Quantitation of Acyl-CoA and Acylcarnitine Esters ... - Semantic Scholar

T H E ~ O U R N A LOF BIOLOGICAL CHEMlSTHY IC:’

Vol. 266, No. 34, Issue of December 5. pp. 22932-22938, 1991

1991 hy The American Society for Biochemistry and Molecular Biolom, Inc.

Printed

in W.S.A.

Quantitation of Acyl-CoA and Acylcarnitine EstersAccumulated during Abnormal Mitochondrial Fatty Acid Oxidation* (Received for Publication, April 9, 1991)

Rajinder Singh Kler$,Sandra Jackson$,Kim BartlettS, LaurenceA. Bindoff$, Simon Eaton$, Morteza PourfarzamS, FrankE. Frermans, StephenI. Goodmans, Nicholas J. Watmoughs, and Douglass M. Turnbull$ll From the $Divisionof Clinical Neurosciences and Department of Child Health, the Medical School, Universityof Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, United Kingdomand the OB. F. Stolinsky Research Laboratories, Department of Pediatrics, University of Colorado Health Science Center, Denuer, Colorado 80262

We have used radio-high pressure liquid chromatog- CoA dehydrogenases are linked to the respiratory chain by raphy to study the acyl-CoA ester intermediates and electrontransfer flavoprotein (ETF)’andETF:ubiquinone the acylcarnitines formed during mitochondrial fatty oxidoreductase. All these enzymes are essential for the comacid oxidation. During oxidation of [U-’4C]hexadeca- plete oxidation of saturated long chain fatty acids. noate by normal human fibroblast mitochondria, only Theorganization of the enzymes of @-oxidationin the the saturatedacyl-CoA and acylcarnitine esters can bemitochondrial matrix is uncertain. There is littleevidence for detected, supporting theconcept that theacyl-CoA de- the formation of amultienzyme complex involvingall the hydrogenase step is rate-limiting in mitochondrial ,& enzymes. Nevertheless, some interaction between the @-oxioxidation. Incubationsof fibroblast mitochondria from Bieber patients with defects of @-oxidation showan entirely dation enzymes seems likely. For example, Kerner and (1990) recently showed that a preparation of mitochondrial different profile of intermediates. Mitochondria from patients with defects in electron transfer flavoprotein carnitine palmitoyltransferase was associated with a complex containing 2-enoyl-CoA hydratase, 3-hydroxyacyl-CoAdeand electron transfer flavoprotein:ubiquinone oxidoreductase are associated with slow flux through 8- hydrogenase, and 3-oxoacyl-CoA thiolase activities. In addioxidation and accumulation of long chainacyl-CoA and tion, we haveshown that adefect of the long chain 3acylcarnitine esters. Increased amounts of saturated hydroxyacyl-CoA dehydrogenase was associatedwith lowered medium chain acyl-CoA and acylcarnitine esters are activity of the long chain 2-enoyl-CoA hydratase and thelong detected in the incubations of mitochondria with me- chain 3-oxoacyl-CoA thiolase (Jackson et al., 1991). The significance of these interactions for the transfer of substrates dium chain acyl-CoA dehydrogenase deficiency, between enzymes is unknown. whereas long chain 3-hydroxyacyl-CoA dehydrogenase deficiency is associated with accumulation of long @-Oxidationof long chain fatty acids is regulated by a chain 3-hydroxyacyl- and 2-enoyl-CoA and carnitine number of different factors, including substrate supply, the esters. These studies showthat thecontrol strength at recycling of cofactors, and the rate of oxidation of acetyl-coA. the siteof the defective enzyme has increased. Radio- The entryof acyl groups into the matrixspace is catalyzed by high pressure liquid chromatography analysis of inter- the carnitine palmitoyltransferases and the carnitine/acylcarmediates of mitochondrial fatty acid oxidation is an nitine translocase. Carnitine palmitoyltransferase I is inhibimportant new technique to study the control, organi- ited by malonyl-CoA, and this is the mechanism by which a zation and defects of the enzymes of &oxidation. fatty acid synthesis/fatty acid oxidation futile cycle is avoided and by which a carbohydrate-rich diet inhibits ketogenesis (McGarry and Foster, 1980). The control of @-oxidationin Mitochondrial fatty acid oxidation of saturated fatty acids the mitochondrial matrix is shared by several sites and deinvolves the repeated sequence of flavoprotein-linked dehy- pends upon product inhibition, the redox state, and the rate of recycling of CoA (Schulz, 1985). drogenation, hydration, NAD+-linked dehydrogenation, and Defects of the enzymes of mitochondrial fatty acid oxidathiolysis to generate acetyl-coA (Schulz, 1990). There aretwo tion are important causes of disease, especially in children. or three enzymes with overlapping chain length specificities for each of these reactions; thus, there are thoughtbe to three The clinical presentations include hypoglycemic coma, myacyl-CoA dehydrogenases, two enoyl-CoA hydratases, two 3- opathy, cardiomyopathy, encephalopathy, and liver disease. hydroxyacyl-CoA dehydrogenases, and two 3-oxoacyl-CoA Defects have been identified in short-chain (Turnbull et al., thiolases (Schulz, 1990). The three chain lengthspecific acyl- 1984), medium chain (Stanley et al., 1983), and long chain (Hale et al., 1985) acyl-CoA dehydrogenases, ETFand ETF:ubiquinone oxidoreductase (Loehr et al., 1990), andlong * This work was supported by Action Research for the Crippled Child, the Muscular Dystrophy Groupof Great Britain, the Wolfson chain 3-hydroxyacyl-CoAdehydrogenase (Jackson et al., Foundation, the Foundation for the Studyof Infant Deaths, Sigma1991).Despitetheir clinical importance,the diagnosis of Tau (D. M. T., K. B.), and National Institutes of Health Grants defects of fatty acid oxidation remains difficult (Bartlett et HDOR315 and HD04024 (F. E. F., S. I. G.). The costs of publication of this article were defrayed in partby the payment of page charges. al., 1991). To study the organization and control of the enzymes of Thisarticlemusttherefore be herebymarked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ’ The abbreviations used are: ETF, electron transfer flavoprotein; 1 T o whom correspondence should he addressed: Divisionof ClinicalNeuroscience,University of Newcastleupon Tyne, Medical HPLC, high pressure liquid chromatography; HEPES, 442-hydroxSchool, Framlington Place, Newcastle upon Tyne NE2 4HH, United yethy1)-1-piperazineethanesulfonic acidEGTA,[ethylenebis(oxyKingdom. Tel.: 011-44-91-222-7051; Fax: 011-44-91-222-7424. ethylenenitri1o)ltetraaceticacid.

22932

Intermediates of Mitochondrial Oxidation Acid Fatty

22933

mitochondrial (3-oxidation and their perturbation by disease 3 mM EDTA, pH 7.5. The fibroblasts were disrupted using a ground and toxins, we have developed methods to measure the acyl- glass hand held homogenizer (10 passes)followed by homogenization in a loose fitting, power-driven Potter-Elvehjem homogenizer (10 CoA esterintermediatesandtheacylcarnitinesgenerated passes). The homogenate was made up to 10 ml and centrifuged (571 during hexadecanoate oxidation (Bhuiyan and Bartlett,1988; X g for 10min) topellet the cell debris. The supernatantwas decanted Watmough et al., 1989). This paperdescribes further improve- and centrifuged (14,290 X g for 10 min) to obtain the mitochondrial ments in our methods to identify acyl-CoA and acylcarnitine pellet. The pellet was rehomogenized as before and again centrifuged esters and the use of these techniques to study fatty acid a t 571 X g for 10 min; the supernatant was decanted and centrifuged oxidation in mitochondriafrom human skin fibroblasts. Using (14,290 X g for 10 min). The mitochondrial pellets were combined, resuspended, and recentrifuged (14,290 X g for 10 min) togive a final these analytical techniques, combined with measurement of mitochondrial pellet thatwas resuspended andused for the measureflux through (3-oxidation, we wished to determine 1) in mito- ment of citratesynthaseactivity(ShepherdandGarland, 1969), chondria from control subjects, which enzyme steps seem to succinate-ferricyanide reductase (Turnbull et al., 1982), protein conbe rate-limiting, 2) in mitochondria from patients with well centration (Lowry et al., 1951) and fatty acidoxidation flux and defined defects of fatty acid oxidation, what is the effect on intermediates. flux and the pattern of intermediates, 3) is the pattern of Incubations with /U-"C]Hexadecanoate intermediates specific for each defect of @-oxidation, and if Incubations were made a t 37 "C in 1 ml of medium containing 110 so, would it be valuable in the diagnosis of the site of the mM KCl, 10 mM HEPES, 5 mM MgCl,, 10 mM potassium phosphate, defect in disordersof (3-oxidation. 1 mM EGTA, 0.2 mg of cytochrome c, 5 mM ATP, 1 mM ADP, 100 CoASH, 1 mM L-carnitine, pH 7.2, in 20 ml of scintillation vials held in a shaking water bath (170 strokes/min). Mitochondrial protein (2-4 mg for detection of intermediatesand 0.2-0.4 mg for flux Patient Material measurements) was preincubated in thismedia for 6 min. Incubations Cultured fibroblasts were grown in Eagle's minimal essential me- for the determination of intermediates were started by addition of 120 nmol of [U-'4C]hexadecanoate complexed tofat-free bovine dium containing Earle's salts, supplemented with 10% fetal bovine serum albumin in a molar ratio of 5:l (specific activity, 52 pCi/pmol) serum, antibiotics, and nonessential amino acids. and were terminated after 30 min by addition of 200 pl of acetic acid Controls-These were cell lines obtained from normal subjects. 50 nmol of heptadecanoyl-CoA and undecanoylcarnitine were then ETF-deficient-Thecase histories of thesechildren havebeen previously described: 1728 (Vergee and Sherwood, 1985), 1901 (Loehr added as internal standards. The incubations to measure flux were initiated by the addition of et al., 1990), 1902 (Niederwieser et al., 1983). The ETF activity for cell line 1728 was 0.32 milliunits. mg of protein"; for cell line 1901 it 40 nmol [U-I4C]hexadecanoate complexed toserumalbuminand was0.01 milliunits.mg of protein"; for cell line 1902 it was 0.01 stopped at appropriate time pointsby addition of 200 pl of 5 M HC10, milliunits.mg of protein" (control range, 1.1-2.50; Loehret al. followed by 120 nMof hexadecanoate complexed to bovine serum (1990)). albuminin a 5:l molar ratio.Precipitatedproteinand remaining E7'F:Ubiquinone Oxidoreductase-deficient-The clinical features substrate were removed by centrifugation (90,000 X g,, for 10 min), have been reported for the patients fromwhom fibroblasts lines 1730 and radioactivity was measured in a 200-pl aliquot of the supernatant (CoudL et al., 1981) and 1808 (Loehr et al., 1990) were derived. The by scintillation counting. ETF:ubiquinone oxidoreductase activity for cell lines 1730 and 1808 Preparation of Fractions Containing Acyl-CoA and Acylcarnitine were 0.1 milliunits. mg of protein" and 0.4 milliunits. mg of protein" Esters-Each sample was transferred from a scintillation vial to an (control range, 5.4-20.5; Loehr et al. (1990)). extraction tube (15 ml) containing 100 pl of saturated NH,SO,. The Medium Chain Acyl-CoA Dehydrogenase-deficient-Both patients scintillation vial was washed twice with 200 p1 of HPLC grade water, with medium chain acyl-CoAdehydrogenase deficiency presented and the washings were added to the extractiontube. The sample was with hypoglycemic episodes associatedwith hypoketonemia. They placed in a water bath for 1 min at 100 "C and then cooled on ice. both excreted dicarboxylic acidsand hadlow medium chain acyl-CoA Each sample was extracted three times with 8 ml of diethyl ether to dehydrogenaseactivity, asdetermined by thefluorometricETFremove the organic acids. Following this, the sample was extracted linked assay (Frerman and Goodman, 1985). for 30 min with 6 ml of methanol/chloroform (2:1, v/v) on a rotating Long Chain 3-Hydroxyacyl-CoA Dehydrogenase-deficient-The pa- table. After centrifugation (40,000 X g,, for 10 min), the supernatant tient E. B. has been previously described in detail (Jackson et al., was retained.Thepellet wasresuspended in 3 ml of methanol/ 1991). chloroform (2:1, v/v) and sonicated in a sonic bath for 10 min and then centrifuged as before. The pellet was discarded, the two superMaterials natants were combined, and the solvent was removed with a stream Bovine serum albumin (Fraction V, fatty acid-free),CoA (grade 11, of nitrogen a t 30 "C. Methanol (2 ml) was added to the dry extract, trilithium salt), ADP, and ATPwere supplied by Boehringer Mann- and the sample sonicated for 15 min in a sonic bath. After centrifugation (20,000 X gav,10 min) the supernatant was decanted and the heim. Acetyl-coA, butyryl-CoA, hexanoyl-CoA, octanoyl-CoA, decanoyl-CoA,dodecanoyl-CoA, andcytochrome c were supplied by pellet was resuspended in 2 ml of methanol and sonicated ina sonic bath for 15 min. This was then recentrifuged, the supernatants were Sigma. DEAE-Sephacel was supplied by Pharmacia LKB Biotechcombined, and 1.3ml of HPLC grade water were added to give a 3:l nology Inc.Acetonitrile (S grade) was purchased from Rathburn Chemicals Ltd., Walkerburn, Scotland. Methanol, Analar grade chlo- methanol/H20 mixture. This mixture was applied to a column (600 75% roform, HPLCgrade water, xylene, Scintron grade Triton X-100 and X 6 mm) of DEAE-Sephacel (acetate form) equilibrated with methanol. The column was washed with 4 ml of 75% methanol, and Scintron grade 2,5-diphenyloxazolewere supplied by BDH, Chemicals the acylcarnitineswere eluated. The acyl-CoA esters were then eluted Ltd. [U-"CC]Hexadecanoate (34.3 GBq/mmol) was obtained from Amer- with 7 ml of 75% methanol containing 0.5 M ammonium acetate and 10 mM acetic acid. sham International, and 1.2 mM [U-"CC]hexadecanoate (155.4 MBq/ Methanol was removedfrom bothfractionswith a stream of mmol) complexed to albuminin a 5:lmolarratio was prepared nitrogen, 3 ml of HPLC grade water was added, and the sampleswere (Sherratt et al., 1988). freeze-dried. The acyl-CoA fraction was dissolved in HPLC grade water, and an aliquotwas analyzed by radio-HPLC. The recovery of Synthesis of Acyl-CoA and Acylcarnitine Esters different chain lengthacyl-CoA esters was assessed using this extracHeptadecanoyl-CoA and undecanoylcarnitine were synthesized as tion method by adding pure acyl-CoA esters to quenched mitochonpreviously described (Bhuiyan and Bartlett, 1988 Watmough et al., drial incubations. The recoveries for acetyl-coA, butyryl-CoA, octa1989). noyl-CoA, hexadec-2-enoyl-CoA, bexadecanoyl-CoA, and heptadecanoyl-CoA were 56.3 k 3.2, 58.6 & 2.8, 60.2 f 4.0, 58.0 f 3.7, 62.9 f Preparation of Mitochondria from Fibroblasts 3.3, and 63.1 4.1, respectively (mean f S.D. for three separate Fibroblasts from three roller bottles (850 cm') were harvested by experiments). Thus, there was no evidence for chain length specific trypsinization, washed twice in phosphate-buffered saline, and resus- losses of acyl-CoA esters; thisallows accurate quantificationusing an pended in medium containing 250 mM sucrose, 20 mM HEPES, and internal standard. EXPERIMENTALPROCEDURES

p~

*

22934


The acylcarnitine fraction wasdissolved in 2 ml of water and TABLE I applied to a Dowex 50W-X8 cationic column (200-400 mesh) in the /U-'4C]Hexadecanoate oxidation in fibroblast mitochondria pyridinium form. The column was washed with 2 ml of 10 mM HCl The rateof mitochondrial fatty acid oxidation was measured using and 2 ml of water, and the acylcarnitineswere eluted with 4 ml of 0.5 [U-''C]hexadecanoate as substrate and measuring acid-soluble radioM pyridinium acetate, pH 4.5, in 50% (v/v) aqueous ethanol and freeze-dried. The residue was redissolved in acetonitrile, and2 mg of activity. The results are expressed as nmol of acetyl units formed KHCO:, and 100 p1 of 4-bromophenacyl bromide were added. The min".mg of protein". QO, ubiquinone oxidoreductase; MCAD, medium chain acyl-CoA dehydrogenase; HOAD, 3-hydroxyacyl-CoA deacylcarnitines were derivitizedby heatingto80 "C for2h. The hydrogenase. recovery of acylcarnitine esters of different chain lengths was measAcid-soluble metabolites line Cell ured using this extraction method by adding pure acylcarnitines to quenched mitochondrial incubations. The recoveries for acetylcarniControl SM 1.73 tine, octanoylcarnitine, undecanoylcarnitine, and hexadecanoylcarControl 1975 1.76 nitine were 31.3 f 5.2, 35.7 f 3.9,37.1 +. 3.2, and 38.2 f 3.6, Control 1980 1.68 respectively (mean & S.D. for three separate experiments). Thus, Mean k SD 1.72 f 0.04 there was no evidence for chain length specific losses of acylcarnitine ETF-deficient 1728 0.25 esters, and accurate quantificationpossible is using an internal stand- 1901 ETF-deficient 0.20 ard. ETF-deficient 1902 0.18 Radio-HPLC Analysis of Acyl-CoA Esters-Radio-HPLC analysis ETF:QO-deficient 1730 0.10 wasperformed as previouslydescribed (Watmoughet al., 1989), ETF:QO-deficient 1808 0.28 except that a Lichrosorb 5RP18 column(250 X 4.6 mm) was used on 1.61 MCAD-deficient S all occasions. Standard mixturesof acyl-CoA, 2-enoyl-CoA,3-hydroxMCAD-deficient P M 1.68 yacyl-CoA, and 3-oxoacyl-CoA esters were chemically synthesized or HOAD-deficient EB 0.65 prepared enzymically (Watmoughet al.,1989), andtheretention times were determined. The identity of each acyl-CoA ester generated during the mitochondrial oxidations was determined by its relative decanoyl-CoA,tetradecanoyl-CoA, dodecanoyl-CoA, decaretention time compared with that of the internal standard. noyl-CoA, octanoyl-CoA, and hexanoyl-CoA(Fig. l ; Table Radio-HPLC Analysis of AcylcarnitinesEsters-Acylcarnitines were analyzed by radio-HPLC usinga 5-pm reversephase Cs Hypersil 11). Hexadecanoyl-CoA was present in much higher concencolumn (250 X 4.6 mm). The following ternary gradientof acetonitrile trations than the other intermediates. However, since hexa(A), water (B), and aqueous 150 mM triethylaminophosphate, pH 5.6 decanoyl-CoA is formed by acyl-CoA synthetase that is situ(C) was used for separation: 0-10 min of isocratic 60% A, 38% B, 2% ated on the outer mitochondrial membrane, it is not known C, 10-25 min of linear gradient to 92%A, 8% C; 25-40 min of linear how much hexadecanoyl-CoA is in the mitochondrial matrix. gradient to 100% A; and 45-55 min of linear gradient to starting Enoyl-CoA, 3-hydroxyacyl-CoA,or 3-oxoacyl-CoA esters were condition followed by 15 min of isocratic equilibration. The flow rate not detected during fatty acid oxidation in fibroblast mitowas 1.2 ml/min. Radioactive analytes were detected using an on-line continuous radioactivity detector, as described previously(Watmough chondria from normal subjects. A similar patternof intermediates is seen in the analysis of et al., 1989). Standard mixtures of acylcarnitine, 2-enoylcarnitine, 3hydroxyacylcarnitine, and 3-oxoacylcarnitine esters were chemically theacylcarnitinefraction.Theintermediatesdetectedare synthesized, their structures confirmed by fast atom bombardment- hexadecanoylcarnitine, tetradecanoylcarnitine, dodecanoylmass spectrometry, and their retention times determined. The iden- carnitine, decanoylcarnitine, octanoylcarnitine,hexanoylcartity of each acylcarnitine ester generated during the mitochondrial nitine,andacetylcarnitine (Fig. 2; Table 111). Wedidnot oxidations was determined by its relative retention time to the interdetect any 2-enoyl-, 3-hydroxyacyl- or 3-oxoacylcarnitine esnal standard.

ters. The proportionof each acyl group present as the carnitine ester increasedwith decreasing chain length. Thus,31% of the CIs acylgroups were present as hexadecanoyl-CoA, whereas only 2% of Cs acyl groups were present ashexanoylPreparation of Mitochondrial Fractions CoA. The proportion of the acyl group present as the acylThe mitochondrial fractions prepared by homogenization CoA or acylcarnitine ester must reflect the amount and kinetic and centrifugation are viable and readily oxidize substrates. characteristics of the carnitine acyltransferases fibroblasts. in Polarographic studies using fibroblast mitochondria show a However, the relatively high concentration of acylcarnitines respiratory control ratio of 3.5 using glutamate as substrate. in the incubations probably does not reflect the situation in Two other methods were used to ensure that the mitochondria the mitochondrial matrix. The acyl-CoA esters (except for were intact. First, citrate synthase activity was measured in hexadecanoyl-CoA) are restrictedto the matrix compartment the presence and absence of Triton X-100 and, in all cases of the mitochondria,whereas the carnitine esters can distribgreater than 90% of activity, was releasedby addition of ute throughout the incubation via the carnitine/acylcarnitine detergent. Second, in some mitochondrial preparations, suc- antiporter. cinate-ferricyanide reductase activity was measured and 9397% was antimycin- or myxothiazol-sensitive. We also conFlux and Intermediates in Fibroblasts from Patients with firmed that the mitochondrial fraction did not contain signifDefects of Fatty Acid Oxidation icant amountsof peroxisomes, since therewas no measurable ETF Deficiency-The rate of fatty acid oxidation in fibroflux through @-oxidation in the presence of the respiratory blast mitochondria from patients with E T F deficiency was chain inhibitors cyanide (1mM), rotenone (2.5 kg.ml"), and 15, 11, and 10% of mean control values for cell lines 1728, myxothiazol (2.5 kg.ml"). 1901, and 1902, respectively. Surprisingly, in viewof the RESULTS

Flux and Intermediatesof Fatty Acid Oxidation of Normal Fibroblasts The rate of @-oxidation, asmeasured by the production of acid-soluble metabolites, was 1.72 k 0.04 (mean k S.D.) acetyl units formed (Table I). The intermediates were determined at 30 min, when flux was linear with time and substrate was not limiting. The only acyl-CoA ester intermediates detected during the @-oxidationof [U-"CC] hexadecanoate were hexa-

differences in activityof E T F between the differentcell lines, the pattern and amount of acyl-CoA and acylcarnitine detected were very similar (Figs. l and 2). Thus, all three cell lines had high concentrations of hexadecanoyl-CoA and hexadecanoylcarnitine (Tables I1 and 111). The concentration of tetradecanoyl-CoA was high in the incubation from the cell line 1901 but normal in the othertwo cell lines. DodecanoylCoA concentrations were low or absent, but dodecanoylcarnitine was detected in all three cell lines. In themore severely


22935

detectable chain-shortened intermediates (Figs. 1 and 2). The concentrations of hexadecanoyl-CoA and hexadecanoylcarnitine were high (Tables I1 and 111). The incubation with mitochondria from the patient with partial ETF:ubiquinone oxidoreductase deficiency had low concentrations of dodecanoyl-CoA. The concentrations of hexadecanoylcarnitine and tetradecanoylcarnitine were high, compared with control val10 were ues. No chain-shortened acyl-CoA or acylcarnitine esters detectable below CI2, although there was a small amount of tricarboxylic acid cycle intermediates andorganic acids. Medium ChainAcyl-CoA Dehydrogenase Deficiency-Acidsoluble metabolites formed during the oxidation of [U-"C] hexadecanoate in thetwo cell lines with medium chain acyl10 CoA dehydrogenase were 97 and96% of control values (Table I). Theapparentlynormal @-oxidationflux is due to the overlapping chain length specificities of the acyl-CoA dehydrogenases and because medium chain acyl-CoA, acylcarnitine esters, and organic acids are acid-soluble. In both cell lines, theconcentrations of hexadecanoyl-CoA andtetradecanoyl-CoA were normal (Figs. 1 and 2). The concentrations of dodecanoyl-CoA and decanoyl-CoA were considerably D elevated, whereastheconcentration of octanoyl-CoA was normal (Table 11).A similar pattern was seen in the acylcarnitine esters with high concentrations of dodecanoylcarnitine, 111). decanoylcarnitine, and octanoylcarnitine (Table Long Chain 3-Hydroxyacyl-CoA Dehydrogenase-Flux 10 through P-oxidation was impaired and the rate of formation of acid-soluble metabolites was 38%of the mean control value. The patternof acyl-CoA esters detected in the incubations of fibroblast mitochondria from this patient are very different fromthat of controls (Fig. 1). There is accumulation of hexadec-2-enoyl-CoA, tetradec-2-enoyl-CoA, and 3-hydroxy10 hexadecanoyl-CoA, which are intermediates not seen in in1 I cubations with fibroblast mitochondria from controls (Fig. 1). The chromatographic separation of 3-hydroxyhexadecanoylCoA and tetradec-2-enoyl-CoA isnot possibleusingonly radiochemical detection; however, there was a peak broadening, showing the presenceof two CoA esters with appropriate retentiontimes(Watmoughet al., 1989), but the relative Minutes amount of eachintermediatecannotbedetermined.The FIG. 1. Radio-HPLC chromatograms of acyl-CoA ester intermediates after incubationof fibroblast mitochondria with concentration of dodecanoyl-CoA is low, and no decanoyl[U-'4C]hexadecanoate. Only clearly identified peaks with appro- CoA, octanoyl-CoA, or hexanoyl-CoA was detected. Analysis priate retention times for acyl-CoA esters are marked. A , control; B , of the acylcarnitine esters revealed an even more dramatic ETF-deficient (1728); C, ETF-deficient (1901); D, ETFubiquinone picture (Fig. 2). The 3-hydroxyacyl-CoA and 2-enoyl-CoA oxidoreductase-deficient (1730);E , medium chain acyl-CoA dehydro- esters seem to beequally good substrates for the acyltransfergenase-deficient(PM);F, long chain 3-hydroxyacyl-CoA dehydrogenase-deficient(EB). The numbers represent the solvent front contain- ases as the saturatedacyl-CoAs, since the 3-hydroxyacylcar(Fig. 2, Table and organic acids ( I ) , nitines and 2-enoylcarnitine esters are detected ingtricarboxylicacidcycleintermediates hexanoyl-CoA ( 3 ) ,octanoyl-CoA ( 4 ) ,decanoyl-CoA (5),dodecanoyl- 111). Unlike defects of ETF or ETF:ubiquinone oxidoreducCoA ( 6 ) ,tetradec-2-enoyl-CoAplus 3-hydroxyhexadecanoyl-CoA(7), tase, the concentrations of hexanoylcarnitine and acetylcartetradecanoyl-CoA (8), hexadec-2-enoyl-CoA(9),and hexadecanoyl- nitine are normal. CoA (IO). 10

1

2: I

affected cell lines, no other chain-shortened intermediates were detectedwhereasdecanoylcarnitine was seeninthe incubation of the mitochondria fromcell line 1728, and in all three cell lines,tricarboxylicacid cycle intermediates and organic acids were detected. ETF:Ubiquinone Oxidoreductase Deficiency-The rate of fatty acid oxidation in fibroblast mitochondria from patients with ETF:ubiquinone oxidoreductasedeficiencywas 6 and 16% of meancontrol valuesfor cell lines 1730 and 1808, respectively. There was considerable differencein the pattern and amount of intermediates seen in the incubations with fibroblast mitochondria from thetwo subjects, which seemed to reflect the severity of the biochemical defect. Thus, the incubation with mitochondria from the patient with severe ETF:ubiquinone oxidoreductase deficiency (line 1730) had no

DISCUSSION

Mitochondrial Preparationsfrom Cultured Skin Fibroblasts-The method described for the preparation of mitochondrial fractions from fibroblasts is both simple and reproducible. The percentage recovery of mitochondria in the final pellet is 25-35%,' and the majority of these mitochondria were intact.Therate of flux through P-oxidation is very similar for the three controls, suggesting that this simple preparative procedure is suitablefor these studies. Measurement of Fluxthrough P-Oxidation-Many radiochemical and spectrophotometric procedures have been developed to measure flux through @-oxidation (Sherratt et al., 1988). In fibroblasts, radiochemical assays that rely on trap-

'L. A. Bindoff and D. M. Turnbull, unpublished work.

22936

Intermediates of Mitochondrial Fatty Acid Oxidation TABLEI1 Acyl-CoA ester intermediates detected during incubation of fibroblast mitochondria with [U-'4C]hexadecanoate Values are nmol of intermediate.mg of protein" for a 30-min incubation. The conditions used are outlined under "Experimental Procedures." The intermediates are TCA + organic acids, solventfront containingtricarboxylic acid cycle intermediatesand organicacids; C , hexanoyl-CoA; C , octanoyl-CoA C,,, decanoyl-CoA; C,,, dodecanoyl-CoA; C14:1C160H, tetradec-2-enoyl-CoA plus 3-hydroxyhexadecanoyl-CoAC,,, tetradecanoyl-CoA; CI6:., hexadec-2-enoyl-CoA; and C,,, hexadecanoyl-CoA. QO, ubiquinone oxidoreductase; MCAD, medium chain acylCoA dehydrogenase; HOAD, 3-hydroxyacyl-CoA dehydrogenase. TCA + organic acids

c6

c8

C,,

Control SM 0.021 0.923 0.022 0.043 Control 1975 0.850 0.014 0.015 0.033 Control 19800.120 1.100 0.097 0.039 0.958 f 0.128 0.018 0.019 0.038 & 0.005 Mean f S.D. ETF-deficient 1728 0.080 ETF-deficient 1901 0.224 ETF-deficient 1902 0.374 ETF:QO-deficient 1730 ETF:QO-deficient 1808 0.200 1.230 0.025 0.165 MCAD-deficient S MCAD-deficient P M 0.920 0.027 0.181 HOAD-deficient EB 0.510

ping I4CO2release from I4C-labeled fatty acids are often used (Kolvraa et al., 1982; Saudubray et al., 1982; Rhead et al., 1983). However, only a small proportion of acetyl units generated by @-oxidation are converted into CO, (Verkamp etal., 1986), and the measurement of flux is difficult and unreliable due to a wide normal range and large interassay variability. We haveshown thatmeasuring acid-solubleradioactivity gives areliable rate of flux that easily identifies defects involving the oxidation of long chain fatty acids. It does not, however, detect defects of medium chain acyl-CoA dehydrogenase, which is due in part to medium chain acyl-CoA and acylcarnitine esters being soluble in perchloric acid. Alternative radiochemical measurements of flux use detritiation of [9,10-3H]hexadecanoate (Moon and Rhead, 1987) or [9,10''Hltetradecanoate (Manning et al., 1990). These assays are valuable in detecting the presenceor absence of a defect, but do not discriminate between abnormalities of the different enzymes. Therefore,theycan be no more thanscreening procedures. Intermediates of Fatty Acid Oxidation in Incubations with Normal Fibroblast Mitochondria-There have been few studies investigating the intermediates of @-oxidation in intact mitochondria.Previousstudieshaveusedradio-gas liquid chromatography(Stewartet al., 1973; Stanley and Tubbs, 1975; Manning etal., 1990), but this haspoor resolving power and does not detect the different classes of intermediates. Our interest in the control and organization of @-oxidation ledus t o develop a system capable of studying the acyl-CoA and acylcarnitineestersformedduringfatty acid oxidationin intact mitochondria (Watmough et al., 1989; Bhuiyan and Bartlett, 1988).Modifications to the original methods now enable us to detect both classes of intermediates from a single incubation. Wehave shown that this techniquereproducible is by preparing a mitochondrial fraction on three separate occasions from thesamefibroblast cell line. Mitochondrial incubations with [U-'4C]hexadecanoate were performed, and the amounts of each acyl-CoA and acylcarnitine ester were determined. The coefficient of variation for each ester was between 5 and 15%. The only intermediates detected in incubations withfibroblast mitochondria from normal subjects were saturated acylCoA esters, which is in good agreement with our findings in human muscle mitochondria (Watmough etal., 1990) and rat liver mitochondria (Stanley and Tubbs, 1975). The sensitivity

C1,

c 1 6

0.093 0.070

0.078 0.060

0.087 f 0.014

0.086 f 0.030 0.042 0.198 0.080

0.590 0.840 0.750 0.726 & 0.127

ClU

C I I .C161 ICL6OH

0.016 0.008 0.012 0.124 0.140 0.032

0.042

0.090 0.072 0.100 0.085

0.065

1.500 1.010 1.200 1.000 0.735 0.750 0.780

of our method is such thatwe could detect CI4 intermediates in amounts down to 20 pmol. The absence of any detectable CI4 intermediate apart from tetradecanoyl-CoA and its corresponding acylcarnitine ester suggests that if 2-enoyl-, 3hydroxy-, or 3-oxoacyl-intermediates are present, they are in extremely low concentrations. Analysis of Intermediates of Fatty Acid Oxidation in the Diagnosis of Defects of This Pathway-The diagnosis of defects of fatty acid oxidation is a difficult but very important clinical problem (Jackson and Turnbull, 1991). Studies of the flux through @-oxidation may detect the presence of an abnormality butwill not give the site. Thus, identifying the site of the defect depends upon the measurementof the enzymes of @-oxidation,which is time-consuming and difficult. We therefore proposed that if a defect of @-oxidation is present, there will be accumulation of intermediates proximal to the block. Since thedefects involve different enzymes with different substrate and chain length specificities, the pattern of intermediates should be characteristic for each enzyme defect. The value of the technique is confirmed by the present studies of cell lines from patients with established defects of the enzymes involved in @-oxidation. Patients with E T F or ETF:ubiquinone oxidoreductase deficiency show not only slow flux through @-oxidation but also undetectable or very low acetylcarnitine. In addition, there were high concentrations of long chain acyl-CoA and acylcarnitine esters, butno medium chain intermediates. This pattern of intermediates is entirely consistent with theknown site of the defect and was similar whether thedefect involved E T F or ETF:ubiquinone oxidoreductase. The intermediates that accumulate inmedium chain acylCoA dehydrogenase deficiency are also characteristic, with analysis showing increased amounts of medium chain acylCoA and acylcarnitine esters. Whereas the organic aciduria in this condition is predominantlyCs products (sebacate and octanoylcarnitine), we observedhigher concentrations of decanoyl-CoA and decanoylcarnitine. Why the Cs products as compared with the are found more frequently in the urine, Cloproducts,isunknown,although possible explanations include additional chain shortening of Clo fatty acids in the peroxisomes, or tissue-specificdifferences suchthatother tissues predominantly formCs products. The intermediates that accumulated in the mitochondria from the cell line with long chain 3-hydroxyacyl-CoA dehy-


'i 0

22937 N tN

N

a0

..

2 12

C

I1.1

I 8

10

A

6

I

n

12

1

0

10

I

I

0 105

15

20

25

30 40 35

45

Minutes

FIG. 2. Radio-HPLC chromatograms of acylcarnitine esters formed after incubating fibroblast mitochondria with U-14C. Only clearlyidentified peaks with appropriate retention times for acyl-CoA esters are marked. A , control; B , ETF-deficient (1728); C, ETF-deficient (1901); D, medium chain acyl-CoA dehydrogenasedeficient (PM); E , long chain 3-hydroxyacyl-CoA dehydrogenase deficiency (EB). The numbers represent acetylcarnitine (1); butyrylcarnitine (2); hexanoylcarnitine ( 3 ) ;octanoylcarnitine ( 4 ) ;dec-2-enoylcarnitine plus 3-hydroxydodecanoylcarnitine( 5 ) ; decanoylcarnitine (6); dodec-2-enoylcarnitine plus 3-hydroxytetradecanoylcarnitine (7); dodecanoylcarnitine (8); tetradec-2-enoylcarnitineplus3-hydroxyhexadecanoylcarnitine (9); tetradecanoylcarnitine (10); hexadec-2-enoylcarnitine (11); hexadecanoylcarnitine (12).

drogenase deficiency were entirely different from those described above with the detectionof both enoyl and 3-hydroxyacyl intermediates. This finding is similar to that seen in rat liver mitochondria treated with a complex I inhibitor (rotenone) and in muscle mitochondria from a patient with complex I deficiency when there is impaired flux through 3hydroxyacyl-CoA dehydrogenase because of the changes in the redox state of the mitochondria (Watmoughet al., 1990). An additional interesting finding in this patient is the presence of normal amounts of acetylcarnitine and hexanoylcarnitine despite the evidence of slow flux and a long chain 3hydroxyacyl-CoA dehydrogenasedefect. These findings are very different to those in other patients with defectsof long chain fatty acid oxidation andsuggests impaired oxidation of acetyl groups. The likely explanation is that the accumulated 3-hydroxyacyl or 2-enoyl intermediates inhibit either the tri-

I

22938

Intermediates of Mitochondrial Fatty Acid Oxidation

A., Amedee-Manesme, O., Saudubray, J. M., and Frezel, J. (1981) carboxylicacid cycle ortherespiratorychain,preventing further oxidation of acetyl groups.This observation is entirely Hum. Genet. 59,263-265 Frerman, F. E., and Goodman, S. I. (1985) Biochem. Med. 33,38-44 compatible with the finding of hyperlactatemia in this and Fukushima, T., Decker, R. V., Anderson, W. M., and Spivey, H. 0. other patients with long chain 3-hydroxyacyl-CoA dehydro(1989) J. Biol. Chem. 264,16483-16488 genase deficiency (Jackson et al., 1991). Hale, D. E., Batshaw, M. L., Coates, P. M., Frerman, F.E., Goodman, S. I., Singh, I., and Stanley, C. A. (1985) Pediatr. Res. 19,666-671 Intramitochondrial Controlof 8-Oxidation-Only saturated acyl-CoA and acylcarnitine esterswere detected in the incu- Jackson, S., and Turnbull, D. M. (1991) in The Molecular and Genetic Basis of Neurological Disease (Rosenberg, R. N., Prusiner, S. B., bations with mitochondria from normal human fibroblasts, Barchi, R. L., Kunkel, L. M., and DiMauro, S., eds) Buttenvorth’s, and this suggests that control is predominantly exhibited a t Stoneham, MA, in press the acyl-CoA dehydrogenase step. Little is known about the Jackson, S., Bartlett, K., Land, J., Moxon, R. E., Pollitt, R. J., K, of the enzymes in fibroblast mitochondria, but pure prep- Leonard, J. V., and Turnbull, D. M. (1991) Pediutr. Res. 2 9 , 406411 arations of the enzymes from other tissuessuggests the K,,, of all the enzymes are likely to be in the low PM range. Meas- Kerner, J., and Bieber, L. (1990) Biochemistry 2 9 , 4326-4334 S., Gregersen, N., Christensen, E., and Hobolth, N. (1982) urement of the enzymes of p-oxidation in cultured skin fibro- Kolvraa, Clin. Chim. Acta 126, 53-67 blast homogenates suggest that theVmaX is much lower for the Loehr, J. P.. Goodman,. S. I.,. and Frerman, F. E. (1990) Pediatr. Res. acyl-CoA dehydrogenases, compared with the other enzymes 27,’311-315 Lowrv. 0. H.. Rosebroueh. N. J.., Farr., A. L.. and Randall. R. J. (1951) (Jackson et al., 1991). Furthermore,similarfindingshave . . J. Bid. Chkm. 193, $65-275 been reported for rat heart andliver (Yang et al., 1987; Melde et al., 1991), again suggesting that control of flux is likely to Manning, N. J., Olpin, S. E., Pollitt, R. J., and Webley, J. (1990) J. Inherited Metab. Dis. 1 3 , 58-68 be predominantly at this step. McCarry, J. D., and Foster, D. W. (1980) Annu. Reu. Biochem. 4 9 , An alternative explanation for our findings is that there is 395-420 substrate channeling occurring between the enzymes of p- Melde, K., Jackson, S., Bartlett, K., Sherratt, H. S. A,, and Ghisla, S. (1991) Biochem. J. 2 7 4 , 395-400 oxidation. There is now evidence of interaction between the 2-enoyl-CoA hydratases, 3-hydroxyacyl-CoA dehydrogenases, Moon, A., and Rhead, W. J. (1987) J. Clin. Znuest. 7 9 , 59-64 Niedenveiser, A,, Steinmann, B., Exner, U., Newheiser, F., Redweik, and 3-oxoacyl-CoA thiolases with carnitine palmitoyltransU., Wang, M., Rampini, S., and Wendel, U. (1983) Helu. Pediatr. ferase (Kerner andBieber, 1990) and of interaction between Acta 38,9-26 3-hydroxyacyl-CoAdehydrogenase and complex I (Fuku- Rhead. W. J.. Amendt. B. A., Fritchman, K. S., and Felts, S. J. (1983) Science 2i1,73-75’ shima et al., 1989). These interactions may favor substrate Saudubrav. J.-M.. CoudB. F.-X.. Demauere. F.. Johnson. C.. Gibson. channeling, although there is no direct evidence that this K. M., and Nyhan, W.‘L. (1982) Pedi&.’Res. 16,877-881 occurs in mammalian mitochondria. Schulz, H. (1985) in Biochemistry of Lipids and Membranes (Vance Defects of fatty acid oxidation were associated with changes D. E., and Vance, J. E., eds) pp. 116-142, The Benjamin/Cummings in the pattern and amountof intermediates, suggesting that Publishing Co., Menlo Park the control strength at the site of the defective enzyme had Schulz, H. (1990) in Fatty Acid Oxidation: Clinical, Biochemical, and Molecular Aspects (Tanaka, K., and Coates, P. M., eds) pp. 23-36, increased. Patients with partial ETF and ETF:ubiquinone Alan R. Liss, Inc., New York oxidoreductase deficiency had very slow flux through p-oxiD., and Garland, P. B. (1969) Methods Enzymol. 13, 11dation. Unlike the enzymes of ,%oxidation, where there are Shepherd, 16 two or three enzymes to catalyze each step withoverlapping Sherratt, H. S. A., Watmough, N. J., Johnson, M. A., and Turnbull, substrate specificity, ETF and ETF:ubiquinone oxidoreducD. M. (1988) Methods Biochem. Anal. 33,243-335 tase are the only enzymes involvedin the transfer of reducing Stanley, K. K., and Tubbs, P. K. (1975) Biochem. J . 150,77-88 Stanley, C. A., Hale, D. E., Coates, P. M., Hall, C. L., Corkey, B. E., equivalents to the respiratory chain, and each plays an essenYang, W., Kelley,R. I., Gonzales, E. L., Williamson, J. R., and tial role. Flux through @-oxidation in the mitochondria from Baker, L. (1983) Pediatr. Res. 17,877-884 the patient with long chain 3-hydroxyacyl-CoA dehydrogen- Stewart, H. B., Tubbs, P. K., and Stanley, K. K. (1973) Biochem. J. ase deficiency was less severely impaired.This is due to short 132,61-76 chain 3-hydroxyacyl-CoA dehydrogenase, which has consid- Turnbull, D. M., Sherratt, H. S. A., Davies, D. M., and Sykes, A. G. erable activity tolong chain substrates in fibroblasts (Jackson (1982) Biochem. J. 206,511-516 Turnbull, D.M., Bartlett, K., Stevens, D. L., Alberti, K.G.M.M., et al., 1991). Acknowledgment-We manuscript.

are grateful to S. Lowe for help with the REFERENCES

Bartlett, K., Aynsley-Green, A., Leonard, J. V., and Turnbull, D. M. (1991) in Inborn Errors of Metabolism (Schob, J., Van Hoof, F., Vis,H.L., eds) Vol. 24, pp. 19-40, Nestle Nutrition Workshop Series, Vevey/Raven Press, New York Bhuiyan, A.K.M. J., and Bartlett, K. (1988) Biochem. SOC.Trans. 16,796-797 Coudi., F. X., Ogier, H., Charpentier, C., Thomassin, G., Checoury,

Gibson, G. J., Johnson, M. A., McCulloch, A. J., and Sherratt, H. S. A. (1984) N. Engl. J. Med. 31 1,1232-1236 Veriee. Z. H.. and Shenvood, W. G. (1985) J. Inherited Metab. Dis. 8 , Suppl. 2,137-138 VerkamD, J. H.. van Moerkerk, H. T. B., Glatz, J. F. C., Zuurveld, J. G. E - M., Jacobs, A. E. M., and Wagenmakers, J. M. (1986) Biochem. Med. Metab. Biol. 35, 248-259 Watmough, N. J., Turnbull, D. M., Sherratt, H. S. A., and Bartlett, K. (1989) Biochem. J . 2 6 2 , 261-269 Watmough, N. J., Bindoff, L. A,, Birch-Machin, M. A., Jackson, S., Bartlett, K., Ragan, C. I., Poulton, J., Gardiner, R. M., Sherratt, H. S. A., and Turnbull, D. M. (1990) J . Clin. Znuest. 86,177-184 Yang, S.-Y., He, X.-Y., and Schulz, H. (1987) J . Biol. Chem. 262, 13027-13032