Characterization of r3H]Adenosine Binding to Fat Cell Membranes*

I >H IAdenosine binding to fat cell plasma membrane preparations was examined by a vacuum filtration technique. Appreciable binding of adenosine to fat cell membranes could only be demonstrated in the presence of the adenosine deaminase inhibitors, erythro-Y-(2-hydroxy-3-nonyljadenine or deoxycoformycin. The binding of adenosine to these membranes was rapid, reaching near equilibrium within 10 min at 37°C. and was reversible. Hound adenosine dissnciated very rapidly fnllnwing a lOO-fold dilution at 0°C or 22°C. Only 7% of equilibrium-bound adenosine remained at 2 min following IOO-fnld dilution at 22°C. Scatchard plots of equilibrium binding data were nonlinear, suggesting at least two populations of adenosine binding sites possessing different affinities for adennsine. The apparent dissociation cnnstants. K,, , and maximum binding capacities, B,,,,, , were 9.5 x 10 Ii M and 28 pmol/mg of protein and 9.5 X 10F1 M and 1700 pmol/mg nf protein for the high and low affinity adenosine binding sites, respectively. The high affinity sites display maximum binding at pH 7.5. above pH 7.5. or below pH 6.5 adennsine binding to the fat cell membranes declines sharply. The divalent cations calcium or magnesium, at concentrations greater than 1 mM. inhibit the binding of adennsine to the high affinity sites of the fat cell membranes. Adennsine binding is reduced by prior exposure of the membranes to trypsin, chymntrypsin, or neuraminidase or by thermal denaturatinn. Purine derivatives compete with adennsine for binding to the membrane site in the following potency order: adenine > ATP = AI)P > cyclic AMP = AMP = innsine. Inhibition studies of high affinity adennsine binding performed with these purine derivatives indicate this binding is to multiple components of differing specificity and not to a single, hnmngenenus class of binding sites. Thenphylline, but not dipyridamnle or p-nitrnbenzylthioguannsine. is a potent inhibitor of adennsine binding to the membranes. One micromolar thenphylline inhibits adennsine binding


Providence,
Rhode Island 02912 I >H IAdenosine binding to fat cell plasma membrane preparations was examined by a vacuum filtration technique. Appreciable binding of adenosine to fat cell membranes could only be demonstrated in the presence of the adenosine deaminase inhibitors, erythro-Y-(2-hydroxy-3-nonyljadenine or deoxycoformycin.
The binding of adenosine to these membranes was rapid, reaching near equilibrium within 10 min at 37°C. and was reversible. Hound adenosine dissnciated very rapidly fnllnwing a lOO-fold dilution at 0°C or 22°C. Only 7% of equilibrium-bound adenosine remained at 2 min following IOO-fnld dilution at 22°C. Scatchard plots of equilibrium binding data were nonlinear, suggesting at least two populations of adenosine binding sites possessing different affinities for adennsine. The apparent dissociation cnnstants. K,, , and maximum binding capacities, B,,,,, , were 9.5 x 10 Ii M and 28 pmol/mg of protein and 9.5 X 10F1 M and 1700 pmol/mg nf protein for the high and low affinity adenosine binding sites, respectively.
The high affinity sites display maximum binding at pH 7.5. above pH 7.5. or below pH 6.5 adennsine binding to the fat cell membranes declines sharply. The divalent cations calcium or magnesium, at concentrations greater than 1 mM. inhibit the binding of adennsine to the high affinity sites of the fat cell membranes.
Adennsine binding is reduced by prior exposure of the membranes to trypsin, chymntrypsin, or neuraminidase or by thermal denaturatinn. Purine derivatives compete with adennsine for binding to the membrane site in the following potency order: adenine > ATP = AI)P > cyclic AMP = AMP = innsine. Inhibition studies of high affinity adennsine binding performed with these purine derivatives indicate this binding is to multiple components of differing specificity and not to a single, hnmngenenus class of binding sites. Thenphylline, but not dipyridamnle or p-nitrnbenzylthioguannsine. is a potent inhibitor of adennsine binding to the membranes. One micromolar thenphylline inhibits adennsine binding 14%. Highly purified fat cell plasma membranes possess the major portion of adenosine binding sites identified with the crude fat cell plasma membranes. The highly purified frartion likewise displayed the two pnpulatinns of adenosine binding sites identified in the crude preparation.
This study demonstrates that fat cell plasma membranes possess sites which bind l"Hladennsine with high affinity.
The relationship between these adennsine binding sites and the influence of adenosine on cyclic AMP levels in intact fat cells and on adenylate cyclase activity is discussed. Dole (1) first demonstrated the ability of adenosine to inhibit epinephrine-stimulated free fatty acid release from fat pads. This observation has since been amply confirmed in adipose tissue (2, 3) and isolated fat cell suspensions (4, 5). Additionally, it has been shown that adenosine and adenosinecontaining compounds such as 5'-AMP, ADP, ATP, and NAD possess a capacity to inhibit glucagon-(6), ACTH-' (7). and norepinephrine-stimulated (3, 7) lipolysis in adipose tissue. Schwabe et al. (8) demonstrated that adenosine is continuously released to the medium by fat cell suspensions. Sattin and Rall (9) first reported the ability of adenosine to elevate adenosine 3':5'-monophosphate (cyclic AMP) levels in brain slices. Shimizu and associates (10,11) confirmed this observation. Subsequent reports have shown adenosine stimulation of cyclic AMP levels in platelets (12)(13)(14). isolated bone cells (15), and cultured cell lines (16-21). Adenosine stimulation of cyclic AMP levels in several clones of mouse neuroblastoma cells required, however, the presence of RO-20-1724, a cyclic AMP phosphodiesterase inhibitor (18). Adenosine also stimulates adenylate cyclase activity in cell membrane preparations (22-271, although some of these preparations display a biphasic stimulation and inhibition of the cyclase by adenosine (22,24,25). In contrast, epinephrine-and glucagonstimulated cyclic AMP accumulation are inhibited by adeno- The perchloric acid extracts were passed over columns of NoritiCelite, washed with water, and the adenosine and derivatives eluted with 50% aqueous pyridine according to Hepp et al. (41). The eluant of the pyridine wash was collected, lyophilized, and dissolved in 75 ~1 of water. Aliquots of this solution (5 ~1) were chromatographed in a single direction on a plastic sheet (20 x 20 cm) using polyethyleneimine-impregnated plates of cellulose MN 300 (Brinkmann CEL 300 PEI) (0.1 mM). The developing solvent was 1-butanol, glacial acetic acid, and water (50:25:251 or 1-butanol, formic acid, and water (77:13:10).
The chromatographic procedure using polyethyleneimine-cellulose plates and the 1-butanol:acetic acid resulted in very discrete spots and reproducible separations except for that of labeled adenine from adenosine. The spots were scraped off the plates with a razor blade and added to a counting vial containing 0.5 ml of 1 N HCl. The contents were mixed and 4 ml of a counting solution containing 1 part of Triton X-100 (Research Products Co.) and 2 parts of 5% Omnifluor (New England Nuclear) in toluene was then added. The vials were counted in a liquid scintillation counter for at least 10 min.

RESULTS
Initial experiments utilizing a rapid filtration assay failed to demonstrate any appreciable binding of 13H1adenosine to crude plasma membranes prepared from rat fat cells. As shown in Table  I. adenosine binding to fat cell membranes could only be demonstrated when an inhibitor of adenosine deaminase was present in the incubation medium. Erythro-9-(2-hydroxy-3-nonylladenine (EHNA). a potent adenosine deaminase inhibitor, at 1 to 100 PM increased the amount of adenosine bound by these membranes. Deoxycoformycin, another potent inhibitor of adenosine deaminase (42). similarly provided the condition necessary to observe 13H]adenosine binding to the membranes (Table I). Since the amount of adenosine deaminase activity of these membrane preparations varied, binding studies were routinely performed in the pres-enCe Of 100 pM EDNA.
The binding of 13H1adenosine at 10 nM radioligand increased linearly with the protein concentration of the fat cell membranes utilized in the assay from 60 to 450 pg of membrane protein/tube (Fig. IA). A 350-to 400-pg aliquot of membrane protein was routinely used in the assay of adenosine binding to the fat cell membranes.
The kinetics of adenosine binding to fat cell membranes at 30 nM adenosine was rapid; more than 50% of the equilibrium binding value was obtained within 60 s (Fig. 1B). Increasing the temperature at which the incubation was performed to 37°C reduced the equilibrium adenosine binding capacity of the membranes by 30%. This reduction in the adenosine binding may be the result of increased metabolism of some critical component of the system at this temperature or perhaps the result of simple heat denaturation of the particulate membranes. Exposing the fat cell membranes alone to 37°C for 15 min reduced the ability of these membranes to subsequently bind I 3H]adenosine following a 20-min incubation at 4°C by 30 to 40% (data not shown). To avoid this complication.
binding assays were routinely performed in an ice bath (O-4") or at 22°C in a shaking water bath.
Saturation studies were next conducted, the results of these are shown in Fig. 2. Adenosine binding by fat cell membranes was concentration-dependent but could not be saturated at concentrations of adenosine as high as 1 mM. Analysis of this binding data by the method of Scatchard (43). as shown in the inset of Fig. 2. produced a nonlinear plot. A simple interpretation of these data suggests the existence of two populations of binding sites with differing affinities for adenosine. The higher affinity sites display an apparent dissociation constant (K,,) of 9.5 ? 3 x lo-" M (n = 3) and a maximum binding capacity (B,,,,) of 28 2 6 pmol of adenosine boundimg of membrane protein. The lower affinity adenosine binding sites display an apparent K,, of 9.5 2 2 x 10m4 M and a B,,,, of 1700 * 300 pmol of adenosine boundimg of membrane protein. The low binding affinity of this population explains the inability of 1 mM adenosine to saturate these sites and suggests these sites have little physiological relevance.
The identity of the radioactivity bound by the fat cell membranes was determined. Under the standard assay conditions using 100 nM 13H]adenosine. at least 80% of the bound radioactivity was identified as adenosine (Table II). Using the two solvent systems, no radioactivity in the form of inosine or adenine was detected by thin layer chromatography.
It was of interest to determine more precisely the subcellular localization of the adenosine binding components of fat cells. Fat cells were isolated and divided into two equal portions. Crude fat cell plasma membranes (similar to those used in the above studies) were prepared from one of these Adenosine Binding to Fat Cell Plasma Membranes 3117 portions. Highly purified fat cell plasma membranes were prepared from the other portion according to the method of McKee1 and Jarett (38). Adenosine binding was then assayed in both this highly purified fraction and the standard crude fat cell plasma membrane preparation.
The results of these studies are shown in Table III. Nearly 70% of the total adenosine binding capacity of the crude membrane fraction measured at 0.1 and 1 PM adenosine was retained in the highly purified plasma membranes. A somewhat higher percentage (78%) of adenosine binding measured at 100 PM was recovered in the highly purified fraction. The highly purified fat cell plasma membranes also display two populations of adenosine binding sites with affinities similar to those displayed in Fig. 2 (data not shown). These data demonstrate that adenosine binding components can be identified and are perhaps localized in the plasma membrane of the fat cell.
Dissociation of adenosine bound by fat cell membranes at equilibrium with 30 nM adenosine was examined following a loo-fold dilution (Fig. 3). Dissociation was very rapid when performed at 22°C or 0°C. More than 50% of the bound adenosine dissociated within 60 s following a loo-fold dilution at either temperature.
Data obtained from the vacuum filtration assay thus probably underestimates actual binding by approximately 10% since some dissociation is possible during the period of filtration and washing (~5 s). This same tech- nique has been used with fat cell membranes to characterize several other rapidly dissociating radioligands (40. 44). The nonlinearity of the semilogarithmic plot of adenosine dissociation probably reflects dissociation from the two populations of binding sites.
The pH dependence of adenosine binding to fat cell membranes was also examined. Adenosine binding was maximum when performed at pH 7.5. declining sharply at pH regions above 7.5 or below 6.5 (data not shown). Binding assays were. therefore. routinely carried out at pH 7.4.

Concentrations
of either calcium or magnesium below 200 PM did not appreciably affect the capacity of the fat cell membranes to bind adenosine (data not shown). Concentrations of either divalent cation greater than 1 mM inhibited adenosine binding, 50% inhibition being attained at 15 mM calcium or 30 mM magnesium.
Fat cell membranes washed twice with 1 mM ethylenediaminetetraacetic acid (EDTA) and resuspended in 10 PM EDTA demonstrated no reduction in their capacity to bind adenosine (data not shown).   (NAD) inhibited epinephrine-stimulated lipolysis of rat adipose tissue.
As shown in Table VI, NAD inhibited the binding of adenosine to the fat cell membranes, whereas nicotinic acid was an ineffective competitor of adenosine binding, as was nicotinamide (data not shown). The effects of several of these compounds on the low affinity adenosine binding was probed using 1 mM l"H]adenosine to test the specificity of the low affinity binding component (Table VII). 2',5'-Dideoxyadenosine, N"-(phenylisopropyl) adenosine. and theophylline at 10 FM failed to appreciably inhibit adenosine binding to the low affinity sites. A lOO-PM concentration of adenine reduced binding by approximately 20%. However, adenine inhibition of the high affinity sites probably accounts for most of this inhibition.
We examined the possibility that adenosine may be inhibiting hormone-stimulated cyclic AMP accumulation via modification of the hormone/receptor interaction. P-Adrenergic hormone binding to fat cell membranes, as probed with (-)l"HJdihydroalprenolo1 (37,51). was not significantly altered by 100 PM adenosine (data not shown). Likewise, neither 10 pM

DISCUSSION
Adenosine regulates adenylate cyclase activity and intracellular cyclic AMP levels in a wide spectrum of tissues and cell types. Specific plasma membrane receptors for adenosine have been postulated as the site of adenosine action in brain slices (9, 11, 52), platelets (14. 221, cultured human astrocytoma and glioma cells (19, 241, isolated bone cells (15). mouse neuroblastoma cells (23), transformed human lung fibroblasts (211,and Leydig tumor cells (27). although studies aimed at direct identification of these sites have not been reported. Plasma membrane adenosine receptors were proposed by Davies (7) as

Adentsinc
Binding to Fat Cell Plasma Membranw the site of adenosine action in adipose tissue. A similar proposal was advanced by other groups on the basis of studies performed with isolated fat cells (5. 34). The present study is. 10 the best of our knowledge. the first report characterizing 1 "H ladcnosine binding to a plasma membrane preparation. The plasma membrane fractions prepared from isolated white fat cells exhibit a capacity to bind adenosine. The binding of adcnosine at nanomolar concentrations to these membranes can be assayed by the vacuum filtration technique herein described.
The kinetics of adenosine binding to fat cell plasma membranes was rapid. rcac.hing 50%. of the equilibrium level within 1 min at 0°C. 22°C. and 37°C. Adenosine inhibition of norcpinephrine-stimulated cyclic AMP accumulation in fat cells is very rapid; if added I min after norepinephrine. adenosinc reduced cyclic AMP accumulation during the next minute (53). Adenosine has also been shown to act rapidly in cultured neuroblastnma cells (18). isolated bone cells (1%. and platelets (14).
obtained between 10 and 100 WM. Pereira and Holland (3) rcportcd a 44% inhibition of norepinephrine-stimulated lipolysis in adipose tissue by I FM adenosine and a similar level of inhibition by 1 to 10 @M ATP. ADP. or 5'-AMP. Adcnosine (1 PM) produced an 80% inhibition of norepinephrine-stimulated cyclic AMP accumulation in the presence of theophylline (53). Dilute fat cell suspensions were more sensitive to inhibition of norepinephrine-stimulated cyclic AMP accumulation by adenosine and by adenosinr-containing compounds (8). In this system. 0.1 PM adcnosine almost completely inhibited the cyclic AMP accumulation due to norepinrphrinc and adenosine-containing compounds displayed the following ordrr of potency: adenosinc > 5'-AMP > ADP > ATP (8). How much of the inhibition of both l:'H ladenosine binding and hnrmoncstimulated lipolysis and cyclic AMP accumulation by these adenosine-containing compounds is due to adcnosine released via their metabolism has not been established.

Addition
of adenosine deaminase to incubated fat cells increases basal cyclic AMP accumulation and pntentiates the norepinephrinc-stimulated increase in cyclic AMP (53. 54). Near maximal effects of adenosinc dcaminasc on norepinephrine-stimulated cyclic AMP accumulation in fat cells were achieved at approximately 20 s of incubation (33). Inhibitors of adenosinc deaminasc were required in the prosent study to detect appreciable adenosine binding to the fat cell plasma membranes. These inhibitors also potentiate adenosine inhibition of norepinephrine-stimulatid cyclic AMP accumulation in fat cells (33. 34). A similar observation was reported in lymphocytes. where adenosine deaminase reverses and EHNA potcntiates the stimulation of cyclic AMP levels by adenosine cm.
Additional evidence suggesting adenosine exerts its effects via a surface receptor is based on cxpcriments utilizing an adenosine derivative which is too large to penetrate ccl1 membranes (55). Adenosinc covalently linked to a watersoluble oligosaccharide.
stachytrsr. with a molecular weight greater than 1500. has been shown to virtually equipotcnt to free adenosine in inhibiting cyclic AMP accumulation by rat fat cells (34). Olsson et al. (55) first demonstrated the ability of this novel adenosine compound to exert an adenosine-like. dose-dependent coronary vasodilation following intracoronary infusion into dogs.
Adenine was a very potent inhibitor of adrnnsinc binding to fat cell membranes (Table V). From a structural standpoint. this observatinn is certainly not unexpected. Dolr (I) reported early that unlike adcnosine. adenine enhanced epincphrinestimulated lipolysis in adipose tissue. Davies (7) subsequently reported a 71% stimulation of norepinephrine-induced lipolysis in adipose tissue by 0.4 rnM adenine. In fat cells. significant elevation of theophylline-induced lipoiysis was demonstrated with 1 pM adenine (5). In contrast to the situation in fat cells. cyclic AMP levels of cultured mouse neuroblastoma cells were stimulated by adenosinr and inhibited by adeninc (18). Z-Chloroadenosine stimulation of adenglatc cyclasc in mouse neuroblastoma preparations was actually antagonized by adenine (23). Likewise. bone cell adenylate cyclase is stimulatid by adenosine and inhibited by adenine (25). Adcnine at both 18 and 73 ~IVI. however. had no effect on norrpinephrincstimulated adenylate cyclase activity of fat cell ghosts (4).

Dipyridamole
is a potent inhibitor of the uptake of I:'Hladcnosine into rat fat cells (5. 48). Kbert and Schwabe (5) observed that dipyridamole (10 P.M) potentiated the ability of adenosinr to inhibit lipolysis in fat cells rather than rcducc it. although it almost completely blocked the uptake of adenosine into the cells. These investigators concluded that the plasma membrane was probably the sitr of this action of adenosine in the fat cell. Dipyridamolr (40 to 100 ELM) likewise failed to block the reduction by adenosine of norcpinephrine-stimulated cyclic AMP accumulation (53). Neither dipyridamole nor NBTG. another potent inhibitor of adenosine uptake (14. 46). were effective inhibitors of adenosine binding to fat cell membranes. Fain et al (4. 56) reported that 6 to 10 PM 2'.6'-dideoxyadenosine inhibited nnrPpincphrine-stimulated cyclic AMP accumulation in fat cells approximately 50%. Fifty pcrccnt inhibition of norepinephrine-stimulated adenylate cyclasr activity in fat cell ghosts was obtained with 5 p.~ 2'.5'-dideoxyadcnosine (4). This adenosinc analog was a potent inhibitor of 1:'HJadenosine binding to fat cell plasma membranes (Table  V). 2'.5'-Didcoxyadcnosine has also been shown to be a potent inhibitor of both glucagon-stimulated adenylate cyclase activity in rat liver membranes (28. 31) and glucagon-and cpinephrine-stimulated cyclic AMP accumulation in intact hrpatocytes (28).
N"-(Phenylisopropyl)adenosine. which cannot bc dcaminated by adenosinc dcaminase, was shown by Fain (53) to be the most potent adenosine analog tested with respect to inhibition of cyclic AMP accumulation in intact fat cells. Cyclic AMP acrumulatiun and lipolysis due to 0.1 PM nortpinephrine in the presence of adenosinc deaminase was almost completely blocked by 0.01 P*M N"-(phenylisoprr)pyI)adcnosinr (S6). However, this adenosine analog. even at 73 PM. was virtually without eflbct on adenylate cyclase activity of fat cell ghosts (4). N',-(PhenyIisttpropyl)adenosinc failed to inhibit adcnosino binding to the fat cell membranes (Table VI). The binding of lYHladenosinr to fat cell plasma membranes Why N"-(phenylisopmpyl)adenosine. the most potent inhibwas inhibited by ATP. ADP. and 5'-AMP, Dole (I) originally itor of norepincphrine-stimulated lipolysis and cyclic AMP reported that addition of ATP. I?'-AMP. or adenosine (at 0.8 to accumulation in the intact cells is so ineffective as an inhibitor 4 mM) to incubated adipose tissue inhibited the lipolytic action of both adenylate cyolase and adenosinc binding remains of epinephrine.
Kappeler (6)  while simultaneously reducing its affinity for N"-(phenylisopropyl)adenosine. Blume and Foster (23) reported the selective loss of 2-chloroadenosine-stimulatable adenylate cyclase activity of neuroblastoma cells following fractionation, although basal cyclase activity did not change. Penit et al. (26) similarly noted that purified membrane fractions prepared from neuroblastoma cells displayed a marked reduction in adenylate cyclase responsiveness to the stimulatory action of adenosine. Alternatively, N"-(phenylisopropyl)adenosine and 2'.5'-dideoxyadenosine could exert their influence on cyclic AMP accumulation in fat cells via separate distinct sites (56). Another common observation concerning adenosine action and mammalian cell cyclic AMP metabolism is that regardless of whether adenosine stimulates (9,12,14,18,19) or inhibits (33, 34) the accumulation of cyclic AMP, methylxanthines appear to antagonize its action in these systems. In rat fat cells, theophylline inhibits uptake of tritiated adenosine and opposes the action of adenosine (5. 48, 56). The present studies demonstrate that theophylline inhibits adenosine binding to fat cell membranes. Significant inhibition of adenosine binding was detected with 0.1 PM theophylline. a concentration of theophylline too low to inhibit cyclic AMP phosphodiesterase (57). Fain et al. (56) found l-methyl-3-isobutyl xanthine to be more potent than theophylline in inhibiting adenosine uptake in the intact fat cell. The present study shows that theophylline inhibits adenosine binding to fat cell membranes.
In contrast, 1-methyl-3-isobutyl xanthine failed to inhibit adenosine binding even at 100 PM. These data suggest that inhibitors of adenosine uptake of intact fat cells. such as dipyridamole and methylxanthines.
are generally poor inhibitors of adenosine binding. Although theophylline does inhibit ["HIadenosine binding, this effect is probably not related to the effect of theophylline on adenosine uptake. Unlike adenosine uptake, inhibition of adenosine binding by theophylline appears to be specific for this methylxanthine.
McKenzie et al. (58) recently similarly reported that the ability of adenosine to relax intestinal smooth muscle is antagonized by theophylline but not 1-methyl-3-isobutyl xanthine. The present study demonstrates that highly purified plasma membranes prepared from fat cells do possess adenosine binding sites. This fat cell plasma membrane preparation was previously shown to be enriched in both adenylate cyclase activity (38) and P-adrenergic receptors (37). Since adenosine antagonizes the action of catecholamines on lipolysis and cyclic AMP accumulation in intact fat cells. it was of interest to explore the possibility that adenosine might, inhibit or modify catecholamine binding to its putative /?-adrenergic receptor. Adenosine possessed no significant ability to alter catecholamine interaction with its receptor as probed with (-1-l "Hldihydroalprenolol.
p-Adrenergic agonists and antagonists had no effect on 13H]adenosine binding to the fat cell membranes.
It is interesting to speculate on the function of these adenosine binding components identified in the fat cell plasma membrane preparations. The inability of dipyridamole to effectively compete for these binding sites with adenosine would argue against these sites being involved in adenosine uptake. The required presence of EHNA or deoxycoformycin to demonstrate adenosine binding would similarly rule out adenosine deaminase. Several characteristics of the binding sites suggest these sites may be involved in the regulation of adenylate cyclase activity. Adenosine has been shown to be an inhibitor of fat cell cyclic AMP phosphodiesterase activity (4. 7). The action of adenosine in fat cells is to inhibit, rather than potentiate, cyclic AMP accumulation arguing against this enzyme being the binding site. In addition. I-methyl-3-isobutyl xanthine. one of the most potent phosphodiesterase inhibitors, failed to compete with adenosine for these sites.
Perhaps some of these adenosine binding components are regulatory sites on the adenylate cyclase complex which modulate the activity of this enzyme. This scheme is only speculative, however, and the precise identity and function of these adenosine binding components remains to be established. Rosenblit and Levy (48) have recently approached this same problem by using a photoreactive derivative of adenosine, 8-azidol 2-JH]adenosine. Photolysis of this derivative with intact fat cells led to its incorporation into several protein components of the membrane (48). However. the specificity of this labeling technique was not addressed in this report (48).