-1 gene transcription is blocked by rapamycin. H4IIE cells were serum-starved overnight and preincubated with or without 10 nM rapamycin for 30 min prior to a 3-h treatment with dexamethasone (500 nM) plus rapamycin (Dex/ Rapa), with dexamethasone plus 10 nM insulin (Dex/ins), or with all three agents (Dex/ins/Rapa). Nuclei were isolated from 5 ϫ 10 cm 2 dishes for each treatment, and nuclear run-on assays were performed to measure IGFBP-1 (BP-1) or ␤-actin transcription as described under "Experimental Procedures." Empty (E) pCR2.1 vector was used as a negative control in each treatment. A representative experiment is shown in A, whereas quantification of two experiments carried out in duplicate was achieved by PhosphorImager analysis in B. In the latter case, the results (means Ϯ S.E.) are presented relative to the induced control. *, p Ͻ 0.05.

FIG. 3. Rapamycin blocks insulin regulation of an IGFBP-1 TIRE-containing promoter. H4IIE cells were transfected with BP-
1WT or BP-1DM5 (insulin-insensitive) plus pEBG2T (A) or pEBG2T-FKHR (B). Cells were incubated with or without insulin Ϯ rapamycin for 20 h, and luciferase assays were performed as described under "Experimental Procedures." Results are presented relative to basal (serum-starved) luciferase expression in the absence of FKHR overexpression and are the means Ϯ S.E. of at least three experiments performed in triplicate. The percent inhibition by insulin relative to the appropriate control is shown, as is the statistical significance between the indicated experimental groups. *, p ϭ 0.11; ***, p Ͻ 0.001. FKHR by insulin, but not the regulation of the endogenous PEPCK promoter (53). Although it has been shown that overexpression of active Akt reduces PEPCK gene expression, this may be explained by inhibition of GSK3 (a physiological target of Akt) (40). Inhibitors of GSK3 strongly reduce PEPCK and G6Pase gene transcription in H4IIE cells (54). The data presented herein demonstrate that the regulation of IGFBP-1 gene transcription by insulin is affected by the presence of rapamycin, a treatment that does not affect insulin regulation of G6Pase (Fig. 1) or PEPCK (51). Thus, mTOR is an important mediator of insulin regulation of the IGFBP-1 promoter. These results demonstrate that distinct mechanisms regulate these three TIREs in vivo.
The IGFBP-1 Promoter May Contain an Element, Other than the TIRE, That Enhances Insulin Repression of Endogenous IGFBP-1 Gene Transcription-Surprisingly, the regulation of the IGFBP-1 TIRE by insulin is completely blocked by rapamycin, whereas endogenous gene regulation is only affected partially (ϳ50%). This suggests that a second insulin response element, or an accessory factor element, may be present within the IGFBP-1 promoter and that this second element is rapamycin-insensitive. Indeed, heterologous promoters containing FIG. 4. Effects of rapamycin on insulin signaling in H4IIE cells. H4IIE cells were serum-starved overnight and preincubated with or without rapamycin for 30 min prior to stimulation with insulin Ϯ rapamycin. A, at the times indicated, cells were lysed, and S6K1 was immunoprecipitated and assayed as described under "Experimental Procedures." Results are the means Ϯ S.E. of two experiments performed in triplicate. A significant effect of rapamycin was observed at 0.5 nM. **, p Ͻ 0.01; NS, not significant. B, alternatively, Western blot analyses was performed on 20 g of cell lysates to assess the phosphorylation status of Thr-202/Tyr-204 p42/p44 MAPK or Ser-9/Ser-21 GSK3␣/␤ following incubation of H4IIE cells with insulin Ϯ 10 nM rapamycin (Rap). C, similarly, the effect of 10 nM rapamycin on the insulin-induced phosphorylation of the three putative Akt target residues of FKHR (Thr-24, Ser-256, and Ser-319) was analyzed by Western blotting of FKHR immunoprecipitated from 200 g of cell lysate.
FIG. 5. Down-regulation of mTOR activity by amino acid deprivation mimics the effects of rapamycin on insulin regulation of IGFBP-1. H4IIE cells were serum-starved overnight and then preincubated for 1 h either in phosphate-buffered saline or left in serumfree DMEM. Subsequently, cells were treated with or without insulin Ϯ DMEM for 30 min (A and B) or 2 h (B and C) and lysed, and the lysates (20 g) were subjected to Western blot analysis. Blots were developed using the following antibodies: phospho-specific ribosomal S6 protein and S6K antibodies (A) and phospho-specific Ser-9/Ser-21 GSK3␣/␤, GSK3␤, and phospho-specific Thr-32 FKHR-L1 antibodies (B). Alternatively, following preincubation with phosphate-buffered saline, cells were treated with hormones (dexamethasone, 500 nM; and insulin, 10 nM) for 2 h in phosphate-buffered saline or in serum-free DMEM, and the total cellular RNA was isolated. An RNase protection assay was performed to assess IGFBP-1 (BP-1) gene expression (C). Results are presented relative to basal IGFBP-1 mRNA levels (serum-free DMEM) and are the means Ϯ S.E. of two experiments performed in duplicate (C, upper panel). A representative experiment is also shown (C, lower panel). ***, p Ͻ 0.001. the IGFBP-1 TIRE are not fully repressed by insulin (Fig. 4) (30,32,37), whereas the endogenous gene is completely repressed (Fig. 1C). Therefore, the binding and/or activity of the TIRE-interacting factor may be enhanced by contiguous IGFBP-1 promoter elements. This is analogous to the PEPCK and G6Pase gene promoter responses to insulin. The PEPCK promoter contains at least one other insulin response element (49), distinct from the TIRE, whereas the G6Pase promoter contains an accessory element that binds hepatocyte nuclear factor-1 and enhances the insulin repression of G6Pase gene expression through its TIRE (55). Thus, these three gene promoters have evolved complex multi-element mechanisms to mediate their repression by insulin in times of plenty. Although they each contain a related TIRE, the structure of each "insulin response unit" is quite different. The IGFBP-1 TIRE consists of an inverted palindrome of the T(G/A)TTT(T/G)(T/G) sequence that is not found in the other two promoters (or any other promoter reported to date). Therefore, we hypothesize that this TIRE structure renders its regulation by insulin completely rapamycin-sensitive, whereas the presence of a single TIRE (PEPCK) or a series of three contiguous TIREs (G6Pase) allows regulation by insulin independently of mTOR.
The Role of mTOR, S6K, and FKHR in the Regulation of IGFBP-1 Gene Expression-The effects of rapamycin (Fig. 2) and amino acid deprivation (Fig. 5) demonstrate a requirement for mTOR activity in the complete repression of IGFBP-1 gene transcription by insulin. However, S6K activity is not sufficient to replace insulin regulation of this gene (Fig. 6). Initial experiments in mice lacking both the S6K1 and S6K2 genes suggest that the regulation of IGFBP-1 expression is not affected by loss of S6K activity (data not shown). This suggests either that S6K is not required for the regulation of this gene by insulin or that these animals have adapted by up-regulating an alternative activity. This requires more detailed analysis, but it remains a distinct possibility that S6K is neither sufficient nor required for the regulation of IGFBP-1 expression by insulin.
In yeast, Tor signaling produces cytoplasmic sequestration of a number of transcription factors known to mediate a response to nutrients (14,56,57). Specifically, starvation down-regulates Tor activity and induces nuclear translocation of factors such as Msn2p, Msn4p, and Gln3p (14,58). Tor has also been shown to regulate the expression of metabolic genes via the factors Tap42p, Mks1p, Ure2p, Gln3p, and Gat1p (57,59). Indeed, rapamycin treatment of yeast for as little as 30 min has profound effects on gene expression, both positively and negatively (57). However, yeast cells do not express an S6K homolog; and thus, the mechanism by which Tor regulates these transcription factors must involve a distinct Tor signaling cascade (11). For example, Tap42p is a phosphatase-associated protein that can interact with mammalian protein phosphatases 2A and 4 (60,61). It is therefore tempting to speculate that mTOR may regulate the mammalian transcriptional machinery through activation of a phosphatase. Meanwhile, the inhibition of mTOR leads to the induction of autophagy in hepatocytes, whereas persistent activation of S6K inhibits autophagy (48,62). The effects that we observed on IGFBP-1 gene transcription in the presence of rapamycin cannot therefore be linked to increased autophagy.
Previously, there was no evidence that an mTOR-activated pathway could regulate FKHR activity. The serine and threonine residues targeted by Akt (and/or SGK) have been well characterized, and mutation of these residues to alanine (AAA-FKHR) renders FKHR almost totally insensitive to insulin (32,37). We have demonstrated that Akt-mediated phosphorylation of FKHR occurs in rapamycin-treated H4IIE cells (Fig. 4). It is difficult to assess the effect of rapamycin on insulin regulation of FKHR transactivating activity due to an inductive effect of rapamycin in the absence of insulin (Fig. 3B). In the presence of insulin, rapamycin returns promoter activity to the base-line level, but not to the rapamycin-induced level. Indeed, there is not a significant difference in the effect of insulin when FKHR is overexpressed in the presence or absence of rapamycin (p Ͼ 0.1), although there is a trend toward a small effect of rapamycin on FKHR transactivating potential. This suggests the existence of an FKHR-independent regulator of the IGFBP-1 TIRE, unless overexpression of FKHR renders it less sensitive to rapamycin. It will require a much more detailed study of endogenous FKHR phosphorylation and activity to determine the relative contributions of these insulin signaling pathways to the regulation of FKHR. Taken together with previous results, it has now been established that insulin-mediated regulation of promoters is profoundly altered following overexpression of FKHR/FKHR-L1. An important issue in this field is to establish whether FKHR and FKHR-L1 are the true regulators of TIREs or whether overexpression alters the TIRE-binding complex. The generation of FKHR/FKHR-L1/AFX-deficient animals may be required to identify the bona fide target gene promoters for these factors.  6. Overexpression of an active form of S6K1 is not sufficient to repress the IGFBP-1 gene promoter. H4IIE cells were incubated with adenovirus (multiplicity of infection ϭ 5) expressing either constitutively active (Ad-S6K-act) or kinase-dead (Ad-S6K KD) S6K in serum-free DMEM for 36 h. A, cells were then treated with or without insulin for 30 min and lysed, and 20 g was subjected to Western blot analysis to determine phosphorylation of ribosomal S6 protein and expression of the recombinant S6K proteins. B, alternatively, following adenoviral incubation, cells were incubated for an additional 3 h with or without hormones (dexamethasone, 500 nM; and insulin, 10 nM) prior to isolation of cellular RNA. IGFBP-1 expression was assessed as described under "Experimental Procedures." Results are presented as -fold activation of basal (serum-starved) IGFBP-1 mRNA levels and are the means Ϯ S.E. of two experiments performed in duplicate. *, p Ͻ 0.05; NS, not significant.
In conclusion, although distinct regulation of the IGFBP-1 and G6Pase gene promoters by phorbol esters has been demonstrated previously (39), this work is the first demonstration of a difference in the specific insulin signaling molecules required for repression of the G6Pase and IGFBP-1 genes. The requirement for mTOR (but not S6K) activity is analogous to Tor regulation of gene transcription in yeast. In addition, we propose the existence of a second insulin response element within the IGFBP-1 promoter that is insensitive to rapamycin and suggest that the transcription factor FKHR may not be the endogenous regulator of the IGFBP-1 TIRE.