346 Biochimica et Biophysica Acta 968 (1988) 346-352 Elsevier

BBA 12207

In vitro glucose and 2-aminoisobutyric acid uptake by rat interscapular brown adipose tissue

Fernando Zamora a Lluis Arola a and Mari~ Alemany b

Departament d'Enginyeria Qulrnica i Bioquirnica, Universitat de Barcelona, Tarragona, and h Departament de Bioqu~rniea i Fisiologia, Universitat de Barcelona, Barcelona (Spain)

(Received 21 May 1987) (Revised manuscript received 20 November 1987)

Key words: Glucose transport; Insulin; Catecholamine (Rat brown adipose tissue)

The dependence upon substrate and insulin concentrations, as well as on sodium and potassium concentra- tions in the medium of the uptake of glucose and 2-aminoisobutyric acid, was determined for fragments of brown and white adipose tissues incubated in vitro. Brown adipose tissue showed a high capacity for glucose uptake at high glucose concentrations, this uptake being dependent on both glucose and insulin concentra- tion. White adipose tissue showed much more limited uptake capabilities. The presence of Na + and K + had little effect on the uptake. The uptake of 2-aminoisobutyric acid was similar in both adipose tissues, being enhanced by physiological levels of insulin and depressed by ouabain. This amino acid transport was dependent on Na + and K + concentrations, and the overall transporting capability was two to three orders of magnitude lower than that for glucose. It was concluded that amino acids could not play a significant role as bulk thermogenic substrates for brown adipose tissue, as their transporters lack the plasticity of response to high substrate and insulin concentrations which characterize brown adipose tissue uptake of glucose.

Introduction

Brown adipose tissue plays a key role in mam- malian thermogenesis [1,2] because of its high blood flow [3] and high capacity for uptake of glucose [4,5] and other substrates [4,6,7]. Brown adipose tissue uses a significant proportion of the available energetic substrates in situations of ex- posure to low temperatures [8,9] and in situations in which there is an excess of nutrients available [10].

Abbreviation: Hepes, 4-(2-hydroxyethyl)-l-piperazineethane- sulfonic acid.

Correspondence: L. Arola i Ferret, Department d'Enginyeria Qulmica i Bioqulmica, Facultat de Citncies Qulmiques, Uni- versitat de Barcelona, P1. Imperial Tarraco, 1, 43005 Tar- ragona, Spain.

Glucose is probably one of the main exogenous substrates used by brown adipose tissue [11], and is transported across the plasma membrane by carrier-facilitated diffusion [12]. This diffusion is regulated by glucose availability as well as by insulin [13].

The regulation of brown adipose tissue activity is in part due to hormone action [14], and prob- ably also to substrate availability [15]. However, a significant part of its ability to function depends upon sympathetic control, through direct innerva- tion [16] and catecholamine-mediated action [17]. These factors create problems for the in vivo study of the transporter function, because animal manipulation or anesthesia can significantly alter the blood flow and /o r adrenergic response of the tissue. For this reason, the use of small pieces of tissue for in vitro incubations has been developed

0167-4889/88/$03.50 (=~ 1988 Elsevier Science Publishers B.V. (Biomedical Division)

to estimate the basal conditions of glucose uptake into the tissue, and the influence of insulin in the absence of catecholamine stimulation.

Materials and Methods

Adult Wistar female rats, weighing 195-205 g were used. The animals were kept under standard light, humidity and temperature conditions in solid-bottomed cages with sepiolite as absorbing material. The animals were fed ad libitum with s tandard pellets ( type A04 from Panlab, Barcelona). The animals were killed by decapita- tion and interscapular brown adipose tissue, as well as retro-lumbar white adipose tissue masses were immediately exposed. Brown adipose tissue was frozen in situ with liquid-nitrogen-cooled tongs. Alternatively the tissue was dissected out, minced into fragments of about 3-5 mg and sus- pended in one of the following media: solution A, Krebs-Ringer bicarbonate buffer (pH 7.4) contain- ing 20 mM Hepes and 10 g / l bovine serum al- bumin (Sigma); solution B, solution A plus 1 mM ouabain (Sigma); solution C, solution A depleted of sodium and potassium. Sucrose was added to hypoosmotic solutions to provide the same osmolality (300 mosmol/1) as solution A after the addition of the required concentrations of alkaline metals. The same strategy was used for hyper- osmotic solutions that were compared with control solutions of sodium and potassium physiological concentration and of the same osmolality.

All manipulations of the non-frozen samples carried out at room temperature (18-20°C) , be- cause cooling of the tissue pieces for a short time to 4 ° C resulted in a significant loss of their ability to take up glucose from the medium.

Glucose uptake was estimated by incubating pieces of the tissue in open tubes gassed with a 95%:5% O2/CO2 gas mixture. The incubation medium contained the desired concentration of salts, glucose and insulin (Novo ultra-rapid, 54 m U / m l and 73 mU// ,g) , and 20 mM Hepes to provide a higher pH buffering capacity. Glucose solutions were prepared with pure D-glucose (Carlo Erba) containing trace amounts of 2-deoxy[3H] glucose (Amersham), at a final spec. radioact, of 39.5 Bq/nmol glucose. Incubations were started by the addition of glucose to the tissue pieces

347

(about 15-20 mg per tube in a final volume of 0.25 ml).

Incubations were carried out for up to 30 min in a shaking (1.3 Hz) water-bath at 37°C; an incubation time of 20 min was selected for most of the experiments. Incubations were terminated by rapidly filtering (under vacuum) the tissue pieces through a glass fiber filter paper (Whatman GF-E), and washing the cells three times with 10 ml buffer containing no radioactivity. The tissue pieces were then added to tubes into which 1 ml 9.7 M hydrogen peroxide/concentrated sulphuric acid (7 : 3, v /v ) was added, in order to completely oxidize all organic matter. This converted all the 3H into 3H20. The tubes were closed and heated for 2 h at l l 0 ° C in an oven, when all digests were clear. After cooling, the tubes were opened and a slurry of 1.6 g of powdered barium hydroxide in 2 ml of distilled water was added to each, im- mediately closing the tubes again. After 10 min with occasional stirring, the tubes were centri- fuged and the clear supernatants were removed and added to Thunberg tubes. The Thunberg tubes were connected to a vacuum pump and then closed, in order to facilitate the distillation of the water of the digest. The tubes were then placed in a hot water-bath, maintaining the bulbous part of the tube cap in ice, to distil into it part of the water (containing 3H20 ). After 10 min, samples of this distilled water were taken and the spec. radioact. was measured by means of liquid scintillation counting of 1-ml samples.

The uptake of 2-aminoisobutyric acid was estimated under the same conditions as those for glucose. The medium contained 7.5 mM glucose and the required concentrations of salts, ouabain and insulin, as well as 0.010 mM 2-aminoisobu- tyric acid (Sigma) containing trace amounts of 2-amino-[1-14C]isobutyric acid (Amersham) at a final spec. radioact, of 3.70 kBq/nmol .

The incubations were carried out as described for glucose for up to 20 min. A 10-rain incubation time was often fixed for most comparative mea- surements. The incubations were also arrested by rapid filtration, and the tissue pieces were washed and then added to scintillation vials. They were then solubilized with 0.250 ml Soluene-350 (Packard) in a 30 °C water-bath. The radioactivity of the digest was then counted and corrected for

348

quenching. 2-Aminoisobutyric acid uptake was then calculated from the amount of 14C retained by the tissue pieces.

The samples used for ATP and ADP estimation were frozen immediately after the incubation pro- cess (or in situ for control samples). The tissue pieces were weighed and homogenized in 3 mM perchloric acid. The precipitates were used for the estimation of protein content [18]. The clear su- pernatants were used for measurement of ATP and ADP [19] concentrations.

Statistical comparisons between the means were established with Student's t-test, as well as analy- sis of variance (ANOVA, BMDP Statistical Soft- ware) [20] programs.

Results and Discussion

In Table I, ATP and ADP concentrations in brown adipose tissue pieces subjected to incuba- tion under standard conditions are presented. The concentrations calculated per unit of protein weight were more consistent than those referring to tissue weight, due to the fact that the incuba- tion medium imbided the tissue pieces and af- fected their weight as compared with frozen in situ

samples. The levels of ATP and ADP were well maintained, as was the A T P / A D P concentrations ratio up to 30 min. At 60 min, this was less effective and, for this reason, the standard 10-20- rain incubation times were used.

Under the standard conditions described above, and for incubation times up to 30 min, correlation coefficients of glucose uptake versus time of 0.987 for brown and 0.932 for white adipose tissues were obtained. These data suggest that both tissues were able to withstand well the process of incuba- tion.

The method used for evaluation of glucose up- take was rather complex and cumbersome but provided more exact and reliable data than the simple tissue solubilization and counting used elsewhere [21]. The quantitative conversion of tri- tium i n t o 3H20 , and the simple counting of this water radioactivity was found to be more efficient (and also more reliabile in the estimation of this efficiency) than other methods tested previously in our laboratory. The use of a non-metabolizable glucose analog, 2-deoxyglucose [22], resulted in the accumulation of the label in the tissue, despite the probable metabolism of glucose by the tissue. Thus, for short times and at low concentrations of the tracer (as was the case in this experimental

TABLE I

ATP AND ADP CONCENTRATIONS OF RAT BROWN ADIPOSE TISSUE INCUBATED U N D E R BASAL CONDITIONS

All data are the mean_+S.E, of 5-6 different samples. The pieces of tissue were incubated with 7.5 mM glucose. Statistical significance of the differences between groups: like superscripts indicate the absence of statistically significant (P > 0.05) differences

between the groups. Different superscripts indicate statistically significant ( P < 0.05) differences. The ANOVA column indicates the P values corresponding to the effect of the incubation time on the given parameter.

Units Incubation time (min)

0 (frozen) 0 15 30 60 ANOVA

ATP ~tmol/g 1.33_+0.23 a 0.77±0.08 a,b 0.77+0.13 b 0.72+0.09 b 0.42±0.04 h 0.0021

tissue ~tmol/g 21.9 ±3.4 a 19.9 ±2.9 a 19.3 _+3.7 a'b 16.4 4_1.7 a'b 9.4 ±1.0 b 0,0276

protein

ADP g m o l / g 1.16.+0.25 a 0.66±0.08 ~,b 0.73.+0.07 a.b 0.64±0.06 a,b 0.38_+0.05 b 0.0054

tissue p, m o l / g 18.7 _+2.9 a 18.3 _+2.4 a 18.0 ±1.6 a 14.8 ± 1 . 7 a'b 9.8 ±1.2 b 0.0071 protein

A T P / A D P ratio 1.24_+0.23 a 1.24-+0.09 a 1.06.+0.15 ~ 1.24±0.31 ~' 0.95+0.12 '̀ 0.8221

setup, in which 2-deoxyglucose was present in a proportion of 1 /44000 with respect to glucose), the ability of the tissue to retain the tracer can b e considered as an adequate measure of total glu- cose uptake.

Brown adipose tissue incorporated more glu- cose than white adipose tissue under standard comparative conditions (Fig. 1). This behavior is consistent with with the higher reputed metabolic activity of brown adipose tissue [23] as a conse- quence of its higher mitochondrial protein content [24,25]. Both types of adipose tissue actively take up glucose [26,27] and this substrate is then used for triacylglycerol synthesis [28] or, in the case of brown adipose tissue, also as a thermogenic sub- strate [11].

Insulin considerably increased the capacity of brown adipose tissue to take up glucose. This suggests that the transporter system was not saturated even at high glucose concentrations. In

glucose

in

o 2~'~/'3 e 0 0 n3M

A B

Fig. i. Glucose uptake (in nmol/s per g tissue) by brown adipose tissue pieces (A) and white adipose tissue (B) in- cubated in the presence of glucose and insulin during 20 min under otherwise standard conditions. Each point is the mean of 5-6 different determinations. Statistical significance of the effects of glucose and insulin concentrations upon glucose uptake by adipose tissues of the rat.

Tissue Two-way ANOVA One-way ANOVAs:

[glucosel [insulin] Glc/I [glucose] (insulin]

WAT 0.000 0.027 0.887 0.145 0.000 IBAT 0.000 0.000 0.000 0.011 0.000

349

contrast, the ability of white adipose tissue to incorporate glucose was soon saturated at physio- logical concentrations of glucose and insulin. The greater potential for glucose incorporation under conditions of excess glucose availability in brown adipose tissue is in agreement with its purported role in the thermogenesis-mediated maintenance of energy homeostasis [15,29]. The response of white adipose tissue to both stimuli was much less marked than that of brown adipose tissue: the highest activities found in brown adipose tissue were 6-7-fold higher than those observed in white adipose tissue. The dependence of brown adipose tissue on insulin was also higher than that of white adipose tissue.

This double dependence on insulin and glucose concentrations for glucose uptake, and oxidation [11], shown by brown adipose tissue bears a direct relationship with diet-induced thermogenesis. In the prandrial state insulin and glucose attain their highest levels and in this prandrial period, the production of heat by brown adipose tissue is maximal [30,31] under conditions of thermal neu- trality.

In Table II, the effects of insulin and ouabain on glucose and 2-aminoisobutyric acid incorpora- tion into both adipose tissue preparations are pre- sented. The uptake of glucose was higher for brown than for white adipose tissue, but the differences were smaller for 2-aminoisobutyric acid. Statisti- cal analysis of the data showed a close dependence of glucose incorporation on insulin. The effects of ouabain, alone or combined with insulin, were apparent only for 2-aminoisobutyric acid uptake in white, but not in brown adipose tissue.

The effect of changing concentrations of sodium and potassium ions in the medium on glucose or 2-aminoisobutyric acid uptake by both adipose tissues were studied. Glucose uptake by brown adipose tissue showed a relative increase at high K + and low Na + concentrations that was reversed when the concentration of both alkaline metals were increased. The addition of insulin to the medium resulted in much lower changes, with minimum activity at high Na + and low K + con- centrations. In white adipose tissue, very few changes were observed in glucose uptake with changing Na + and K + concentrations, either in the presence or absence of insulin. These results

350

TABLE II

GLUCOSE A N D 2-AMINOISOBUTYRIC ACID U P T A K E BY F R A G M E N T S OF W H I T E A N D BROWN ADIPOSE TISSUES. I N F L U E N C E OF INSULIN A N D OUABAIN

All data are the mean+S .E , of 5 -6 different samples. Glucose uptake is expresed in / zmol /g tissue, and 2-aminoisobutyric acid uptake is expressed in n m o l / g tissue. Group C, control; I, insulin; O, ouabain; and OI, ouabain plus insulin; WAT, white adipose tissue; IBAT; interscapular brown adipose tissue. Two-way ANOVA P values for control versus insulin, control versus ouabain and control versus insulin + ouabain for glucose (Glc) and 2-aminoisibutyric acid (AIB) uptake in adipose tissues are also shown.

Group Time (min) and tissue

0 0 10 10 20 20 30 30 WAT IBAT W AT IBAT WAT IBAT WAT IBAT

Glucose uptake C 0.06_+0.01 0.07+0.01 0.58_+0.05 1.31-+0.19 0.75-+0.05 2.01_+0.29 1 0.06-+0.01 0.06-+0.01 0.64-+0.04 1.55_+0.21 0.90-+0.06 2.62_+0.23 O 0.04±0.01 0.08-+0.01 0.56+0.05 1.20-+0.22 0.73_+0.06 1.91 ±0.26 OI 0.06±0.01 0.07±0.01 0.52_+0.03 1.11 -+0.13 0.80_+0.08 2.13+0.32

2-Aminoisobutyric acid uptake C 0.08±0.01 0.05_+0.01 1.40+_0.22 1.41 ±0.15 3.02-+0.33 2.21+_0.27 I 0.06±0.01 0.06_+0.01 1.82_+0.37 1.81 -+0.24 3.31 +_0.26 2.84 ± 0.59 O 0.06_+0.01 0.06_+0.01 1.35_+0.13 1.14+0.08 1.53-+0.12 1.77-+0.24 OI 0.04-+0.01 0.05_+0.01 1.19_+0.23 1.10+0.22 1.22_+0.11 1.75_+0.53

1.15+0.02 2.93_+0.15 1.34_+0.08 3.51 _+0.32 1.16+0.10 2.84_+0.28 1.19+0.08 2.67_+0.34

tissue substrate C vs. I C vs. O C vs. I + O

time I t /1 time O t / O time Ol t / O l

W A T Glc 0.000 0.005 0.187 0.000 0.774 0.986 0.000 0.797 0.678 AIB 0.000 0.268 0.657 0.000 0.002 0.001 0.000 0.001 0.001

IBAT Glc 0.000 0.021 0.379 0.000 0.606 0.990 0.000 0.566 0.796 AIB 0.000 0.158 0.564 0.000 0.095 0.395 0.000 0.253 0.683

suggest that it is unlikely that changes within the physiological range [32] in the concentrations of Na + and K + in the cells' external medium could significantly modulate the uptake of glucose by either adipose tissue. In both tissues, the uptake of glucose does not seem to be very closely related to an ouabain-sensitive sodium transport. This can also be extended to the uptake of 2-aminoisobu- tyric acid in brown adipose tissue.

In Table lII, the combined effects of changing sodium and potassium ion concentrations in the medium, as well as the presence or absence of insulin in the medium, are presented. The pres- ence of both sodium and potassium ions in- fluenced the uptake of glucose and 2-aminoisobu- tyric acid in both tissues. Analysis of the results showed that, for white adipose tissue, the uptake of glucose was dependent on the sodium ion con- centration, and that of brown adipose tissue was dependent on the sodium ion concentration and

insulin. The uptake of 2-aminoisobutyric acid was dependent on the potassium on concentration and the ratio of potasssium versus sodium ion con- centrations in white adipose tissue. In brown adipose tissue, there was a significant dependence on potassium and sodium ion concentrations modulated by insulin at a higher level than in white adipose tissue. In both adipose tissues, the maximum 2-aminoisobutyric acid uptake was at 6 mM K + and 145 mM Na +, showing lowest values at either higher or lower K + concentrations. These concentrations are those found in Krebs-Ringer bicarbonate buffer [33]. It seems that the sensitiv- ity of the A system of amino acid transport to the alkaline metals concentration in the medium is much higher that that of the glucose transporter system.

Glucose transport into some tissues, such as the gastrointertial tract and kidney, is sodium ion dependent [12]. However, this is not a universal

351

TABLE III

GLUCOSE AND 2-AMINOISOBUTYRIC ACID UPTAKE BY FRAGMENTS OF WHITE AND INTERSCAPULAR BROWN ADIPOSE TISSUE: INFLUENCE OF [Na + ] [K ÷ ] AND INSULIN

All data are the mean_+ S.E. of 5-6 different samples. Glucose uptake is expressed in ttmol/s per tissue (in 20 mln of incubation) 2-aminoisobutyric acid uptake is expressed in nmol/s per g tissue (in 10 min of incubation). Abbreviations as in Table II. Three-way ANOVA P values for the effect on either glucose or 2-aminoisobutyric acid uptake by rat adipose tissues are also shown.

[K + ] (mM) Substrate ]Insulin] [Na + ] (mM) and tissue

25 25 145 145 265 265 WAT IBAT WAT IBAT WAT !BAT

0.21 G1 0 0.57_+0.04 1.93_+0.19 0.58_+0.03 2.37_+0.19 0.65_+0.01 1.48_+0.33 0.2 0.69_+0.07 2.53_+0.38 0.64_+0.08 2.48_+0.18 0.62_+0.10 1.74_+0.22

AIB 0 1.65_+0.17 1.72_+0.30 1.49_+0.10 1.60_+0.22 1.49-+0.15 2.10_+0.14 0.2 1.40_+0.18 1.77_+0.17 1.30_+0.17 1.65_+0.10 1.44_+0.18 2.51 _+0.30

6 GI 0 0.83_+0.11 2.13_+0.71 0.62_+0.04 1.58_+0.13 0.48-+0.05 1.48_+0.17 0.2 0.63_+0.02 2.40-+0.33 0.73_+0.06 1.88-+0.10 0.58-+0.04 2.29_+0.51

AIB 0 1.85_+0.03 1.89_+0.07 2.09-+0.37 2.35_+0.25 2.04-+0.07 2.29_+0.25 0.2 2.07_+0.28 1.77+_0.20 3.04_+0.56 3.02_+0.40 2.34_+0.27 3.11_+0.17

21 G1 0 0.74+0.04 2.81 _+0.49 0.64-+0.06 1.52_+0.33 0.74_+0.15 1.57_+0.16 0.2 0.69_+0.14 2.48_+0.28 0.70_+0.08 2.02_+0.49 0.44_+0.02 1.91 _+0.18

AIB 0 2.25_+0.17 1.50_+0.17 1.99_+0.40 2.27_+0.13 2.07_+0.13 2.61_+0.20 0.2 2.44_+0.17 1.50_+0.32 1.44_+0.05 1.99_+0.10 2.27_+0.07 2.15_+0.30

t issue substrate [insulin] [Na + ] [K + ] I/Na + I /K + Na+/K + I/Na +/K +

WAT glucose 0.427 0.024 0.271 0.360 0.070 0.671 0.125 AIB 0.928 0.867 0.000 0.806 0.140 0.029 0.209

IBAT glucose 0.028 0.016 0.621 0.533 0.701 0.187 0.720 AIB 0.320 0.000 0.001 0.431 0.013 0.032 0.325

situation, as there are other glucose t ranspor ter

systems that are not dependent on sodium ions [34]. In the case of b rown and white adipose tissues, the low response of glucose uptake to

sodium ion, compared to that of 2-aminoisobu- tyric acid incorporat ion, suggests that a significant par t of the glucose t ransport would be indepen-

dent of sodium. The higher sensitivity of 2-aminoisobutyr ic acid

t ransport to physiological concent ra t ions of in- sulin and to changes in alkaline metal ions, to- gether with the comparable activities found in

both adipose tissues, suggest a secondary role of amino acid uptake in these tissues, definitely not

comparable to glucose uptake, that takes place at rates two orders of magni tude higher. These data agree with a minor role for amino acids as meta- bolic substrates in both tissues. It is likely that sod ium-dependent t ransport of amino acids is not extensively used to provide substrates to fuel either thermogenesis or lipogenesis, despite the known

high capacity of brown adipose tissue to metabo- lize amino acids [35].

In conclusion, b rown adipose tissue is very similar to white adipose tissue with respect to

sodium ion-dependen t an imo acid uptake. How- ever, its versatile and adaptive response to glucose

and insul in give to this thermogenic site the abil i ty to extract very impor tan t amounts of glucose from the blood stream to fuel its heat product ion. This

capabil i ty would probably be of the highest im- por tance in the control of die t - induced thermo- genesis in prandia l states.

References

1 Cannon, B. and Nedergaard, J. (1982) in The Biochemical Development of the Fetus and the Neonate (Jones, C.T., ed.), pp. 697-727 Elsevier, Amsterdam.

2 Foster, D.O. (1986) in Brown Adipose Tissue (Trayhurn, P. and Nicholls, D.G., eds.), pp. 31-51, Edward Arnold, London.

3 Foster, D.O. and Frydman, M,L. (1978) Can. J. Physiol. Pharmacol. 56, 110-112.

352

4 Mampel, T. and Herrera, E. (1985) Diabet. Metab. 11, 118-124.

5 Young, J.B., Saville, E., Rothwell, N.J. and Stock, M.J. (1984) FEBS Lett. 176, 16-20.

6 Radomski, M.W. and Orme, T. (1971) Am. J. Physiol. 220, 1852-1856.

7 Agius, L. and Williamson, D.H. (1981) Biochim. Biophys. Acta 666, 127-132.

8 Matsusashita, M., Kobayashi, K. and Kutsumi, H. (1980) Adv. Physiol. Sci. 32, 523-526.

9 Foster, D.O. and Frydman, M.L. (1978) Can. J. Physiol. Pharmacol. 57, 257-270.

10 Rothwell, N.J. and Stock, M.J. (1979) Nature 281, 31-35. 11 Cooney, G.D. and Newsholme, E.A. (1984) Trends Bio-

chem. Sci. 9, 303-305. 12 Glieman, J. (1982) Biochem. Soc. Trans. 10, 7-9. 13 Newsholme, E.A. and Leech, A.R. (1983) in Biochemistry

for the Medical Sciences, pp. 562-578, John Wiley & Sons, Chichester.

14 Desautels, M. and Himms-Hagen, J. (1979) Can. J. Bio- chem. 57, 968-976.

15 Cooney, G.D. and Newsholme, E.A. (1982) FEBS Lett. 148, 198-200.

16 Landsberg, L. and Young, J.B. (1983) in Mammalian Ther- mogenesis (Girardier, L. and Stock, M.J., eds.), pp. 126-128, Chapman & Hall, London.

17 Rothwell, N.J. and Stock, M.J. (1981) Pfliiger's Archs. 389, 237-242.

18 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R. (1951) J. Biol. Chem. 135, 265-275.

19 Lowry, O.H. and Passoneau, J.V. (1972) A Flexible System of Enzymatic Analysis, pp. 147-156, Academic Press, New York.

20 Dixon, W.J., et al. (1983) BMDP Statistical Software, Uni- versity of California Press, Berkeley.

21 Peng, C.T. (1977) Sample Preparation in Liquid Scintilla- tion Counting, RCC Review 17, pp. 42-48, Amersham, Wembley.

22 Craig, B.W. and Treadway, J. (1986) J. Appl. Physiol. 60, 1704-1709.

23 Cannon, B. (1979) Eur. J. Biochem. 23, 125-135. 24 Smith, R.E. and Horwitz, R.A. (1969) Physiol. Rev. 49,

330-425. 25 Girardier, L. (1983) in Mammalian Thermogenesis (Girar-

dier, L. and Stock, U.J., eds.), pp. 50-91, Chapman & Hall, London.

26 Leturque, A., Burnol, A. Ferr6, P. and Girard, J. (1984) Am. J. Physiol. 246, E24-E31.

27 Czech, M.D., Lawrence, J.C. and Lynn, W.C. (1974) J. Biol. Chem. 249, 5421-5427.

28 McCormack, J.G. (1982) Prog. Lipid Res. 21, 195-223. 29 Himms-Hagen, J. (1983) in The Adipocyte and Obesity:

Cellular and Molecular Mechanisms (Hallenberg, A.A. and Roncari, D.A.K., eds.), pp. 259-270, Raven Press, New York.

30 Glick, Z., Teague, R.J. and Bray, A. (1981) Science 213, 1125-1127.

31 Glick, Z., Bray, A. and Teague, R.J. (1984) J. Nutr. 114, 286-291.

32 Romeu, A., Arola, L. and Alemany, M. (1986) Biol. Res. Pregnancy 7, 52-55.

33 Umbreit, W.W., Burris, R.H. and Stauffer, J.F. (1964) in Manometric Techniques, p. 132, Burger, Minnesota.

34 Carruthers, A. (1984) Prog. Biophys. Mol. Biol. 41, 33-69. 35 L6pez-Soriano, F. and Alemany, M. (1986) Biochem. Int.

12, 471-478.