Biochem. J. (2009) 420, 429–438 (Printed in Great Britain) doi:10.1042/BJ20081932 429

Modulation of the hepatic malonyl-CoA–carnitine palmitoyltransferase1A partnership creates a metabolic switch allowing oxidation of de novofatty acids1

Marie AKKAOUI*†, Isabelle COHEN*†, Catherine ESNOUS*†, Veronique LENOIR*†, Martin SOURNAC*†, Jean GIRARD*† andCarina PRIP-BUUS*†2

*Institut Cochin, Departement d’Endocrinologie, Metabolisme et Cancer, Universite Paris Descartes, CNRS Unite Mixte de Recherche 8104, and †INSERM U567, 24 rue du FaubourgSaint-Jacques, 75014 Paris, France

Liver mitochondrial β-oxidation of LCFAs (long-chain fattyacids) is tightly regulated through inhibition of CPT1A (carnitinepalmitoyltransferase 1A) by malonyl-CoA, an intermediate oflipogenesis stimulated by glucose and insulin. Moreover, CPT1Asensitivity to malonyl-CoA inhibition varies markedly dependingon the physiopathological state of the animal. In the presentstudy, we asked whether an increase in CPT1A activity solely or inassociation with a decreased malonyl-CoA sensitivity could, evenin the presence of high glucose and insulin concentrations, main-tain a sustained LCFA β-oxidation and/or protect from triacylgly-cerol (triglyceride) accumulation in hepatocytes. We have shownthat adenovirus-mediated expression of rat CPT1wt (wild-typeCPT1A) and malonyl-CoA-insensitive CPT1mt (CPT1AM593Smutant) in cultured fed rat hepatocytes counteracted the inhibitionof oleate β-oxidation induced by 20 mM glucose/10 nM insulin.Interestingly, the glucose/insulin-induced cellular triacylglycerol

accumulation was prevented, both in the presence and absence ofexogenous oleate. This resulted from the generation of a metabolicswitch allowing β-oxidation of de novo synthesized LCFAs,which occurred without alteration in glucose oxidation and glyco-gen synthesis. Moreover, CPT1mt expression was more effectivethan CPT1wt overexpression to counteract glucose/insulin effects,demonstrating that control of CPT1A activity by malonyl-CoA isan essential driving force for hepatic LCFA metabolic fate. Inconclusion, the present study highlights that CPT1A is a primetarget to increase hepatic LCFA β-oxidation and that actingdirectly on the degree of its malonyl-CoA sensitivity may be arelevant strategy to prevent and/or correct hepatic steatosis.

Key words: carnitine palmitoyltransferase 1, fatty acid oxidation,lipogenesis, liver, mitochondrion, triacylglycerol.

INTRODUCTION

The hepatic isoform of CPT1 (carnitine palmitoyltransferase 1;EC 2.3.1.21) (termed CPT1A) is known to be the key regulatoryenzyme in liver mitochondrial β-oxidation of LCFAs (long-chainfatty acids) [1,2]. By converting LC-CoA (long-chain acyl-CoA)into acylcarnitine, CPT1A catalyses the rate-limiting step in theentry of cytosolic LC-CoA into mitochondria where β-oxidationtakes place. Under physiological conditions, lipogenesis andLCFA β-oxidation are tightly regulated in the liver [1]. Malonyl-CoA, the first intermediate in lipogenesis, is synthesized by ACC(acetyl-CoA carboxylase) and is the substrate of FAS (fatty acidsynthase) for de novo LCFA synthesis. Malonyl-CoA is alsothe physiological allosteric inhibitor of CPT1A [3]. Therefore,after feeding a carbohydrate-rich meal, the presence of both highplasma glucose and insulin concentrations stimulates liver glucoseoxidation, glycogen storage and lipogenesis, allowing convertionof excess glucose into LCFAs. The resulting increase in malonyl-CoA level inhibits CPT1A activity. Both exogenous LCFAs takenup by the liver and endogenous LCFAs generated by lipogenesisare then esterified into TAGs [triacylglycerols (triglycerides)]and partly secreted as VLDLs (very-low-density lipoproteins).Conversely, in the fasted state when lipogenesis is low, CPT1Ais retrieved from malonyl-CoA inhibition. Hepatic β-oxidation of

LCFAs released from adipose tissue can then occur to produceenergy, cofactors required for optimal gluconeogenesis and ketonebodies used as fuels by extrahepatic tissues [1,2].

Disturbance of this key regulatory malonyl-CoA–CPT1Apartnership might contribute to hepatic steatosis. InheritedCPT1A deficiency in humans is associated with hypoglycaemiaand hypoketonaemia, as well as hepatic steatosis during fasting[4,5]. Similarly to drug-induced impairment of mitochondrialβ-oxidation [6], pharmacological inhibition of CPT1A activityby irreversible inhibitors, such as tetradecylglycidic acid [7] oretomoxir [8], induces mitochondrial injury leading to steatosisand inflammation. Whereas the consequences of a CPT1A activitydefect are relatively well known, the metabolic impact of anincreased CPT1A activity itself had, until recently, never beenreported in liver cells. During the course of the present study,it was published that CPT1A overexpression in rat hepatocytescultured in the presence of a low glucose concentration increasedLCFA β-oxidation capacity, leading to metabolic reorientation ofexogenous LCFAs taken up by the cells toward oxidation at theexpense of esterification [9]. However, whether such an increasedcapacity to oxidize exogenous LCFAs could be maintained in con-ditions under which this pathway is usually abolished, e.g. in thepresence of high glucose and insulin concentrations known tostimulate de novo lipogenesis [10], had never been investigated

Abbreviations used: ACC, acetyl-CoA carboxylase; ASP, acid-soluble products; βgal, β-galactosidase; CPT, carnitine palmitoyltransferase; CPT1mt,CPT1AM593S mutant; CPT1wt, wild-type CPT1A; FAS, fatty acid synthase; LC-CoA, long-chain acyl-CoA; LCFA, long-chain fatty acid; MCD, malonyl-CoAdecarboxylase; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium; PL, phospholipid; TAG, triacylglycerol;TOFA, 5-(tetradecyloxy)-2-furoic acid; VLDL, very-low-density lipoprotein.

1 This manuscript is dedicated to Professor Denis J. McGarry.2 To whom correspondence should be addressed (email carina.prip@inserm.fr).

c© The Authors Journal compilation c© 2009 Biochemical Society

430 M. Akkaoui and others

in any model. Moreover, a unique feature of CPT1A is that itssensitivity to malonyl-CoA inhibition varies markedly dependingon the physiological state in adult rats. For example, it is increasedby refeeding carbohydrate to fasted rats, by obesity state or afterinsulin administration to diabetic rats, whereas it is decreased bystarvation and diabetes [11–15].

Thus the aim of the present study was to clearly decipherwhether a decrease in CPT1A malonyl-CoA sensitivity representsan efficient strategy to enhance mitochondrial LCFA β-oxidationin liver cells. By overexpressing CPT1As with distinct malonyl-CoA sensitivity in cultured rat hepatocytes, we have demonstratedthat expression of a malonyl-CoA-insensitive CPT1A is moreeffective than overexpressing CPT1wt (wild-type CPT1A) tocounteract the glucose/insulin inhibitory effect on exogenousLCFA oxidation flux and to prevent glucose/insulin-inducedTAG accumulation. Moreover, modulation of the malonyl-CoA–CPT1A partnership generates a metabolic switch allowingβ-oxidation of de novo synthesized LCFAs, hence preventing theiresterification into TAGs. Taken together, these results highlightthat control of CPT1A activity by malonyl-CoA is an essentialdriving force for hepatic LCFA metabolic fate and that, actingdirectly on the degree of CPT1A malonyl-CoA sensitivity may bea relevant strategy to prevent and/or reduce liver steatosis.

EXPERIMENTAL

Materials

Collagenase used to isolate rat hepatocytes was purchasedfrom Roche Diagnostics. Cell culture reagents (mediumM199 with Earl salts, glutamine and Ultroser G) were ob-tained from Invitrogen. [14C]sodium bicarbonate was purchasedfrom PerkinElmer. [1-14C]oleate, D-[U-14C]glucose, L-[methyl-3H]carnitine and [1-14C]acetate were purchased from GE Health-care. TOFA [5-(tetradecyloxy)-2-furoic acid] was a gift fromDr A. Richardson (Merrel National Laboratories, Cincinnati, OH,U.S.A.). TLC silica plates and dexamethasone were purchasedfrom Merck Chemicals. Insulin was obtained from Novo Nordisk.Other biochemicals were purchased from Sigma–Aldrich.

Construction of recombinant adenovirus

CPT1mt (the CPT1AM593S mutant) was constructed using theQuikChange® site-directed mutagenesis kit (Stratagene) usingpYeDP1/8-10 containing the full-length rat CPT1A cDNA [16] asa template, and the forward (5′-CCTCACATATGAGGCCTCCA-GTACCCGGCTCTTCCGAGAAGG-3′) and reverse (5′-CCTTC-TCGGAAGAGCCGGGTACTGGAGGCCTCATATGTGAG-G-3′) primers. Human adenovirus serotype 5 vectors (denoted asAd) encoding either β-galactosidase (Ad-βgal), CPT1wt (Ad-CPT1wt) or CPT1mt (Ad-CPT1mt) under the control of the CMV(cytomegalovirus) promoter were produced by the Laboratoirede Therapie Genique (INSERM U649, Nantes, France).

Isolation of rat hepatocytes

Male Wistar rats (200–300 g) were purchased from ElevageJanvier and adapted to the environment for at least 1 weekprior to the experiment. Rats were housed in plastic cages ina temperature-controlled (21 ◦C) and ventilated room with a12 h light/dark cycle (light from 15:00 h to 03:00 h). Animalshad free access to water and were fed ad libitum a standardrodent chow (A03, SAFE). All procedures were carried outaccording to the French guidelines for the care and use of

experimental animals, and were approved by the DirectionDepartementale des Services Veterinaires de Paris. Hepatocyteswere isolated at 09:00 h by collagenase perfusion into theportal vein as described previously [10]. Isolated hepatocyteswere resuspended in M199 medium containing 5 mM glucose,100 i.u/ml penicillin, 100 μg/ml streptomycin, 0.1% (w/v) BSAand 2 mM glutamine (basal medium). From cell attachment,irrespective of the experiment performed, 1 mM carnitine wassystematically added to avoid a limitation in CPT1A activitywhich could alter LCFA oxidation. Cell viability estimated byTrypan Blue exclusion was always greater than 80 %.

Hepatocyte culture and adenovirus infection

Hepatocytes were plated at a density of 8 × 106 cells/100 mmPetri dish (for isolation of mitochondria), at 0.5 × 106 cells/well of12-well plates {for the MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] as-say} or at 2.5 × 106 cells/25 cm2 flask (for all other experiments)and cultured for 4 h at 37 ◦C in an incubator equilibrated withair/CO2 (95:5) in basal medium supplemented with 100 nMdexamethasone, 10 nM insulin and 2 % (v/v) Ultroser G beforeinfection with adenovirus. After cell attachment, the medium wasremoved and hepatocytes were incubated for 2 h with 1 ml (flask),250 μl (12-well plate) or 3 ml (dish) of basal medium containingeither 0, 2.5 or 5 infectious particles/cell of Ad-βgal, Ad-CPT1wtor Ad-CPT1mt. Thereafter, the infection medium was removedand cells were further cultured for another 40 h, the experimentsbeing performed during the last 24 h.

Immunoblot analysis

Aliquots of proteins were subjected to SDS/PAGE [17] (7% gel)and transferred on to nitrocellulose membrane. Detection ofproteins was performed as described previously [16] using theECL (enhanced chemiluminescence) Pierce detection system(Perbio Sciences SAS). The antibodies used were against ratCPT1A [16], Escherichia coli βgal (Rockland), rat FAS (a giftfrom Dr I. Dugail, INSERM U465, Paris, France), and humanACC2 (Upstate). For the generation of an anti-CPT2 polyclonalantibody, a peptide corresponding to the last 20 C-terminalresidues of human CPT2 was synthesized, conjugated to keyholelimpet haemocyanin, and used to immunize New Zealand Whiterabbits (Neosystem). The immunoblots were quantified using achemigenius apparatus (Syngene).

Immunofluorescence assay

Hepatocytes (7 × 105 cells/well of six-well plates) were culturedon coverslips coated with 4 % (v/v) collagen in 0.001 % aceticacid/PBS. At 40 h after infection (5 infectious particles/cell),cells were fixed with 2% (v/v) formaldehyde, permeabilized with0.1% SDS and non-specific binding of antibodies was blockedby incubation with 10% (v/v) FBS (foetal bovine serum)/PBSfor 20 min. Cells were stained with polyclonal anti-rat CPT1Aand mouse monoclonal anti-cytochrome c (BD Biosciences)antibodies which were detected by a goat anti-rabbit IgGconjugated with Alexa Fluor® 488 and a goat anti-mouse IgG con-jugated with Alexa Fluor® 594 (Molecular Probes) respectively.After incubation with 2 μM Hoechst 33342 for DNA staining,coverslips were mounted on to glass slides with fluoromount G(Clinisciences). Fluorescent staining was viewed in a confocallaser-scanning microscope (Leica TCS SP2 AOBS). The capturedimages were processed using ImageJ software (Wayne Rasband).

c© The Authors Journal compilation c© 2009 Biochemical Society

Modulation of the liver malonyl-CoA–CPT1 partnership 431

Isolation of mitochondria and CPT activity assay

At 40 h after infection, cells were washed and scraped withice-cold PBS. Mitochondria were isolated in an isolation buffer[0.3 M sucrose, 5 mM Tris/HCl and 1 mM EGTA (pH 7.4)] usingdifferential centrifugation [16]. The protein concentration wasdetermined using the Lowry method with BSA as a standard[18]. CPT1 activity was assayed in intact isolated mitochondriaas palmitoyl-L-[methyl-3H]carnitine formed from L-[methyl-3H]carnitine (400 μM; 10 Ci/mol) and palmitoyl-CoA (600 μM)in the presence of 1 % (w/v) BSA. Estimation of the malonyl-CoA IC50 value (the concentration of malonyl-CoA required for50% inhibition) and CPT2 activity were measured as describedpreviously [16].

Measurement of fatty acid metabolism

During the last 24 h of culture, the medium was replaced bya medium containing either 5 mM glucose or 20 mM glucoseplus 10 nM insulin. Fatty acid oxidation and esterification weredetermined during the last 2 h of culture in the presence of0.3 mM [1-14C]oleate (0.5 Ci/mol) bound to 1% (w/v) defattedBSA. After 2 h of incubation, medium was collected to determine14CO2, [14C]ASP (acid-soluble products), unlabelled ketone bodyproduction (acetoacetate and β-hydroxybutyrate) and lactate andpyruvate concentrations. Cells were washed and scraped in PBSto determine [14C]TAG and [14C]PL (phospholipid) productionas previously described [19,20]. For the determination of VLDLsecretion, medium (3 ml) was collected, immediately adjustedto 0.005 % gentamicin, 1 mM EDTA, 0.04% sodium azide and0.02% sodium (ethylomercurithio)-2 benzoate, and was layeredunder 5 ml of 0.15 M NaCl. After centrifugation at 43000 rev./minfor 18 h at 10 ◦C (in a Beckman 70 Ti rotor), the 1 ml top fractioncontaining VLDLs was counted in 10 ml of scintillation liquid ina scintillation counter (Packard).

Measurement of intracellular TAG and Oil Red O staining

For measurement of unlabelled TAGs, cellular lipids wereextracted in chloroform/methanol (2:1, v/v) with vigorous shakingfor 10 min. After centrifugation for 25 min at 1200 g, the lowerorganic phase was collected, dried, and solubilized in chloroform/methanol. Lipid classes were separated by TLC on silica-gel platesby using petroleum ether/diethyl ether/acetic acid (85:15:0.5,v/v/v) as the mobile phase. Lipids were visualized with iodinevapour. Bands were scraped and TAGs were extracted from silicain chloroform/methanol. After centrifugation for 45 min at 1500 gto precipitate silica, and evaporation, TAGs were measured with aPAP 150 TAG kit (Biomerieux). Lipids droplets were detected inhepatocytes fixed with 3 % (w/v) paraformaldehyde in PBS andcoloured with Oil Red O [21].

Measurement of ACC, FAS and MCD (malonyl-CoA decarboxylase)enzyme activity

Cells were scraped in ice-cold 0.25 M saccharose, 1 mM DTT(dithiothreitol), 1 mM EDTA and protease inhibitors, and cytoso-lic fractions obtained by differential centrifugation (10000 gfor 10 min at 4 ◦C, followed by 100000 g for 1 h at 4 ◦C) wereused for ACC [22] and FAS activity [23] assays. MCD activitywas assayed as described previously [24].

Measurement of glucose metabolism

During the last 24 h of culture, the medium was replaced bya medium containing either 5 mM glucose or 20 mM glucoseplus 10 nM insulin in the absence or presence of 200 μM TOFA,

a specific ACC inhibitor [25]. Glucose oxidation and glycogensynthesis were determined during the last 2 h of culture in the pres-ence of either 5 mM [U-14C]D-glucose (133 μCi/mmol) or 20 mM[U-14C]D-glucose (87.5 μCi/mmol) plus 10 nM insulin, withoutor with 0.3 mM oleate bound to 1% (w/v) defatted BSA. After2 h of incubation, the medium was collected to determine 14CO2

and unlabelled β-hydroxybutyrate production, whereas cells werewashed three times with ice-cold PBS and scraped in 1 ml of 5 MKOH. Non-radioactive glycogen carrier (2.5 mg) was added to thelysates (600 μl) and samples were boiled for 30 min. Glycogenwas precipitated overnight at −20 ◦C with two volumes of ice-cold100% ethanol. Precipitated glycogen was centrifuged at 10000 gfor 10 min. Pellets were washed once with 70 % ethanol, resus-pended in 0.5 ml of water, and counted by scintillation counting.The protein concentration was determined using the Bradfordmethod (Bio-Rad) using BSA as a standard. The lipogenesis ratefrom 5 mM [1-14C]acetate (10 μCi per flask) was determinedduring the last 2 h of culture as described previously [21]. Briefly,cells were rinsed twice with ice-cold PBS, immediately frozen inliquid nitrogen and scraped off in 30% (v/v) KOH. Labelled fattyacids were then extracted and counted by scintillation counting.

Measurement of mitochondrial activity

The MTS assay was performed in 12-well plates using theCellTiter 96 Aqueous Non-Radioactive Cell Proliferation Assay(Promega) according to the manufacturer’s instructions.

Statistical analysis

Results are expressed as means +− S.E.M. ANOVA was carried outusing the StatView program (Abacus) to determine differencesbetween groups. When significant differences were detected, aposteriori comparisons between means were conducted using theFisher LSD (least significant difference) test (α = 0.05).

RESULTS

Adenovirus-mediated expression of CPT1wt and CPT1mtin rat hepatocytes

To investigate whether an increase in CPT1A activity solely or inassociation with a decreased malonyl-CoA sensitivity could mod-ulate hepatic LCFA oxidation, primary rat hepatocytes were infec-ted with 2.5 or 5 infectious particles/cell of adenovirus encodingeither CPT1wt or CPT1mt that is active but insensitive to malonyl-CoA inhibition [26–28]. This led to an increase in CPT1A proteinexpression in an adenovirus concentration-dependent mannerin comparison with uninfected hepatocytes (Figure 1a). Theprotein level of another enzyme of the mitochondrial carnitine-shuttle system, CPT2, whose activity is insensitive to malonyl-CoA inhibition [1], was not modified (Figure 1a). As a control,hepatocytes were also infected with an adenovirus-encoding βgal.The resulting βgal expression had no effect on CPT1A and CPT2protein levels when compared with uninfected cells (Figure 1a).Considering these results, the amount of 5 infectious particles/cellwas chosen for subsequent infection experiments. Immuno-fluorescence analysis of infected hepatocytes showed that bothoverexpressed CPT1wt and CPT1mt proteins were co-localizedwith cytochrome c, a mitochondrial marker protein (Figure 1b),indicating, as expected, a mitochondrial localization. Moreover,infection with Ad-CPT1wt and Ad-CPT1mt respectively led to a15- and 11-fold increase in mitochondrial CPT1A protein leveland a 7- and 3-fold increase in CPT1A activity when comparedwith mitochondria isolated from either Ad-βgal-infected oruninfected hepatocytes (Figure 2a). In agreement with Figure 1(a),

c© The Authors Journal compilation c© 2009 Biochemical Society

432 M. Akkaoui and others

Figure 1 Mitochondrial localization of the overexpressed CPT1wt andCPT1mt proteins in cultured rat hepatocytes

Primary rat hepatocytes were infected with either 0, 2.5 or 5 infectious particles/cell of theindicated adenovirus and cultured for 40 h in 5 mM glucose. (a) Immunoblot analysis ofcellular extracts (30 μg of protein) using specific antibodies for βgal, CPT1A and CPT2.Western blots are representative of four independent experiments performed in duplicate.(b) Immunofluorescence analysis detection of the overexpressed CPT1 proteins. Hepatocyteswere double-immunostained with antibodies raised against cytochrome c (Cyt c) and CPT1A.The right-hand panels (merge) represent the overlay of the two left-hand panels. Co-localizationof cytochrome c and CPT1A appears as yellow spots reflecting the merge of red (Cyt c) andgreen (CPT1A) fluorescence. Hoechst 33342 was used to stain the hepatocyte nuclei. Scalebar = 50 μm.

CPT1wt and CPT1mt overexpression had no repercussion on bothmitochondrial CPT2 protein level and activity (Figure 2a).

To evaluate the capacity of CPT1mt to exhibit enzyme activitydespite the presence of malonyl-CoA, malonyl-CoA sensitivitywas then measured in isolated mitochondria. In uninfected hep-atocytes and Ad-βgal- or Ad-CPT1wt-infected hepatocytes, CPTactivity was almost completely suppressed (80%) by the highestmalonyl-CoA concentration used (150 μM) (Figure 2b). Thiswas indicative of a good membrane integrity of the isolated mito-chondria and suggested that only CPT1A activity was measuredwithout any significant contribution of CPT2. Their correspondingIC50 values for malonyl-CoA were not significantly different(uninfected, 2.5 +− 0.7 μM; Ad-βgal, 2.4 +− 0.6 μM; Ad-CPT1wt,12.2 +− 8.7 μM). By contrast, in Ad-CPT1mt-infected hepato-cytes, CPT1A activity was not inhibited by 1–10 μM malonyl-CoA (Figure 2b). Moreover, even in the presence of 150 μM ofmalonyl-CoA, only 43% of the activity was inhibited, whichprobably corresponds to the inhibition of endogenous CPT1A.Taken together, these results confirm that CPT1wt and CPT1mtoverexpression in rat hepatocytes increases mitochondrial CPT1Aprotein levels and activity, CPT1mt being insensitive to malonyl-CoA inhibition.

Figure 2 Effects of CPT1wt and CPT1mt overexpression on CPT1A proteinlevel and activity (a) and malonyl-CoA sensitivity (b) in mitochondria isolatedfrom infected hepatocytes

Mitochondria were isolated 40 h after infection of hepatocytes with 5 infectious particles/cell ofthe indicated adenovirus. (a) Immunoblot analysis of mitochondrial proteins using specificantibodies for CPT1A and CPT2. Western blots are representative of three independentexperiments with mitochondria isolated from separate hepatocyte cultures. CPT1A and CPT2activities were measured in either intact (CPT1A) or solubilized (CPT2) mitochondria. Resultsare means +− S.E.M. of three independent experiments with separate isolated mitochondria.(b) Malonyl-CoA sensitivity of CPT1A was measured in intact isolated mitochondria. Resultsare means +− S.E.M. of two independent experiments with separate isolated mitochondria.*P < 0.01, compared with Ad-βgal; #P < 0.01, CPT1wt compared with CPT1mt.

CPT1wt and CPT1mt overexpression counteracts glucose andinsulin effects on lipid metabolism

CPT1wt and CPT1mt overexpression in cultured rat hepatocytesincreased LCFA β-oxidation capacity under basal conditions ofculture (5 mM glucose), leading to metabolic reorientation of ex-ogenous LCFAs toward oxidation at the expense of esterifica-tion (Supplementary Figure S1 at http://www.BiochemJ.org/bj/420/bj4200429add.htm). We then evaluated their metabolic ef-fects in the presence of high concentrations of glucose and insulin,conditions known to promote hepatic lipogenesis [10], hencepreventing mitochondrial LCFA oxidation [1]. In uninfectedhepatocytes, a 24 h-exposure to 20 mM glucose plus 10 nMinsulin markedly stimulated the expression and the activity of lipo-genic ACC and FAS enzymes (Supplementary Figures S2a–S2cat http://www.BiochemJ.org/bj/420/bj4200429add.htm), de novolipogenesis (see below Figure 6b) and suppressed [1-14C]oleateoxidation by more than 90% (Figures 3a and 3b). Similar results

c© The Authors Journal compilation c© 2009 Biochemical Society

Modulation of the liver malonyl-CoA–CPT1 partnership 433

Figure 3 CPT1wt and CPT1mt overexpression counteracts the glucose/insulin effects on [1-14C]oleate metabolism

At 16 h after infection of hepatocytes with 5 infectious particles/cell of the indicated adenovirus, cells were cultured for 24 h in the presence of either 5 mM glucose (G5) or 20 mM glucose plus 10 nMinsulin (G20 + ins). During the last 2 h of culture, 0.3 mM [1-14C]oleate bound to 1 % defatted BSA was added. Oleate oxidation into CO2 (a) and ASP (b), and oleate esterification into cellularTAGs (TG; c) and PLs (d) were measured at the end of the 2 h-incubation period. (e) Metabolic orientation of oleate as a percentage of the total nmol of metabolized [1-14C]oleate. (f) [1-14C]Oleateincorporation into VLDLs secreted in the culture medium during the 2 h incubation period. Results are means +− S.E.M. of duplicate flasks from four independent experiments. *P < 0.05 and**P < 0.01, compared with Ad-βgal; #P < 0.05, CPT1wt compared with CPT1mt.

were obtained in Ad-βgal-infected hepatocytes. Overexpressionof CPT1wt and CPT1mt did not alter the glucose/insulin-stimulated expression and activity of ACC and FAS, nor theenzymatic activity of MCD which catalyses the degradationof malonyl-CoA into acetyl-CoA (Supplementary Figure S2).However, their overexpression led to a marked increase in[1-14C]oleate oxidation. Indeed, 14CO2 and [14C]ASP productionwere increased by 4- and 6.8-fold respectively, for Ad-CPT1wt-infected hepatocytes, and by 6.6- and 13.6-fold respectively,for Ad-CPT1mt-infected hepatocytes, when compared with Ad-βgal-infected hepatocytes (Figures 3a and 3b). Conversely, onlyCPT1mt expression significantly decreased by 44 % [1-14C]oleateesterification into TAGs when compared with Ad-βgal-infectedcells (Figure 3c), without any significant modification of its

esterification into PLs (Figure 3d). In the presence of glucoseand insulin, the total amount of [1-14C]oleate metabolized wassimilar for all infections (uninfected, 103 +− 21 nmol/2 h per mgof protein; Ad-βgal, 104 +− 18 nmol/2 h per mg of protein; Ad-CPT1wt, 115 +− 12 nmol/2 h per mg of protein; Ad-CPT1mt,138 +− 17 nmol/2 h per mg of protein). Whereas [1-14C]oleateoxidation rate represented only 5% of [14C]oleate metabolizedin Ad-βgal-infected hepatocytes, CPT1wt and CPT1mt over-expression increased this rate up to 30% and 47% respectively(Figure 3e). To determine whether the decrease in [14C]TAGproduction, and hence in esterification flux, was not due to anincreased export of TAG as VLDL, [1-14C]oleate incorporationinto VLDLs secreted into the culture medium was also measured.[1-14C]VLDL secretion was increased by 50% and 80% in

c© The Authors Journal compilation c© 2009 Biochemical Society

434 M. Akkaoui and others

Figure 4 CPT1mt overexpression decreases the glucose/insulin-inducedTAG accumulation in cultured hepatocytes

At 16 h after infection with 5 infectious particles/cell of the indicated adenovirus, hepatocyteswere cultured for 24 h in the presence of either G5 or G20 + ins. (a) Oil Red O staining ofhepatocytes to detect neutral lipid droplets. Oleate (0.3 mM) was added during the last 2 hof culture. Original magnification × 200. (b) Measurement of cellular TAG content. Whenindicated, 0.3 mM oleate was added during the last 2 h of culture. Results are means +− S.E.M.of duplicate flasks from three independent experiments. *P < 0.05 and **P < 0.01 comparedwith Ad-βgal in G20 + ins; #P < 0.05, CPT1wt compared with CPT1mt.

hepatocytes overexpressing CPT1wt and CPT1mt respectively(Figure 3f). However, the corresponding increased amount ofincorporated oleate into VLDL (1.5 and 1.8 nmol/2 h per mgof protein) was 10- and 20-fold lower than the observed reductionin oleate incorporation into TAGs (18 and 35 nmol/2 h per mgof protein). Consequently, oleate incorporated into VLDLs canprobably explain no more than 10 % of the decrease in TAGesterification. Thus, in the presence of glucose and insulin, CPT1Aoverexpression decreased exogenous oleate esterification intoTAGs by enhancing hepatic LCFA oxidation capacity, the effectsof CPT1mt expression always being significantly higher thanthose observed for CPT1wt.

To determine whether this decrease in exogenous oleateesterification was reflected in intracellular lipid content, Oil RedO staining of hepatocytes, which allows visualization of neutrallipids within the cells, and measurement of intracellular TAGcontent were performed. A 24 h-exposure to glucose and insulininduced a huge increase in lipid droplets in Ad-βgal-infectedcells, which was much lower in hepatocytes expressing CPT1mt(Figure 4a). Moreover, whereas the presence of glucose and in-sulin induced a 5-fold increase in TAG content both in uninfectedand Ad-βgal-infected hepatocytes, CPT1mt expression decreasedthe TAG content by 43% when compared with Ad-βgal-infectedhepatocytes (Figure 4b, left-hand panel). CPT1wt overexpressionled to an intermediate decrease in TAG content (Figure 4b, left-hand panel). In these culture conditions, cellular TAG couldresult from esterification of both LCFAs produced by de novolipogenesis and exogenous oleate. To determine the contributionof TAGs coming from lipogenesis, intracellular TAGs were

quantified after a 24 h-treatment with glucose and insulin, butin the absence of oleate. In these conditions, the TAG content wasalso markedly increased in both uninfected and Ad-βgal-infectedhepatocytes (Figure 4b, right-hand panel). CPT1wt and CPT1mtoverexpression led to a 25% and 50% decrease respectively in theglucose/insulin-stimulated TAG accumulation when comparedwith Ad-βgal-infected hepatocytes, with a significant differencebetween CPT1wt and CPT1mt (Figure 4b, right-hand panel).Whether exogenous oleate was present or not, the TAG accumu-lation induced by glucose and insulin was not statisticallydifferent in hepatocytes infected with Ad-βgal (Figure 4b). Thisindicated that accumulated cellular TAGs mostly resulted fromesterification of LCFAs newly synthesized from glucose. Thusthe increased capacity for LCFA oxidation due to CPT1wt andCPT1mt overexpression counteracts the glucose/insulin-inducedaccumulation of TAGs coming from lipogenesis.

Effect of CPT1wt and CPT1mt overexpression onglucose metabolism

To understand how the level of cellular TAGs resulting fromlipogenesis could be decreased, two hypotheses were explored.The first one was that CPT1wt and CPT1mt overexpression couldinduce a decrease in glucose utilization, leading to a lower cellularTAG content. To answer this question, [14C]glucose oxidationand [14C]glycogen synthesis were measured during the last 2 hof culture in the absence of exogenous oleate. As expected,glucose and insulin markedly stimulated glucose oxidation intoCO2 and glycogen synthesis (Figures 5a and 5b, left-hand panels).Infection of hepatocytes with either Ad-βgal, Ad-CPT1wt orAd-CPT1mt had no significant effect on 14CO2 production and[14C]glycogen synthesis (Figures 5a and 5b, left-hand panels). Atrend of decreased 14CO2 production was observed in hepatocytesinfected with Ad-CPT1wt or Ad-CPT1mt. However, pyruvate,lactate, malate and oxaloacetate levels were not altered by CPT1wtand CPT1mt overexpression (results not shown). These resultssuggested that, in our experimental conditions, CPT1wt andCPT1mt overexpression did not alter glucose utilization. Similarresults were observed in the presence of oleate (Figures 5a and 5b,right-hand panels), indicating that, even when LCFA oxidationflux was further enhanced by addition of exogenous LCFAs,glucose metabolism was also not altered.

The second hypothesis was that CPT1wt and CPT1mt over-expression, which leads to an increase in LCFA oxidationcapacity, might allow the oxidation of LCFAs newly synthesizedfrom glucose. To test this hypothesis, β-hydroxybutyrateproduction was measured in hepatocytes cultured in the absenceor presence of 0.3 mM oleate during the last 2 h of culture. Asexpected, in basal conditions of culture (5 mM glucose) and inthe absence of exogenous oleate, the β-hydroxybutyrate levelremained low in uninfected hepatocytes, whereas it increased by2.8-fold upon oleate addition (Figure 6a). In uninfected and Ad-βgal-infected hepatocytes, a 24 h-exposure to glucose and insulindecreased β-hydroxybutyrate production to the same basal levelwhether oleate was present or not (Figure 6a). By contrast, inthe presence of oleate, the β-hydroxybutyrate level increasedby 2.5- and 3.2-fold respectively in CPT1wt- and CPT1mt-overexpressing hepatocytes, hence abrogating the glucose/insulininhibitory effect on ketogenesis (Figure 6a, right-hand panel).The observation that a 1.7-fold (CPT1wt) and 2.3-fold (CPT1mt)increase in β-hydroxybutyrate production occurred in the absenceof exogenous oleate (Figure 6a, left-hand panel) stronglysuggested that LCFAs arising from lipogenesis are oxidized toproduce ketone bodies. To confirm this, hepatocytes were culturedfor 24 h in the presence of TOFA, a specific ACC inhibitor

c© The Authors Journal compilation c© 2009 Biochemical Society

Modulation of the liver malonyl-CoA–CPT1 partnership 435

Figure 5 CPT1wt and CPT1mt overexpression does not alter the glucose/insulin effects on [U-14C]glucose metabolism

At 16 h after infection with 5 infectious particles/cell of the indicated adenovirus, hepatocytes were cultured for 24 h in the presence of either G5 or G20 + ins. During the last 2 h, a tracer amount of[U-14C]glucose was added in the absence or in the presence of 0.3 mM oleate. The medium was then collected to measure [U-14C]glucose oxidation into CO2 (a). Cells were washed and scraped tomeasure incorporation of [U-14C]glucose into glycogen (b). Results are means +− S.E.M. of duplicate flasks from five independent experiments. §P < 0.01, compared with G5.

[25]. As expected, TOFA treatment of uninfected hepatocytestotally inhibited both basal and glucose/insulin-stimulated de novolipogenesis (Figure 6b), as well as glucose/insulin-induced TAGaccumulation (Figure 6c). In these conditions, the increase inβ-hydroxybutyrate production previously observed in CPT1wt-and CPT1mt-overexpressing hepatocytes was fully abrogated(Figure 6a), confirming that the produced β-hydroxybutyratearose from the oxidation of de novo synthesized LCFAs. Thusthe higher induced increase in β-hydroxybutyrate productionobserved in the presence of exogenous oleate (Figure 6a) resultedfrom the oxidation of both exogenous and de novo synthesizedLCFAs. As observed previously, CPT1mt expression had, bothin the absence and presence of oleate, a significantly highereffect than CPT1wt overexpression. Taken together, these resultsindicate that the observed decrease in cellular TAG content inthe absence of exogenous LCFA (Figure 4b) is not due to analteration in glucose utilization following CPT1wt or CPT1mtoverexpression, but results from the oxidation of LCFAs newlysynthesized from glucose.

DISCUSSION

The present study provides the first proof of concept that, evenin conditions under which lipogenesis is stimulated, a directmodulation of the malonyl-CoA–CPT1A partnership in liver cellsis a relevant strategy to maintain a high mitochondrial LCFAβ-oxidation flux which protects against TAG accumulation.

In agreement with previous in vitro studies performed inpancreatic β-cells [27], muscle cells [29] and hepatocytes [9],the present study showed that overexpression of rat CPT1wtor expression of CPT1mt in cultured rat hepatocytes increasedLCFA β-oxidation capacity under basal conditions of culture.To challenge whether this increased oxidative capacity couldbe maintained in conditions under which LCFA β-oxidationis usually abolished, we chose to expose primary cultured rathepatocytes to 20 mM glucose plus 10 nM insulin, conditionsknown to elicit hepatic lipogenesis [10]. As expected, thismarkedly increased glucose oxidation and glycogen synthesis,stimulated ACC and FAS expression and enzymatic activity,as well as de novo lipogenesis, and led to TAG accumulation,whereas it abolished exogenous oleate oxidation and ketogenesis.We have shown that CPT1A overexpression partially (CPT1wt) orfully (CPT1mt) abolished the inhibitory effect of glucose/insulinon oleate oxidation. This allowed, in the case of CPT1mtexpression, the maintenance of an oxidation flux and a ketonebody production from exogenous LCFAs similar to those observedin basal conditions of culture. It is noteworthy that this wasaccomplished without alteration in the stimulatory effects ofglucose/insulin on lipogenic enzyme expression and activityand on glucose utilization. Moreover, the higher effectivenessof CPT1mt than CPT1wt overexpression to counteract theglucose/insulin effects on hepatic lipid metabolism occurreddespite a lower increase in both mitochondrial CPT1A proteinlevel and activity. The biological relevance of this finding isthat it represents the converse experimental demonstration of the

c© The Authors Journal compilation c© 2009 Biochemical Society

436 M. Akkaoui and others

Figure 6 CPT1wt and CPT1mt overexpression allows oxidation of de novo synthesized LCFAs

At 16 h after infection with 5 infectious particles/cell of the indicated adenovirus, hepatocytes were cultured for 24 h in the presence of either G5 or G20 + ins, in the absence or presence of 200 μMTOFA. When indicated, 0.3 mM oleate was added during the last 2 h of culture. (a) The medium was collected to measure β-hydroxybutyrate production. (b) The lipogenesis rate from 5 mM[1-14C]acetate was measured, in the absence of exogenous oleate, in uninfected cells during the last 2 h of culture. (c) Measurement of cellular TAG content. Results are means +− S.E.M. of duplicateflasks from three to seven independent experiments. †P < 0.01, compared without oleate; *P < 0.05 and **P < 0.01, compared with Ad-βgal in G20 + ins; #P < 0.01, as indicated on the Figure.

physiological role of malonyl-CoA in controlling hepatic LCFAmetabolic fate [1–3]. Although liver mitochondrial glycerol-3-phosphate acyltransferase 1 isoform, which catalyses the firststep in glycerolipid synthesis, may be important for LCFAmetabolic channeling by competing with CPT1A for cytosolicLC-CoAs [30,31], our results clearly demonstrated that controlof CPT1A activity by malonyl-CoA is an essential drivingforce for exogenous LCFA β-oxidation flux. Additionally, theystrengthened our previous findings that changes in the degreeof CPT1A malonyl-CoA sensitivity, which occur in vivo indifferent physiopathological situations [11–15], play a criticalrole in the regulation of hepatic LCFA oxidation flux [19,32].Therefore this observation allowed us to predict that strategiesaiming to increase only CPT1A activity might be limited insituations where malonyl-CoA levels remain high. This mayexplain why a compensatory increase in CPT1A gene expressionmight be insufficient to prevent hepatic TAG accumulation whenlipogenesis is exacerbated, such as in mice overexpressing acyl-CoA:diacylglycerol acyltransferase 2 in the liver [33].

The key finding in the present study is that, in the absence ofexogenous LCFAs, CPT1mt expression still markedly protectsliver cells from glucose/insulin-induced TAG accumulation. Inthese culture conditions, TAG can only be produced by esteri-fication of LCFAs de novo synthesized from glucose throughlipogenesis. By deciphering the metabolic impact of CPT1mt

expression, we clearly showed that the decreased TAG content dir-ectly resulted from β-oxidation of the newly synthesized LCFAs,as reflected by the increased β-hydroxybutyrate production. Thusan increased CPT1A activity in association with a decreasedmalonyl-CoA sensitivity allowed hepatocytes to oxidize LCFAswhatever their origin, i.e. exogenous or endogenous (Figure 7).According to Randle’s glucose/fatty acid cycle [34], one wouldpredict that increased LCFA β-oxidation may have an effect onthe metabolic fate of glucose. However, previous studies havereported that CPT1wt and CPT1mt overexpression did not affectglucose metabolism in pancreatic β-cells [27] and in muscularcells [28]. Similarly, in the present study, an increased LCFAβ-oxidation flux in hepatocytes did not alter the insulin-inducedglucose oxidation and glycogen synthesis. Additionally, oxida-tion of both glucose and LCFAs is associated with conversionof oxidized cofactors into reduced cofactors which are thenre-oxidized by the mitochondrial respiratory chain in order topermit other cycles of fuel oxidation [35]. The lactate/pyruvateand β-hydroxybutyrate/acetoacetate ratio respectively reflect thecytosolic and mitochondrial NADH/NAD+ ratio. In the presenceof glucose plus insulin, CPT1wt and CPT1mt overexpression didnot modify the lactate/pyruvate ratio, suggesting no alterationin the cytosolic redox state. The β-hydroxybutyrate/acetoacetateratio increased upon CPT1wt and CPT1mt overexpression,but did not exceed the one measured in basal conditions of

c© The Authors Journal compilation c© 2009 Biochemical Society

Modulation of the liver malonyl-CoA–CPT1 partnership 437

Figure 7 Importance of the malonyl-CoA–CPT1A partnership in the liver

In physiological situations, inhibition of CPT1A by malonyl-CoA preserves liver cells from wasting energy by preventing β-oxidation of LCFAs de novo synthesized from glucose. Indeed, retrievingCPT1A from malonyl-CoA inhibition (expression of a malonyl-CoA-insensitive CPT1A, i.e. CPT1mt) induces a metabolic switch towards β-oxidation of LCFAs independently of their origin(exogenous or de novo synthesized), and prevents TAG (TG) accumulation in liver cells. Therefore the malonyl-CoA–CPT1A partnership greatly contributes to the physiological function of the liverto convert excess dietary carbohydrates into TAGs.

culture (results not shown). Moreover, the mitochondrial activityreflected by an MTS assay was not modified by CPT1wt andCPT1mt overexpression (Supplementary Figure S3 at http://www.BiochemJ.org/bj/420/bj4200429add.htm). This suggests thatmitochondria, in our experimental conditions, are able to handleefficiently the re-oxidation of NADH when both glucose andLCFAs are oxidized. As illustrated in the present study, thefundamental role of the hepatic malonyl-CoA–CPT1A partner-ship is to provide a powerful regulatory mechanism for themetabolic switch between carbohydrate and lipid utilization as anenergy fuel. In the present study, we report for the first time that themalonyl-CoA–CPT1A interaction is also essential to preservethe cells from wasting energy by diverting the newly de novosynthesized LCFAs from mitochondrial oxidation (Figure 7). Thismay greatly contribute to the physiological function of the liverto convert excess dietary carbohydrates into TAGs.

Because hepatic steatosis is a risk factor for non-alcoholicsteatohepatitis and Type 2 diabetes, research efforts have focusedon a decreased capacity for de novo lipogenesis in an attemptto prevent and/or reduce hepatic steatosis. Among them, liver-specific suppression of rat ACC isoform expression [36],inhibition of the nuclear factor carbohydrate-responsive element-binding protein gene expression [37], which decreased ACC andFAS gene expression, and hepatic MCD overexpression [38]were effective in decreasing TAG content in fed animals and/orhepatic steatosis and insulin resistance in obese animals. Owingto the existence of the malonyl-CoA–CPT1A partnership,these strategies also stimulated LCFA β-oxidation flux throughthe release of CPT1A from malonyl-CoA inhibition [36–38].Direct evidence that an enhanced capacity to oxidize LCFAs,independently of a reduced lipogenesis flux, may contribute todecrease liver TAG content was recently provided by in vivooverexpression of CPT1wt in rat liver [9]. However, in high-fat-diet-induced obese rats, hepatic CPT1A overexpression ledonly to a slight increase in LCFA β-oxidation capacity and asubstantial reduction in hepatic TAG accumulation, whereas noimprovement of insulin sensitivity could be observed [9]. Suchmoderate effects might be due to the presence of an elevatedmalonyl-CoA concentration which, despite an increased CPT1A

expression, can limit in vivo CPT1A activity since the latterremains malonyl-CoA sensitive. Therefore we propose that, inaddition to an increase in CPT1A expression, acting directlyon the degree of CPT1A malonyl-CoA sensitivity might be amore relevant strategy in situations of insulin resistance whichare associated with an increase in LCFA delivery to the liver(increased consumption of a high-fat diet and/or lipolysis) and inendogenous LCFAs (excessive lipogenesis).

In summary, a direct modulation of the malonyl-CoA–CPT1Ainteraction in liver cells, independently of the lipogenic pathway,allows mitochondrial oxidation of both exogenous LCFAs takenup by the liver and endogenous LCFAs generated by lipogenesis,without alteration in glucose utilization. Thus, whereas CPT1Ainhibition by malonyl-CoA is essential to avoid, in physiologicalsituations, a waste of energy, modulation of its malonyl-CoAsensitivity may be a useful therapeutic strategy to prevent and/orreduce liver steatosis.

AUTHOR CONTRIBUTION

Marie Akkaqui, Jean Gira and Carina Prip-Buus designed the research; Marie Akkaqui,Isabelle Cohen, Catherine Esnous, Veronique Lenoir, Martin Sournac and Carina Prip-Buusperformed the research; Isabelle Cohen contributed new analytic tools; Marie Akkaqui,Isabelle Cohen and Carina Prip-Buus analysed results and wrote the paper.

ACKNOWLEDGEMENTS

We thank France Demaugre (INSERM U785, Villejuif, France) and Abdelhak Mansouri forcritical review of the manuscript prior to submission, and Sylvie Demignot (CNRS UMR505, Paris, France) for helpful suggestions for TAG and VLDL measurements. We thankthe Vector Core of the University Hospital of Nantes supported by the AFM (AssociationFrancaise contre les Myopathies) for providing the adenovirus vectors.

FUNDING

This work was supported, in part, by the Ministere de la Recherche “ACI Biologiedu Developpement et Physiologie Integrative” [grant number 0220527], ALFEDIAM-AstraZeneca, INSERM “Programme National de Recherches sur le Diabete” [grant numberASE04179KSA] and the Agence Nationale de la Recherche “Cardiovasculaire, Obesite,Diabete” [grant number APV05051KSA]. M.A. is a recipient of a MENRT (Ministere del’Education Nationale, de la Recherche et de la Technologie) fellowship.

c© The Authors Journal compilation c© 2009 Biochemical Society

438 M. Akkaoui and others

REFERENCES

1 McGarry, J. D. and Brown, N. F. (1997) The mitochondrial carnitine palmitoyltransferasesystem. From concept to molecular analysis. Eur. J. Biochem. 244, 1–14

2 Zammit, V. A. (2008) Carnitine palmitoyltransferase 1: central to cell function. IUBMB Life60, 347–354

3 McGarry, J. D., Mannaerts, G. P. and Foster, D. W. (1977) A possible role for malonyl-CoAin the regulation of hepatic fatty acid oxidation and ketogenesis. J. Clin. Invest. 60,265–270

4 Bougneres, P. F., Saudubray, J. M., Marsac, C., Bernard, O., Odievre, M. and Girard, J. R.(1980) Decreased ketogenesis due to deficiency of hepatic carnitine acyl transferase. N.Engl. J. Med. 302, 123–124

5 Bonnefont, J. P., Djouadi, F., Prip-Buus, C., Gobin, S., Munnich, A. and Bastin, J. (2004)Carnitine palmitoyltransferases 1 and 2: biochemical, molecular and medical aspects.Mol. Aspects Med. 25, 495–520

6 Fromenty, B. and Pessayre, D. (1995) Inhibition of mitochondrial beta-oxidation as amechanism of hepatotoxicity. Pharmacol. Ther. 67, 101–154

7 Grefhorst, A., Hoekstra, J., Derks, T. G., Ouwens, D. M., Baller, J. F., Havinga, R., Havekes,L. M., Romijn, J. A. and Kuipers, F. (2005) Acute hepatic steatosis in mice by blockingβ-oxidation does not reduce insulin sensitivity of very-low-density lipoproteinproduction. Am. J. Physiol. Gastrointest. Liver Physiol. 289, G592–G598

8 Vickers, A. E., Bentley, P. and Fisher, R. L. (2006) Consequences of mitochondrial injuryinduced by pharmaceutical fatty acid oxidation inhibitors is characterized in human andrat liver slices. Toxicol. In Vitro 20, 1173–1182

9 Stefanovic-Racic, M., Perdomo, G., Mantell, B. S., Sipula, I. J., Brown, N. F. andO’Doherty, R. M. (2008) A moderate increase in carnitine palmitoyltransferase 1a activityis sufficient to substantially reduce hepatic triglyceride levels. Am. J. Physiol.Endocrinol. Metab. 294, E969–E977

10 Prip-Buus, C., Perdereau, D., Foufelle, F., Maury, J., Ferre, P. and Girard, J. (1995)Induction of fatty-acid-synthase gene expression by glucose in primary culture of rathepatocytes. Dependency upon glucokinase activity. Eur. J. Biochem. 230, 309–315

11 Ontko, J. A. and Johns, M. L. (1980) Evaluation of malonyl-CoA in the regulation oflong-chain fatty acid oxidation in the liver. Evidence for an unidentified regulatorycomponent of the system. Biochem. J. 192, 959–962

12 Bremer, J. (1981) The effect of fasting on the activity of liver carnitine palmitoyltransferaseand its inhibition by malonyl-CoA. Biochim. Biophys. Acta 665, 628–631

13 Grantham, B. D. and Zammit, V. A. (1986) Restoration of the properties of carnitinepalmitoyltransferase I in liver mitochondria during re-feeding of starved rats. Biochem. J.239, 485–488

14 Grantham, B. D. and Zammit, V. A. (1988) Role of carnitine palmitoyltransferase I in theregulation of hepatic ketogenesis during the onset and reversal of chronic diabetes.Biochem. J. 249, 409–414

15 Faye, A., Borthwick, K., Esnous, C., Price, N. T., Gobin, S., Jackson, V. N., Zammit, V. A.,Girard, J. and Prip-Buus, C. (2005) Demonstration of N- and C-terminal domainintramolecular interactions in rat liver carnitine palmitoyltransferase 1 that determine itsdegree of malonyl-CoA sensitivity. Biochem. J. 387, 67–76

16 Prip-Buus, C., Cohen, I., Kohl, C., Esser, V., McGarry, J. D. and Girard, J. (1998)Topological and functional analysis of the rat liver carnitine palmitoyltransferase 1expressed in Saccharomyces cerevisiae. FEBS Lett. 429, 173–178

17 Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature 227, 680–685

18 Lowry, O. H., Rosebrough, N. J., Lewis Farr, A. and Randall, R. J. (1951) Proteinmeasurement with the Folin phenol reagent. J. Biol. Chem. 193, 265–275

19 Prip-Buus, C., Pegorier, J. P., Duee, P. H., Kohl, C. and Girard, J. (1990) Evidence that thesensitivity of carnitine palmitoyltransferase I to inhibition by malonyl-CoA is an importantsite of regulation of hepatic fatty acid oxidation in the fetal and newborn rabbit. Perinataldevelopment and effects of pancreatic hormones in cultured rabbit hepatocytes.Biochem. J. 269, 409–415

20 Ferre, P., Pegorier, J. P., Williamson, D. H. and Girard, J. (1979) Interactions in vivobetween oxidation of non-esterified fatty acids and gluconeogenesis in the newborn rat.Biochem. J. 182, 593–598

21 Guillet-Deniau, I., Pichard, A. L., Kone, A., Esnous, C., Nieruchalski, M., Girard, J. andPrip-Buus, C. (2004) Glucose induces de novo lipogenesis in rat muscle satellite cellsthrough a sterol-regulatory-element-binding-protein-1c-dependent pathway. J. Cell Sci.117, 1937–1944

22 Fujioka, T., Tsujita, Y. and Shimotsu, H. (1997) Induction of fatty acid synthesis bypravastatin sodium in rat liver and primary hepatocytes. Eur. J. Pharmacol. 328, 235–239

23 Halestrap, A. P. and Denton, R. M. (1973) Insulin and the regulation of adipose tissueacetyl-coenzyme A carboxylase. Biochem. J. 132, 509–517

24 Antinozzi, P. A., Segall, L., Prentki, M., McGarry, J. D. and Newgard, C. B. (1998)Molecular or pharmacologic perturbation of the link between glucose and lipidmetabolism is without effect on glucose-stimulated insulin secretion. A re-evaluation ofthe long-chain acyl-CoA hypothesis. J. Biol. Chem. 273, 16146–16154

25 Panek, E., Cook, G. A. and Cornell, N. W. (1977) Inhibition by 5-(tetradecyloxy)-2-furoicacid of fatty acid and cholesterol synthesis in isolated rat hepatocytes. Lipids 12, 814–818

26 Morillas, M., Gomez-Puertas, P., Bentebibel, A., Selles, E., Casals, N., Valencia, A.,Hegardt, F. G., Asins, G. and Serra, D. (2003) Identification of conserved amino acidresidues in rat liver carnitine palmitoyltransferase I critical for malonyl-CoA inhibition.Mutation of methionine 593 abolishes malonyl-CoA inhibition. J. Biol. Chem. 278,9058–9063

27 Herrero, L., Rubi, B., Sebastian, D., Serra, D., Asins, G., Maechler, P., Prentki, M. andHegardt, F. G. (2005) Alteration of the malonyl-CoA/carnitine palmitoyltransferase Iinteraction in the beta-cell impairs glucose-induced insulin secretion. Diabetes 54,462–471

28 Sebastian, D., Herrero, L., Serra, D., Asins, G. and Hegardt, F. G. (2007) CPT Ioverexpression protects L6E9 muscle cells from fatty acid-induced insulin resistance.Am. J. Physiol. Endocrinol. Metab. 292, E677–E686

29 Perdomo, G., Commerford, S. R., Richard, A. M., Adams, S. H., Corkey, B. E., O’Doherty,R. M. and Brown, N. F. (2004) Increased β-oxidation in muscle cells enhancesinsulin-stimulated glucose metabolism and protects against fatty acid-induced insulinresistance despite intramyocellular lipid accumulation. J. Biol. Chem. 279,27177–27186

30 Linden, D., William-Olsson, L., Ahnmark, A., Ekroos, K., Hallberg, C., Sjogren, H. P.,Becker, B., Svensson, L., Clapham, J. C., Oscarsson, J. and Schreyer, S. (2006)Liver-directed overexpression of mitochondrial glycerol-3-phosphate acyltransferaseresults in hepatic steatosis, increased triacylglycerol secretion and reduced fatty acidoxidation. FASEB J. 20, 434–443

31 Linden, D., William-Olsson, L., Rhedin, M., Asztely, A. K., Clapham, J. C. and Schreyer,S. (2004) Overexpression of mitochondrial GPAT in rat hepatocytes leads to decreasedfatty acid oxidation and increased glycerolipid biosynthesis. J. Lipid Res. 45, 1279–1288

32 Prip-Buus, C., Bouthillier-Voisin, A. C., Kohl, C., Demaugre, F., Girard, J. and Pegorier,J. P. (1992) Evidence for an impaired long-chain fatty acid oxidation and ketogenesis inFao hepatoma cells. Eur. J. Biochem. 209, 291–298

33 Monetti, M., Levin, M. C., Watt, M. J., Sajan, M. P., Marmor, S., Hubbard, B. K., Stevens,R. D., Bain, J. R., Newgard, C. B., Farese, Sr, R. V. et al. (2007) Dissociation of hepaticsteatosis and insulin resistance in mice overexpressing DGAT in the liver. Cell Metab. 6,69–78

34 Randle, P. J. (1998) Regulatory interactions between lipids and carbohydrates: theglucose fatty acid cycle after 35 years. Diabetes Metab. Rev. 14, 263–283

35 Fromenty, B., Robin, M. A., Igoudjil, A., Mansouri, A. and Pessayre, D. (2004) The insand outs of mitochondrial dysfunction in NASH. Diabetes Metab. 30, 121–138

36 Savage, D. B., Choi, C. S., Samuel, V. T., Liu, Z. X., Zhang, D., Wang, A., Zhang, X. M.,Cline, G. W., Yu, X. X., Geisler, J. G. et al. (2006) Reversal of diet-induced hepaticsteatosis and hepatic insulin resistance by antisense oligonucleotide inhibitors ofacetyl-CoA carboxylases 1 and 2. J. Clin. Invest. 116, 817–824

37 Dentin, R., Benhamed, F., Hainault, I., Fauveau, V., Foufelle, F., Dyck, J. R., Girard, J.and Postic, C. (2006) Liver-specific inhibition of ChREBP improves hepatic steatosis andinsulin resistance in ob/ob mice. Diabetes 55, 2159–2170

38 An, J., Muoio, D. M., Shiota, M., Fujimoto, Y., Cline, G. W., Shulman, G. I., Koves, T. R.,Stevens, R., Millington, D. and Newgard, C. B. (2004) Hepatic expression of malonyl-CoAdecarboxylase reverses muscle, liver and whole-animal insulin resistance. Nat. Med. 10,268–274

Received 24 September 2008/2 March 2009; accepted 20 March 2009Published as BJ Immediate Publication 20 March 2009, doi:10.1042/BJ20081932

c© The Authors Journal compilation c© 2009 Biochemical Society

Biochem. J. (2009) 420, 429–438 (Printed in Great Britain) doi:10.1042/BJ20081932

SUPPLEMENTARY ONLINE DATAModulation of the hepatic malonyl-CoA–carnitine palmitoyltransferase1A partnership creates a metabolic switch allowing oxidation of de novofatty acids1

Marie AKKAOUI*†, Isabelle COHEN*†, Catherine ESNOUS*†, Veronique LENOIR*†, Martin SOURNAC*†, Jean GIRARD*† andCarina PRIP-BUUS*†2

*Institut Cochin, Departement d’Endocrinologie, Metabolisme et Cancer, Universite Paris Descartes, CNRS Unite Mixte de Recherche 8104, and †INSERM U567, 24 rue du FaubourgSaint-Jacques, 75014 Paris, France

Figure S1 CPT1wt and CPT1mt overexpression increases [1-14C]oleate oxidation in hepatocytes cultured in basal conditions

Hepatocytes infected with 5 infectious particles/cell of the indicated adenovirus were cultured for 40 h in the presence of 5 mM glucose. During the last 2 h of culture, 0.3 mM [1-14C]oleate boundto 1 % defatted BSA was added. Oleate oxidation into CO2 (a) and ASP (b), and oleate esterification into cellular TAGs (TG; c) and PLs (d) were measured at the end of the 2 h incubation period.(e) Metabolic orientation of oleate as a percentage of the total nmol of metabolized [1-14C]oleate. (f) [1-14C]Oleate incorporation into VLDLs secreted in the culture medium during the 2 h incubationperiod. Results are means +− S.E.M. of duplicate flasks from four independent experiments. *P < 0.05 and **P < 0.01 compared with Ad-βgal.

1 This manuscript is dedicated to Professor Denis J. McGarry.2 To whom correspondence should be addressed (email carina.prip@inserm.fr).

c© The Authors Journal compilation c© 2009 Biochemical Society

M. Akkaoui and others

Figure S2 Effect of glucose plus insulin on the expression and activity of lipogenic enzymes in hepatocytes overexpressing CPT1wt and CPT1mt

At 16 h after infection with 5 infectious particles/cell of the indicated adenovirus, hepatocytes were cultured for 24 h in the presence of either 5 mM glucose (G5) or 20 mM glucose plus 10 nMinsulin (G20 + ins). Immunoblot analysis of cellular extracts (30 μg of protein) was performed using specific antibodies for CPT1A, CPT2, ACC and FAS (a). Western blots are representative of fourindependent experiments performed in duplicate. Enzymatic activity of ACC (b), FAS (c) and MCD (d). Results are means +− S.E.M. of three to four independent experiments. §P < 0.05 comparedwith G5.

Figure S3 CPT1wt and CPT1mt overexpression does not alter mitochondrialactivity

At 16 h after infection with 5 infectious particles/cell of the indicated adenovirus, hepatocyteswere cultured for 24 h in the presence of either 5 mM glucose (G5) or 20 mM glucose plus10 nM insulin (G20 + ins). Cell culture medium was then replaced by fresh medium containingMTS reagent. After 1 h incubation at 37◦C, the amount of MTS formazan product releasedinto the medium was measured at 490 nm absorbance. Data were normalized to the averagevalue from the control in the basal condition. Results are means +− S.E.M. of duplicate 12-wellplates from four independent experiments. §P < 0.01 compared with G5.

Received 24 September 2008/2 March 2009; accepted 20 March 2009Published as BJ Immediate Publication 20 March 2009, doi:10.1042/BJ20081932

c© The Authors Journal compilation c© 2009 Biochemical Society