Expression of the insulin receptor with a recombinant vaccinia virus. Biochemical evidence that the insulin receptor has intrinsic serine kinase activity.

We have previously reported the tight association of a serine kinase activity with the human insulin receptor (Lewis, R. E., Wu, G. P., MacDonald, R. G., and Czech, M. P. (1990) J. Biol. Chem. 265, 947-954). We tested the possibility that the associated serine kinase activity was intrinsic to the receptor catalytic domain. The ratio of phosphoserine to phosphotyrosine on insulin receptors phosphorylated in vitro was used as an index of the associated serine kinase activity. Phosphorylation and phosphoamino acid analysis of insulin proreceptors revealed associated serine kinase activity early in receptor synthesis. Insulin receptors were expressed in HeLa cells using a recombinant vaccinia virus. The ratio of phosphoserine to phosphotyrosine on insulin receptors expressed by the recombinant vaccinia virus was determined relative to endogenous insulin receptors in cells treated with alpha-amanitin to block host cell mRNA synthesis. alpha-Amanitin treatment had no effect on the ratio of phosphoserine to phosphotyrosine on insulin receptors expressed from the recombinant virus even though they were present in a 4000-fold excess above endogenous receptors. We conclude that the serine kinase activity associated with the insulin receptor is intrinsic to the receptor catalytic domain. Receptor-catalyzed autophosphorylation of serine may play an important role in modulating insulin receptor signaling.

Insulin-stimulated phosphorylation of the insulin receptor cytoplasmic domain plays a central role in the transmission and regulation of insulin receptor signal transduction. Insulin activation of the receptor tyrosine kinase results in a cascade of tyrosine autophosphorylation. Autophosphorylation of tyrosines 1158, 1162, and 1163 1 maintain receptor kinase activation in the absence of bound insulin (1,2). Phosphorylation of tyrosines 1328 and 1334 in the receptor carboxyl-terminal tail occurs upon activation of the insulin receptor tyrosine kinase (3)(4)(5). Serine phosphorylation of the insulin receptor cytoplasmic domain also occurs following insulin activation but subsequent to receptor tyrosine autophosphorylation (6,7). Phorbol ester or forskolin addition to cells has been reported to stimulate the serine phosphorylation of the insulin receptor ␤ sub-unit (8 -10). A decrease in insulin receptor tyrosine kinase activity coincides with the phosphorylation of insulin receptors by these agents (10,11). Protein kinase C (12) and cyclic AMPdependent protein kinase (13) are each capable of phosphorylating the insulin receptor in cell-free systems. Receptor phosphorylation catalyzed by each of these kinases results in a decrease in insulin-stimulated tyrosine kinase activity toward exogenous substrates (12,13). Additional studies with kinaseinactive insulin receptor mutants indicate that the receptor tyrosine kinase is important for the complete biological activity of insulin. Consequently, inhibition of receptor tyrosine kinase activity by serine phosphorylation may be an important mechanism for regulating receptor signaling in intact cells.
An insulin-sensitive serine kinase activity is associated with the insulin receptor in eluates from wheat germ-agglutinin affinity columns (14)(15)(16). This insulin-stimulated serine kinase activity remains tightly associated with highly purified insulin receptors eluted from insulin-agarose affinity columns (17,18). The tightly associated serine kinase activity is also capable of phosphorylating highly purified insulin receptors on serine and threonine sites within the cytoplasmic domain that are also phosphorylated in intact cells (17). The serine kinase activity tightly associated with the insulin receptor phosphorylates synthetic peptides identical to sites of insulin receptor serine and threonine phosphorylation in vivo. The conclusion that the insulin receptor activates the receptor-associated serine kinase is supported by the observation that peptide phosphorylation is enhanced by insulin addition to the affinity-purified receptor preparations (17).
The presence of serine kinase activity associated with insulin receptors immunoprecipitated from Sf9 cells infected with a recombinant baculovirus containing the human insulin receptor cDNA has been reported (6). Although these receptor preparations contain primarily unprocessed proreceptor that is unresponsive to insulin activation, they suggest the possibility that the insulin receptor may contain intrinsic serine/threonine as well as tyrosine kinase activity. Consistent with this possibility, the tyrosine kinase inhibitor (hydroxy-2-naphthalenylmethyl)phosphonic acid blocks the ability of insulin receptor preparations from baculovirus-infected Sf9 cells containing associated serine kinase activity from phosphorylating a synthetic peptide containing an insulin receptor serine phosphorylation site (6).
Biochemical analysis has provided evidence for the existence of several "dual specificity" kinases (19). The amino acid sequences of those kinases considered to be capable of phosphorylation on tyrosine, serine, and threonine most closely resemble the family of serine/threonine kinases. We tested the hypothesis that the insulin receptor kinase had the capability of phosphorylating its cytoplasmic domain on serine and tyrosine. We expressed the insulin receptor under control of the bacteriophage T7 promoter in HeLa cells that were also in-fected with a recombinant vaccinia virus that expresses T7 polymerase. Receptor-associated serine kinase activity can be detected in insulin proreceptors early in synthesis. Furthermore, we demonstrate that the serine kinase activity associated with insulin receptors expressed under control of the T7 promoter is not altered in receptor preparations isolated under conditions that block synthesis of host cell mRNA.

EXPERIMENTAL PROCEDURES
Materials-Antibodies CT-1 and 83-14 were gracious gifts from Ken Siddle (Cambridge, UK). Lentil lectin-agarose and wheat germ agglutinin-agarose were obtained from E-Y Labs, Inc. [ 32 P]phosphate and L-[ 35 S]methionine were purchased from DuPont NEN. [␥-32 P]ATP was prepared using a Gamma-Prep A kit (Promega Biotech) as described by the manufacturer. Goat affinity-purified antibody to mouse IgG was obtained from Cappel. Dulbecco's modified Eagle medium, penicillin, streptomycin and neomycin antibiotic mixture, fetal bovine serum, and restriction endonuclease SalI were purchased from Life Technologies, Inc. L-1-(Tosylamino)-2-phenylethyl chloromethyl ketone-treated trypsin was obtained from Worthington Biochemical Corporation; restriction endonuclease NcoI was from New England Biolabs, Inc., insulin was from Calbiochem, and ␣-amanitin and phosphoamino acid standards were from Sigma. Thin layer chromatography plates used for phosphoamino acid analysis were obtained from Machery-Nagel.
Cell Culture-HeLa and COS-1 cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum. All cells were incubated at 37°C in 5% CO 2 .
Construction of the pTM1hIR Expression Plasmid and Virus Infection/Plasmid Transfection of HeLa Cells-The cDNA for the human insulin receptor (20) was subcloned into the expression plasmid pTM-1 (21) at sites NcoI and SalI immediately downstream of the bacteriophage T7 promoter. The resultant vector was named pTM1hIR. A recombinant vaccinia virus (VTF7) that expresses T7 polymerase (22) was allowed to infect HeLa cells in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum for 1 h at a multiplicity of 10. The virus was removed and pTM1hIR was transfected into HeLa cells by a modified version of the calcium phosphate precipitation method (23). A precipitate was formed in a 1.02-ml volume of 1 ϫ HBS (140 mM NaCl, 25 mM HEPES, and 0.75 mM Na 2 PO 4 , pH 7.10) including 29 g of calf thymus DNA, 20 g of pTM1hIR, and 106 l of 2.5 M CaCl 2 . The precipitate was applied to the cells and incubated for 3 h at 37°C in 5% CO 2 . The precipitate was removed, and a solution of 15% glycerol in 1 ϫ HBS was applied for 2 min. The glycerol solution was removed, and the cells were washed two times in 1 ϫ HBS and fed again with Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum with or without 10 g/ml ␣-amanitin. The transfected cells were then incubated at 37°C in 5% CO 2 for 18 h.
Construction of Vaccinia Virus Recombinants-A recombinant virus encoding the human insulin receptor cDNA was created by infecting CV-1 cells with vaccinia virus (strain WR) and then treating with calcium phosphate-precipitated pTM1hIR DNA as described above. TK Ϫ recombinant virus was isolated by plaque assay on TK Ϫ cells in the presence of 5-bromo-2Ј-deoxyuridine and identified by DNA slot blot hybridization. Virus was plaque purified three additional times, and stocks of the recombinant virus were prepared in HeLa cells without selection.
Insulin Receptor Immunoprecipitation-After incubating with or without 100 nM insulin for 10 min at 37°C, the cultures were immediately frozen on liquid N 2 and solubilized in Buffer A (50 mM HEPES, pH 7.8, 10 mM EDTA, 30 mM sodium pyrophosphate, 10 mM sodium fluoride, 1 mM sodium vanadate, 20 g/ml aprotinin, 25 mM benzamidine, and 0.1 mM phenylmethylsulfonyl fluoride) containing 1% Triton X-100. Insoluble matter was removed by centrifugation for 10 min at 15,000 rpm in a microcentrifuge. Cell supernatants (1 ml) were added to 1 l of CT-1, a monoclonal antibody that recognizes the carboxyl terminus of the insulin receptor ␤ subunit (24) or insulin receptor antibody 83-14 (25), and 10 l of goat anti-mouse IgG agarose or protein G-Sepharose. The cell supernatants were incubated for 2 h at 4°C with mixing to allow adsorption of insulin receptor to the antibody⅐IgG-agarose complex.
[ 35 S]Methionine Labeling of Insulin Receptors-Cells were incubated in 1 ml of methionine-free Dulbecco's modified Eagle medium containing 25 mM HEPES and 5% calf serum at 37°C. After 1 h, 100 Ci of [ 35 S]methionine was added, and the incubation was continued for 30 min. The cells were incubated for an additional 4 h in complete media, and the cells were lysed in Buffer A containing 1% Triton X-100. The lysate was clarified, and insulin receptors were immunoprecipitated as described above.
Insulin Receptor in Vitro Kinase Assay-Immunoadsorbed insulin receptor was washed three times by centrifugation with Buffer A containing 1% Triton X-100 and once with 50 mM HEPES, pH 7.8. Insulin receptor immune complexes were phosphorylated at 22°C in 50-l reaction volumes containing 1 mM dithiothreitol, 10 mM MgCl 2 , 3 mM MnCl 2 , 5 M [␥-32 P]ATP (200 Ci/nmol, Amersham Corp.), and 38 mM Tris, pH 7.4. When insulin receptors were precipitated with antibody CT-1, insulin was added to a final concentration of 100 nM. Reactions were terminated by the addition of 50 l of electrophoresis sample buffer containing 100 mM dithiothreitol. The phosphorylated insulin receptor ␤ subunit was identified by autoradiography following SDSpolyacrylamide gel electrophoresis of the affinity-purified preparation on a 7% gel.
Phosphoamino Acid Analysis-Phosphoamino acid analysis was performed as described (17). The 32 P-labeled insulin receptor ␤ subunit was localized by autoradiography and excised. The labeled bands were counted for Cerenkov radiation and washed alternately in acetone and water three times for 10 min each time. Gel fragments were placed in a 1.5-ml Eppendorf tube with 0.3 ml of 0.25% ammonium bicarbonate, pH 8.2, containing 30 g/ml trypsin and digested at 37°C for 18 h. The tryptic eluate containing ␤ subunit phosphopeptides was removed and lyophilized to dryness in a clean 1.5-ml Eppendorf tube. Dried phosphopeptides were reconstituted in 0.2 ml of 6 M HCl. Partial hydrolysis of the phosphopeptides was performed at 110°C for 1 h, after which the hydrolysate was diluted 6-fold with water, lyophilized, reconstituted with 100 l of water, and lyophilized a second time. The entire contents of each tube were spotted on a 20 ϫ 20-cm thin layer cellulose plate in 5 l of 30% formic acid containing 1 mg/ml each of phosphothreonine, phosphoserine, and phosphotyrosine. Samples were separated in one dimension by electrophoresis at 1500 V for 25 min in pH 1.9 buffer (formic acid/acetic acid/water, 25:78:897). The plates were dried, and the samples were separated by electrophoresis at 1300 V for 20 min in pH 3.5 buffer (water/pyridine/acetic acid, 189:1:10) containing 0.5 mM EDTA. Separation in pH 3.5 buffer was performed at right angles to the initial direction of separation. The plates were dried, and phosphoamino acid standards were visualized with ninhydrin. Radiolabeled phosphoamino acids were identified by autoradiography. The relative amounts of phosphate incorporated into tyrosine or serine were determined on a Betagen Betascope and a Molecular Dynamics PhosphorImager.
High Pressure Liquid Chromatography Phosphopeptide Mapping-For phosphopeptide mapping, the phosphorylated insulin receptor was separated on an 8% SDS-polyacrylamide gel. The insulin receptor ␤ subunit was localized by autoradiography and digested twice with 100 g/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin in 0.25% ammonium bicarbonate. Phosphopeptides were separated on a Brownlee Aquapore OD300 column with a gradient of 50% acetonitrile in 0.1% trifluoroacetic acid at 0.25 ml/min. Fractions were collected at 30-s intervals and counted for Cerenkov radiation. Recovery of 32 P was 70 -84%. Two-dimensional Peptide Mapping-HPLC 2 phosphopeptide peaks were further resolved in two dimensions on cellulose thin layer plates. HPLC fractions were pooled, lyophilized, reconstituted in 5 l of pH 1.9 buffer, and spotted on thin layer plates. The tracking dye ⑀-2,4-dinitrophenyl lysine was spotted adjacent to radioactive samples. The samples were separated by electrophoresis at 1000 V for 35 min in pH 1.9 buffer. After electrophoresis each plate was allowed to dry thoroughly, and then chromatography was performed at a right angle to the direction of electrophoresis in n-butanol/pyridine/acetic acid/water (15:10:3:12) as described previously (17). After chromatography the plates were dried and exposed to x-ray film (XAR-5, Kodak) to localize 32 P-labeled peptides.
Lentil Lectin Chromatography-Infected/transfected HeLa cells were lysed and collected in Buffer A with 1% Triton X-100. The lysate was diluted 10-fold with Buffer A before applying it to a wheat germ agglutinin-agarose column equilibrated in Buffer A containing 0.1% Triton X-100. The samples were recycled once more through the column, and the final flow-through was saved for later application to a lentil lectinagarose column. Sample was eluted from the wheat germ agglutininagarose column with Buffer A containing 0.1% Triton X-100 and 0.3 M N-acetylglucosamine. Final flow-through from the wheat germ agglutinin-agarose column was applied to a lentil lectin-agarose column equilibrated in Buffer A containing 0.1% Triton X-100. The samples were recycled once more through the column, and the proteins were eluted in Buffer A containing 0.1% Triton X-100 and 0.3 M ␣-methyl-mannopy-ranoside. Insulin receptors within eluates from each lectin column were immunoprecipitated as described above.

RESULTS
The purpose of this study was to test whether insulin receptor-associated serine kinase activity is intrinsic to the receptor. To test this possibility, human insulin receptor was initially expressed in COS-1 cells. Extracts from transfected cells were partially purified on wheat germ agglutinin-agarose. The flowthrough fraction from the wheat germ agglutinin-agarose was applied to lentil lectin-agarose. Insulin receptors bound to the lectin affinity resins were eluted and then immunoprecipitated with the anti-insulin receptor antibody CT-1 (24). Insulin receptors were phosphorylated in the immune complex by incubation with [␥-32 P]ATP for 30 min and then analyzed by denaturing polyacrylamide gel electrophoresis (Fig. 1A). Insulin receptor antibodies precipitated two proteins from the wheat germ agglutinin-agarose eluate that could be phosphorylated in the immune complex. The autoradiogram in Fig. 1A demonstrates phosphoproteins of 200 and 97 kDa that represent the uncleaved insulin receptor precursor and the mature insulin receptor ␤ subunit, respectively.
Biosynthetic labeling studies demonstrated that the insulin receptor is initially synthesized in the endoplasmic reticulum as a high mannose proreceptor (190 -210 kDa) that undergoes subsequent maturation of its N-linked carbohydrate side chains and proteolytic processing to produce mature ␣ and ␤ subunits (26 -28). To isolate insulin receptor precursors in the endoplasmic reticulum containing high mannose carbohydrates, we incubated the flow-through from wheat germ agglutinin-agarose with lentil lectin-agarose. The lentil lectin-agarose eluate was immunoprecipitated with antibody CT-1 as described above. Phosphorylation of the immune complex and SDS-polyacrylamide gel electrophoresis revealed a single phosphoprotein of approximately 200 kDa (Fig. 1A). This phosphoprotein most likely represents high mannose insulin receptor ␣␤ monomer isolated from early in the synthetic pathway.
Phosphorylation of insulin proreceptors and mature insulin receptor ␤ subunits immunoprecipitated from wheat germ agglutinin-agarose revealed not only phosphotyrosine but also phosphoserine (Fig. 1B, left panels). Interestingly, serine phosphorylation was also detected in high mannose insulin receptor precursors (Fig. 1B, right panel). Trace amounts of phosphothreonine could also be detected in receptor precursors immunoprecipitated from wheat germ agglutinin-agarose or lentil lectin-agarose eluates (Fig. 1B). The phosphoserine content of receptor precursor and ␤ subunit immunoprecipitated from wheat germ agglutinin-agarose was 11-12% of the phosphotyrosine released by hydrolysis from these phosphoproteins (Table I). Similarly, phosphoserine detected in proreceptor from lentil lectin-agarose eluates was approximately 15% of detectable phosphotyrosine (Table I).
The data in Fig. 1 suggest that serine kinase activity becomes associated with the insulin receptor early in its biosynthetic pathway. The level of serine phosphorylation detected with the high mannose form of the insulin receptor precursor appears comparable to that co-precipitating with biosynthetically mature forms. These observations suggest that either a distinct serine kinase associates with insulin receptor precursors early in synthesis or the insulin receptor tyrosine kinase has the intrinsic ability to phosphorylate itself on serine residues. To distinguish between these two possibilities, we developed a system to express the insulin receptor in cells in which endogenous mRNA synthesis had been blocked for prolonged periods of time. We subcloned the insulin receptor cDNA into the transfer vector pTM-1 (21) to create the plasmid pTM1hIR. This construct contains a T7 promoter and 5Ј-untranslated sequences from encephalomyocarditis virus for efficient transcription and translation of uncapped insulin receptor mRNA. Insulin receptors were detected by immune complex kinase assay of insulin receptor immunoprecipitates from HeLa cells transfected with pTM1hIR and infected with a recombinant vaccinia virus (VTF7) encoding T7 polymerase (Fig. 2A, lane 5).
To determine the amount of endogenous insulin receptor contributed by HeLa cells, control infections and transfections were performed. Similar amounts of phosphorylated insulin receptor precursor and insulin receptor ␤ subunit were detected in the immune complex from HeLa cells infected with wild-type vaccinia virus (WR) and transfected with recombinant plasmid pTM1HIR (Fig. 2A, lane 1  plasmid pTM-1 ( Fig. 2A, lane 2). Insulin receptors translated from T7-generated transcripts could be distinguished from insulin receptors derived from endogenous HeLa mRNA by the addition of the RNA polymerase II inhibitor ␣-amanitin (29). Treatment of control cells with 10 g/ml ␣-amanitin for 18 h prior to lysis blocked endogenous insulin receptor synthesis ( Fig. 2A, lanes 3 and 4). ␣-Amanitin had no effect, however, on the level of insulin receptor expressed in cells infected with VTF7 and transfected with pTM1hIR ( Fig. 2A, lane 6). In multiple experiments insulin receptors expressed in VTF7-infected and pTM1hIR-transfected cells were 8 -140-fold more abundant than endogenous insulin receptors from control cells.
Phosphoamino acid analysis was performed on insulin receptors from control-treated HeLa cells and from HeLa cells infected with VTF7 and transfected with pTM1hIR. Quantitative analysis of the 32 P incorporated on to tyrosine and serine in the insulin receptor ␤ subunit demonstrated that treatment with ␣-amanitin has no effect on the relative amount of serine phosphorylation of insulin receptors expressed from T7 polymerasegenerated transcripts (Fig. 2B and Table I). Thus, serine kinase activity remained associated with the insulin receptors synthesized from T7 polymerase-generated transcripts even though endogenous mRNA synthesis was inhibited.
The efficiency of insulin receptor production was improved by the construction and co-infection of a recombinant vaccinia virus, VTF7hIR, that contained the human insulin receptor cDNA un-der control of the T7 promoter. Co-infection of HeLa cells with VTF7 and VTF7hIR followed by treatment with ␣-amanitin resulted in a 4000-fold increase in mature, phosphorylated insulin receptors over the level of receptor observed in ␣-amanitintreated cells infected with VTF7 and the control virus WR (Fig. 3,  A and B). Phosphoamino acid analysis of 32 P-labeled insulin receptors revealed that phosphoserine persisted in both the proreceptor and mature ␤ subunit produced in VTF7hIR-infected and ␣-amanitin-treated cells (Fig. 3C). Quantitative analysis of phosphoamino acids present in the 32 P-labeled insulin receptors produced in VTF7hIR-infected and ␣-amanitin-treated cells revealed phosphotyrosine:phosphothreonine ratios of 6.3 and 10.7 in the proreceptors and mature ␤ subunits, respectively. These ratios are comparable to those seen at lower levels of insulin receptor expression in COS-1 cells and in pTM1hIR-transfected HeLa cells (Fig. 4 and Table I).
Phosphopeptide mapping was performed to determine if treatments that inhibited the synthesis of endogenous insulin receptors specifically altered the phosphorylation of individual sites within the insulin receptor cytoplasmic domain. Hela cells co-infected with VTF7 and VTF7hIR were treated with ␣-amanitin or left untreated. Insulin receptors were immunoprecipitated, labeled with [␥-32 P]ATP in the presence of 100 nM insulin, and resolved by SDS-polyacrylamide gel electrophoresis. 32 P-labeled receptor ␤ subunits were localized by autoradiography and digested with trypsin. The resulting phos- phopeptides were resolved by HPLC into three sets of peaks (Fig. 5A). Two-dimensional analysis revealed that all phosphopeptides were conserved between insulin receptors isolated from ␣-amanitin-treated and untreated cells. Furthermore, three phosphopeptides within HPLC peak 2 contained phosphoserine and were conserved between ␣-amanitin-treated and untreated cells (Fig. 5, B and C). All other phosphopeptides resolved in each HPLC peak contained only phosphotyrosine (data not shown). These data demonstrate that although ␣-amanitin treatment blocks endogenous insulin receptor expression, it has no effect on the ability of over-expressed insulin receptors to be phosphorylated on specific serine phosphorylation sites in vitro. DISCUSSION We investigated the nature of the insulin receptor-associated serine kinase activity by examining its temporal association with the insulin receptor during synthesis. Our observations indicate that insulin receptor serine kinase activity is detectable in insulin receptors at early stages of receptor synthesis and persists during treatments that block endogenous mRNA expression. Serine kinase activity was detected in high mannose forms of the insulin proreceptor, a form of the receptor found in the endoplasmic reticulum and medial Golgi (26,27). The extent of proreceptor serine phosphorylation was comparable to that associated with fully processed receptor ␤ subunits ( Fig.  1B and Table I). If a distinct serine kinase associates with the insulin receptor, this observation suggests that it must do so prior to maturation of carbohydrate side chains and proteolytic processing of receptor subunits.
We also determined whether serine kinase activity would remain associated with insulin receptors translated from re-ceptor mRNA transcripts generated by T7 polymerase in cells treated with and without 10 g/ml ␣-amanitin for 18 h. ␣-Amanitin blocks de novo mRNA synthesis by inhibiting RNA polymerase II (29). ␣-Amanitin has no effect however, on T7 polymerase. Thus, ␣-amanitin treatment of cells infected with VTF7 and transfected with pTM1hIR will yield insulin receptors produced only by T7 polymerase provided by the recombinant virus (Fig. 2A, compare lanes 5 and 6). Insulin receptors expressed in VTF7-infected and pTM1hIR-transfected cells were at least 8-fold more abundant than insulin receptors in control cells (note times of exposure for lanes 5 and 6 versus lanes 1-4 in Fig. 2A).
Insulin receptor levels were elevated greatly with a recombinant vaccinia virus, VTF7hIR, encoding the human insulin receptor under control of the T7 promoter. Co-infection of VTF7hIR with the virus VTF7 allowed T7-mediated expression of insulin receptors in ␣-amanitin-treated cells that was 4000fold above endogenous receptor expression in ␣-amanitin cells infected with control virus (Fig. 3). Despite the ability of  1-7) and insulin receptor ␤ subunits (lanes 8 -12) were isolated and phosphorylated from HeLa cells transfected with plasmids pTM1hIR (lanes 1, 3, 5, 6, 8, 10, and 11) or control plasmid pTM-1 (lanes 2, 4, and 9) and infected with the wild-type control vaccinia virus WR (lanes 1, 3, and 8) or the recombinant vaccinia virus VTF7 (lanes 2, 4, 5, 6, 9, 10, and 11). Insulin receptor expression was also generated by co-infection with the recombinant viruses VTF7hIR and VTF7 (lanes 7 and 12). Following transfection and infection procedures, cells were incubated with (lanes 3, 4, 6, 7, 11, and 12) or without (lanes 1, 2,5,8,9, and 10) ␣-amanitin for 18 h. 32 P-labeled insulin receptor levels in proreceptors (lanes 2-6) and in ␤ subunits (lanes 9 -11) were normalized to control treatments in lanes 1 and 8, respectively. 32 P-labeled insulin receptor levels indicated in lanes 7 and 12 were normalized relative to receptor expression in cells co-infected with control viruses WR and VTF7 (Fig. 3). The ratio of phosphotyrosine to phosphothreonine was determined from phosphoamino acid analysis of each receptor sample and is indicated as an open circle. The presence of multiple circles denotes additional experiments for that condition.

FIG. 5. Phosphopeptide mapping of insulin receptors isolated from VTF7hIR-infected cells. A, HeLa cells infected with VTF7 and
VTF7hIR were treated with or without ␣-amanitin, labeled with 32 P, and treated with insulin as described in the legend to Fig. 3. Phosphopeptides from 32 P-labeled receptor ␤ subunits were separated by HPLC, and individual fractions were counted for radioactivity. B, phosphopeptides within HPLC peak 2 from insulin receptor treated with and without ␣-amanitin were analyzed in two dimensions on thin layer plates. The location of the 32 P-labeled phosphopeptides was determined by autoradiography of the thin layer plates. C, phosphoamino acid analysis of phosphopeptides marked A, B, and C in panel B. The relative migration of ninhydrin-stained phosphoamino acid standards is indicated.
␣-amanitin to block endogenous insulin receptor production, VTF7hIR-generated insulin receptors retained serine kinase activity comparable to that of endogenous insulin receptors isolated from control-treated cells (Figs. 3C and 4). The fact that the amount of serine phosphate on the insulin receptor precursor and mature ␤ subunit does not change appreciably in cells over-expressing the receptor could be explained by the presence of an excess amount of an associating serine kinase. However, persistence of insulin receptor-associated serine kinase activity of VTF7/pTM1hIR cells treated with ␣-amanitin for 18 h would mean that a distinct serine kinase activity would also have an extremely slow turnover rate compared with the endogenous insulin receptors that were abolished by ␣-amanitin. Consequently, we believe a more likely explanation for the insulin receptor-associated serine kinase is that the kinase domain of the insulin receptor ␤ subunit has the intrinsic ability to transfer phosphate to serine as well as tyrosine. Phosphopeptide mapping and phosphoamino acid analysis (Fig. 5) demonstrated that the pattern of serine phosphorylation on insulin receptors in VTF7hIR-infected cells is not altered by ␣-amanitin treatment. Thus, the serine kinase activity associated with the insulin receptor would not appear to be composed of multiple activities both distinct and intrinsic to the insulin receptor catalytic domain.
Previous reports have demonstrated that an insulin-sensitive serine kinase activity was associated with human insulin receptors partially purified from human placental membranes on wheat germ agglutinin-agarose (14 -17). An associated serine kinase is reportedly dissociated from the insulin receptor by 1 M NaCl (15,30), and that replacement of the NaCl eluate could reconstitute serine phosphorylation of the receptor. Other investigations found that a serine kinase activity remained associated with insulin receptors purified with 1 M NaCl included in the chromatography buffer (17,18). We demonstrated previously that affinity-purified preparations of insulin receptor retained an insulin-stimulated serine kinase activity that phosphorylated the insulin receptor at serines 1293 and 1294 (17). Our observation that affinity-purified receptor preparations could also phosphorylate synthetic peptides identical to sequences containing serines 1293/1294 indicated that the associated serine kinase activity was activated in response to insulin. However, this serine kinase activity associated with the insulin receptor was not capable of phosphorylating peptide substrates of other known kinases (17). The associated serine kinase activity could not be dissociated from the insulin receptor by gel filtration or density gradient centrifugation. 3 These data suggest that there may be at least two serine kinase activities associated with the insulin receptor. One of these kinases may be a distinct enzyme separable upon treatment with high salt concentrations. The data presented here suggest that at least one component of the insulin receptor-associated kinase activity is intrinsic to the insulin receptor. Our data indicate that the insulin receptor is a dual specificity kinase (19) capable of phosphorylating serine as well as tyrosine residues. Previous experiments with the tyrosine kinase inhibitor (hydroxy-2-naphthalenylmethyl)phosphonic acid support this conclusion (6).
The majority of eukaryotic protein kinases can be classified as protein serine/threonine kinases or protein tyrosine kinases. The predicted amino acid sequences of cloned kinase genes demonstrate structural differences that aid in this categorization. Several proteins with overall homology that is closer to the serine/threonine family of protein kinases have been shown to be able to autophosphorylate serine, threonine, and tyrosine residues. These dual specificity kinases may form a distinct family of enzymes, and their primary amino acid sequences have been suggested to contain information predictive of their catalytic specificity (19). The insulin receptor, however, shares greater homology with protein tyrosine kinases than with dual specificity kinases. This difference demonstrates the importance of directly determining phosphoacceptor specificity.
Phosphorylation of the insulin receptor by the associated serine kinase occurs on sites within the juxtamembrane domain (31,32) and the carboxyl-terminal tail (17). Phosphopeptide maps suggest that as many as four additional serine phosphorylation sites for the receptor-associated serine kinase may exist (17). Determination of the role of serine autophosphorylation in the enzymatic function of the insulin receptor and in insulin-mediated control of cellular metabolism may require identification of these sites. Analysis of mutant receptors lacking all serine phosphorylation sites may be necessary to evoke an effect on receptor enzymatic activity and signaling.