Utilization of Long Chain Fatty Acids by Rat Liver: Studies of the Role of Fatty Acid Binding Protein DAVID A. BURNETT, ROBERT K. OCKNER
JOAN A. MANNING,
Department of Medicine and Liver Center, University of California, San Francisco San Francisco, California
To investigate the possible role of fatty acid binding protein (FABP) in the utilization of long chain fatty acids by rat liver we studied uptake and metabolic disposition of albumin-bound Y-oleate by isolated rat liver cell suspensions, both under control conditions and in the presence of flavaspidic acid, a known inhibitor of fatty acid binding to FABP. In incubations lasting up to 4 min, 0.02 mM flavaspidic acid effected a modest but consistent, significant and reversible inhibition of oleate oxidation (22%), esterification of phospholipids (17%) and triglycerides (15%), and uptake (I%%), but had no effect on 0, consumption, ATP concentration, or oxidation of acetate or octanoate. In a second series of experiments, the mechanism of these effects was investigated in cell-free systems by studying mitochondrial and microsomal enzymes involved in the initial stages of fatty acid utilization. In the presence of physiologic concentrations of partially purified FABP, the activities of these enzymes were significantly enhanced in response to small changes in FABP concentrations, and 0.02 mM flavaspidic acid was not inhibitory. Finally, FABP partially purified from rat liver cytosol was found to contain 38.4 nmol of endogenous long chain fatty acid/mg protein, a more than Xi-fold enrichment relative to the 205,000 g supernatant. Of these endogenous FABP-associated fatty Received September 18,1978. Accepted March 15.1979. Address requests for reprints to: Robert K. Ockner, M.D., Gastrointestinal Unit, University of California, San Francisco, San Francisco, California 94143. This study was supported in part by research grants AM-13328, P50 AM..18520, and GM-07546 from the National Institutes of Health. Dr. Burnett’s present address is: Liver Study Unit, Veterans Administration Hospital, 4191 Woolworth Avenue, Omaha, Nebraska 68105. 0 1979 by the American Gastroenterological Association 99x-5985/79/989241-99$92.96
acids, 26% were accounted for by 18:2 and 23% by 20:4. a substantial enrichment in these essential polyunsaturates compared with whole homogenate and serum. These experiments show that a low concentration of flavaspidic acid inhibits hepatocyte utilization of oleate. It is suggested that flavaspidic acid interferes with the entry of fatty acids into oxidation and esterification pathways by inhibiting formation of the fatty acid-FABP complex. Although its possible role as a fatty acid “carrier” remains uncertain, FABP influences the activity of enzyme reactions in which fatty acid and fatty acid acyl CoA are substrates, is associated with large quantities of endogenous unsaturated fatty acids, and may be important in the utilization of fatty acids formed within the cell, de novo or by ester hydrolysis. These and other studies support the concept that FABP plays a broad role in the cellular economy of long chain fatty acids.
Fatty acid binding protein (FABP)’ consists of one or more 12,000 mol wt acidic proteins present in the 105,000 g supernatant fraction of homogenate of liver, intestinal mucosa, myocardium, kidney, and adipose tissue.lw3 FABP, which is probably the same as the “Z” protein of Levi et al.,’ binds long chain fatty acids and acyl CoA thioesters with high affinity, and available evidence suggests that it may participate in the cellular utilization of long chain fatty scids_l-3.5-7 The precise nature of this participation remains unclear, however. In the rat small intestine, the mucosal distribution of FABP and its response to dietary fat intake are consistent with a relationship of the protein to fatty acid absorption.’ While intestinal FABP does not directly influence enterocyte uptake of luminal fatty acids, it appears to facilitate access of fatty acid to sites of acyl Coenzyme A (CoA) synthesis in the
BURNETT ET AL.
endoplasmic reticulum.7 Furthermore, although intestinal FABP has not been shown to exhibit enzyme activity, it markedly stimulates acyl CoA synthesis7 and diglyceride acyltransferase activity’ in intestinal microsomes. In liver, as in intestine, there has been no direct demonstration of an effect of FABP on cellular fatty acid uptake, but a relationship to fatty acid utilization was suggested by studies in whole animals and perfused livers.’ With regard to its effect on hepatic enzymes of glycerolipid synthesis, however, published data are somewhat in conflict. Thus, FABP has been shown to stimulate the activity of microsomal a-glycerophosphate acyltransferase” and diglyceride acyltransferaseP but did not appear to affect microsomal acyl CoA synthetase activity.” Recently published evidence indicates that as the major cytosolic acceptor for long chain acyl CoA thioesters, FABP modulates the activity of acetyl CoA carboxylase,” the rate-limiting enzyme in fatty acid synthesis. The present studies were designed to explore in greater detail the possible role of FABP in hepatic utilization of long chain fatty acids by means of three different experimental approaches. In the first, suspensions of isolated rat hepatocytes were studied with regard to cellular uptake and utilization of albumin-bound Y-oleate. In these experiments, flavaspidic acid-N-methyl glucaminate, a known inhibitor of fatty acid binding to FABP,‘,’ was used at low, nontoxic concentrations to observe the effect of decreased formation of the fatty acid-FABP complex on fatty acid utilization. In the second approach, the effects of FABP and of flavaspidic acid on hepatic microsomal and mitochondrial lipid synthesizing enzymes were examined in cell-free systems. Finally, partially purified hepatic FABP was analyzed for its content of endogenous long chain fatty acids and was compared with whole cytosol, whole homogenate, and serum. The results of these experiments provide support for the concept that FABP plays a broad role in the fatty acid economy of the liver cell. Results of the present studies have been reported in part in published abstracfs.‘2.‘3 Methods Materials 1-[‘%]Oleic acid, l-[‘%I sodium octanoate, l[Wlsodium acetate, [Wlglycerol-3-phosphate, and [3H]inulin were purchased from New England Nuclear (Boston, Mass.). Unlabeled oleic acid was purchased from Calbiochem (San Diego, Calif.). Octanoic acid, sodium acetate, Tris buffer, palmitoyl CoA, ATP, CoA, dithiothreito1 (DTT), Triton x-100, glycerol-3-phosphate, and fireflylantern luciferase were obtained from Sigma Chemical Co.
(St. Louis, MO.). Fatty acid free albumin, purchased from Sigma, was found to contain less than 0.02 nmol fatty acid per nmol albumin. Flavaspidic acid-N-methylglucaminate (hereafter flavaspidic acid) was generously provided by Dr. A. Aho (Turku, Finland).7 Male Sprague-Dawley rats, 250-300 g, maintained on standard laboratory chow (Feedstuffs Processing Co., San Francisco, Calif.) were used in all experiments.
of Hepatic FABP
Hepatic FABP was partially purified as the 12,000 mol wt fraction from the 105,000 g supernatant of liver homogenate by gel filtration on Sephadex G-50, as previously described.’
Studies Utilizing Isolated
Suspensions of isolated rat liver cells were prepared by Ontko’s modification of the method of Berry and Friend.14 Incubations, carried out in triplicate for each liver, contained 6-14 mg cell protein, and 30 mg albumin in a final volume of 2 ml. Reactions were started by adding 1 milliliter of fatty acid-albumin complex to 1 ml of cell suspension. More than 85% of cells excluded trypan blue at the end of the longest incubations. Hepatocyte oxygen consumption was measured over l3 min at 3i”C in a Gilson polarograph with a 1.5-ml chamber, utilizing a calibrated Clarke electrode. Hepatocyte ATP levels were measured by adding an aliquot of the incubated cell suspension directly to ethanol in a dry ice/acetone bath. The ATP was extracted into the ethanol, and its concentration was measured by the luciferase method as described by Holmsen et al.‘” Uptake of [‘%]oleate by hepatocytes was defined as cell-associated ‘% at the end of the incubation, and was determined by the following modification of the method of Spector et al.‘? A l-milliliter aliquot of the incubation mixture containing cells, suspension medium, albuminbound [‘%]oleate, and [“Hlinulin was rapidly diluted in 20 ml of ice-cold albumin-free suspension medium, immediately centrifuged at 1000 g for 30 set, and the supernatant was removed. Radioactivity in the cell pellet was measured, and [‘%]oleate in adherent medium was corrected for on the basis of the [3H]inulin present in the pellet and supernatant. By this method, uptake at zero time was not different from zero, and at later times approximated the sum of the 14C in measured products of oxidation and esterification (see Results). In studies of ‘%O, production, cells were incubated in flasks containing a center well. At the end of the incubation, “%O, was collected for 60 min as described previously.’ Incorporation of [‘%]oleate into lipids and water-soluble products was measured under conditions identical to those used in studies of uptake and ‘X0, formation. At the end of the incubation, lipids were extracted by the method of Folch et a1.17 Lipid classes were separated by thin-layer chromatography on 0.25 mm silica gel 60 (EM Laboratories, Elmsford, N.Y.) in a solvent system of petro-
FATTY ACID BINDING PROTEIN AND FATTY ACID UTILIZATION IN RAT LIVER
leum ether: diethyl ether:acetic acid: 90:15:1.5, and identified as described previously.” Appropriate zones were scraped directly into counting vials and assayed for radioactivity (,see below). After 4 min incubations, more than 95% of the radioactivity was in the fatty acid zone, indicating that substrate concentration had not changed appreciably during the experiment. For each group of experiments, “blank” samples were prepared by adding cell suspension and albumin-bound [14C]oleate directly to the Folch extraction mixture; radioactivity levels in the various lipid zones after thin-layer chromatography were taken as blank values for that experiment. Under these conditions there was significant radioactivity in both phospholipid and diglyceride zone. In the latter, because of the short incubation periods and the large amount of radioactivity in the fatty acid zone, blank radioactivity was high relative to that in the samples (see Results). Watersoluble radioactivity, consisting of ketone bodies, citric acid cycle intermediates, and acetyl CoA,” was measured directly in the upper (water-methanol) phase of the Folch extraction system.
Gas-Liquid Free Fatty
Samples of FABP fraction, partially purified from the 105,006 g supernatant of liver homogenate by Sephadex G-50 gel filtration (see above) or of whole 105,006 g supernatant, liver homogenate, or serum, were subjected to lipid extraction by the method of Folch et a1.,17and fatty acids were isolated by thin-layer chromatography. Fatty acid methylesters were prepared by the addition of diazomethane to the fatty acids eluted from the thin-layer chromatogram and were subjected to gas-liquid chromatography on a Hewlett-Packard 402B gas chromatograph employing a 6-foot glass column packed with 10% SP-2330 on lOO/120 mesh Chromosorb W AW (Supelco) at 166’18O’C. Pentadecanoic acid was employed as an internal standard and fatty acid mass was determined by means of a Hewlett-Packard model 3380 digital integrator.
7.4,2 mM EDTA, 50 mM MgCl,, 20 mM ATP, 0.6 mM CoA, 1.0 mM DTT, and 0.02 mM l-[‘4C]oleate. The reaction was started by the addition of 10 pg microsomal or mitochondrial protein, and activity was expressed as nmol [‘Xloleate incorporated into acyl CoA per mg microsomal or mitochondrial protein at 1 min. Acyl CoA:glycerol-3-phosphate acyltransferase (EC 18.104.22.168) was assayed by the method of Jamdar and Fallon,** as the incorporation of [‘4C]glycerol-3-phosphate into phosphatidate. The final assay volume of 0.7 ml contained 25 mM Tris, pH 7.5, 50 mM KCl, 0.42 mM [‘4C]glycerol-3-phosphate, 0.7 mM DTT, 0.02 mM palmitoy1 CoA and microsomal protein, 0.3 mg/ml. In this assay, albumin is normally added to prevent inhibition of the enzyme by the long chain acyl CoA substrate. In some of the present experiments, FABP fraction was employed instead. The reaction, linear to 5 min. was started by the addition of microsomes and was stopped at 2 min by adding 0.5 ml of the incubation mixture to 10 ml chloroform methano1.17 The product was identified as phosphatidate by thin-layer chromatography of the organic phase. Activity was expressed as nmol glycerol-3-phosphate incorporated into phosphatidate per mg microsomal protein per min.
Radioassays Samples were assayed for radioactivity in Liquifluor toluene solution (New England Nuclear) containing 10% Biosolv (Beckman Instruments, Inc., Fullerton, Calif.) in a Beckman liquid scintillation system model LS-250. For lipid-soluble extracts, Biosolv was not added. Quenching was corrected for by an automatic external standard.
Significance of differences among experimental groups was determined by the paired or unpaired t-tests.23
Results Preparation of Microsomal and Mitochondrial Fractions of Whole Rat Livers Rat liver microsomes were prepared by the method of Pande and Mead.*’ For mitochondria, livers were prepared in 2 volumes of 0.25 M sucrose in 10 mM phosphate, pH 7.4. The homogenate was diluted to 10: 1 with the same solution and centrifuged at 600 g for 10 min. The supernatant was then spun at 8500 g for 10 min, and the mitochondrial pellet was resuspended in 0.25 M sucrose. Microsomal Assays
Long chain acyl CoA synthetase (EC 22.214.171.124) activity was measured in microsomes and mitochondria by the method of Bar-Tana et al.‘l modified in that microsomes were diluted in 0.25 M sucrose instead of water. The final assay volume of 0.2 ml contained 150 mM Tris-HCl, pH
Effect of Flavaspicid Acid on Oxygen Consumption, ATP Concentration, and Oxidation of Short and Medium Chain Fatty Chain Fatty Acids in Isolated Hepatocytes Flavaspidic acid has been shown to uncouple oxidative phosphorylation in rat liver mitochondria,24 and to effect a concentration-dependent increase in oxygen consumption and decrease in ATP concentration in isolated liver cells.13 To avoid these effects, flavaspidic acid concentration in the present studies was limited to 0.02 mM and had no effect on hepatocyte oxygen consumption vs. controls (5.5 f 0.5 vs. 5.4 + 0.3 nmol O,/mg protein/min) or ATP concentration vs. controls (8.5 fi 0.8 vs. 8.2 + 0.8 nmol/mg protein). Similarly, the oxidation of [“Cloctanoate or [“Clacetate to “CO, was unaffected (not shown).
Vol. 77. No. 2
1. Utilization of l-[‘4C]oleate by isolated hepatocytes. Cells were incubated for up to 4 min in medium containing albumin-bound 1-[‘%]oleate, 0.02 mM. Uptake was measured as cell-associated 14C, corrected for adherent extracellular fluid, as described in the Methods section. Esterification is the sum of incorporation of 14C into phospholipids, diglycerides, and triglycerides. Oxidation is the sum of 14C in CO, and water-soluble metabolites. Mean * SE: n = 5 for uptake and the 4 min esterification points; n = 3 for all others. Each of these three or five observations was based on a separate cell preparation and represents the mean of duplicate or triplicate incubations.
Effect of Flavaspidic Acid on Oleate Utilization by Isolated Hepatocytes The utilization of albumin-bound 0.02 mM l[Wloleate by isolated rat hepatocytes under control conditions is shown in Figure 1. It can be seen that uptake (cell-associated ‘YZ) and oxidation (sum of “CO, and water-soluble products) are approximately linear over 4 min. Total esterification, i.e., the sum of incorporation into phospholipids, diglycerides, and triglycerides, is generally linear except at
Acid on the Oxidation by Isolated Hepatocytes
Control Flavaspidic acid
6.69 + 1.27 4.43 + 0.92”
Water-soluble products 39.4 f 5.4 31.4 f 3.9”
Cells were incubated in suspension medium and albumin-bound l-[“%I oleate, 0.02 mM, -C cquimolar flavaspidic acid. Radioactivity in 14C0, and water-soluble oxidation products was measured as described in the Methods section. Mean + SEM. n = 4 to
/. ” P < 0.01 vs. controls.
2. Flavaspidic acid inhibition of 1-[‘%]oleate utilization by isolated hepatocytes: effect of oleate concentration. Cells were incubated for 4 min in 0.02 mM flavaspidic acid as in Figure 1 at three different oleate concentrations. As described in the Methods section, oxidation was measured as the sum of “C0, and ‘%-water soluble products, and esterification as the sum of [%]oleate incorporation into phospholipid and trin = 3 for 0.1 mM and 0.44 mM oleate and n= glyceride. 5 to 7 for 0.02 mM oleate.
1 min. For this reason, subsequent experiments principally involved observations at 4 min. Under all conditions, incorporation of [Yloleate into cholesterol esters was negligible. Flavaspidic acid inhibited the incorporation of [Yloleate into “CO, and water-soluble oxidation products by 34% and 20% respectively, at 4 min (Table 1); similar inhibition was observed at 1 and 2 min. The 22% overall inhibition of oleate oxidation was substantially reversed at higher oleate concentrations (Figure 2). These inhibitory effects of low concentrations of flavaspidic acid stand in marked contrast to earlier studies which showed that at higher concentrations flavaspidic acid actually increased the conversion of l-[‘4C]oleate to “CO, because of an uncoupling of oxidative phosphorylation.‘z.‘3 The effect of flavaspidic acid on fatty acid esterification is shown in Table 2. Relative to controls, incorporation of 1-[‘%]oleate into phospholipids and triglycerides was decreased by 17% and 15%, respectively, at 4 min (P < 0.025). Although incorporation into diglyceride appeared unaffected, blank radioactivity in the diglyceride zone of the thin-layer chromatography system was relatively high, a result of its proximity to the fatty acid zone and the fact that these experiments were necessarily conducted over very short intervals, and may have obscured significant changes (see Methods). As with the effects of flavaspidic acid on uptake and oxidation, in-
FATTY ACID BINDING PROTEIN AND FATTY ACID UTILIZATION
IN RAT LIVER
Table 2. Effect of Flavaspidic Acid on the Esterification of I-P’CjOleate by Isolated Hepatocytes l-[‘4C]Oleate incorporation (pmol/mg prot/4 min)
Control Flavaspidic acid, 0.02 mh4
50.7 f 8.0
63.8 -t 8.2
80.3 f 7.8
41.9 f 6.3”
61.5 + 7.9
68.1 + 6.9O
Cells were incubated for were extracted, isolated sayed for radioactivity as k SEM. n = 5 or 6. ” P <
4 min as described in Figure 2. Lipids by thin-layer chromatography, and asdescribed in the Methods section. Mean 0.025 vs. controls.
hibition of esterification was reversed at higher oleate concentrations (Figure 2). The impaired oxidation and esterification of oleate were associated with similar inhibition of oleate uptake, which also was reversed at higher oleate concentration. Thus, while uptake of 0.02 mM oleate at 4 min was inhibited by 18% (P < O.Ol), that of 0.44 mM oleate was essentially unaffected (Figure 2). It should be noted that although the inhibitory effects of flavaspidic acid on cellular fatty acid utilization were small under these conditions, this was the inevitable result of the requirement that low concentrations’ of this agent (and, therefore, of oleate) be employed and that the duration of the incubations be limited.
ALBUMIN OR FABP FRACTION (mg/ml)
Figure 3. Effect of albumin and FABP fraction on microsomal acyl CoA synthetase activity. Rat liver microsomes were assayed for the incorporation of [W]oleate into acyl CoA as described in the Methods section, with 0.02 mM l-[‘4C]oleate, in the presence or absence of albumin or FABP fraction. Mean f SE: n = 3 for each point.
spuriously low because of dilution of the [“Cloleate substrate by unlabeled FABP-associated fatty acid. Figure 4 shows that FABP fraction and albumin also stimulate microsomal acyl CoA:glycerol-3-phosphate acyltransferase activity in a manner similar to their effects on acyl CoA synthetase. Thus, stimulation was noted at lower concentrations, but at higher protein concentrations, the effect is diminished and lost, presumably because of competition between protein and enzyme for the palmitoyl CoA substrate. In the assay for acyl CoA synthetase in isolated mitochondria (Figure 5), the presence of FABP frac-
Studies of Fatty Acid Activation and Transacylation in Subcellular Fractions It is possible that the above effects primarily reflected inhibition of microsomal and mitochondrial enzymes for acyl CoA and glycerolipid synthesis. To investigate this possibility, appropriate assay conditions were established to examine the effect of flavaspidic acid on the activity of certain of these enzymes. Although albumin is used in most standard assays, FABP is the principal cytosolic acceptor Iof fatty acid and acyl CoA, and, for this reason, the effects of FABP and albumin on enzyme activities were compared. Both FABP fraction and albumin significantly increased the activity of microsomal acyl CoA synthetase (Figure 3). Activity was maximal at 0.15 mg albumin per ml, and declined as the albumin concentration was increased. With FABP fraction, enzyme activity was similarly increased but persisted over a much broader concentration range. Moreover, unlike the albumin, which was fatty acid-free, FABP fraction contains large amounts of endogenous fatty acid (see below, and Reference 7) suggesting that enzyme activity measured in the presence of FABP is
FABP -5 ‘P
2.0 OR FABP FRACTION
Figure 4. Effect of albumin and FABP fraction on microsomal long chain acyl CoA:glycerol-3-phosphate acyltransferase activity. Rat liver microsomes were assayed for the incorporation of [Wlglycerol-3-phosphate into phosphatidate as described in the Methods section, with 0.42 mM l-[‘4C]glycerol-3-phosphate and 0.02 mM palmitoyl CoA, in the presence or absence of albumin or FABP fraction. Mean f SE; n = 3 for each point.
BURNETT ET AL.
ALBUMIN cl I
ALBUMIN OR FABP FRACTION (mg/ml)
of albumin and FABP fraction on mitochondrial acyl CoA synthetase activity. Rat liver mitochondria were assayed for the incorporation of [W]oleate into acyl CoA as described in the Methods section, with 0.02 mM l-[‘4C]oleate, in the presence or absence of albumin or FABP fraction. Mean f SE; n = 3 for each point. Effect
in the reaction produced an effect tion or albumin very similar to that seen with the microsomal enzyme (Figure 3), and perhaps even more pronounced in the case of FABP. It is noteworthy that in all of these assays, small changes in FABP concentrations, especially in the lower range, effected relatively large changes in enzyme activity, suggesting that FABP concentration itself may influence the activity of these reactions in the intact cell. In Table 3 are shown studies of the effect of 0.02 mM flavaspidic acid on the activity of microsomal and mitochondrial acyl CoA synthetase both in the
Long Chain AcyJ CoA Synthetase in Rat Liver Microsomes and Mitochondria: Effect of FJavaspidic Acid Additions None
Acyl CoA formed (nmol/mg prot/min) Microsomes Control Flavaspidic 0.02 mM Mitochondria Control Flavaspidic 0.02 mM
64.0 -+ 1.9
65.1 f. 2.2
122.7 f 6.8
30.5 + 1.1”
41.7 + 1.1”
118.3 + 5.5b
30.3 + 1.5
33.0 + 0.7
76.3 + 2.2
20.2 + 1.7”
22.5 f 1.6”
105.0 + 2.1”
Assays were performed as described in the Methods section using 0.02 mM l-[‘4C]oleate in the presence or absence of tlavaspidic acid and of albumin or FABP fraction (1.0 mg/ml). Mean f SE. n = 3 for all groups. a P < 0.01 vs. control. b Not significantly different from control. ’ 0.01
0.05 vs. control.
77, No. z
presence of added albumin or FABP and in their absence. Flavaspidic acid was used at 0.02 mM since this was its extracellular concentration in the hepatocyte studies, and was regarded as the maximum concentration to which the intracellular organelles were likely to have been exposed. Although 0.02 mM flavaspidic acid significantly inhibited both enzymes in the absence of soluble acceptors and in the presence of albumin, 1.0 mg per ml FABP fraction entirely prevented this inhibition. With the mitochondrial enzyme, flavaspidic acid actually increased its activity in the presence of FABP fraction. In the presence of albumin or FABP fraction, 0.02 mM flavaspidic acid did not inhibit microsomal acyl CoA: glycerol-3-phosphate acyltransferase (results not shown). Together, these studies suggest that in the intact cell, i.e., in the presence of soluble acceptor (FABP), 0.02 mM flavaspidic acid did not inhibit the activities of three key membrane-associated enzymes involved in the initial cellular utilization of long chain fatty acids.
Endogenous Long Chain Fatty Acid Associated with FABP Since FABP binds FFA and participates in their cellular utilization, it might be anticipated that endogenous FFA also would be associated with this protein. To investigate this, FABP fraction was analyzed for its content of long chain fatty acids (Methods), and was compared with 105,000 g supernatant, whole liver homogenate, and serum. It was found that FABP contained 38.4 + 0.9 nmol fatty acid per mg protein, a more than 25fold enrichment compared with whole homogenate (1.4 f 0.1). This represents an average fatty acid: protein mole ratio of 0.46 for the entire FABP fraction. Since it is likely that the fatty acids present in this fraction are relatively concentrated on the fatty acid binding protein(s) within it, the actual mole ratio is probably even greater. Of the endogenous fatty acids associated with the FABP fraction, two-thirds are unsaturated, and nearly 50% of the total is accounted for by the essential polyunsaturates 18: 2 and 20:4 (Figure 6). This unusual distribution is similar to that for the 105,000 g supernatant, consistent with the fact that the FFA present in the 105,000 g supernatant are accounted for chiefly by FABP, but is in marked contrast to that for the FFA associated with serum albumin and whole liver homogenate. This finding is consistent with earlier evidence that FABP preferentially binds unsaturated long chain fatty acids.’ Phospholipid was not detectable in this fraction.
FATTY ACID BINDING PROTEIN AND FATTY ACID UTILIZATION
Discussion These experiments show that flavaspidic acid, which competes with long chain fatty acids for binding to FABP in vitro, inhibited fatty acid utilization by liver cell suspensions. Since the flavaspidic acid concentration employed in these experiments was necessarily low, inhibition was accordingly modest. Nonetheless, the differences were significant and reproducible, and the demonstration of their reversibility together with normal cellular ATP concentration, oxygen consumption, and oxidation of ‘Goctanoate and Y-acetate to “CO, suggest that this effect did not represent nonspecific toxicity. Two theoretically possible explanations for the effect were not directly tested in these studies, since published evidence suggests that they are highly unlikely. The first is that there exists positive cooperativity between flavaspidic acid and oleate in binding to albumin, so that the dissociation of oleate and its entry into the hepatocyte is impaired. However, this is not consistent with the findings of Kamisaka et a1.*5who showed weak competition between flavaspidic acid and oleate, indicating that interaction between them at the level of albumin-binding would be expected to increase, not decrease, the availability of oleate to the cell. The second possibility, i.e., that the observed inhibition could have reflected a primary impairment of fatty acid uptake by the hepatocyte plasma membrane, is also unlikely since fatty acid uptake does not depend on specific surface membrane receptors for which flavaspidic acid might compete,” and Mishkin et al.” showed that binding of oleate to a liver plasma membrane preparation was not inhibited by flavaspidic acid. On the basis of these considerations, flavaspidic acid was regarded as acting at an intracellular location. Within the cell, inhibition of the mitochondrial and microsomal enzymes of fatty acid oxidation and esterification reactions could lead to fatty acid accumulation and, secondarily, to diminished net uptake. However, in the presence of FABP fraction, flavaspidic acid, at a concentration equal to that present in the extracellular medium in the isolated cell experiments, did not inhibit mitochondrial acyl CoA synthetase, or the microsomal enzymes acyl CoA synthetase and acyl CoA:glycerol-3-phosphate acyltransferase. Selective inhibition of later steps in the microsomal esterification pathways, such as phosphatidate phosphohydrolase, diglyceride acyltransferase, or choline phosphotransferase, was not excluded but seems unlikely since this might have been expected to result in accumulation of labeled fatty acid in phosphatide or diglyceride. The present findings, then, are interpreted as suggesting that flavaspidic acid partially blocked fatty
IN RAT LIVER
r 50 a k
8 10 SERUM
Figure 6. Fatty acid composition of rat serum, liver homogenate, 105,CUKlg supernatant, and FABP fraction. Samples were subjected to lipid extraction, thin-layer chromatography, and gas-liquid chromatography of fatty acid methyl esters as described in the Methods section. Mean f SE, expressed as percent of total long chain fatty acids in the designated material; n = 3 for a11 groups.
acid utilization at some site within the cell, prior to activation and esterification. In other words, flavaspidic acid interfered with the access of fatty acid to mitochondrial and microsomal activating and esterifying enzymes, similar to earlier studies of its effect in rat intestine’ and consistent with an interaction at the level of FABP. Since it is probable that fatty acid moves through cytosol by free diffusion either as fatty acid monomer or bound to FABP, it is appropriate to consider the determinants of flux rates for these two possibilities. For diffusion in a given medium over a distance, x, net flux, J, is proportional to D, the diffusion coefficient of the diffusing species, and dc/dx, its concentration gradient: J a D(dc/dx) Because D for monomeric fatty acid anions is approximately 4 x lo-’ cm2 per sec,27 i.e., about three times the value for 12,ooO mol wt globular proteins of 1.3-1.5 X10-’ cm’ per set,*” diffusion of fatty acid monomer would be the more efficient transfer mechanism if concentration gradients were identical. However, available evidence suggests that the concentration of FABP in cell water is approximately O.l- 0.2 mM,3*” i.e., at least l-2 orders of magnitude greater than the maximal concentration of long chain fatty acid monomer in aqueous media.30-33 Therefore, it is likely that the fatty acid-FABP complex diffuses along a steeper concentration gradient than does fatty acid monomer and therefore may constitute the more efficient mechanism for net fatty acid movement within the cell. On the basis of these considerations, it is possible
BURNETT ET AL.
that access of fatty acid to intracellular membranebound enzymes does not depend simply on free diffusion of fatty acid monomer through cytosol. Rather, it is proposed that FABP, as the principal cytosolic acceptor for fatty acids, in some manner facilitates this process. Apart from these considerations, there is increasing evidence that FABP is not merely a fatty acid “carrier.” Rather, both the present studies and earlier data from this and other laboratories suggest that this protein participates in the interaction of fatty acid and long chain acyl CoA substrates with a number of mitochondrial and microsomal enzymes. Thus, FABP increased the activity of both microsomal and mitochondrial acyl CoA synthetase, similar to earlier studies of intestinal microsomes.7 Although Suzue and Marcel” did not observe enhancement of hepatic microsomal acyl CoA synthetase by FABP, the presence of Triton X-100 in their assay system may have obscured an FABP effect. As noted, the endogenous fatty acids associated with the FABP fraction were mostly unsaturated, especially 18 : 2 and 20 : 4. This finding suggests the important possibility that a significant portion of FABP-associated fatty acids may be either derived from, or destined for, pathways in which unsaturated fatty acids predominate, such as the cleavage of phospholipid fatty acids by phospolipase A,, or the synthesis of phospholipids or prostaglandins. In other studies, it was shown that hepatic FABP served as the principal cytosolic acceptor for FFA liberated in vitro by the acyl CoA thioesterase reaction.” FABP also is the major cytosolic acceptor for long chain acyl COA,~.” and the present studies show that FABP enhances the activity of microsomal acyl CoA glycerol-3-phosphate acyltransferase. Mishkin and Turcotte made similar observations,’ and O’Doherty and Kuksis found that FABP (“Z”) stimulated the diglyceride acyltransferase reaction in liver and intestinal microsomes.’ These observations suggest that the FABP-acyl CoA complex may be a preferred substrate in these transacylation reactions. Furthermore, recent work from this laboratory indicates that FABP modulates the inhibitory effect of long chain fatty acids and acyl CoA on rat liver acetyl CoA carboxylase in vitro, and may thereby participate in the short-term control of fatty acid synthesis.” On the basis of the available evidence, then, some tentative conclusions may be drawn regarding the role of FABP in hepatic long chain fatty acid metabolism. First, a carrier function of the protein is neither established nor excluded.
GASTROENTEROLOGY Vol. 77, No. 2
Second, FABP serves in vitro as a “co-factor” for the reactions of several membrane-bound enzymes in which long chain fatty acid or acyl CoA is substrate (References 7-9 and present studies) and in addition is the principal cytosolic acceptor for FFA liberated by the acyl CoA thioesterase reaction.” The present studies do not provide evidence bearing on a possible effect of FABP on the compartmentalization of fatty acids between oxidation and esterification pathways, but this possibility is not excluded. Third, FABP modulates the inhibitory effect of long chain fatty acid and acyl CoA on acetyl CoA carboxylase activity and fatty acid synthesis and may modulate the effects of these substances on other enzymes as well.” Finally, because it is likely that FABP and “Z” are the same, it is appropriate to reassess the concept that this protein is related to the cytoplasmic transport of bilirubin and organic anions other than long chain fatty acids. In this connection, quantitative aspects of the metabolism of bilirubin (the principal physiologic organic anion studied) and of fatty acids are particularly significant. Thus, it has been estimated that in adult humans, net bilirubin production, which at steady-state approximates net bilirubin flux from plasma to liver to bile, is approximately 300 mg/24 hr, i.e., about 0.5 mmo1.34.35 In contrast, net release of endogenous long chain fatty acids from adipose tissue into plasma is about 200 g/24 hr, i.e., 700 mmol of which roughly onethird or 200 mmol enter the liver.3s On a molar basis, therefore, the ratio of endogenous fatty acid to bilirubin entering the liver from plasma is approximately 4OO:l. If that fraction of exogenous (alimentary) fatty acid which enters liver is also included, the ratio is even greater. Although this enormous excess of fatty acid does not entirely preclude participation by FABP in the hepatic transport of bilirubin or other organic anions, it suggests that such a role is likely to be secondary at best. Additional support for the primary relationship of FABP to fatty acid metabolism is provided by recent evidence that, in rats, treatment with clofibrate causes increased hepatic concentration of FABP,29*37.38 and hepatic utilization of plasma FFA3’ and that sex differences in the incorporation of albumin-bound oleate into triglyceride by isolated rat hepatocytes are associated with corresponding differences in FABP concentration3’ and by earlier evidence from a number of laboratories as summarized previously.’ In conclusion, available evidence suggests that FABP participates in several aspects of the cellular utilization, synthesis, and regulatory effects of long chain fatty acids and acyl CoA thioesters. Further studies will be required to define the nature and con-
trol of these diverse tissues.
FATTY ACID BINDING PROTEIN AND FATTY
in liver and other
Ocknec RK, Manning JA, Poppenhausen RB, et al: A binding protein for fatty acids in cytosol of intestinal mucosa, liver, myocardium, and other tissues. Science 177:56-58,197Z Mishkin S, Stein L, Gatmaitan Z, et al: The binding of fatty acids to cytoplasmic proteins: binding to Z protein in liver and other tissues of the rat. Biochem Biophys Res Commun 47:997-1883,197z Ockner RK, Manning JA: Fatty acid-binding protein in small intestine. Identification, isolation, and evidence for its role in cellular fatty acid transport. J Clin Invest 54:326-338,1974 Levi AJ, Gatmaitan Z. Arias IM: Two hepatic cytoplasmic protein fractions, Y and Z, and their possible role in the hepatic: uptake of bilirubin, sulfobromophthalein, and other anions. J Clin Invest 48:2156,1969 Mishkin S, Turcotte R: The binding of long-chain fatty acid CoA to Z, cytoplasmic protein present in liver and other tissues of the rat. Biochem Biophys Res Commun 57:918-926, 1974 Mishkin S, Stein L, Fleischner G, et al: Z protein in hepatic uptake and esterification of long-chain fatty acids. Am J Physiol228:1634-16461975 Ockner RK, Manning JA: Fatty acid binding protein. Role in esterification of absorbed long chain fatty acid in rat intestine. J Clin Invest 53:632-641,1976 O’Doherty PJA, Kuksis A: Stimulation of triacylglycerol synthesis by a protein in rat liver and intestinal mucosa. FEBS Lett 68256-258.1975 Mishkin S, Turcotte R: Stimulation of monoacylglycerophosphate formation by Z protein. Biochem Biophys Res Commun 60:376-381,1974 Suzue G, Marcel YL: Studies on the fatty acid binding proteins in cytosol of rat liver. Can J Biochem 53:884-869.1975 Lunzer MA, Manning JA. Ockner RK: Inhibition of rat liver acetyl coenzyme A carboxylase by long chain acyl coenzyme A and fatty acid. Modulation by fatty acid binding protein. J Biol Chem 252:5483-5487,1977 Burnett DA, Ockner RK: Studies of the role of fatty acid binding protein (FABP) in the utilization of free fatty acid (FFA) by isolated hepatocytes (abstr). Clin Res 24:432A, 1976 Burnett DA, Ockner RK: Flavaspidic acid (FLAV): effects on oxidative phosphorylation and fatty acid (FA) utilization in isolated hepatocytes (abstr). Clin Res 25:467A, 1977 Berry MN, Friend DS: High-yield preparation of isolated rat liver parenchymal cells. A biochemical and fine structural study. J Cell Biol43:566-520.1969 Holmsen H, Holmsen I, Bernhardsen A: Microdetermination of adenosine diphosphate and adenosine triphosphate in plasma with the firefly luciferase system. Anal Biochem 17:456-473.1966 Spector AA, Steinberg D, Tanaka A: Uptake of free fatty acids by Ehrlich ascites tumor cells. J Biol Chem 240:10321041,1965 Folch J, Lees M, Sloane Stanley GM: A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 226497-509.1957 Ockner RK, Pittman JP, Yager JL: Differences in the intestinal absorption of saturated and unsaturated long-chain fatty acicls. Gastroenterology 82981992.1972
ACID UTILIZATIONIN RAT LIVER
Sauer F, Mahadevan S, Erfle JD: The accumulation of citrate cycle intermediates
in rat liver cells oxidizing palmitate.
chim Biophys Acta 239:26-32,197l 20. Pande SV, Mead JF: Long chain fatty acid activation in subcellular preparations from rat liver. J Biol Chem 243:352-361, 1968 21. Bar-Tana J, Rose G, Shapiro B: The purification and properties of microsomal palmitoyl-CoA synthetase. Biochem J 122:353-362,197l 22. Jamdar SC, Fallon HJ: Glycerolipid biosynthesis in rat adipose tissue. I. Properties and distribution of glycerophosphate acyltransferase and effect of divalent cations on neutral lipid formation. J Lipid Res 14:509-516,1973 23. Snedecor GW, Cochran WG: Statistical Methods. Sixth edition. Ames, Iowa, Iowa State University Press. 1967. p 593 24. Runeberg L: Uncoupling of oxidative phosphorylation in rat liver mitochondria with desaspidin and related phlorobutyrophenone derivatives. Biochem Pharmacol11:237-242.1962 25. Kamisaka K, Litowsky I, Gatmaitan Z, et al: Circular dichroism analysis of the secondary structure of Z protein and its complexes with bilirubin and other organic anions. Biochim Biophys Acta 393:24-30,1975 26. Kuhl WE, Spector AA: Uptake of long-chain fatty acid methyl ester by mammalian cells. J Lipid Res 11:458-465,1978 27. Patil GS, Matthews RH, Cornwell DG: Kinetics of the processes of desorption from fatty acid monolayers. J Lipid Res 14:26-31,1973 28. Sober HA: Handbook of Biochemistry. Cleveland, Chemical Rubber Co, 1970, p C-10 29. Fleischner G, Meijer DKJ, Levine WG, et al: Effect of hypolipidemic drugs, nafenopin and clofibrate on the concentration of ligandin and Z protein in rat liver. Biochem Biophys Res Commun 67:1041-1947,1975 30. Eagland D, Franks F: Association equilibria in dilute aqueous solution of carboxylic acid soaps. Trans Faraday Sot 61:24682477.1965 31. Mukerjee P: The nature of the association equilibria and hydrophobic bonding in aqueous solutions of association colloids. Adv Colloid Interface Sci 1:241-275.1967 of long chain-alkyl car32. Smith R, Tanford C: Hydrophobicity boxylic acids, as measured by their distribution between heptane and aqueous solutions. Proc Nat1 Acad Sci 70:289-293, 1973 33. Sallee VL: Apparent monomer activity of saturated fatty acids in micellar bile salt solutions measured by a polyethylene partitioning system. J Lipid Res 15:56-64.1974 in man. N Engl J Med 34. Schmid R: Bilirubin metabolism 287:703-769.1972 35. Berk P, Martin JF, Blaschke TF. et al: Unconjugated hyperbilirubinemia. Physiologic evaluation and experimental approaches to therapy. Ann Intern Med 82:552-570,1975 and management of 36. Have1 RJ: Pathogenesis, differentiation hypertriglyceridemia. Adv Intern Med 15:117-154,1969 37. Foliot A, Housset E, Fiessinger J-N, et al: Influence of clofibrate (ethylchorophenoxyisobutyrate) on hepatic Z protein in rats (abstr). Digestion 12:256,1975 38. Renaud G, Foliot A, Infante R: Increased uptake of fatty acid by the isolated rat liver after raising the fatty acid binding protein concentration with clofibrate. Biochem Biophys Res Commun 86:327-334,1978 39. Ockner RK, Burnett DA. Lysenko N, et al: Sex differences in free fatty acid utilization and triglyceride biosynthesis in hepatocytes (abstr). Clin Res 26:531A, 1978