Induction of aP2 gene expression by nonmetabolized long-chain fatty ...

Proc. Nati. Acad. Sci. USA Vol. 89, pp. 10930-10934, November 1992 Biochemistry

Induction of aP2 gene expression by nonmetabolized long-chain fatty acids (adipocyte/fatty-acyl-CoA synthase/differentiation)

PAUL A. GRIMALDI*, SUSAN M. KNOBELt, RICHARD R. WHITESELLt,

AND

NADA A. ABUMRAD*tt

tDepartment of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, TN 37232; and *Centre National de la Recherchd Scientifique, Parc Valrose 06034, Nice Cedex, France Communicated by Charles R. Park, July 10, 1992

Long-chain fatty acids (FA) have been shown ABSTRACT to regulate expression of the gene for the adipocyte FA-binding protein aP2. We examined whether this effect was exerted by FA themselves or by a FA metabolite. The a-bromo derivative of palmitate, an inhibitor of FA oxidation, was synthesized in the radioactive form, and its metabolism was investigated and correlated with its ability to induce aP2 in Ob1771 preadipocytes. a-Bromopalmitate was not utilized by preadipocytes. It was not cleared from the medium over a 24-hr period and was not incorporated into cellular lipids. Short incubations indicated that a-bromopalmitate exchanged across the preadipocyte membrane but remained in the free form inside the cell. In line with this, preadipocyte homogenates did not activate a-bromopalmitate to the acyl form. However, although it was not metabolized, bromopalmitate was much more potent than native FA in inducing aP2 gene expression. Induction exhibited the characteristics previously described for native FA, indicating that a similar if not identical mechanism was involved. The data indicated that induction of aP2 was exerted by unprocessed FA. Finally, in contrast to preadipocytes, adipocytes metabolized bromopalmitate. This reflected increased activity with cell differentiation of a palmitoyl-CoA synthase that could activate palmitate and bromopalmitate at about one-fifth the rate for palmitate. In preadipocytes, the predominant fattyacyl-CoA synthase, arachidonyl-CoA synthase, had very low affinity for both FA. Increased activity of the palmitoyl-CoA synthase, which has a wider substrate range, is likely to be important for initiation of lipid deposition. Fatty acids (FA) are important energy substrates and precursors of triglycerides, phospholipids, and prostaglandins. FA can also affect many aspects of cell function. For example, FA activate guanylate cyclase (1), protein kinase C (2), and K+ channels (3) and inhibit chloride channels (4). Some effects of FA involve FA metabolites such as products of arachidonic acid metabolism and fatty acyl-coenzyme A complexes (5,6). Recently, however, multiple effects of FA have been shown to be exerted by unprocessed FA directly

(2,3,7).

We have described regulatory effects of long-chain FA on gene expression. When added to the culture media of two lines of preadipocytes, Ob1771 (8) and BFC-1 (9), FA induced or enhanced expression of the cytosolic FA binding protein [aP2; also called adipocyte long-chain FA binding protein (ALBP)]. This low molecular weight protein is postulated to play various roles in intracellular transport of FA and might be involved in the insulin signaling pathway in adipocytes (10-12). In this report, we investigated whether the effect of long-chain FA to induce aP2 in Ob1771 was exerted by FA themselves or by FA metabolism. This information would be essential to elucidate the molecular mechanism of the FA

effect. Furthermore, a similar regulatory effect of FA might apply to other genes induced during adipose differentiation. For example, in Ob1771 preadipocytes, FA also induced the gene for fatty-acyl-CoA synthase (EC 2.3.1.86) (8). In BFC-1 preadipocytes, an increase in membrane transport of FA was an early event during cell differentiation, which would be consistent with FA playing a role in modulating subsequent differentiation steps (9). Our results, using a-bromopalmitate, indicate that induction of aP2 in preadipocytes is exerted by unprocessed FA. The data also demonstrate increased expression, with cell differentiation, of a palmitoyl-CoA synthase that has a wide range of FA substrates (which include a-bromopalmitate) and that is likely to be important for the initiation of adipogenesis.

MATERIALS AND METHODS Materials. Guanidinium isothiocyanate and films for autoradiography (X-Omat, AR) were from Kodak. Cesium chloride was from Fisher Scientific. The multiprime labeling kit for cDNA, the [32P]dCTP, and the nitrocellulose filters were from Amersham. The cDNA for aP2 (13) was a gift from H. Green (Harvard Medical School, Cambridge, MA). Enzymes for nucleic acid manipulations were from Promega. Thinlayer plates were from Analtech. All organic chemicals were from Fisher Scientific and were of the highest grade available. Lipid standards for TLC and all FA were from Sigma. Unlabeled a-bromopalmitate was from K & K Laboratories. [3H]Oleate and [14C]palmitate were from NEN. Anti-aP2 antibody was a gift from D. Bernlohr (University of Minnesota). Experimental Procedures. Cell culture. Ob1771 (14) cells were plated at a density of 2 x 103 per cm2 and grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 8% fetal bovine serum, 200 units of penicillin per ml, 50 ug of streptomycin per ml, 33 juM biotin, and 17 nM panthotenate. Confluence (about 1-1.5 x 106 cells per 35-mm dish) was reached within 5 days. To promote differentiation of Ob1771, 2 nM triiodothyronine, 1.2 nM growth hormone, and 20 nM insulin were added at confluence (day 0). Cell treatments. FA were added 1-2 days after cell confluence and for the time periods indicated. They were dissolved in a small volume of ethanol (final concentration, 90% was recovered in solution. The concentrations of FA added varied between 50 and 200 .M for palmitate and between 1 and 200 AM for bromopalmitate. An additional 40 ,uM was contributed by the serum. In experiments where 0.1% serum was used, FA and albumin carried over from the serum were about 0.5 and 0.75 AM, respectively, so the amounts of FA added were correspondingly scaled down. The molar ratios of FA to albumin ranged between 0.8 and 4.4, with the unbound FA, calculated as described (15), ranging between 0.04 and 5

jLM.

Synthesis of radioactive bromopalmitate. Unlabeled and/or labeled (14C or 3H) palmitate (1-10 mg) were allowed to react with bromine (2-fold molar excess) and phosphorus tetrachloride in a 50-/l glass capillary at 1200C for 3 hr. The products were recovered in acetone (5 ml) and separated by chromatography on silica gel G with chloroform/methanol/ acetic acid, 80:20:0.02 (vol/vol). Under these conditions of very low acetic acid, a-bromopalmitate could be separated from palmitate and from dibromopalmitate (Rf values were 0.51, 0.66, and 0.31, respectively). However, when the ratio of acetic acid was increased, bromopalmitate migrated with palmitate. Bromopalmitate was recovered by several extractions of the silica with chloroform/methanol, 80:20 (vol/vol). The extract was washed extensively with water (about 10 times) to remove all traces of silicic acid. Product purity was >90%o as tested by a second chromatography on silica G, and the yield was -80%. Specific activity of synthesized bromopalmitate ranged between 10 and 100AGCi (1 ACi = 1 kBq) per ,mol. Utilization assays. Cells were incubated at 370C with the indicated concentrations of labeled and unlabeled FA for the indicated times. At the end of the incubation period, the medium was sampled for assay of radioactivity. The cells were washed in cold phosphate-buffered saline (PBS; 137 mM NaCl/1.5 mM KH2PO4/7 mM Na2HPO4/2.7 mM KCl) to remove FA and albumin and were extracted for lipid by the method of Bligh and Dyer (16). Lipids were subjected in some experiments to TLC on silica gel G (petroleum ether/diethyl ether/acetic acid, 80:20:1 (vol/vol), to determine the distribution of radioactivity into the various lipid fractions (17). The concentration of FA in the medium or in the cells was determined by the 63Ni assay of Ho and Meng (18) as described (17). Measurements of the uptake and equilibration (0-2 min) of bromopalmitate were conducted as described for oleate (9). Intracellular water was estimated in 90-min incubations from the difference between sugar-accessible space for 3-0[3H]methylglucose and [14C]sucrose space, as described for cultured cells (17). At confluence, Ob1771 cells contained about 2.0 1.d of intracellular water (per 1 x 106 cells). Measurement offatty-acyl-CoA synthase activity. Before the assay the cells were washed at 370C with a Krebs-Ringer Solution buffered with Hepes (KRH) containing 1% bovine serum albumin (BSA) and then kept in this buffer at 370C for 20 min to deplete intracellular FA and fatty-acyl-CoA. The cells were washed with albumin-free KRH at 37C and then twice with ice-cold PBS. This treatment was found to remove efficiently all FA from the medium and the cells. The cells were then homogenized in cold buffer containing 25 mM Tris HCl (pH 8.0), 150 mM KCl, and 1 mM dithiothreitol. Fatty-acyl-CoA synthase activity was assayed (19) at 37°C (0.1-10 min) by using 0.05-0.3 mg of homogenate protein in a total volume of 100 ,u containing 100 mM Tris HCI (pH 8.0), 20 mM MgCl2, 5 mM dithiothreitol, 6.7 mM ATP, 0.7 mM CoA, and the indicated concentrations of [3H]- or [14C]palmitate, a-bromopalmitate, or arachidonate (200,000 cpm) added in Triton X-100 (final concentration, 0.1%). Enzyme activity was determined from the linear portions of the reaction time courses (0.2-5 min).

10931

x g for 45 min, and 100 ,ug of the supernatant was used for Western blots and immunodetection of aP2 as described by Bernlohr et al. (21) except for the use of a Vectastain ABC-AP kit (Vector Laboratories).

RESULTS

Metabolism of a-Bromopalmitate by Ob1771 Preadipocytes and Adipocytes. a-Bromopalmitate is a known inhibitor of FA oxidation in liver and heart tissues (22) that acts at the level of the enzyme carnitine palmitoyltransferase (23). A study by Pandy and colleagues (24) with rat liver preparations suggested that a-bromopalmitate at high concentrations can also inhibit FA activation. We investigated the metabolism of

a-bromopalmitate by Ob1771 preadipocytes and adipocytes. Preadipocytes (Fig. 1 Upper) did not utilize any a-bromo[C14]palmitate during a 24-hr incubation period, whereas 50% of medium [3H]palmitate was utilized in about 4 hr (Fig. 1 Upper Left). Cell-associated radioactivity in cultures of preadipocytes, incubated with [14C]bromopalmitate, was very low (Fig. 1 Upper Right) and was all recovered as the unprocessed FA (data not shown). In contrast to preadipocytes, differentiated Ob1771 (Fig. 1 Lower) utilized a-bromo[14C]palmitate as evidenced from its disappearance from the medium (Fig. 1 Lower Left) and from its incorporation into cell lipids (Fig. 1 Lower Right). However, its rate of utilization was slower than that of palmitate. In adipocytes incubated with bromopalmitate for 6 hr, the percent of radioactivity recovered in phospholipids was similar to that in cells incubated with palmitate; however, the percent radioactivity recovered in triglycerides was decreased about 30%6, while percentages recovered in FA and in diglyceride 100 75

50

.e

25 0

V) 0

0

6

12

18

24

D

0 To

0

0

ED0.

0)

12..5 101 7,F5 5( 218

io

W60e

a

6

12

18

24

E

C

C

100 E 0 75 50

U-:1

25 0

6

12

18

24

Time, hr

FIG. 1. Utilization of 200 ,uM bromopalmitate (-) or palmitate (-) by Ob1771 preadipocytes (Upper) and adipocytes (Lower). Radioactivities remaining in the medium (Left) and those incorporated into cell lipids (Right) are shown for the incubation times indicated on the x axes. Data are typical of two other experiments.

10932

Biochemistry: Grimaldi et al.

were increased by 10- and 20-fold, respectively. This suggested that conversion of diglycerides containing bromopalmitate to triglycerides was slow. In addition, utilization of a-bromopalmitate by adipocytes appeared to stop when >50%o of the substrate remained in the medium (Fig. 1 Lower). This contrasted with palmitate utilization, which continued until >75% of the FA had been removed. This discrepancy most likely reflected a higher Km for bromopalmitate activation by the fatty-acyl-CoA synthase. Although preadipocytes did not utilize bromopalmitate, its cellular uptake could be demonstrated in short 0.1- to 2-min incubations (Fig. 2) by a rapid equilibration of a-bromo[14C]palmitate across the cell membrane. At a FA/BSA ratio of 4.0, the amount of bromo[14C]palmitate accumulated at equilibrium and corrected for that remaining in the extracellular space and for that bound to the cells averaged 3 nmol/,ld of intracellular water space, which was equivalent to an intracellular concentration of about 3 ,tM (intracellular water = 2 ,ud). This approximates the concentration of unbound palmitate in solution at a FA/BSA ratio of 4.0 and would suggest that bromopalmitate binds to BSA with an affinity that is similar to that of palmitate. Fig. 2 also illustrates why no detectable uptake of bromopalmitate could be measured in Fig. 1, as the free bromopalmitate present in the cell (6 nmol) was minute compared with that present in the medium (400 nmol). Activity of Fatty-Acyl-CoA Synthase Against a-Bromopalmitate. Preadipocyte homogenates activated (i.e., formed CoA complexes with) palmitate but not bromopalmitate (Table 1). No detectable formation of acyl-bromopalmitate could be measured even when the concentration of homogenate protein and the specific activity of bromo[14C]palmitate were increased >5-fold above those used for palmitate. In contrast, homogenates from adipocytes activated bromopalmitate at about one-fifth the rate of palmitate activation. The predominant fatty-acyl-CoA synthase in preadipocytes activated arachidonate (Table 1). This enzyme had low affinity for palmitate, bromopalmitate, or oleate because high concentrations of these FA did not compete for activation with tracer arachidonate (data not shown). In adipocytes, a large fraction of the fatty-acyl-CoA synthase activity present activated palmitate, oleate (not shown), and to a lesser extent bromopalmitate (Table 1). Induction of aP2 Gene Expression by a-Bromopalnitate. The effects of a-bromopalmitate and palmitate on the level of aP2 gene expression were compared in 1-day postconfluent Ob1771 cells. These cells were committed to differentiate since they expressed mRNA for pOb24 (25); however, they

Proc. Natl. Acad. Sci. USA 89 (1992) Table 1. Activity of fatty-acyl-CoA synthases in preadipocytes and adipocytes, measured with palmitate, bromopalmitate, and arachidonate as substrates Fatty acyl-CoA, nmol/min per mg of protein Arachidonate Palmitate a-Bromopalmitate ± 68 ± 4 3.2 0.4 0 Preadipocytes 55 ± 3 66 ± 1.7 10.7 + 0.4 Adipocytes Fatty-acyl-CoA synthase activity was assayed with 50-300 ,jg of homogenate protein in the presence of 200 j.M [14C]palmitate, bromo[14C]palmitate or [14C]arachidonate. Activity of fatty-acylCoA synthase against a-bromopalmitate in preadipocytes was undetectable at all protein concentrations used. Data are means (n = 3-5) ± SEM.

still lacked expression of the aP2 gene (Fig. 3, lane a) and that of other differentiation markers (25). Addition of palmitate led to the appearance of aP2 mRNA. The effect of an optimal concentration (200 .uM) is shown in Fig. 3, lane b. Exposure of the cells to the same concentration of a-bromopalmitate resulted in a much stronger induction of aP2 mRNA (Fig. 3, lane c). Both treatments did not affect mRNA levels for pOb24 and GAPDH used as internal standards. The increases in aP2 mRNA were associated with parallel and proportional increases in aP2 protein (Fig. 3 Right). In contrast to bromopalmitate, another inhibitor of FA oxidation, 2-mercaptopropionic acid (10-300 ,uM) was ineffective in inducing aP2. This indicated that induction by bromopalmitate was not mediated by inhibition of FA oxidation and was consistent with the requirement for a long hydrocarbon chain (8). Accumulation of aP2 mRNA was observed with bromopalmitate starting at 10 ,M, whereas 50 jM palmitate was barely effective (Fig. 4). The higher potency of bromopalmitate was not related to different affinities of albumin for the two FA, as this would imply that unbound bromopalmitate was 15-20 times higher than unbound palmitate at the same FA/BSA ratio. Such an interpretation is not supported by the similar and low solubilities of the two FA. The time course of aP2 mRNA accumulation in cells exposed to 50 ;LM bromopalmitate is shown in Fig. 5. The amount of aP2 mRNA, after an initial lag period of a few hours, increased linearly with time to levels that were 75-fold higher at 48 hr than those in control cells. Removal of bromopalmitate after a 24-hr treatment led to a decrease in aP2 mRNA with a half time of 12-14 hr, which is similar to the turnover time of aP2 mRNA in cultured adipocytes (26). The time course and reversibility of aP2 induction by bromopalmitate were similar to previous observations with native FA (27). In addition, as with native a o u1 GAPDH a

pOb24 Time, min FIG. 2. Uptake and equilibration of bromopalmitate during short incubations. Ob1771 cells were incubated with 40 gM bromo[3H]palmitate (3000 cpm/,ul) bound to BSA (FA/BSA = 4). The medium also contained [14C]sucrose (800 cpm/,ul). Intracellular bromopalmitate was obtained from cell-associated tritium values corrected for bromo[3H]palmitate that is trapped in the extracellular medium or bound to the outside of the cells. This was estimated from the remaining 14C radioactivity as well as from zero-time controls to which isotope and stop solution were added simultaneously. Data are pooled from two experiments.

cc -

23 kDa

aP2

FIG. 3. Effect of bromopalmitate and palmitate on induction of aP2 mRNA and protein in Ob1771 preadipocytes. One-day postconfluent cells were maintained for 24 hr in medium containing 8% BSA in the absence (lanes a) or presence of 200 ,uM palmitate (lanes b) or bromopalmitate (lanes c). (Left) Northern blots (5 Ag of RNA per lane) showing aP2 mRNA and the mRNAs for pOb24 and glyceraldehyde-phosphate dehydrogenase (GAPDH) used as controls. (Right) Levels of aP2 protein under the same conditions. Data are typical of at least two other experiments.

Proc. Natl. Acad. Sci. USA 89 (1992)


10

0

10

0

30

100

300

CONCENTATION (R

FIG. 4. Comparison of the potency of bromopalmitate and palmitate in inducing aP2 mRNA. One-day postconfluent cells were kept for 24 hr in DMEM with 8% BSA and without (o) or with the indicated concentrations of bromopalmitate (o) or palmitate (v). aP2 mRNA was quantitated and normalized to mRNAs for pOb24 and GAPDH as described in text. Data are typical of those from three experiments.

FA, bromopalmitate acted synergistically with 1 ,uM dexamethasone to induce aP2 gene expression (Table 2). Finally, induction by 50 uM bromopalmitate for 8 hr, as previously demonstrated for natural FA (27), was completely abolished by cycloheximide (data not shown). These observations indicated that induction of aP2 by bromopalmitate involved a mechanism that was similar, if not identical, to that involved in the effect of native FA. Additional experiments were conducted to determine if part of the bromopalmitate effect was mediated by interaction with FA in the medium. For example, a-bromopalmitate would compete with natural FA for binding to BSA, resulting in higher concentrations of unbound native FA in the medium. However, reducing the serum in the medium to 0.1%, which reduced concentrations of both FA and BSA by 99%o, resulted in a leftward shift in the dose-response curve for bromopalmitate (Table 3). This most likely reflected higher levels of free unbound bromopalmitate in incubations with 0.1% versus 8% serum. These results demonstrated that interactions with medium FA did not mediate the effect of bromopalmitate on aP2.

DISCUSSION This study demonstrated that a nonmetabolized long-chain FA induced expression of the aP2 gene in preadipocytes. We 10

.3

8

1 64 0

6

u 611.

4

10933

Table 2. Synergistic effects of a-bromopalmitate and of dexamethasone in inducing aP2 mRNA in Ob1771 preadipocytes aP2 mRNA, arbitrary Treatment integrator units Control 0.08 Dexamethasone (1 uM) 0.31 a-Bromopalmitate (50 ,uM) 3.9 + dexamethasone 9.9 One-day postconfluent cells were maintained in DMEM containing 8% BSA with or without the indicated agent(s). Total RNA was analyzed for aP2 mRNA by scanning with a laser densitometer. Scans were normalized for mRNAs for pOb24 or for GAPDH. Data are representative of two other experiments.

have presented evidence to indicate that the FA derivative, a-bromopalmitate, was not metabolized by preadipocytes. Preadipocytes did not incorporate any radioactive bromopalmitate and preadipocyte homogenates did not catalyze the formation of any bromopalmitoyl-CoA. However, a-bromopalmitate was more potent than palmitate in inducing the aP2 gene. This reflected bromopalmitate accumulation in the unprocessed form inside the cell, and it explained why relatively high concentrations of palmitate and other native FA, which are metabolized by preadipocytes, were needed to induce aP2 (8). Thus, the results rule out the possibility that induction by FA was mediated by FA metabolites or by general effects of FA metabolism on cellular energetics. The findings suggest that the increase in FA transport, observed early during differentiation of preadipocytes (9), might play a role in aP2 induction. This report describes an action of unprocessed FA on gene expression. Direct effects of FA have been documented on membrane K+ and Cl- channels (3,4) and on membrane protein kinase C (2) and guanylate cyclase (1). FA have also been reported to inhibit tyrosine kinase activity of the insulin receptor (28). It is possible that one of these mechanisms mediates the action of FA to induce aP2. Alternatively, interaction of FA with the signaling pathways for insulin or insulin-like growth factor 1, which induce aP2 in differentiating preadipocytes (29,30), might be involved. aP2 is also induced by dexamethasone in cultured preadipocytes (31); however, effects of FA and of dexamethasone on aP2 exhibited synergism (Table 2 and ref. 8) and different time courses (8) and thus would appear to involve separate mechanisms. Our data suggest a potentially important regulatory role for unprocessed FA. Glycerolipid synthesis, the main route of FA disposal in preadipocytes and adipocytes, saturates at low concentrations of unbound FA (Km = 0.01-0.06 ,uM) that are within the range of physiological values (32). At saturating unbound FA, unprocessed FA would accumulate inside the cell and, as shown, would promote induction of aP2 and

.4

9

Table 3. Induction of aP2 mRNA by a-bromopalmitate in preadipocytes maintained in DMEM with 0.1% bovine serum aP2 mRNA, arbitrary Bromopalmitate, uM integrator units 0 0.02 1 0.03

O ..1

2 0 0

12

24

36

48

TIME (hours)

FIG. 5. Time course for induction of aP2 mRNA by a-bromopalmitate. Cells were kept for the times indicated in DMEM with 8% BSA without (o) or with 50 MuM a-bromopalmitate (0). a-Bromopalmitate was removed by two washes (o) with DMEM containing 8% BSA. The cells were then incubated in the same medium for 8 or 16 hr. Data are representative of three experiments.

5 10 25

0.10 0.41 0.95 One-day postconfluent cells were washed twice with 10 ml of serum-free DMEM and then incubated for 15 hr in DMEM with 0.1% bovine serum without or with bromopalmitate at the indicated concentrations. mRNA for aP2 was quantitated and normalized for pOb24 mRNA as described in the text and in Table 2. Data are representative of two experiments from two different cell series.

10934


possibly that of other differentiation markers such as fattyacyl-CoA synthase (8). This suggests that FA can promote adipose conversion of preadipocytes in addition to enhancing lipid accumulation in adipocytes by providing a limiting substrate. Differentiation of preadipocytes was associated with an increase in the activity of a fatty-acyl-CoA synthase that could activate palmitate and to a lesser extent bromopalmitate. This enzyme appeared similar to the one present in isolated rat adipocytes that metabolized bromopalmitate at about one-fourth the rate for palmitate (unpublished observations). In preadipocytes, the predominant synthase present appeared highly specific for arachidonic acid (unpublished observations) and is likely to be similar to the enzyme discovered by Wilson et al. (33) in human platelets. Activity of this synthase was not altered by cell differentiation, whereas activity of the palmitoyl-CoA synthase was greatly increased (Table 1). This might explain how adipose differentiation can be associated with a 15-fold increase in the synthesis of diglycerides and triglycerides while that of phospholipids remains largely the same (17). The increase in palmitoyl-CoA synthase activity would contribute, together with that in stearoyl-CoA desaturase (34), to the changes in FA composition of membrane phospholipid observed during cell differentiation (35). These changes would alter membrane properties (35) and possibly the function of membrane proteins. More significantly, the increase in palmitoyl-CoA synthase is likely to be a determining step for initiation oflipid deposition since this enzyme activates a wide range of FA substrates, which include those (oleate and palmitate) preferentially incorporated into triglycerides. The cellular distribution of palmitoyl-CoA synthase (36) would be consistent with this interpretation. Its activity is high in adipose and liver cells, which are concerned with synthesis of triacylglycerols, in contrast to most other cells where activity of the arachidonyl-CoA synthase predominates. This work was supported by a grant from the National Institutes of Health (DK 33301) and by a private grant from the Taher Foundation for Medical Research. 1. Braughler, J. M., Mittal, C. K. & Murad, F. (1979) Proc. Nadl. Acad. Sci. USA 76, 219-222. 2. Murakami, K., Chan, S. Y. & Routtenberg, A. (1986) J. Biol. Chem. 261, 15424-15429. 3. Ordway, R. W., Walsh, J. V., Jr., & Singer, J. J. (1989) Science 244, 1176-1179. 4. Andersen, M. P. & Welsh, M. J. (1990) Proc. Natl. Acad. Sci. USA 87, 7334-7338. 5. Kim, D. & Clapham, D. (1989) Science 244, 1174-1176. 6. Block, K. & Vance, D. (1977) Annu. Rev. Biochem. 46,

263-298. 7. Ordway, R. W., Singer, J. J. & Walsh, J. V., Jr. (1991) Trends Neurosci. 14, 96-100.

Proc. Natl. Acad Sci. USA 89 (1992) 8. Amri, E. Z., Bertrand, B., Ailhaud, G. & Grimaldi, P. A. (1991) J. Lipid Res. 32, 1449-1456. 9. Abumrad, N. A., Forest, C. C., Regens, D. M. & Sanders, S. (1991) Proc. Natl. Acad. Sci. USA 88, 6008-6012. 10. Waggoner, D. W. & Bernlohr, D. A. (1990) J. Biol. Chem. 265, 11417-11420. 11. Bernier, M., Laird, D. M. & Lane, D. M. (1987) Proc. Natl. Acad. Sci. USA 84, 1844-1848. 12. Lane, D. M., Flores-Riveros, J. R., Hresko, R. C., Kaestner, K. H., et al. (1990) Diabetes Care 13, 567-575. 13. Spiegelman, B. M., Frank, M. & Green, H. (1983) J. Biol. Chem. 258, 10083-10089. 14. Negrel, R., Grimaldi, P. & Ailhaud, G. (1978) Proc. Natl. Acad. Sci. USA 75, 6054-6058. 15. Abumrad, N. A., Perkins, R. C., Park, J. H. & Park, C. R. (1981) J. Biol. Chem. 256, 9183-9191. 16. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 17. Abumrad, N. A., Forest, C. C., Regen, D. M., Melki, S. A. & Barnella, U. S. (1991) Am. J. Physiol. 261, E76-E86. 18. Ho, R. J. & Meng, H. C. (1969) Anal. Biochem. 31, 426-436. 19. Samuel, D. & Ailhaud, G. (1969) FEBS Lett. 2, 213-217. 20. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156-159. 21. Bernlohr, D. A., Doering, T. L., Kelly, T. J. & Lane, M. D. (1985) Biochem. Biophys. Res. Commun. 132, 850-855. 22. Burges, R. A., Butt, W. D. & Baggaley, A. (1968) Biochem. J. 109, 39-42. 23. Schultz, H. (1990) in FA Oxidation, Clinical, Biochemical and Molecular Aspects, eds. Tanaka, K. & Goates, P. (Liss, New York), pp. 153-165. 24. Pande, S. V., Siddiqui, A. W. & Gattereau (1971) Biochim. Biophys. Acta 248, 156-166. 25. Dani, C., Doglio, A., Amri, E. Z., Bardon, S., Fort, P., Grimaldi, P. & Ailhaud, G. (1989) J. Biol. Chem. 264, 1011910125. 26. Cook, K. S., Hunt, C. R. & Spiegelman, B. M. (1985) J. Cell Biol. 100, 514-520. 27. Amri, E. Z., Ailhaud, G. & Grimaldi, P. A. (1991) J. Lipid Res. 32, 1457-1463. 28. Buelt, M. K., Shekels, L. L., Jarvis, B. W. & Bernlohr, D. A. (1991) J. Biol. Chem. 266, 1226-1231. 29. Smith, P. J., Wise, L. S., Berkowitz, R., Wan, C. & Rubin, C. S. (1988) J. Biol. Chem. 263, 9402-9408. 30. Blake, W. L. & Clarke, S. D. (1990) Biochem. Biophys. Res. Commun. 173, 87-91. 31. Cook, J. S., Lucas, J. J., Bolanowski, M. A., Christy, R. J. & Kelly, T. J. (1988) Proc. Natl. Acad. Sci. USA 85, 2949-2953. 32. Melki, S. A. & Abumrad, N. A. (1992) J. Lipid Res. 33, 669-678. 33. Ntambi, J. M., Buhrow, S. A., Kaestner, K. H., Christy, R. J., Sibley, E., Kelly, T. J. & Lane, D. M. (1988) J. Biol. Chem. 263, 17291-17300. 34. Storch, J., Shulman, S. L. & Kleinfeld, A. M. (1989) J. Biol. Chem. 264, 10527-10533. 35. Wilson, D. B., Prescott, S. M. & Majerus, P. W. (1982) J. Biol. Chem. 257, 3510-3515. 36. Laposta, M., Reich, E. L. & Majerus, P. (1985) J. Biol. Chem. 260, 11016-11020.