Subcellular Distribution of and Epinephrine-induced Changes in Hormone-sensitive Lipase, Phosphorylase, and Phosphorylase Kinase in Rat Adipocytes*

A systematic study was carried out of the subcellular distribution of hormone-sensitive lipase, glycogen phosphorylase, and phosphorylase kinase in rat adipocytes and the changes in these enzymes when the cells were incubated with epinephrine. Under the conditions of homogenization used, approximately one-third of the total hormone-sensitive lipase activity was associated with the fat cake floated to the top of a homogenate by low speed centrifugation, although as discussed in detail, this value remains uncertain in view of technical problems faced in assaying lipase tightly bound to fat. Subfractionation of the fat-free homogenate showed that over 80% of the hormone-sensitive lipase was recovered in the soluble fraction and the specific activity of that fraction was considerably higher than it was in particulate fractions. The plasma membrane fraction contained less than 2% and the microsomal fraction less than 3% of the activity in the fat-free homogenate.

From the Division of Metabolic Disease and Division of Pharmacology, Departmed of Medicine, School of &fedicine, University of California at San Diego, La Jolla, California 92037, and Division of Laboratory M edicine, Washington University, St. Louis, Missouri 6'3110 SUMMARY A systematic study was carried out of the subcellular distribution of hormone-sensitive lipase, glycogen phosphorylase, and phosphorylase kinase in rat adipocytes and the changes in these enzymes when the cells were incubated with epinephrine.
Under the conditions of homogenization used, approximately one-third of the total hormone-sensitive lipase activity was associated with the fat cake floated to the top of a homogenate by low speed centrifugation, although as discussed in detail, this value remains uncertain in view of technical problems faced in assaying lipase tightly bound to fat. Subfractionation of the fat-free homogenate showed that over 80% of the hormone-sensitive lipase was recovered in the soluble fraction and the specific activity of that fraction was considerably higher than it was in particulate fractions.
The plasma membrane fraction contained less than 2% and the microsomal fraction less than 3% of the activity in the fat-free homogenate.
When adipocytes were lysed, over 80% of the total lipase activity was recovered in the supematant fraction and less than 20 % in the fat cell "ghosts." The distribution of lipase activity among subcellular fractions of the fat-free homogenate was not altered by previous treatment of the adipocytes with epinephrine.
Lipase activity in the soluble fraction from control cells was activated by treatment with cyclic adenosine 3',5'-monophosphate-dependent protein kinase and MgATP.
When the cells had been previously treated with epinephrine, the percentage of activation by protein kinase was reduced in this and in all fractions.
Phosphorylase and phosphorylase kinase activity were found predominantly in the soluble fraction; small amounts (21 and 14%, respectively) were associated with the fat cake assayed after resuspension in Triton X-100. The distribution of these enzymes in cells previously treated with epinephrine was not significantly different from that in control cells.
Total phosphorylase activity (assayed in the presence of 5'-AMP) was significantly increased in homogenates prepared from cells previously treated with epinephrine as was the ratio of phosphorylase activities measured in the absence and in the pres-* This research WRS supported by Grants HL-05899, HL-12373, HL-14197, and AM-11892 from the United States Public Health Service. ence of 5'-AMP.
On the other hand, there was no change in total phosphorylase kinase activity after epinephrine treatment and no change in the ratio of activities measured at low pH values (6.2 or 6.8) and high pH value (8.2).
Further characterization of the phosphorylase and phosphorylase kinase systems in adipose tissue is needed to assess the regulatory role of phosphorylase kinase in this tissue.
,4dipose tissue contains at least two different systems that, are rapidly activated by epinephrine, one leading to free fatty acid mobilization (1,2) and one to glycogenolysis (3)(4)(5)(6). Both systerns are stimulated not only by catecholamines but also by ilCTH, glucagon, and many other so-called lipolytic hormones. The significance of the parallel activation of lipolysis and glycogenolysis is not clear nor is it certain that they are functionally or obligatorily linked.
It is now well established that activation of hormone-sensitive lipase is effected via adenyl cyclase and cyclic AMP'-dependent protein kinase (7-9). The mechanism of phosphorylase activation in adipose tissue has not been studied in detail but it seems likely that, as in other tissues, this is also mediated by cyclic AMP (6). Further characterization of these two hormonally regulated enzyme systems is essential as a basis for exploring the interrelationships between them and the possibility of differential regulation occurring after the receptoradenyl cyclase level.
No systematic studies on the subcellular distribution of hormone-sensitive lipase have been reported and t,he limited data available are conflicting (10-13).
In part this may reflect the presence in adipose tissue of several lipases and esterases and the uncertainty regarding criteria for specific assay of hormone-sensitive lipase. Crude extracts of adipose tissue contain, in addition to hormone-sensitive lipase: lipoprotein lipase (14) ; a high level of activity against monoglycerides and diglycerides (6, 11); and activity against short chtlin glyceryl esters (esterase) (15,16). The properties of these enzyme activities suggest that they are at least in part referrable to 4813 different enzyme proteius (17, 18) but none has been prepared in pure form. Only receutly has llormone-sensitive lipase been partially purified and characterized (19). Thus it. is difficult to assess the siguificnnce of some of the earlier results. Rizack (10) studied hormone-sensitive lipase actiyily in a psrticulate fraction sedimented from a 10,000 x g supernatnnt fraction of fat pad homogenates by eentrifugation at 105,000 x g for 12 hours. No data were presented on the occurrence or properties of hormone-sensitive lipase in other subcell&t1 fractions. Vaughan et al. (11) found the enzyme to be largely associated with the fat layer after homogenization in 0.15 M KCl. Huttunen et aE. (12) showed t.ha.t the fract,ion of hormone-sensitive lipase associated with the fat layer was variable, being a function of ionic composition of t,he homogenizing medium and of temperature at which the homogenate was prepared. When binding to the fat layer was minimized, most of the enzyme was found in the 100,000 x g supernatant fraction prepared from the layer beneath the fat cake. Most recently, Crum and Calvert (13) have suggested that the enzyme is associated predominantly with the plasma membrane.
In the present study we have utilized assay conditions shown to be optimal for the assay of partially purified hormone-sensitive lipase (20) and in addition have demonstrated the responsiveness of the enzyme in the various subcellular fractions studied to activation by cyclic AMP-dependent protein kinase. Activation of phosphorylase has been shown in hormonestimulated fat pads (3,21) and ndipocytes (22); the phosphorylase activity was measured in the supernatant fraction obtained by low speed centrifugation of homogenates. No attempt has been made to study the subcellular distribution of phosphorylase or its possible trssociatioll with glycogen particles, such as has been described by Meyer et al. in muscle homogenates (23). Stull and Mayer (24) have recently reported that in skeletal muscle after either electrical stimulation or isoproterenol administratiolr, there can be a dissociation between the conversion of phosphorylase b to phosphorylase a on the one hand and the activatiou of phosphorylase 6 kinase on the other. The intimate nature of the phosphorylase actiration system iu adipose tissue has not been previously studied.
In the present paper we report a systematic study of the subcellular distribution of hormone-sensitive lipase, phosphorylase kinase, and phosphorylase. Subcellular fractions of rat adipocytes were prepared by methods that have been carefully characterized on morphological and biochemical grouiids by .
In addition, we have studied the response of these several enzymes to epinephrine treatment of adipocytes.

METHOIZi AND MATERIAL8
Cellular Preparations-Epididymal fat pads were excised from fed Sprague-Hawley rats and adipocytes were prepared according to Rodbell (29). The washed cells were resuspended (about 2 mg of cell protein per ml) in Krebs-Rillger bicarbonate buffer (pH 7.4) containing 3(;, fatty acid-poor bovine serum albumin (Cohn fraction V) and glucose (2 mg per ml). The suspended cells were equilibrated with 957; 02-5(;(:, CO2 and incubated for 15 min at 37" prior to the start of each experiment. Aliquots of the fat cell suspension were taken for cell count and determination of protein concentration as previously described (25). The cell suspension was then dispensed into siliconized vials with and without epinephrine (10 PM) and iucubated for exactly 5 min. At the beginning and end of the incubation period an aliquot from each cell suspension was pipett,ed into trichloroacetic acid (final concentrat,ion 5%) for subsequent det.ermination of glycerol and cyclic AMP coneentrations.
The rest of the cell suspension was transferred to a siliconized tube and centrifuged at 400 x g for 30 s at room temperature. The infranatant solution was aspirated and the cells were washed with 5 to 6 volumes of llledium I (10 mM Tris, pH 7.4; 1 .O m&c EDTB; and 0.25 11 sucrose). The cell suspension was recentrifuged and the packed cells were resuspended in 3 to 4 volumes of the same medium for cell fractionation by a slight modification of the method of McKee1 and Jarett (26). The cell suspension, which contained approximately 2 mg of cell protein per ml, was homogenized as originally described and the homogenate centrifuged at 16,000 X g for 15 min at 4" to yield a pellet (PI), an infranatant fluid (S1), and a floating fat cake layer which was discarded. 81 was further centrifuged at 160,000 x g for 70 min to yield the microsomal supernatant fraction while the pellet, was gently rinsed, resuspended in Medium I and recentrifuged at, 160,000 x g for 70 min to yield the microsomal pellet.
The pellet from the 16,000 X g centrifugation of the crude homogenate (PI) was resuspended in Medium I and centrifuged at 1000 X g for 10 min to yield a pellet containing nuclei and large cell debris (P,). The supernatant fluid from this centrifugation was centrifuged at 17,000 x g for 20 min to yield a pellet (P3) which was resuspended in 8 ml of Medium I and placed on a discontinuous Ficoll gradient. consisting of 10 ml of 15% Ficoll beneath 10 ml of 9yc Ficoll, both in 0.25 M sucrose. The gradient was centrifuged in a Beckman SW 25.1 rotor at 24,000 rpm for 45 min. The plasma membranes banded at the sample-g%, Ficoll interface and t,he mitochondria formed a pellet at the bottom of the tube. XI1 particulate fractions were resuspended in 1 to 2 ml of Medium I. Prot,ein content of the various subcellular fractions was determined by t,he method of Lowry et al. (30).

Preparation of Fat Cell '
LGhosts"---Fat cell ghost,s were prepared as described by Rirnbaumer et nl. (31) except that the lysing medium contained only 10 XnM Tris, pH 7.4, and 1 MM EDTA.
This preparation was carried out at 0". Five milliliters of t,he lysing medium was added to approximately 1.7 x 10' fat cells in a plastic tube. The tube was inverted 20 times, 3 s per cycle. The lysate w&s collected by floating the released fat by centrifugation at 200 x g for 1 mill and aspirating the turbid infranatant material.
To the remaining fat plus intact fat cells, 5 ml of the fresh lysing medium \I-ere added and the procedure was repeated three additional times. The lysate: were pooled and centrifuged at 900 x g for 15 mirr to yield a pellet (fat cell ghosts) and the supernatanb fluid.
The fat cell ghosts were then homogenized in 2 ml of lysing medium prior to enzyme assays.
Assay Procedures-Hormone-sensitive lipase was assayed by the [14C]triolein method of IIut,tunen and Steinberg (20), t.he free fatty acid being isolated by the method of Kelley (32). When hormone-sensitive lipase was activated with skeletal muscle protein kinase prior to assay, the activation system was as previously described (20), except that an ATP-regenera,ting system was added.
One-tenth milliliter of the activat,ion mixture was added to 0.1 ml of the sample. The final mixture during activation contained: 10 m&f Tris, pH 7.4; 1.0 mu dithiothreitol; 0.5 mM EGTA; 0.25 mg of bovine serum albumin per ml ; (34); the dist.ribution pattern found was similar in the presence microsomes was not reduced when this fraction was resuspended or in the absence of SaCI.
Inhibition of total lipase activity in a large volume of Medium I and recentrifuged at 160,000 x b?; T\'aCl ranged from 20 t,o 40%, which is similar to the degree 8, suggesting that it was not simply adsorbed or trapped. of NaCl inhibition observed with purified hormone-sensitive In light of the above data, triglyceride lipase activity in sublipase (18).
sequent studies was measured only in selected fractions. Table  Triglyceride content of fat cell homogenates was determined III summarizes the data on distribution of activity found with the method of Kessler and Lederer (35). Glycerol (36) among the various subcellular fractions in a series of experiand cyclic ilRiP (37) were determined as previously described. ments. Over 607; of the total activity in control cells was Glycogen phosphorylase was assayed in the direction of glu-recovered in the P1 and S1 fractions as measured by the ['"Clcase l-phosphate formation in the presence and absence of 5'-triolein assay. This is consistent with the data in Table I A?\IP as described previously (38  counted for 409, of the original activity. Results from epi-tivated to a much lesser extent by protein kinase than were nephrine-treated cells are discussed later. corresponding fractions from control cells. Kinase-In preliminary experiments, activation by protein kinase was examined with the system described by Huttunen et al. (20). This yielded consistent activation of the enzyme in the microsomal supernatant fraction but, little or no activation in other fractions.
Previous studies by Jarett and McKee1 (25) showed that the crude homogenate and the various particulate fractions contained an active Mg++-ATPase but that the microsomal supernatant fraction did not, suggesting that ATP might be rate-limiting in the case of the particulate fractions.
When an ATP-regenerating system was added to the system previously used, a consistent and rapid activation was observed in all the fractions, reaching a maximum by 8 min. All subsequent assays included the ATP-regenerating system.
The triglyceride lipase in control fat cell homogenates was activated 56% by cyclic AMP-dependent protein kinase (Table  III).
The enzyme associated with the & fraction, the microsomal fraction and the microsomal supernatant fraction was activated to about the same extent while the enzyme in the PI and plasma membrane fractions was only activated about 30%.
The degree of activation of lipase in the total homogenate, the & fraction and the microsomal supernatant fraction was of the same magnitude as that found by Huttunen et al. (20) with purified hormone-sensitive lipase. This would indicate that the triglyceride lipase activity measured in the present experiments represents primarily hormone-sensitive triglyceride lipase activity with very little contribution from or interference by lipoprotein lipase or other lipases.
Triglyceride Lipase in Adipocyk Ghosts-The studies described above showed that the purified plasma membranes from adipocytes contained very little triglyceride lipase activity whereas Crum and Calvert (13) have suggested that hormonesensitive lipase is associated primarily with the inner surface of the plasma membrane.
To explore this further, ghosts were prepared according to the method of Birnbaumer et al. (31) by centrifuging the lysate for 15 min at 900 X g rather than 15 min at 100,000 x g as in the studies of Crum and Calvert. Both the ghosts and the 900 x g supernatant fraction were assayed with the [14C]triolein assay method in order to permit a direct comparison of the activities in the two fractions.
Crum and Calvert used endogenous substrate to measure the activity in the crude homogenate and in the fraction containing fat plus 100,000 X g infranatant material, but used tributyrin as substrate to measure the activity in the ghosts. As shown in Table IV, the 900 x g supernatant fraction from the fat cell lysate contained 83% of the total activity while the ghosts contained only 177,. The specific activities of the 900 x g supernatant fraction and of the ghosts were about equal.
Cyclic AMPdependent protein kinase activation of the enzyme in the 900 x g supernatant was similar to that reported in Table III for the X1 fraction; the enzyme associated with the ghosts was also activated although to a lesser extent. Table III  Lipase--Incubation of adipocytes with 10 PM epinephrine increased the rate of glycerol release (5 min incubation) from 24 f 6 to 1821 f 183 nmoles per mg of protein per hour (n = 4); the level of cyclic AMP (cells plus medium) from 12. 3   TABLE   III   TABLE   II   Distribution of triglyceride lipase among subcellular fractions The homogenate of adipocytes was fractionated into subcellular components as described under "Methods and Materials" and the lipase activity was measured by the [%]triolein assay. PI represents the 16,000 X g pellet from the homogenate: Pe the 1000 X g pellet from PI; Pa the 17,000 X g pellet from PI; 81 the 16,000 X g supernatant from the original homogenate. Plasma membranes and mitochondria were derived from PS by discontinuous density gradient centrifugation. S1 was further centrifuged at 160,006 X g for 70 min to yield microsomes and the microsomal supernatant fraction.     I  -550  1300  750  580  750  110  2  +  1450  2100  650  580  650  100  3  -370  750  380  370  380  60  4  -370  1000  630  880  630  100  5  +  540  1380  840 1120 840 120 f 0.9 to 32.6 & 2.1 pmoles per mg per protein (n = 14). Epinephrine treatment of the adipocytes prior to homogenizat.ion did not cause a change in the percentage of lipase activity associated with P,, plasma membrane, or microsomal fractions (Table III). However, recovery of activity in the S, fraction a, * Total activity calculated for homogenate prepared from 8 ml of packed cells in 35 ml of Medium I. was significzmtly and consistently lower in homogenates prepared from epinephrine-treated cells than in homogenates prepared from control cells. Recoveries in the microsomal supernat.ant fraction prepared from epinephrine-treated cells also tended t,o be lower but the difference was not statistically significant.

As shown in
The degree of protein kinase activation of lipase in all fractions prepared from epinephrine-treated cells was significantly less tha.n in corresponding fractions from control cells (Table  III).
These findings are consistent with earlier results both in crude fractions and with purified hormone-sensitive lipase (9, 20). Lipase ilctivity Associated with Pat Cake-Measurement of lipase activity in the crude, triglyceride-rich homogenate or in the fat cake itself poses special problems.
Previous studies show that the rate of release of fatty acids from endogenous substrate in crude homogenates is increased by prior treatment of fat pads with epinephrine (11). In the present studies the rate of release of free fatty acid from endogenous triglyceride substrate was increased 113 f 12O/;, (n = 9) in homogenates from epinephrine-treated cells as compared wit,h cont.rols. When, however, lipase activity was measured in terms of the rate of release of [l"C]oleic acid from added [14C]triolein, the increase due to epinephrine treatment was only 17 f 5oh (n = 15). Evidently the added radioactive substrate and the endogenous substrate are not equivalent.
To explore this further, studies were done comparing net free fatty acid release in total homogenates with and without the addition of exogenous triolein substrate. As shown in Table V, net release was increased by addition of extra substrate (triolein) both to control homogenates and to homogenates from epinephrine-treated cells. This occurred although the increment in substrate concentration (0.74 mg per ml) was K&use-As shown in Table VI, the phosphorylase activity of the fat cell homogenate was recovered only in Fraction S1 and in the microsomal supernatant fraction; the particulate fractions (2'1 and microsomes) were found to be devoid of activity. However, only two-thirds of the phosphorylase act,ivity in t,he original homogenate was accounted for. Phosphorylase kinase was also found to be primarily associated with the S1 and microsomal supernatant fractions; 50/;, or less was recovered in PI and in the microsomal fraction. Only one-half of the original activity was accounted for. Assay of the fat cake posed dificulties; when it was rehomogenized in 0.50/;, Triton X-100, 21% of the phosphorylase and 14% of the phosphorylase kinase activity were found in this fraction. Effect of flpinephrine on Phosphorylase and Phosphorylase Kinase Activity-Epinephrine treatment of adipocytes did not alter the distribution of either phosphorylase or phosphorylase kinase among the adipocyte subcellular fractions and only the activity in total fat-free homogenates (low speed centrifugation) was subsequently measured.
Tot.al phosphorylase activity (measured in the presence of 5'.AMP) in whole homogenates prepared from epinephrine-treated adipocytes was increased almost a-fold (p < 0.01) over that from nontreated cells (Table VII). This activation was accompanied by a marked increase in the relative activity assayed in the absence of 5'-AMP (p < 0.01). In contrast, epinephriue treatment of adipocytes resulted in no change in tot,al phosphorylase kinase act.ivity, nor in the ratio of activities assayed at low pH (6.2 or 6.8) and at high pH (8.2) (Table VII). DISCUSSION A large fraction of the hormone-sensitive lipase in rat adipose tissue can be recovered in a large, phospholipid-rich "particle" of high molecular weight (approxima.tely 7 x 106) (19). This has suggested the possibility tha.t the enzyme in the intact cell might be associated with a lipid-rich matris. Electron microscopic evidence for an ordered complex of filaments (44) or a fenestrated envelope of thin cytoplasm (45) surrounding the central fat droplet has been presented, although these structures have never been fully characterized.
Localizat.ion of hormone-sensitive lipase at the perimeter of the fat droplet would seem to be functionally advantageous. One might postulate that the vigorous homogenization needed to disrupt intact adipose tissue might result in fragmentation of a membrane-like structure present in the intact cell. However, the present studies were carried out with the gentlest available technique for disrupting isolated adipocytes (26) and a.gain the hormone-sensitive lipase was found predominantly in the cytosol, 89% of it if one neglects the enzyme associated with the fat cake. Less than 27" was associated with the plasma membrane and the specific activity in this fraction was rather low. Furthermore, lysis of the cells to prepare ghosts by the method of Birnbaumer et al. (31) released over 805; of the total hormone-sensitive lipase into the supernatant fraction, only 17'3> being recovered in the ghosts. These results strongly suggest to us that hormone-sensitive lipase is predomiuwntly a oytosol enzyme. The nature of the enzyme activity associated with the fat cake is difficult to evaluate. Previous studies have shown that the fraction of enzyme activity bound to the fat cake depends in part on the conditions used for preparing the homogenate (11,12). In the present studies, using a sucrose-EDT&Tris medium and homogenizing at room temperature, about onethird of the hormone-sensitive lipase was bound to the fat layer. Washing procedures did not remove the lipase from the fat and attempts to remove lipase from the fat by solvent extraction procedures have in the past been unsuccessful (46). All that can be said is that a significant fraction of hormonesensitive lipase is associated with the fat layer and that the absolute enzyme activity in that fraction is markedly increased when the adipocytes have been exposed to epinephrine. This may account for the observed decrease in lipolytic activity in the microsomal supernatant fraction in epinephrine-treated cells. Whether that enzyme is particle-bound and trapped in the fat layer or whether it represents firmly adsorbed "soluble" enzyme cannot be determined.
The results of the studies in which exogenous triolein substrate was added to whole homogenates are compatible with the conclusion that the enzyme tightly bound to endogenous substrate acts on it exclusively (i.e. is already saturated with substrate) and that the exogenous substrate is hydrolyzed by lipase in other fractions of the whole homogenate.
At the very least, the data indicate that the added exogenous substrate does not mix uniformly with the endogenous substrate during the assay. The results indicate the potential hazards involved in relying on measurements of the rates of hydrolysis of added radioactive triglyceride to assay lipase activity in mixtures containing significant amounts of endogenous unlabeled triglyceride.
The present results confirm and extend previous studies on cyclic AMP-dependent protein kinase activation of hormonesensitive lipase (20). Activation was shown with the enzyme in each of the subcellular fractions.
When the adipocytes had previously been exposed to epinephrine, the percentage activation effected by protein kinase treatment was sharply reduced, indicating that the activation process in the intact cell is very probably the same as or closely related to that effected by protein kinase.
Both phosphorylase and phosphorylase kinase were found almost exclusively in the soluble fractions.
Since, however, as much as 507; of the total activity was lost during the preparation of the subcellular fractions, the presence of these enzymes in other fractions cannot be completely ruled out. Prior treatment of adipocytes with epinephrine gave a clear-cut increase in total phosphorylase activity, as well as an increase in the ratio of activities measured in the absence and in the presence of 5'-AMP.
On the other hand, phosphorylase kinase activity showed no change at all as a result of epinephrine treatment, nor did the ratio of activities at low and high pH values change as a result of epinephrine treatment.
These results raise the question of whether there are in adipose tissue mechanisms for epinephrine stimulation of phosphorylase that do not depend upon a concurrent transformation of phosphorylase kinase to an activated form. However, neither phosphorylase nor phosphorylase kinase in adipose tissue has been previously purified or characterized.
The methods used for assay here have been drawn from previous experience with these enzymes in heart (39) and skeletal muscle (24). Consequently, interpretation of the result must be guarded.
The methods for preparing phosphorylase kinase for assay or the conditions necessary to distinguish between activated and nonactivated forms may be quite different in adipose tissue from those in muscle or heart. This question is under further investigation.