Identification of the site in the cGMP-inhibited phosphodiesterase phosphorylated in adipocytes in response to insulin and isoproterenol.

Stimulation of rat adipocytes with insulin and isoproterenol results in serine phosphorylation and activation of the adipocyte cGMP-inhibited phosphodiesterase (cGI PDE), events believed to be important in the antilipolytic action of insulin (Degerman, E., Smith, C.J., Tornqvist, H., Vasta, V., Manganiello, V.C., and Belfrage, P. (1990) Proc. Natl. Acad. Sci. U.S.A. 87,533-537). Here we demonstrate, by two-dimensional phosphopeptide mapping, that the major phosphopeptide generated by trypsin, or trypsin followed by Asp-N protease digestion of [32P]cGI PDE phosphorylated in adipocytes in response to isoproterenol and/or insulin, in each case co-migrates with the phosphopeptide released by the same treatment of M297FRRPS(P)LPCISREQ310. This peptide was synthesized based on the deduced sequence of the cloned rat adipocyte cGI PDE and phosphorylated by cAMP-dependent protein kinase (protein kinase A). Radiosequencing of authentic and synthetic tryptic 32P-peptides showed that a single site in cGI PDE (Ser302) was phosphorylated in adipocytes incubated with isoproterenol and/or insulin. The more than additive phosphorylation and activation of cGI PDE in response to the two hormones found in this report and previously (Smith, C.J., Vasta, V., Degerman, E., Belfrage, P., and Manganiello, V.C. (1991) J. Biol. Chem. 266, 13385-13390) is proposed to reflect cross-talk between their respective signal transduction pathways at the level of the cGI PDE serine protein kinase or upstream regulatory component(s).

Insulin controls cell metabolism and proliferation through several multicomponent diverging pathways, which are only partially understood (1)(2)(3). One important and well known metabolic effect is to antagonize hormone-activated lipolysis in adipose tissue, brought about mainly via insulin-mediated activation of a membrane-associated cGMP-inhibited phosphodi-esterase (cGI PDE, 1 referred to as the PDE 3 gene family; Ref. 4), which leads to a reduction in cAMP, a decrease in cAMPdependent protein kinase (protein kinase A) activity, and a lowering of the phosphorylation and thereby the activity of the hormone-sensitive lipase, the rate-limiting enzyme in the regulation of lipolysis (5)(6)(7)(8)(9)(10)(11).
Insulin and cAMP-increasing agents induce serine phosphorylation and activation of cGI PDE in rat adipocytes (12,13). An insulin-stimulated cGI PDE serine protein kinase activity in rat adipocytes has previously been partially characterized (14,15). In the presence of both isoproterenol and insulin, the phosphorylation/activation of cGI PDE is more than additive (13,16), suggesting cross-talk between the two signal transduction pathways.
To further dissect the components of the antilipolytic signaling pathway controlling the cGI PDE, we used two-dimensional tryptic phosphopeptide mapping and other methods to analyze the site(s) phosphorylated in this enzyme in 32 P-labeled rat adipocytes incubated with insulin, isoproterenol, or both. Since the amount of labeled phosphopeptides that could be obtained from experiments with intact cells was far less than that required for amino acid sequencing, the phosphorylation site(s) was identified by comparison of properties of phosphopeptides from cGI PDE phosphorylated in adipocytes, or from recombinant cGI PDE expressed in NIH 3006 human fibroblasts 2 (rcGI PDE) phosphorylated by protein kinase A, with peptides synthesized on the basis of the deduced sequence of rat adipocyte cGI PDE (17) and phosphorylated by protein kinase A. Peptides chosen for these studies contained consensus sequences for phosphorylation by protein kinase A (18), since preliminary results indicated that the site(s) phosphorylated during incubation of adipocytes with insulin or isoproterenol was located in the same tryptic phosphopeptide.  1) were synthesized (Biomolecular Core Facilities, Lund University) on an Applied Biosys-tems model 430A using the N-(9-fluorenyl)methoxycarbonyl (Fmoc) program and purified by preparative reversed phase chromatography (Kromasil C8 column). We previously reported that the major site in adipocyte cGI PDE phosphorylated by protein kinase A in vitro is Ser 427 (18), localized within P 423-440 (Fig. 1). The identity and purity of the synthetic peptides were assessed by amino acid analysis. The peptides were phosphorylated with protein kinase A (Sigma, P-2645), subjected to chromatography on C18 disposable minicolumns (Chromabond; Macherey-Nagel), and vacuum centrifuged (18). The phosphopeptides (PP) were suspended in 50 l of 50 mM ammonium bicarbonate buffer and trypsinized with 1 g of trypsin (modified sequencing grade; Promega) overnight at 37°C. The digests were subjected to vacuum centrifugation, oxidized, and re-trypsinized as described for the cGI PDE tryptic phosphopeptides below. All tryptic phosphopeptides (tPP) used in this study are presented in Fig. 1.
Tryptic and Asp-N Protease Phosphopeptide Mapping of 32 PcGI PDE-Adipocytes from Sprague-Dawley rats (35-37 days old) were prepared (21,22) and incubated with 32 P i (0.5-1 mCi/ml) (12,13). Solubilized cGI PDE from hormone-stimulated 32 P-labeled adipocytes (12,13) and protein kinase A-phosphorylated cGI PDE (see above) was incubated overnight with anti-P 423-440 , a polyclonal antiserum raised against P 423-440 (12,18). The immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (23), followed by electroblotting to nitrocellulose membranes (Amersham Life Science) (30 V, 16 h, in 5% methanol, 96 mM glycine, 12.5 mM Tris). Solubilized rcGI PDE was phosphorylated in vitro by protein kinase A and directly subjected to SDS-PAGE, followed by electroblotting to nitrocellulose membranes. 32 P-Phosphorylated cGI PDE was localized by digital imaging of 32 P (Fujix BAS 2000, Fuji Inc., Tokyo, Japan). That portion of the membranes was excised and the phosphoproteins digested with trypsin essentially as described (24). Briefly, the membrane pieces were incubated for 30 min at 37°C in 0.6% acetic acid (HAc) and 0.5% polyvinylpyrrolidone, washed three times with distilled water, and incubated with trypsin (1 g of trypsin in 200 l of 50 mM ammonium bicarbonate buffer/piece) for 16 h at 37°C. More than 95% of the 32 P radioactivity was released from the nitrocellulose membranes. The eluates were vacuum centrifuged, and the dried peptides were thereafter oxidized with formic acid (final volume of 50 l) for 60 min at 4°C. The oxidized phosphopeptides were diluted in 450 l of water, frozen and subjected to vacuum centrifugation, carefully dissolved in 50 l of 50 mM ammonium bicarbonate buffer, and trypsinized for 16 h. After the second trypsinization, 140 l of pH 1.9 buffer (formic acid:HAc:H 2 O (44:156:1800; v/v/v)) were added to the tryptic digests, and, after centrifugation (10,000 ϫ g, 5 min), 180 l were subjected to vacuum centrifugation. Finally, the phosphopeptides were dissolved in 7 l of pH 1.9 buffer and subjected to two-dimensional phosphopeptide mapping (25), using the Hunter thin layer electrophoresis apparatus (HTLE-7000, CBS Scientific). First dimension thin layer electrophoresis (TLE) was performed in pH 1.9 buffer for 25 min at 2000 V; second dimension ascending thin layer chromatography (TLC) was performed in isobutyric acid buffer (isobutyric acid:n-butanol:pyridine:HAc:H 2 O (1250:38:96:58:558; v/v/v/v/v)). After localization of 32 P by digital imaging, in some experiments 32 P-phosphopeptides were scraped from the TLC plates, eluted in pH 1.9 buffer, and lyophilized. The eluates were then either re-run in the two-dimensional system to assess co-migration with other peptides, immunoprecipitated with anti-P 423-440 followed by Tricine-SDS/urea-PAGE (26), or subjected to additional digestion with endoproteinase Asp-N (sequencing grade, 2 g reconstituted in 50 l of H 2 O, giving a final concentration of 10 mM Tris; Boehringer Mannheim). Briefly, the eluted tryptic phosphopeptides were suspended in 50 l of 50 mM ammonium bicarbonate buffer and further digested with 2 g of Asp-N protease (0.4 g added 5 times at 2 h intervals, followed by incubation overnight at 37°C). Trifluoroacetic acid (700 l of 0.1%) and 1 g of carrier-P 297-310 were added and the digests were applied to C18 disposable minicolumns and subjected to two-dimensional phosphopeptide mapping.
Estimation of the Stoichiometry of Hormone-induced Phosphorylation of cGI PDE-cGI PDE (0.06 pmol/sample as estimated from enzyme activity; Ref. 13), 32 P-phosphorylated in adipocytes as described above, was immunoprecipitated, subjected to SDS-PAGE, and analyzed by digital imaging and volume integration analysis of 32 P. 32 P radioactivity was determined from a standard curve relating volume integration values linearly to in vitro protein kinase A phosphorylated rcGI PDE (2-600 pmol) ( 32 P radioactivity obtained by liquid scintillation counting of pieces cut out from SDS-polyacrylamide gels). The specific activity of intracellular [␥-32 P]ATP at steady state was assumed to be half that in the extracellular medium (27). The total in vitro phosphorylation stoichiometry of cGI PDE was determined as in Ref. 18. Phosphorylation stoichiometry for individual tryptic phosphopeptides was estimated by comparing their volume integration values and relating them to total stoichiometry.

Tryptic Phosphopeptide Mapping of cGI PDE Phosphorylated in Adipocytes in
Response to Isoproterenol and/or Insulin-In intact adipocytes, the 135 kDa rat adipocyte cGI PDE was maximally phosphorylated to a similar extent in the presence of isoproterenol ( Fig. 2A, lane 2) or insulin (lane 3) (12) with an estimated total phosphorylation stoichiometry of 0.2 Ϯ 0.06 (mean Ϯ S.E., n ϭ 5 for each hormone). More than additive phosphorylation ( Fig. 2A, lanes 5-7) and activation (data not shown) occurred during incubation with submaximal concentrations of both hormones at shorter incubation times (13,16). [ 32 P]cGI PDE was immunoprecipitated, subjected to SDS-PAGE, and transferred to nitrocellulose membranes before tryptic digestion and analysis by two-dimensional phosphopeptide mapping. As shown in Fig. 2 (B-D), the 32 P-phosphorylated cGI PDE in each case yielded one major tryptic phosphopeptide (tPP iso , tPP ins , or tPP isoϩins ) and one minor tryptic phosphopeptide, which in all cases appeared to migrate to the same positions. Similar results were obtained in 18 separate experiments (7 experiments each for isoproterenol or insulin alone and 4 experiments for stimulation by the combination of the hormones). The amount of the minor phosphopeptide varied considerably between experiments and contributed to 0 -30% of the total amount of tryptic phosphopeptides. The small amount of this phosphopeptide precluded further analysis.
We have previously identified Ser 427 as the major site in the solubilized rat adipocyte cGI PDE phosphorylated by protein kinase A in vitro (18). Ser 427 is located within one of the three FIG. 1. Linear structure of cGI PDE. Boxes indicate putative protein kinase A consensus sequences; tryptic cleavage sites are marked by arrows. Tryptic fragments derived from protein kinase A-phosphorylated synthetic peptides P 423-440 and P 297-310 are denoted tPP 425-440 /tPP 426 -440 and tPP 300 -308 , respectively. Tryptic fragments derived from protein kinase A-phosphorylated P 272-299 are not indicated. consensus sequences for phosphorylation by protein kinase A (-RRXS-) in the rat adipocyte cGI PDE (17) (Fig. 1). This serine was not, however, phosphorylated in cGI PDE in intact adipocytes during isoproterenol stimulation (Fig. 2). Fig. 3 (A and B) is a representative example of two-dimensional tryptic phosphopeptide maps of isolated, solubilized adipocyte cGI PDE and rcGI PDE, maximally phosphorylated by protein kinase A to an estimated total phosphorylation stoichiometry of 0.9 Ϯ 0.2 (mean Ϯ S.E., n ϭ 5) (cf. Ref. 18). When compared to the two-dimensional tryptic phosphopeptide map from [ 32 P]cGI PDE phosphorylated in intact adipocytes (Fig. 2B), two major in vitro phosphopeptides tPP5 and tPP6 (Fig. 3, A and B) clearly migrated differently from the tPP iso . tPP5 and tPP6 were identified as tPP 426 -440 and tPP 425-440 (Fig. 1) by two approaches. (i) After elution from the TLC plates, tPP5 and tPP6 cross-reacted with anti-P 423-440 as judged from Tricine-SDS/urea-PAGE of immunopellets and immunosupernatants (Fig. 3C, lanes 5 (tPP 426 -440 ) and lane 6 (tPP 425-440 )). Minor phosphopeptides (denoted tPP1-tPP4) did not react (Fig. 3C,  lanes 1, 2, and 4). (ii) tPP5 and tPP6 co-migrated with tPP 426 -440 and tPP 425-440 respectively, as judged from mixing experiments in which the respective phosphopeptide was eluted from the TLC plate, mixed, and re-run together using two-dimensional phosphopeptide mapping (data not shown). tPP5 (tPP 426 -440 ) and tPP6 (tPP 425-440 ) were previously shown to be generated from P 423-440 by tryptic cleavage C-terminal to Arg 424 or Arg 425 and Arg 440 , respectively (18). Thus, tPP iso was clearly distinct from tPP 425-440 or tPP 426 -440 , i.e. the major site phosphorylated in cells by isoproterenol was different from that FIG. 2. Two-dimensional tryptic phosphopeptide maps of cGI PDE from intact rat adipocytes, stimulated with isoproterenol and/or insulin. A, solubilized cGI PDE from 32 P-labeled rat adipocytes (2 ml of 10% by volume of cells/condition) stimulated with isoproterenol (ISO), insulin (INS) or in combination (ISOϩINS), as indicated, was immunoprecipitated and subjected to SDS-PAGE, followed by digital imaging of 32 P. B-D, solubilized cGI PDE from 32 Plabeled rat adipocytes (20 ml of 10% by volume of cells with 0, 5-1 mCi/ml) stimulated with isoproterenol and/or insulin was immunoprecipitated, subjected to SDS-PAGE, transferred to a nitrocellulose membrane, and digested with trypsin. 32 P-Phosphopeptides were separated by two-dimensional TLE/TLC and visualized by digital imaging of 32 P. The major tryptic phosphopeptides are indicated as tPP iso , tPP ins , or tPP isoϩins . Minor tryptic phosphopeptides are indicated with arrows. X marks the origin. The electrophoresis direction indicated is from anode to cathode.

FIG. 4. Comparison of tryptic phosphopeptides from rat adipocyte cGI PDE phosphorylated in cells
with tPP 300 -308 . P 297-310 was phosphorylated by protein kinase A, trypsinized, and subjected to two-dimensional TLE/ TLC (A). Major tryptic phosphopeptides tPP iso , tPP ins , and tPP isoϩins derived from adipocyte cGI PDE phosphorylated in cells were eluted from the TLC plates and re-run without (B-D) or with (E-G) eluted tPP 300 -308 (A). X marks origin. phosphorylated in vitro by protein kinase A (Ser 427 ).
One of the minor phosphopeptides (tPP1) (Fig. 3) appeared to migrate similarly to tPP iso . Eluted tPP1, derived from solubilized rcGI PDE phosphorylated with protein kinase A in vitro, co-migrated with tPP iso in mixing experiments (data not illustrated). Thus, although most of the protein kinase A-catalyzed phosphorylation of solubilized authentic adipocyte cGI PDE took place on Ser 427 as reported previously (18), with an estimated phosphorylation stoichiometry of 0.4 Ϯ 0.04 (mean Ϯ S.E., n ϭ 5), some also occurred in the same tryptic peptide as that phosphorylated in intact cells during isoproterenol stimulation (estimated in vitro phosphorylation stoichiometry 0.2 Ϯ 0.04 (mean Ϯ S.E., n ϭ 7)). tPP1 was therefore used to further identify the site phosphorylated in cells, as described below.
The minor tryptic phosphopeptides tPP2, tPP3, and tPP4 migrated differently from the major tryptic phosphopeptide derived from cGI PDE phosphorylated in cells. The presence of tPP3 was variable, being absent in some experiments. The positions of these tryptic phosphopeptides in the enzyme were not identified.
Identification of Tryptic and Asp-N Protease-derived Phosphopeptides from cGI PDE Phosphorylated in Adipocytes-Since the amount of phosphopeptides that could be isolated from cGI PDE phosphorylated in adipocytes was far below that required for amino acid sequencing, tryptic phosphopeptides derived from adipocyte cGI PDE were instead evaluated for co-migration in several separation systems with the tryptic phosphopeptides obtained from protein kinase A-phosphorylated synthetic cGI PDE peptides (Fig. 1). This approach has been used previously in our laboratories for identification of the site of protein kinase A-catalyzed phosphorylation in vitro (18). As shown in Fig. 4, tPP iso co-migrated with tPP 300 -308 in the two-dimensional separation system, whereas this was not the case with any of the tryptic cleavage products from PP 272-299 (data not shown). Thus, the serine phosphorylated in cells during stimulation with isoproterenol appeared to reside in the cGI PDE amino acid sequence Arg 300 -Arg 308 .
As shown in Fig. 4, tPP ins and tPP isoϩins , which migrated very similarly to tPP iso in the two-dimensional phosphopeptide mapping system (Fig. 2, B-D), co-migrated with tPP 300 -308 . Thus, stimulation of adipocytes with isoproterenol, insulin, or the combination of both appeared to result in phosphorylation of cGI PDE within the amino acid sequence Arg 300 -Arg 308 . To verify this, the tryptic phosphopeptides were cleaved once more with Asp-N protease (known to cleave N-terminal to cysteic acid or aspartic acid). As shown in Fig. 5, this treatment generated one new phosphopeptide, which in each case again comigrated with the one obtained from Asp-N protease-treated tPP 300 -308 .
Identification by Other Methods of the Tryptic Phosphopeptides from cGI PDE Phosphorylated in Adipocytes-tPP1 comigrated with tPP iso (see above) and tPP ins (data not shown). Since tPP1 could be obtained in a sufficiently large amount from rcGI PDE phosphorylated in vitro by protein kinase A, which was not the case with the phosphopeptides derived from cGI PDE phosphorylated in adipocytes, tPP1 was used to verify further the identity of tPP iso , tPP ins , and tPP isoϩins using other separation systems. Using Mono Q, reversed phase C8 chromatography, or Tricine-SDS/urea-PAGE, tPP1 was shown in each case to co-fractionate with tPP 300 -308 (Fig. 6). These results, taken together, demonstrate that the cGI PDE amino acid sequence Arg 300 -Arg 308 contains the serine phosphorylated in situ during incubation with the different hormones.
Identification of the Serine Residue in cGI PDE Phosphorylated in Response to Hormone Stimulation of Adipocytes-Two serines, Ser 302 and Ser 307 , are located in the cGI PDE sequence Arg 300 -Arg 308 . To identify the specific phosphorylation site, tPP iso , tPP ins , and tPP isoϩins were coupled to a Sequelon-AA membrane and subjected to radiosequencing. In each case the radioactivity released was found in cycle 3 (Fig. 7). Similar results were found when tPP1 was analyzed (data not shown). This clearly demonstrates that Ser 302 was the site phosphorylated in cGI PDE during incubation of adipocytes with insulin or isoproterenol alone or in combination.

DISCUSSION
Using several experimental approaches and fractionation procedures, we have demonstrated that stimulation of adipocytes with insulin, isoproterenol, or the combination of both results in phosphorylation of cGI PDE at a single site (Ser 302 ) and activation of the enzyme. Ser 302 is localized within the sequence MFRRPS 302 LPCISREQ, one of three consensus sequences (-RRXS-) for putative protein kinase A-catalyzed phosphorylation of rat adipocyte cGI PDE (Fig. 1). The insulininduced phosphorylation was presumably mediated by the previously described (14,15) insulin-stimulated cGI PDE kinase activity found in cytosolic extracts of rat adipocytes following their stimulation by insulin. For technical reasons, it has not yet been possible to identify the serine(s) in cGI PDE phosphorylated in vitro during incubation with such extracts FIG. 5. Comparison of Asp-N protease-treated tryptic phosphopeptides from cGI PDE phosphorylated in rat adipocytes with Asp-N protease-digested tPP 300 -308 . tPP 300 -308 and tryptic phosphopeptides from cGI PDE phosphorylated in cells stimulated with isoproterenol, insulin, or both hormones were prepared and subjected to two-dimensional TLE/TLC. Major tryptic phosphopeptides, corresponding to tPP 300 -308 , tPP iso , tPP ins , and tPP isoϩins , were eluted from the TLC plates, re-run (A, D, G, and J), or digested with Asp-N protease and subjected to two-dimensional TLE/TLC as indicated (B, E, H, and K). These Asp-N protease-generated tryptic phosphopeptides from adipocytes (E, H, and K) were eluted and mixed with eluted Asp-N proteasedigested tPP 300 -308 (B) and subjected to two-dimensional TLE/TLC as indicated (F, I, and L). The generation of a new phosphopeptide is illustrated by mixing of Asp-N protease-digested and undigested tPP 300 -308 (C). X marks origin. from insulin-treated adipocytes. Several insulin-stimulated serine protein kinases, which appear to have one or several properties in common with insulin-stimulated cGI PDE kinase (e.g. activation by serine/threonine phosphorylation that is blocked by wortmannin treatment of cells), have been found in adipose tissue (28 -30).
Protein kinase A is likely to be responsible for the isoproterenol-induced phosphorylation of cGI PDE in intact cells. In cell-free preparations, Ser 427 rather than Ser 302 was found, as described previously (18), to be the major target for protein kinase A-catalyzed phosphorylation of cGI PDE. Ser 302 was found to be phosphorylated by protein kinase A in vitro, less extensively than Ser 427 but to the same extent as in intact cells in response to isoproterenol or insulin (phosphorylation stoichiometry approximately 0.2). This, and the fact that membrane-associated adipocyte cGI PDE has been reported to become activated (31), and platelet and hepatocyte cGI PDEs phosphorylated and activated in vitro by protein kinase A (32-34), indicates that this kinase directly (and not through FIG. 6. Separation of tryptic phosphopeptides by Mono Q, C8 reversed phase chromatography, or Tricine-SDS/urea-PAGE. rcGI PDE and P 297-310 were phosphorylated in vitro by protein kinase A in the presence of [␥-32 P]ATP. Tryptic phosphopeptides were separated by two-dimensional TLE/TLC. tPP1 (ϫ), derived from protein kinase A-phosphorylated rcGI PDE and tPP 300 -308 (E), were eluted from the TLC plates and subjected to chromatography on a Mono Q HR 5/5 (Pharmacia) column (A) or on a reversed phase C8 column (Kromasil C8 5, 4.6 ϫ 150 mm) (B), either separately or in combination (q). The Mono Q column was equilibrated in 10 mM Tris-HCl, pH 8.0, and eluted with a linear gradient of 0 -500 mM NaCl (1 ml/min). The C8 reversed phase column was equilibrated in 0.1% trifluoroacetic acid and eluted with a linear gradient of 0 -60% acetonitrile (0.5 ml/min). The elutionposition of the 32 P-labeled peptides were detected by liquid scintillation counting. C, tPP1 from protein kinase A-phosphorylated rcGI PDE and tPP 300 -308 were eluted from the TLC plates and subjected to Tricine-SDS/urea-PAGE. FIG. 7. Radiosequencing of tryptic phosphopeptides obtained from cGI PDE phosphorylated in adipocytes. Tryptic phosphopeptides from solubilized cGI PDE from 32 P-labeled rat adipocytes stimulated with isoproterenol and/or insulin were analyzed as described under "Experimental Procedures" and in Fig. 2. The major phosphopeptides (tPP iso , tPP ins , and tPP isoϩins ) were eluted and coupled to Sequelon-AA membranes and sequenced on an Applied Biosystems gas phase sequencer (model 470A). Released phenylthiohydantoin amino acidderivatives from each cycle were spotted on TLC plates, and the radioactivity was quantified by digital imaging and volume integration of each spot after subtraction of background 32 P. another kinase) phosphorylates and activates the enzyme in intact cells, although this still remains to be definitely established. The reason for the preferential phosphorylation of Ser 427 rather than Ser 302 by protein kinase A in vitro is not known. However, adipocyte cGI PDE is membrane-bound and in fat cell extracts requires salt and detergent for efficient solubilization. The cDNA structure of the enzyme predicts it to be an intrinsic membrane protein with six putative transmembrane sequences in its N-terminal region (17). In spite of several attempts with intact and solubilized membrane preparations of cGI PDE, we have so far been unable to establish conditions during which Ser 427 is not extensively phosphorylated by protein kinase A in vitro. 3 This may indicate that correct membrane insertion of the enzyme is a critical determinant for exposing Ser 302 and not Ser 427 for phosphorylation.
Previous work showing different time-courses for phosphorylation and activation of cGI PDE in rat adipocytes stimulated with isoproterenol or insulin (12,13) and more than additive phosphorylation/activation during stimulation with both hormones in combination (13,16), was tentatively interpreted as indicating phosphorylation at two sites in cGI PDE (10). Clearly, the present result is not consistent with this hypothesis and instead indicates that the signal pathways interact upstream of cGI PDE. Since the insulin-stimulated cGI PDE kinase appears to be activated in fat cells via phosphorylation during insulin stimulation (Ref. 10;cf. Ref. 35), this kinase could be dually regulated by an insulin-stimulated kinase kinase (or an insulin-inactivated phosphatase) and protein kinase A, respectively. However, phosphatidylinositol 3-kinase (36) and insulin receptor substrate-1 (and related proteins) (1, 37), known upstream links in the antilipolytic signal chain, could also be merging points. cAMP-enhancing agents have recently been reported to modulate the signals from insulin and other growth factors (38 -43), generally by attenuating them, but at least in one case by enhancing the effect of growth factors (43), consistent with the findings in this report. The identification reported here of the site in cGI PDE phosphorylated in cells in response to both insulin and cAMP-enhancing agents should be useful for the further dissection of the signal transduction pathway, linking the hormone receptor events to the activation of cGI PDE in the control of adipose tissue lipolysis.