Phosphorylation of Purified InsulinReceptor by cAMP KinaseRICHARD A. ROTH AND JACQUELINE BEAUDOIN

SUMMARYHighly purified insulin receptor was shown to be asubstrate for cAMP kinase. Approximately 1 phosphatewas incorporated per molecule of receptor, and thecAMP kinase's affinity for the receptor was at least ashigh as its affinity for histone. The sites phosphoryl-ated by cAMP kinase seemed distinct from thosephosphorylated by the protein kinase C. Phosphoryla-tion by cAMP kinase had no effect on the ability ofseveral monoclonal antibodies to recognize the recep-tor or on the insulin-binding activity of the receptor.However, cAMP phosphorylation partially inhibited thetyrosine kinase activity of the receptor (-25%). Theseresults suggest that catecholamine-induced resistanceto insulin may be partly due to a direct phosphoryla-tion of the receptor by cAMP kinase and a subsequentinhibition of the ability of the receptor kinase to be ac-tivated by insulin. Diabetes 36:123-26, 1987

Recent data indicate that the insulin receptor hasan intrinsic tyrosine-specific protein kinase activ-ity that is important in mediating responses to in-sulin (1-3). Thus, the regulation of the receptor

kinase would serve to modulate the ability of insulin to stim-ulate its different biologic responses. Two studies haveshown that elevation of intracellular levels of cAMP causeda decrease in the ability of insulin to stimulate the receptor'sintrinsic tyrosine kinase activity (4,5). The decrease in re-ceptor kinase activity would explain the ability of catechol-amines to induce insulin resistance in isolated adipocytes(6,7) and in vivo (8).

Because cAMP can activate a serine-specific kinase (9),one explanation for the results described would be that thecAMP-dependent kinase directly phosphorylates the insulin

From the Department of Pharmacology, Stanford University School of Medi-cine, Stanford, California 94305-5332.Send correspondence and reprint requests to Dr. Richard A. Roth at the aboveaddress.Received for publication 17 September 1986 and accepted 10 October 1986.

receptor and regulates its kinase activity. This hypothesis issupported by the finding that elevation of intracellular cAMPlevels increases the amount of serine phosphorylation of thereceptor in intact cells (4). However, the cAMP kinase mightalso phosphorylate another serine-specific kinase that in turnphosphorylates the receptor. Indirect evidence for this al-ternate hypothesis has come from studies suggesting thatthe isolated insulin receptor is not phosphorylated by purifiedcAMP kinase (10,11). Our studies were therefore undertakento further investigate whether cAMP kinase would phosphor-ylate the insulin receptor in vitro. We found the highly purifiedreceptor to be a substrate for the cAMP kinase. This phos-phorylation did not affect the insulin-binding activity of thereceptor but did inhibit the intrinsic kinase activity of thereceptor.

MATERIALS AND METHODSMaterials. Insulin receptor was purified from human pla-centa on a monoclonal antibody affinity column and a wheatgerm lectin column and exhibited only two protein bands(the a- and p-subunits of the receptor) on SDS-PAGE (12).The catalytic subunit of cAMP kinase (9) and C-kinase (13)were purified by procedures from bovine heart and rat brain,respectively. Histone type IIA, poly(Glu4,Tyr), protein kinaseinhibitor, and Staphylococcus aureus V8 protease werepurchased from Sigma (St. Louis, MO). [32P-7]ATP (3000Ci/mmol) was purchased from Amersham (ArlingtonHeights, IL).Methods. Phosphorylation reactions were performed by in-cubation of the insulin receptor with cAMP kinase for 10-60min at 24°C in a final volume of 25 jxl of buffer containing150 mM NaCI, 50 mM Hepes, pH 7.6, 0.02% Triton X-100,2 mM MgCI2, 0.2 mg/ml bovine serum albumin, and 10 |xM[32p_7]ATP (10-30 Ci/mmol). Reaction mixtures were thenanalyzed by SDS-gel electrophoresis and autoradiography(12). To test the phosphorylated receptor for its insulin-bind-ing and kinase activities, it was adsorbed to microtiter wellscoated with anti-mouse immunoglobulin and monoclonalanti-receptor antibody. Parallel wells were then incubated

DIABETES, VOL. 36, JANUARY 1987 123

INSULIN RECEPTOR PHOSPHORYLATION BY cAMP KINASE

a b c d e f g h

95 - ^^^^m^^^m** 4

Ba b c

95 -

a b c d e f g h i j

95 -

Histone —

FIG. 1. Phosphorylation of insulin receptor bycAMP kinase. A: either untreated (a-d) or heat-treated (60 min at 37°C; e-h) insulin receptor wasincubated with either 20 (a, e), 8 (b, f), 4 (c, g), or0 (d, h) ng/ml of cAMP kinase in presence of 2mM MgCI2 and 10 (AM ATP (10 mCi/jtmol). After 1h at 24°C, reaction mixtures were analyzed on7.5% SDS-PAGE; autoradiograph of 95,000-Mr

portion of gel is shown. Number of cpm incorpo-rated into receptor were a, 2054; b, 1097; c, 439;d, 350; e, 909; f, 438; g, 195; and h, 95. B: effectof cAMP kinase inhibitor on receptor phosphoryl-ation. cAMP kinase was incubated with either 5(a), 1.5 {b), or 0 (c) fig of inhibitor before incuba-tion with insulin receptor, MgCI2, and [32P]ATP. C:comparison of receptor and histone phosphoryla-tion by cAMP kinase. cAMP kinase (10 i^g/ml)was incubated with either 20 (a), 10 (b), 5 (c), or2.5 (d) nM heat-treated insulin receptor or 2500(e), 1200 (f), 600 (g), or 300 (h) nM histone HA.Controls with either insulin receptor alone (/) orcAMP kinase alone (/) were also performed. Re-action and analysis was performed as in A.Amounts of radioactivity incorporated into recep-tor (a-d) and histone (e-h) bands were a, 3031;b, 1591; c, 781; d, 513; e, 11,649; f, 2260; g, 1823;and h, 1147 cpm.

either with 125l-labeled insulin in the presence of various con-centrations of unlabeled insulin (to assess the receptor'sinsulin-binding ability) or with poly(Glu.Tyr) in the presenceof 5 mM MnCI2 and 10 jxM[32P-7]ATP (to assess the re-ceptor's kinase activity) and analyzed as described (14).

RESULTSPhosphorylation of the insulin receptor by cAMP kinase.Incubation of purified insulin receptor with various concen-trations of purified cAMP kinase in the presence of Mg2+

and [32P]ATP resulted in a dose-dependent phosphorylationof the receptor's p-subunit {Mr = 95,000) (Fig. '{A). With20 fxg/ml cAMP kinase, 1.2 molecules of phosphate wereincorporated per molecule of receptor. In the absence ofcAMP kinase, there was still some phosphorylation of the95,000-Mr band because of the intrinsic kinase activity ofthe receptor (the autophosphorylating activity of the re-ceptor is low under these conditions; i.e., no insulin or MnCI2is present).

To test whether the increased phosphorylation observedin the presence of cAMP kinase could be due to an activationof the receptor kinase, several experiments were performed.First, because the receptor kinase is heat labile, the receptorwas preincubated for 1 h at 37°C and then incubated withcAMP kinase. With this heat-treated receptor, phosphoryl-ation of the 95,000-/V/r band was still observed in the pres-ence of the cAMP kinase, but the amount of phosphorylationin its absence was decreased (Fig. 1A bands e-h). Second,antibodies that inhibit the receptor kinase (2) were includedin the reaction mixture. Again, the 95,000-Mr band was phos-phorylated in the presence but not the absence of the cAMP

kinase. Third, receptor phosphorylated in the presence ofcAMP kinase was tested for its ability to be recognized byantibodies to phosphotyrosine. Although these antibodiesreadily recognize autophosphorylated receptor, they did notrecognize receptor phosphorylated in the presence of cAMPkinase. Finally, an inhibitor of cAMP kinase (15) was foundto block the phosphorylation of the receptor by the cAMPkinase preparation (Fig. 18). The results indicated that thereceptor was being phosphorylated by the cAMP kinase.

To obtain significant levels of receptor phosphorylation, aconcentration of at least 3 |xg/ml cAMP kinase was required.For comparison, the cAMP kinase was examined for its abilityto phosphorylate various concentrations of histone IIA, aknown substrate for this kinase (9). With 1200 nM histone inthe reaction, an almost equal amount of counts was incor-porated into the histone band as into receptor when 20 nMreceptor was present in the reaction (Fig. 1C).Ability of various monoclonal antibodies to recognizecAMP-phosphorylated receptor. Because autophosphor-ylation of the receptor affects the ability of certain monoclonalantibodies to bind to the receptor (16), we investigated

TABLE 1Effect of cAMP kinase phosphorylation on receptor kinase

cAMP kinase added(jig/ml)

Poly(Glu,Tyr) phosphorylation Inhibition(cpm) (%)

40102.50

4091 ± 4055029 ± 4776482 ± 7526484 ± 778

37220

124 DIABETES, VOL. 36, JANUARY 1987

a b c d e f g h i j

- 93- 66

- 45

- 31

- 22

- 14

FIG. 2. Protease mapping of C-kinase and cAMP kinase phosphoryl-ated receptor. Insulin receptor was phosphorylated by either cAMP(a-e) or C-kinase (f-j), electrophoresed on 7.5% SDS-PAGE, and phos-phorylated p-subunit bands were eluted from gel. p-Subunits werethen digested with either 10 (a,f), 1 (b,g), 0.1 (c,h), 0.01 (d,i), or 0 (e,y)(jig of Staphylococcus aureus V8 protease and electrophoresed on 15%SDS-PAGE. Autoradiograph of gel is shown.

whether a similar effect was observed with cAMP phosphor-ylation of the receptor. However, receptor phosphorylatedby cAMP kinase was found to be recognized identically tocontrol receptor by seven different monoclonal antibodiesthat recognize four distinct antigenic regions of the cyto-plasmic domain of the receptor p-subunit (16).Effect of cAMP phosphorylation on insulin-binding andtyrosine kinase activities of receptor. Receptor was phos-phorylated by cAMP kinase and tested for its ability to bind125l-insulin and phosphorylate an exogenous substrate,poly(Glu.Tyr). Phosphorylated receptor was found to bindinsulin as well as nonphosphorylated receptor. In contrast,the receptor's intrinsic kinase activity was reduced dose de-pendently by prior incubation with cAMP kinase (Table 1).The maximum inhibition observed in five separate experi-ments was 37%, and the average inhibition was 24%.Comparison of sites phosphorylated by C-kinase andcAMP kinase. Because purified insulin receptor is alsophosphorylated by C-kinase (14), we investigated whethercAMP kinase was phosphorylating the receptor at the samesites. A comparison of the Staphylococcus aureus V8 digestsof receptor phosphorylated by these two enzymes indicatesthat the enzymes phosphorylate distinct sites on the receptorp-subunit (Fig. 2).

DISCUSSIONOur studies demonstrate that incubation of highly purifiedinsulin receptor with purified cAMP kinase results in an in-creased phosphorylation of the p-subunit of the insulin re-ceptor. This increased phosphorylation probably results fromthe cAMP kinase phosphorylating the receptor, because 7)phosphorylation was not inhibited by inactivating the recep-tor kinase by either heat treatment or inclusion of monoclonal

antibodies to the receptor kinase; 2) this phosphorylationwas not on a tyrosine residue, unlike receptor autophos-phorylation (1); 3) the phosphorylation had no effect on theability of several monoclonal antibodies to recognize thereceptor, although these antibodies do not recognize theautophosphorylated receptor (16); and 4) a specific inhibitorof the cAMP kinase blocked the phosphorylation (15).

The phosphorylation of the receptor by excess cAMP ki-nase resulted in ~1 molecule of phosphate incorporated permolecule of receptor. The Vmax and Km for the phosphorylationcould not be determined because of the limited amount ofreceptor available. However, the receptor was a better sub-strate for the cAMP kinase than histone 11 A, a known sub-strate for this enzyme (9). The inability of several others tofind any phosphorylation of the receptor by cAMP kinase(10,11) may be due to 7) the use of low concentrations ofimpure receptor that would favor the intramolecular auto-phosphorylation reaction over the bimolecular cAMP kinasephosphorylation of the receptor, 2) the presence of highconcentrations of detergent in these other studies, or 3) theuse of receptor that has a partially proteolyzed p-subunit.

The phosphorylation of the receptor by cAMP kinase hadno effect on the insulin-binding activity of the receptor butpartly inhibited its intrinsic kinase activity. The level of thisinhibition was —25%. This extent of inhibition may be suffi-cient to account for the decreases in the biologic responsesto insulin observed in catecholamine-treated adipocytes.However, a greater inhibition of the receptor kinase wasobserved in receptors isolated from cells that have had el-evated levels of cAMP (4,5). This may be due to a greaterlevel of phosphorylation of the receptor in vivo by cAMPkinase. Numerous attempts at obtaining a greater inhibitionof the receptor kinase by varying the conditions of the invitro phosphorylation were not successful. Alternatively,there may be another protein in vivo that preferentially rec-ognizes receptor phosphorylated by cAMP kinase and in-hibits the receptor's tyrosine kinase activity. Such a proteinhas been described for phosphorylated rhodopsin (17). Fi-nally, another kinase may be activated in catecholamine-treated cells, and this kinase could also phosphorylate thereceptor and modify its tyrosine kinase activity. The phos-phorylation sites recognized by this hypothetical kinase mayresemble those recognized by C-kinase, another serine-spe-cific kinase that phosphorylates the receptor and has agreater inhibitory effect on the receptor kinase (14).

ACKNOWLEDGMENTSWe thank H. Schulman, C. Csernansky, and J. Lloyd forgenerous gifts of protein kinase C and cAMP kinase, J. Wangfor anti-phosphotyrosine antibody, and Karen Bird for prep-aration of the manuscript.

This work was supported by Biomedical Research SupportGrant RR-5353, National Institutes of Health Grant AM-34962, and Research Career Development Award AM-01393.

REFERENCES1. Kahn CR: The molecular mechanisms of insulin action. Annu Rev Med

36:429-51, 19852. Morgan DO, Ho L, Korn LJ, Roth RA: Insulin action is blocked by a

monoclonal antibody that inhibits the insulin receptor kinase. Proc NatlAcad Sci USA 83:328-32, 1986

DIABETES, VOL. 36, JANUARY 1987 125

INSULIN RECEPTOR PHOSPHORYLATION BY cAMP KINASE

3. Ellis L, Clauser E, Morgan DO, Edery M, Roth RA, RutterWJ: Replacementof insulin receptor tyrosine residues 1162 and 1163 compromises insulin-stimulated kinase activity and uptake of 2-deoxyglucose. Cell 45:721-32, 1986

4. Stadtmauer L, Rosen OM: Increasing the cAMP content of IM-9 cellsalters the phosphorylation state and protein kinase activity of the insulinreceptor. J Biol Chem 261:3402-407, 1986

5. Haring H, Kirsch D, Obermaier B, Ermel B, Machicao F: Decreased ty-rosine kinase activity of insulin receptor isolated from rat adipocytes ren-dered insulin-resistant by catecholamine treatment in vitro. Biochem J234:59-66, 1986

6. Kirsch D, Kemmler W, Haring HU: Cyclic AMP modulates insulin bind-ing and induces post-receptor insulin resistance of glucose transportin isolated rat adipocytes. Biochem Biophys Res Commun 115:398,1983

7. Pessin JE, Gitomer W, Oka Y, Oppenheimer CL, Czech MP: p-Adrenergicregulation of insulin and epidermal growth factor receptors in rat adi-pocytes. J Biol Chem 258:7386-94, 1983

8. Deibert C, DeFronzo RA: Epinephrine-induced insulin resistance in man.J Clin Invest 65:717-21, 1980

9. Krebs EG, Beavo JA: Phosphorylation-dephosphorylation of enzymes:review. Annu Rev Biochem 48:923-59, 1979

10. Obberghen EV, Rossi B, Kowalski A, Gazzano H, Ponzio G: Receptor-mediated phosphorylation of the hepatic insulin receptor: evidence that

the M, 95,000 receptor subunit is its own kinase. Proc Natl Acad Sci USA80:945-49, 1983

11. Joost HG, Steinfelder HJ, Schmitz-Salue C: Tyrosine kinase activity ofinsulin receptors from human placenta. Biochem J 233:677-81, 1986

12. Roth RA, Morgan DO, Beaudoin J, Sara V: Purification and characteri-zation of the human brain insulin receptor. J Biol Chem 261:3753-57,1986

13. Gould KL, Woodgett JR, Cooper JA, Buss JE, Shalloway D, Hunter T:Protein kinase C phosphorylates pp60src at a novel site. Cell 42:849-57,1985

14. Bollag GE, Roth RA, Beaudoin J, Mochly-Rosen D, Koshland DE Jr: Pro-tein kinase C directly phosphorylates the insulin receptor in vitro andreduces its protein-tyrosine kinase activity. Proc Natl Acad Sci USA83:5822-24, 1986

15. Ashby CD, Walsh DA: Characterization of the interaction of a proteininhibitor with adenosine 3',5'-monophosphate-dependent protein kin-ases. J Biol Chem 247:6637-42, 1972

16. Morgan DO, Roth RA: Mapping surface structures of the human insulinreceptor with monoclonal antibodies: localization of main immunogenicregions to the receptor kinase domain. Biochemistry 25:1364-71, 1986

17. Wilden U, Hall SW, Kuhn H: Phosphodiesterase activation by photoexcitedrhodopsin is quenched when rhodopsin is phosphorylated and binds theintrinsic 48-kDa protein of rod outer segments. Proc Natl Acad Sci USA83:1174-78, 1986

126 DIABETES, VOL. 36, JANUARY 1987