An Effect of Insulin on Cyclic Adenosine 3’: SMonophosphate Phosphodiesterase Activity in Fat Cells*

Abstract Homogenates of rat fat cells were separated into three fractions by centrifugation: P1 (sedimented at 10,000 x g for 7 min), P2 (sedimented from the 10,000 x g supernatant after 20 min at 100,000 x g), and S (the 100,000 x g supernatant). All fractions contained adenosine 3':5'-monophosphate (cAMP) phosphodiesterase activity. The S fraction was enriched in higher Km (∼15 µm) phosphodiesterase and the P2 fraction in low Km (∼0.2 µm) activity, although both exhibited two apparent Michaelis constants for cAMP. When assayed with 62 nm cAMP, the specific activity of P2 was 2 to 3 times that of the unfractionated homogenate, and this fraction contained more than one-half the total phosphodiesterase activity. Incubation of fat cells with 1.0 milliunit per ml of insulin invariably increased the phosphodiesterase activity assayed in homogenates with l10 µm cAMP. The increment in activity was for the most part confined to the P2 fraction, the specific activity of which was increased 71 ± 5.5% (mean ± S.E.) in 24 experiments. The specific activity of 5'-AMP nucleotidase in P2 (which was 3 to 4 times that in the whole homogenate) was not altered by insulin. The effect on phosphodiesterase activity was essentially maximal after exposure of cells to insulin, 1 milliunit per ml, for 6 to 8 min. It was not diminished by washing the cells four times without insulin but was completely reversed by incubation for 30 min without insulin after washing. Neither prostaglandin E1, 2.8 µm, nor nicotinic acid, 10 µm, mimicked the effect of insulin on fat cell phosphodiesterase activity. The effect of insulin was prevented by anti-insulin serum and insulin treated with dithiothreitol was inactive. When present in the phosphodiesterase assay, dithiothreitol, 0.0042 to 4.2 mm, increased the activity of both P2 and S fractions isolated from control cells, whereas heparin, 8.4 to 420 ng per ml, and ethylene glycol bis(β-aminoethyl ether-N,N'-tetraacetic acid (EGTA), 0.05 to 2.0 mm, increased P2 activity with little or no effect on the activity of the supernatant fraction. Dithiothreitol, heparin, and EGTA increased, and sodium deoxycholate, 25 to 850 µg per ml, inhibited the activity of P2 fractions isolated from control and insulin-treated cells. Triton X-100, 0.0025 to 0.25%, was without effect. Incubation at 45° for 5 min reduced the phosphodiesterase activity of P2 fractions from control and insulin-treated cells to the same level which was about 50% of the initial control value. The activities then were stable for at least 25 min at 45°. Treatment of rats with dexamethasone for 20 hours decreased the phosphodiesterase activity of the isolated fat cells. The decrement in activity of whole homogenates was apparently accounted for by a loss of activity in the P2 fraction. These observations are consistent with the view that the effects of insulin and perhaps corticosteroids on cAMP-mediated processes in fat cells may be the result of alterations in the activity of a membrane-associated phosphodiesterase which has a relatively high affinity for cAMP.


From the Molecular
Disease Branch, Xational Heart and Lung Institute, National Maryland 2OOlg SUMMARY Homogenates of rat fat cells were separated into three fractions by centrifugation: Pi (sedimented at 10,000 x g for 7 min), Pz (sedimented from the 10,000 x g supernatant after 20 min at 100,000 x g), and S (the 100,000 x g supernatant).
All fractions contained adenosine 3':5'-monophosphate (CAMP) phosphodiesterase activity. The S fraction was enriched in higher K, (-15 pM) phosphodiesterase and the Pe fraction in low K, (-0.2 pM) activity, although both exhibited two apparent Michaelis constants for CAMP. When assayed with 62 nM CAMP, the specific activity of PLz was 2 to 3 times that of the unfractionated homogenate, and this fraction contained more than one-half the total phosphodiesterase activity. Incubation of fat cells with 1.0 milliunit per ml of insulin invariably increased the phosphodiesterase activity assayed in homogenates with <lO PM CAMP. The increment in activity was for the most part confined to the Pz fraction, the specific activity of which was increased 71 + 5.5% (mean =L-S.E.) in 24 experiments.
The specific activity of 5'-AMP nucleotidase in PZ (which was 3 to 4 times that in the whole homogenate) was not altered by insulin. The effect on phosphodiesterase activity was essentially maximal after exposure of cells to insulin, 1 milliunit per ml, for 6 to 8 (4). Rats were allowed free access to food and water until food was removed about 18 hours before decapitation. Samples of cells were incubated (37") in 4.0 ml of Krebs-Ringer phosphatc, medium containing bovine serum albumin, 30 mg per ml. After 10 min, insulin was added to some cells and the incubation continued, usually for 15 min. All cells were then washed twice with approximately 8 ml of medium without albumin and twice with similar volumes of 0.25 1\1 sucrose containing 10 mM Tris, pH 7.4. Solutions for washing cells were kept at 37". In early esperimcnts, insulin, 1 milliunit per ml, was included in the media used for washing the insulin-treated cells but in later espcrimcnts all cells were washed with the media described \yithout other additions.
After washing, the cells, suspended in 5 ml of 0.25 M sucrose, 10 mAr l'ris, pI1 7.4, were homogenized in a 7-ml Dounce homogenizer (tight strokes with a U pestle). A sample of each homogenate was stored in ice and the remainder was centrifuged at 10,000 x g for 7 min at O-5".
The fat cake was removed with a spatula, the supernatant decanted, and the sides of the tubes wiped dry, taking care not to disturb the pellet (PI). The supernatant was t,hen centrifuged at 100,000 x g for 20 mill, the supernatant' (S) decanted, and the sides of the tube wipc,d dry, again taking care not to disturb the pellet (Pt). Both the 10,000 x g pellet (PI) and the 100,000 x g pellet (PZ) TX-ere suspended in 0.25 M sucrose, 10 rnM Tris, pH 7.4, and dispersed using a Dounce homogenizer.
For assay of phosphodiesterase activity, samples were incubated at 30" in a total volume of 0. The incubation with 5'-nucleotidase proceeded for 30 min at 37" after which the samples were diluted, centrifuged, and ["Hladenosine isolated for radioassay, as pre-T-iousl\-described (5). Under the conditions employed, less than 15y0 of substrate was hydrolyzed in 10 min and for many experiments the activities reported represent verified initial rates of hydrolysis.
One and Triton X-100 from Packard Instrument Co.

RESULTS
The distribution of phosphodiesterase activity in the three homogenate fractions, PI (sedimented at 10,000 x g for 7 min), PZ (sedimented from the 10,000 x g supernatant after 20 min at 100,000 x g), and S (the 100,000 x g supernatant) is shown in Fig. 1 (Fig. 3). An effect of insulin on tllr activity in Pz was detected with assays at all three substrate concentrations but the percentage effect was smaller with the higher concentra-Cons of CAMP.
In four experiments in which 1 milliunit per ml of insulin increased the phosphodiesterasc activity by 57.8 rt Syb (mean =t S.E.), 0.1 milliunit per ml produced an increase of 26.5 & 5.97& In two experiments, 0.01 milliunit per ml produccd little or no increase in phosphodicsterase activity. Inca bation of fat cells with 1 milliunit per ml of insulin produced an increase in the velocity of the Pz I)llosp~rodiesteraRe activity with little or no change in the apparent I(, (approximately 0.2 pM) for CAMP.
As shown in to 8 min of washing caused little or no further increasr (Fig. 4).
After incubation of cells for 15 milr with illsulill, \\-ashing four times in medium TCthout insulin over a period of 6 to 8 min did not reverse the c#cct of insulilr.
Whell, bon-CJYT, the cells were incubated further for 30 min without illsulin, the phosphodiesterase activities were llot diffcrcnt from thos(' of cells ne\-er exposed to illsulill (l'ablc II, Esperimcnts 2 a11cl 3). Essentially all of the 5'-A11\11' nucslcotidasc activity of tllc, fat by guest on March 24, 2020 http://www.jbc.org/ cell homogenates was recovered in the P1 and PZ fractions, with the highest specific activity in the latter. Nucleotidase activity was not altered by incubation of fat cells with insulin (Table  III).
The specific activity of adenylate cyclase was somewhat ( < 100%) higher in the Pz fraction than it was in the PI fraction but the percentage effects of isoproterenol and NaF were similar in both. Adenylate cyclase activity (basal or in the presence Fat cells were incubated for 15 min, and washed four times with or without insulin, 1 milliunit per ml, as indicated.
In Experiments 1 and 2, they were then homogenized in, and fractions suspended in, medium without insulin.
In Experiment 3, after the first four washes, the cells were incubated for 15 min in medium of the same composition as the last wash, then insulin or diluent was added as indicated in parentheses.
After incubation for 15 min more the cells were washed four times without insulin and homogenized.
Phosphodiesterase activities of whole homogenates and S, PI, and Pt fractions were assayed. Since the effect of insulin was essentially confined to the homogenate and PZ fractions, only the values for PP phosphodiesterase activity are presented.

7167
of isoproterenol or NaF) was not altered by incubation of cells with insulin (data not shown).
Neither PGE1 nor nicotinic acid mimicked the effects of insulin on fat cell phosphodiesterase activity. Addition of cycloheximide or ouabain to the incubation medium did not prevent or modify the increase in phosphodiesterase activity produced by insulin (data not shown).
Anti-insulin serum, which, in this experiment but not in others, appeared to have a small effect, completely abolished the effect of insulin and insulin incubated with dithiothreitol for 15 hours before addition to fat cells produced no increase in phosphodiesterase activity (Table IV). When added to the phosphodiesterase assay, on the other hand, 42 or 420 PM dithiothreitol markedly increased phosphodiesterase activity in Pz fractions from control and from insulin-treated In Experiment 1, 0.5 ml of guinea pig anti-insulin (porcine) serum or 0.5 ml of buffer was added to flasks containing 3.5 ml of fat cell suspension.
After incubation for 15 min, insulin, 1 milliunit per ml, was added to some of the flasks and incubation continued for 15  6. Effect of dithiothreitol, EGTA, and heparin on phosphodiesterase activity in P, and S fractions. PP fractions (0) and S fractions (0) from control cells were assayed wit,h 62 nrvr (--) and 1.1 JJL-"I CAMP (---).
cells. With 4.2 m&f dithiothreitol, the stimulation of activity in the control Pt fractions was less and in some instances the effect of insulin was apparently at least partially reversed (Fig.  5). Dithiothreitol, 42 to 420 phi, also increased activity in the S fraction from control cells (Fig. 6).
,4ddition to the assays of, EGTA, 0.05 to 2 InM, or hepariq4 42 to 420 ng per ml, increased phosphodiesterase activity in Pz fractions and the iucrements were similar with fractions from control and insulin-treated cells (Fig. 5). As shown in Fig. 6, these agents produced little or no stimulation of phosphodiestrrase in the S fraction.
In fact 0.02 mM EGTB was inhibitory when the S fraction n-as assayed with 1.1 PM CAMP.
Sodium deoxycholate, at concentrations betwren 250 ..ir(! I ,3 ,:g per ml, 4 Incubation of fat cells with heparin, foilc;vc:l by washing of cells in the absence of heparin, did not alter phosphodiesterase activity in the P? fraction. Heparin also did not inhibit the lipolytic effect of epinephrine. Addition to the assay of hyaluronic arid (0.2 to 10 pg per ml), chondroitin sulfate (0.2 to 10 pg per ml), polyaspnrtic acid (0.083 to 42 rg per ml), or polyglutamic acid (0.083 to 42 pg per ml) did not stimulate P, activity. Ps fractions from control and insulin-treated cells to roughly the same extent (Fig. 5), whereas Triton X-100, 0.0025 to 0.25%, was without effect.
As shown in Fig. 7, phosphodiestcrase (assayed with 62 nM CAMP) was rapidly inactivated by incubation of the Pp fractions at 45". Within about 5 min the fractions from control and insulin-treated cells had reached the same specific activity which was about 50y0 of the initial control level and there was little further inactivation during the next 25 min at 45".
(The activity remaining after 20 min at 45" still exhibited two K, values for CAMP.) Treatment of rats with dexamethasone decreased the phosphodiesterase activity of the isolated fat cells. The effect, which was demonstrable only when assays were carried out with <lO pM CAMP, was essentially maximal within about 20 hours after the first injection of the steroid hormone. It was most marked in the P, fraction (Fig. 8) and, in fact, the entire decrement in activit'y of the whole homogenates was apparently accounted for by the loss of act'ivity in the P2 fraction. The phosphodiesterase activity of fat cells from rats treated with dexamethasone for 20 or 40 hours could st'ill be enhanced by incubation of the cells with insulin (data not shown). DISCUSSIOK We have found, as did Loten and Snegd (3), that incubation of fat cells with insulin, 0.1 t,o 1.0 milliunit per ml, for 6 to 8 min or less, products an increase in the phosphodiesterase activity measured in homogenates of t,hese cells using CAMP concernrations of less than about 10 PM (1, 3). Fat ccl1 homogenates contain phosphoclicsterase act,ivity with more than one apparern Michaelis constant for cAhlP (I, 3, 5) and more than one-half of the low K, activity, i.e. the activit,y assayed with subrnicromolar concentrations of cAMP, is associated with particulate clcments of the hornogenate. The fraction referred t'o above as Ps (sedimented in 20 min at 100,000 X g after re- moval of material sedimented in 7 mm at 10,000 x g) contains 50 to 60% of the low K, activity with a specific activity about 3 times that of the whole homogenate.
Iiy electron microscopy this fraction appears to consist largely of smooth membranous structures arranged in vesicles and sheets.
It contains adenylate cyclase and 5'.AMP nucleotidasc (enzymes presumably localized to the plasma membrane), t,he latter with a specific activity about 4 times that of the whole homogenate.
Essentially all of the increase in low K, phosphodiesterasc activity produced by incubation of cells with insulin occurs in this fraction. About 50% of the low I<, phosphodiesterase activity in P, fractions from control cells and all of the insulin-induced increment was destroyed after 5 min at 45" while t,he remaining activity was stable for at least 30 min at this temperature.
If by this treatment we have distinguished a heat-labile enzyme whose activity is enhanced by insulin from a heat-stable one whose activity is unaffected by insulin, it is apparent that insulin treatment. may produce a 2000/, increase in the activity of this enzyme.  8,9) and on the CAMP content of fat cells (2).
Thus far, our attempts to increase phosphodiesterase activity by the addition of insulin to homogenates or Pq fractions before or during assay have been unsuccessful.
Similarly, Carter and Martin (10) were unable, by the direct addition of insulin, to stimulate glucose uptake in a "microsomal" fraction from fat cells prepared by a procedure very similar to that> used for preparation of Pz although the microsomal fractions prepared from cells incubated with insulin took up glucose at an enhanced rate compared to that of fractions from control cells. Loten and Sneyd (3) also found that the addition of insulin to fat cell homogenates was without effect on phosphodiesterasc activity. An effect of insulin on a rnembrane fraction from rat liver has been reported (II), but from the data presented it is not clear whether insulin increased phosphodiesterase activit,y or prevented inactivation during the assay period. The observations reported hcrc are consistent w-ith our view that at least one of the consequences of the interaction of insulin with its specific binding sites on the fat cell surface is an increase in the activity of a phosphodiesterase in the plasma membrane, a phosphodiesterase with a relatively high affinity for CAMP, positioned perhaps in close proximity to the adenylate cgclase. Thus, degradation of CAMP could be regulated at a site very near that at which its synthesis takes place and is regulated. The rapid and reversible effect of insulin on membrane phosphodiesterase activity would appear sufficient to account for its effects on fat cell CAMP content (2) and on lipolysia (8,9). These could, of course, also result from an insulin-induced depression of adenylate cyclase activity. Although we and others have consistently failed to observe such an effect (1 I, la), it has been reported that insulin can prevent or diminish the stimulatory effects of glucagon on adenylate cyclase activity from liver (13) and fat cells (14).
Since the cell fractions in which adenylate cyclase activity is measured consist of more or less purified plasma membranes, the phosphodiesterase that is stimulated by insulin is presumably also present. To detci,mine whether insulin can in fact modulate the synthesis as well as the degradation of CAMP in fat cells, it will be nccossary to investigate the effects of insulin on an adenylatc cyclase preparation that is free of phosphodiesterase activity. It appears that the membrane-bound low Ii, diestcrasr may also play a role in the action of corticostcroid hormones on ci2MPmediated processes in fat cells. Senft et al. (15) originally reported that treatment of adrenalectomized rats with 6-methylprednisolone decreased the elevated phosphodiestcrasc activity of adipose and several other tissues.
The effect on liver ljhosphodiesterase activity was notrd 15 hours after a single subcutaneous injection.
As reported above, the decrcmerrt in fat ccl1 phosphodiestcrasc activity observed after 10 to 40 hours of dcxamethasonc treatment in intact rats was apparently