- Email: [email protected]
Cyclic AMP metabolism by swine adipocyte microsomal and plasma membranes
Comparative Biochemistry and Physiology Part B 124 (1999) 61 – 71 www.elsevier.com/locate/cbpb
Cyclic AMP metabolism by swine adipocyte microsomal and plasma membranes L.A. Zacher, G.B. Carey * Department of Animal and Nutritional Sciences, Uni6ersity of New Hampshire, Durham, NH 03824, USA Received 17 February 1999; received in revised form 31 May 1999; accepted 8 June 1999
Abstract Extracellular cyclic AMP is source of extracellular adenosine in brain and kidney. Whether this occurs in adipose tissue is unknown. The present study evaluated the capacity of swine adipocyte plasma membranes to metabolize cyclic AMP to AMP and adenosine, via phosphodiesterase (PDE) and 5%-nucleotidase (5%-NT), respectively. Plasma membranes (PM) and microsomal membranes (MM) were isolated from over-the-shoulder subcutaneous adipose tissue of 3 month-old male miniature swine. The purity of the membrane fractions was determined and PDE and 5%-NT activities in PM and MM fractions were corrected for cross-contamination. The maximal activity of MM-PDE was 7-fold greater than that of PM-PDE. MM-PDE was 100% inhibited by 5 mM cilostamide, while PM-PDE was unaffected by this PDE3B inhibitor. Inhibitors of PDE1, PDE2, PDE4 and PDE5 also failed to inhibit PM-PDE. However, 1 mM DPSPX inhibited PM-PDE activity by 72%. When PM were incubated with 0.8 mM cyclic AMP for 20 min, AMP accumulation was four times that of adenosine. These data demonstrate that cyclic AMP can be converted to AMP and adenosine by the PM-bound enzymes 5%-NT and PDE, and suggest that the PM-PDE responsible for extracellular cyclic AMP metabolism to AMP is distinct from the intracellular MM-PDE. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Adenosine; Adipocytes; AMP; Cyclic AMP; Phosphodiesterase; Plasma membrane; Microsomal membrane; Swine
1. Introduction Adenosine inhibits lipolysis in adipocytes by binding to the A1 receptor and inhibiting adenylate cyclase [17] and adipocyte sensitivity to adenosine is altered when animals are subjected to modifications in diet and exercise [6,31]. However, the origin of extracellular adenosine in adipose tissue is unknown. Extracellular adenine nucleotides have been suggested as a source of extracellular adenosine in other tissues. In arterial smooth muscle cells and cardiac myocytes, the hydrolysis of extracellular ATP is a significant source of extracellular adenosine [27]. In brain cells, brain slices and perfused kidney, cyclic AMP (cAMP) is exported then hydrolyzed by the ‘ecto’ enzymes phosphodiesterase (PDE) and 5%-nucleotidase (5%-NT) to produce extracellular adenosine [4,32,39]. * Corresponding author. Tel.: +1-603-862-4628; fax: + 1-603-8623758. E-mail address: [email protected] (G.B. Carey)
Isolated adipocytes export cAMP [5,13], and intact adipose tissue has been shown to metabolize exogenous cAMP to extracellular adenosine [16]. If exported cAMP is metabolized to adenosine, a negative-feedback loop may exist in adipocytes where this adenosine could bind to the A1 receptor on the plasma membrane and inhibit adenylate cyclase activity, thereby decreasing formation and export of intracellular cAMP. Such a feedback loop was first described in 1991 in the juxtaglomerular cell of the kidney [18] and has since been described in brain [4,39]. Extracellular cAMP metabolism to adenosine would require the presence of outwardly facing ‘ecto’ enzymes PDE and 5%-NT on the surface of adipocytes. The existence of the plasma membrane (PM)-bound 5%-NT in adipocytes is well documented [15,34–37,45]. However, research on the PM-bound PDE in adipocytes is limited. The majority of PDE research has focused on intracellular, microsomal membrane (MM)-bound PDE. Adipocyte MM-PDE is sensitive to certain hormones [1] as well as to exercise treatment of animals
0305-0491/99/$ - see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 3 0 5 - 0 4 9 1 ( 9 9 ) 0 0 0 9 8 - X
62
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
[20]. PM-bound PDE has been identified in adipocytes, but research on this enzyme has focused primarily on its activity in basal states [24]. PM-bound PDE activity is altered by treating isolated adipocytes with phospholipids [23], and is dependent on the site of adipose tissue deposition as well as the diet fed to the animal [3]. The possibility that extracellular adenosine in adipose tissue arises from extracellular cAMP was explored in the present study. The goal of this study was to determine whether the PDE bound to the PM of swine adipocytes is distinct from the MM-bound PDE, and to evaluate the capacity of swine adipocyte plasma membranes to metabolize cAMP to AMP and adenosine. We demonstrate that the PM-bound PDE is cilostamide-insensitive, suggesting that it is distinct from the cilostamide-sensitive MM-bound PDE. We also show that although the maximum activity of PM5%NT is approximately 7-fold higher than that of the PM-PDE, the 5%-NT limits adenosine production under pre-steady state conditions in which the low Km PDE is saturated with substrate. Our results are consistent with the idea that a pathway for the negative-feedback regulation of the cAMP second messenger pathway may exist in adipocytes.
2. Materials and methods
2.1. Animals Yucatan miniature swine (Sus scrofa) were bred at the University of New Hampshire’s Burley Demeritt swine facility. All swine were 2.5 – 3.5 month old males, with a mean body weight9S.D. of 11.139 1.59 kg. The swine were housed with their littermates and were group-fed miniature swine ration (Agway, Syracuse, NY) twice a day with free access to water. All procedures were approved by the Institutional Animal Care and Use Committee.
2.2. Tissue biopsy, adipocyte isolation, and membrane fractionation Adipose tissue was removed from subcutaneous overthe-shoulder adipose tissue deposits after the animals were euthanized. The adipocytes were isolated as previously described with modifications [6]. Collagenase D was chosen because of its relatively low trypsin levels compared to other types of collagenase; using collagenase with high trypsin levels to isolate fat cells has been shown to decrease phosphodiesterase activity [11]. The adipocyte membranes were fractionated using the method of Simpson et al. [42] with modifications. Briefly, isolated adipocytes were washed twice in a 10 mM Tris–HCl–255 mM sucrose – 2 mM ngCl2 –0.1 mM phenylmethylsulfonyl fluoride (PMSF) homogeniz-
ing medium at 37°C to remove any Krebs–Ringer buffer. Adipocytes were then resuspended in 2 volumes of the same homogenizing medium, and homogenized with 3 up-and-down strokes in a 55 ml size C Teflon pestle homogenizer (Arthur H. Thomas and Co.). All subsequent steps took place at 4°C. De-fatted homogenate was centrifuged for 15 min at 16 000×g and the resulting pellet (containing PM) was resuspended in 20 ml homogenizing medium and centrifuged for 15 min at 16 000 × g. The resulting pellet was resuspended in 5 ml homogenizing medium, placed on 5 ml of a 1.12 M sucrose cushion containing 20 mM Tris–HCl–0.1 mM PMSF, and centrifuged at 101 000× g for 70 min. The plasma membranes were removed from the interface using a transfer pipette, and the mitochondria, nuclei, and cell debris were collected as a pellet. The mitochondrial pellet and the plasma membranes were resuspended in 30 ml homogenizing medium and centrifuged for 30 min at 30 000× g. The plasma membrane and mitochondrial pellets were resuspended in buffer containing 40 mM Tris–HCl–2.5 mM MgCl2 –1 mM EGTA–0.3 mM PMSF–0.001 mg leupeptin (resuspension buffer) to a volume of approximately 2 mg protein/ml, and approximately 4 mg protein/ml, respectively. The initial supernatant from the first 16 000×g spin was centrifuged for 70 min at 212 000× g. The pellet (containing microsomes) was resuspended in resuspension buffer to a final volume of approximately 2 mg protein/ml. Protein was determined by the method of Lowry et al. [22]. All cell fractions were immediately frozen upon preparation and stored at −80°C. Membrane marker enzyme assays were performed within 1 week; PDE and 5%-NT assays were generally performed within 6 weeks. The stability of PDE activity was verified in the membranes stored up to 6 weeks. In a small subset of experiments PDE was assayed from 8 weeks to 5 months after preparation of membranes. Although we did not verify PDE stability in the same preparation over 5 months, there was no evidence of a loss of PDE activity during this time. 5%-NT stability was not assessed.
2.3. Membrane marker enzyme assays 5%-NT marker activity was assayed as described [2] in the presence of 0.05% Triton X-100. Cytochrome c reductase activity was measured by the method of Dallner et al. [9]. Citrate synthase was assayed by the method of Srere [43].
2.4. Experimental incubations PDE activity was measured by the method of Thompson [44]. Plasma or microsomal membranes (5 mg ) in resuspension buffer were incubated at 37°C at a
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
final volume of 50 ml. Either 0.8 mM or 20 mM [3H]cAMP was added to start the reaction. In some cases, PDE activity was inhibited by pre-incubation with cilostamide (ranging in concentrations from 0.05 to 10 mM) at 37°C for 2 min prior to the addition of cAMP. Cilostamide was prepared in DMSO, and the final DMSO concentration did not exceed 0.2% (v/v). Other PDE inhibitors used were 1,3-dipropyl-8-p-sulfophenyl-xanthine (DPSPX), rolipram, erythro-9-(2-hydroxy-3-nonyl)adenine hydrochloride (EHNA), 8-methoxymethyl-3-isobutyl-1-methylxanthine (8methyl-IBMX) and zaprinast. 8-Methyl-IBMX was prepared in 100% ethanol, and the final ethanol concentration did not exceed 0.5% (v/v). Zaprinast was prepared in 100% methanol, and the final methanol concentration did not exceed 0.5% (v/v). DPSPX, rolipram, and EHNA were prepared in resuspension buffer. Reactions were halted by the addition of 100 ml 0.1 M HCl, and neutralized by the subsequent addition of 0.1 M Trizma base. Snake venom (croatus atrox) (10 ml) was added to each vial to convert any AMP formed to adenosine. Less than 50% of the substrate was hydrolyzed during the incubations, and the rates of adenosine formation remained linear for 20 min. 5%-NT activity was measured in the same system as the PDE (as distinct from the marker assay of Ref. [2]). Plasma membranes (5 mg) were used, and resuspension buffer and 50 mM AMP were added to yield a final volume of 50 ml. 5%-NT activity was inhibited using a,b-methyleneadenosine-5%-diphosphate (AMP-CP), ranging in concentration from 50 to 900 mM. AMP-CP, dissolved in resuspension buffer, was added to the plasma membranes and allowed to incubate at 37°C for 7 min prior to the addition of [3H]AMP. After a pre-determined time, the reactions were halted by boiling for exactly 1 min. Chromatography columns for the PDE assay were filled with 1 ml DEAE-Sephadex A25 resin, and equilibrated with 8 ml high salt buffer (20 mM Tris – HCl, 0.5 M NaCl, pH 6.8) followed by 8 ml low salt buffer (20 mM Tris–HCl, pH 6.8) for the PDE assays. Columns for the 5%-NT assays were filled with 3 ml DEAE-Sephadex A25 resin, and equilibrated with 25 ml high salt buffer (pH 6.8) followed by 25 ml low salt buffer (adjusted to pH 8.4). These volumes of DEAE-Sephadex ensured that less than 1.12% unreacted [3H]cAMP or [3H]AMP flowed through the columns when washed with the low salt buffer, and that all [3H]cAMP or [3H]AMP flowed through the columns when washed with the high salt buffer. The samples were applied to the DEAE-Sephadex columns and washed with 2 ml low salt buffer (pH 6.8) for the PDE samples, or 6 ml low salt buffer (pH 8.4) for the NT samples, and the eluate was collected. Scintillation cocktail in the amount twice that of the low salt washes was added to each vial, and the vials were vortexed and
63
counted in a liquid scintillation counter for radioactive adenosine. Graphpad Prism™ (Graphpad Software, San Diego, CA) was used for non-linear curve fitting. Goodness-offit was determined by calculating a P-value from the F ratio (the relative increase in sum-of-squares and the relative increase in degrees of freedom, for each curve being considered). Statistical significance was set at PB 0.05.
2.5. Chemicals Cilostamide was purchased from Tocris Cookson, St. Louis, MO. The columns were purchased from Evergreen Scientific. Collagenase D (lot c 83615722) was purchased from Boehringer Mannheim Biochemicals, Indianapolis, IN. DPSPX, EHNA, and rolipram were purchased from Research Biochemicals International, Natick, MA. All other chemicals and supplies were supplied by Sigma, St. Louis, MO or Fisher Scientific, Springfield, NJ.
3. Results Previous work has shown that the PDE enzyme in human adipocytes consists of two isoforms, one that hydrolyzes relatively low levels of substrate—referred to as low Km PDE—and one that hydrolyzes higher levels of substrate—referred to as high Km PDE [11]. PDE activity from swine adipose tissue also demonstrates a biphasic nature consistent with the notion that this tissue contains two isoforms [29]. We assayed for both isoforms: the low Km PDE was assayed at B1 mM cAMP and relatively short time points, and the high Km PDE was assayed at \ 1 mM cAMP and longer time points. Data for the low and high Km assays were combined for non-linear curve fitting analyses.
3.1. PDE time course To evaluate PDE activity over time, 4.5 mg PM or MM protein were incubated for 1–20 min with 0.8 mM cAMP to assess low Km PDE activity, or incubated with 20 mM cAMP for 5–90 min to assess high Km PDE activity. Levels of cAMP that saturate the low and high Km activities were determined by dose response experiments described in the next section. Low Km PDE activity in PM and MM fractions remained linear for up to 20 min and hydrolyzed 21 and 47% substrate, respectively, at 20 min (Fig. 1A). High Km PDE activity in PM and MM remained linear for up to 90 min and hydrolyzed 12 and 26% substrate, respectively, at 90 min (Fig. 1B).
64
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
3.2. PDE dose response to cAMP To determine a dose of cAMP that maximally stimulates low Km PDE, 5 mg of PM-protein or MM-protein were incubated for 12 and 6 minutes, respectively, with cAMP concentrations ranging from 0.05 to 0.80 mM. PM and MM were also incubated for 60 min with cAMP concentrations ranging from 2 to 26 mM to measure high Km PDE activity. Non-linear curve fitting analysis of PM-PDE activity over the entire cAMP concentration range (0.05 – 26 mM) showed that the data fit a two-component curve significantly better than a one-component curve, implying the existence of two isozymes (Fig. 2). The Km and Vmax values for the two-component curve are Km1 =0.15 mM cAMP and
Fig. 2. PDE activity in the PM fraction of swine adipocytes in response to increasing concentrations of cAMP. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: PM protein (5 mg) was incubated for 12 min at 37°C with doses of cAMP ranging from 0.05 to 0.8 mM and for 60 min at doses ranging from 2 to 26 mM cAMP. The dotted line represents the fit of data to a one-component curve, and the solid line represents the fit of data to a two-component curve, using the software Graphpad Prism™. The solid line for the two-component curve fits significantly better than the one-component curve.
Vmax1 = 54.1 pmol/min/mg protein, and Km2 = 2.0 mM cAMP and Vmax2 = 106.4 pmol/min/mg protein. Vmax1 accounts for 34% of the total activity, and Vmax2 accounts for 66% of the total activity. The Km and Vmax values for the PM one-component curve are 0.66 mM cAMP and 152.8 pmol/min/mg protein, respectively. Non-linear curve fitting of MM-PDE activity data also implied the existence of two isozymes (Fig. 3). The Km and Vmax values for the two-component curve are
Fig. 1. Panel A: time course of low Km PDE activity in adipocyte PM fraction (circles) and MM fraction (triangles). Membrane fractions were prepared from adipocytes isolated from three different pigs on three separate days. Fractions were subsequently pooled for assay. PM or MM protein (4.5 mg) was incubated with 0.8 mM cAMP for the indicated times. Values represent the mean9 S.D. for four replicate samples from the pool. The results were verified using membrane fractions prepared from a single pig. Panel B: time course of high Km PDE activity in adipocyte PM fraction (circles) and MM fraction (triangles). PM or MM (4.5 mg protein) was incubated with 20 mM cAMP for the indicated times. Membrane fraction preparation and pooling, data expression, and experiment verification are the same as described for Panel A.
Fig. 3. PDE activity in the MM fraction of swine adipocytes in response to increasing concentrations of cAMP. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: MM protein (5 mg) was incubated for 6 min at 37°C with doses of cAMP ranging from 0.05 to 0.8 mM and for 60 min at doses ranging from 2 to 26 mM cAMP. The dotted line represents the fit of the data to a one-component curve, and the solid line represents the fit of the data to a two-component curve, using the software Graphpad Prism™. The solid line for the two-component curve fits significantly better than the one-component curve.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
65
Table 1 Purity of fractions obtained from differential centrifugation of swine adipocytesa Fraction
Homogenate
5%-NT (nmol/h/ mg protein)
21.09 0.61 (2564) PM 56.6 9 1.5 (300) MM 50.8 9 1.4 (314) mitochondria, nuclei and 41.69 0.8 (385) cell debris Cytosol 0 (0)
Cytochrome c reductase (mmol/min/mg protein)
Citrate synthase (mmol/min/mg protein)
Total protein (mg)
0.369 0.01 (43.6)
0.15 90.01 (18.8)
122
0.669 0.03 (3.5) 1.969 0.03 (12.1) 0.679 0.01 (6.2)
0.07 90.01 (0.4)95.3 0.13 90.01 (0.8) 1.16 90.09 (10.8)
0.02 90.01 (1.1)
0.03 90.01 (1.6)
6.2 9.3 54
a
The membrane fractions were prepared from adipocytes isolated from three different pigs on three separate days. The fractions were subsequently pooled and the single pool was assayed for marker enzymes to determine membrane purity. The values are mean 9 S.D. of enzyme activity for three replicate samples from the pool. The values in parentheses represent the total enzyme activity in each fraction: nmol/h for 5%-NT and mmol/min for cytochrome c reductase and citrate synthase. The 5%-NT assay was conducted by incubating 20 mg protein with 400 mM [3H]AMP for 1 h. The cytochrome c reductase and citrate synthase assays were conducted using 5 mg membrane protein.
Km1 =0.13 mM cAMP and Vmax1 =138.6 pmol/min/mg protein, and Km2 =3.3 mM cAMP and Vmax2 =306.4 pmol/min/mg protein. Vmax1 accounts for 31% of the total activity, and Vmax2 accounts for 69% of the total activity. The Km and Vmax values for the one-component curve are 0.82 mM cAMP and 405.8 pmol/min/mg protein, respectively. The maximum activity of PDE in the adipocyte PM fraction is approximately 161 pmol/min/mg protein; this is approximately one-third that in the MM fraction. However, neither the PM nor the MM fractions are pure. To more accurately determine the kinetic parameters of the PDE in each fraction, fraction purity was determined.
3.3. Cross-contamination of adipocyte fractions Attempts were made to prepare the purest PM fraction possible by optimizing centrifugation speeds, buffer components, and buffer concentrations. MM contamination of the PM fraction was reduced from approximately 70% contamination to approximately 30% contamination by this optimization. We then used three enzymes as markers for PM, MM, and mitochondrial membranes (Table 1). These markers were 5%-NT for the PM fraction, cytochrome c reductase for the MM, and citrate synthase for the mitochondrial membranes. Using the 5%-NT marker to determine the amount of PM present in each fraction, we found the PM fraction had been enriched almost 3-fold from homogenate levels, however, similar enrichments of 5%NT were found in the mitochondrial and MM fractions. The similarity of 5%-NT enrichment in these three fractions questions the value of 5%-NT as a plasma membrane marker, as pointed out by Mersmann and Shparber [28]. Cytochrome c reductase showed a 5-fold enrichment in the MM fraction, and only one-third of the cytochrome c reductase activity in the MM fraction was present in the PM and mitochondrial fractions.
The fractionation procedure was very effective at removing mitochondrial membranes from MM and PM fractions, as citrate synthase was only 6 and 11%, respectively, of the activity present in the mitochondrial membrane fraction. The cytosol fraction had negligible levels of the three marker enzymes. Recovery of homogenate enzyme activity, determined by summing total enzyme activity in each fraction, was 39% 5%-NT, 53% cytochrome c reductase and 72% citrate synthase. Cross-contamination data were used to calculate uncontaminated PM-PDE and MM-PDE activities.
3.4. Microsomal and plasma membrane PDE response to cilostamide PDE activity has been reported in PM and MM fractions of isolated adipocytes [1,23] and the MMPDE is known to be completely inhibited by cilostamide [3]. To compare the cilostamide sensitivity of PDE in the PM and MM fractions, the membranes were incubated with concentrations of cilostamide ranging from 0.04 to 10 mM (Fig. 4). In the adipocyte MM fraction, PDE activity was maximally inhibited by 5 mM cilostamide (Fig. 4, dashed line), reducing PDE to 10% of its maximal activity (to 19.4 from 189.0 pmol/min/mg protein). Cilostamide also inhibited the PM fraction, reducing PDE to 32% of its maximal activity (to 26.5 from 84.0 pmol/min/mg protein; Fig. 5, dashed line). The remainder of the PDE activity (26.5 pmol/min/mg protein) is presumably PM-PDE that is insensitive to cilostamide. The cilostamide-insensitive PDE activity in the MM fraction (19.4 pmol/min/mg protein) can be completely attributed to PDE activity from plasma membrane contamination. PM-PDE activity was determined using data from the 5%-NT marker assay (89% contamination of MM fraction by PM= 0.89× 26.5 pmol/min/mg protein= 23.6 pmol/min/mg protein). After correcting for PM contamination, the microsomal membrane PDE
66
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
([S]/Km)) [8] where IC50 = 0.063 mM cilostamide, [S]= 0.8 mM cAMP, and Km = 0.13 mM cAMP. PM-PDE response to cilostamide was calculated by subtracting PDE activity due to MM contamination, using cytochrome c reductase marker assay data (33.5% of the cytochrome c reductase activity in the MM fraction was present in the PM fraction). At each cilostamide concentration, 33.5% of the corresponding MM fraction PDE activity was subtracted from the PM fraction PDE activity. Correcting for microsomal contamination of the PM fraction, PM-PDE showed no inhibition by cilostamide (Fig. 5, solid line). Fig. 4. Low Km PDE activity in MM in response to increasing concentrations of cilostamide. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: MM protein (5 mg) was pre-incubated with cilostamide (0.04 – 10 mM) for 2 min, followed by the addition of 0.8 mM cAMP for 6 min. The dashed line represents total MM fraction activity and the solid line represents calculated MM-PDE activity in the fraction after correcting for PM contamination using 5%-NT marker assay data (Table 1). Lines are fit on a point-to-point curve. Inset is a linear regression of the MM-PDE activity after correcting for PM contamination vs. log nM cilostamide; these data were used to determine the IC50.
(MM-PDE) showed complete inhibition by cilostamide by a dose of 5 mM (Fig. 4, solid line). Plotting PDE activity versus the log of cilostamide concentration, we determined the IC50 of cilostamide in the microsomal membrane fraction (after correcting for contamination) to be 63 nM (inset). The Ki of cilostamide was 8.8 nM, and was determined using the equation Ki =IC50/(1+
3.5. Plasma membrane PDE acti6ity in the presence of other known PDE inhibitors Five known PDE inhibitors were examined for their effectiveness to inhibit adipocyte PM-PDE activity. These included 8-methyl-IBMX to inhibit PDE1 [46], EHNA to inhibit PDE2 [30], rolipram to inhibit PDE4 [41], zaprinast to inhibit PDE5 [33], and DPSPX to inhibit the ecto-PDE [32]. After a 2 min pre-incubation with 5 mM cilostamide to inhibit any MM-PDE contaminating the PM fraction, 5 mg of PM protein were pre-incubated with either 1, 100, or 1000 mM DPSPX, 52 mM 8-methyl-IBMX, 10 mM EHNA, 10 or 100 mM rolipram, or 7.6 mM zaprinast. These concentrations are 5–100-fold above the published IC50 values for each inhibitor [30,32,33, 41,46]. The reactions were started with the addition of 0.8 mM cAMP, and were stopped after 12 min. The inhibitors 8-methyl-IBMX, EHNA, rolipram, and zaprinast showed no inhibition of PM-PDE (data not shown). Although 1 mM DPSPX failed to inhibit PMPDE, 100 and 1000 mM DPSPX significantly reduced PM-PDE activity by 25 and 73%, respectively (Fig. 6).
3.6. Uncontaminated PDE dose response to cAMP
Fig. 5. PDE PM activity in response to increasing concentrations of cilostamide. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: PM protein (5 mg) was pre-incubated with cilostamide (0.04 – 10 mM) for 2 min, followed by the addition of 0.8 mM cAMP for 12 min (n= 4). Dashed line (triangles) represents total PM fraction activity and is fit to a one-component curve. The solid line (circles) represents the calculated PM-PDE activity after correcting for microsomal contamination, using cytochrome c reductase marker assay data, and is fit to a linear regression model.
To evaluate PM-PDE enzyme kinetics in the absence of contaminating MM-PDE, the PM fraction was preincubated with cilostamide to completely inhibit any MM-PDE activity present. The cAMP dose response of PDE in the PM fraction in the presence of cilostamide showed that the two-component curve (Fig. 7A, upper solid line) fit significantly better than the one-component curve (Fig. 7A, dashed line). The Km and Vmax values for the two-component curve (PM+ cilostamide) are Km1 = 0.12 mM cAMP and Vmax1 = 23. 8 pmol/min/ mg protein, and Km2 = 5.0 mM cAMP and Vmax2 =30.3 pmol/min/mg protein. Vmax1 accounts for 44% of the total activity and Vmax2 accounts for 56% of the total PM-PDE activity. The Km and Vmax values for the one-component curve are 0.49 mM cAMP and 45.1 pmol/min/mg protein, respectively.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
67
mately 4.5 mg PM protein was incubated for 30–120 min with 50 mM AMP. Fig. 8 shows that PM activity remained linear for up to 120 min and hydrolyzed 31% of substrate at the end of this time. When 5 mg PM protein was incubated with increasing concentrations of AMP, the activity of the enzyme 5%-NT was biphasic (Fig. 9). An initial plateau at approximately 200 pmol/min/mg protein was seen between 50 and 250 mM AMP; the Km for this portion was at 4.9 mM AMP. At AMP concentrations \500
Fig. 6. PM-PDE activity in the presence of increasing doses of DPSPX. Membrane fractions were prepared from adipocytes isolated from two different pigs on two separate days. The fractions were subsequently pooled for assay. PM protein (5 mg) was pre-incubated with 5 mM cilostamide for 2 min, followed by a 7 min pre-incubation with 0, 1, 100 or 1000 mM DPSPX. The membranes were then incubated with 0.8 mM cAMP for 12 min. Values represent the mean9 S.D. for four replicate samples from the pool. * Significantly different from control (0 mM DPSPX).
When PM fractions were incubated with increasing doses of cAMP, and PDE activity was corrected for MM-PDE contamination, the data were more variable, especially at the higher cAMP doses (Fig. 7). The two-component curve model did not fit significantly better than the one-component curve model. The latter model revealed a Km and Vmax 0.42 mM cAMP and 37.2 pmol/min/mg protein, respectively (Fig. 7A, lower solid line). When MM fractions were incubated with increasing doses of cAMP, and PDE activity was corrected for PM-PDE contamination, MM-PDE data fit a two-component curve (Fig. 7B, solid line) significantly better than a one-component curve (Fig. 7B, dashed line). The Km and Vmax values for the two-component curve are Km1 =0.12 mM cAMP and Vmax1 =151.4 pmol/min/mg protein, and Km2 =4.4 mM cAMP and Vmax2 =237.4 pmol/min/mg protein. Vmax1 accounts for 39% of the total activity, and Vmax2 account for 61% of the total MM-PDE activity. The Km and Vmax values for the one-component curve are 0.62 mM cAMP and 342.2 pmol/min/mg protein, respectively.
3.7. 5 %-NT time course and dose response The time course and dose response of 5%-NT to AMP was examined in the adipocyte PM fraction. Approxi-
Fig. 7. PDE activity in the PM fraction (panel A) and MM fraction (panel B) of swine adipocytes in response to increasing concentrations of cAMP, after correcting for membrane purity. Membrane fraction preparation, data expression and experiment verification are as described in Fig. 1. Panel A is the cAMP dose response when 5 mg PM protein was pre-incubated with 5 mM cilostamide — the two-component curve (solid line-stars) fits the data significantly better than a one-component reaction curve (dashed line-stars). Also included is PDE activity when the PM fraction was corrected for microsomal contamination using CCR marker assay data (solid line-triangles)— the two-component curve does not fit the data significantly better than a one-component curve. Panel B is the cAMP dose response when 5 mg MM protein was assayed and corrected for PM contamination using 5%-NT marker assay data — the two-component curve (solid line-squares) fits significantly better than the single component reaction curve (dashed line).
68
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
Fig. 8. Time course of 5%-NT activity in the adipocyte PM fraction. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (4.5 mg) was incubated with 50 mM AMP for 30 – 120 min. Each point represents the mean9S.D. of four replicate samples. The findings were verified using a PM fraction prepared from adipocytes isolated from a different pig. Data were fit to linear regression analysis.
mM, the 5%-NT activity increased further to approximately 380 pmol/min/mg protein. The reason for this complex behavior is not yet understood.
3.8. 5 %-NT response to AMP-CP AMP-CP is a known 5%-NT inhibitor [25,26,40]. To determine if AMP-CP inhibits swine adipocyte PM 5’NT, 5 mg PM protein were pre-incubated with AMPCP for 7 min, then incubated with 50 mM AMP for 1 h. Fig. 10 shows that \ 700 mM AMP-CP inhibits \ 90% of the 5%-NT in swine adipocyte PM. The IC50 for AMP-CP was 47.3 mM, determined by plotting 5%-NT activity versus log mM AMP-CP. The Ki of AMP-CP was 4.22 mM and was determined using the equation
Fig. 9. 5%-NT AMP dose response in adipocyte PM. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (5 mg) was incubated with concentrations of AMP ranging from 10 to 750 mM for 3 h. Each point represents the mean 9 S.D. of four replicate samples. The findings were verified using a PM fraction prepared from adipocytes isolated from a different pig.
Fig. 10. Inhibition of 5%-NT activity in the PM fraction in response to AMPCP. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (5 mg) was pre-incubated with the indicated concentrations of AMPCP for 7 min, followed by the addition of 50 mM AMP for 1 h. Each point represents the mean9S.D. of four replicate samples. The results were verified with a PM fraction prepared from adipocytes isolated from a different pig. The line is a linear regression of the 5%-NT activity vs. log mM AMPCP and was used to determine the IC50.
Ki = IC50/(1+ ([S]/Km)) [8] where IC50 = 47.3mM AMPCP, [S]= 50 mM AMP, and Km = 4.9 mM AMP. Thus, AMP-CP appears to be an effective inhibitor of 5’NT activity in PM fractions.
3.9. Adipocyte plasma membrane metabolism of cAMP To evaluate the capacity of swine adipocyte plasma membranes to metabolize cAMP to AMP and subsequently to adenosine, 10 mg PM protein was incubated with 0.8 mM cAMP for 1–20 min. After halting the reaction, the contents of the tubes were divided in half to determine adenosine and AMP levels. AMP and adenosine levels increased linearly from 1 to 20 min, and at 20 min AMP accumulation was four times higher than that of adenosine (Fig. 11). Under these pre-steady state conditions when the low Km PDE is
Fig. 11. Time course of cAMP metabolism to AMP and adenosine in pig adipocyte PM. The PM fraction was prepared from adipocytes isolated from two different pigs on two separate days. The fractions were subsequently pooled for assay. PM was incubated with 0.8 mM cAMP for 1 – 20 min. The AMP and adenosine formed are expressed in picomoles formed per 5 mg protein, and each point represents the mean 9S.D. of three replicate samples.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
saturated with substrate, 5%-NT activity limits the production of adenosine.
4. Discussion Our results present three significant findings. First, the PDE associated with the PM fraction of swine adipocytes appears distinct from the PDE associated with the MM fraction, as only the latter is inhibited by cilostamide. Second, the xanthine derivative DPSPX, which indirectly has been shown to inhibit ecto-PDE in kidney [32] significantly inhibits PM-bound PDE. Third, when swine adipocyte PM are supplied with exogenous cAMP, both AMP and adenosine are produced, as has been described for brain [39], vascular smooth muscle cells [10] and kidney [19]. This latter finding has implications for a possible fate of cAMP exported from adipocytes: exported cAMP metabolized to adenosine may act in a negative feedback loop to bind to the A1 receptor on the adipocyte PM and inhibit adenylate cyclase, thereby preventing the overstimulation of adenylate cyclase. Although the existence of a negative feedback loop has not been explored in adipocytes, it has been shown to exist in brain and kidney [18, 39]. The finding that the PDE bound to the PM of swine adipocytes is cilostamide-insensitive, unlike the MMbound PDE3B, has not been previously reported. It is possible that a lipid or protein unique to the PM fraction binds cilostamide so it cannot interact with PM-PDE. However, other differences between adipocyte PM-PDE and MM-PDE activities have been reported. In subfractionated rat adipocytes, the PDE associated with the MM fraction showed significant stimulation by both insulin and isoproterenol [1], whereas PDE activity in the PM fraction was not altered by incubation with insulin or isoproterenol. Also, MM-bound PDE was nearly 100% inhibited by 2 mM cGMP, whereas PM-bound PDE decreased only 30%. In contrast, Macaulay et al. [23], reported that PM-PDE in rat adipocytes is insulin-sensitive but marker enzymes were not measured to correct for purity of the membrane fractions. PM-PDE and MMPDE sensitivity to cilostamide has been previously examined in brown adipocytes. A portion of both the PM and MM fractions showed cilostamide sensitivity [38], although, once again, membrane purity was not measured. Taken together with our results, data are accumulating to suggest that the PDE isoform in adipocyte PM may not be PDE3 as previously thought [3]. The PDE isoform found on the PM is unknown. Enoksson et al. [12] report that 41% of total PDE activity in adipose tissue may be due to PDE1, PDE2, PDE5, PDE6 PDE7 or to a ‘‘still unidentified subclass
69
of PDE’’. This unidentified subclass may include an ecto-PDE, sensitive to inhibition by DPSPX, a membrane-impermeable xanthine derivative. In the kidney, 1 mM DPSPX maximally inhibits conversion of exogenous 30 mM cAMP to AMP [32] but the possibility was not ruled out that intracellular PDE, which had leaked into the extracellular fluid, was being inhibited. We have also reported indirect evidence that DPSPX inhibits PM-bound PDE: incubation of intact adipocytes with forskolin plus 1 mM DPSPX causes a 2-fold increase in extracellular cAMP levels (from 0.01 to 0.02 mM) with no change in intracellular cAMP levels [13]. In addition, incubating intact adipocytes with forskolin plus 1 mM DPSPX and 1 mM cilostamide resulted in an additive effect on extracellular cAMP level [13]. Using the Ki for cilostamide generated from our experiments, we calculate that 1 mM cilostamide in the presence of 0.02 mM cAMP will completely block any leaked MMPDE, suggesting that the DPSPX does block the ‘ecto’ form of the PM-PDE. More direct evidence, however, is the present finding that 1 mM DPSPX causes a 73% inhibition of PM-PDE. The need for greater levels of DPSPX to inhibit the PM-PDE in the plasma membrane fraction compared to whole cells may be a consequence of the 40-fold greater concentration of cAMP in the PM experiments. These data, taken with the previous whole cell data from our laboratory, are consistent with the notion that PDE bound to the PM of swine adipocytes is outwardly-facing. There are several possible explanations for the lack of PM-PDE sensitivity to low micromolar levels of DPSPX. First, it is conceivable that during the membrane fractionation procedure, the PDE enzyme site is chemically or mechanically damaged. Second, albumin, a component of the Krebs–Ringer bicarbonate buffer used to isolate adipocytes [13], slows down the rate of cAMP degradation in liver cells, presumably by binding to cAMP and making it unavailable for metabolism [16]. The binding of cAMP to albumin could result in less cAMP competing with DPSPX for PM-PDE binding sites, making DPSPX appear more potent in whole cells and tissue than previously thought. Third, 1 mM DPSPX may be inhibiting MM-PDE that was released from leaky cells [13]; we did not examine inhibition of MM-PDE by DPSPX. Fourth, in intact cells DPSPX can antagonize the adenosine receptor [7], which could result in increased intracellular cAMP formation and its subsequent efflux. Thus, the 2-fold increase in extracellular cAMP seen by Finnegan and Carey [13] could be due to this antagonistic effect of DPSPX, and not to blocking ecto-PDE activity. Lastly, because adenosine receptors are directly coupled to adenylate cyclase and adenylate cyclase may serve to transport cAMP out of the cell [21], the possibility exists that blocking adenosine receptors with DPSPX could keep adenylate cyclase activated, which could result not only in ele-
70
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
vated intracellular cAMP levels, but also increased cAMP export as well. It should be pointed out that the analytical method used to quantitate the conversion of cAMP to AMP and subsequently to adenosine relies on the measurement of [3H]adenosine eluted from an exchange column. The presence of ecto-adenosine deaminase [14], adenosine kinase and AMP deaminase, which were not measured in our PM fractions, would potentially lead to an underestimation of that conversion in swine adipocytes. Cyclic AMP metabolism to adenosine has been shown to occur in a variety of tissues and whole cells including rat kidney, brain, and adipose [4,24,32,39]. In rat kidney, perfusing with cAMP resulted in the production of AMP, adenosine, and inosine, with levels of AMP triple that of adenosine [32], implying 5%-NT is limiting the production of adenosine. In the brain, incubating hippocampus slices with 20 mM cAMP resulted in levels of extracellular adenosine that act in a physiological manner to inhibit synaptic responses [4]. Brundege et al. [4] suggested that ecto-PDE may be the limiting enzyme in cAMP metabolism: administering AMP to the hippocampus resulted in higher levels of adenosine production than when cAMP was administered. Similar results were reported by Gorin and Brenner [16]: incubating adipose tissue with AMP resulted in a faster disappearance of AMP from the incubation medium than when tissue was incubated with cAMP, implying that PDE is the rate limiting enzyme under these conditions. Our data suggest that 5%-NT is limiting the production of adenosine in pre-steady state conditions, when the low Km PDE is saturated with substrate. The kinetic parameters for the two enzymes support this finding. Although 5%-NT has seven times the capactity to metabolize its substrate compared to the PDE, the 5%-NT is producing less total adenosine compared to AMP because its substrate is limiting, in relation to its Km. If the PDE could produce mM rather than nM quantities of AMP, the 5%-NT would be able to effectively convert it to adenosine. In this regard, physiological changes that might increase the metabolism of cAMP to adenosine could result from an increased capacity of the PDE or a decreased Km of the 5%-NT. Changes that may serve to alter the synthesis of adenosine may include diet and exercise. PDE activity in adipocytes has been shown to be altered by both diet and exercise: a diet consisting of higher levels of lipid, specifically linoleic acid, results in a lowering of swine adipocyte PM-PDE activity [3], while exercise training increases rat adipocyte PM-PDE activity [20]. Nicolas et al. [37] have shown that diet does not appear to alter swine adipocyte 5%-NT activity. In summary, these data demonstrate that cAMP can be converted to AMP and adenosine on the PM fraction of swine adipocytes by the PM-bound enzymes
5%-NT and PDE, and that the PM-PDE responsible for the extracellular cAMP metabolism to AMP appears distinct from the intracellular MM fraction PDE. Future studies will determine if cAMP can be metabolized to adenosine in whole cells, and will examine physiological conditions that may alter this metabolism.
Acknowledgements The authors gratefully acknowledge the assistance of Tom Oxford, UNH swine barn manager, for his care of the experimental subjects and Dr Rick Cote, UNH Associate Professor, for his willingness to share his laboratory and its expertise for the PDE assays, and his constructive comments on the manuscript. This work was completed by L.A. Zacher in partial fulfillment of the requirements for an MS degree. Portions of this work were reported at the 1998 Federation of the American Societies for Experimental Biology meeting in San Francisco, CA. Scientific contribution no. 2020 from the New Hampshire Agricultural Experiment Station.
References [1] Anderson NG, Kilgour E, Houslay MD. Subcellular localization and hormone sensitivity of adipocyte cyclic AMP phosphodiesterase. Biochem J 1989;262:867 – 72. [2] Avruch J, Wallach DF. Spectroscopic evidence for beta structure in rat adipocyte plasma membrane protein. Biochim Biophys Acta 1971;241:249 – 53. [3] Benmansour N, Demarne Y, Lecourtier MJ, Lhuillery C. Effects of castration, dietary fat and adipose tissue sites on adipocyte plasma-membranes cyclic AMP phosphodiesterase activity in the pig. Int J Biochem 1991;23:1205 – 11. [4] Brundege JM, Diao L, Proctor WR, Dunwiddie TV. The role of cyclic AMP as a precursor of extracellular adenosine in the rat hippocampus. Neuropharmacology 1997;36:1201 – 10. [5] Carey GB, Kaufmann J. Human adipocytes export cAMP. FASEB J 1998;12:256 Abstract. [6] Carey GB, Sidmore KA. Exercise attenuates the anti-lipolytic effect of adenosine in adipocytes isolated from miniature swine. Int J Obes Relat Metab Disord 1994;18:155 – 60. [7] Chen J, Dinger B, Fidone SJ. cAMP production in rabbit carotid body: role of adenosine. J Appl Physiol 1997;82:1771–5. [8] Cheng Y-C, Prusoff WH. Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50% inhibition (IC50) of an enzymatic reaction. Biochem Pharmacol 1973;22:3099 – 108. [9] Dallner G, Siekevitz P, Palade E. Biogenesis of endoplasmic reticulum membranes. II. Synthesis of constitutive microsomal enzymes in developing rat hepatocyte. J Cell Biol 1966;30:97– 117. [10] Dubey RK, Mi Z, Gillespie DG, Jackson EK. Cyclic AMPadenosine pathway inhibits vascular smooth muscle cell growth. Hypertension 1996;28:765 – 71. [11] Engfeldt P, Arner P, Ostman J. Influence of adipocyte isolation by collagenase on phosphodiesterase activity and lipolysis in man. J Lipid Res 1980;21:443 – 8.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71 [12] Enoksson S, Degerman E, Hagstrom-Toft E, Large V, Arner P. Various phosphodiesterase subtypes mediate the in vivo antilipolytic effect of insulin on adipose tissue and skeletal muscle in man. Diabetologia 1998;41:560–8. [13] Finnegan RB, Carey GB. Characterization of cyclic AMP efflux from swine adipocytes in vitro. Obes Res 1998;6:292–8. [14] Franco R, Mallol J, Casado V, et al. Ecto-adenosine deaminase: an ecto-enzyme and a costimulatory protein acting on a variety of cell surface receptors. Drug Dev Res 1998;45:261–8. [15] Gandarias JM, Santa Maria S, Galdiz B, Fernandez B. Activities and some properties of 5’ nucleotidase in isolated hamster fat cells. Comp Biochem Physiol B 1985;82:549–53. [16] Gorin E, Brenner T. Extracellular metabolism of cyclic AMP. Biochim Biophys Acta 1976;451:20–8. [17] Green A, Milligan G, Dobias SB. Gi down-regulation as a mechanism for heterologous desensitization in adipocytes. J Biol Chem 1992;267:3223–9. [18] Jackson EK. Adenosine: a physiological brake on renin release. Ann Rev Pharmacol Toxicol 1991;31:1–35. [19] Jackson EK, Mi Z, Gillespie DG, Dubey RK. Metabolism of cAMP to adenosine in the renal vasculature. J Pharm Exp Ther 1997;283:177 – 82. [20] Kenno KA, Durstine JL, Shepherd RE. Distribution of cyclic AMP phosphodiesterase in adipose tissue from trained rats. J Appl Physiol 1986;61:1546–51. [21] Krupinski J, Coussen F, Bakalyar HA, et al. Adenylyl cyclase amino acid sequence: possible channel- or transporter-like structure. Science 1989;244:1558–64. [22] Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265 – 75. [23] Macaulay SL, Kiechle FL, Jarett L. Comparison of phospholipid effects on insulin-sensitive low Km cyclic AMP phosphodiesterase in adipocyte plasma membranes and microsomes. Biochim Biophys Acta 1983;760:293–9. [24] Macmanus JP, Whitfield JF, Braceland B. The metabolism of exogenous cyclic AMP at low concentrations by thymic lymphocytes. Biochem Biophys Res Commun 1971;42:503 – 9. [25] Mallet RT, Sun J, Fan WL, Kang YH, Bunger R. Magnesium activated adenosine formation in intact perfused heart: predominance of ecto 5%-nucleotidase during hypermagnesemia. Biochim Biophys Acta 1996;1290:165–76. [26] Manganiello VC, Smith CJ, Degerman E, Belfrage P. Cyclic GMP-inhibited cyclic nucleotide phosphodiesterates. In: Beavo J, Houslay MD, editors. Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action. New York: Wiley, 1990:87 – 116. [27] Meghji P, Pearson JD, Slakey LL. Regulation of extracellular adenosine production by ectonucleotidases of adult rat ventricular myocytes. Am J Physiol 1992;263:40–7. [28] Mersmann HJ, Shparber A. Is the lipolytic response in porcine adipose tissue slices equivalent to the lipolytic response in isolated adipocytes? Int J Biochem 1993;25:1673–9. [29] Mersmann HJ, Phinney G, Brown LJ, Steffen DG. Ontogeny of adenylate cyclase and phosphodiesterase activities in swine tissues. Biol Neonate 1977;32:266–74.
.
71
[30] Mery PF, Fischmeister R. EHNA reverses the inhibitory effect of cGMP on cardiac L-type Ca current via PDE-II inhibition. FASEB J 1994;8 (Abstract). [31] Meservey C, Carey GB. Dietary fat saturation and endurance exercise alter lipolytic sensitivity of adipocytes isolated from Yucatan miniature swine. J Nutr 1994;124:2335 – 43. [32] Mi Z, Jackson EK. Metabolism of exogenous cyclic AMP to adenosine in the rat kidney. J Pharmacol Exp Ther 1995;273:728 – 33. [33] Nemecek GM, Honeyman TW. The role of cyclic nucleotide phosphodiesterase in the inhibition of cyclic AMP accumulation by carbachol and phosphatidate. J Cycl Nucleotide Res 1982;8:395 – 408. [34] Newby AC, Luzio JP, Hales CN. The properties and extracellular location of 5%-nucleotidase of the rat fat-cell plasma membrane. Biochem J 1975;146:625 – 33. [35] Newsholme EA, Blomstrand E, Newell J, Pitcher J. Maximal activities of enzymes involved in adenosine metabolism in muscle and adipose tissue of rats under conditions of variations in insulin sensitivity. FEBS Lett 1985;181:189– 92. [36] Nicolas C, Demarne Y, Lecourtier MJ, Lhuillery C. Dependence between 5%-nucleotidase activity and phosphatidylcholine content in pig adipocyte plasma membrane. Absence of relationship with membrane physical state. Comp Biochem Physiol B 1990; 96:195 – 9. [37] Nicolas C, Demarne Y, Lecourtier MJ, Lhuillery C. Effect of castration and anatomical site on the plasma membrane structure and the activities of adenylate cyclase and 5%-nucleotidase in pig adipocytes. Horm Metab Res 1991;23:522 – 9. [38] Omatsu-Kanbe M, Cushman SW, Manganiello VC, Taira M. Insulin stimulates hormone-sensitive cyclic GMP-inhibited cyclic nucleotide phosphodiesterase in rat brown adipose cells. FEBS Lett 1995;374:187 – 91. [39] Rosenberg PA, Dichter MA. Extracellular cAMP accumulation and degradation in rat cerebral cortex in dissociated cell culture. J Neurosci 1989;9:2654 – 63. [40] Schutz W, Schrader J, Gerlach E. Different sites of adenosine formation in the heart. Am J Physiol 1981;240:963 – 70. [41] Sheppard H, Wiggan G, Tsien WH. Structure-activity relationships for inhibitors of phosphodiesterase from erythrocytes and other tissues. Adv Cycl Nucleotide Res 1972;1:103 – 12. [42] Simpson IA, Yver DR, Hissin PJ, et al. Insulin-stimulated translocation of glucose transporters in the isolated rat adipose cells: characterization of subcellular fractions. Biochim Biophys Acta 1983;63:393 – 407. [43] Srere PA. Citrate synthase. Methods Enzymol 1969;13:3–5. [44] Thompson WJ, Brooker G, Appleman MM. Assay of cyclic nucleotide phosphodiesterases with radioactive substrates. Methods Enzymol 1974;38:205 – 12. [45] Vernon RG, Taylor E. Enzymes of adenosine metabolism of sheep adipose tissue: changes in activity with season, pregnancy and lactation. Horm Metab Res 1986;18:369 – 73. [46] Wells JN, Miller JR. Methylxanthine inhibitors of phosphodiesterases. Methods Enzymol 1988;159:489 – 96.
Cyclic AMP metabolism by swine adipocyte microsomal and plasma membranes L.A. Zacher, G.B. Carey * Department of Animal and Nutritional Sciences, Uni6ersity of New Hampshire, Durham, NH 03824, USA Received 17 February 1999; received in revised form 31 May 1999; accepted 8 June 1999
Abstract Extracellular cyclic AMP is source of extracellular adenosine in brain and kidney. Whether this occurs in adipose tissue is unknown. The present study evaluated the capacity of swine adipocyte plasma membranes to metabolize cyclic AMP to AMP and adenosine, via phosphodiesterase (PDE) and 5%-nucleotidase (5%-NT), respectively. Plasma membranes (PM) and microsomal membranes (MM) were isolated from over-the-shoulder subcutaneous adipose tissue of 3 month-old male miniature swine. The purity of the membrane fractions was determined and PDE and 5%-NT activities in PM and MM fractions were corrected for cross-contamination. The maximal activity of MM-PDE was 7-fold greater than that of PM-PDE. MM-PDE was 100% inhibited by 5 mM cilostamide, while PM-PDE was unaffected by this PDE3B inhibitor. Inhibitors of PDE1, PDE2, PDE4 and PDE5 also failed to inhibit PM-PDE. However, 1 mM DPSPX inhibited PM-PDE activity by 72%. When PM were incubated with 0.8 mM cyclic AMP for 20 min, AMP accumulation was four times that of adenosine. These data demonstrate that cyclic AMP can be converted to AMP and adenosine by the PM-bound enzymes 5%-NT and PDE, and suggest that the PM-PDE responsible for extracellular cyclic AMP metabolism to AMP is distinct from the intracellular MM-PDE. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Adenosine; Adipocytes; AMP; Cyclic AMP; Phosphodiesterase; Plasma membrane; Microsomal membrane; Swine
1. Introduction Adenosine inhibits lipolysis in adipocytes by binding to the A1 receptor and inhibiting adenylate cyclase [17] and adipocyte sensitivity to adenosine is altered when animals are subjected to modifications in diet and exercise [6,31]. However, the origin of extracellular adenosine in adipose tissue is unknown. Extracellular adenine nucleotides have been suggested as a source of extracellular adenosine in other tissues. In arterial smooth muscle cells and cardiac myocytes, the hydrolysis of extracellular ATP is a significant source of extracellular adenosine [27]. In brain cells, brain slices and perfused kidney, cyclic AMP (cAMP) is exported then hydrolyzed by the ‘ecto’ enzymes phosphodiesterase (PDE) and 5%-nucleotidase (5%-NT) to produce extracellular adenosine [4,32,39]. * Corresponding author. Tel.: +1-603-862-4628; fax: + 1-603-8623758. E-mail address: [email protected] (G.B. Carey)
Isolated adipocytes export cAMP [5,13], and intact adipose tissue has been shown to metabolize exogenous cAMP to extracellular adenosine [16]. If exported cAMP is metabolized to adenosine, a negative-feedback loop may exist in adipocytes where this adenosine could bind to the A1 receptor on the plasma membrane and inhibit adenylate cyclase activity, thereby decreasing formation and export of intracellular cAMP. Such a feedback loop was first described in 1991 in the juxtaglomerular cell of the kidney [18] and has since been described in brain [4,39]. Extracellular cAMP metabolism to adenosine would require the presence of outwardly facing ‘ecto’ enzymes PDE and 5%-NT on the surface of adipocytes. The existence of the plasma membrane (PM)-bound 5%-NT in adipocytes is well documented [15,34–37,45]. However, research on the PM-bound PDE in adipocytes is limited. The majority of PDE research has focused on intracellular, microsomal membrane (MM)-bound PDE. Adipocyte MM-PDE is sensitive to certain hormones [1] as well as to exercise treatment of animals
0305-0491/99/$ - see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 3 0 5 - 0 4 9 1 ( 9 9 ) 0 0 0 9 8 - X
62
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
[20]. PM-bound PDE has been identified in adipocytes, but research on this enzyme has focused primarily on its activity in basal states [24]. PM-bound PDE activity is altered by treating isolated adipocytes with phospholipids [23], and is dependent on the site of adipose tissue deposition as well as the diet fed to the animal [3]. The possibility that extracellular adenosine in adipose tissue arises from extracellular cAMP was explored in the present study. The goal of this study was to determine whether the PDE bound to the PM of swine adipocytes is distinct from the MM-bound PDE, and to evaluate the capacity of swine adipocyte plasma membranes to metabolize cAMP to AMP and adenosine. We demonstrate that the PM-bound PDE is cilostamide-insensitive, suggesting that it is distinct from the cilostamide-sensitive MM-bound PDE. We also show that although the maximum activity of PM5%NT is approximately 7-fold higher than that of the PM-PDE, the 5%-NT limits adenosine production under pre-steady state conditions in which the low Km PDE is saturated with substrate. Our results are consistent with the idea that a pathway for the negative-feedback regulation of the cAMP second messenger pathway may exist in adipocytes.
2. Materials and methods
2.1. Animals Yucatan miniature swine (Sus scrofa) were bred at the University of New Hampshire’s Burley Demeritt swine facility. All swine were 2.5 – 3.5 month old males, with a mean body weight9S.D. of 11.139 1.59 kg. The swine were housed with their littermates and were group-fed miniature swine ration (Agway, Syracuse, NY) twice a day with free access to water. All procedures were approved by the Institutional Animal Care and Use Committee.
2.2. Tissue biopsy, adipocyte isolation, and membrane fractionation Adipose tissue was removed from subcutaneous overthe-shoulder adipose tissue deposits after the animals were euthanized. The adipocytes were isolated as previously described with modifications [6]. Collagenase D was chosen because of its relatively low trypsin levels compared to other types of collagenase; using collagenase with high trypsin levels to isolate fat cells has been shown to decrease phosphodiesterase activity [11]. The adipocyte membranes were fractionated using the method of Simpson et al. [42] with modifications. Briefly, isolated adipocytes were washed twice in a 10 mM Tris–HCl–255 mM sucrose – 2 mM ngCl2 –0.1 mM phenylmethylsulfonyl fluoride (PMSF) homogeniz-
ing medium at 37°C to remove any Krebs–Ringer buffer. Adipocytes were then resuspended in 2 volumes of the same homogenizing medium, and homogenized with 3 up-and-down strokes in a 55 ml size C Teflon pestle homogenizer (Arthur H. Thomas and Co.). All subsequent steps took place at 4°C. De-fatted homogenate was centrifuged for 15 min at 16 000×g and the resulting pellet (containing PM) was resuspended in 20 ml homogenizing medium and centrifuged for 15 min at 16 000 × g. The resulting pellet was resuspended in 5 ml homogenizing medium, placed on 5 ml of a 1.12 M sucrose cushion containing 20 mM Tris–HCl–0.1 mM PMSF, and centrifuged at 101 000× g for 70 min. The plasma membranes were removed from the interface using a transfer pipette, and the mitochondria, nuclei, and cell debris were collected as a pellet. The mitochondrial pellet and the plasma membranes were resuspended in 30 ml homogenizing medium and centrifuged for 30 min at 30 000× g. The plasma membrane and mitochondrial pellets were resuspended in buffer containing 40 mM Tris–HCl–2.5 mM MgCl2 –1 mM EGTA–0.3 mM PMSF–0.001 mg leupeptin (resuspension buffer) to a volume of approximately 2 mg protein/ml, and approximately 4 mg protein/ml, respectively. The initial supernatant from the first 16 000×g spin was centrifuged for 70 min at 212 000× g. The pellet (containing microsomes) was resuspended in resuspension buffer to a final volume of approximately 2 mg protein/ml. Protein was determined by the method of Lowry et al. [22]. All cell fractions were immediately frozen upon preparation and stored at −80°C. Membrane marker enzyme assays were performed within 1 week; PDE and 5%-NT assays were generally performed within 6 weeks. The stability of PDE activity was verified in the membranes stored up to 6 weeks. In a small subset of experiments PDE was assayed from 8 weeks to 5 months after preparation of membranes. Although we did not verify PDE stability in the same preparation over 5 months, there was no evidence of a loss of PDE activity during this time. 5%-NT stability was not assessed.
2.3. Membrane marker enzyme assays 5%-NT marker activity was assayed as described [2] in the presence of 0.05% Triton X-100. Cytochrome c reductase activity was measured by the method of Dallner et al. [9]. Citrate synthase was assayed by the method of Srere [43].
2.4. Experimental incubations PDE activity was measured by the method of Thompson [44]. Plasma or microsomal membranes (5 mg ) in resuspension buffer were incubated at 37°C at a
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
final volume of 50 ml. Either 0.8 mM or 20 mM [3H]cAMP was added to start the reaction. In some cases, PDE activity was inhibited by pre-incubation with cilostamide (ranging in concentrations from 0.05 to 10 mM) at 37°C for 2 min prior to the addition of cAMP. Cilostamide was prepared in DMSO, and the final DMSO concentration did not exceed 0.2% (v/v). Other PDE inhibitors used were 1,3-dipropyl-8-p-sulfophenyl-xanthine (DPSPX), rolipram, erythro-9-(2-hydroxy-3-nonyl)adenine hydrochloride (EHNA), 8-methoxymethyl-3-isobutyl-1-methylxanthine (8methyl-IBMX) and zaprinast. 8-Methyl-IBMX was prepared in 100% ethanol, and the final ethanol concentration did not exceed 0.5% (v/v). Zaprinast was prepared in 100% methanol, and the final methanol concentration did not exceed 0.5% (v/v). DPSPX, rolipram, and EHNA were prepared in resuspension buffer. Reactions were halted by the addition of 100 ml 0.1 M HCl, and neutralized by the subsequent addition of 0.1 M Trizma base. Snake venom (croatus atrox) (10 ml) was added to each vial to convert any AMP formed to adenosine. Less than 50% of the substrate was hydrolyzed during the incubations, and the rates of adenosine formation remained linear for 20 min. 5%-NT activity was measured in the same system as the PDE (as distinct from the marker assay of Ref. [2]). Plasma membranes (5 mg) were used, and resuspension buffer and 50 mM AMP were added to yield a final volume of 50 ml. 5%-NT activity was inhibited using a,b-methyleneadenosine-5%-diphosphate (AMP-CP), ranging in concentration from 50 to 900 mM. AMP-CP, dissolved in resuspension buffer, was added to the plasma membranes and allowed to incubate at 37°C for 7 min prior to the addition of [3H]AMP. After a pre-determined time, the reactions were halted by boiling for exactly 1 min. Chromatography columns for the PDE assay were filled with 1 ml DEAE-Sephadex A25 resin, and equilibrated with 8 ml high salt buffer (20 mM Tris – HCl, 0.5 M NaCl, pH 6.8) followed by 8 ml low salt buffer (20 mM Tris–HCl, pH 6.8) for the PDE assays. Columns for the 5%-NT assays were filled with 3 ml DEAE-Sephadex A25 resin, and equilibrated with 25 ml high salt buffer (pH 6.8) followed by 25 ml low salt buffer (adjusted to pH 8.4). These volumes of DEAE-Sephadex ensured that less than 1.12% unreacted [3H]cAMP or [3H]AMP flowed through the columns when washed with the low salt buffer, and that all [3H]cAMP or [3H]AMP flowed through the columns when washed with the high salt buffer. The samples were applied to the DEAE-Sephadex columns and washed with 2 ml low salt buffer (pH 6.8) for the PDE samples, or 6 ml low salt buffer (pH 8.4) for the NT samples, and the eluate was collected. Scintillation cocktail in the amount twice that of the low salt washes was added to each vial, and the vials were vortexed and
63
counted in a liquid scintillation counter for radioactive adenosine. Graphpad Prism™ (Graphpad Software, San Diego, CA) was used for non-linear curve fitting. Goodness-offit was determined by calculating a P-value from the F ratio (the relative increase in sum-of-squares and the relative increase in degrees of freedom, for each curve being considered). Statistical significance was set at PB 0.05.
2.5. Chemicals Cilostamide was purchased from Tocris Cookson, St. Louis, MO. The columns were purchased from Evergreen Scientific. Collagenase D (lot c 83615722) was purchased from Boehringer Mannheim Biochemicals, Indianapolis, IN. DPSPX, EHNA, and rolipram were purchased from Research Biochemicals International, Natick, MA. All other chemicals and supplies were supplied by Sigma, St. Louis, MO or Fisher Scientific, Springfield, NJ.
3. Results Previous work has shown that the PDE enzyme in human adipocytes consists of two isoforms, one that hydrolyzes relatively low levels of substrate—referred to as low Km PDE—and one that hydrolyzes higher levels of substrate—referred to as high Km PDE [11]. PDE activity from swine adipose tissue also demonstrates a biphasic nature consistent with the notion that this tissue contains two isoforms [29]. We assayed for both isoforms: the low Km PDE was assayed at B1 mM cAMP and relatively short time points, and the high Km PDE was assayed at \ 1 mM cAMP and longer time points. Data for the low and high Km assays were combined for non-linear curve fitting analyses.
3.1. PDE time course To evaluate PDE activity over time, 4.5 mg PM or MM protein were incubated for 1–20 min with 0.8 mM cAMP to assess low Km PDE activity, or incubated with 20 mM cAMP for 5–90 min to assess high Km PDE activity. Levels of cAMP that saturate the low and high Km activities were determined by dose response experiments described in the next section. Low Km PDE activity in PM and MM fractions remained linear for up to 20 min and hydrolyzed 21 and 47% substrate, respectively, at 20 min (Fig. 1A). High Km PDE activity in PM and MM remained linear for up to 90 min and hydrolyzed 12 and 26% substrate, respectively, at 90 min (Fig. 1B).
64
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
3.2. PDE dose response to cAMP To determine a dose of cAMP that maximally stimulates low Km PDE, 5 mg of PM-protein or MM-protein were incubated for 12 and 6 minutes, respectively, with cAMP concentrations ranging from 0.05 to 0.80 mM. PM and MM were also incubated for 60 min with cAMP concentrations ranging from 2 to 26 mM to measure high Km PDE activity. Non-linear curve fitting analysis of PM-PDE activity over the entire cAMP concentration range (0.05 – 26 mM) showed that the data fit a two-component curve significantly better than a one-component curve, implying the existence of two isozymes (Fig. 2). The Km and Vmax values for the two-component curve are Km1 =0.15 mM cAMP and
Fig. 2. PDE activity in the PM fraction of swine adipocytes in response to increasing concentrations of cAMP. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: PM protein (5 mg) was incubated for 12 min at 37°C with doses of cAMP ranging from 0.05 to 0.8 mM and for 60 min at doses ranging from 2 to 26 mM cAMP. The dotted line represents the fit of data to a one-component curve, and the solid line represents the fit of data to a two-component curve, using the software Graphpad Prism™. The solid line for the two-component curve fits significantly better than the one-component curve.
Vmax1 = 54.1 pmol/min/mg protein, and Km2 = 2.0 mM cAMP and Vmax2 = 106.4 pmol/min/mg protein. Vmax1 accounts for 34% of the total activity, and Vmax2 accounts for 66% of the total activity. The Km and Vmax values for the PM one-component curve are 0.66 mM cAMP and 152.8 pmol/min/mg protein, respectively. Non-linear curve fitting of MM-PDE activity data also implied the existence of two isozymes (Fig. 3). The Km and Vmax values for the two-component curve are
Fig. 1. Panel A: time course of low Km PDE activity in adipocyte PM fraction (circles) and MM fraction (triangles). Membrane fractions were prepared from adipocytes isolated from three different pigs on three separate days. Fractions were subsequently pooled for assay. PM or MM protein (4.5 mg) was incubated with 0.8 mM cAMP for the indicated times. Values represent the mean9 S.D. for four replicate samples from the pool. The results were verified using membrane fractions prepared from a single pig. Panel B: time course of high Km PDE activity in adipocyte PM fraction (circles) and MM fraction (triangles). PM or MM (4.5 mg protein) was incubated with 20 mM cAMP for the indicated times. Membrane fraction preparation and pooling, data expression, and experiment verification are the same as described for Panel A.
Fig. 3. PDE activity in the MM fraction of swine adipocytes in response to increasing concentrations of cAMP. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: MM protein (5 mg) was incubated for 6 min at 37°C with doses of cAMP ranging from 0.05 to 0.8 mM and for 60 min at doses ranging from 2 to 26 mM cAMP. The dotted line represents the fit of the data to a one-component curve, and the solid line represents the fit of the data to a two-component curve, using the software Graphpad Prism™. The solid line for the two-component curve fits significantly better than the one-component curve.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
65
Table 1 Purity of fractions obtained from differential centrifugation of swine adipocytesa Fraction
Homogenate
5%-NT (nmol/h/ mg protein)
21.09 0.61 (2564) PM 56.6 9 1.5 (300) MM 50.8 9 1.4 (314) mitochondria, nuclei and 41.69 0.8 (385) cell debris Cytosol 0 (0)
Cytochrome c reductase (mmol/min/mg protein)
Citrate synthase (mmol/min/mg protein)
Total protein (mg)
0.369 0.01 (43.6)
0.15 90.01 (18.8)
122
0.669 0.03 (3.5) 1.969 0.03 (12.1) 0.679 0.01 (6.2)
0.07 90.01 (0.4)95.3 0.13 90.01 (0.8) 1.16 90.09 (10.8)
0.02 90.01 (1.1)
0.03 90.01 (1.6)
6.2 9.3 54
a
The membrane fractions were prepared from adipocytes isolated from three different pigs on three separate days. The fractions were subsequently pooled and the single pool was assayed for marker enzymes to determine membrane purity. The values are mean 9 S.D. of enzyme activity for three replicate samples from the pool. The values in parentheses represent the total enzyme activity in each fraction: nmol/h for 5%-NT and mmol/min for cytochrome c reductase and citrate synthase. The 5%-NT assay was conducted by incubating 20 mg protein with 400 mM [3H]AMP for 1 h. The cytochrome c reductase and citrate synthase assays were conducted using 5 mg membrane protein.
Km1 =0.13 mM cAMP and Vmax1 =138.6 pmol/min/mg protein, and Km2 =3.3 mM cAMP and Vmax2 =306.4 pmol/min/mg protein. Vmax1 accounts for 31% of the total activity, and Vmax2 accounts for 69% of the total activity. The Km and Vmax values for the one-component curve are 0.82 mM cAMP and 405.8 pmol/min/mg protein, respectively. The maximum activity of PDE in the adipocyte PM fraction is approximately 161 pmol/min/mg protein; this is approximately one-third that in the MM fraction. However, neither the PM nor the MM fractions are pure. To more accurately determine the kinetic parameters of the PDE in each fraction, fraction purity was determined.
3.3. Cross-contamination of adipocyte fractions Attempts were made to prepare the purest PM fraction possible by optimizing centrifugation speeds, buffer components, and buffer concentrations. MM contamination of the PM fraction was reduced from approximately 70% contamination to approximately 30% contamination by this optimization. We then used three enzymes as markers for PM, MM, and mitochondrial membranes (Table 1). These markers were 5%-NT for the PM fraction, cytochrome c reductase for the MM, and citrate synthase for the mitochondrial membranes. Using the 5%-NT marker to determine the amount of PM present in each fraction, we found the PM fraction had been enriched almost 3-fold from homogenate levels, however, similar enrichments of 5%NT were found in the mitochondrial and MM fractions. The similarity of 5%-NT enrichment in these three fractions questions the value of 5%-NT as a plasma membrane marker, as pointed out by Mersmann and Shparber [28]. Cytochrome c reductase showed a 5-fold enrichment in the MM fraction, and only one-third of the cytochrome c reductase activity in the MM fraction was present in the PM and mitochondrial fractions.
The fractionation procedure was very effective at removing mitochondrial membranes from MM and PM fractions, as citrate synthase was only 6 and 11%, respectively, of the activity present in the mitochondrial membrane fraction. The cytosol fraction had negligible levels of the three marker enzymes. Recovery of homogenate enzyme activity, determined by summing total enzyme activity in each fraction, was 39% 5%-NT, 53% cytochrome c reductase and 72% citrate synthase. Cross-contamination data were used to calculate uncontaminated PM-PDE and MM-PDE activities.
3.4. Microsomal and plasma membrane PDE response to cilostamide PDE activity has been reported in PM and MM fractions of isolated adipocytes [1,23] and the MMPDE is known to be completely inhibited by cilostamide [3]. To compare the cilostamide sensitivity of PDE in the PM and MM fractions, the membranes were incubated with concentrations of cilostamide ranging from 0.04 to 10 mM (Fig. 4). In the adipocyte MM fraction, PDE activity was maximally inhibited by 5 mM cilostamide (Fig. 4, dashed line), reducing PDE to 10% of its maximal activity (to 19.4 from 189.0 pmol/min/mg protein). Cilostamide also inhibited the PM fraction, reducing PDE to 32% of its maximal activity (to 26.5 from 84.0 pmol/min/mg protein; Fig. 5, dashed line). The remainder of the PDE activity (26.5 pmol/min/mg protein) is presumably PM-PDE that is insensitive to cilostamide. The cilostamide-insensitive PDE activity in the MM fraction (19.4 pmol/min/mg protein) can be completely attributed to PDE activity from plasma membrane contamination. PM-PDE activity was determined using data from the 5%-NT marker assay (89% contamination of MM fraction by PM= 0.89× 26.5 pmol/min/mg protein= 23.6 pmol/min/mg protein). After correcting for PM contamination, the microsomal membrane PDE
66
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
([S]/Km)) [8] where IC50 = 0.063 mM cilostamide, [S]= 0.8 mM cAMP, and Km = 0.13 mM cAMP. PM-PDE response to cilostamide was calculated by subtracting PDE activity due to MM contamination, using cytochrome c reductase marker assay data (33.5% of the cytochrome c reductase activity in the MM fraction was present in the PM fraction). At each cilostamide concentration, 33.5% of the corresponding MM fraction PDE activity was subtracted from the PM fraction PDE activity. Correcting for microsomal contamination of the PM fraction, PM-PDE showed no inhibition by cilostamide (Fig. 5, solid line). Fig. 4. Low Km PDE activity in MM in response to increasing concentrations of cilostamide. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: MM protein (5 mg) was pre-incubated with cilostamide (0.04 – 10 mM) for 2 min, followed by the addition of 0.8 mM cAMP for 6 min. The dashed line represents total MM fraction activity and the solid line represents calculated MM-PDE activity in the fraction after correcting for PM contamination using 5%-NT marker assay data (Table 1). Lines are fit on a point-to-point curve. Inset is a linear regression of the MM-PDE activity after correcting for PM contamination vs. log nM cilostamide; these data were used to determine the IC50.
(MM-PDE) showed complete inhibition by cilostamide by a dose of 5 mM (Fig. 4, solid line). Plotting PDE activity versus the log of cilostamide concentration, we determined the IC50 of cilostamide in the microsomal membrane fraction (after correcting for contamination) to be 63 nM (inset). The Ki of cilostamide was 8.8 nM, and was determined using the equation Ki =IC50/(1+
3.5. Plasma membrane PDE acti6ity in the presence of other known PDE inhibitors Five known PDE inhibitors were examined for their effectiveness to inhibit adipocyte PM-PDE activity. These included 8-methyl-IBMX to inhibit PDE1 [46], EHNA to inhibit PDE2 [30], rolipram to inhibit PDE4 [41], zaprinast to inhibit PDE5 [33], and DPSPX to inhibit the ecto-PDE [32]. After a 2 min pre-incubation with 5 mM cilostamide to inhibit any MM-PDE contaminating the PM fraction, 5 mg of PM protein were pre-incubated with either 1, 100, or 1000 mM DPSPX, 52 mM 8-methyl-IBMX, 10 mM EHNA, 10 or 100 mM rolipram, or 7.6 mM zaprinast. These concentrations are 5–100-fold above the published IC50 values for each inhibitor [30,32,33, 41,46]. The reactions were started with the addition of 0.8 mM cAMP, and were stopped after 12 min. The inhibitors 8-methyl-IBMX, EHNA, rolipram, and zaprinast showed no inhibition of PM-PDE (data not shown). Although 1 mM DPSPX failed to inhibit PMPDE, 100 and 1000 mM DPSPX significantly reduced PM-PDE activity by 25 and 73%, respectively (Fig. 6).
3.6. Uncontaminated PDE dose response to cAMP
Fig. 5. PDE PM activity in response to increasing concentrations of cilostamide. Membrane fraction preparation, data expression and experiment verification are identical to that described in Fig. 1, Panel A: PM protein (5 mg) was pre-incubated with cilostamide (0.04 – 10 mM) for 2 min, followed by the addition of 0.8 mM cAMP for 12 min (n= 4). Dashed line (triangles) represents total PM fraction activity and is fit to a one-component curve. The solid line (circles) represents the calculated PM-PDE activity after correcting for microsomal contamination, using cytochrome c reductase marker assay data, and is fit to a linear regression model.
To evaluate PM-PDE enzyme kinetics in the absence of contaminating MM-PDE, the PM fraction was preincubated with cilostamide to completely inhibit any MM-PDE activity present. The cAMP dose response of PDE in the PM fraction in the presence of cilostamide showed that the two-component curve (Fig. 7A, upper solid line) fit significantly better than the one-component curve (Fig. 7A, dashed line). The Km and Vmax values for the two-component curve (PM+ cilostamide) are Km1 = 0.12 mM cAMP and Vmax1 = 23. 8 pmol/min/ mg protein, and Km2 = 5.0 mM cAMP and Vmax2 =30.3 pmol/min/mg protein. Vmax1 accounts for 44% of the total activity and Vmax2 accounts for 56% of the total PM-PDE activity. The Km and Vmax values for the one-component curve are 0.49 mM cAMP and 45.1 pmol/min/mg protein, respectively.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
67
mately 4.5 mg PM protein was incubated for 30–120 min with 50 mM AMP. Fig. 8 shows that PM activity remained linear for up to 120 min and hydrolyzed 31% of substrate at the end of this time. When 5 mg PM protein was incubated with increasing concentrations of AMP, the activity of the enzyme 5%-NT was biphasic (Fig. 9). An initial plateau at approximately 200 pmol/min/mg protein was seen between 50 and 250 mM AMP; the Km for this portion was at 4.9 mM AMP. At AMP concentrations \500
Fig. 6. PM-PDE activity in the presence of increasing doses of DPSPX. Membrane fractions were prepared from adipocytes isolated from two different pigs on two separate days. The fractions were subsequently pooled for assay. PM protein (5 mg) was pre-incubated with 5 mM cilostamide for 2 min, followed by a 7 min pre-incubation with 0, 1, 100 or 1000 mM DPSPX. The membranes were then incubated with 0.8 mM cAMP for 12 min. Values represent the mean9 S.D. for four replicate samples from the pool. * Significantly different from control (0 mM DPSPX).
When PM fractions were incubated with increasing doses of cAMP, and PDE activity was corrected for MM-PDE contamination, the data were more variable, especially at the higher cAMP doses (Fig. 7). The two-component curve model did not fit significantly better than the one-component curve model. The latter model revealed a Km and Vmax 0.42 mM cAMP and 37.2 pmol/min/mg protein, respectively (Fig. 7A, lower solid line). When MM fractions were incubated with increasing doses of cAMP, and PDE activity was corrected for PM-PDE contamination, MM-PDE data fit a two-component curve (Fig. 7B, solid line) significantly better than a one-component curve (Fig. 7B, dashed line). The Km and Vmax values for the two-component curve are Km1 =0.12 mM cAMP and Vmax1 =151.4 pmol/min/mg protein, and Km2 =4.4 mM cAMP and Vmax2 =237.4 pmol/min/mg protein. Vmax1 accounts for 39% of the total activity, and Vmax2 account for 61% of the total MM-PDE activity. The Km and Vmax values for the one-component curve are 0.62 mM cAMP and 342.2 pmol/min/mg protein, respectively.
3.7. 5 %-NT time course and dose response The time course and dose response of 5%-NT to AMP was examined in the adipocyte PM fraction. Approxi-
Fig. 7. PDE activity in the PM fraction (panel A) and MM fraction (panel B) of swine adipocytes in response to increasing concentrations of cAMP, after correcting for membrane purity. Membrane fraction preparation, data expression and experiment verification are as described in Fig. 1. Panel A is the cAMP dose response when 5 mg PM protein was pre-incubated with 5 mM cilostamide — the two-component curve (solid line-stars) fits the data significantly better than a one-component reaction curve (dashed line-stars). Also included is PDE activity when the PM fraction was corrected for microsomal contamination using CCR marker assay data (solid line-triangles)— the two-component curve does not fit the data significantly better than a one-component curve. Panel B is the cAMP dose response when 5 mg MM protein was assayed and corrected for PM contamination using 5%-NT marker assay data — the two-component curve (solid line-squares) fits significantly better than the single component reaction curve (dashed line).
68
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
Fig. 8. Time course of 5%-NT activity in the adipocyte PM fraction. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (4.5 mg) was incubated with 50 mM AMP for 30 – 120 min. Each point represents the mean9S.D. of four replicate samples. The findings were verified using a PM fraction prepared from adipocytes isolated from a different pig. Data were fit to linear regression analysis.
mM, the 5%-NT activity increased further to approximately 380 pmol/min/mg protein. The reason for this complex behavior is not yet understood.
3.8. 5 %-NT response to AMP-CP AMP-CP is a known 5%-NT inhibitor [25,26,40]. To determine if AMP-CP inhibits swine adipocyte PM 5’NT, 5 mg PM protein were pre-incubated with AMPCP for 7 min, then incubated with 50 mM AMP for 1 h. Fig. 10 shows that \ 700 mM AMP-CP inhibits \ 90% of the 5%-NT in swine adipocyte PM. The IC50 for AMP-CP was 47.3 mM, determined by plotting 5%-NT activity versus log mM AMP-CP. The Ki of AMP-CP was 4.22 mM and was determined using the equation
Fig. 9. 5%-NT AMP dose response in adipocyte PM. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (5 mg) was incubated with concentrations of AMP ranging from 10 to 750 mM for 3 h. Each point represents the mean 9 S.D. of four replicate samples. The findings were verified using a PM fraction prepared from adipocytes isolated from a different pig.
Fig. 10. Inhibition of 5%-NT activity in the PM fraction in response to AMPCP. The PM fraction was prepared from adipocytes isolated from one pig. PM protein (5 mg) was pre-incubated with the indicated concentrations of AMPCP for 7 min, followed by the addition of 50 mM AMP for 1 h. Each point represents the mean9S.D. of four replicate samples. The results were verified with a PM fraction prepared from adipocytes isolated from a different pig. The line is a linear regression of the 5%-NT activity vs. log mM AMPCP and was used to determine the IC50.
Ki = IC50/(1+ ([S]/Km)) [8] where IC50 = 47.3mM AMPCP, [S]= 50 mM AMP, and Km = 4.9 mM AMP. Thus, AMP-CP appears to be an effective inhibitor of 5’NT activity in PM fractions.
3.9. Adipocyte plasma membrane metabolism of cAMP To evaluate the capacity of swine adipocyte plasma membranes to metabolize cAMP to AMP and subsequently to adenosine, 10 mg PM protein was incubated with 0.8 mM cAMP for 1–20 min. After halting the reaction, the contents of the tubes were divided in half to determine adenosine and AMP levels. AMP and adenosine levels increased linearly from 1 to 20 min, and at 20 min AMP accumulation was four times higher than that of adenosine (Fig. 11). Under these pre-steady state conditions when the low Km PDE is
Fig. 11. Time course of cAMP metabolism to AMP and adenosine in pig adipocyte PM. The PM fraction was prepared from adipocytes isolated from two different pigs on two separate days. The fractions were subsequently pooled for assay. PM was incubated with 0.8 mM cAMP for 1 – 20 min. The AMP and adenosine formed are expressed in picomoles formed per 5 mg protein, and each point represents the mean 9S.D. of three replicate samples.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
saturated with substrate, 5%-NT activity limits the production of adenosine.
4. Discussion Our results present three significant findings. First, the PDE associated with the PM fraction of swine adipocytes appears distinct from the PDE associated with the MM fraction, as only the latter is inhibited by cilostamide. Second, the xanthine derivative DPSPX, which indirectly has been shown to inhibit ecto-PDE in kidney [32] significantly inhibits PM-bound PDE. Third, when swine adipocyte PM are supplied with exogenous cAMP, both AMP and adenosine are produced, as has been described for brain [39], vascular smooth muscle cells [10] and kidney [19]. This latter finding has implications for a possible fate of cAMP exported from adipocytes: exported cAMP metabolized to adenosine may act in a negative feedback loop to bind to the A1 receptor on the adipocyte PM and inhibit adenylate cyclase, thereby preventing the overstimulation of adenylate cyclase. Although the existence of a negative feedback loop has not been explored in adipocytes, it has been shown to exist in brain and kidney [18, 39]. The finding that the PDE bound to the PM of swine adipocytes is cilostamide-insensitive, unlike the MMbound PDE3B, has not been previously reported. It is possible that a lipid or protein unique to the PM fraction binds cilostamide so it cannot interact with PM-PDE. However, other differences between adipocyte PM-PDE and MM-PDE activities have been reported. In subfractionated rat adipocytes, the PDE associated with the MM fraction showed significant stimulation by both insulin and isoproterenol [1], whereas PDE activity in the PM fraction was not altered by incubation with insulin or isoproterenol. Also, MM-bound PDE was nearly 100% inhibited by 2 mM cGMP, whereas PM-bound PDE decreased only 30%. In contrast, Macaulay et al. [23], reported that PM-PDE in rat adipocytes is insulin-sensitive but marker enzymes were not measured to correct for purity of the membrane fractions. PM-PDE and MMPDE sensitivity to cilostamide has been previously examined in brown adipocytes. A portion of both the PM and MM fractions showed cilostamide sensitivity [38], although, once again, membrane purity was not measured. Taken together with our results, data are accumulating to suggest that the PDE isoform in adipocyte PM may not be PDE3 as previously thought [3]. The PDE isoform found on the PM is unknown. Enoksson et al. [12] report that 41% of total PDE activity in adipose tissue may be due to PDE1, PDE2, PDE5, PDE6 PDE7 or to a ‘‘still unidentified subclass
69
of PDE’’. This unidentified subclass may include an ecto-PDE, sensitive to inhibition by DPSPX, a membrane-impermeable xanthine derivative. In the kidney, 1 mM DPSPX maximally inhibits conversion of exogenous 30 mM cAMP to AMP [32] but the possibility was not ruled out that intracellular PDE, which had leaked into the extracellular fluid, was being inhibited. We have also reported indirect evidence that DPSPX inhibits PM-bound PDE: incubation of intact adipocytes with forskolin plus 1 mM DPSPX causes a 2-fold increase in extracellular cAMP levels (from 0.01 to 0.02 mM) with no change in intracellular cAMP levels [13]. In addition, incubating intact adipocytes with forskolin plus 1 mM DPSPX and 1 mM cilostamide resulted in an additive effect on extracellular cAMP level [13]. Using the Ki for cilostamide generated from our experiments, we calculate that 1 mM cilostamide in the presence of 0.02 mM cAMP will completely block any leaked MMPDE, suggesting that the DPSPX does block the ‘ecto’ form of the PM-PDE. More direct evidence, however, is the present finding that 1 mM DPSPX causes a 73% inhibition of PM-PDE. The need for greater levels of DPSPX to inhibit the PM-PDE in the plasma membrane fraction compared to whole cells may be a consequence of the 40-fold greater concentration of cAMP in the PM experiments. These data, taken with the previous whole cell data from our laboratory, are consistent with the notion that PDE bound to the PM of swine adipocytes is outwardly-facing. There are several possible explanations for the lack of PM-PDE sensitivity to low micromolar levels of DPSPX. First, it is conceivable that during the membrane fractionation procedure, the PDE enzyme site is chemically or mechanically damaged. Second, albumin, a component of the Krebs–Ringer bicarbonate buffer used to isolate adipocytes [13], slows down the rate of cAMP degradation in liver cells, presumably by binding to cAMP and making it unavailable for metabolism [16]. The binding of cAMP to albumin could result in less cAMP competing with DPSPX for PM-PDE binding sites, making DPSPX appear more potent in whole cells and tissue than previously thought. Third, 1 mM DPSPX may be inhibiting MM-PDE that was released from leaky cells [13]; we did not examine inhibition of MM-PDE by DPSPX. Fourth, in intact cells DPSPX can antagonize the adenosine receptor [7], which could result in increased intracellular cAMP formation and its subsequent efflux. Thus, the 2-fold increase in extracellular cAMP seen by Finnegan and Carey [13] could be due to this antagonistic effect of DPSPX, and not to blocking ecto-PDE activity. Lastly, because adenosine receptors are directly coupled to adenylate cyclase and adenylate cyclase may serve to transport cAMP out of the cell [21], the possibility exists that blocking adenosine receptors with DPSPX could keep adenylate cyclase activated, which could result not only in ele-
70
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71
vated intracellular cAMP levels, but also increased cAMP export as well. It should be pointed out that the analytical method used to quantitate the conversion of cAMP to AMP and subsequently to adenosine relies on the measurement of [3H]adenosine eluted from an exchange column. The presence of ecto-adenosine deaminase [14], adenosine kinase and AMP deaminase, which were not measured in our PM fractions, would potentially lead to an underestimation of that conversion in swine adipocytes. Cyclic AMP metabolism to adenosine has been shown to occur in a variety of tissues and whole cells including rat kidney, brain, and adipose [4,24,32,39]. In rat kidney, perfusing with cAMP resulted in the production of AMP, adenosine, and inosine, with levels of AMP triple that of adenosine [32], implying 5%-NT is limiting the production of adenosine. In the brain, incubating hippocampus slices with 20 mM cAMP resulted in levels of extracellular adenosine that act in a physiological manner to inhibit synaptic responses [4]. Brundege et al. [4] suggested that ecto-PDE may be the limiting enzyme in cAMP metabolism: administering AMP to the hippocampus resulted in higher levels of adenosine production than when cAMP was administered. Similar results were reported by Gorin and Brenner [16]: incubating adipose tissue with AMP resulted in a faster disappearance of AMP from the incubation medium than when tissue was incubated with cAMP, implying that PDE is the rate limiting enzyme under these conditions. Our data suggest that 5%-NT is limiting the production of adenosine in pre-steady state conditions, when the low Km PDE is saturated with substrate. The kinetic parameters for the two enzymes support this finding. Although 5%-NT has seven times the capactity to metabolize its substrate compared to the PDE, the 5%-NT is producing less total adenosine compared to AMP because its substrate is limiting, in relation to its Km. If the PDE could produce mM rather than nM quantities of AMP, the 5%-NT would be able to effectively convert it to adenosine. In this regard, physiological changes that might increase the metabolism of cAMP to adenosine could result from an increased capacity of the PDE or a decreased Km of the 5%-NT. Changes that may serve to alter the synthesis of adenosine may include diet and exercise. PDE activity in adipocytes has been shown to be altered by both diet and exercise: a diet consisting of higher levels of lipid, specifically linoleic acid, results in a lowering of swine adipocyte PM-PDE activity [3], while exercise training increases rat adipocyte PM-PDE activity [20]. Nicolas et al. [37] have shown that diet does not appear to alter swine adipocyte 5%-NT activity. In summary, these data demonstrate that cAMP can be converted to AMP and adenosine on the PM fraction of swine adipocytes by the PM-bound enzymes
5%-NT and PDE, and that the PM-PDE responsible for the extracellular cAMP metabolism to AMP appears distinct from the intracellular MM fraction PDE. Future studies will determine if cAMP can be metabolized to adenosine in whole cells, and will examine physiological conditions that may alter this metabolism.
Acknowledgements The authors gratefully acknowledge the assistance of Tom Oxford, UNH swine barn manager, for his care of the experimental subjects and Dr Rick Cote, UNH Associate Professor, for his willingness to share his laboratory and its expertise for the PDE assays, and his constructive comments on the manuscript. This work was completed by L.A. Zacher in partial fulfillment of the requirements for an MS degree. Portions of this work were reported at the 1998 Federation of the American Societies for Experimental Biology meeting in San Francisco, CA. Scientific contribution no. 2020 from the New Hampshire Agricultural Experiment Station.
References [1] Anderson NG, Kilgour E, Houslay MD. Subcellular localization and hormone sensitivity of adipocyte cyclic AMP phosphodiesterase. Biochem J 1989;262:867 – 72. [2] Avruch J, Wallach DF. Spectroscopic evidence for beta structure in rat adipocyte plasma membrane protein. Biochim Biophys Acta 1971;241:249 – 53. [3] Benmansour N, Demarne Y, Lecourtier MJ, Lhuillery C. Effects of castration, dietary fat and adipose tissue sites on adipocyte plasma-membranes cyclic AMP phosphodiesterase activity in the pig. Int J Biochem 1991;23:1205 – 11. [4] Brundege JM, Diao L, Proctor WR, Dunwiddie TV. The role of cyclic AMP as a precursor of extracellular adenosine in the rat hippocampus. Neuropharmacology 1997;36:1201 – 10. [5] Carey GB, Kaufmann J. Human adipocytes export cAMP. FASEB J 1998;12:256 Abstract. [6] Carey GB, Sidmore KA. Exercise attenuates the anti-lipolytic effect of adenosine in adipocytes isolated from miniature swine. Int J Obes Relat Metab Disord 1994;18:155 – 60. [7] Chen J, Dinger B, Fidone SJ. cAMP production in rabbit carotid body: role of adenosine. J Appl Physiol 1997;82:1771–5. [8] Cheng Y-C, Prusoff WH. Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50% inhibition (IC50) of an enzymatic reaction. Biochem Pharmacol 1973;22:3099 – 108. [9] Dallner G, Siekevitz P, Palade E. Biogenesis of endoplasmic reticulum membranes. II. Synthesis of constitutive microsomal enzymes in developing rat hepatocyte. J Cell Biol 1966;30:97– 117. [10] Dubey RK, Mi Z, Gillespie DG, Jackson EK. Cyclic AMPadenosine pathway inhibits vascular smooth muscle cell growth. Hypertension 1996;28:765 – 71. [11] Engfeldt P, Arner P, Ostman J. Influence of adipocyte isolation by collagenase on phosphodiesterase activity and lipolysis in man. J Lipid Res 1980;21:443 – 8.
L.A. Zacher, G.B. Carey / Comparati6e Biochemistry and Physiology, Part B 124 (1999) 61–71 [12] Enoksson S, Degerman E, Hagstrom-Toft E, Large V, Arner P. Various phosphodiesterase subtypes mediate the in vivo antilipolytic effect of insulin on adipose tissue and skeletal muscle in man. Diabetologia 1998;41:560–8. [13] Finnegan RB, Carey GB. Characterization of cyclic AMP efflux from swine adipocytes in vitro. Obes Res 1998;6:292–8. [14] Franco R, Mallol J, Casado V, et al. Ecto-adenosine deaminase: an ecto-enzyme and a costimulatory protein acting on a variety of cell surface receptors. Drug Dev Res 1998;45:261–8. [15] Gandarias JM, Santa Maria S, Galdiz B, Fernandez B. Activities and some properties of 5’ nucleotidase in isolated hamster fat cells. Comp Biochem Physiol B 1985;82:549–53. [16] Gorin E, Brenner T. Extracellular metabolism of cyclic AMP. Biochim Biophys Acta 1976;451:20–8. [17] Green A, Milligan G, Dobias SB. Gi down-regulation as a mechanism for heterologous desensitization in adipocytes. J Biol Chem 1992;267:3223–9. [18] Jackson EK. Adenosine: a physiological brake on renin release. Ann Rev Pharmacol Toxicol 1991;31:1–35. [19] Jackson EK, Mi Z, Gillespie DG, Dubey RK. Metabolism of cAMP to adenosine in the renal vasculature. J Pharm Exp Ther 1997;283:177 – 82. [20] Kenno KA, Durstine JL, Shepherd RE. Distribution of cyclic AMP phosphodiesterase in adipose tissue from trained rats. J Appl Physiol 1986;61:1546–51. [21] Krupinski J, Coussen F, Bakalyar HA, et al. Adenylyl cyclase amino acid sequence: possible channel- or transporter-like structure. Science 1989;244:1558–64. [22] Lowry OH, Rosenbrough NJ, Farr AL, Randall RJ. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193:265 – 75. [23] Macaulay SL, Kiechle FL, Jarett L. Comparison of phospholipid effects on insulin-sensitive low Km cyclic AMP phosphodiesterase in adipocyte plasma membranes and microsomes. Biochim Biophys Acta 1983;760:293–9. [24] Macmanus JP, Whitfield JF, Braceland B. The metabolism of exogenous cyclic AMP at low concentrations by thymic lymphocytes. Biochem Biophys Res Commun 1971;42:503 – 9. [25] Mallet RT, Sun J, Fan WL, Kang YH, Bunger R. Magnesium activated adenosine formation in intact perfused heart: predominance of ecto 5%-nucleotidase during hypermagnesemia. Biochim Biophys Acta 1996;1290:165–76. [26] Manganiello VC, Smith CJ, Degerman E, Belfrage P. Cyclic GMP-inhibited cyclic nucleotide phosphodiesterates. In: Beavo J, Houslay MD, editors. Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action. New York: Wiley, 1990:87 – 116. [27] Meghji P, Pearson JD, Slakey LL. Regulation of extracellular adenosine production by ectonucleotidases of adult rat ventricular myocytes. Am J Physiol 1992;263:40–7. [28] Mersmann HJ, Shparber A. Is the lipolytic response in porcine adipose tissue slices equivalent to the lipolytic response in isolated adipocytes? Int J Biochem 1993;25:1673–9. [29] Mersmann HJ, Phinney G, Brown LJ, Steffen DG. Ontogeny of adenylate cyclase and phosphodiesterase activities in swine tissues. Biol Neonate 1977;32:266–74.
.
71
[30] Mery PF, Fischmeister R. EHNA reverses the inhibitory effect of cGMP on cardiac L-type Ca current via PDE-II inhibition. FASEB J 1994;8 (Abstract). [31] Meservey C, Carey GB. Dietary fat saturation and endurance exercise alter lipolytic sensitivity of adipocytes isolated from Yucatan miniature swine. J Nutr 1994;124:2335 – 43. [32] Mi Z, Jackson EK. Metabolism of exogenous cyclic AMP to adenosine in the rat kidney. J Pharmacol Exp Ther 1995;273:728 – 33. [33] Nemecek GM, Honeyman TW. The role of cyclic nucleotide phosphodiesterase in the inhibition of cyclic AMP accumulation by carbachol and phosphatidate. J Cycl Nucleotide Res 1982;8:395 – 408. [34] Newby AC, Luzio JP, Hales CN. The properties and extracellular location of 5%-nucleotidase of the rat fat-cell plasma membrane. Biochem J 1975;146:625 – 33. [35] Newsholme EA, Blomstrand E, Newell J, Pitcher J. Maximal activities of enzymes involved in adenosine metabolism in muscle and adipose tissue of rats under conditions of variations in insulin sensitivity. FEBS Lett 1985;181:189– 92. [36] Nicolas C, Demarne Y, Lecourtier MJ, Lhuillery C. Dependence between 5%-nucleotidase activity and phosphatidylcholine content in pig adipocyte plasma membrane. Absence of relationship with membrane physical state. Comp Biochem Physiol B 1990; 96:195 – 9. [37] Nicolas C, Demarne Y, Lecourtier MJ, Lhuillery C. Effect of castration and anatomical site on the plasma membrane structure and the activities of adenylate cyclase and 5%-nucleotidase in pig adipocytes. Horm Metab Res 1991;23:522 – 9. [38] Omatsu-Kanbe M, Cushman SW, Manganiello VC, Taira M. Insulin stimulates hormone-sensitive cyclic GMP-inhibited cyclic nucleotide phosphodiesterase in rat brown adipose cells. FEBS Lett 1995;374:187 – 91. [39] Rosenberg PA, Dichter MA. Extracellular cAMP accumulation and degradation in rat cerebral cortex in dissociated cell culture. J Neurosci 1989;9:2654 – 63. [40] Schutz W, Schrader J, Gerlach E. Different sites of adenosine formation in the heart. Am J Physiol 1981;240:963 – 70. [41] Sheppard H, Wiggan G, Tsien WH. Structure-activity relationships for inhibitors of phosphodiesterase from erythrocytes and other tissues. Adv Cycl Nucleotide Res 1972;1:103 – 12. [42] Simpson IA, Yver DR, Hissin PJ, et al. Insulin-stimulated translocation of glucose transporters in the isolated rat adipose cells: characterization of subcellular fractions. Biochim Biophys Acta 1983;63:393 – 407. [43] Srere PA. Citrate synthase. Methods Enzymol 1969;13:3–5. [44] Thompson WJ, Brooker G, Appleman MM. Assay of cyclic nucleotide phosphodiesterases with radioactive substrates. Methods Enzymol 1974;38:205 – 12. [45] Vernon RG, Taylor E. Enzymes of adenosine metabolism of sheep adipose tissue: changes in activity with season, pregnancy and lactation. Horm Metab Res 1986;18:369 – 73. [46] Wells JN, Miller JR. Methylxanthine inhibitors of phosphodiesterases. Methods Enzymol 1988;159:489 – 96.