O-GlcNAc Regulates FoxO Activation in Response to Glucose*

FoxO proteins are key transcriptional regulators of nutrient homeostasis and stress response. The transcription factor FoxO1 activates expression of gluconeogenic, including phosphoenolpyruvate carboxykinase and glucose-6-phosphatase, and also activates the expression of the oxidative stress response enzymes catalase and manganese superoxide dismutase. Hormonal and stress-dependent regulation of FoxO1 via acetylation, ubiquitination, and phosphorylation, are well established, but FoxOs have not been studied in the context of the glucose-derived O-linked β-N-acetylglucosamine (O-GlcNAc) modification. Here we show that O-GlcNAc on hepatic FoxO1 is increased in diabetes. Furthermore, O-GlcNAc regulates FoxO1 activation in response to glucose, resulting in the paradoxically increased expression of gluconeogenic genes while concomitantly inducing expression of genes encoding enzymes that detoxify reactive oxygen species. GlcNAcylation of FoxO provides a new mechanism for direct nutrient control of transcription to regulate metabolism and stress response through control of FoxO1 activity.

Nutrient homeostasis is tightly regulated to ensure the health of sensitive cells and tissues. Molecular mechanisms to counter the adverse effects of nutrient excess are overwhelmed in diabetes and result in diseases of the nervous system, heart, vasculature, and kidneys (1). Key molecules that respond to, and regulate, nutrient and stress levels include the forkhead transcription factor FoxO1 (FKHR) 2 and the post-translational modification O-GlcNAc.
The FoxO transcription factors are important regulators of broad gene expression programs, which include metabolism (e.g. gluconeogenesis, amino acid catabolism, glycolysis, pentose phosphate shunt, and fatty acid/triglyceride/sterol synthesis) (2), cell cycle (3), stress response (4,5), and longevity in Caenorhabditis elegans (6). Ubiquinitation and acetylation (7) are post-translational regulators of FoxO; however, phosphorylation by the insulin signaling pathway is the most well characterized signal transduction pathway that impinges upon FoxO (8). In the presence of insulin, this highly conserved pathway results in the nuclear exclusion of FoxO1 upon protein kinase B (PKB)/AKT phosphorylation.
Three key sites of AKT phosphorylation have been identified for FoxO1: threonine 24, serine 256, and serine 319. Homologs of FoxO1, including FoxO3 and FoxO4, contain conserved AKT phosphorylation sites and are regulated by insulin signaling in a similar manner (9). Nuclear exclusion of FoxO eventually leads to its ubiquitination and degradation (10). Stress stimuli also alter FoxO3 activation and turnover via deacetylation by SIRT1 (7).
The highly abundant and dynamic post-translational modification, O-GlcNAc, is implicated in stress responses, insulin signaling, nutrient sensing, cell cycle progression, protein turnover, translation, and transcription, in addition to other biological processes (11,12). O-GlcNAc rapidly cycles on serine and threonine residues of nuclear and cytoplasmic proteins in a fashion analogous to phosphorylation. Unlike phosphorylation, however, the addition of O-GlcNAc is performed by a single catalytic subunit (O-GlcNAc transferase; OGT) as opposed to hundreds of kinases and requires interactions with adaptor molecules for specificity to the hundreds of target proteins (11). O-GlcNAc functions as a nutrient sensor because the intracellular concentration of the donor sugar for GlcNAcylation, UDP-GlcNAc, rapidly responds to flux through multiple metabolic pathways, and OGT activity is highly dependent upon the concentration of this donor substrate (13). Elevated flux through the UDP-GlcNAc synthetic pathway results in insulin resistance (the hallmark of type II diabetes) of peripheral tissues (14). In murine adipocytes, elevated O-GlcNAc directly causes insulin resistance as measured by reduced 2-deoxyglucose uptake (15). Overexpression of OGT in muscle or adipose tissues of transgenic mice results in diabetes (16). However, elevated O-GlcNAc appears to be protective against heat shock (17). In the heart, O-GlcNAc is protective against ischemic stress (18), highlighting the importance of this post-translational modification to the stress response. In mammals and plants, deletion of OGT is lethal (19,20). However, in C. elegans, an OGT knock-out was found to have altered glycogen and trehalose storage and dauer formation induced by a temperature-sensitive daf-2 allele, indicating a role for OGT in insulin signaling (21).
Preliminary studies have shown FoxO1 to be GlcNAcylated (22,23); however, detailed mechanistic insights and their relevance to diabetes have not been elucidated. Here, we show that FoxO1 is GlcNAcylated in liver and that the modification is increased in diabetic rats. Furthermore, the hexosamine biosynthetic pathway serves to sense glucose levels and regulate FoxO1 transcriptional activation through GlcNAcylation of key amino acids in a complex interplay with FoxO phosphorylation.
Cell Culture and Treatments-Fao rat hepatoma cells were cultured in Coon's modification of Ham's F-12 medium (2 g/liter glucose; Sigma) supplemented with 5% fetal bovine serum. MEFs were cultured in DMEM (1 g/L glucose) supplemented with 10% fetal bovine serum. The cells were treated with 100 M O-(2-acetamido-2-deoxy-Dglusopyranoslidene) amino-N-phenylcarbamate (PUGNAc; Toronto Research Chemicals) overnight. The cells were treated with 1, 5, or 25 mM glucose or 25 mM glucose ϩ10 mM glucosamine in serum-free DMEM overnight. Insulin was used at indicated concentrations for 1 h.
Transcription Activation Assays-FoxO1 transcriptional activation was measured using a luciferase reporter construct pGL3 promoter, Promega) containing three copies of the FoxO1-binding site from the IGFBP promoter (29). 10 ng/well of the FoxO1 reporter construct as well as 10 ng/well of a ␤-galactosidase control construct were transfected into HEK293 cells in 24-well dishes using Lipofectamine 2000 (Invitrogen). pShuttle-OGT was used at 100 ng/well. The cells were transfected for 4 h in serum-free DMEM (Invitrogen). The medium was then replaced with serum-free DMEM containing varying amounts of glucose and incubated overnight. Luminescence was normalized to ␤-galactosidase activity.
Gene Expression Analysis-RNA was collected using TRIzol (Invitrogen) and cDNAs were reverse transcribed using Super Script II (Invitrogen). Real time PCR was performed on a Stratagene MX3000 using SYBR mix (Invitrogen). The data were normalized to 18 S rRNA.
Protein Interaction Analysis-Co-immunoprecipitation assays were performed on lysates from Fao cells infected with Ad-FLAG-HA-PGC-1a. Anti-FLAG agarose (Sigma) or anti-OGT (AL28) was applied to cells lysed in Tris-buffered saline with 1%Nonidet P-40. The immunoprecipitates were washed three times and separated by SDS-PAGE gel electrophoresis and blotted using anti-HA or anti-OGT (DM-17; Sigma).
UDP-GlcNAc Determination-UDP-GlcNAc levels were measured by capillary zone electrophoresis (32). OGT and Kinase Assays-Recombinant AKT was obtained from Cell Signaling (7500). GST-FoxO1 was labeled on beads for 1 h at room temperature in the presence of 200 M ATP as indicated by the manufacturer. Plasmid expressing recombinant OGT was a kind gift from Suzanne Walker (33). OGT assays following AKT labeling were performed on bead-bound FoxO1 for 2 h at room temperature in 50 mM Tris, pH 7.5. with 1 M UDP-GlcNAc. 0.25 Ci of [ 3 H]UDP-GlcNAc (American Radiolabeled Chemicals) was added to the reaction and incubated for 1 h at room temperature. The beads were then washed, and incorporated label was measured by scintillation counting or SDS-PAGE and autoradiography.
ETD MS/MS-GST-FoxO1 bound to glutathione-conjugated beads (Amersham Biosciences) was carbamidomethylated with dithiothreitol (Amersham Biosciences) and iodoacetamide (Sigma-Aldrich) at room temperature as previously described (34). The protein-bound beads were enzymatically digested with either endoproteinase LysC (Roche Applied Science) or endoproteinase AspN (Roche Applied Science) at an enzyme-to-substrate ratio of 1:20 in 100 mM ammonium bicarbonate, pH 8, at room temperature while shaking for 6.5 h. Supernatant containing the resulting proteolytic peptides was removed from the spun-down beads and acidified to pH 3.5 with glacial acetic acid (Sigma-Aldrich).
An aliquot of the supernatant from the digested sample was loaded onto a polyimide-coated, fused silica capillary reverse phase precolumn (360-m outer diameter ϫ 75-m inner diameter; Polymicro Technologies) packed with C18 resin (5-20-m diameter, 120-Å pore size; YMC Inc.) and desalted with 0.1% acetic acid. The precolumn was connected to a capillary analytical column (360-m outer diameter ϫ 50-m inner diameter) packed with C18 resin (5-m diameter, 120-Å pore size; YMC Inc.) and equipped with an integrated, electrospray emitter as described in Ref. 35. The peptides were eluted into the mass spectrometer at a flow rate of 60 nl/min with a gradient of 0 -60% B in 60 min and 60 -100% B in 5 min (A ϭ 0.1 M acetic acid, B ϭ 70% acetonitrile, 0.1 M acetic acid) using a 1100 series binary high pressure liquid chromatography (Agilent Technologies) (35). Aliquots of samples were analyzed on a Thermo Electron LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific). Full mass spectra were acquired with the Orbitrap analyzer operated at a resolving power of 30,000 (at m/z 400). Collisionactivated dissociation tandem mass spectra were acquired datadependently with the quadrupole linear ion trap analyzer. ETD MS/MS spectra were acquired on a Finnigan LTQ (Thermo Fisher Scientific) modified in house with a chemical ionization source to generate fluoranthene radical anions for fragmentation via ETD (45 msec ETD reaction) (36).Targeted ETD MS/MS were acquired on a second Finnigan LTQ (Thermo Fisher Scientific) equipped with a prototype Thermo Fisher ETD upgrade (100-ms ETD reaction, AGC target for ETD reagent ions 1E5). All of the data were interpreted manually.
Galactose Labeling Assays-O-GlcNAc levels were measured by [ 3 H]UDP-galactose labeling of GlcNAc (37). FoxO1 was immunoprecipitated using cold-labeled antibodies (Cell Signaling Technology) from 0.1 g of liver lysed in 10 ml of radioimmune precipitation assay buffer. The immunoprecipitates were then labeled overnight, separated by SDS-PAGE, and stained with Coomassie G-250. Following destaining, the gels were treated with En 3 Hance (PerkinElmer Life Sciences) and exposed to autoradiography film.
Animal Experiments-Male Sprague-Dawley rats (ϳ16 weeks old) were given a single injection of streptozotocin (50 mg of STZ/kg of body weight) to induce hyperglycemia. After 2 weeks the animals were sacrificed, and the livers were harvested. Control animals received vehicle alone.
For OGT overexpression studies, eight 9-week-old male balb/c mice (Taconic) were injected with 8.0 ϫ 10 8 adeno-OGT or adeno-GFP virus particles. Gene expression was corrected by 36B4 and normalized to GFP.

FoxO1 GlcNAcylation Is Elevated in Diabetes-Given that
FoxO1 is a transcription factor that controls many aspects of metabolism and stress responses in the liver and O-GlcNAc functions as a nutrient sensor, we tested whether FoxO1 is O-GlcNAc modified in the liver. Fig. 1 shows that FLAG-tagged FoxO1 expressed in Fao rat hepatoma cells is GlcNAcylated using the anti-O-GlcNAc antibody, CTD110.6, and the terminal GlcNAc-specific lectin, succinylated wheat germ agglutinin (Fig. 1A). Specificity was confirmed by GlcNAc competition and increased O-GlcNAc levels following treatment of the cells with the O-GlcNAcase inhibitor, PUGNAc. Endogenous FoxO1 immunoprecipitated from rat liver was shown to be O-GlcNAc modified by using a terminal GlcNAc-specific galactosyltransferase probe (Fig. 1B) (37). We next asked whether GlcNAcylation of FoxO1 is dysregulated by hyperglycemia in an model for diabetes. Using an O-GlcNAc-specific antibody (CTD110.6), we found that liver FoxO1 from STZinduced diabetic rats was more heavily GlcNAcylated ( Fig. 2A  and supplemental Fig. S1). The reduction in antibody reactivity by free GlcNAc further demonstrates antibody specificity. The donor sugar for GlcNAcylation is UDP-GlcNAc, a product of the hexosamine biosynthetic pathway (HBP). Elevated glucose can increase flux through the HBP resulting in increased UDP-GlcNAc (14), upon which OGT activity is dependent (13). We therefore measured UDP-GlcNAc levels in the control and STZ diabetic rat livers and found an increase in the sugar nucleotide ( Fig. 2B; 55.5 Ϯ 2.5 versus 88.6 Ϯ 12.3 pmol/mg liver, p Ͻ 0.05 by Student's t test). These data suggest that the HBP may act as a nutrient sensor by regulating OGT activity toward transcription factors, such as FoxO1. The STZ diabetic rats also had elevated mRNA expression of the FoxO1 targets Pepck and G6pc, consistent with previous reports (Fig. 2C) (38). Additionally, we measured the GlcNAcylation of FoxO1a from high fat fed diabetic mice and found an increase similar to the STZinduced diabetic rats (Fig. 2D).

FoxO Transcription Factors Integrate Nutrient and Hormone
Information-To confirm that hyperglycemia elevates GlcNAcylation of FoxO1, we employed a hepatoma cell culture model. Fig. 3A shows that culturing Fao hepatocytes in increasing glucose concentrations resulted in a dose-dependent increase in O-GlcNAc levels on FLAG-FoxO1 as detected by the CTD110.6 antibody. Interestingly, the addition of 100 nM insulin for 1 h down-regulated O-GlcNAc specifically on FoxO1 but did not reduce total cellular O-GlcNAc levels (Fig. 3, A and B) or alter O-GlcNAcase activity (Fig. 3C). We next assayed the O-GlcNAc levels of the insulin-insensitive mutant FoxO1 (T24A,S256A,S319A, termed "3A"). The lack of AKT phosphorylation sites results in much higher O-GlcNAc levels (Fig.  4A). To further address the interplay between O-GlcNAc and O-phosphate, we employed in vitro OGT and AKT assays. In vitro GlcNAcylation of GST-FoxO1 had no effect on subsequent in vitro AKT phosphorylation at Ser 256 (Fig.  4B). Additionally, in vitro AKT phosphorylation of GST-FoxO1 did not affect subsequent in vitro GlcNAcylation (Fig. 4C) Fig. 3A, it appears that AKT-mediated phosphorylation is dominant to GlcNAcylation with regards to FoxO. Another GST-tagged FoxO isoform, FoxO3 (FKHRL1), was also an in vitro target of OGT (Fig. 2E). A truncated FoxO3 (amino acids 1-525) incorporated less O-GlcNAc, showing that some sites are found between amino acids 526 and 673. A mutant lacking three AKT phosphorylation sites (T32A,S253A,S315A, termed "TM") of the truncated form labeled to the same extent as the truncation containing the hydroxyamino acids, consistent with the model that phosphorylation by AKT is not directly reciprocal with GlcNAcylation.
O-GlcNAc Regulates FoxO1 Activity-To investigate the role of O-GlcNAc on FoxO1, we first asked whether glucose could activate FoxO1-dependent transcription. Luciferase reporter assays in HEK293 cells (Fig. 5A) demonstrate that overnight incubation in 25 mM glucose resulted in increased FoxO1 transactivation. Immunoblotting showed that FoxO1 protein levels did not increase with increasing glucose. The AKT-insensitive FoxO1 mutant is also activated by glucose independent of insulin signaling. To further demonstrate that the hexosamine biosynthetic pathway can regulate FoxO1 in a mechanism that is distinct from nuclear localization, we tested FoxO1 3A activation while blocking the rate-limiting step in UDP-GlcNAc synthesis (Fig. 5B) with the glutamine:fructose-6-phosphate amidotransferase inhibitor, 6-diazo-5-oxonorleucine (DON). DON attenuated FoxO1 activation by 25 mM glucose (Fig.   5C), whereas the addition of glucosamine, which enters the UDP-GlcNAc synthesis pathway after glutamine: fructose-6-phosphate amidotransferase, restored FoxO1 activation, indicating that constitutively nuclear FoxO1 is activated by O-GlcNAc. The addition of glucosamine to the culture medium of Fao cells was confirmed to increase FoxO1 O-GlcNAc levels, whereas treatment with DON reduced FoxO1 glycosylation (Fig.  5D). Overexpressing OGT enhanced glucose-induced FoxO1 transcriptional activation (Fig. 5E) and GlcNAcylation (Fig. 5F).
To examine the effects of glucose sensing via O-GlcNAc on FoxO1 target genes, we performed RT-PCR analysis on mRNA isolated from Fao hepatoma cells cultured overnight in 1, 5, or 25 mM glucose. mRNA levels of Pepck and G6pc were elevated in a dose-dependent manner following overnight glucose treatment (Fig. 6A). We next asked whether other FoxO1 target genes are also elevated, specifically those involved in the oxidative stress responses, and found both catalase and MnSOD expression to be increased (Fig. 6B). This RT-PCR analysis was performed with either Ad-GFP or an adenovirus expressing a dominant negative FoxO1 mutant (⌬256) (39). The Ad-FoxO1 ⌬256-infected cells did not respond to glucose, demonstrating that the glucose sensing is specifically enhancing FoxO1-dependent transcription of gluconeogenesis and stress response genes. To confirm that the increase in catalase and MnSOD expression requires OGT, we employed MEFs that have loxP recombination sites flanking the OGT gene (19). mRNA expression of the stress response genes was not increased by culturing in high glucose conditions when the cells had been infected with an adenovirus expressing Cre recombinase causing a reduction in OGT compared with adeno-GFP controls (Fig. 6C). The reduced mRNA expression in the 25 mM glucose samples may be due to an effect on basal transcriptional machinery (40,41). Additionally, OGT overexpression (via adenovirus) in Fao cells was found to increase Pepck and G6pc expression as compared with GFP controls (Fig. 6D).

FIGURE 5. O-GlcNAc increases FoxO-dependent luciferase reporter transcription in HEK293 cells via the HBP.
A, luciferase reporter activity is increased by high glucose for both wild type (wt) and 3A FoxO1 (all panels are plotted as relative luciferase activity normalized to ␤-galactosidase activity; error bars indicate standard errors; *, p Ͻ 0.05 by Student's t test). FoxO1 protein levels are unaffected. B, schematic of the UDP-GlcNAc synthesis pathway in which 2-5% of glucose that enters the cells is used for production of the donor sugar nucleotide. C, addition of 50 M DON, which inhibits the rate-limiting enzyme in UDP-GlcNAc synthesis (glutamine:fructose-6-phosphate amidotransferase, GFAT), reduces high glucose activation of luciferase activity. To test how OGT affects gene expression in animals, we infected mice with adeno-OGT virus and preformed RT-PCR analysis from mRNA isolated from the livers. Fig. 6E shows that Pepck and G6pc expression are elevated by ϳ50% compared with adeno-GFP infected control mice, whereas Cox7a1 did not change significantly. These data, consistent with previous reports (42), suggest that, in response to glucose, FoxO1 and O-GlcNAc are involved in the ability of the cell to regulate transcription of gluconeogenic and stress response genes.
To better understand how O-GlcNAc regulates FoxO1, we investigated the DNA binding properties of this transcription factor. Using the FoxO-binding portion of the luciferase reporter vector used in Fig. 5, we performed electrophoretic mobility shift assays on recombinant FoxO1. No difference in DNA binding was found between naked or GlcNAc-FoxO1 nor between FoxO and FoxO incubated with OGT (Fig. 6, F and G). These results suggest that mechanisms other than nuclear localization or DNA binding, such as recruitment of transcription machinery, mediate the activation of FoxO1 by glucose.
To identify which of the 131 serine and threonine residues in human FoxO1 might be O-GlcNAcylated, we employed a relatively new technique, ETD MS/MS. The O-GlcNAc modification is lost under conventional, collision-activated dissociation mass spectrometry but retained with ETD (43). As a result, we were able to determine that FoxO1 residues Ser 550 , Thr 648 , Ser 654 , and either Thr 317 or Ser 318 were GlcNAcylated ( Fig. 7A and supplemental Fig. S3). We then mutated several of these sites and tested whether they were activated by glucose in a luciferase reporter assay. Mutation of threonine 317 to alanine reduced transcriptional activation by high glucose, whereas mutation of serine 318 had no effect (Fig.  7B). Under the conditions tested, mutation of the other identified sites had no effect in the FoxO1 luciferase reporter assay (supplemental Fig. S4). Western blot analysis showed reduced O-GlcNAc on the T317A mutant versus wild type after overnight culture in 25 mM glucose (Fig. 6C).

DISCUSSION
The highly abundant post-translational modification O-GlcNAc has been proposed to be a nutrient sensor as the donor sugar, UDP-GlcNAc, receives input from multiple metabolic pathways and that OGT activity depends upon UDP-GlcNAc concentration (11). We have shown that hyperglycemia elevates UDP-GlcNAc levels in rat liver, resulting in augmented FoxO1 GlcNAcylation and activation of this transcription factor. Early work showed that O-GlcNAc is enriched in chromatin (40) and at active sites of gene transcription in polytene chromosomes. Many transcription factors have been shown to be GlcNAcylated (for review see Ref. 12), and the RNA polymerase II C-terminal domain is extensively GlcNAcylated in a competitive manner with phosphorylation (41). The modification has been shown to regulate transactivation by potentially altering stability, nuclear localization, recruitment of transcriptional machinery, or DNA binding. In the case of Sp1, activation is regulated by O-GlcNAc and depending on cell type or promoter via DNA binding or other mechanisms (44,45). Another mechanism by which O-GlcNAc regulates transcription factors is through nuclear localization. Pancreatic NeuroD1 translocates to the nucleus under high glucose or inhibition of O-GlcNAcase (46). However, we have shown that a constitutively nuclear mutant FoxO1 (3A) is also regulated by glucose via the hexosamine biosynthetic pathway (Fig. 5), indicating that additional, intranuclear mechanisms of transactivation exist. It is possible that this mechanism involves GlcNAcylation of upstream regulators of FoxO because AKT and insulin receptor substrate 1 are also GlcNAcylated (47). This mechanism is unlikely, however, because the mutation of a very specific hydroxyamino acid (Thr 317 but not Ser 318 ), which we have shown to be GlcNAcylated, reduces FoxO1 stimulation by hyperglycemia. The fact that steady-state levels of FoxO1 protein did not vary in high glucose conditions speaks against FoxO regulation by the proteasome, a possibility given that FoxO is ubiquitinated (48) and proteasome function is inhibited by O-GlcNAc (49). Another possible mechanism for O-GlcNAc activation of FoxO1 is through mediating the recruitment of basal transcription factors, many of which are also GlcNAcylated (50). GlcNAcylation of Sp1 in its transactivation domain inhibits the interaction with TATA-binding-associated factor (TAF110) (51).
The discovery that GlcNAcylation of serine and threonine residues also occurs at phosphorylation sites led to the "Yin-Yang" hypothesis where O-GlcNAc moieties could directly oppose O-phosphate. We now know that this model is overly simple. When the three AKT phosphorylation sites are mutated in FoxO1, the O-GlcNAc levels become amplified.   's t test). C, HEK293 cells were transfected with FLAG-FoxO1 vectors containing the indicated site mutations, subjected to SDS-PAGE, and blotted using anti-O-GlcNAc antibodies (CTD110.6). IP, immunoprecipitation; IB, immunoblotting. assay, but determining the function of the other sites will require different assays and/or experimental conditions. If a GlcNAc site is also phosphorylated, mutation of a hydroxyamino acid to alanine will not be able to determine whether a phenotype for this mutant is due to the lack of phosphorylation or GlcNAcylation. What is clear is that multiple post-translational modifications at either shared or distinct sites give the cell greatly expanded molecular diversity.
This complex molecular diversity may allow a liver cell to sense and respond to multiple stimuli and respond with changes in transcription target activation and specificity. Here we show that glucose up-regulates mRNA expression of Pepck and G6pc via increasing GlcNAcylation of FoxO1 in the absence of insulin. This potentially establishes a dangerous positive feedback loop of gluconeogenesis and demonstrates that inappropriate gluconeogenesis (52) is not merely an effect of lack of insulin but a pathologically activated pathway (53). It has been asked why, from an evolutionary perspective, would an organism up-regulate glucosenegenesis under conditions of hyperglycemia (54)? The answer may lie in the finding that GlcNAcylation of FoxO1 also activates reactive oxygen species detoxification enzyme expression. Given that glucose metabolism leads to reactive oxygen species production, direct glucose control of this stress response pathway would be advantageous to the cell. O-GlcNAc senses and protects cells from stress (17,18,55), perhaps by activating expression of protective genes. Therefore, investigation into the pathological activation of gluconeogenesis as a drug target must also consider increased oxidative damage as a potential side effect.
In addition to gluconeogenesis, FoxO transcription factors regulate cell cycle, apoptosis, and longevity in C. elegans in addition to their roles in liver metabolism. O-GlcNAc is also implicated in cell cycle progression (56), apoptosis, and dauer formation in C. elegans, suggesting there may be other processes controlled by GlcNAcylation of FoxO1. Altogether, these data suggest a novel form of FoxO1 regulation relevant not only to diabetes, but to a wide array of cellular processes.