Enzyme reactor for urinary acylcarnitines assay by reversed-phase high-performance liquid chromatography

Clinica Chimica Acta. 216 (1993} 135-143 © 1993 Elsevier Science Publishers B.V. All riots reserved. 0009-8981/93/$06.00

135

CCA 05540

Enzyme reactor for urinary acylcarnitines assay by reversed-phase high-performance liquid chromatography Kojiro M a t s u m o t o a, M a m o r u Takahashi b, N o b u a k i T a k i y a m a c, Hideo Misaki b, N o b u t a k e M a t s u o c, Sawao M u r a n o d a n d H i d e t a k a Yuki a aClinical Chemistry. School of Piu~,maceutical Sciences. Toho University. Miyama 2-2-1. Funahashi. Chiba 274. bResearch Laboratories. Diagnostic Division. Asahi Chemical Indutry Co. Ltd.. Mifuku 632-1. Ohito-cho. Tagatagun. Shizuoka 410-23. CPediatrics. School of Medicine. Keio University. Shinanomachi 35. Shinjuku-ku. Tokyo 160 and dDeportment of Applied Microbiological Technology. The Kumamoto Institute of Technology. lkeda 4-22-1. Kumamoto 860 (Japan) (Received 19 October 1992; revision received 16 February 1993, ac,".epted 20 February 1993) Key words: Acylcarnitines: Enzyme reactor: HPLC: Acylcarnitine hydrolase

Summary An immobilized enzyme reactor, made up acylcarnitine hydrolase, carnitine dehydrogenase and diaphorase in sequence, was developed for the sensitive and selective determination of urinary free and individual acylcarnitines by a reversedphase high-performance liquid chromatography. A 100-~l urine sample was directly injected onto the TSKgel ODS 80Ts column and eluted by a step-gradient procedure. The eluent was mixed with the substrate solution of O-NAD + (1.0 mmol/i), resazurin (25 omol/I) and Tris acetate (0.2 mol/I, pH 9.0). The mixture was passed through the immobilized enzyme reactor at 40°C. Acylcarnitines were hyd,:olyzed and then converted to rezorufin which was measured by monitoring the fluorescence intensity at )~EX= 560 nm and )~EM= 580 nm. Free, acetyl-, glutaryl-, propionyl-, butyryl-, isobutyryl-, valeryl- and isovalerylcarnitine were determined within 55 min with detection limits (< 1 ~mol/i) and within-run and day-to-day imprecision (C.V. < 6%). Free, acetyl- and isobutyrylcarnitine were found in normal urine. On the other hand, propionylcarnitine was detected in the urine of children with propionic aciduria and methylmalonic aciduria and multiple acylcarnitines were found in the urine of children with glutaric aciduria (type li). Correspondence to: Kojiro Matsumoto, Clinical Chemistry, School of Pharmaceutical Sciences, Toho University, Miyama 2-2-1, Funabashi, Chiba 274, Japan.

136

Introduction The identification and determination of acylcarnitines in urine is useful to the diagnosis of secondary carnitine deficiency caused by various inborn errors and acquired disorders of fatty acid and amino acid oxidation [1-3]. The excretion of total, free and acylcarnitines has been measured by radiometric [4] and spectrophotometric [5-8] procedures. A flow injection method that uses immobilized carnitine dehydrogenase (carnitine/NAD + oxidoreductase, EC I. 1. I. 108) and diaphorase (dihydrolipoamide/ NAD* oxidoreductase, EC 1.8.l.4) has already been reported [9,10]. A new enzyme, acylcarnitine hydrolase, has now been prepared and it has been applied in an immobilized reactor to the separative determination of acylcarnitines. This method can directly analyze acylcarnitines in human urine samples with high sensitivity and selectivity.

Experimental

Principle and flow diagram The flow diagram for the present assay system is shown in Fig. 1 and the reaction principle of the immobilized enzyme reactor is as follows: Acylcarnitine hydrolas¢ Acyl-L-carnitine + H20 ~ L-carnitine + RCOOH Carnitine dehydrogenase L-Carnitine +/3-NAD + = dehydrocarnitine +/~-NADH + H +

NADH + resazurin

Diaphorase ~ NAD + + rezorufin

Acyl-L-carnitines separated by the reversed-phase column are hydrolyzed to Lcarnitine in the first acylcarnitine hydrolase-immobilized column, liberated Lcarnitine converts/~-NAD* to/~-NADH in the second carnitine dehydrogenaseimmobilized column and then/~-NADH reduces resazurin to rezorufin in the third

Column ~

Detector~ Recorder

I omoI I A B C

Eluent

Drain

Substrate

Fig. I. Flow diagram for the determination of acylcarnitines with the immobilized enzyme reactor.

137

diaphorase-immobilized column. The fluorescence intensity 0~EX= 560 nm, )~EM= 580 nm) of rezorufin produced by these enzyme reactions is proportional to the concentrations of the acylcarnitines in the sample.

Enzymes Acylcarnitine hydrolase (Alcaligenes sp., 30 units/mg protein), carnitine dehydrogenase (AIcaligenes sp., 210 units/mg protein) and diaphorase (Bacillus megaterium, 53 units/mg protein) were from Asahi Chemical lndu,.;try Co. (Tokyo, Japan). Newly prepared acylcarnitine hydrolase has a molecular weight of 67 kDa, an isoelectric point of 5.1 and a pH optimum between 6.5 and 9.5. The Michaelis constants for acyl-L-carnitines in Tris-HCI buffer (0. I mold, pH 9.0) are as follows: acetyl-, 40 ~mol/l; propioayl-, 30 ~mol/l; hexanoyl-, octanoyl-, decanoyl-, lauroyland myristoyl-, 20 ~mol/l; palmitoyl-, 30 ~mol/l; stearoyl-, 50 ~mol/i. Other detailed enzymatic properties will be published elsewhere (unpublished). Reagents DL-Carnitine HCI, acetyl-DL-carnitine HCI, octanoyl-DL-carnitine HCi and resazurin (Sigma Chemical Co., St Louis, MO), /3-NAD+ sodium salt (Oriental Yeast Co., Osaka, Japan), propionyi chloride, glutaric anhydride, isobutyryl chloride, butyryl chloride, valeryl chloride and isovaleryl chloride (Tokyo Kasei Kogyo Co., Tokyo, Japan) were used. Other chemicals were of analytical grade. Urine samples Urine samples were obtained from healthy adults (4 males and 9 females, 18-73 years) and children (4 males and 6 females, 0-15 years). Urine samples of patients with propionic aciduria, methylmalonic aciduria and glutaric aciduria (type ll) were collected before and after carnitine administration. All cases were diagnosed by gas chromatography-mass spectrometric (GC-MS) analysis of urinary organic acids. The patient with propion!c aciduria (2 years, female) was rehospitalized with convulsions and hyperammonemia (284 ~mol/I) and treated with L-carnitine. The patient with methylmalonic aciduria (10 days, female) had vomiting and hyperammonemia (786 ~mol/I) 5 days after birth, then medication of vitamin Bl~ and Lcarnitine was started. The patient with glutaric aciduria (type II, 5 years, female) was hospitalized with convulsions and dim consciousness. Because of subsequent hypertransaminasemia (sAST: 192 units/I, sALT: 631 units/I), hyperammonemia (350 ~mol/I) and hypoglycemia (1.4 mmol/1) Reye's-like syndrome was suspected and then L-carnitine and riboflavin were administered. Her brother (2 years) also died suddenly with diarrhoea and convulsions and suspected Reye's-like syndrome. These samples were stored at -20°C. Before assay, 0.1 mol/i of phosphoric acid was added to the urine sample (1:9, by vol.) to adjust the pH to 4. Urine samples with a high concentration of carnitines can be diluted with solvent A of the HPLC system. Immobilization of enzymes Acylcarnitine hydrolase (10 mg), carnitine dehydrogenase (15 mg) and diaphorase (20 mg) were separatedly immobilized on i g of an AF FormyI-Toyopearl 650 gel

138

(Tosoh Co., Tokyo, Japan), as previously described [9]. The gel was washed with water and sodium phosphate buffer (0.1 mol/l, pH 8.0) containing 0.5 mol/I NaCI. The enzyme in the phosphate buffer (5 m~) was added to ~he gel and the suspension was gently stirred overnight at room temperature in the presence of NaCNBH3 (30 mg). After washing the gel with water and 0.5 mol/l NaCi, the gel was blocked by ethanolamine HCi buffer (I.0 tool/i, pH 8.0) containing NaCNBH3 (3.5 mg) for 2 h at room temperature. The yields of immobilization calculated by absorbance at 280 nm were 45%, 53% and 50% for acylcarnitine hydrolase, carnitine dehydrogenase and diaphorase, respectively. The gels immobilized with the enzymes were packed into separate stainless-steel tubes (4.0 i.d. × 10 mm).

Synthesis of acylcarnitines Propionyl-, glutaryi-, butyryl-, isobutyryl-, valeryl- and isovaleryl-DL-carnitine HC! were synthesized according to the method of Bohmer and Bremer [11]. DLCarnitine HCI (300 mg) was dissolved in trifluoroacetic acid (5 ml), mixed with the correspending acid chlorides and left at room temperature overnight. After add'~ng ml of acetone and cooling on ice for 2 h, the precipitates were removed by centrifugation. Acylcarnitines were crystallized by adding ethyl ether. The products were recrystallized in ethanol-acetone-ethyl ether. In the case of glutaryi-DLcarnitine, DL-carnitine HC! (300 mg) was dissolved in pyridine (0.5 mi) and mixed with glutaric anhydride (I g). HPLC system The HPLC system was compo~d of a Tosoh CP-8000 series (Tosoh Co.): a Model CCPM-8000 dual-type pump, a Model PT-8000 solvent switching valve, a Model AS-48 autosampler equipped with a 100-~d loop injector and a cooling device, a Model FS-8010 fluorescence detector, a Model CO-8000 column oven controlled at 40°C and a Model CP-8000 chromatoprocessor. Acidified or diluted urine samples were centrifuged and injected onto a TSKgel ODS 80Ts (4.6 i.d. x~il50 mm, 40°C, Tosoh Co.) column. Urinary acylcarnitines were eluted by a step-~radient system with three solvents at a flow rate of 0.5 ml/min; solvent A: I% (v/v) methanol in sodium phosphate buffer (5 mmoi/I, pH 4.0t containing 0.25 mmol/I of sodium heptanesulfonate and 0.15 mol/I of sodium sulfate, 0-8 min; solvent B: 6% (v/v) methanol in the phosphate buffer, 8-22 rain; solvent C: 18% (v/v) ~nethanol in the phosphate buffer, 22-40 min. The eluent was mixed with the substxate solution at a flow rate of 0.5 ml/min. The substrate solution was 1.0 mmol/l Of/~-NAD + and 25 ~mol/I of resazurin in Tris acetate buffer (0.2 mol/I, pH 9.0), which was cooled on ice. The ciuent mixed with the substrate solution was passed throngh the immobilized enzyme reactor at 40°C and the fluorescenct intensity of rezorufin was monitored at gEX = 560 nm and ~EM = 580 nm. After the column ~,as equilibrated with solvent A for 15 min, the next sample was injected. Results

Reaction conditions for enzyme reactor To examine the optimum reaction conditions of the immobiliz,:d enzyme reactor, 100/d each of free, acetyi- and octanoylcarnitine (10/~moi/I) mixed with 0.2 mol/l

139

of Tris acetate buffer (pH range 8.0-9.5),/~-NAD* and resazurin solution were introduced into the immobilized enzyme reactor, Maximum peak height responses of the immobilized enzyme reactor for free, acetyl- and octanoylcarnitines were similar in the pH range 8.5-9.25. Relative hydrolytic activities of the immobilized acylcarnitine hydrolase calculated from the peak height ratios of acetyl/free and octanoyl/free carnitine were highest below pH 8.75. A pH of 8.75-9.0 was selected as the working range. Acylcarnitine hydrolase and carnitine dehydrogenase are not stable in organic solvents, but organic solvents were required in the mobile phase for the elution of acylcarnitines from the reversed-phase column. Methanol inhibited the activities of both enzymes when its concentration was over 20% (v/v) and a high concentration of methanol irreversively inactivated the enzymes. Acetonitrile inhibited the enzyme activities at lower concentrations than methanol. These results indicated that the mobile phase for the HPLC should contain less than 20% (v/v) methanol.

Separation of acylcarnitines Sodium heptanesulfonate was added to the mobile phase at a concentration of 0.25 mmol/I to retain and separate carnitine and acetylcarnitine on the reversedphase column. Since peak broadening occurred, sodium sulfate was also added at a concentration of 0.15 mol/I. Glutarylcarnitine was retained on the column at the acidic pH and overlapped acetylcarnitine when the mobile phase pH was above 5.0. The pH of the phosphate buffer was therefore maintained at 4.0. A typical elution chromatogram for the standard mixture (each 20/=mol/l) is shown in Fig. 2. Sufficient separation of free carnitine and seven acylcarnitines were obtained within 50 min. The small peaks at the retention times of 19 and 33 min were solvent peaks caused by the step-gradient elution.

14-

~

23~

56

12' 10'

g8" 6.

2.

oo

2b

3"o 4b

retention time (min)

so

Fig. 2. Typical elution chromatoggam of a standard mixlure ofcamitine and acylcarnitines. I, cart~Jtine: 2, acetylcarnitine; 3, glutarylcarnitine; 4, propionylcarnitlne; 5, isobutyrylcarnitine; 6, butyrylcarnitine; 7, isovalerylcarnitine; 8, valeryicarnitine. A 100-pl standard mixture (each 20 tAmol/l, 2 nmol as injection amounts) was injected onto the column.

140 TABLE i

Within-run and day-to-day imprecision (n = 5) Imprecision (C.V. %)

Camitines

Within-run

Free Acetyl Glutaryl Propionyl isobutyryl Butyryl lsovaleryl Valeryl

Day-to-day

A

B

A

B

2.2 3.1 1.5 3.2 2.2 2.2 2.0 2.9

0.5 2.6 i.6 2.0 I. I 1.7 0.2 1.5

5.8 3.4 2.6 1.9 3.2 1.9 1.6 2.6

3.8 !.0 4.0 3.5 4.3 3.8 2.4 3.0

Samples A and B contained I0 and 50 t~mol/1, respectively, of eac'a standard acylcarnitine.

Linearity, within-run and day-to.day imprecision and recovery Non-linear calibration curves were obtained in this method, which were caused by the large Michaelis constant of carnitine dehydrogenase for L-carnitine (9.3 retool/I). The responses for acetyl-, glutaryl- and propionylcarnitines were similar to those for free carnitine, but the responses for isobutyryl- and butyrylcarnitines and isovaleryl, and valerylcarnitines were 2-10"/o and 30-35% lower than those for free carnitine, respectively. The detection limits of this method (signal/noise = 2) were 0,5-1.0 ~mol/I (50-100 pmol injected). Within-run and day-to.day imprecisions (n = 5) for analysis of the standard mixture are shown iq Table i, Good imprecisions (C,V. < 6%) were obtained both in the assay of the low (10 ~mol/I) and high (50 0traol/I) concentrations of acyicarTABLE II

Analytical recoveries of acylcarnitines (, = 3) Carnitines

Recovery (%, mean 4. S,D,) A

B

99.2 4. $.5 100.3 4. 1.7 96.1 .*- 3,0 97.~ 4. 1,6 96,8 4. 2.0 97.9 4. 1.6 97.9 4. 2,7 100,1 4-2,7

96,'/4. 53 100,8 4. |.5 98.8 4. 1,2 99,3 4. 0,9 99,8 s: 0,$ 99.8 :t: O.S 98,3 4. i.0 98,1 4. I,I

nm

Free Acelyl Glutaryl Propionyl Isobutyryl Butyryl lsovaleryl Valeryl

Sample~; A and B were prepared by adding each of the standard acylcarnitines to human urine (1:4, by vol.) at a final concentratio, of I0 and 50 ;tmol/I, respectively,

141

20-

20-

"" 16 E

16

.- 12

12

.m 0

.c

8

8'

0

4 0

0 20'

2b

4b

0o

4b

'°1 16

•~ 16' E -, 12

12:

o~

8

"8 0

4

n. 4

0

2b

4b

retention time (mini

0

0

2'o

4'o

retention time (min)

Fig, 3, Typical elution chromatograms of urine samples. (A) normal child; (B) propionic aciduria; (C) methylmalonic aciduda; (D) glutaric aciduria (type I!).

nitines. Sufficient analytical recoveries (n -- 3) of the standard mixture from human urine samples were also obtained, as shown in Table II.

Application to patient urine samrles Typical elution chromatograms of the urine samples of the healthy child (no dilution) and the patients with propionic acidemia (x20 dilution), methylmalonic acidemia (x 20 dilution) and glutaric aciduria (type 11, x 5 dilution) are shown in Fig. 3. In the urine of the healthy child, free and acetyl carnitines were observed to be the major peaks and isobutyrylcarnitine and several minor peaks were also detected. On the other hand, propionylcarnitine was a major peak in the urine of propionic acidemia and methylmalonyl acidemia after supplements of L.carnitine. In the case of glutaric aciduria (type ll), glutaryl., isobutyryl-, isovaleryi-, valerylcarnitine and several unidentified acylcarnitines were also observed. Discussion

The determination of individual acylcarnitines in biological fluids is important for the study of the metabolic state in seco~ldary carnitine deficiencies caused by organic acidurias. Gas chromatography (GC) of liberated acids following hydrolysis [12] is

142

insensitive and it is difficult to avoid contamination from other acyl-containing compounds. GC-mass spectrometry (MS) involving conversion of acylcarnitines into their corresponding volatile lactones [ 13] provides excellent separation of acyicarnitines; however, it needs some clean-up procedure and cannot determine dicarboxylcarnitines. Fast atom bombardment (FAB~-MS [14,15] and proton nuclear magnetic resonance (NMR) [16] require expensive instruments. HPLC is a convenient method for analysis of biological samples. The radioisotope-exchange method [I 7,18] and the coe.nzyme A esters determination method [19,20] are limited by the carnitine acetyl transferase (CAT) reaction, since mediumchain acylcarnitines are poor substrates for CAT and dicarboxylic acyicarnitines are not enzymatically exchanged. The former is sensitive but requires ['~H]carnitine, the latter is insensitive and cannot be applied to the determination of urinary acylcarnitines. Thermospray HPLC-MS [21] requires an expensive instrument. HPLC with a carboxylic acid analyzer [22,23] is too insensitive to determine subnormal levels of urinary acylcarnitines and concomitant organic acids interfere with the p:ocedure. Precolumn derivatization HPLC with 9-anthryldiazomethane [24] or 4'-bromophenacyl triflate [25,26] provide interesting approaches but have not yet been applied to biological samples. A HPLC method combined with a specific and sensitive enzyme reactor has been developed. This is the first report of the isolation of acylcarnitine hydrolase from a bacterial species, which has quite different enzymatic properties from that isolated from rat liver [27]. The enzyme from rat liver hydrolyzed only octanoyl-, decanoyland palmitoyicarnitines with large Michaelis constants (3.0, 2.0 and 5.0 retool/I, respectively), but did not hydrolyze acetyl- and propionylcarnitine [27]. On the other hand, bacterial acylcarnitine hydrolase exhibited a broad reactivity with small Michaelis constants (20-50/~mol/I) for acylcarnitines including short-chain acylcarnitincs and it also hydrolyzed dicarboxylcarnitine such as glutaryh.,~ nitine. The sensitivity of our method was compatible with that of the radioisotopicexchange method [17,18], which was sufficient to analyze acylcarnitines in normal human urine. Glutaryicarnitin¢ could also be determined with the s~,ne sensitivity as those of free and other acylcarnitines; however, we could not determine other dicarboxylcarnitines because these standard samples were unavailable. This method can be directly applied to the determination of acylcarnitines in human urine and it is useful for diagnosis of organic acidurias.

References I 2 3 4 S

StanleyCA. New genetic defects in mitochondrial fatty acid oxidation and carniline deficiency. Adv Pediatr 1987;34:59-88, Siliprandi N, Sartorelli L, Cim'.m M, di Lisa I,'. Carnitinc: metat~)lism and clinical chemistry. Clin Chim Acta 1989:183:3-12. Duran M, Leer NE, Ketting D, Dorland L. St,~ondary carnitinc deficiency. J £'lin ('hem Clin Biochcm 1990;28:359-36.1. Ro.~sleC, Koh~e KP, Franz HE, Furst P, An improved method I'~)r the determination of free and esterified carnitine. Clin Chim Acta 198S;149:263-268, Rodriguez.,~gad¢ S. de la Pena CA, Paz JM, [:)cl Rio R. Determination of [.-carnitinc in serum, and implementation on the ABA-100 and CemrifiChcm 600, Clin Chcm 1985;31:754-757,

143 6 Cederblad G, Harper P, Lindgren K. Spectrophotometry ofcarnitine in biological fluids and tissue with a Cobas Bio centrifugai analyze~. Clin Chem 1986:32:342-346. 7 Schafer J, Reichmann H. A spectrophotometric method for the determination of free and esterified carnitine. Clin Chim Acta 1989:182:87-94. 8 Deufel T. Determination of L-carnitine in biological fluids and tissues. J Clin Chem Clin Biochem 1990:28:307-31 I. 9 Matsumoto K, Yamadd Y, Takahashi Met al. Fluorometric determination of carnitine in serum with immobilized carnitine dehydrogenase and diaphorase. Clin Chem 1990:36:2072-2076. 10 Matsumoto K, Yamada Y, Yuki H et al. Fluorometric determination ofcarnitine in serum and urine with immobilized carnitine dehydrogenase and its clinical application. J Pharmacobio-Dyn 1991;14:s-98. I! Bohmer T, Bremer J. Pr•pionylcarnitine, physiological variations in vivo. Biochim Biophys Acta 1968:! 52:559-567. 12 Duran M, Ketting D, van Vossen R et al. Octanoylglucuronide excretion in patients with a defective oxidation of medium-chain fatty acids. Clin Chim Acta 1985;152:253-260. 13 Lowes So Rose ME. Simple and unambiguous methods for identifying urinary acylcarnitines using gas chromatography-mass spectrometry. Analyst 1990:I 15:51 I-516. 14 Montgomery JA, Mamer OA. Measurement of urinary free and acylcarnitines: quantitative acylcarnitine profiling in normal humans and in several patients with metabolic errors. Anal Biochem 1989:i76:85-95. 15 Millington DS, Norwood DL, Kodo H, Roe CR, l noue F. Application of fast atom bombardment with tandem mass spectrometry and liquid chromatography/mass spectrometry to the analysis of acylcarnitines in human urine, blood, and tissue. Anal Biochem 1989:180:331-339. 16 lies RA, Hind A J, Chalmers RA. Use of proton nuclear magnetic resonance spectroscopy in detection and study of organic acidurias. Clin Chem 1985:31:1795-1801. 17 Kerner J, Bieber LL, A radioisotopic exchange method for quantitation of short-chain acid-soluble acylcarnitines, Anal Biochem 1983:134:459-466. 18 Schmidt-Sommerfeld E, Penn D, Kerner J, Bieber LL. Analysis of acylcarnitine in normal human urine with the radioisotopic exchange-high performance liquid chromatographic method. Clin Chim Acta 1989:181:231-238. 19 Takeyama N, Takagi D, Adachi K, Tanaka T. Measurement of free and esterified carnitine in tissue extracts by high-performance liquid chromatography. Anal Biochem 1986;158:346-354, 20 Dugan RE, Schmidt MJ, Hoganson GE, Steele J, Colles BA, Shug AL. High-performance liquid chromatography of coenzyme A esters formed by transesterification of short-chain acylcarnitines: diagnosis of acidemias by urinary analysis. Anal Biochem 1987:160:275-280. 21 Yergey AL, Liberato DJ, Millington DS. Thermospray liquid chromatography/mass spectrometry for the analysis of L-carnitine and its short-chain acyl derivatives. Anal Biochem 1984:139:278-283. 22 Kidouehi K, Sugiyama N, Morishita H, Wada Y, Nagai S, Sakakibara J. Analytical method for urinary glutarylcarnitine, acetylcarnitine and propionylcarnitine with a carbo~tylic acid al~alyzer and a reversed-phase column. J Chromatogr 1987:423:297-303. 23 Kidouchi K, Niwa T, Nohara D et al. Urinary acylcarnitines in patients with neonatal multiple acylCoA dehydrogenation deficiency, quantified by a carboxylic acid analyzer with a reversed-phase column. Clin Chim Acta 1988:173:263-272. 24 Yoshida T, Aetake A, Yamaguchi H, Nimura N, Kinoshita T. Determination of carnitine by highperformance liquid chromatography using 9.anthryldiazomethane. J Chromatogr 1988:445: 175-182. 25 Minkler PE, IngaUs ST, Hopr~,l CL. Determination of total carnitine in human urine by highperformance liquid chromatography. J Chromatogr 1987:420:385-393. 26 Pourfrarzan M, Bartlett K. Synthesis, characterization and high.performance liquid chromatography of C~-Cle dicarboxylyl-mono-coenzyme A and mono-carnitine esters. J Chromatogr 1991:570:253-276. 27 Mahadevan S, Sauer F. Carnitine ester hydrolase of rat liver. J Biol Chem 1969:244:4448-4453.