O-Linked N-Acetylglucosamine Modification on CCAAT Enhancer-binding Protein β

Abstract

CCAAT enhancer-binding protein (C/EBP)β is a basic leucine zipper transcription factor family member, and can be phosphorylated, acetylated, and sumoylated. C/EBPβ undergoes sequential phosphorylation during 3T3-L1 adipocyte differentiation. Phosphorylation on Thr¹⁸⁸ by MAPK or cyclin A/cdk2 primes the phosphorylations on Ser¹⁸⁴/Thr¹⁷⁹ by GSK3β, and these phosphorylations are required for the acquisition of DNA binding activity of C/EBPβ. Here we show that C/EBPβ is modified by O-GlcNAc, a dynamic single sugar modification found on nucleocytoplasmic proteins. The GlcNAcylation sites are Ser¹⁸⁰ and Ser¹⁸¹, which are in the regulation domain and are very close to the phosphorylation sites (Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹) required for the gain of DNA binding activity. Both in vitro and ex vivo experiments demonstrate that GlcNAcylation on Ser¹⁸⁰ and Ser¹⁸¹ prevents phosphorylation on Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹, as indicated by the decreased relative phosphorylation and DNA binding activity of C/EBPβ delayed the adipocyte differentiation program. Mutation of both Ser¹⁸⁰ and Ser¹⁸¹ to Ala significantly increase the transcriptional activity of C/EBPβ. These data suggest that GlcNAcylation regulates both the phosphorylation and DNA binding activity of C/EBPβ.

CCAAT enhancer-binding protein (C/EBP)² β is a basic leucine zipper transcription factor, expressed in variety of cells including adipocytes (1), hepatocytes (2), keratinocytes (3), epithelial cells (4), and blood cells (5). It plays an important role in adipocyte development (6), gluconeogenesis (7), liver regeneration (8), mammary gland development (9), hematopoietic system (10), and immune response to interferon γ (11). The functional importance of C/EBPβ during adipocyte development has already been demonstrated: overexpression of C/EBPβ in the nonpreadipocyte line of NIH 3T3 causes the commitment to the adipocyte lineage (12), and overexpression of C/EBPβ in 3T3-L1 preadipocytes is sufficient to induce adipocyte differentiation without hormonal inducers normally required (13).

C/EBPβ contains a C-terminal basic region leucine zipper DNA-binding domain, a dimerization domain, and an N-terminal transactivation domain together with a regulation domain (RD), which contains multiple putative modification sites in the middle. C/EBPβ assumes a tightly folded conformation in which the activation domain (N-terminal) and DNA-binding domain (C-terminal) are obscured by interaction with the RD. Appropriate modifications disrupt the RD interaction rending the binding domain accessible for DNA binding and thereby facilitating transactivation. Three phosphorylation sites (Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹) have been identified in this RD region of C/EBPβ during 3T3-L1 adipocyte differentiation (14). Both ex vivo and in vitro experiments demonstrate that phosphorylation on Thr¹⁸⁸ (first by MAPK during G₁ phase and late by CyclinA/cdk2 (15)) primes C/EBPβ for the phosphorylation on Ser¹⁸⁴ or Thr¹⁷⁹ by GSK3β, and these phosphorylations are required for the gain of DNA binding activity of C/EBPβ (14, 16).

O-Linked β-N-acetylglucosamine modification (GlcNAcylation) (17) is a novel glycosylation in which the monosaccharide β-N-acetylglucosamine (GlcNAc) attaches to serine/threonine residues via an O-linked glycosidic bond and is found mostly within the cytoplasm or nucleoplasm (18). GlcNAcylation regulates several important cellular processes, such as signal transduction, protein expression, degradation, and trafficking (19). This post-translational modification is regulated by two highly conserved enzymes: OGT and O-GlcNAcase (β-N-acetylglucosaminidase), whereas OGT catalyzes the attachment and O-GlcNAcase catalyzes the removal of O-GlcNAc to proteins (20, 21). Cycling of O-GlcNAc is dynamically regulated in a manner similar to phosphorylation. O-GlcNAc and O-phosphate site-mapping studies suggest that there are several different types of dynamic interplay between O-GlcNAc and O-phosphate: 1) competitive occupancy at the same site; 2) competitive and alternative occupancy at proximal sites; 3) a complex interplay whereby some O-phosphate attachment sites on a given protein are the same as some O-GlcNAc sites, whereas others are adjacent to, or even distant from each other; and 4) phosphorylation regulates the O-GlcNAc cycling enzymes and O-GlcNAc regulates kinases (19). The C-terminal domain of RNA polymerase II is extensively GlcNAcylated (22), and almost all RNA polymerase II transcription factors are modified by the sugar. GlcNAcylation can either suppress or enhance transcription, depending on the promoter involved, on the cell system, and on associations with different co-factors (23, 24).

C/EBPβ can be phosphorylated (25), acetylated (26), and sumoylated (27). In this study, we find that C/EBPβ is GlcNAcylated on Ser¹⁸⁰ and Ser¹⁸¹, which is in the RD region and adjacent to the phosphorylation sites (Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹) required for the gain of DNA binding activity. GlcNAcylation on Ser¹⁸⁰ and Ser¹⁸¹ prevent the phosphorylation on Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹. Mutations of Ser¹⁸⁰ and Ser¹⁸¹ to alanine residues significantly increase the transcriptional activity of C/EBPβ. This suggests that GlcNAcylation regulates both the phosphorylation and DNA binding activity of C/EBPβ, and GlcNAcylation and phosphorylation regulate the transcriptional activities of C/EBPβ in a competitive way and by alternative occupancy at adjacent sites.

EXPERIMENTAL PROCEDURES

Cell Culture, Induction of Differentiation, and Transfection

3T3-L1 preadipocytes were propagated and maintained in DMEM containing 10% (v/v) calf serum as described (30). To induce differentiation, 2-day postconfluent preadipocytes (designated day 0) were fed DMEM containing 10% (v/v) fetal bovine serum, 1 μg/ml insulin, 1 μm dexamethasone, and 0.5 mm 3-isobutyl-1-methylxanthine until day 2. The cells were then fed DMEM supplemented with 10% fetal bovine serum and 1 μg/ml insulin for 2 days, after which they were fed every other day with DMEM containing 10% fetal bovine serum. To test the effect of GlcNAcylation on the C/EBPβ phosphorylation and subsequently adipocyte differentiation, the post-confluent 3T3-L1 preadipocytes were treated with O-GlcNAcase inhibitor, GlcNAc thialzoline (GT; Johns Hopkins Medical Chemical Core) 1 h before and at the time exposed to the differentiation inducers. At the various times indicated, whole cell lysates were prepared and subjected to SDS-PAGE and detected with antibodies against total C/EBPβ, phospho-C/EBPβ, PPARγ, and 422/aP2 antibodies; relative protein levels were quantitated and plotted.

A C/EBPα promoter report construct was made by cloning 2044 bp of the mouse C/EBPα promoter into luciferase vector pGL3.0. C/EBPβ expression vectors were prepared by cloning wild type mouse C/EBPβ cDNA or mutants (Ser¹⁸⁰ → Ala and/or Ser¹⁸¹ →Ala) into pcDNA3.1, and the mutants were generated by using Stratagene Quick change site-directed mutagenesis kit. After reaching 90% confluency, 3T3-L1 cells were transfected with equal amount of report construct and/or C/EBPβ expression vectors, a thymidine kinase Renilla luciferase construct (Promega) was co-transfected and used as transfection efficiency control. Two days after transfection, the cells were induced as above, and 36 h after induction cells were lysed, and luciferase activity was determined by dual luciferase assay kit (Promega).

Western Blotting

Equal amounts of protein were subjected to SDS-PAGE and immunoblotted with antibodies against C/EBPβ, Thr(P)-188-C/EBPβ (Cell Signaling Technology, Beverley, MA), CTD110.6, 422/aP2, and PPARγ (Santa Cruz, Santa Cruz, CA). Antibody against C/EBPβ and 422a/P2 were prepared in this laboratory. CTD110.6 was used for O-GlcNAc detection (from Dr. G. W. Hart's lab).

Preparation of Recombinant C/EBPβ-LAP and C/EBPβ-LIP Protein

The cDNA encoding full-length C/EBPβ-LAP protein and the C-terminal 152–289 amino acids of C/EBPβ-LIP were cloned into the downstream of GST gene in pGEX-6P (Amersham Biosciences) and transformed into Escherichia coli strain BL21 (DE3) pLysS (Novagen). A single colony was propagated overnight in 3 ml of LB medium containing ampicillin and chloramphenicol, and then diluted (1:100) into 500 ml of fresh LB medium the next day and cultured until an A₆₀₀ of 0.6–0.7 was reached. Expression of the fusion proteins was induced by addition of 0.5 mm isopropyl β-d-thiogalactopyranoside for 3 h, and the cells were harvested and resuspended in 25 ml of phosphate-buffered saline containing 1% Triton X-100. After lysis by one cycle of freezing-thawing, the cell suspension was treated with DNase I and RNase A and then incubated on ice for 10 min in the presence of 0.5 m NaCl and 5 mm dithiothreitol to dissociate DNA-protein complex. After centrifugation at 12,000 rpm for 10 min, 250–500 μl of bed volume of GSH-Sepharose (GE Healthcare) was added to the supernatant and mixed overnight in a cold room. The beads were washed three times with phosphate-buffered saline, washed once with PreScission cleavage buffer (50 mm Tris-Cl, pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm dithiothreitol), and then treated with 80 units of PreScission Protease (GE Healthcare) in 960 μl of buffer. The cleaved recombinant C/EBPβ protein was further purified by CM-Sepharose column (GE Healthcare), and soluble, highly purified recombinant protein was eluted at 250–300 mm of NaCl. The protein concentration was determined by Bradford assay, and the purity of protein was verified by SDS-PAGE.

In Vitro Phosphorylation and Mass Analysis of C/EBPβ

Two micrograms of recombinant C/EBPβ-LAP were incubated with activated MAPK (Calbiochem) and GSK3β (Upstate) in buffer containing 50 mm HEPES (pH 7.0), 10 mm MgCl₂, 1 mm dithiothreitol, and 100 mm cold ATP or 20 μCi of [γ-³²P]ATP (1 Ci = 37 GBq) at 30 °C for 30 min. The phosphorylation was detected with Thr(P)¹⁸⁸ phospho-specific antibody or by autoradiography after the protein was separated by SDS-PAGE. For identification of phosphorylation sites on C/EBPβ, similar experiments were performed with 0.5 mm ATP. The protein band containing C/EBPβ was excised from the gel, digested with trypsin, and analyzed by nano liquid chromatography-tandem mass spectrometry as previously described (26).

To investigate the effect of O-GlcNAcylation on the O-phosphorylation of C/EBPβ, GST-C/EBPβ-LAP fusion protein (which bound to the GSH-Sepharose) was used as substrate, GST-C/EBPβ-LAP was GlcNAcylated by OGT first and then used for substrate for in vitro phosphorylation as described above, at times indicated reactions were stopped by adding loading buffer and separated by SDS-PAGE, the phosphorylation was detected by Western blotting with Thr(P)¹⁸⁸ phospho-specific antibody or autoradiography (detecting of all phosphorylation).

In Vitro O-GlcNAcylation and Identification of O-GlcNAc Sites of C/EBPβ

Two μg of recombinant C/EBPβ was incubated with 1 μg of recombinant OGT in buffer containing 50 mm sodium cacodylate (pH 6.5), 1 mg/ml bovine serum albumin, 35 mm NaF, 2 μm 2-acetamido-2-deoxy-d-gluconohydroximolacetone, 2.5 mm AMP, and 1 mm UDP-GlcNAc (or 1 μCi of [³H[UDP-GlcNAc) at 25 °C for 60 min. The GlcNAcylation was detected by Western blotting with O-GlcNAc-specific antibody CTD110.6 or autoradiography after the protein was separated by SDS-PAGE. For identification of O-GlcNAcylation sites on C/EBPβ, the protein band containing C/EBPβ was excised from the gel and destained. The protein bands were digested overnight with trypsin (Promega, Madison, WI), and then the peptides were extracted and lyophilized. For some samples, the generated tryptic peptides were subjected to a second digestion using chymotrypsin. In those cases the tryptic peptides were redissolved in 16 μl of 100 mm ammonium bicarbonate containing 10 ng/μl of high purity bovine pancreas chymotrypsin (Calbiochem, La Jolla, CA). The second part of this sequential digestion was also allowed to proceed overnight. All of the digestions were stopped with acetic acid. Lyophilized peptides was redissolved in 20 μl of 3% acetonitrile, 0.1% formic acid prior to liquid chromatography-tandem mass spectrometry analysis.

Mass spectrometry analyses were performed using a ion trap fitted with a chip cube LC system and electron transfer dissociation (ETD) capability (Agilent Technologies, Santa Clara, CA), as previously described. The ion trap was operated in 1) MS followed by alternating CID and ETD, 2) MS only; or 3) ETD only on selected precursors.

Detection of GlcNAcylation on Endogenous C/EBPβ

3T3-L1 cells were differentiated as above and harvested at the indicated time. C/EBPβ was immunoprecipitated, washed with lysis buffer, and washed twice with reaction buffer containing 20 mm HEPES, pH 7.9, 50 mm NaCl, 1 μm 2-acetamido-2-deoxy-d-gluconohydroximolacetone, and 5 mm MnCl₂ with protease and phosphatase inhibitors. Next, 2 μl of Gal-T1 Y289L (Invitrogen) and 2 μl of 0.5 mm UDP-GalNAz (Invitrogen) were added to a reaction volume of 20 μl. The reaction was performed overnight at 4 °C. The beads were washed twice with reaction buffer to remove excess UDP-GalNAz. The sampled were then reacted with biotin-alkyne (Invitrogen) according to the manufacturer's instructions. The enzymatic reactions were eluted with Laemmli buffer for immunoblot analysis.

Chromatin Immunoprecipitation (ChIP) Analysis

ChIP analysis was performed essentially as described in the protocol of ChIP assay kit (Upstate Biotechnology). Briefly, 3T3-L1 preadipocytes were maintained and induced to differentiate as above. O-GlcNAcase inhibitor (GT; Johns Hopkins Medical Chemical Core) was added 1 h before and added again at the time of induction. At various times indicated, the cells were cross-linked with 1% formaldehyde in phosphate-buffered saline buffer. The cross-linked cells were harvested, lysed in SDS lysis buffer, and sonicated. After incubation with salmon sperm DNA-protein A overnight at 4 °C, the DNA-protein complexes in the supernatant were immunoprecipitated with antibodies against C/EBPβ, and the immune complexes were recovered by adding protein A-agarose. After series washing, DNA was extracted with phenol/chloroform, precipitated, redissolved, and used as a template for PCR with the following primers. The primers used to amplify the DNA fragments containing the C/EBP-binding site in C/EBPα promoter: TCCCTAGTGTTGGCTGGAAG and CAGTAGGATGGTGCCTGCTG.

RESULTS

GlcNAcylation of Ser¹⁸⁰ and Ser¹⁸¹ in the RD of C/EBPβ

C/EBPβ has two isoforms (liver activation protein (LAP) and liver inhibitory protein (LIP); Fig. 1A) both of which are derived from the same mRNA by using different translation start codons (28). These two isoforms share the same DNA-binding domain and leucine zipper domain in their C terminus and the RD in the upstream of DNA-binding domain, whereas LAP has a transactivation domain in the N terminus and LIP does not. After treatment with OGT and unlabeled UDP-GlcNAc, both recombinant LAP and LIP are GlcNAcylated as evidenced by Western blotting with O-GlcNAc specific antibody CTD110.6 (Fig. 1, B and C, middle panels), and incubation with OGT and ³H-UDP-GlcNAc labels both LAP and LIP well, as indicated by autoradiography (Fig. 1, B and C, bottom panels). These data indicate that C/EBPβ can be GlcNAcylated and that the residue or residues very likely are located in the DNA-binding and/or regulation domain of C/EBPβ, which is shared by both LAP and LIP.

FIGURE 1.

C/EBPβ is a substrate for O-GlcNAc transferase. A, difference between C/EBPβ-LAP and C/EBPβ-LIP is illustrated; B and C, C/EBPβ-LAP or C/EBPβ-LIP was incubated with OGT in the presence of cold UDP-GlcNAc or [³H]UDP-GlcNAc as described. After separation by SDS-PAGE, GlcNAcylation on C/EBPβ was evidenced by Western blotting with O-GlcNAc specific antibody CTD110.6 (middle panels) or by autoradiography (bottom panels).

By combining ETD and CID mass spectrometry with liquid chromatography, we were able to map GlcNAcylation sites as well as O-linked phosphorylations of C/EBPβ. We used CID to identify peptides showing a characteristic fragment pattern for O-GlcNAc-modified peptides and ETD to pin-point the modified residue(s) of these peptides. After treatment with OGT and cold UDP-GlcNAc, C/EBPβ (LAP) were purified by SDS-PAGE. The C/EBPβ-LAP protein band was excised and trypsinized. A triply charged tryptic peptide AYLGYQATPSGSSGSLSTSSSSSPPGTPSPADAK was found to be a O-GlcNAc-modified peptide. However, because of the high number of potential modified residues (14 residues) neither CID nor ETD allowed us to accurately pin-point the modified residue. By subjecting the already trypsinized sample to a second digestion, this time using chymotrypsin, we were able to create smaller peptides carrying five and nine potential modified residues, respectively, thereby greatly simplifying our task. Peptides with the sequence STSSSSSPPGTPSPADAK showed the GlcNAcylation fragmentation characteristics and were selected for ETD-based GlcNAcylation site mapping. The same sequence was found to carry either one or two O-GlcNAc-modified residues. When carrying one O-GlcNAc, the modified residue was assigned as Ser¹⁸⁰. When carrying two O-GlcNAc modifications, the modified residues were pin-pointed as Ser¹⁸⁰ and Ser¹⁸¹. Fig. 2 shows the ETD spectra of the modified peptides. Fig. 2A shows the unmodified peptide, whereas the spectra of peptides carrying one and two O-GlcNAc units are shown in B and C.

FIGURE 2.

ETD to determine modified residues. Recombinant C/EBPβ-LAP was incubated with OGT in the presence of cold UDP-GlcNAc or incubated with MAPK and GSK3β, after separation with SDS-PAGE, C/EBPβ-LAP band was digested, and mass spectrometry analysis was performed. Peptides carrying one (B) and two (C) O-GlcNAc units or one (D), two (E), and three (F) O-linked phosphorylations were subjected to ETD. For comparison, A shows a spectrum of the unmodified peptides. For all five modified peptides the modification residues could accurately be determined. The dotted lines connecting panels vertically indicate mass shifts of fragment ions caused by the modifications. Ser¹⁸⁰ and Ser¹⁸¹ of C/EBPβ are identified as being modified by GlcNAcylation. O-Linked phosphorylation were found on Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹.

GlcNAcylation Prevents the Phosphorylation of C/EBPβ

GlcNAcylation on Ser¹⁸⁰ and Ser¹⁸¹ are in the RD region of C/EBPβ and adjacent to three phosphorylation sites (Thr¹⁷⁹, Ser¹⁸⁴, and Thr¹⁸⁸) we identified before (14). To confirm and extend our previous study, C/EBPβ-LAP was phosphorylated with MAPK and GSK3β, and phosphorylations on Thr¹⁸⁸, Ser¹⁸⁴ (Fig. 2, D and E), and Thr¹⁷⁹ (data not shown) were confirmed. Because the phosphorylations on Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹ are required for C/EBPβ to gain the DNA binding activity, the question was raised as to whether GlcNAcylation will affect the DNA binding activity. Recombinant C/EBPβ-LAP was phosphorylated with MAPK and GSK3β (or GlcNAcylated with OGT), and the phosphorylation was confirmed by Western blotting with Thr(P)¹⁸⁸ phospho-specific antibody (Fig. 3B) and mass spectrometry analysis (similar to Fig. 2, D and E, and data not shown). GlcNAcylation was confirmed by Western blotting with O-GlcNAc specific antibody CTD110.6 (Fig. 3C). Electrophoretic mobility shift analysis was performed to compare the DNA binding activity; consistent with our previous findings (14), unmodified C/EBPβ-LAP had little or no DNA binding activity, whereas the DNA binding activity dramatically increased after phosphorylation by MAPK and GSK3β (Fig. 3D). GlcNAcylated C/EBPβ-LAP did not bind to the C/EBP consensus sequence similar to unmodified C/EBPβ-LAP (Fig. 3D).

FIGURE 3.

Phosphorylation of C/EBPβ on Thr¹⁸⁸/Ser¹⁸⁴/Thr¹⁷⁹ by MAPK and GSK3β leads to the gain of DNA binding activity, whereas GlcNAcylation on Ser¹⁸⁰/Ser¹⁸¹ does not affect its DNA binding activity. Recombinant C/EBPβ-LAP was modified by MAPK and GSK3β or OGT in vitro, and an equal amount of protein was used as indicated by Coomassie Blue staining (A). The phosphorylation was confirmed by Western blotting with phospho-specific antibody (B). GlcNAcylation was confirmed by Western blotting with O-GlcNAc specific antibody CTD110.6 (C). DNA binding activity was detected by electrophoretic mobility shift analysis and compared (D). Modifications on C/EBPβ-LAP are illustrated (E). P-LAP, phosphorylated LAP; G-LAP, GlcNAcylated LAP.

GlcNAcylation (Ser¹⁸⁰ and Ser¹⁸¹) and phosphorylation (Thr¹⁷⁹, Ser¹⁸⁴, and Thr¹⁸⁸) do not occur on the same sites but occur very close to each other; next we asked whether these two different post-translational modifications could influence each other. To test this, GST-fused C/EBPβ-LAP was first GlcNAcylated and then used as the in vitro substrate for MAPK and GSK3β. We found that phosphorylation was significantly decreased and delayed as indicated by Western blotting (detecting Thr(P)¹⁸⁸) (Fig. 4) and autodiography (detecting all phosphorylations; data not shown).

FIGURE 4.

GlcNAcylation prevents the phosphorylation of C/EBPβ in vitro. GST-C/EBPβ-LAP was first GlcNAcylated by incubation with OGT and then used as the substrate for further phosphorylation. At various time points the reactions were stopped, and the protein was separated by SDS-PAGE. The GlcNAcylated C/EBPβ was detected by Western blotting with CTD110.6 antibody (middle panel), and phosphorylated protein was detected by phospho-C/EBPβ (Thr¹⁸⁸) antibody (top panel). The Coomassie Blue stain shows the loading control (bottom panel).

GlcNAcylation Modulates the DNA Binding Activity, and Mutations on Ser¹⁸⁰ and Ser¹⁸¹ Significantly Increase the Transcriptional Activity of C/EBPβ

A chemoenzymatic approach (29) was used to confirm GlcNAcylation on the endogenous C/EBPβ. Growth-arrested 3T3-L1 preadipocytes were induced to differentiation, and endogenous C/EBPβ was purified with immunoaffinify column (14), labeled by using a mutant galactosyltransferase GalT1 Y289L and an azide derivative of UDP-Gal (UDP-GalNAz) as donor substrate. The resulting azide-labeled C/EBPβ was chemically tagged with biotin-alkyne, and GlcNAcylation was detected by Western blotting with streptavidin-horseradish peroxidase. We found that C/EBPβ-LAP was strongly labeled with biotin, confirming the presence of O- GlcNAc in this protein (Fig. 5). To further test whether GlcNAcylation prevents phosphorylation of C/EBPβ in vivo, the GlcNAcylation of whole cell proteins was elevated, and the effects on the phosphorylation and DNA binding activity were investigated. After treatment with O-GlcNAcase inhibitor (GlcNAc-Thiazoline, GT), the GlcNAcylation of total cell proteins was significantly increased as detected by Western blotting with O-GlcNAc specific antibody CTD110.6 (Fig. 6C), we found that whole C/EBPβ protein level is increased (Fig. 6A) while relative phospho-C/EBPβ protein is decreased (Fig. 6, B and E), which indicates the ratio of phosphorylated C/EBPβ to total C/EBPβ is decreased (Fig. 6E). The ratio change was not due to the change in MAPK activity, which is responsible for the phosphorylation, and the phosphorylation and activation of MAPK indicated by p-MAPK, since p-MAPK was not affected by the O- GlcNAcase inhibitor (GT) (Fig. 6F).

FIGURE 5.

C/EBPβ is GlcNAcylated in vivo. Post-confluent 3T3-L1 preadipocytes were induced to differentiation as described, and nuclear extracts were prepared at 4h and 24h after induction. C/EBPβ was successfully immunoprecipitated as indicated by Western blotting with specific antibody (A), and GlcNAclation was detected as described (B).

FIGURE 6.

O-GlcNAcase inhibitor (GT) decreases the ratio of phospho-C/EBPβ to total C/EBPβ in 3T3-L1 preadipocytes. The post-confluent 3T3-L1 preadipocytes were treated with O-GlcNAcase inhibitor (GT) 1 h before and at the time exposed to the differentiation inducers. At various times indicated, whole cell lysates were prepared and subjected to SDS-PAGE and detected with antibodies against total C/EBPβ (A), phospho-C/EBPβ (B), O-GlcNAc (C), and MAPK and p-MAPK (F). The Coomassie Blue stain image was used to indicate the loading control (D); and phospho-C/EBPβ and total C/EBPβ were quantified and relative ratio was plotted in E. All of the experiments were repeated three times, and the data were statistical analyzed by using the Student's t test. *, 0.01 < p values < 0.05; **, p values < 0.01.

ChIP analysis was performed to test whether O-GlcNAcase inhibitor (GT) treatment also affected the DNA binding activity of C/EBPβ after the phospho-C/EBPβ was decreased. Postconfluent 3T3-L1 preadipocytes were treated with/out GT and 24 h later cells were cross-linked, and ChIP analysis was performed on the DNA binding activity of C/EBPβ on the C/EBPα promoter. We found that the binding of C/EBPβ to the C/EBP element in C/EBPα promoter was significantly decreased (Fig. 7A) after relative phospho-C/EBPβ was decreased, and treatment with the O-GlcNAcase inhibitor (GT) delays the 3T3-L1 adipocyte differentiation program as evidenced by decreasing the expression of PPARγ and 422/aP2 ∼20–30% after 48 and 72 h of MDI induction; however, the expression catches up at 96 h after MDI induction (Fig. 7, C and D).

FIGURE 7.

Increasing the GlcNAcylation of C/EBPβ decreases its DNA binding and transactivation activities and subsequently delays the adipocyte differentiation program. A, increasing the GlcNAcylation of C/EBPβ decreases its DNA binding activity. Postconfluent 3T3-L1 preadipocytes were treated with O-GlcNAcase inhibitor (GT) 1 h before and at the time exposed to the differentiation inducers. 24 h after induction ChIP analysis was performed to detect the C/EBPβ binding activity on the promoter of C/EBPα. B, mutations on both Ser¹⁸⁰ and Ser¹⁸¹ potentiate the transactivation activity. C/EBPα promoter report construct was co-transfected with wild type (wt) C/EBPβ-LAP or its mutants (Ser¹⁸⁰ → Ala and/or Ser¹⁸¹ → Ala) expression vectors into 3T3-L1 preadipocytes at 90% confluency, cell lysates were prepared 2 days late, and luciferase activity was determined and plotted. C, increasing the GlcNAcylation of C/EBPβ delays the adipocyte differentiation program. 3T3-L1 preadipocytes were treated with GT as above, cells were harvested 48, 72, and 96 h after induction, and immunoblotting analysis was performed to detect the expression of PPARγ (top panel) and 422/aP2 (middle panel). β-Actin was used as loading control (bottom panel). D, relative levels were quantified and plotted. All of the experiments were repeated three times, and the data were statistical analyzed by using the Student's t test. *, 0.01 < p values < 0.05; **, p values < 0.01.

Ser¹⁸⁰ and Ser¹⁸¹ were also mutated to further investigate the function role of GlcNAcylation on C/EBPβ. We found that C/EBPβ can significantly transactivate C/EBPα promoter reporter, but neither mutant of S180A nor mutant of S181A alone had any effect. However, mutations on both Ser¹⁸⁰ and Ser¹⁸¹ dramatically increased the transactivation of C/EBPβ (Fig. 7B).

DISCUSSION

We have used tandem mass spectrometry to identify the O-GlcNAc residues on C/EBPβ. Although this approach in most cases is very sensitive and accurate, O-GlcNAc-modified residues can be challenging to pin-point. First, enrichment of peptides carrying this modification is far from trivial, and second the O-GlcNAc modification is very labile when subjected to CID, making it difficult to accurately assign the modified residue(s) (31). To address the latter point, we decided to combine ETD and CID with liquid chromatography to map GlcNAcylation sites as well as O-linked phosphorylations of C/EBPβ. This newer fragmentation technique has proven to be an excellent choice to study so-called labile modifications exemplified by O-linked phosphorylations and O-GlcNAc (24). We provide clear mass spectrometry data to show that C/EBPβ is GlcNAcylated on Ser¹⁸⁰ and Ser¹⁸¹ by OGT and phosphorylated on Thr¹⁸⁸/Ser¹⁸⁴/Thr¹⁷⁹ by MAPK and GSK3β (Fig. 2).

Increasing evidence indicates that most RNA polymerase II transcriptional factors are modified by both phosphorylation and GlcNAcylation, whereas phosphorylation and GlcNAcylation often reciprocally regulate the activities of these transcriptional factors (32). c-Myc is a helix-loop-helix leucine zipper transcription factor. Ser⁶² and Thr⁵⁸ are identified as phosphorylation sites, and phosphorylation on Ser⁶² by MAPK primes the phosphorylation on Thr⁵⁸ by GSK3β. Thr⁵⁸ is also identified as the major GlcNAcylation site of c-Myc. Thr⁵⁸ is located in the transactivation domain, and the competitive occupancy at this amino acid (Thr⁵⁸) plays an important role in regulation of the functions of c-Myc (33). FoxO1 is an important transcription factor regulating broad gene expression programs including metabolism, cell cycle, longevity, and stress response. The phosphorylation of FoxO1 by insulin signaling pathway has been well characterized, and phosphorylations on Thr²⁴/Ser²⁵⁶/Ser³¹⁹ exclude the FoxO1 from nucleus. FoxO1 is also GlcNAcylated, and the GlcNAcylation sites are mapped to Thr³¹⁷/Ser⁵⁵⁰/Thr⁶⁴⁸/Ser⁶⁵⁶. Further evidence demonstrates that the phosphorylation sites do not overlap directly with but are adjacent to the GlcNAcylation sites, suggesting a complex interplay between phosphorylation and GlcNAcylation of FoxO1. Interestingly, the GlcNAcylation of FoxO1 increases its transactivation capability in response to high glucose (24).

Our previous data demonstrate that the transcription factor C/EBPβ is sequentially phosphorylated on Thr¹⁸⁸/Ser¹⁸⁴/Thr¹⁷⁹, which is similar to c-Myc; phosphorylation on Thr¹⁸⁸ by MAPK or CyclinA/cdk2 primes the phosphorylations on Ser¹⁸⁴/Thr¹⁷⁹ by GSK3β; and phosphorylations on Thr¹⁸⁸/Ser¹⁸⁴/Thr¹⁷⁹ are required for the acquisition of DNA binding activity by C/EBPβ. In this paper we demonstrate that C/EBPβ is also modified by O-GlcNAc (Fig. 1), and the GlcNAcylation sites are on Ser¹⁸⁰ and Ser¹⁸¹ (Fig. 2), which are very close to its phosphorylation sites (Thr¹⁸⁸, Ser¹⁸⁴, and Thr¹⁷⁹) and both of which are located in the RD of C/EBPβ. Although GlcNAcylation itself does not affect the DNA binding activity in vitro (Fig. 3), it prevents the phosphorylations (Fig. 4). C/EBPβ is GlcNAcylated during 3T3-L1 adipocyte differentiation (Fig. 5), and elevation of GlcNAcylation significantly decreases the phosphorylation (Fig. 6) as well as the DNA binding activity of C/EBPβ in vivo (Fig. 7A), subsequently delaying the adipocyte differentiation program (Fig. 7, B and C). Importantly mutations on Ser¹⁸⁰ and Ser¹⁸¹ significantly increase the transactivation activity of C/EBPβ detected by C/EBPα promoter reporter assay (Fig. 7B), indicating that blocking of GlcNAcylation favors the phosphorylations. We conclude that GlcNAcylation and phosphorylation act in a competitive manner and by alternative occupancy at adjacent sites to regulate the function of C/EBPβ, which is similar to the situation with FoxO1.