Proc. Nati. Acad. Sci. USAVol. 86, pp. 8368-8372, November 1989Biochemistry

Isoproterenol stimulates phosphorylation of the insulin-regulatableglucose transporter in rat adipocytes

(insulin receptor/insulin action/catecholamines/cAMP-dependent protein kinase)

DAVID E. JAMES*, JEFFREY HIKENt, AND JOHN C. LAWRENCE, JR.tDepartments of *Cell Biology and Physiology and tPharmacology, Washington University School of Medicine, 660 South Euclid Avenue,Saint Louis, MO 63110

Communicated by Stuart Kornfeld, August 10, 1989

ABSTRACT We have examined the acute effects of insulinand isoproterenol on the phosphorylation state of the insulin-regulatable glucose transporter (IRGT) in rat adipocytes. TheIRGT was immunoprecipitated from either detergent-solubilized whole-cell homogenates or subcellular fractions of32P-labeled fat cells and subjected to sodium dodecylsulfate/polyacrylamide gel electrophoresis. The 32P-labeledIRGT was detected by autoradiography as a species ofapparentMr 46,000. Insulin stimulated translocation of the IRGT fromlow-density microsomes to the plasma membrane but did notaffect phosphorylation of the transporter in either fraction.Isoproterenol inhibited insulin-stimulated glucose transport by40% but was without effect on the subcellular distribution ofthe transporter in either the presence or absence of insulin.Isoproterenol stimulated phosphorylation of the IRGT 2-fold.Incubating cells with dibutyryl-cAMP and 8-bromo-cAMP alsostimulated phosphorylation 2-fold, and the transporter wasphosphorylated in vitro when IRGT-enriched vesicles wereincubated with cAMP-dependent protein kinase and [Y-32P]ATP. These results suggest that isoproterenol stimulatesphosphorylation of the IRGT via a cAMP-dependent pathwayand that phosphorylation of the transporter may modulate itsability to transport glucose.

Glucose transport is the first step in the complex pathway ofintracellular glucose utilization, and it is rate limiting forglucose uptake in muscle and fat (reviewed in refs. 1 and 2).Furthermore, it is a focal point for acute regulation ofglucosemetabolism by hormones such as insulin and epinephrine andother stimuli. Insulin has been shown to stimulate the move-ment of glucose transporters from an intracellular domain tothe cell surface, and this appears to be a major mechanism foractivation of glucose transport (3-8). However, this may notbe the only mechanism for acutely regulating glucose trans-port. /3-Adrenergic agonists, such as epinephrine or isopro-terenol, inhibit insulin-stimulated glucose transport in bothmuscle and adipose tissue (reviewed in refs. 9-12). That suchinhibition was not accompanied by a corresponding change inthe subcellular distribution of the glucose transporter, asmeasured by cytochalasin B binding (4), led to the conclusionthat isoproterenol modifies the intrinsic ability of the carrierprotein to transport glucose (10). In previous experiments inwhich an anti-human erythrocyte glucose transporter anti-body was used to perform immunoprecipitations from 32p-labeled rat (13) and 3T3-L1 (14) adipocytes, phosphorylationof the transporter was not detected either in control cells (13,14) or in cells incubated with insulin (13, 14) or isoproterenol(13). Consequently, it has been concluded that the inhibitionof transport in response to isoproterenol does not involvephosphorylation of the transporter (13).

Molecular cloning has led recently to the identification ofa number of glucose transporter species among varioustissues (15-18). The human erythrocyte glucose transporterwas first cloned from Hep G2 cells (17). This transporter isexpressed in many tissues including muscle, fat, and centralnervous tissue (16, 19). However, the Hep G2-type glucosetransporter is not the major insulin-sensitive glucose trans-porter in muscle and adipose tissue (20). These cells expressa unique glucose transporter [insulin-regulatable glucosetransporter (IRGT)] that undergoes insulin-dependent trans-location to the cell surface (15, 20). The amino acid sequencesof the IRGT and Hep G2-type glucose transporter are 65%identical (17). The most striking differences between the twotransporters occur within presumed intracellular domains. Inthe case of the IRGT sequence, these regions contain severalpotential phosphorylation sites that are not present in theHep G2-type glucose transporter (17). In the present study,we have investigated the effects of insulin and isoproterenolon the phosphorylation and subcellular distribution of IRGT.

MATERIALS AND METHODSIsolation and Incubation of Adipocytes. Adipocytes were

prepared by incubating epididymal fat pads of 190 to 210-gSprague-Dawley rats with crude collagenase (type I, lot46J057, Cooper Biomedical) by the method of Rodbell (21).Cells were incubated essentially as described (22). The in-cubation medium contained 135 mM NaCl, 5.4 mM KCl, 1.4mM MgSO4, 1.4 mM CaC12, 0.18 mM sodium phosphate, 3%(wt/vol) bovine serum albumin (fraction V, lot 58F0581,Sigma), and 10 mM Hepes (pH 7.4). Cells were suspended (5ml of medium per g of original adipose tissue) in medium andincubated at 37°C for 2 hr with 32p; (0.1 mCi/mi; 1 Ci = 37GBq). This time is sufficient to achieve steady-state labelingof [y-32P]ATP (22). Cells were then incubated with insulin andother agents for 1-30 min before homogenization. Underthese conditions, neither insulin nor isoproterenol affects thespecific activity of [y-32P]ATP (22).

Preparation of Cells for Immunoprecipitation. When as-sessing the effects of agents on total transporter phospho-rylation, cells were homogenized in buffer A, which con-tained 100 mM NaF, 2 mM sodium pyrophosphate, 10 mMEDTA, 2 mM EGTA, and 50 mM Tris HCl (pH 7.8 at 23°C).Total membranes were pelleted by centrifugation at 150,000x g for 90 min at 4°C. Pellets were solubilized in 100 ,ul ofbuffer A containing 1% sodium dodecyl sulfate (SDS) andfrozen at -70°C before immunoprecipitation. Buffer A con-tains inhibitors of both phosphatases and kinases. To becertain that the inhibitors blocked posthomogenizationchanges in the phosphorylation state of the IRGT, someexperiments were performed in which cells were homoge-

Abbreviations: IRGT, insulin-regulatable glucose transporter; LDM,low-density microsomes; HDM, high-density microsomes; PM,plasma membrane; mAb, monoclonal antibody.

8368

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 86 (1989) 8369

nized in buffer A containing 1% SDS. Inclusion of SDSensured rapid denaturation of proteins and complete extrac-tion of the glucose transporter. The same amounts of 32P-labeled transporter were recovered by using both methods.When subcellular fractionation was performed, adipocytes

were rinsed once at 370C in incubation medium preparedwithout CaC12, MgSO4, or albumin (hormones and otherconstituents were present where appropriate). The cells werethen homogenized in buffer B, which contained 25 mM NaF,1 mM sodium pyrophosphate, 1 mM EDTA, 0.5 mM EGTA,250 mM sucrose, and 1 mM ammonium molybdate (pH 7.4).Sodium vanadate (1 mM) was included in some experiments.Plasma membranes (PM), high-density microsomes (HDM),low-density microsomes (LDM), and mitochondria/nucleiwere prepared essentially as described (23). Membrane frac-tions were suspended in buffer B and stored overnight at-700C.Immunoprecipitation. The protein contents of membrane

samples were measured by using bicinchoninic acid (24) andadjusted to the same value. SDS-solubilized membranes (100sul; 100-200 Ag of protein) were added to 900 Al of buffer Bcontaining 1% Triton X-100 and 100mM NaCl. Samples wereincubated for 30 min at 22°C and then centrifuged for 5 minat 13,000 x g. Supernatants were retained for immunoprecip-itation with a monoclonal antibody (mAb 1F8) previouslyshown to be specific for the IRGT (15). The mAb was firstimmobilized by incubation at 220C with goat anti-mouse IgGcoupled to Sepharose beads (Cappel Laboratories; 1 ,ug ofsecondary antibody per 1 ,ug of mAb 1F8) suspended in bufferB containing 2% bovine serum albumin. After 60 min thebeads were pelleted by centrifugation and washed three timeswith buffer B containing 1% Triton X-100. Beads were thenadded to solubilized membrane samples (typically 125 ,1 ofbeads per mg of membrane protein). After incubation for 60min at 22°C, the samples were centrifuged at 13,000 x g for10 s to pellet the beads containing the immune complexes. Nodetectable IRGT (assayed by immunoblotting; see below)remained in the supernatant. The samples were washed fourtimes with buffer B containing 1% Triton X-100 (1 ml perwash) before SDS sample buffer (50 ,l) was added to eluteIRGT (25).

Electrophoretic Analysis. Samples were subjected to SDS/PAGE (7.5% polyacrylamide resolving gel) by the method ofLaemmli (25). In some experiments, 32P-labeled sampleswere transferred to nitrocellulose sheets (Schleicher &Schuell). Autoradiograms were prepared by exposing nitro-cellulose sheets or dried gels to preflashed X-Omat AR film(Kodak) and Cronex Lightning Plus enhancing screens at-700C. Relative amounts of 32P-labeled IRGT were deter-mined by optical density scanning of the autoradiograms.Direct measurement of the 32p content of immunoprecipi-tated IRGT was accomplished by scintillation counting of gelslices containing the 32P-labeled protein.A rabbit polyclonal antibody (R820) specific for a 12-amino

acid peptide based on the deduced carboxyl-terminal se-quence of the IRGT (15) was used to detect IRGT byimmunoblotting. The antiserum (20 gg of IgG per ml) wasprepared in phosphate-buffered saline (PBS) containing 1%Triton X-100 and 1% powdered milk (pH 7.4; Carnation) for1 hr at 220C. The sheets were washed three times (10 mineach) with PBS containing 1% Triton X-100 and then wereincubated with 1251-labeled protein A (2 ,uCi/ml; Amersham).The sheets were then washed four times (10 min each) withPBS containing Triton X-100 and dried.The amount of 1251 present in the IRGT band under these

conditions was large relative to the 32p present. Therefore,32P did not interfere with quantitation of 1251 by opticaldensity scanning of autoradiograms. This was confirmed bymeasuring 1251 emissions in slices of nitrocellulose (contain-ing the IRGT) with a y counter. The use of R820 for

immunoblotting instead of mAb 1F8 eliminated a problem inassessing the amount ofIRGT. 1251-labeled protein A does notbind to mAb 1F8 under these conditions. Thus, the heavychain ofmAb 1F8, which migrates close to the IRGT in 7.5%acrylamide gels, did not interfere with the detection orquantitation of 1251 associated with IRGT.To determine whether the amount of detected 1251 accu-

rately reflected the amount of IRGT immunoprecipitated,adipocytes were labeled with [35S]methionine (26) in thepresence and absence of insulin. PM, LDM, HDM, andmitochondria/nuclei fractions were isolated, and immuno-precipitations were performed with mAb 1F8. Samples of theimmunoprecipitated IRGT were then divided equally andsubjected to SDS/PAGE. One gel was impregnated with 1 Msodium salicylate to enable detection of the 35S-labeledtransporter (27); the other was used for immunoblotting withR820. Based on optical density scanning of autoradiograms,the ratios of 35S/125I remained constant among the subcellularfractions obtained from insulin- and noninsulin-treated cells.These results indicate that immunoblotting provides an ac-curate measure of the relative amount of IRGT.

RESULTSPhosphorylation of IRGT in Rat Adipocytes. To investigate

the phosphorylation of the IRGT, fat cells were incubated inmedium supplemented with 32p;. After homogenizing thecells, membrane fractions were prepared by high-speed cen-trifugation, and immunoprecipitations were performed withmAb 1F8. A phosphorylated protein with an apparent mo-lecular weight of 46,000 was detected by autoradiographyafter SDS/PAGE (see, for example, Fig. la). The mobility ofthe phosphorylated species corresponded exactly with that ofthe IRGT detected by immunoblotting with R820 (Fig. lb). Inother experiments (not shown here), we found that thisphosphorylated species could be immunoprecipitated withthe R820 antisera but not with a nonimmune IgG or anantibody against the Hep G2-type glucose transporter. Theseresults strongly suggest that the Mr 46,000 species is theIRGT and that under basal conditions it is phosphorylated.

It has not been possible to directly measure the number ofIRGT molecules immunoprecipitated. However, by usingmeasurements of glucose-inhibitable cytochalasin B bindingsites as an estimate of transporter number (4, 28) and assum-ing that the specific activity of 32p in the transporter was equalto that of the intracellular pool of [y-32P]ATP, we havecalculated that there is at least 0.2 mol of phosphate per molof IRGT. It should be emphasized that this is an approxima-tion and could represent an underestimate of the true stoi-chiometry. Irrespective of this, a significant proportion of theintracellular pool of IRGT appears to be phosphorylated,suggesting that phosphorylation of this protein may be ofphysiological relevance.

Effect of Insulin on Phosphorylation of IRGT. Insulin stim-ulates the translocation of the IRGT from the LDM fractionof the cell to the PM (15). Therefore, it seemed possible thatinsulin affected phosphorylation of IRGT within a specificcompartment of the cell. To investigate this possibility,membrane fractionation was performed after incubating 32p_labeled cells in the absence or presence of insulin. Thespecific activity of transporters in the PM of insulin-treatedcells was no different from that in the LDM of nonstimulatedcells (Fig. ic). Whereas there was no difference in the specificactivity of the IRGT between LDM from nonstimulated cellsand PM from insulin-treated cells, the specific activity of theIRGT in the PM from nonstimulated cells was higher. Be-cause of the low amount of transporters in this fractionobtained from nonstimulated cells, we cannot rule out thepossibility that phosphorylation of this small subset (i.e.,5-10% of the total; see Fig. lb) of transporters was affected

Biochemistry: James et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

M/N HDM PM LDM

a

58 10485 1

365 o

b58 0485 0

cmE

0

I-

CM

ah__sa

365 P

INSULIN

C

-

FIG. 1. Effect of insulin on phosphorylation of the IRGT. Ratadipocytes were incubated with 32p; (0.1 mCi/ml) for 2 hr. Cells wereincubated for a further 20 min in the absence (lanes -) or presence(lanes +) of insulin (200 milliunits/liter and then homogenized.Subcellular fractionation was performed to obtain mitochondria/nuclei (M/N), HDM, LDM, and PM. The 32P-labeled IRGT wasimmunoprecipitated from each fraction by using mAb 1F8. Sampleswere split into equal parts, and these were subjected to SDS/PAGEwith a 7.5% resolving gel. (a) Autoradiogram of a dried gel. (b)Detection of IRGT (after transfer of proteins to nitrocellulose) byusing the anti-peptide antibody R820 and 125I-labeled protein A. (c)Specific activity of IRGT (means ± SEM) from seven experiments.Immunoprecipitated transporter was subjected to SDS/PAGE andtransferred to nitrocellulose. An autoradiogram of the nitrocellulosewas prepared to quantitate [32P]IRGT. The nitrocellulose sheetswere then incubated with R820 and 1251-labeled protein A to estimatethe relative amounts of transporter. Autoradiograms of the 32P- and1251-labeled species were scanned for optical density, and areasbeneath the peaks corresponding to the IRGT were determined.Specific activities in arbitrary units were calculated by dividing thevalues for 32p peaks by those of 1251 peaks. To compare results fromdifferent experiments, the specific activity of the transporter in theLDM of control cells was assigned a value of 1.

by insulin. In other experiments we investigated the effectsof incubating cells with insulin for different times (1-30 min)and with increasing concentrations (2.5-2500 milliunits/liter)of the hormone. To be sure that phosphorylation/dephos-phorylation was inhibited, cells were homogenized in buffercontaining SDS before immunoprecipitation. In agreementwith the subcellular fractionation studies (Fig. 1c), no effectof insulin on IRGT phosphorylation was detected.

Stimulation of IRGT Phosphorylation by Adrenergic Ago-nists and cAMP Derivatives. In contrast to insulin, /3-

adrenergic agonists were found to increase phosphorylationof IRGT. In the experiments shown in Fig. 2, cells wereincubated for 10 min with different adrenergic agonists. The/3-adrenergic agonist isoproterenol stimulated the phospho-rylation of IRGT -'2-fold. Epinephrine, which binds to both

100.

75 -

50-

25 -

0

CON ISO MET EPI

FIG. 2. Stimulation of IRGT phosphorylation by f-adrenergicagonists. Rat adipocytes were incubated with 32P; and 2.5 ,Ag ofadenosine deaminase per ml for 2 hr. Incubations were continued for10 min in the absence [control (CON)] or presence of 1 A±M isopro-terenol (ISO), 100 /uM methoxamine (MET), or 10 uM epinephrine(EPI). Adipocytes were homogenized in buffer A, and samples werecentrifuged at 150,000 x g for 90 min. The pellets, which containedtotal cellular IRGT, were solubilized with SDS, and IRGT wasimmunoprecipitated with mAb 1F8. Immunoblotting indicated thatunder these conditions the same amount of transporter per unit ofprotein was present in each treatment group. After SDS/PAGE,autoradiograms were prepared by exposing gels to film for 24 hr at-70°C. Autoradiograms were scanned for optical density, and thepeak area corresponding to the IRGT band was measured. Radio-activity in each sample was calculated by using peak areas ofradioactive standards. The results represent the cpm of [32P]IRGTper g ofadipose tissue and are the means ± SEM of five experiments.*, Not significantly different from the control by Dunnett's t test (29),but P < 0.01 compared with the control as determined by Student'st test; **, P < 0.01 compared with the control as judged by analysisof variance and Dunnett's t test for comparing multiple values to asingle control (29).

a- and ,B-adrenergic receptors, also stimulated phosphoryla-tion of IRGT, although its effect was slightly less than that ofisoproterenol. Methoxamine, an a-adrenergic receptor ago-nist, was without effect on IRGT phosphorylation (Fig. 2).These data indicate that IRGT phosphorylation may bestimulated via ,3-adrenergic receptor activation.

Results from subcellular fractionation experiments indi-cated that isoproterenol stimulated the phosphorylation ofIRGT in both the LDM and PM fractions. In the experimentspresented in Fig. 3, cells were first incubated with insulin for10 min, a time sufficient to stimulate translocation of trans-porters to the PM. Incubations were then continued for anadditional 5 min in the absence or presence of isoproterenol.Isoproterenol stimulated phosphorylation of IRGT 2-foldboth in the LDM and in the PM (Fig. 3a). Isoproterenol-dependent phosphorylation of IRGT was significantly en-hanced by including adenosine deaminase (2.5 ,ug/ml) in theincubation medium.Many of the effects of A3-adrenergic agonists are mediated

by increased intracellular cAMP (30, 31). To investigatefurther the possible role of cAMP on increasing IRGT phos-phorylation, cells were incubated with the cAMP derivatives8-bromo-cAMP and dibutyryl-cAMP. Both cAMP deriva-tives increased IRGT phosphorylation -2-fold (Fig. 4). In

contrast, 8-bromoadenosine monophosphate and butyratewere without effect (Fig. 4).cAMP-dependent protein kinase mediates many of the

actions of cAMP in eukaryotic cells (30, 31). Hence, itseemed possible that the IRGT was a substrate for cAMP-dependent kinase. To test this hypothesis, IRGT-enrichedLDM were incubated with [y-32P]ATP in the absence orpresence of the catalytic subunit of cAMP-dependent proteinkinase. The catalytic subunit catalyzed the phosphorylationof the IRGT (Fig. 3a). These results suggest that IRGT is a

32p / 125ICONTROL INSULIN

LDM 1.0 0.8±0.1

PM 2.0±0.3 0.9+r0.2. __

8370 Biochemistry: James et al.

Proc. Natl. Acad. Sci. USA 86 (/989) 8371

a58 0

485 P

36 5 l

1 2 3 4 5 6 7 8

b58 0485 0

365 P

1 2 3 4 5 6 7 8

FIG. 3. Effect of isoproterenol and adenosine deaminase onIRGT phosphorylation in insulin-treated cells. 32P-labeled cells wereincubated with insulin (200 milliunit/liter) for 15 min followed by a5-min incubation with no other additions (lanes 3 and 6), withisoproterenol (lanes 4 and 7), or isoproterenol and adenosine deam-inase (2.5 jLg/ml) (lanes 5 and 8). Cells were then homogenized, andsubcellular fractions were prepared. LDM (lanes 3-5) or PM (lanes6-8) fractions (100 ,ug of protein each) were solubilized, and IRGTwas immunoprecipitated with mAb 1F8. Samples were subjected toSDS/PAGE. (a) Autoradiogram of a dried gel showing [32P]IRGT(lanes 3-8). (b) Autoradiogram of the corresponding region of animmunoblot prepared with the anti-peptide antiserum R820 and1251-labeled protein A. In b, an aliquot of rat adipocyte microsomalmembrane was electrophoresed and immunoblotted as a standard(lane 1). In a, samples of IRGT-enriched LDM vesicles (5 ,ug)prepared from unlabeled cells as described (23) were incubated in 10mM Hepes (final volume, 100 ,l; pH 7.4) containing 1 mM EDTA,0.1 mM EGTA, 250 mM sucrose, 6 mM MgCl2, and 0.2 mM[y-32P]ATP (3000 cpm/pmol) for 1 hr in the absence or presence ofthe catalytic subunit (0.4 ,M) of cAMP-dependent protein kinasefrom beef heart. After 1 hr the reaction was terminated by adding 100,ul of 200 mM NaF and 10 mM EDTA. Immunoprecipitation wasperformed with mAb 1F8, and samples were subject to SDS/PAGE.An autoradiogram (12-hr exposure) of the dried gel was prepared,and the region surrounding the IRGT phosphorylated in the absence(lane 2) or presence (lane 1) of the catalytic subunit is shown.

direct substrate for cAMP-dependent kinase but do notexclude possible indirect actions of the kinase.

Effect of Isoproterenol on Subcellular Distribution of IRGT.One mechanism by which isoproterenol-dependent phos-phorylation of IRGT might result in inhibition of insulin-stimulated cellular glucose transport is by inhibiting translo-cation of transporters to the cell surface. To address thisissue, cells were incubated with hormones before subcellularfractionation. The relative amount of IRGT in each fractionwas measured after immunoprecipitation by immunoblottingwith R820. In the absence of insulin, the IRGT was locatedpredominantly in the LDM, and this distribution was unaf-fected when cells were incubated with 1 ,uM isoproterenol for5 min (Fig. 5). Insulin stimulated the translocation of IRGTto the PM. Isoproterenol did not affect the amount of IRGTdetected in the PM after insulin treatment.

DISCUSSIONInsulin-dependent glucose transport in muscle and fat is mostlikely mediated by the IRGT because it appears to be thepredominant transporter in these tissues, and it undergoesinsulin-dependent translocation to the plasma membrane (15,

-

CM7I--

CE,

125 -

100 -

75-

50-

25 -

0-CON 8BrcA 8BrA dbcA but

FIG. 4. Increased phosphorylation of IRGT by cAMP deriva-tives. Adipocytes were incubated with 32P for 2 hr followed by a10-min incubation with no addition (CON), 5 mM 8-bromo-cAMP(8BrcA), 5 mM 8-bromo-AMP (8BrA), 5 mM dibutyryl-cAMP(dbcA), or 5 mM butyrate (but). The IRGT was immunoprecipitatedafter homogenization and solubilization of the total membrane frac-tions as described in Fig. 2. Means + SEM for five experiments areshown. **, P < 0.01 compared with control (29).

20). The present results show that the IRGT is phosphoryl-ated in fat cells. By using 32P-labeled epitrochlearis muscle(results not shown), we have found that the IRGT is alsophosphorylated in skeletal muscle.

Insulin is perhaps the most important regulator of glucosetransport (1, 2). It is now evident that a major mechanism bywhich insulin stimulates glucose transport involves translo-cation of glucose transporters from inside the cell to the PM(3-7, 20). Protein phosphorylation has been implicated ininsulin action (32, 33). Thus, it was of interest to examinewhether the regulation of glucose transport by insulin in-volved phosphorylation of IRGT. The results presented heresuggest that the ability of insulin to stimulate translocation ofthe glucose transporter from the LDM fraction to the PMdoes not involve phosphorylation of the transporter per se.However, an important role for phosphorylation in the acti-vation of glucose transport by insulin cannot be excluded. Infact, the recent observation that glucose transport is in-creased by incubating adipocytes with okadaic acid, an

c1*

0=i.,

E-

INS

*S- A LDM

0 T

o5

ISO - +

+ - + - +

+ - - + +

FIG. 5. Failure of isoproterenol to affect insulin-dependent trans-location of the IRGT. Adipocytes were incubated in low phosphatemedium containing 2.5 ,ug of adenosine deaminase per ml for 2 hr.Incubations were continued as follows: 20 min without additions, 20min with insulin (INS; 200 milliunits/liter), 15 min without additionsfollowed by 5 min with 1 ,uM isoproterenol (ISO), or 15 min withinsulin and a further 5 min with insulin plus isoproterenol. LDM andPM were prepared. The IRGT was immunoprecipitated from eachfraction by using mAb 1F8, and samples were subjected to SDS/PAGE. Proteins were transferred to nitrocellulose sheets, whichwere incubated with the anti-peptide antiserum R820 and 125I-labeledprotein A. Autoradiograms were prepared, and the relative amountsof IRGT were determined by comparing peak areas obtained fromoptical density scanning. Each value was normalized to the amountof IRGT in LDM from control cells, which was assigned a value of1.0. Means ± SEM of five experiments are shown.

** **

nfl FT]

B PM

Biochemistry: James et al.

Lo -

Proc. Natl. Acad. Sci. USA 86 (1989)

inhibitor of type I and Ila phosphatases, suggests that proteinphosphorylation might be linked to activation of transport(34). This could potentially involve phosphorylation of otherproteins in, or associated with, the intracellular vesiclescontaining the glucose transporter.

Isoproterenol has been shown to inhibit glucose transportin both muscle and adipose tissue (9, 10, 12, 13). Our resultsindicate that these effects might be mediated via phospho-rylation of IRGT. The magnitude of the isoproterenol-mediated increase in IRGT phosphorylation (2-fold, Figs. 2and 3) was within the range of its inhibitory effect on cellularglucose transport (data not shown; refs. 9 and 10). Further-more, adenosine appears to be a major determinant of themagnitude of the effect of isoproterenol on inhibiting glucosetransport and stimulating phosphorylation of IRGT (Fig. 3a).This nucleoside accumulates in the incubation mediumthroughout the course ofthe incubation and inhibits the effectof isoproterenol on glucose transport. It has been shown (10)that the effect of isoproterenol on inhibiting transport can berestored by adding adenosine deaminase. At the relativelyhigh cell concentrations used in the present studies, the effectof isoproterenol on stimulating IRGT phosphorylation wassimilarly sensitive to adenosine deaminase (Fig. 3a). Inparticular, adenosine deaminase appeared to be required todemonstrate effects of isoproterenol on IRGT phosphoryla-tion in the PM.

Isoproterenol did not significantly alter the subcellulardistribution of the IRGT in either the absence or presence ofinsulin (Fig. 5). These results agree with previous findingsthat the concentration of PM glucose transporters, as mea-sured by cytochalasin B binding, was unaffected by isopro-terenol (10). Under the conditions of the experiments de-scribed in Fig. 5, insulin was found to stimulate 2-deoxyglucose uptake an average of 23-fold, whereas in thepresence of isoproterenol, the insulin effect was reduced-50%o. Based on these findings we suggest that stimulation ofIRGT phosphorylation by isoproterenol may decrease theintrinsic transport rate of the glucose transporter.

Phosphorylation ofIRGT was also increased by incubatingadipocytes with derivatives ofcAMP (Fig. 4). Together withthe ability of cAMP-dependent kinase to phosphorylateIRGT in vitro (Fig. 3a), these data strongly suggest that theeffects of isoproterenol on stimulation of IRGT phosphoryl-ation are mediated by cAMP-dependent protein kinase. Thesurrounding sequence of amino acids is important in deter-mining whether a particular serine/threonine will be phos-phorylated by this kinase. The consensus sequence forcAMP-dependent protein kinase is (Arg or Lys)-Xaa-(Ser orThr) (31). While future studies will be required to identifysites of phosphorylation in the IRGT, it is interesting that theputative cytoplasmic domains of the IRGT contain severalconsensus sites for cAMP-dependent protein kinase (forexample, Ser-243, Thr-351, and Ser-497). The fact that thesesites are not present in the Hep G2-type glucose transportermight explain the inability to detect phosphorylation of theHep G2-type glucose transporter with isoproterenol (13).

We thank Kerri James for artwork. Support for this work wasprovided by grants from the National Institutes of Health (AR34815)to J.C.L., the Juvenile Diabetes Foundation to D.E.J., and theWashington University Diabetes Research and Training Center.

1. Morgan, H. E. & Whitfield, C. F. (1973) Curr. Top. Membr.Transp. 4, 255-303.

2. Elbrink, J. & Bihler, I. (1975) Science 188, 1177-1184.3. Suzuki, K. & Kono, T. (1980) Proc. Natl. Acad. Sci. USA 77,

2542-2545.4. Cushman, S. W. & Wardzala, L. J. (1980) J. Biol. Chem. 255,

4758-4762.5. Lienhard, G. E., Kim, H. H., Ransome, K. J. & Gorga, J. C.

(1982) Biochem. Biophys. Res. Commun. 105, 1150-1156.6. Oka, Y. & Czech, M. P. (1984) J. Biol. Chem. 259, 8125-8133.7. Blok, J., Gibbs, E. M., Lienhard, G. E., Slot, J. W. & Geuze,

H. J. (1988) J. Cell Biol. 106, 69-76.8. Simpson, l. A. & Cushman, S. W. (1986) Annu. Rev. Biochem.

55, 1059-1089.9. Smith, U., Kuroda, M. & Simpson, l. A. (1984) J. Biol. Chem.

259, 8758-8763.10. Kuroda, M., Honnor, R. C., Cushman, S. W., Londos, C. &

Simpson, I. A. (1987) J. Biol. Chem. 262, 245-253.11. Kashiwagi, A., Huecksteadt, P. & Foley, J. E. (1983) J. Biol.

Chem. 258, 13685-13692.12. James, D. E., Burleigh, K. M. & Kraegen, E. W. (1986) J. Biol.

Chem. 261, 6366-6374.13. Joost, H. G., Weber, T. M., Cushman, S. W. & Simpson, I. A.

(1987) J. Biol. Chem. 262, 11261-11267.14. Gibbs, E. M., Allard, W. J. & Lienhard, G. E. (1986) J. Biol.

Chem. 261, 16597-16603.15. James, D. E., Strube, M. & Mueckler, M. (1989) Nature

(London) 338, 83-87.16. Birnbaum, M. J., Haspel, H. C. & Rosen, 0. M. (1986) Proc.

Natl. Acad. Sci, USA 83, 5784-5788.17. Mueckler, M., Caruso, C., Baldwin, S. A., Panico, M., Blench,

I., Morris, H. R., Allard, W. J., Lienhard, G. E. & Lodish,H. F. (1985) Science 229, 941-945.

18. Thorens, B., Sarkar, H. K., Kaback, H. R. & Lodish, H. F.(1988) Cell 55, 281-290.

19. Flier, J. S., Mueckler, M., McCall, A. L. & Lodish, H. F.(1987) J. Clin. Invest. 79, 657-661.

20. James, D. E., Brown, R., Navarro, J. & Pilch, P. F. (1988)Nature (London) 333, 183-185.

21. Rodbell, M. (1964) J. Biol. Chem. 239, 375-380.22. Lawrence, J. C., Jr., & James, C. (1984) J. Biol. Chem. 259,

7975-7982.23. James, D. E., Lederman, L. & Pilch, P. F. (1987) J. Biol.

Chem. 262, 11817-11824.24. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K.,

Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke,N. M., Olson, B. J. & Klenk, D. C. (1985) Anal. Biochem. 150,76-85.

25. Laemmli, U. K. (1970) Nature (London) 227, 680-685.26. Lawrence, J. C., Jr., Hiken, J. F., Inkster, M., Scott, C. W. &

Mumby, M. C. (1986) Proc. Natl. Acad. Sci. USA 83, 3649-3653.

27. Chamberlain, J. P. (1979) Anal. Biochem. 98, 132-135.28. Joost, H. G., Weber, T. M. & Cushman, S. W. (1988) Bio-

chem. J. 249, 155-161.29. Dunnett, C. A. (1955) J. Am. Stat. Assoc. 50, 1096-1121.30. Cohen, P. (1985) Eur. J. Biochem. 151, 439-448.31. Krebs, E. G. & Beavo, J. (1978) Annu. Rev. Biochem. 48,

923-959.32. Avruch, J., Alexander, M. C., Palmer, J. L., Pierce, M. W.,

Nemenoff, R. A., Blackshear, P. J., Tipper, J. P. & Witters,L. A. (1982) Fed. Proc. Fed. Am. Soc. Exp. Biol. 41, 2629-2633.

33. Denton, R. M. (1986) Adv. Cyclic Nucleotide Protein Phos-phorylation Res. 20, 293-341.

34. Haystead, T. A. J., Sim, A. T. R., Carling, D., Honnor, R. C.,Tsukitani, Y., Cohen, P. & Hardie, D. G. (1989) Nature(London) 337, 78-81.

8372 Biochemistry: James et al.