K

Archives of Biochemistry and BiophysicsVol. 389, No. 2, May 15, pp. 166–175, 2001doi:10.1006/abbi.2001.2331, available online at http://www.idealibrary.com on

Hyperglycemia and the O-GlcNAc Transferase in RatAortic Smooth Muscle Cells: Elevated Expressionand Altered Patterns of O-GlcNAcylation

Yoshihiro Akimoto,*,1 Lisa K. Kreppel,† Hiroshi Hirano,‡ and Gerald W. Hart†,2

*Department of Biochemistry and Molecular Genetics Schools of Medicine/Dentistry, University of Alabama atBirmingham Station, Alabama 35294; †Department of Biological Chemistry, The Johns Hopkins UniversitySchool of Medicine, Baltimore, Maryland 21205; and ‡Department of Anatomy,

yorin University School of Medicine, Mitaka Tokyo 181-8611, Japan

Received December 6, 2000, and in revised form February 7, 2001; published online April 24, 2001

Hyperglycemia leads to vascular disease specific todiabetes mellitus. This pathology, which results fromabnormal proliferation of smooth muscle cells in arte-rial walls, may lead to cataract, renal failure, and ath-erosclerosis. The hexosamine biosynthetic pathway isexquisitely responsive to glucose concentration andplays an important role in glucose-induced insulin re-sistance. UDP-GlcNAc: polypeptide O-N-acetylglu-cosaminyltransferase (O-GlcNAc transferase; OGTase)catalyzes the O-linked attachment of single GlcNAcmoieties to serine and threonine residues on manycytosolic or nuclear proteins. Polyclonal antibodyagainst OGTase was used to examine the expression ofOGTase in rat aorta and aortic smooth muscle (RASM)cells. OGTase enzymatic activity and expression at themRNA and protein levels were determined in RASMcells cultured at normal (5 mM) and at high (20 mM)glucose concentrations. OGTase mRNA and proteinare expressed in both endothelial cells and smoothmuscle cells in the aorta of normal rats. In both celltypes, the nucleus is intensely stained, while the cyto-plasm stains diffusely. Immunoelectron microscopyshows that OGTase is localized to euchromatin andaround the myofilaments of smooth muscle cells. InRASM cells grown in 5 mM glucose, OGTase is alsolocated mainly in the nucleus. Hyperglycemic RASMcells also display a relative increase in OGTase’s p78subunit and an overall increase protein and activityfor OGTase. Biochemical analyses show that hypergly-cemia qualitatively and quantitatively alters the gly-

1 To whom correspondence should be addressed. Fax: 410-614-8804. E-mail: gwhart@jhmi.edu.

2

Current address: Department of Anatomy, Kyorin UniversitySchool of Medicine, Mitaka Tokyo 181 Japan.

166

cosylation or expression of many O-GlcNAc-modifiedproteins in the nucleus. These results suggest that theabnormal O-GlcNAc modification of intracellular pro-teins may be involved in glucose toxicity to vasculartissues. © 2001 Academic Press

Key Words: O-GlcNAc transferase; aortic smoothmuscle cells; hyperglycemia; diabetes; immunohisto-chemistry; rat; glucosamine.

The hexosamine biosynthetic pathway is one of thepossible mechanisms mediating the induction of insu-lin resistance (1–7). Glutamine:fructose-6-phosphateamidotransferase (GFAT)3 is the rate-limiting enzymeof the hexosamine pathway and is a key regulator ofthis pathway. Overexpression of GFAT induces an in-crease in the hexosamine (UDP-GlcNAc) concentrationand results in insulin resistance (8–10). UDP-GlcNAcis the final product of the hexosamine pathway and isa donor for both N-linked and O-linked protein glyco-sylation. O-linked N-acetylglucosamine moieties (O-GlcNAc) are attached to serine or threonine residues ofmany nuclear and cytoplasmic proteins (11). O-GlcNAc

3 Abbreviations used: GFAT, glutamine:fructose-6-phosphateamidotransferase; UDP-GlcNAc, Uridine diphospho-N-acetylglu-cosamine; O-GlcNAc, O-linked N-acetylglucosamine moiety; OG-Tase, O-GlcNAc transferase; IRS, insulin receptor substrate; RASM,rat aortic smooth muscle; FBS, fetal bovine serum; DMEM, Dulbec-co’s modified Eagle’s medium; PBS, phosphate-buffered saline; BSA,bovine serum albumin; DAPI, 49,69-diamidine-2-phenylindole hydro-chloride; DIG, digoxigenin; DEPC, diethylpyrocarbonate; SSC, stan-dard sodium citrate; NBT, nitroblue tetrazolium chloride; BCIP,5-bromo-4-chloro-3-indoylphosphate; PMSF, phenylmethylsulfonyl

fluoride; PVDF, polyvinylidene difluoride; DTT, dithiothreitol; PKC,protein kinase C; STZ, streptozotocin.

0003-9861/01 $35.00Copyright © 2001 by Academic Press

All rights of reproduction in any form reserved.

tPpwd5

167HYPERGLYCEMIA AND O-GlcNAc IN RASM CELLS

modification, which is termed O-GlcNAcylation, is aregulatory modification that has a complex dynamicinterplay with O-phosphorylation (12).

O-GlcNAc transferase (OGTase: EC 2.4.1) has beenpurified from the cytosol of rat liver (13). This enzymeis a unique nuclear and cytosolic glycosyltransferasethat contains multiple tetratricopeptide repeats (14).The liver enzyme contains two immunologically re-lated subunits of Mr 110 kDa (a-subunit) and 78 kDa(b-subunit). Other tissues, such as brain, contain onlythe a-subunit, which contains the active site. Bothsubunits of the enzyme are O-GlcNAcylated and ty-rosine-phosphorylated. The a-subunit forms homo- andheterotrimers that appear to have different bindingaffinities for UDP-GlcNAc over the entire physiologicalrange (15). The cDNA encoding the OGTase a-subunithas been cloned from rat, Caenorhabditis elegans, andhumans (14, 16). OGT is highly conserved at the pri-mary sequence level from C. elegans to humans. TheOGTase protein is unlike any glycosyltransferase pre-viously described (17, 18). The OGTase gene resides onthe X chromosome and is essential for embryonic stemcell viability and for normal development (19). Eventhough OGTase is ubiquitously expressed in all tissuesexamined so far, pancreatic cells are highly enriched inOGTase mRNA (12-fold over other tissues), furthersuggesting the involvement of O-GlcNAc in the main-tenance of glucose homeostasis (16). We previouslydemonstrated that OGTase and O-GlcNAc are abun-dant in almost all pancreatic cells and also abundant inthe islets of Langerhans (20). We also showed thatexpression and activity of OGTase and the extent ofO-GlcNAcylation on proteins were elevated in thestreptozotocin (STZ)-induced diabetic rat pancreas(21). Glucose and STZ substantially enhance the O-GlcNAcylation of a 135-kDa protein in the pancreaticislets (22). Especially in the pancreatic beta cells, glu-cose dramatically stimulates the O-GlcNAcylation ofproteins (23). The O-GlcNAcylation of the p62 nuclearpore protein is also responsive to glucose (24). OGTasemay be one of the key enzymes downstream fromGFAT that is involved in insulin resistance.

Prolonged exposure to hyperglycemia leads to a vas-cular disease specific to diabetes mellitus (25, 26).Much of the pathology results from abnormal prolifer-ation of smooth muscle cells in arterial walls. Over-time, this glucose-induced hyperproliferation may leadto increased risk of atherosclerosis. A very high expres-sion of GFAT is found in vascular smooth muscle cells(27). This indicates that GFAT is possibly involved inthe development of the vascular complications of dia-betes (28). O-GlcNAc levels on intracellular proteins,including insulin receptor substrates 1 and 2 (IRS, 1and 2) of vascular smooth muscle and skeletal muscle

cells are elevated in response to glucose and glu-cosamine (28–30).

In the present study, the expression of OGTase atboth the protein and the mRNA levels was examinedimmunohistochemically by using a polyclonal antibodyspecific for OGTase protein and by in situ hybridizationanalysis. To further clarify the involvement of OGTasein glucose toxicity, we investigated the effects of highglucose on the expression of OGTase in rat aorticsmooth muscle cells. To detect hyperglycemia-inducedalterations in O-GlcNAcylated proteins in the nucleus,where O-GlcNAc proteins are particularly abundant,we exogenously labeled O-GlcNAc-modified proteins byusing purified galactosyltransferase and UDP-[3H]galactose and subsequently examined them by 2D-gel electrophoresis. We report that the O-GlcNAcyla-tion level of many proteins is substantially altered inresponse to high glucose. High glucose also elevatesboth the expression and activity of OGTase.

MATERIALS AND METHODS

Cell culture. Rat aortic smooth muscle (RASM) cells were iso-lated from the aorta (250–300 g, male Sprague–Dawley) (31). Thecells were cultured routinely in Dulbecco’s modified Eagle’s medium(DMEM; 5 mM glucose) in the presence of 10% (v/v) fetal bovineserum (FBS), passaged every 5 days at 1:10, and not used after morethan six passages.

Antibodies. Rabbit polyclonal anti-OGTase antibody (AL-25, pu-rified IgG) was raised against purified recombinant 110-kDa subunitof OGTase that was expressed in E. coli (14). AL-25 recognizes both110- and 78-kDa subunits of OGTase. Moreover, OGTase enzymaticactivity is precipitated by AL-25 (14). Cy3-conjugated donkey anti-rabbit IgG antibody and FITC-conjugated donkey anti-mouse IgGwere obtained from Jackson Immunoresearch (West Grove, PA).FITC-conjugated phalloidin was purchased from Molecular Probes(Eugene, OR).

Immunostaining for light microscopic observation. Aorta wasfixed in 4% (v/v) formaldehyde in 0.1 M phosphate buffer (pH 7.3) for1 h at 4°C, immersed in 2.3 M sucrose–phosphate-buffered saline(PBS) for 1 h at 4°C, and then frozen with liquid nitrogen. Semithinfrozen sections of 1–2 mm thickness were cut, washed with PBS, andreated for 10 min with 1% (w/v) bovine serum albumin (BSA) inBS. The sections were then incubated with AL25 (1:250) or withreimmune rabbit IgG (1:250) for 1 h at room temperature, washedith PBS, and subsequently incubated for 1 h with Cy3-conjugatedonkey anti-rabbit IgG antibody (1:1000) and FITC–phalloidin (1:0). Nuclei were stained with 49,69-diamidine-2-phenylindole hydro-

chloride (DAPI, Boehringer Mannheim).RASM cells were fixed in 4% (v/v) formaldehyde in 0.1 M phos-

phate buffer (pH 7.3) for 1 h at 4°C, permeabilized with 0.5% (v/v)Triton X-100 in PBS for 5 min, and treated with 1% (w/v) BSA in PBSfor 10 min. The specimens were then incubated with the AL25 (orwith preimmune rabbit IgG for 1 h at room temperature), washedwith PBS, and subsequently incubated for 1 h with Cy3-conjugateddonkey anti-rabbit IgG antibody (Jackson Immunoresearch).

After a wash with PBS, the specimens were mounted in 90% (v/v)glycerol–0.1 M Tris–HCl buffer (pH 8.5) containing 0.5 mM p-phe-nylene diamine and observed under a Leica microscope equippedwith an epifluorescence system and chilled 3CCD color camera(C5810, Hamamatsu Photonics Systems, Bridgewater, NJ).

Preparation of cRNA probes. Antisense and sense RNA probeswere prepared by in vitro transcription from a fragment of OGTase

cDNA (from nt 1193 to nt 1942 of rat OGTase cDNA sequence in thedatabase [GenBankTM/EBI Data Bank, Accession No. U76557]) by

168 AKIMOTO ET AL.

opyD((Ht(Tmnmiwtwbbs5h

m(aUnb1cmcm

RRRfis

mTam

(ppcatr(o

b

dnn1w2u

td

ramSbpie

Rb

Ebnm4dKap

pwF(efiC

169HYPERGLYCEMIA AND O-GlcNAc IN RASM CELLS

use of T3 or T7 RNA polymerase in the presence of digoxigenin(DIG)-linked UTP in the reaction mixture as described pre-viously (20).

In situ hybridization. In situ hybridization was performed withDIG-cRNA on cryosections as described previously (20). Cryosectionsof unfixed rat aorta were cut at a 4-mm thickness, thaw-mountednto silane-coated slide glasses, allowed to air dry, fixed in 4%araformaldehyde–PBS for 1 h, incubated twice in 0.1% (w/v) dieth-lpyrocarbonate (DEPC)–PBS (15 min each time), and rinsed withEPC-treated citrate-buffered saline at a concentration of 43 SSC

32). The sections were then prehybridized in prehybridization buffer50% formamide, 13 Denhardt solution, 0.6 M NaCl, 10 mM Tris–Cl, 1 mM EDTA, 100 mg/ml salmon sperm DNA, 100 mg/ml yeast

RNA) at 45°C for 3 h and then hybridized in hybridization buffer50% (v/v) formamide, 13 Denhardt solution, 0.6 M NaCl, 10 mMris–HCl (pH 7.4), 1 mM EDTA, 100 mg/ml salmon sperm DNA, 0.4g/ml DIG-labeled probe RNA) at 50°C for 15 h. The sections wereext sequentially rinsed with DEPC-treated 43 SSC at 45°C for 15in and with 50% formamide–23 SSC at 45°C for 15 min, and then

ncubated in 1 mg/ml RNase A solution at 37°C for 15 min. Next, theyere rinsed with DEPC-treated 23 SSC at 45°C for 15 min, and then

wice with DEPC-treated 0.53 SSC at 45°C for 15 min. The sectionsere incubated in a solution of polyclonal sheep anti-DIG Fab anti-ody conjugated with alkaline phosphatase [1:500 diluted in DIGuffer (0.1 M Tris–HCl [pH 9.5], 0.1 M NaCl, 50 mM MgCl2)]. Theignal was detected with nitroblue tetrazolium chloride (NBT) and-bromo-4-chloro-3indolylphosphate (BCIP) (Boehringer-Mann-eim).Postembedding immunostaining for electron microscopy. Speci-ens were fixed in 4% (v/v) formaldehyde in 0.1 M phosphate buffer

pH 7.3) for 1 h at 4°C. The fixed aorta was dehydrated in ethanolnd then infiltrated with LR White (London Resin, Basingstoke,K). Ultrathin sections were cut and picked up on Formvar-coatedickel grids, treated for 10 min with 1% (w/v) BSA, and then incu-ated with AL 25 or rabbit preimmune IgG at room temperature forh. The grids were washed with PBS and then incubated for 1 h with

olloidal gold (10 nm)-conjugated goat anti-rabbit IgG (Jackson Im-unoresearch). Sections were stained with uranyl acetate and lead

itrate and examined with a JEM-1200EX transmission electronicroscope (JEOL, Tokyo, Japan).

Northern blotting for detecting expression of OGTase mRNA inASM cells. Total RNA was isolated from RASM cells by using anNeasy Total RNA kit (Qiagen, CA). Twenty-five micrograms of eachNA was resolved on a 1% (w/v) agarose gel, transferred to a Nytranlter, and probed with OGTase cDNA PCR-22b fragment as de-cribed previously (14).Western blot analysis. RASM cells cultured in the presence of 5.5M glucose or 20 mM glucose were washed with TBS–EDTA (20 mMris–HCl at pH 7.4, 155 mM NaCl, 1 mM EDTA) and then lysed inlysis buffer containing 20 mM Tris–HCl (pH 7.4), 2% (w/v) SDS, 5M EDTA, 5 mM EGTA, 1 mM phenylmethylsulfonyl flouride

FIG. 1. Immunohistochemical localization by light microscopy of OGortion of the wall of aorta (a cross section through the aorta). Semithiith FITC-phalloidin (green), and for nuclear DNA with DAPI (blue). (a-actin and nuclear DNA. (c) Triple exposure image of OGTase, F-actin

S), nuclei show intense staining, whereas cytoplasm emits weak andndothelial cells and smooth muscle cells in rat aorta. (d) Phase-contraxed with 4% (v/v) paraformaldehyde, permeabilized with 0.5% Try3-donkey anti-rabbit IgG. (f) Double exposure image of phase-contras

FIG. 2. OGTase mRNA is expressed in both endothelial cells andexamined by in situ hybridization histochemistry using DIG-labe

endothelial cells (EC) and smooth muscle cells (S). b) No hybridization slamella; L, lumen. Bar, 10 mm.

PMSF), protease inhibitor cocktail 1 and 2 (1:1000 dilution). Theroteins were separated by 10% SDS–PAGE and transferred to aolyvinylidene difluoride (PVDF) membrane. Purified rabbit poly-lonal IgG AL-25 (1:5000) was used as the primary antibody, andnti-rabbit IgG coupled to horseradish peroxidase (Amersham), ashe secondary antibody (1:20,000 dilution). Detection of the horse-adish peroxidase activity was enhanced by chemiluminescenceECL), as described by the manufacturer (Amersham). The intensityf protein bands was quantified by scanning densitometry.OGTase assay of RASM cells. RASM cells were homogenized in

uffer (10 mM Tris–HCl, pH 7.5, 10 mM MgCl2, 1 mM EDTA, 1 mMPMSF) in a Dounce homogenizer and sonicated. The insoluble ma-terial was pelleted by centrifugation at 27,500g for 30 min at 4°C and

iscarded. The supernatant was made 30% saturated with ammo-ium sulfate, after which the supernatants were allowed to sit over-ight on ice. The precipitate was then collected by centrifugation at2,000g for 20 min. The supernatants were discarded and the pelletsere resuspended in buffer (20 mM Tris–HCl, pH 7.8, containing0% (v/v) glycerol). The insoluble material was removed by centrif-gation at 12,000g for 15 min at 4°C.The assay was performed as described previously (13). The reac-

ion mixture for the standard assay contained 50 mM sodium caco-ylate, pH 6.0, 150 mg of the synthetic peptide YSDSPSTST, 2.5 mM

59-adenosine monophosphate, and 0.5 mCi of UDP-[3H]GlcNAc. Theeaction was started by the addition of enzyme and continued for 15nd 30 min at 20°C. The reaction was stopped by the addition of 50M formic acid, and the mixture was then loaded onto a 0.5-mlP-Sephadex (SP-C25-120, Sigma) column equilibrated in the sameuffer. The column was washed with 50 mM formic acid, and theeptides were eluted with 0.5 M NaCl. Incorporation of [3H]GlcNAcnto the peptide was quantified by liquid scintillation spectrophotom-try.

Two-dimensional gel of RASM cell nuclear extract proteins.ASM cells were collected, washed, and suspended in hypotonicuffer (10 mM Hepes at pH 7.9, 1.5 mM MgCl2, 10 mM KCl, 0.2 mM

PMSF, 0.5 mM dithiothreitol (DTT), 15 mM GlcNAc). The swollencells were homogenized, and the nuclei were pelleted by centrifuga-tion at 4000 rpm for 15 min at 4°C. The nuclei were resuspended ina low-salt buffer [20 mM Hepes at pH 7.9, 25% (v/v) glycerol, 1.5 mMMgCl2, 20 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT, 15mM GlcNAc]. Gentle dropwise addition of a high-salt buffer (20 mMHepes at pH 7.9, 25% glycerol, 1.5 mM MgCl2, 1.2 M KCl, 0.2 mM

DTA, 0.2 mM PMSF, 0.5 mM DTT, 15 mM GlcNAc) released solu-le proteins from the nuclei without lysing the nuclei. After theuclei had been extracted for 30 min at 4°C with continuous gentleixing, they were removed by centrifugation for 30 min at 25,000g at

°C. The supernatant (nuclear extract) was then dialyzed againstialysis buffer (20 mM Hepes at pH 7.9, 20% (v/v) glycerol, 100 mMCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT, 15 mM GlcNAc)nd centrifuged for 20 min at 14,500 rpm at 4°C, after which theellet was discarded.

se in the rat aorta and the RASM cells. Enlarged view of the luminalozen sections are triple-stained for OGTase with Cy3 (red), for F-actin

munofluorescence staining for OGTase. (b) Double-exposure image ofd nuclear DNA. In both endothelial cells (EC) and smooth muscle cellsuse signals. OGTase is localized in the nucleus and cytoplasm of bothmage. (e) RASM cells cultured in the presence of 5.5 mM glucose were

X-100, reacted with anti-OGTase antibody (AL25) and then withage and nuclear DNA (blue). EL, elastic lamella; L, lumen. Bar, 10 mm.ooth muscle cells in rat aorta. a) Distribution of OGTase mRNAanti-sense cRNA probe. Hybridization signals are seen in both

Tan fr) Im

, andiffst iitont im

smled

ignals are observed with DIG-labeled sense cRNA probe. EL, elastic

wdm erv

170 AKIMOTO ET AL.

Terminally O-GlcNAc-modified proteins among the nuclear ex-

FIG. 3. Immunohistochemical localization by electron microscopy ohite-embedded sections of aorta were reacted with AL25 (a) oriameter)-conjugated goat anti-rabbit IgG. The cell nuclei, especiayofilaments are labeled (a). The colloidal gold label is scarcely obs

tracted proteins were detected by the galactosyltransferase probemethod using purified Galb (1–4) galactosyltransferase and UDP-

[3H] galactose as described (33). The proteins were resolved by two-

GTase in an aortic smooth muscle cell. Formaldehyde-fixed and LRbbit preimmune IgG (b), and then with colloidal gold (10 nm inthe euchromatin, and the cytoplasm in the immediate vicinity ofed in the cytochemical control (b). N, nucleus. Bar, 1 mm.

f Orally

dimensional gel electrophoresis (first dimension, isoelectric focusinggel electrophoresis, ampholyte pH range of 3–10; second dimension,

ps(

DaO(Tieme*

laO

171HYPERGLYCEMIA AND O-GlcNAc IN RASM CELLS

10% SDS–acrylamide gel electrophoresis). The [3H]galactose-taggedroteins were detected by fluorography of the gel treated with 1 Modium salicylate. The spots were analyzed with PDQuest softwareBio-Rad Lab, CA).

RESULTS AND DISCUSSION

Immunolocalization of OGTase in Aorta and inRASM cells. The distribution of OGTase in the rataorta was examined by immunofluorescence micros-

FIG. 4. Northern blot analysis of expression of OGTase mRNA inRASM cells. (A) Northern blotting of OGTase transcripts in the cellscultured with 5.5 mM (control) or 20 mM glucose (high glucose).Total RNA (25 mg) was probed with the OGTase cDNA mPCR-22b

NA fragment. The arrows at the left indicate the positions of 28Snd 18S rRNAs. The bars (1–4) at the right indicate the positions ofGTase transcripts. (B) Northern blotting of beta actin transcript.

C) The intensity of bands is quantified by scanning densitometry.he OGTase mRNA band was normalized by the actin band. The

ntensity of each band is plotted as the percentage of the intensity ofach corresponding control band. h, 5.5 mM glucose (control); ■, 20M glucose (high glucose). Data represent the mean 6 standard

rror in three independent experiments. * P , 0.01 versus control;* P , 0.05 versus control.

copy (Figs. 1a–1d). Intense staining signals for OGTasewere observed in the nuclei of endothelial cells and

smooth muscle cells, whereas the cytoplasm of thesecells showed a weak staining pattern. To further ex-amine the localization of OGTase in aortic smoothmuscle cells, we also performed an immunohistochem-ical study using cultured RASM cells. OGTase waslocalized in both the nucleus and cytoplasm of culturedRASM cells (Figs. 1e and 1f). The nucleolus wasscarcely stained. The nuclear staining was more in-tense than the cytoplasmic staining, while the cyto-plasm showed a weak and diffuse staining pattern(Figs. 1e and 1f). In the control experiment, in whichprimary antibody (AL25) was omitted or replaced withthe preimmune rabbit IgG, no positive staining was

FIG. 5. Immunoblot analysis of OGTase in RASM cells. (A) Cellscultured in the presence of 5.5 mM (control) or 20 mM glucose (highglucose). The same amount of protein (10 mg) was loaded per track inysates of control and high glucose, electrophoresed on a 10% poly-crylamide gel, and immunoblotted with antibody (AL25) againstGTase. OGTase antibody shows two bands, one at 78 kDa (b-

subunit) and one at 110 kDa (a-subunit). (B) Quantification of bandintensity by scanning densitometry. In the high glucose, the inten-sity of p78 was significantly increased over the control value. Theintensity of each band (p110 and p78) is expressed as the meanpercentage of the control 6 standard error of the band intensity

ratio. h, 5.5 mM glucose (control); ■, 20 mM glucose (high glucose).* P , 0.05 versus control. Experiments were performed in triplicate.

v

msbtaRsms

itltafioc

b3nnsNa

tGT

mrcnk

mgbigapN61gtsaa

cpynsTRa(5bs(t

gicrrTpho

sihtOd

172 AKIMOTO ET AL.

observed (data not shown). We then examined immu-nohistochemically in more detail the OGTase expres-sion in the vascular smooth muscle cells (in vivo and initro) of experimentally diabetic animals.In situ hybridization. The localization of OGTaseRNA in sections of rat aorta was examined by in

itu hybridization histochemistry using the DIG-la-eled antisense RNA probe (Fig. 2). The hybridiza-ion signals were observed in the endothelial cellsnd smooth muscle cells (Fig. 2a). When the senseNA probe was used, no positive signals were ob-erved (Fig. 2b). These results indicate that OGTaseRNA is expressed in both endothelial cells and

mooth muscle cells.Immunoelectron microscopic localization of OGTase

n aortic smooth muscle cells. The precise location ofhe OGTase was visualized at the electron microscopicevel. In the smooth muscle cells of the tunica media ofhe aorta, the cell nuclei, especially the euchromatin,nd the cytoplasm in the immediate vicinity of myo-laments were labeled (Fig. 3a). The label was scarcelybserved in the smooth muscle cells in the cytochemi-al control (Fig. 3b).

As described previously, many nuclear proteins haveeen demonstrated to be modified with O-GlcNAc (34,5). Some of them have been identified, for example,uclear pore proteins, chromatin-associated proteins,uclear oncogene and tumor suppressor proteins, andteroid receptors (35). We have proposed that O-Glc-Acylation of nuclear protein is involved in variousspects of gene expression (35).Cytoskeletal proteins, such as MAPs, neurofilaments

au, ankyrin G, talin, vinculin, and synapsin are O-

FIG. 6. OGTase activity (DPM/mg protein) in RASM cells. RASMcells were cultured in the presence of 5.5 mM glucose (control, h) or20 mM glucose (high glucose, ■) for 5 days. Reaction time is 15 min(left) and 30 min (right). * P , 0.01 versus control. Data representthe mean 6 standard error (n 5 3).

lcNAcylated and also phosphorylated (12, 36–39).hus, we think that OGTase may be involved in the

ie

odulation of the contraction of smooth muscle cells byegulating the O-GlcNAcylation/O-phosphorylation ofytoskeletal proteins in conjunction with specific ki-ases, such as calmodulin kinase or myosin light-chaininase.Effects of glucose on the level of expression of OGTaseRNA in RASM cells. To evaluate the effect of a high

lucose level on the OGTase expression, we analyzedy Northern blotting the expression of OGTase mRNAn RASM cells cultured in the presence of 5.5 or 20 mMlucose (Fig. 4). Total RNA was isolated from the cells,nd a cDNA clone for the transferase was used as arobe to determine the size of the RASM cell mRNA.orthern blot analysis indicated four transcripts (8.0,.0 4.2, and 1.7 kb; Fig. 4A). The amount of 6.0, 4.2, and.7-kb transcripts significantly increased in the highlucose, whereas no significant change was observed inhe amount of the 8.0-kb transcript (Fig. 4C). In theame experiment, the expression level of beta-actin (asreference protein) mRNA was almost the same in 5.5nd 20 mM glucose (Fig. 4B).Immunoblot analysis of OGTase expression in RASM

ells. The effect of high glucose on the level of OGTaserotein expression was examined by immunoblot anal-sis of OGTase in RASM cells that were grown inormal medium (5.5 mM glucose) or in the mediumupplemented with 20 mM glucose for 5 days (Fig. 5).he same amount of cell lysate obtained from theASM cells under the two glucose conditions was ex-mined. AL-25 antibody recognized both the 78-kDabeta) and 110-kDa (alpha) subunits of OGTase (Fig.A). When the density of protein bands was quantifiedy scanning densitometry, the amount of the 78-kDaubunit was significantly increased in the high glucoseP , 0.05); but no significant increase in the amount ofhe 110-kDa subunit was observed (Fig. 5B).

OGTase assay in RASM cells. The effect of highlucose on the OGTase enzymatic activity was exam-ned in the RASM cells. The 30% ammonium sulfateytosolic pellets were assayed for enzyme activity toemove UDP-GlcNAc hydrolases which prevent accu-ate assay of OGTase. The results are shown in Fig. 6.he enzyme activity of the RASM cells cultured in theresence of 20 mM glucose was significantly (P , 0.01)igher than that in the 5.5 mM glucose after either 30r 60 min of assay time.In the present study, we found that OGTase expres-

ion and activity increased in the RASM cells culturedn the 20 mM glucose. At present, it is unclear whyyperglycemia elevates both the levels of OGTase andhe overall levels of O-GlcNAc. Perhaps the increased-GlcNAcylation of transcription factors, such as SP1,ue primarily to elevated UDP-GlcNAc pools, results

n selective transactivation of the OGTase gene? Anlevated glucose concentration in vascular cells was

173HYPERGLYCEMIA AND O-GlcNAc IN RASM CELLS

reported to cause an increase in the activity of proteinkinase C (PKC) (40–43), which would certainly alterthe expression of several genes. In the RASM cell es-pecially, protein kinase isoforms that bind to diacyl-glycerol are activated by high glucose (44). Vasculardysfunctions in diabetic rats are ameliorated by PKCinhibitor (45). Interestingly PKC a, b, and e are in-duced to translocate into the nucleus by high glucose(44), and PKC is known to modulate gene expression(46). Phosphorylation and O-GlcNAcylation are closelycorrelated with each other (12, 34, 47). A high glucoseconcentration also induces an activation of the Na1/H1

antiport and increases the NHE-1 mRNA level andosteopontin expression via a glucose-induced PKC-de-pendent mechanism (48, 49). Previously, it was shownthat core2 b1,6-N-acetylglucosamine structures are el-evated in hyperglycemic rat cardiac tissues (50), sug-gesting that the expression of other unrelated GlcNAcglycosyltransferases are also regulated by glucose.

FIG. 7. Two-dimensional gel analysis of high-glucose concentrationextract. RASM cells were cultured for 5 days in the presence of 5.O-GlcNAc-modified proteins were detected by the galactosyl tranUDP-[3H]galactose as described. The proteins were resolved by twofocusing gel electrophoresis, and the second, 10% SDS–acrylamide gfluorography of gels treated with 1 M sodium salicylate. Patterns artop and acidic proteins to the left. We compared the intensity of 96 mProteins that show marked changes are designated with arrowheadsblack arrowheads, spots that increased in intensity in the 20 mM glsimilar experiments.

High glucose and glucosamine also induced an increasein the expression of growth factors in RASM cells,

including TGF alpha (28), which resulted in an in-crease in O-GlcNAcylation on intracellular proteins asrevealed by Western blotting and indirect immunoflu-orescence using O-GlcNAc-specific monoclonal anti-body (RL2) (51). Our present data are consistent withthese data.

Exogenous glucosamine, which greatly elevatesUDP-GlcNAc pools in cells, causes insulin resistance inisolated adipocytes by impairing insulin-induced glu-cose transporter (GLUT4) translocation to the plasmamembrane (1, 7). In skeletal muscle glucosamine alsocauses insulin resistance in vivo by affecting GLUT4translocation (2). In vivo glucosamine infusion inducesinsulin resistance in normoglycaemic rat (3). Moreover,in the muscle of insulin-resistant ob/ob mice the hex-osamine synthesis pathway showed increased activity(6). On the other hand, glycogen synthase and proteinphophatase 1 are downregulated by high glucose inRat-1 fibroblasts, and this is regulated by the hex-

duced changes in O-GlcNAc-modified proteins in RASM cell nuclearr 20 mM glucose. Nuclear proteins were extracted and terminallyrase probe method using Gal b (1–4) galactosyltransferase andmensional gel electrophoresis, the first dimension being isoelectriclectrophoresis. The [3H]galactose-tagged proteins were detected by

hown in the standard orientation, with high molecular mass at ther spots (white figures) between 5.5 mM glucose and 20 mM glucose.ite arrowheads, spots that decreased in intensity in 20 mM glucose;se relative to 5.5 mM glucose. The figure is representative of three

-in5 osfe-diel ee sajo

: whuco

osamine biosynthesis pathway (52). The insulin resis-tance induced by glucosamine and uridine is mediated

1

1

1

1

1

1

1

1

11

2

2

2

2

2

22

2

2

2

174 AKIMOTO ET AL.

by increased accumulation of muscle UDP-N-acetyl-hexosamines, probably via altered glycosylation of pro-teins in GLUT4-containing vesicles (53).

Two-dimensional gel analysis of high glucose-in-duced changes in O-GlcNAc-modified RASM cell nu-clear proteins. The striking localization of OGTasewithin nuclei in the RASM cells and its apparent re-distribution in response to high glucose are suggestiveof a role for O-GlcNAc metabolism in the cell function.Most of the O-GlcNAcylated proteins are found in thenucleus, such as RNA polymerase II, transcription fac-tors and p62 (54). So we examined the effect of highglucose on the O-GlcNAcylation of nuclear proteins.O-GlcNAc-bearing nuclear proteins were probed usingthe well-established galactosyltransferase probemethod and were analyzed by two-dimensional gelelectrophoresis. As seen in Fig. 7, the levels of O-GlcNAc on many proteins changed strikingly. We com-pared the intensity of 96 major spots between 5.5 mMglucose and 20 mM glucose. The level of O-GlcNAc in atleast 26 proteins decreased in 20 mM glucose, whereasthat in at least 34 proteins increased in the high glu-cose. This glucose-dependent alteration in O-GlcNAcmodified proteins may be due to elevated UDP-GlcNAcpools elevating glycosylation due to glucose-inducedsignaling events that impinge on OGTase regulation.Alternatively, the high-glucose induced changes mayalso be, in part, due to altered O-glycosylation chang-ing the turnover rates of these proteins. Recently, hy-perglycemia was found to induce activation of nucleartranscription factor kappa B (NF-kB) in vascularsmooth muscle cells (55), and the authors suggestedthat hyperglycemia-induced activation of NF-kB invascular smooth muscle cells may be a key mechanismfor the accelerated vascular diseases observed in dia-betes. From the amino acid sequence data (56), NF-kBwould be expected to be O-GlcNAcylated. Recent, pre-liminary data, indeed show that NF-kB is O-GlcNAcy-lated (Vossler and Hart, unpublished). Systematicidentification of O-GlcNAc modified nuclear proteins,including NF-kB, is currently underway using pro-teomic analyses.

ACKNOWLEDGMENTS

We thank Ms. Betty Jean Hart for peptide synthesis. Dr. RichardMarchase, Dr. Jeffrey E Kudlow, and all the members of the HartLab for helpful discussions. This work was supported in part by theJuvenile Diabetes Foundation and Fifty 50 Foods Inc., by NIH GrantHD13563 awarded to G.W.H. and by a gift from Oxford GlycosystemsInc. We also express our appreciation to Mr. M. Fukuda, Ms. S.Matsubara, Ms. M. Kanai, and Ms. S. Shibata (Laboratory for Elec-tron Microscopy and Department of Anatomy, Kyorin UniversitySchool of Medicine) for their technical assistance. This work wassupported in part by grants-in-aid for scientific research from the

Ministry of Education, Science, Sports, and Culture of Japan and thePromotion and Mutual Aid Corporation for Private Schools of Japan.

3

REFERENCES

1. Marshall, S., Bacote, V., and Traxinger, R. R. (1991) J. Biol.Chem. 266, 4706–4712.

2. Baron, A. D., Zhu, J.-S., Zhu, J.-H., Weldon, H., Maianu, L., andGarvey, W. T. (1995) J. Clin. Invest. 96, 2792–2801.

3. Rossetti, L., Hawkins, M., Chen, W., Gindi, J., and Barzilai, N.(1995) J. Clin. Invest. 96, 132–140.

4. Robinson, K. A., Weinstein, M. L., Lindenmayer, G. E., and Buse,M. G. (1995) Diabetes 44, 1438–1446.

5. McClain, D. A., and Crook, E. D. (1996) Diabetes 45, 1003–1009.6. Buse, M. G., Robinson, K. A., Gettys, T. W., McMahon, E. G., and

Gulve, E. A. (1997) Am. J. Physiol. 272, E1080–E1088.7. Hawkins, M., Barzilai, N., Liu, R., Hu, M., Chen, W., and Ros-

setti, L. (1997) J. Clin. Invest. 99, 2173–2182.8. Crook, E. D., Daniels, M. C., Smith, T. M., and McClain, D. A.

(1993) Diabetes 42, 1289–1296.9. Hebert, L. F., Daniels, M. C., Zhou, J., Crook, E. D., Turner,

R. L., Simmons, S. T., Neidigh, J. L., Zhu, J.-S., Baron, A. D., andMcClain, D. A. (1996) J. Clin. Invest. 98, 930–936.

0. Cooksey, R. C., Hebert, L. F., Zhu, J.-H., Wofford, P., Garvey,W. T., and McClain, D. A. (1999) Endocrinology 140, 1151–1157.

1. Torres, C. R., and Hart, G. W. (1984) J. Biol. Chem. 259, 3308–3317.

2. Comer, F. I., and Hart, G. W. (2000) J. Biol. Chem. 275, 29179–29182.

3. Haltiwanger, R. S., Blomberg, M. A., and Hart, G. W. (1992)J. Biol. Chem. 267, 9005–9013.

4. Kreppel, L. K., Blomberg, M. A., and Hart, G. W. (1997) J. Biol.Chem. 272, 9308–9315.

5. Kreppel, L. K., and Hart, G. W. (1999) J. Biol. Chem. 274,32015–32022.

6. Lubas, W. A., Frank, D. W., Krause, M., and Hanover, J. A.(1997) J. Biol. Chem. 272, 9316–9324.

7. Paulson, J. C., and Colley, K. J. (1989) J. Biol. Chem. 264,17615–17618.

8. Joziasse, D. H. (1992) Glycobiology 2, 271–277.9. Shafi, R., Iyer, S. P., Ellies, L. G., O’Donnell, N., Marek, K. W.,

Chui, D., Hart, G. W., and Marth, J. D. (2000) Proc. Natl. Acad.Sci. USA 97, 5735–5739.

0. Akimoto, Y., Kreppel, L. K., Hirano, H., and Hart, G. W. (1999)Diabetes 48, 2407–2413.

1. Akimoto, Y., Kreppel, L. K., Hirano, H., and Hart, G. W. (2000)Diabetologia 43, 1239–1247.

2. Konrad, R. J., Janowski, K. M., and Kudlow, J. E. (2000) Bio-chem. Biophys. Res. Commun. 267, 26–32.

3. Liu, K., Paterson, A. J., Chin, E., and Kudlow, J. E. (2000) Proc.Natl. Acad. Sci. USA 97, 2820–2825.

4. Han, I., Oh, E.-S., and Kudlow, J. E. (2000) Biochem. J. 350,109–114.

5. Lorenzi, M. (1992) Diabetes Metab. Rev. 8, 85–103.6. King, G. L., Kunisaki, M., Nishio, Y., Inoguchi, T., Shiba, T., and

Xia, P. (1996) Diabetes 45(Suppl. 3), S105–S108.7. Nerlich, A. G., Sauer, U., Kolm-Litty, V., Wagner, E., Koch, M.,

and Schleicher, E. D. (1998) Diabetes 47, 170–178.8. McClain, D. A., Paterson, A. J., Roos, M. D., Wei, X., and Kudlow,

J. E. (1992) Proc. Natl. Acad. Sci. USA 89, 8150–8154.9. Daniels, M. C., Kansal, P., Smith, T. M., Paterson, A. J., Kudlow,

J., and McClain, D. A. (1993) Mol. Endocrinol. 7, 1041–1048.

0. Patti, M. E., Virkamaki, A., Landaker, E. J., Kahn, C. R., and

Yki-Jarvinen, H. (1999) Diabetes 48, 1562–1571.

3

3

3

33

4

4

4

44

4

4

5

5

55

55

5

175HYPERGLYCEMIA AND O-GlcNAc IN RASM CELLS

31. Gunther, S., Alexander, R. W., Atkinson, W. J., and Gimbrone,M. A., Jr. (1982) J. Cell Biol. 92, 289–98.

2. Braissant, O., Foufelle, F., Scotto, C., Dauca, M., and Wahli, W.(1996) Endocrinology 137, 354–366.

3. Roquemore, E. P., Chou, T. Y., and Hart, G. W. (1994) MethodsEnzymol. 230, 443–460.

34. Hart, G. W. (1997) Annu. Rev. Biochem. 66, 315–335.35. Comer, F. I., and Hart, G. W. (1999) Biochim. Biophys. Acta

1473, 161–171.6. Arnold, C. S., Johnson, V. W., Cole, R. N., Dong, D. L.-Y., Lee, M.,

and Hart, G. W. (1996) J. Biol. Chem. 271, 28741–28744.37. Ding, M., and Vandre, D. D. (1996) J. Biol. Chem. 271, 12555–

12561.8. Cole, K. N., and Hart, G. W. (1999) J. Neurochem. 73, 418–429.9. Arnold, C. S., and Hart, G. W. (1999) Trends Glycosci. Glycotech.

11, 355–370.0. Ayo, S. H., Radnik, R., Garoni, J. A., Troyer, D. A., and Kreis-

berg, J. I. (1991) Am. J. Physiol. 261, F571–F577.1. Wolf, B. A., Willamson, J. R., Easom, R. A., Chang, K., Sherman,

W. R., and Turk, J. (1991) J. Clin. Invest. 87, 31–38.2. Inoguchi, T., Xia, P., Kunisaki, M., Higashi, S., Feener, E. P., and

King, G. L. (1994) Am. J. Physiol. 267, E369–379.3. Williams, B., and Schrier, R. W. (1992) Diabetes 41, 1464–1472.4. Haller, H., Baur, E., Quass, P., Behrend, M., Lindschau, C.,

Distler, A., and Luft, F. C. (1995) Kidney Int. 47, 1057–1067.

5. Ishii, H., Jirousek, M. R., Koya, D., Takagi, C., Xia, P., Clermont,

A., Bursell, S. E., Kern, T. S., Ballas, L. M., Heath, W. F.,

Stramm, L. E., Feener, E. P., and King, G. L. (1996) Science 272,728–231.

46. Nishizuka, Y. (1988) Nature 334, 661–665.47. Hart, G. H., Kreppel, L. K., Comer, F. I., Arnold, C. S., Snow,

D. M., Ye, Z., Cheng, X., DellaManna, D., Caine, D. S., Earles,B. J., Akimoto, Y., Cole, R. N., and Hayes, B. K. (1996) Glycobi-ology 6, 711–716.

48. Williams, B., and Howard, R. L. (1994) J. Clin. Invest. 93, 2623–2631.

9. Takemoto, M., Yokote, K., Yamazaki, M., Ridall, A. L., Butler,W. T., Matsumoto, T., Tamura, K., Saito, Y., and Mori, S. (1999)Biochem. Biophys. Res. Commun. 19, 722–726.

0. Nishio, Y., Warren, C. E., Buczek-Thomas, J. A., Rulfs, J., Koya,D., Aiello, L. P., Feener, E. P., Miller, T. B., Dennis, J. W., andKing, G. L. (1995) J. Clin. Invest. 4, 1759–1767.

1. Sayeski, P. P., and Kudlow, J. E. (1996) J. Biol. Chem. 271,15237–15243.

2. Crook, E. D., and McClain, D. A. (1996) Diabetes 45, 322–327.3. Hawkins, M., Angelov, I., Liu, R., Barzilai, N., and Rossetti, L.

(1997) J. Biol. Chem. 272, 4889–4895.4. Kelly, W. G., and Hart, G. W. (1989) Cell 57, 243–251.5. Yerneni, K. K., Bai, W., Khan, B. V., Medford, R. M., and Nat-

arajan, R. (1999) Diabetes 48, 855–864.6. Meyer, R., Hatada, E. N., Hohmann, H.-P., Haiker, M., Bartsch,

C., Roethlisberger, U., Lahm, H.-W., Schlaeger, E. J., van Loon,

A. P., and Scheidereit, C. (1991) Proc. Natl. Acad. Sci. USA 88,966–970.