MCD - Wiley Online Library

Eur. J. Biochem. 271, 2831–2840 (2004) FEBS 2004

doi:10.1111/j.1432-1033.2004.04218.x

Malonyl-CoA decarboxylase (MCD) is differentially regulated in subcellular compartments by 5¢AMP-activated protein kinase (AMPK) Studies using H9c2 cells overexpressing MCD and AMPK by adenoviral gene transfer technique Nandakumar Sambandam, Michael Steinmetz, Angel Chu, Judith Y. Altarejos, Jason R. B. Dyck and Gary D. Lopaschuk Department of Pediatrics, Faculty of Medicine & Dentistry, University of Alberta, Edmonton, Canada

Malonyl-CoA, a potent inhibitor of carnitine pamitoyl transferase-I (CPT-I), plays a pivotal role in fuel selection in cardiac muscle. Malonyl-CoA decarboxylase (MCD) catalyzes the degradation of malonyl-CoA, removes a potent allosteric inhibition on CPT-I and thereby increases fatty acid oxidation in the heart. Although MCD has several Ser/ Thr phosphorylation sites, whether it is regulated by AMPactivated protein kinase (AMPK) has been controversial. We therefore overexpressed MCD (Ad.MCD) and constitutively active AMPK (Ad.CA-AMPK) in H9c2 cells, using an adenoviral gene delivery approach in order to examine if MCD is regulated by AMPK. Cells infected with Ad.CAAMPK demonstrated a fourfold increase in AMPK activity as compared with control cells expressing green fluorescent protein (Ad.GFP). MCD activity increased 40- to 50-fold in Ad.MCD + Ad.GFP cells when compared with Ad.GFP control. Co-expressing AMPK with MCD further augmented MCD expression and activity in Ad.MCD + Ad.CA-AMPK cells compared with the Ad.MCD + Ad.GFP control. Subcellular fractionation

further revealed that 54.7 kDa isoform of MCD expression was significantly higher in cytosolic fractions of Ad.MCD + Ad.CA-AMPK cells than of the Ad.MCD + Ad.GFP control. However, the MCD activities in cytosolic fractions were not different between the two groups. Interestingly, in the mitochondrial fractions, MCD activity significantly increased in Ad.MCD + Ad.CA-AMPK cells when compared with Ad.MCD + Ad.GFP cells. Using phosphoserine and phosphothreonine antibodies, no phosphorylation of MCD by AMPK was observed. The increase in MCD activity in mitochondria-rich fractions of Ad.MCD + Ad.CA-AMPK cells was accompanied by an increase in the level of the 50.7 kDa isoform of MCD protein in the mitochondria. This differential regulation of MCD expression and activity in the mitochondria by AMPK may potentially regulate malonyl-CoA levels at sites nearby CPT-I on the mitochondria.

Malonyl-CoA is a potent inhibitor of carnitine palmitoyl transferase-I (CPT-I), thereby playing a pivotal role in fuel selection in cardiac muscle [1]. CPT-I, localized on the outer mitochondrial membrane, is the rate-limiting enzyme of fatty acid transport into mitochondria for b-oxidation [2–4]. As b-oxidation of fatty acids contributes the majority of energy produced by the normal aerobic heart [5,6], malonylCoA has a key role in regulating cardiac energy metabolism. Tissue levels of malonyl-CoA are determined by its rate of

synthesis by acetyl-CoA carboxylase (ACC) and by its rate of degradation by malonyl-CoA decarboxylase (MCD) [1]. Various physiological and pathological conditions result in rapid changes in malonyl-CoA levels [7–9]. For instance, malonyl-CoA levels drop rapidly and dramatically during ischemia and reperfusion, which is associated with a significant increase in fatty acid oxidation [8]. Similarly, rapid maturation of fatty acid oxidation in the developing heart is associated with a significant decrease in malonylCoA levels in the myocardium [7]. While decreased synthesis of malonyl-CoA by ACC is partly responsible for these changes in malonyl-CoA, a simultaneous degradation by MCD also has an important role in lowering malonyl-CoA levels [10]. MCD was originally identified in the uropygial gland of the goose [11]. We also showed MCD to be highly expressed in mammalian cardiac muscle [12], and provided evidence to suggest that cardiac MCD plays an important role in regulating fatty acid metabolism in the heart [10,13]. Regulation of MCD occurs both at the level of transcription and post-translation [14,15]. MCD has several serine and threonine residues that can potentially be phosphorylated. Previous studies in our lab and other groups have shown

Correspondence to G. Lopaschuk, 423 Heritage Medical Research Building, University of Alberta, Edmonton, Alberta T6G 2S2, Canada. Fax: + 1 780 492 9753, Tel.: + 1 780 4922170, E-mail: [email protected] Abbreviations: ACC, acetyl-CoA carboxylase; AICAR, 5-aminoimidazole-4-carboxamide riboside; AMPK, 5¢AMP-activated protein kinase; CPT-I, carnitine pamitoyl transferase-I; GFP, green fluorescent protein; Itu, 5¢-iodotubercidin; MCD, malonyl-CoA decarboxylase; moi, multiplicity of infection. (Received 26 February 2004, revised 14 April 2004, accepted 14 May 2004)

Keywords: malonyl-CoA decarboxylase; AMPK; cardiac cells.

FEBS 2004

2832 N. Sambandam et al. (Eur. J. Biochem. 271)

that MCD can either be inhibited or activated by phosphorylation [16,17]. One potential kinase that could control MCD activity is 5¢AMP-activated protein kinase (AMPK). AMPK is a cellular fuel gauge, and acts to simultaneously shut down ATP consuming biosynthetic processes and facilitate ATP producing catabolic processes during periods of metabolic stress [18]. One important stress that can occur in the heart is ischemia. AMPK is rapidly activated during myocardial ischemia [8,19,20], leading to rapid changes in the control of glucose and fatty acid metabolism. AMPK stimulation of fatty acid metabolism occurs as a result of AMPK phosphorylation and inhibition of ACC [18,20–24]. This activation of AMPK and inhibition of ACC results in a dramatic drop in malonyl-CoA levels during and following ischemia [8,20]. Alterations in myocardial malonyl-CoA levels can not be solely explained by suppression of ACC activity unless simultaneous degradation of malonyl-CoA is occurring. It has therefore been hypothesized that AMPK could also play a dual role by activating MCD to facilitate malonylCoA degradation [12]. However, the existing literature on MCD regulation by AMPK is inconsistent in this regard. Although we [12] and others [25] have demonstrated that MCD is not a direct substrate for AMPK in vitro, other studies suggest that MCD is activated by phosphorylation by AMPK [16,17]. The inconsistencies in the literature regarding AMPK’s role on MCD regulation may be partly due to the fact that the above studies have either used nonspecific means to activate AMPK [16,17] or have used in vitro conditions that do not mimic conditions seen in the intact cell [25]. Two alternate translational start sites on MCD appear to give rise to two isoforms of molecular weight 54.7 kDa and 50.7 kDa, respectively [11,13,26]. MCD could potentially exist in different subcellular compartments, including cytoplasm, peroxisome or mitochondria [27]. In cardiac myocytes, the majority of the MCD is the 50.7 kDa isoform, which is primarily expressed in the mitochondria [1,28]. How compartmentalization regulates cardiac MCD activity is not clearly understood. In the present study we examined whether cardiac MCD is regulated by AMPK, by cooverexpressing a constitutively active mutated form of the catalytic subunit of AMPK and the full length human MCD in H9c2 cells (a rat cardiac ventricular cell line) using an adenoviral gene delivery technique. As MCD is localized in various subcellular compartments, we also examined whether AMPK differentially regulates MCD in mitochondrial and cytosolic fractions of these cardiac cells.

Materials and methods H9c2 cell culture H9c2 cells (ATCC, Rockville, MD, USA) were grown as myoblasts to confluency in 60-mm diameter cell culture dishes in Dulbecco’s modified Eagles’ medium (DMEM; Sigma) containing 10% (v/v) fetal bovine serum, 1% (w/v) PenStrep (Sigma) and 0.25 mM L-carnitine (Sigma). Dishes were incubated in a water-jacketed CO2 incubator maintained at 37 C with 95% O2 and 5% CO2 (v/v/v). Cells were replenished with fresh media every 48 h. Cells were seeded at approximately 4000–5000 cells per cm2. On

reaching approximately 90% confluency, myoblasts were allowed to differentiate into myotubes in DMEM containing 1% (v/v) fetal bovine serum, 1% (w/v) penstrep, and 0.25 mM L-carnitine. In the presence of 0.25 mM L-carnitine, full differentiation of myoblast to myotubes occurred within 7 days of adding 1% (v/v) fetal bovine serum, using peak levels of myo-d expression as a marker of muscle cell differentiation (data not shown). Passages 12–25 were used for experiments described in this study. AICAR treatment H9c2 cells were treated with 2 mM 5-aminoimidazole4-carboxamide riboside (AICAR) for 2 h, as described previously [29]. Briefly, DMEM containing 1% (v/v) fetal bovine serum was removed and cells were incubated with Krebs’ Henseleit (KH) solution (118 mM NaCl, 3.5 mM KCl, 1.3 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4) for 20 min at 37 C. At the end of 20 min, fresh KH solution with or without AICAR (2.0 mM final concentration) was added to each dish, and cells were incubated for 2 h. Some cells were also treated with the AMPK antagonist 5¢-iodotubercidin (Itu, 50 lM) for 2 h, either with or without 2.0 mM AICAR. Four groups were included: (a) control, (b) AICAR treated, (c) Itu treated, and (d) Itu + AICAR treated cells. At the end of the 2-h incubation, cells were rapidly lysed as described previously [29]. Cell lysates were then used for measurement of AMPK activities. Construction of recombinant adenovirus encoding MCD, AMPK, and GFP and infection of H9c2 cells To construct recombinant adenovirus, full length human MCD cDNA containing the two putative start sites [30] was subcloned into a pAdTrack-CMV shuttle vector, linearized with Pme 1 and inserted into adenovirus using pAdEasy-1 system for homologous recombination in Escherichia coli [31]. The full-length hMCD with two start sites can express two isoforms of MCD (a 50 kDa and 54.7 kDa isoforms). The longer form has a putative mitochondrial targeting sequence, as well as peroxisomal targeting sequence [32]. The pAdTrack-CMV shuttle vector also contained a gene encoding enhanced green fluorescent protein (GFP). Therefore, the adenovirus used to express MCD protein also expressed GFP, which served as a marker of successful viral infection and protein overexpression. A similar protocol was used to construct adenoviruses encoding a myc-tagged constitutively active (T172D) catalytic l1 subunit (1–312 amino acid residues) of AMPK (CA-AMPKa1(312)) [33], as well as an adenovirus encoding GFP alone (used as a control). Differentiated H9c2 cells cultured in DMEM with 1% (v/v) fetal bovine serum were infected with either five multiplicity of infection (moi) per cell of Ad.MCD, 25 moi per cell of Ad.GFP or 25 moi per cell of Ad.CA-AMPK. Ad.CA-AMPK (25 moiÆcell)1) were determined to yield optimum CA-AMPKa1(312) expression and activity from series of Ad.CA-AMPK concentrations (5, 10, 25 and 50 moiÆcell)1). Some cells were double infected with Ad.MCD (5 moi) and Ad.CA-AMPK (25 moi) to study the effect of overexpressed AMPK on overexpressed MCD activity (Ad.MCD + Ad.CA-AMPK). Cells infected with

FEBS 2004

Cardiac malonyl-CoA decarboxylase (Eur. J. Biochem. 271) 2833

Ad.MCD (5 moi) and Ad.GFP (25 moi) served as the control (Ad.MCD + Ad.GFP) to the above group. Cells were allowed to express the proteins for 48 h and lysed rapidly as described below. Cell lysis and sample preparation for MCD and AMPK assays Cells were subjected to a rapid lysis procedure to avoid activation of endogenous AMPK, as slow lysis of cells has been shown previously to increase cellular AMP levels [29]. Culture dishes were placed on ice, ice-cold lysis buffer was added, cells were scraped carefully with a rubber scraper and transferred to microfuge tubes. Samples were then immediately homogenized by ultrasonication (Sonifier, Model W185D, Heat Systems-Ultrasonics, Inc., NY, USA) and centrifuged at 17 000 g for 3 min [29]. Supernatants were subsequently collected and stored at )80 C. For AMPK and ACC assays, cells were lysed in buffer containing 50 mM Tris-base, 250 mM mannitol, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5.0 mM NaPPi, 1 mM dithiothreitol, mammalian protease inhibitor cocktail (Sigma) and 1% (v/v) Triton X-100. For MCD assays, lysis buffer containing 75 mM KCl, 20 mM sucrose, 10 mM Hepes, 1 mM EGTA, 50 mM NaF, 5 mM NaPPi, 1 mM dithiothreitol, and a protease inhibitor cocktail was used. Samples were subjected to ultrasonication on ice for 5 s and whole cell lysates were used for MCD assay. Protein concentrations of the cell lysates were determined by a Bradford protein assay kit. Subcelluar fractionation to isolate cytosol and mitochondrial fractions To prepare mitochondrial and cytoplasmic fractions, three 60 mm dishes were pooled. Cytoplasmic fractions were obtained by permeabilization of plasma membrane by digitonin (30 lM) treatment for 20 min at 37 C [34]. Each 60 mm dish was treated with buffer containing 30 lM digitonin, 0.15 mM MgCl2, 10 mM KCl, 10 mM Tris/HCl, pH 6.7). Following incubation the buffer was removed, and centrifuged at 1500 g for 5 min. Supernatant was concentrated using Amicon Ultrafree-MCTM ultrafiltration (30 kDa molecular mass cut-off) units, centrifuged at 5500 g for 1 h in 4 C. Mitochondrial fraction was prepared from the above digitonin permeabilized cells, as described previously [35]. Cells were quickly washed with ice-cold NaCl/Pi and scraped into ice cold NaCl/Pi in 15 mL centrifuge tubes. Cells were pelleted by centrifuging at 1000 g for 10 min. The pellet was then re-suspended in approximately six volumes of homogenizing buffer (0.15 mM MgCl2, 10 mM KCl, 10 mM Tris/HCl, pH 6.7), transferred to a glass-Teflon homogenizer (Potter-Elvehjem, between 0.10 and 0.15 mm clearance), and homogenized by 10–15 up and down strokes while revolving at 500 r.p.m. Homogenate was then transferred to a microfuge tube, and sucrose was added to the homogenate to a final concentration of 0.25 M and dissolved. The homogenate was centrifuged at 1500 g for 3 min to remove nuclei and larger fragments. The supernatant was then centrifuged at 5000 g for 10 min to pellet mitochondria. The pellet was resuspended in 10 mM

Tris-acetate (pH 6.7) buffer containing 0.15 mM MgCl2, 250 mM sucrose and re-centrifuged at 5000 g for 10 min. The pellet was then suspended in 10 mM Tris-acetate (pH 7.0) buffer containing 250 mM sucrose. This procedure is known to yield a mitochondrial-rich fraction of high purity and functional integrity [36]. Voltage dependent anion-selective channel protein 1 (VDAC-1), a mitochondrial porin, was used as a marker to check the mitochondrial fractions [37]. Digitonin permeabilization followed by mitochondrial fractionation did not affect mitochondrial integrity as determined by negligible amounts of cytochrome C released into cytosol. Western blot and SDS/PAGE for AMPK, MCD and mitochondrial markers To identify AMPK and MCD in the samples, SDS/PAGE and Western blot analysis was peformed. Thirty micrograms of either whole cell lysates or subcellular fractions were loaded in each well of 10% SDS gel. Following electrophoresis, proteins were transferred to nitrocellulose membranes which were then blocked overnight with either 5% (w/v) bovine serum albumin (for MCD) or in 5% (w/v) skimmed milk powder (for AMPK) in NaCl/Tris. For CAAMPKa1(312) which is myc-tagged, polyclonal anti-myc (Santa Cruz Biotechnology Inc., CA, USA); and for MCD, rabbit polyclonal anti-MCD IgG [12,13] were used. Enhanced chemiluminscence detection was carried out to visualize the protein bands on an autoradiograph. Western blot analyses for VDAC1, cytochrome C oxidase and ubiquinone-cytochrome C core 2 subunit of complex III were performed using respective primary antibodies (polyclonal goat anti-VDAC1, Santa Cruz Biotechnology Inc.; monoclonal mouse anti-cytochrome c, BD Biosciences Pharmingen, San Diego, CA, USA; monoclonal mouse anti-core 2 subunit, Molecular Probes, Eugene, OR, USA). AMPK assay Both endogeous AMPK and overexpressed CA-AMPKa1(312) activities were measured as previously described [8]. Samples were diluted to a concentration of 1 mgÆmL)1 in re-suspension buffer containing 100 mM Tris-base, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5 mM NaPPi, 10% (v/v) glycerol, 1 mM dithiothreitol, 0.1% (w/v) mammalian protease inhibitor cocktail and 0.12% (v/v) Triton X-100. Two microlitres of the above sample was then incubated with the synthetic 200 lM AMARA (AMARAASAAALARRR) peptide, 200 lM [32P]ATP[c-P], 0.8 mM dithiothreitol, 5 mM MgCl2, 200 lM AMP in buffer (pH 7.0) containing 40 mM Hepes/NaOH, 80 mM NaCl, 8% (w/v) glycerol for 5 min at 30 C (total volume 25 lL). This incubation leads to incorporation of 32P into the AMARA peptide. At the end of 5 min, 15 lL of the incubation mixture was blotted onto a 1 cm2 phosphocellulose paper. The paper was then washed three times for 10 min in 150 mM phosphoric acid followed by a 5 min final wash in acetone. The papers were then dried and counted in 4 mL of scintillation fluid (EcoLiteTM, ICN, CA, USA). AMPK activity was expressed as picomoles of 32 P incorporated into AMARA peptide per minute per milligram protein.

FEBS 2004


MCD assay MCD activity was determined by radiometric assay that was slightly modified from a previously described method [10]. Acetyl-CoA, the product of malonyl-CoA degradation by MCD, was converted to [14C]citrate by incubation with [14C]oxaloacetate in the presence of citrate synthase (0.73 lUÆlL)1). [14C]Oxaloactetate in turn was produced from [U-14C]aspartate (5 lCiÆmL)1) and a-ketoglutarate (2 mM) by transamination in the presence of glutamic oxaloacetate transaminase. One hundred microliters of whole cell lysates or cytoplasmic and mitochondrial fractions of either undiluted samples for endogenous MCD in nonoverexpressing cells (2.0–3.0 mgÆmL)1 protein concentration) or 20–40 times diluted samples for cells overexpressing MCD were incubated with 90 lL incubation buffer containing phosphatase inhibitors 50 mM NaF, 5 mM NaPPi, 1 mM dithiothreitol and 100 mM Tris-base (pH 8.0). The timed reaction was started by adding 1.0 mM malonyl-CoA to the incubation mixture and incubated at 37 C for 20 min to allow formation of acetyl-CoA. The reaction was stopped with 40 lL of 0.5 M perchloric acid, neutralized with 10 lL of 2.2 M KHCO3 (pH 10) and centrifuged at 1500 g at 4 C for 5 min to remove precipitated proteins. Supernatants containing formed acetyl-CoA were incubated with 22 lL of a mixture of 0.01 mM dithiothreitol, 1.0 mM CuSO4, and 400 mM potassium acetate solution, 20 lL of 60 mM EDTA and 30 lL of 30 mM N-ethylmaleimide to remove excess CoA remaining in the later stages of the reaction so that the citrate present could not generate non-MCD derived acetylCoA. The unreacted [14C]oxaloacetate was converted back to aspartate by the addition of glutamic oxaloacetate transaminase (0.533 lUÆlL)1) in the presence of 6.8 mM sodium glutamate. The resulting reaction mixture was then added to 1 mL of a 1 : 1 suspension of Dowex 50 W-X8 (100–200 mesh, hydrogen form) in distilled water. Dowex binds the aspartate while leaving citrate in the supernatant. 0.5 mL of supernatant was collected after centrifuging the slurry at 1000 g for 5 min, mixed with 4 mL of scintillation fluid (EcoLiteTM, ICN, CA, USA) and counted in a liquid scintillation counter. The radioactivity was converted to nanomoles of acetyl-CoA formed in the reaction using a standard curve generated from 0 to 20 nM range of standard acetyl-CoA which underwent similar treatment as that of samples. Preliminary experiments established that 20 min incubation and the amount of samples used were in the linear range of MCD enzyme activity.

In vitro phosphorylation of MCD by AMPK using lysates of cells overexpressing either MCD or CA-AMPKa1(312) H9c2 cells overexpressing MCD were lysed with buffer containing 75 mM KCl, 20 mM Sucrose, 10 mM Hepes, 1 mM EGTA, 1 mM dithiothreitol, and a protease inhibitor cocktail on ice by ultrasonication for 5 s and whole cell lysates were used. Cells overexpressing CA-AMPKa1(312) were lysed with buffer containing 50 mM Tris-base, 250 mM mannitol, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5.0 mM NaPPi, 1 mM dithiothreitol, mammalian protease inhibitor cocktail (Sigma) and 1% (v/v) Triton X-100 by ultrasonication as mentioned above. Whole cell lysates of

cells overexpressing MCD were incubated with the lysates of cells overexpressing CA-AMPKa1(312) for 20, 30, 60, 120 and 180 min. At the end of the indicated time points samples were immunoprecipitated for MCD with rabbit polyclonal anti-MCD IgG bound to protein-A sepharose beads. Immunoprecipitates were subjected to SDS/PAGE and Western blotting and probed with antiphosphoserine or antiphosphothreonine antibodies. In some experiments cell extracts were incubated in the presence of 100 lCi of [32P]ATP[c-P] for the above-indicated duration followed by immunoprecipitation as above and autoradiographed for 1 week in )20 C. Statistical analysis Data are presented as means ± SEM. Statistically significant differences between groups of two were assessed using the paired Students t-test. A two-tailed value of P < 0.05 was considered to be significant.

Results Effect of AICAR on endogenous AMPK and MCD activities Incubation of H9c2 cells with AICAR increased AMPK activity significantly (916 ± 130 pmolÆmgÆmin)1) when compared with untreated control cells (588 ± 81 pmolÆmgÆ min)1). AICAR treatment also increased MCD activity only modestly (to 126% of untreated controls, 1.3 ± 0.3 nmolÆmin)1Æmg)1, one tailed P < 0.05). Itu, an inhibitor of AMPK, inhibited AMPK activity (381 ± 66 pmolÆmin)1Æmg)1) but did not affect MCD activity (1.5 ± 0.6 nmolÆmin)1Æmg)1). However, Itu did inhibit AICAR-stimulatable AMPK activity significantly (394 ± 46 pmolÆmin)1Æmg)1, P < 0.05), as well as the small increase in MCD activity (1.9 ± 0.2 nmolÆ min)1Æmg)1 in AICAR-treated cells vs. 1.3 ± 0.03 nmolÆ min)1Æmg)1 in Itu + AICAR-treated cells, P < 0.05). Endogenous MCD activity by overexpressed CA-AMPKa1(312) Overexpression of CA-AMPKa1(312) using recombinant adenovirus (Ad.CA-AMPK) resulted in an increase in AMPK expression and activity in a concentration-dependent manner when compared with control cells expressing GFP (Fig. 1Ai,ii). As a concentration of 25 moiÆcell)1 Ad.CA-AMPK yielded maximum activity, we used the above concentration of Ad.CA-AMPK for all further studies. Control cells were infected with an equivalent amount of Ad.GFP virus per cell. Overexpression of CA-AMPKa1(312) did not increase endogenous cytoplasmic MCD activity measured when compared with Ad.GFP cells (Fig. 1Bi). In mitochondrial rich fractions, there was trend towards an increase in MCD activity in response to CA-AMPKa1(312) overexpression, which was not statistically significant when compared with Ad.GFP cells (Fig. 1Bii, P < 0.07). As shown, the endogenous MCD activities were very low and difficult to obtain a reproducible result in subcellular fractions. Therefore, due to low endogenous MCD activities in H9c2 cell fractions, we decided to increase the expression of MCD in

FEBS 2004


108.5 ± 14.2 nmolÆmin)1Æmg)1 in Ad.MCD + Ad.CAAMPK cells, P ¼ 0.058; Fig. 2Biv). MCD expression and activity in subcellular fractions of H9c2 cells overexpressing both MCD and AMPK

Fig. 1. AMPK overexpression by adenoviral gene transfer and endogenous MCD activity. (A) AMPK expression (i) and activity (ii) in H9c2 cells infected with Ad.CA-AMPK or Ad.GFP. Control H9c2 cells had no viral infection while Ad.GFP cells had 25 moiÆcell)1 of Ad.GFP virus. As AMPK is myc tagged, anti-myc antibody was used to probe overexpressed CA-AMPKa1(312). Western blot is a representative of n ¼ 2 experiments, AMPK activity values are average of n ¼ 2 experiments. (B) Endogenous MCD activity in cytosolic (i) and mitochondrial (ii) fractions of H9c2 cells infected with Ad.CA-AMPK or Ad.GFP. Values are mean ± SE of n ¼ 5 experiments.

these cells along with CA-AMPKa1(312) to examine the role of AMPK in the regulation of MCD. MCD activity in H9c2 cells coinfected with Ad.CA-AMPK and Ad.MCD Infection of H9c2 cells with Ad.MCD resulted in a significant increase in MCD protein and activity when compared with Ad.GFP cells (Fig. 2Ai,ii). In order to study the effect of AMPK on MCD, we coinfected H9c2 cells with Ad.MCD and Ad.CA-AMPK viruses (Ad.MCD + Ad.CA-AMPK) and compared our results to control cells coinfected with an equivalent number of viral particles/cells of Ad.MCD and Ad.GFP (Ad.MCD + Ad.GFP). Our Western blot analysis show that there was a significant increase in both the 50.7 and 54.7 kDa isoform of MCD protein levels in Ad.MCD + Ad.CA-AMPK cells compared with Ad.MCD + Ad.GFP cells (Fig. 2Bi–iii). The enzyme activity of MCD showed a trend towards increase which was not statistically significant (from 80.7 ± 7.3 nmolÆmin)1Æmg)1 in Ad.MCD + Ad.GFP cells to

Previous studies have shown that MCD exists in both the cytoplasmic and mitochondrial compartments [25]. Hence, we wanted to determine if there is a differential expression and activation in various subcellular compartments. We therefore isolated cytoplasmic and mitochondrial rich fractions to determine MCD distribution and its regulation in different compartments in response to increased AMPK activity. Mitochondrial-rich fractions showed enrichment of a mitochondrial specific protein VDAC1 that was absent in cytoplasmic fractions (Fig. 3Ai). Further, most of the cytochrome C was confined to mitochondrial rich fractions and very little of cytochrome C was released into the cytoplasmic fractions (Fig. 3Aii) suggesting that digitonin permeabilization resulted in a negligible damage to mitochondria. Taken together, our data suggest that the subcellular fractions were relatively pure. Figure 3Aiii shows that overexpressed MCD was present in both cytoplasmic and mitochondrial rich fractions. While the majority of over expressed MCD activity was present in mitochondria, about 30–40% of total MCD activity was measured in cytoplasmic fractions (33 ± 18 nmolÆmin)1Æ mg)1 in cytoplasmic fractions vs. 81 ± 7 nmolÆmin)1Æmg)1 in whole cell lysates). This distribution is consistent with previously published studies [1]. Cytoplasmic MCD Figure 3Bi–iv shows the effect of CA-AMPKa1(312) overexpression on cytoplasmic MCD protein levels and activities. As observed in Western blot analysis, cytoplasmic fractions show both isoforms of MCD. In Ad.MCD + Ad.CA-AMPK cells, there was increase in MCD protein levels (both long and short isoforms; Fig. 3Bi) when compared with Ad.MCD + Ad.GFP cells. This increase was more pronounced with the long isoform in Ad.MCD + Ad.CA-AMPK cells (optical density 1.29 ± 0.11 AU vs. 0.16 ± 0.02 AU in Ad.MCD + Ad.GFP cells, P < 0.0001; Fig. 3Bi,ii). Interestingly, the increase in short isoform of MCD protein was not statistically significant. Despite the increased expression of MCD protein, MCD activity in cytoplasmic fraction (normalized to milligrams of total protein) was not different between two groups (Fig. 3iv). Mitochondrial MCD MCD activity was augmented in mitochondrial fractions obtained in response to co-overexpression of AMPK in Ad.MCD + Ad.CA-AMPK cells compared with Ad.MCD + Ad.GFP cells (Fig. 4Aiii). Unlike the cytoplasmic fractions, almost all of the MCD was the shorter form (50.7 kDa). The increase in MCD activity in the mitochondrial rich fractions was accompanied by a corresponding increase in MCD protein levels (Fig. 4Ai,ii,iii). Figure 4Bi,ii shows that levels of other mitochondrial-


FEBS 2004

Fig. 2. MCD overexpression and activity in H9c2 cells co-expressing Ad.CA-AMPK or Ad.GFP. (A) MCD expression (i) and activity (ii) in H9c2 cells infected with Ad.MCD or Ad.GFP. In the Western blot, lanes 1 and 2 ¼ Ad.GFP and lanes 3 and 4 ¼ Ad.MCD. Western blot is a representative of n ¼ 3 experiments. Activity values are means ± SE of n ¼ 5 experiments. *Significantly different from Ad.GFP control, P < 0.05. (B) MCD expression (i), optical density of 54.7 kDa isoform (ii), optical density of 50.7 kDa isoform (iii) and activity (iv) in whole cell lysates of H9c2 cells coinfected with Ad.MCD + Ad.GFP or Ad.MCD + Ad.CA-AMPK virus. In the representative Western blot, lanes 1 and 3 ¼ Ad.MCD + Ad.GFP and lanes 2 and 4 ¼ Ad.MCD + Ad.CA-AMPK. The relative intensity and activity values are means ± SE of n ¼ 5 experiments. *Significantly different from Ad.MCD + Ad.GFP group, P < 0.05.

associated proteins like VDAC1 and ubiquinone-cytochrome C-core 2 subunit of complex III are not affected by increased CA-AMPKa1(312).

Discussion Regulation of MCD by AMPK remains controversial [16,17,25]. In this study, we demonstrate that AMPK regulates MCD by increasing levels of mitochondrial MCD protein and activity whereas, cytoplasmic MCD protein levels increased without a change in enzyme activity. In vitro incubation of purified enzymes confirms that MCD may not

be a direct substrate for AMPK [12,25]. However, it cannot be excluded that AMPK could indirectly modulate MCD activity in intact cells. Stimulation of MCD activity with AICAR in intact cells was very modest. This is probably due to the fact that AICAR stimulation of AMPK is only modest and also the level of endogenous MCD activity was very low in H9c2 cells to see a significant change in activity. We therefore, overexpressed CA-AMPKa1(312) in H9c2 cells and examined the regulation of MCD activity by AMPK. However, low levels of endogenous MCD in H9c2 cells posed a practical problem to measure either the protein or the enzyme activity in subcellular fractions. Hence, we also

FEBS 2004


Fig. 3. MCD expression and activity in cytosolic fractions of H9c2 cells. (A) Western blots for VDAC1 (i), cytochrome C (ii) and MCD expression (iii) in cytoplasmic and mitochondrial fractions from H9c2 cells infected with Ad.GFP or Ad.MCD. Western blots are representative of n ¼ 2 experiments. (B) MCD expression (i), optical density of 54.7 kDa isoform (ii), optical density of 50.7 kDa isoform (iii) and activity (iv) of cytoplasmic fractions obtained from H9c2 cells coinfected with Ad.MCD + Ad.GFP or Ad.MCD + Ad.CA-AMPK virus. Western blot is representative of n ¼ 6 experiments and relative intensities are means ± SE of n ¼ 6 experiments. Lanes 1 and 3 for Ad.MCD + Ad.GFP and lanes 2 and 4 for Ad.MCD + Ad.AMPK cells. Activity values are means ± SE of n ¼ 5 experiments. *Significantly different from Ad.MCD + Ad.GFP control, P < 0.05.

overexpressed MCD. Overexpression of CA-AMPKa1(312) resulted in a three- to fourfold increase in AMPK protein and activity. Similarly, MCD overexpression yielded a several-fold increase in MCD expression and activity. When AMPK was co-overexpressed, there was an increase in MCD activity in the mitochondrial fraction, which was due to an increase in the amount of MCD localized to the mitochondria. On the other hand, cytoplasmic fractions exhibited increases only in MCD protein levels and no change in activity compared with control conditions. In the heart, we and others [1,12] have previously demonstrated that the majority of MCD protein is in the short form ( 50.7 kDa) associated with mitochondria. Whether this short form is as a result of alternate splicing at the level of transcription or as a result of post-translational

modification of full length protein ( 54.7 kDa), is not yet known. It was proposed that once MCD is targeted to mitochondria, it may lose the mitochondrial target sequence by proteolytic cleavage and exists in the short form [38]. Our data support this concept. When we overexpressed human recombinant MCD in H9c2 cells both the short and the long forms of MCD were expressed. While the majority of the overexpressed MCD was the short form and was localized to mitochondria, the long form was expressed in Ad.MCD cells and was observed primarily in cytoplasm. As mitochondria are a rich source of MCD, it is possible that the short isoform could have leached out of mitochondria into cytoplasm during the fractionation procedures. In spite of an increased MCD protein in the cytoplasm, the activity did not increase in response to AMPK overexpression. In fact,


FEBS 2004

Fig. 4. MCD expression and activity in mitochondrial fractions of H9c2 cells. (A) MCD expression (i), optical density of 50.7 kDa isoform (ii) and activity (iii) of mitochondrial fractions obtained from H9c2 cells coinfected with Ad.MCD + Ad.GFP or Ad.MCD + Ad.CA-AMPK virus. In the Western blot, lanes 1 and 2 are for Ad.MCD + Ad.GFP cells and lanes 3 and 4 are for Ad.MCD + Ad.CA-AMPK cells. Western blot is a representative of n ¼ 3 experiments and relative intensity values are means ± SE of n ¼ 3 experiments. Activity values are means ± SE of n ¼ 5 experiments. *Significantly different from Ad.MCD + Ad.GFP control, P < 0.05. (B) Western blots for VDAC1 protein (i) and cytochrome c Core 2 subunit of complex III (ii) in mitochondrial fractions obtained from H9c2 cells infected with Ad.MCD + Ad.GFP or Ad.MCD + Ad.CAAMPK virus. Lanes 1, 2, 5 and 6 represent Ad.MCD + Ad.GFP cells and lanes 3, 4, 7 and 8 represent Ad.MCD + Ad.CA-AMPK cells. Results represent n ¼ 4 different passages from each group.

the specific activity per amount of protein was lower when compared with control cells suggesting that the long isoform, which contributes to most of the increases in cytoplasmic MCD protein, may be less active than the short form. Our data suggest that AMPK augments levels of both isoforms of MCD. Whether this increase in MCD expression by AMPK is a result of post-transcriptional regulation either affecting mRNA stability or protein stability is not known. Although evidence suggests that AMPK may regulate MCD transcription via PGC1 and PPARa [14,15,39,40], it may not be applicable here as MCD overexpression per se is driven by the cytomegalo virus promoter present in the recombinant Ad.MCD virus. The heart predominantly expresses the 50 kDa isoform of MCD [12]. In this study, we observed that this short isoform is mainly associated with mitochondria. In this study we demonstrated that AMPK overexpression faci-

litated an increase in the short MCD isoform in mitochondria, with a parallel increase in MCD activity. Although we did not screen for all the mitochondrial proteins, the increased CA-AMPKa1(312) activity did not affect the levels of other mitochondria-associated proteins like VDAC1 and ubiquinone–cytochrome c–core 2 subunit of complex III. This suggests that the role of AMPK in increasing MCD protein and activity in the mitochondria may be selective to MCD when compared with the other proteins tested above. Contrary to our findings, Habinowski et al. observed no differences in MCD activities between cytoplasmic and mitochondrial fractions [25]. In their study, an islet cell line was used, where a greater expression of the longer isoform of MCD is observed. Previous studies have shown that pancreatic MCD is post-translationally processed and regulated differently than either heart or muscle MCD

FEBS 2004


[12,38]. Pancreatic MCD appears in both longer and shorter forms while heart and muscle show mainly the shorter form of MCD [12,38]. This greater distribution of MCD in the cytoplasmic compartment may explain the lack of AMPK regulation of MCD in pancreatic islets in the above study. MCD protein has several potential Ser/Thr sites, phosphorylation of which could result in either a decrease or increase in activity [12,16,17]. Previously we have shown that dephosphorylation of MCD using alkaline phosphatase increased MCD activity suggesting that MCD is down regulated by phosphorylation [12]. However, recent studies in skeletal muscle demonstrated that phosphorylation of MCD increases its activity and that dephosphorylation by PP2A decreases or prevents the raise in MCD activity in response to activation of AMPK [16]. On the other hand in vitro incubation of purified MCD with heterotrimeric AMPK holoenzyme as well as constitutively active a1 subunit found that there was no phosphorylation of MCD [12,25]. In the present study, when we incubated the lysates from cells overexpressing MCD with those overexpressing CA-AMPKa1(312), we did not observe any phosphorylation of MCD (data not shown). Also, when immunoprecipitated MCD was probed for the myc-AMPK by Western blot analysis, we did not observe AMPK suggesting that there may be no physical interaction between the two proteins (data not shown). Taken together, this indicates that MCD may not be a direct substrate for AMPK in vivo. However, this does not rule out that AMPK can regulate MCD via other intermediary protein and by other post-translational modifications. In this regard, previous studies suggested that a 40 kDa protein that coprecipitated with MCD could be an MCD-inhibitory protein [41]. Although this study has limitations in that (a) a nonphysiological model system overexpressing MCD as well as AMPK was used, and (b) a constitutively active fragment of catalytic subunit of AMPK rather than physiological heterotrimeric form was used, the observations are interesting and support the possibility of differential regulation of MCD in different subcellular compartments. Of particular interest, basal malonyl-CoA levels in tissues are well above the inhibitory concentration for CPT-1 [42], suggesting a compartmentalization of cardiac malonyl-CoA. Thus, it is possible that malonylCoA levels in the vicinity of CPT-I (on the outer mitochondrial membrane) could undergo changes sufficient enough to either activate or inhibit CPT-I. In support of this, a recent study in human skeletal muscle observed a moderate increase in malonyl-CoA concentrations (20% of control) led to significant decrease in fatty acid oxidation (41% of control) [43]. Therefore, it is tempting to speculate that an AMPK mediated increase in MCD expression and activity selectively in mitochondria could potentially decrease malonyl-CoA levels sufficiently in the vicinity of CPT-I to increase CPT-I activity. This in turn would increase fatty acid uptake and oxidation. In summary, our results demonstrate that increasing AMPK activity by overexpression of constitutively active AMPK increases both MCD expression and activity. Whereas cytoplasmic MCD levels rise without any change in activity, both mitochondrial MCD levels and activity increase. Whether this differential regulation of MCD by AMPK is at the post-

transcriptional or post-translational level needs further investigation.

Acknowledgements This study was funded by a grant from the Canadian Institute for Health Research. N.S. is a postdoctoral fellow of the Alberta Heritage Foundation for Medical Research and Heart and Stroke Foundation of Canada. J.R.B.D. is a Scholar of the Alberta Heritage Foundation for Medical Research and a Canadian Institutes of Health Research New Investigator. G.D.L. is a Medical Scientist of the Alberta Heritage Foundation for Medical Research.

References 1. Hamilton, C. & Saggerson, E.D. (2000) Malonyl-CoA metabolism in cardiac myocytes. Biochem. J. 350, 61–67. 2. McGarry, J.D. & Foster, D.W. (1980) Regulation of hepatic fatty acid oxidation and ketone body production. Annu. Rev. Biochem. 49, 395–420. 3. McGarry, J.D., Mills, S.E., Long, C.S. & Foster, D.W. (1983) Observations on the affinity for carnitine, and malonyl-CoA sensitivity, of carnitine palmitoyltransferase I in animal and human tissues: demonstration of the presence of malonyl-CoA in nonhepatic tissues of the rat. Biochem. J. 214, 21–28. 4. Weis, B.C., Esser, V., Foster, D.W. & McGarry, J.D. (1994) Rat heart expresses two forms of mitochondrial carnitine palmitoyltransferase I. The minor component is identical to the liver enzyme. J. Biol. Chem. 269, 18712–18715. 5. Lopaschuk, G.D. (2001) Optimizing cardiac energy metabolism: how can fatty acid and carbohydrate metabolism be manipulated? Coron. Artery Dis. 12, S8–S11. 6. Sambandam, N., Lopaschuk, G.D., Brownsey, R.W. & Allard, M.F. (2002) Energy metabolism in the hypertrophied heart. Heart Failure Rev. 7, 161–173. 7. Lopaschuk, G.D., Witters, L.A., Itoi, T., Barr, R. & Barr, A. (1994) Acetyl-CoA carboxylase involvement in the rapid maturation of fatty acid oxidation in the newborn rabbit heart. J. Biol. Chem. 269, 25871–25878. 8. Kudo, N., Barr, A.J., Barr, R.L., Desai, S. & Lopaschuk, G.D. (1995) High rates of fatty acid oxidation during reperfusion of ischemic hearts are associated with a decrease in malonyl-CoA levels due to an increase in 5¢-AMP-activated protein kinase inhibition of acetyl-CoA carboxylase. J. Biol. Chem. 270, 17513– 17520. 9. Saddik, M., Gamble, J., Witters, L.A. & Lopaschuk, G.D. (1993) Acetyl-CoA carboxylase regulation of fatty acid oxidation in the heart. J. Biol. Chem. 268, 25836–25845. 10. Sakamoto, J., Barr, R.L., Kavanagh, K.M. & Lopaschuk, G.D. (2000) Contribution of malonyl-CoA decarboxylase to the high fatty acid oxidation rates seen in the diabetic heart. Am. J. Physiol. Heart Circ. Physiol. 278, H1196–H1204. 11. Courchesne-Smith, C., Jang, S.H., Shi, Q., DeWille, J., Sasaki, G. & Kolattukudy, P.E. (1992) Cytoplasmic accumulation of a normally mitochondrial malonyl-CoA decarboxylase by the use of an alternate transcription start site. Arch. Biochem. Biophys. 298, 576–586. 12. Dyck, J.R., Barr, A.J., Barr, R.L., Kolattukudy, P.E. & Lopaschuk, G.D. (1998) Characterization of cardiac malonylCoA decarboxylase and its putative role in regulating fatty acid oxidation. Am. J. Physiol. 275, H2122–H2129. 13. Dyck, J.R., Berthiaume, L.G., Thomas, P.D., Kantor, P.F., Barr, A.J., Barr, R., Singh, D., Hopkins, T.A., Voilley, N., Prentki, M. & Lopaschuk, G.D. (2000) Characterization of rat liver malonylCoA decarboxylase and the study of its role in regulating fatty acid metabolism. Biochem. J. 350, 599–608.

FEBS 2004

2840 N. Sambandam et al. (Eur. J. Biochem. 271) 14. Lee, G.Y., Cho, J.W., Lee, H.C. & Kim, Y.S. (2002) Genomic organization and characterization of the promoter of rat malonylCoA decarboxylase gene. Biochim. Biophys. Acta 1577, 133–138. 15. Campbell, F.M., Kozak, R., Wagner, A., Altarejos, J.Y., Dyck, J.R., Belke, D.D., Severson, D.L., Kelly, D.P. & Lopaschuk, G.D. (2002) A role for peroxisome proliferatoractivated receptor alpha (PPARalpha) in the control of cardiac malonyl-CoA levels: reduced fatty acid oxidation rates and increased glucose oxidation rates in the hearts of mice lacking PPARalpha are associated with higher concentrations of malonylCoA and reduced expression of malonyl-CoA decarboxylase. J. Biol. Chem. 277, 4098–4103. 16. Saha, A.K., Schwarsin, A.J., Roduit, R., Masse, F., Kaushik, V., Tornheim, K., Prentki, M. & Ruderman, N.B. (2000) Activation of malonyl-CoA decarboxylase in rat skeletal muscle by contraction and the AMP-activated protein kinase activator 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside. J. Biol. Chem. 275, 24279–24283. 17. Park, H., Kaushik, V.K., Constant, S., Prentki, M., Przybytkowski, E., Ruderman, N.B. & Saha, A.K. (2002) Coordinate regulation of malonyl-CoA decarboxylase, sn-glycerol-3-phosphate acyltransferase, and acetyl-CoA carboxylase by AMPactivated protein kinase in rat tissues in response to exercise. J. Biol. Chem. 277, 32571–32577. 18. Hardie, D.G. & Carling, D. (1997) The AMP-activated protein kinase – fuel gauge of the mammalian cell? Eur. J. Biochem. 246, 259–273. 19. Marsin, A.S., Bertrand, L., Rider, M.H., Deprez, J., Beauloye, C., Vincent, M.F., Van den Berghe, G., Carling, D. & Hue, L. (2000) Phosphorylation and activation of heart PFK-2 by AMPK has a role in the stimulation of glycolysis during ischaemia. Curr. Biol. 10, 1247–1255. 20. Kudo, N., Gillespie, J.G., Kung, L., Witters, L.A., Schulz, R., Clanachan, A.S. & Lopaschuk, G.D. (1996) Characterization of 5¢AMP-activated protein kinase activity in the heart and its role in inhibiting acetyl-CoA carboxylase during reperfusion following ischemia. Biochim. Biophys. Acta 1301, 67–75. 21. Gamble, J. & Lopaschuk, G.D. (1997) Insulin inhibition of 5¢-adenosine monophosphate-activated protein kinase in the heart results in activation of acetyl coenzyme A carboxylase and inhibition of fatty acid oxidation. Metabolism 46, 1270–1274. 22. Kemp, B.E., Mitchelhill, K.I., Stapleton, D., Michell, B.J., Chen, Z.P. & Witters, L.A. (1999) Dealing with energy demand: the AMP-activated protein kinase. Trends Biochem. Sci. 24, 22–25. 23. Kim, K.H., Lopez-Casillas, F., Bai, D.H., Luo, X. & Pape, M.E. (1989) Role of reversible phosphorylation of acetyl-CoA carboxylase in long-chain fatty acid synthesis. FASEB J. 3, 2250– 2256. 24. Park, S.H., Paulsen, S.R., Gammon, S.R., Mustard, K.J., Hardie, D.G. & Winder, W.W. (2002) Effects of thyroid state on AMPactivated protein kinase and acetyl-CoA carboxylase expression in muscle. J. Appl. Physiol. 93, 2081–2088. 25. Habinowski, S.A., Hirshman, M., Sakamoto, K., Kemp, B.E., Gould, S.J., Goodyear, L.J. & Witters, L.A. (2001) Malonyl-CoA decarboxylase is not a substrate of AMP-activated protein kinase in rat fast-twitch skeletal muscle or an islet cell line. Arch. Biochem. Biophys. 396, 71–79. 26. Lee, G.Y., Bahk, Y.Y. & Kim, Y.S. (2002) Rat malonyl-CoA decarboxylase; cloning, expression in E. coli and its biochemical characterization. J. Biochem. Mol. Biol. 35, 213–219. 27. Mulder, H., Lu, D., Finley, J.T., An, J., Cohen, J., Antinozzi, P.A., McGarry, J.D. & Newgard, C.B. (2001) Overexpression of a

28.

29.

30.

31.

32.

33.

34.

35.

36. 37. 38.

39.

40.

41.

42. 43.

modified human malonyl-CoA decarboxylase blocks the glucoseinduced increase in malonyl-CoA level but has no impact on insulin secretion in INS-1-derived (832/13) beta-cells. J. Biol. Chem. 276, 6479–6484. Dyck, J.R. & Lopaschuk, G.D. (1998) Glucose metabolism, H+ production and Na+/H+-exchanger mRNA levels in ischemic hearts from diabetic rats. Mol. Cell Biochem. 180, 85–93. Hardie, D.G., Salt, I.P. & Davies, S.P. (2000) Analysis of the Role of the AMP-Activated Protein Kinase in the Response to Cellular Stress. Humana Press Inc, Totowa, NJ. Gao, J., Waber, L., Bennett, M.J., Gibson, K.M. & Cohen, J.C. (1999) Cloning and mutational analysis of human malonyl-coenzyme A decarboxylase. J. Lipid Res. 40, 178–182. He, T.C., Zhou, S. & da Costa, L.T., Yu J., Kinzler, K.W. & Vogelstein, B. (1998) A simplified system for generating recombinant adenoviruses. Proc. Natl Acad. Sci. USA 95, 2509– 2514. Sacksteder, K.A., Morrell, J.C., Wanders, R.J., Matalon, R. & Gould, S.J. (1999) MCD encodes peroxisomal and cytoplasmic forms of malonyl-CoA decarboxylase and is mutated in malonylCoA decarboxylase deficiency. J. Biol. Chem. 274, 24461–24468. Woods, A., Azzout-Marniche, D., Foretz, M., Stein, S.C., Lemarchand, P., Ferre, P., Foufelle, F. & Carling, D. (2000) Characterization of the role of AMP-activated protein kinase in the regulation of glucose-activated gene expression using constitutively active and dominant negative forms of the kinase. Mol. Cell Biol. 20, 6704–6711. McMillin, J.B., Edgar, K. & Buja, L.M. (1993) Long chain acylCoA metabolism by mitochondrial carnitine palmitoyltransferase: A cell model for pathological studies. In Methods in Toxicology, pp. 301–309. Academic Press Inc, San Diego, CA, USA. Rice, J.E. & Lindsay, J.G. (1997) Subcellular fractionation of mitochondria in subcellular fractionation. In A Practical Approach Series (Graham, J.M. & Rickwood, D., eds), pp. 117–118. Oxford University Press, Oxford, UK. Rice, J.E. & Lindsay, J.G. (1997) Subcellular Fractionation of Mitochondria. Oxford University Press, Oxford, UK. Green, D.R. & Reed, J.C. (1998) Mitochondria and apoptosis. Science 281, 1309–1312. Voilley, N., Roduit, R., Vicaretti, R., Bonny, C., Waeber, G., Dyck, J.R., Lopaschuk, G.D. & Prentki, M. (1999) Cloning and expression of rat pancreatic beta-cell malonyl-CoA decarboxylase. Biochem. J. 340 (1), 213–217. Lehman, J.J., Barger, P.M., Kovacs, A., Saffitz, J.E., Medeiros, D.M. & Kelly, D.P. (2000) Peroxisome proliferator-activated receptor gamma coactivator-1 promotes cardiac mitochondrial biogenesis. J. Clin. Invest. 106, 847–856. Zong, H., Ren, J.M., Young, L.H., Pypaert, M., Mu, J., Birnbaum, M.J. & Shulman, G.I. (2002) AMP kinase is required for mitochondrial biogenesis in skeletal muscle in response to chronic energy deprivation. Proc. Natl Acad. Sci. USA 99, 15983– 15987. Young, M.E., Goodwin, G.W., Ying, J., Guthrie, P., Wilson, C.R., Laws, F.A. & Taegtmeyer, H. (2001) Regulation of cardiac and skeletal muscle malonyl-CoA decarboxylase by fatty acids. Am. J. Physiol. Endocrinol. Metab. 280, E471–E479. Eaton, S. (2002) Control of mitochondrial beta-oxidation flux. Prog. Lipid Res. 41, 197–239. Bavenholm, P.N., Pigon, J., Saha, A.K., Ruderman, N.B. & Efendic, S. (2000) Fatty acid oxidation and the regulation of malonyl-CoA in human muscle. Diabetes 49, 1078–1083.