Bioluminescent determination of free fatty acids

ANALYTICAL

BIOCHEMISTRY

140, 349-353 (1984)

Bioluminescent HORST

Determination

of Free Fatty Acids

RATHER AND EBERHARD WIELAND

Clinical Institute for Myocardial Infarction Research of the Medical Clinic of the University of Heidelberg, BergheimerstraJe 58, D-6900 Heidelberg, West Germany Received November 28, 1983 A simple, highly specific, and sensitive bioluminescent method for determination of free fatty acids in unextracted plasma or serum has been developed. The method is based on the activation of free fatty acids by acyl-CoA synthetase (EC 6.2.1.3). The pyrophosphate formed is used to phosphorylate fructose 6-phosphate in a reaction catalyzed by the enzyme pyrophosphate-f6-phosphate phosphotransferase (EC 4.1.2.13). The triosephosphates produced from fructose 1,6-bisphosphate by aldolase are oxidized by NAD in the presence of arsenate to 3-phosphdycerate. The NADH is detected via the bacterial NADH-linked luciferase system. Excellent agreement has been obtained by comparison with accepted methods. In addition, for the determination of serum free fatty acids, the method is particularly applicable for following lipolysis of isolated adipocytes. KEY WORDS: free fatty acids; acyl-CoA synthetase; NADH-linked luciferase system; serum; isolated rat adipocytes.

The two most commonly used methods for free fatty acid determination in serum are not enzymatic, and have disadvantages because of their lack of specifity, their limited sensitivity, and the complexity of the procedure (1,2,3). They both require extraction of free fatty acids into an organic solvent. Recently, four enzymatic methods for the

determination of free fatty acids have been described (4-7). They use the bacterial acylCoA synthetase (EC 6.2.1.3) and are performed on unextracted plasma. We describe a simple sensitive bioluminescent method for the assay of free fatty acids in serum or plasma and incubation media with the use of the bacterial NADH-linked luciferase system.

Principle ACS

I a) FFA’ + CoA + ATP

*

Acyl-CoA + AMP + PPi

PPi-PFK

b) PPi + F-6-P

e

F-I-6-P2 + Pi

Aldolase

c) F- 1-6-P,

=

DHAP+GAP

TIM

d) DHAP e) GAP + NAD+ + AsO:II a) NADH b) FMNH*

+ H+ + FMN + R-CHO

5

Gz

GAP 3-PG + NADH

s;

FMNH2

+ O2 Lucife”

’ Abbreviations used: FFA, free fatty acids; ACS, acylCOA synthetase (EC 6.2.1.3); PPi-PF?& pyrophosphatefructose-6-phosphate 1-phosphotransferase; F-6-P, fructose

+ H+ + HAsO:-

+ NAD+

R-COOH+FMN+~*V

6-phosphate; F- 1-6-P2, fructose 1,6-bisphosphate; DHAP, dihydroxyacetonphosphate; GAP, glyceraldehyde 3-phosphate; Aldolase, Fructose-bisphosphate aldolase (EC 349

0003-2697184 $3.00 Copyright

8 I984 by Academic Pns, Inc. All rights of reproduction in any form reserved.

350

KATHER

AND WIELAND

As originally described by Shimizu et al. fatty acids are activated in the presence of ATP and CoA by acyl-CoA synthetase (4). The pyrophosphate formed is used to phosphorylate fructose 6-phosphate in a reaction catalyzed by the enzyme pyrophosphate-fructose-6phosphate 1-phosphotransferase (89). The triosephosphates produced from fructose 1,6bisphosphate by aldolase are oxidized by NAD in the presence of arsenate to ?)-phosphoglycerate as previously published ( 10). The NADH formed is monitored by using bacterial NADH-like luciferase.

Fat cell isolation. Epididymal rat fat cells were isolated by collagenase digestion (2 mg/ml, 45 min, 37°C) (1 l), washed, and suspended in Krebs-Henseleit bicarbonate buffer, pH 7.4, containing 4% (w/v) bovine serum albumin, essentially free of fatty acids. Aliquots of the cell suspension (0.05 ml, 2001000 cells) were incubated for 180 min at 37 “C in the presence of 5 mmol/liter glucose and isoproterenol ( 1O-l’- 1O-l4 mol/liter). Incubations were stopped by heating (95°C 5 min.). The deproteinized media were diluted 2- to lo-fold with deionized water. Glycerol release was estimated as reported previously MATERIALS AND METHODS (10,ll). Conversion of free fatty acids. A 50-~1 aliChemicals. ATP, NAD, triosephosphatequot of the diluted sample was added to an isomerase, and enzymes for luminescence equal volume of medium composed of 23 were from Boehringer-Mannheim, Mannmmol/liter Hepes, pH 8.0, 1.1 mmol/liter heim, FRG. Acyl-CoA synthetase, all other KCl, 20 mmol/liter Na3As04, 1.1 mmol/liter enzymes for conversion of free fatty acids, and dithiothreitol (DTI’), 2.9 mmol/liter MgCl, a6 bovine serum albumin, essentially fatty acid H20, 0.4 mmol/liter ATP, 8 mmolfiter NAD, free, were from Sigma Chemical Company, 0.065 mmol/liter coenzyme A (CoA), 0.53 St. Louis, Missouri. Enzymes were used withpmol/liter EDTA, 33 U/ml triosephosphate out further purification. Tetradecanal and rafisomerase (TIM), 14 U/ml glyceraldehyde-3finose were from EGA Chemie, Steinheim, phosphate dehydrogenase (GAPDH), 0.28 FRG, and Merck AG, Darmstadt, FRG, reU/ml aldolase, 0.10 U/ml PPi-phosphofrucspectively. Coenzyme A, capric acid, lauric tokinase (PPi-PFK), 0.04 mmol/liter fructose acid, palmitic acid, oleic acid, linoleic acid, 6-phosphate (F-6-P), and 6.7 mu/ml acyl-CoA linolenic acid, and buffer reagents were from synthetase (ACS). After incubation for 180 Serva GmbH, Heidelberg, FRG. All other remin at 25°C the samples were further diluted agents were of analytical grade. (10 times), and 0.03- to 0.05-ml aliquots of Pretreatment of serum and plasma. A lo- the diluted medium were assayed for NADH ~1 aliquot of native serum or plasma was added content. to 10 ~1 of 1 N HCl to inactivate possibly Bioluminescent assay. Luciferase ( 1.5 mU/ interfering dehydrogenases. After 1 min the ml) and NAD(P)H:FMN oxidoreductase (1.7 sample was neutralized by adding 10 ~1 of 1 U/ml) were dissolved in a potassium phosN NaOH, and diluted 100 times with deionized phate buffer (0.2 mol/liter) containing 0.4 water containing 0.25% Triton X-100. mmol/liter DTT and 67 mmol/liter raffinose. Tetradecanal (4.7 mmol/liter) was dissolved in a solution containing 50 g/liter of bovine 4.1.2.13); TIM, triosephosphateisomerase, D-glyceraldeserum albumin, essentially fatty acid free, and hyde-3-phosphate ketol-isomerase (EC 5.3.1.1); GAPDH, glyceraldehyde-3-phosphate dehydrogenase, o-glyceral10 g/liter Triton X-100, pH 7.0, at 50°C. Sodehyde-3-phosphate:NAD oxidoreductase (phosphorylatlutions of tetradecanal and of luciferase were ing) (EC 1.2.1.12); NAD(P)H:FMN oxred., NAD divided into small portions and stored fro(P)H:FMN oxidoreductase (EC 1.6X.1); lucifemse, a&anal, zen at -20°C. Flavinmononucleotide (0.11 reduced-FMN-oxygen oxidoredwtase ( 1-hydroxylating, mmol/liter) was dissolved in a potassium luminescing) (EC 1.14.14.3); Hepes, 4-(2-hydroxyethyl)I-pipemzineethanesulfonic acid; DTT, ditbiothreitol. phosphate buffer (2 mmol/liter, pH 7.0) and

BIOLUMINESCENT

DETERMINATION

OF FREE FATTY

351

ACIDS

kept in a dark bottle on ice. The solution was made up fresh daily. The assay cocktail ( 100 ml) contained 0.5 mmol/liter tetradecanal, 1.1 pmol/liter FMN, 5.3 mu/ml luciferase, and 83 U/ml NAD(P)H:FMN oxidoreductase. Portions (0.5 ml) of the assay cocktail were brought to 25 “C by a 20-min preincubation period. Bioluminescence assays were started by addition of 0.03-0.05 ml of sample. Production of light was measured by a Berthold Biomulat 9500 T with the photometer settings at rate. RESULTS

Conversion of free fatty acids were time and concentration dependent (Fig. 1). At a concentration of 10 pmol/liter, palmitic acid NADH formation was completed within 40 min compared to 90 min required for conversion of 50 pmol/liter. In serum, the reaction lasted approximately 120 min. A similar time course was obtained using a mixture of capric acid, lauric acid, palmitic acid, oleic acid, linoleic acid, and linolenic acid (5 pmol/liter of each fatty acid). As shown in Fig. 1, light emission remained constant over 180 min. To

Time in min.

FIG. 1. Time course of FFA conversion. 0, 50 pmol/ liter palmitic acid; H, 10 pmol/liter palmitic acid; A, blank (no WA). Values are means of duplicate determinations. In order to be in the linear range, samples were diluted 15-fold instead of lo-fold (see Materials and Methods) prior to determination.

1i

t’ I

o0

5

10

15

20 25 FFA(wrd/l)

30

FIG. 2. Standard curve for palmitic acid. Values are means of duplicate determinations. Experimental conditions were as described under Materials and Methods.

assure completion of free fatty acid reaction, an incubation period of 3 h has been chosen. Figure 2 shows a standard curve for palmitic acid. Light production is linearly related to fatty acid concentrations up to 30 pmol/liter. Based on a signal-to-noise ratio of 1, the detection limit was about 5 pmol/liter. The recovery of two concentrations ( 10 and 5 gmol/liter) of sodium palmitate added to an appropriately diluted serum sample was 95%. The recoveries for capric acid, lauric acid, palmitic acid, oleic acid, linoleic acid, and linolenic acid were in a similar range (96- 104%). At a serum free fatty acid concentration of 1752.8 hmol/liter, within-day precision averaged 4% (SD, 70.53; n = 18). Comparison with two commercially available kit methods for determination of free fatty acids in serum (Boehringer-Mannheim, nonesterified fatty acids, calorimetric method, No. 12055; and Wako NEFA C, ACSACOD method No. 27375049) is shown in Fig. 3. The linear regression data show good correlations for the chemical calorimetric method (Boehringer kit) and ex-

352

KATHER

AND WIELAND

DISCUSSION

~~~~~~~~

kk

L I I 8 0 500 loo0 1500 FFA ( pmol/l 1 by the wsent method

; k

0 500 loo0 15w 2owl FFA ( pmolll I by the present method

FIG. 3. Comparison of serum FFA concentrations measured by the present method and by the chemical colorimetric method (A; r = 0.96, Y intercept = -4.103, Slope = 0.92, n = 12) or the enzymatic calorimetric method (B; r = 0.98, Y intercept = -4.67, slope = 1.082, n = 20). Values are means of duplicate determinations. -, Regression line.

cellent agreement with the enzymatic colorimetric procedure (Wako kit) (r = 0.96, Y intercept = -4.103, slope = 0.92; and r = 0.99, Y intercept = -4.75, slope = 1.082, respectively). It should be mentioned that untreated serum was used for the enzymatic calorimetric method whereas the samples analyzed by the bioluminescent method had been subjected to acid denaturation of serum enzymes. The excellent agreement between both methods indicates, therefore, that acid treatment caused negligible hydrolysis of serum triglycerides. The method has also been used to follow lipolysis of rat cells by measuring the release of free fatty acids and glycerol. Figure 4 shows the effect of increasing concentrations of isoproterenol on the release of free fatty acids and of glycerol in isolated rat fat cells. The &adrenergic amine induced a dose-dependent increase of free fatty acids and glycerol release. The qualitative pattern of isoproterenol action was almost identical for both parameters of lipid mobilization. Half-maximal activation occurred at approximately 10 nmol/liter of isoproterenol, maximal release of free fatty acids and glycerol was observed at an isoproterenol concentration of 10 pmol/liter. The ratio of free fatty acids to glycerol was approximately 3, indicating that reesterification was negligible under the conditions employed.

The bacterial luciferase/FMN-reductase system has been used for the analysis of various compounds. However, many of the reported applications have not been verified on actual samples or have not been compared to accepted methods. We present a method for fatty acid determination which combines the fatty acid specificity of ‘he acyl-CoA synthetase reaction with the excellent sensitivity of the bacterial NADH-linked luciferase system. Because acylCoA synthetase is specific for fatty acids (47), the method can be applied to unextracted plasma samples or incubation medium. Due to the use of bacterial luciferase, our approach is at least 10 times more sensitive than recently published enzymatic calorimetric or fluorometric methods. Even if small samples (10 ~1) are employed, serum fatty acid concentration markedly exceeded the upper detection limit of the method. In order to obtain correct estimates of free fatty acid levels, the sera have to be diluted 50- to loo-fold prior to fatty acid determination. The high sensitivity of the bioluminescent fatty acid determination offers particular advantages for metabolic studies with cells and tissues. As shown in this report,

-

13-

AFreeFattyAcds - i2-@ll- ilo-B -2 -aa-

9-

FIG. 4. Effect of increasing isoproterenol concentrations on the FFA (A) and glycerol (B) release of isolated rat adipocytes. Values are means of duplicate determinations. Experimental conditions were as described under Mate&Is and Methods.

BIOLUMINESCENT

DETERMINATION

lipolysis of isolated fat cells can easily be followed by means of this technique. Theoretically only a few cells would be necessary. Due to relatively high blank values, approximately 500 cells per sample (50 jd) are presently needed as a minimum. Nevertheless, the method presented here is sensitive enough to permit systematic studies on fat cell lipolysis with milligram amounts of human adipose tissue, which can easily be obtained by needle biopsies. Cells isolated from 100-300 mg of tissue are sufficient for more than 300 determinations. In summary, we describe a highly sensitive bioluminescent method for free fatty acids which has been validated by comparison with accepted methods. In addition to the determination of serum free fatty acids and the estimation of lipolysis in isolated fat cells, this method should be applicable to the assay of lipase and lipoprotein lipase in biological fluids. Besides its simplicity, the method offers

OF FREE FATTY

ACIDS

353

the advantage of being performed with commercially available enzymes and chemicals. REFERENCES 1. Dole, V. P. (1956) J. Clin. Invest. 35, 150-154. 2. Duncombe, W. G. (1963) Biochem. J. 88,7-10. 3. Duncombe, W. G. (1964) Clin. Chim. Acta 9, 122125. 4. Shim& S., Inoue, K., Tani, Y., and Yamada, H. (1979) Anal. B&hem. 98, 341-345. 5. Shim& S., Tani, Y., Yamada, H., Tabata, M., and Murachi, T. (1980) Anal. Biochem. 107, 193-198. 6. Okabe, H., Uji, Y., Nagashima, K., and Noma, M. (1980) Clin. Chem. 26, 1540-1543. 7. Miles, J., Glasscock, R., Aikens, J., Gerich, J., and Haymond, M. (1983) J. Lipid Res. 24, 96-99. 8. O’Brien, W. E. (1976) Anal. B&hem. 76,423-430. 9. Cook, G. A., O’Brien, W. E., Wood, H. G., King, M. T., and Veech, R. L. (1978) Anal. Biochem. 91,557-565.

10. Kather, H., Schrader, F., and Simon, B. (1982) Clin. Chim. Acta 120, 295-300. 11. Kather, H., Schriider, F., and Simon, B. (1982) in Luminescent Assays:Perspectivesin Endocrinology and Clinical Chemistry (Serio, M., and Pazzagali, M., eds.), pp. 53-56, Raven Press, New York.