• acetyl-l-carnitine;
  • NMR spectroscopy;
  • carbon-13;
  • cholesterol;
  • lipogenesis


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The biochemical pathways involved in acetyl-l-carnitine utilization were investigated in conscious, freely moving rats by 13C NMR spectroscopy. Following 4-h [(1,2-13C2)acetyl]-l-carnitine infusion in fasted animals, the free carnitine levels in serum were increased, and an efflux of unlabelled acetyl-l-carnitine from tissues was observed. [(1,2-13C2)Acetyl]-l-carnitine was found to enter biosynthetic pathways in liver, and the acetyl moiety was incorporated into both cholesterol and 3-hydroxybutyrate carbon skeleton. In accord with the entry of [(1,2-13C2)acetyl]-l-carnitine in the mitochondrial acetylCoA pool associated with tricarboxylic acid cycle, the 13C label was also found in liver glutamate, glutamine, and glutathione. The analysis of the 13C-labelling pattern in 3-hydroxybutyrate and cholesterol carbon skeleton provided evidence that the acetyl-l-carnitine-derived acetylCoA pool used for ketone bodies synthesis in mitochondria was homogeneous, whereas cholesterol was synthesized from two different acetylCoA pools located in the extra- and intramitochondrial compartment, respectively. Furthermore, cholesterol molecules were shown to be preferentially synthesized by the metabolic route involving the direct channelling of CoA-activated mitochondria-derived ketone bodies into 3-hydroxy-3-methylglutarylCoA pathway, prior to equilibration of their acyl groups with extramitochondrial acetylCoA pool via acetoacetylCoA thiolase.



Acetyl-l-carnitine is a major esterified carnitine congener in most mammalian tissues [1,2]. It is synthesized by carnitine O-acetyltransferase reaction (EC2.3.1.7), which reversibly converts acetylCoA and free carnitine to acetyl-l-carnitine and CoASH without ATP utilization. Carnitine O-acetyltransferase activity was reported to be present almost in all tissues [3,4], and in the liver, it was found in peroxisomes, mitochondria, and microsomes [5,6].

Previous investigations on acetyl-l-carnitine metabolism have focused on its role in energy-producing processes [7,8]. Evidence has been also presented that supports the conclusion that the acetylcarnitine/carnitine system buffers the CoASH/acetylCoA ratio in the mitochondrial matrix [9].

To date, however, no detailed study has investigated both the role of acetyl-l-carnitine in acetyl groups trafficking between intracellular organelles and its involvement in oxidative and biosynthetic processes in vivo.

As a product of peroxisomal β-oxidation, acetyl-l-carnitine may be shuttled out from peroxisomes to the cytosol [10,11], while the acetylCoA synthesized in mitochondria by glucose or fatty acid oxidation may be transported to the cytosolic compartment as acetyl-l-carnitine both in liver [12,13] and heart myocytes [14]. Studies on exogenous administration of ([1-14C]acetyl)-l-carnitine in 24-h fasted mice reported that most acetyl moiety was oxidized in mitochondria and radioactivity was released as 14CO2[15]. Moreover, the liposoluble fraction of tissue extracts retained most of the [14C] label, whereas the aqueous and protein fractions were labelled to a minor extent [15].

From the above, it seems reasonable to claim for acetyl-l-carnitine a possible pivotal role in modulating the activity of both biosynthetic and oxidative acetylCoA pools, which are differently located and metabolically compartmentalized in the cells.

13C NMR spectroscopy and the administration of 13C-labelled compounds have been demonstrated to be a powerful tool for investigating cell metabolism [16].

Measurements of 13C label distribution in carbon skeleton of metabolic intermediates provided useful information about different biochemical pathways. In addition, by analyzing the 13C homonuclear spin coupling patterns, a property unique to contiguously labelled compounds, the metabolic fate of substrates containing 13C-13C units can be followed [16].

In this paper, in order to gain more knowledge about different aspects of intracellular acetyl-l-carnitine metabolism, [(1,2-13C2)acetyl]-l-carnitine was infused in conscious, freely moving rats, and the 13C label distribution in hydrosoluble and liposoluble metabolites of serum and liver extracts was analyzed by 13C NMR spectroscopy. We shown that exogenous acetyl-l-carnitine is transported to and metabolized in the liver. The acetyl moiety of [(1,2-13C2]acetyl)-l-carnitine enters oxidative and biosynthetic pathways and is involved in feeding two differently compartmentalized cholesterogenic and ketogenic acetylCoA pools.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References


[(1,2-13C2]acetyl)-l-carnitine was synthesized by the Chemistry Department, Sigma-Tau Research Labs, by O-acetylation of l-carnitine. Briefly, 1 g of l-carnitine was dissolved at 0 °C with 1 mL trifluoracetic acid. To this solution, 0.5 mL of 99.9% [1,2-13C2]acetyl chloride (Isotec Inc., Miamisburg, OH, USA) was added and the mixture was left at room temperature for 4 h. The reaction product was precipitated with 10 mL ethyl ether, recovered on filter paper, and dried overnight at 30 °C. The reaction yield was 92% and the purity of [(1,2-13C2)acetyl]-l-carnitine, tested by HPLC and NMR, was 99.9%.

Animals and surgical procedures

Male Fischer 344 rats (Charles River, Como, Italy) aged 6 months, were used. Animals were housed in a light- and temperature-controlled room (lights on 07:00–19:00; temperature, 21 °C) and fed on normal rat chow (4RF 21, Mucedola srl, Settimo Milanese, Italy) and water ad libitum. Under chloral hydrate anaesthesia, the left internal jugular and the right femoral veins were isolated from the surrounding tissues. Silmedic® catheters (Gessil, St. Quentin-Fallavier, France) were implanted into each vein and anchored at the venotomy site with silk ligatures. The free ends of the catheters were tunnelled subcutaneously around the side of the body or of the neck to the top of the skull, where they were exteriorized via a skin incision. Catheters were then connected to a special device (Danuso Umberto srl, Bresso, Italy) placed on the top of the neck, which allowed blood sampling (jugular catheter), and substances infusion (femoral catheter) in conscious, freely moving rats. At the end of the procedure and then daily until the study onset, catheters were flushed with isotonic saline containing heparin (100 U·mL−1). Animals were allowed 5 days to recover from the effect of surgery and anaesthesia. Care and treatment of experimental animals was in accordance with European Communities Council Directive of 24 November 1986 (86/609/EEC).

Acetyl-L-carnitine infusion and blood sampling

All rats were fasted 24 h before the infusion experiments. We used conscious, freely moving rats because an in-vivo model was critical for maintaining all the complex mechanisms regulating the whole body carnitines homeostasis. Also, the use of anaesthesia [17] and the restraint stress [18] have been shown to interfere with cholesterol metabolism.

On the day of the study, [(1,2-13C2]acetyl)-l-carnitine (n = 3) or unlabelled acetyl-l-carnitine (n = 3) were dissolved to a final concentration of 105 mg·mL−1 in 2.5 mL of a solution containing NaCl (128 mm), NaHCO3 (24 mm), KCl (4.2 mm), NaH2PO4 (2.4 mm), CaCl2 (1.5 mm), MgSO4 (0.9 mm). The acetyl-l-carnitine solutions were drawn up in a plastic syringe in a computer-controlled variable rate Harvard 22 infusion pump (Harvard Apparatus Inc., Holliston, MA, USA). The pump rate was set to quickly reach and maintain a high plasma concentration of acetyl-l-carnitine throughout the experiments. The total amount of acetyl-l-carnitine administered during 4 h infusion was 620 mg·kg−1 body mass. Blood samples were collected at 1 h before and at 1, 2.5, and 4 h after the start of acetyl-l-carnitine infusion. In order to avoid dilution of the blood sample with the dead space fluid in the catheter and connection tubing, the following routine was strictly observed: 0.5 mL of blood was drawn into 1-mL syringe and set aside. This volume exceeded the dead space by 25-fold. A fresh syringe was then attached and 400 µL of blood sample withdrawn. Finally, the first syringe was reattached and its content reinfused. The time required for this three-step procedure was approximately 15 s. To prevent catheter obstruction, 100 µL of a saline solution containing 10 U·mL−1 of heparin was administered following each blood sampling.

At the end of the 4-h infusion, animals were anaesthetized with halothane 1.5% in N2O/O2 70/30%. The abdominal wall was opened, the liver was rapidly freeze-clamped with cooled tongs, and dropped into a Dewar flask containing liquid nitrogen.

Sample processing

Liver and plasma samples were extracted with 5 mL·g−1 tissue (or mL plasma) of chloroform/methanol/water at the final proportion of 2 : 2 : 1 v/v as previously described [19]. Following 20 min centrifugation at 4 °C and 10 000 g, the aqueous phases of the extracts were separated. For the liver samples, the remaining lower liposoluble fraction and tissue residue were re-extracted overnight with 50 mL·g−1 tissue of chloroform/methanol/water 0.9% NaCl, at proportions of 2 : 1 : 3, v/v, to ensure complete removal of lipids. Aqueous and liposoluble extracts were dried under a nitrogen stream and stored at −80 °C until NMR measurements.

NMR spectroscopy

1H and 13C NMR spectra of liver and serum extracts were obtained on a Bruker AM 500 spectrometer (Bruker Spectrospin, Milano, Italy). Dried samples were dissolved in 300 µL of D2O (hydrosoluble fraction) or in 300 µL of deuterated chloroform/methanol 2 : 1 (v/v) (liposoluble fraction). Furthermore, 15 µL of acetonitrile in an external capillary tube were used in 13C spectra as a reference for chemical shift and concentration determination. Proton decoupled 13C spectra were accumulated using a 45° pulse angle with a spectral width of 31 kHz and 32 000 data points. The acquisition time was 0.52 s and an additional relaxation delay of 7.48 s was used. To avoid nuclear Overhauser effect, 13C spectra were broad-band decoupled only during acquisition using an inverse-gated sequence. 1H NMR spectra were acquired with the following parameters: 45° pulse, 6000 Hz spectral width and 64 000 data points. The acquisition time was 5.44 s and an additional delay of 10 s was used to minimize effects of the different relaxation times of protons bound either to 12C or 13C carbon neighbours. The NMR data were processed off-line by using win-nmr software (Bruker-Franzen Analitik GmbH, Bremen, Germany). Peak areas were used to quantify the amount of 1H or 13C, and resonance assignments of the cholesterol carbons as well as those of the other metabolites were made by a comparison with the literature [16,20]. The percentage 13C enrichment of the acetyl-l-carnitine in serum was calculated in 1H spectra from the ratio between the satellite peak areas resulting from 1H-13C spin couplings (proton linked to 13C atoms) and the total peak areas (resonances of protons linked to both 12C and 13C atoms). The CH3 resonance of acetyl moiety (protons-bound to the C2 carbon) centered at 2.14 p.p.m. was used. In 13C spectra the total amount of 13C in each carbon resonance of the metabolites was determined using acetonitrile as a concentration standard. The percentage 13C enrichment in each carbon position of the metabolites was calculated by subtracting the naturally abundant 13C (1.08% of total carbon) from the total amount of 13C label found in that position, and dividing by the total amount of the specific metabolite [21]. The following equation was used:

  • Percent13C enrichment = ([13C]t -[13C]a)·100/[m](1)

where [13C]t and [13C]a are the total amounts and naturally abundant 13C at a carbon position of the metabolite, respectively, and [m] is the pool size of that metabolite.

When 13C-13C units were incorporated in metabolite carbon skeleton, the doublets arising from 13C-13C spin couplings between two adjacent 13C atoms were measured and the percentage 13C enrichment for each carbon position was calculated.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The 13C spectra of hydrosoluble and liposoluble fractions of liver and plasma contain a good deal of information reflecting their complex metabolite composition. Table 1 reports the time course of changes in serum acetyl-l-carnitine and free carnitine levels, as well as the percentage 13C enrichment of acetyl-l-carnitine at 1 h before and at 1, 2.5, and 4 h during infusion with acetyl-l-carnitine. Serum acetyl-l-carnitine content was rapidly increased to ≈1 mm after the start of infusion, and a concomitant increase in free carnitine levels was observed (Table 1). Interestingly, despite the continuous infusion of 99.9% 13C-enriched acetyl-l-carnitine, the percentage 13C enrichment of acetyl-l-carnitine in serum maintained a steady-state value of about 80% throughout the experiment (Table 1).

Table 1. Serum carnitine and acetyl-l-carnitine levels, and percentage13C enrichment of acetyl-l-carnitine during [(1,2-13C2)acetyl]-l-carnitine infusion. Rats were infused 4 h with [(1,2-13C2)acetyl]-l-carnitine and serum samples were collected at 1 h before and at 1, 2.5, and 4 h during infusion. The percentage 13C enrichment of acetyl-l-carnitine was calculated as described in Materials and methods. Values are means ± SD of three rats.
Time (hours)[Carnitine] (µmol· L−1)[Acetyl-l-carnitine] (µmol· L−1)Acetyl-l-carnitine % 13C enrichment
–117 ± 224 ± 2
+1197 ± 131108 ± 15881 ± 4
+2.5222 ± 341090 ± 14477 ± 3
+4194 ± 321010 ± 16181 ± 2

The intravenous administration of acetyl-l-carnitine, both in animals and humans, has been previously reported to induce an efflux of free carnitine from tissues to plasma [22,23]. Nevertheless, an operative acylcarnitine/carnitine exchange system has been suggested to occur in different tissues [24,25]. Our findings do not allow to unequivocally identify the origin of the free carnitine in serum, because the 13C label is present in the acetyl moiety of acetyl-l-carnitine. However, they appear consistent with the activity of an exchange system between exogenous acetyl-l-carnitine from blood to tissues and free carnitine in the opposite direction.

Furthermore, with respect to the percentage 13C enrichment of serum acetyl-l-carnitine, it is hypothesized that following the entry of 13C-labelled acetyl-l-carnitine and its equilibration with intracellular acetylCoA pools via carnitine O-acetyltransferase reaction, unlabelled acetyl-l-carnitine may escape from tissues, thus contributing to the dilution of 13C enrichment in serum acetyl-l-carnitine.

13C NMR spectra of liver hydrosoluble extracts contained resonances relative to different carbon positions of glucose, lactate, glutamate, glutamine and glutathione.

Signals of the 3-hydroxybutyrate carbons were reported in liver as well as in serum extracts. In serum, incorporation of 13C label from [(1,2-13C2)acetyl]-l-carnitine into all carbon atoms of 3-hydroxybutyrate was observed (Table 2). A similar 13C enrichment was also found in C2, C3, and C4 carbons of liver 3-hydroxybutyrate (Table 2), whereas the resonance relative to the C1 position overlapping other signals was not determined.

Table 2. Percent13C enrichment in the different carbon positions of 3-hydroxybutyrate molecules from serum and liver extract. Rats were infused 4 h with [(1,2-13C2)acetyl]-l-carnitine. Serum and liver samples were collected at 4 h following infusion. Percent 13C enrichment was calculated as described in materials and methods. Values are means ± SD of three rats. ND, not detectable.
 % 13C enrichment
Serum0.84 ± 0.030.89 ± 0.040.82 ± 0.080.89 ± 0.05
LiverND0.91 ± 0.050.87 ± 0.131.06 ± 0.18

Note that 13C label in the different carbon positions of 3-hydroxybutyrate was even present in the 13C NMR spectra as doublets due to 13C-13C fragments. Incorporation of singlets due to unpaired 13C atoms was not observed. Moreover, no significant differences in the 13C enrichment found in the first two carbon atoms (C1-C2 fragment) with respect to that measured in the C3-C4 fragment were detected.

The acetyl moiety of acetyl-l-carnitine was also incorporated as an intact 13C-13C fragment into C4-C5 carbons of liver glutamate and glutamine (Table 3). Furthermore, incorporation of 13C-13C fragments into C4-C5 carbons of glutamyl moiety of glutathione was found (Table 3).

Table 3. Percent13C enrichment in C-4 carbon positions of liver glutamate, glutamine and glutathione. Rats were infused 4 h with [(1,2-13C2)acetyl]-l-carnitine and liver samples were collected at the end of infusion. The 13C label were incorporated as intact 13C-13C units into C4-C5 positions. Percent 13C enrichment was calculated as described in Materials and methods. Values are means ± SD of three rats.
Compound% 13C enrichment
Glutamate1.64 ± 0.61
Glutamine1.30 ± 0.46
Glutathione0.80 ± 0.34

Signals relative to cholesterol, fatty acids and other lipid species were observed in 13C spectra of liposoluble fractions. As an example, some expansions of 13C NMR spectra relative to cholesterol resonances of the liver liposoluble extracts from rats infused with unlabelled (bottom) and 13C-labelled (top) acetyl-l-carnitine are shown in Fig. 1.


Figure 1. Portions of13C NMR spectra of liposoluble liver extracts from rats infused for 4 h with13C-labelled (top) and unlabelled (bottom) acetyl-l-carnitine. Expansions display the different types of 13C labelling found in cholesterol carbon positions. Doublets or type I (C17 and C20 carbons), singlets from former doublets or type II (C14 carbon), and singlets or type III (C22 carbon), respectively. Note that the incorporation of 13C-13C units leads to the presence of satellite bands (*) adjacent to the main peak of the carbon resonances, while an increase in signal intensity was observed when unpaired 13C atoms were incorporated.

Infusion with [(1,2-13C2)acetyl]-l-carnitine led to a 13C label distribution into different cholesterol carbons, and the appearance of doublets due to 13C-13C couplings near the main peak of naturally occurring 13C (1.08%) (Fig. 1, top). In addition, the enhancement of the main peak in specific positions of cholesterol carbon skeleton due to incorporation of unpaired 13C atoms is also observed (Fig. 1, top).

To facilitate understanding about the different 13C labelling pattern observed in cholesterol carbon skeleton, a short description of the biochemical pathways involved in its synthesis is depicted in Fig. 2. Briefly, in the extramitochondrial compartment, two molecules of acetylCoA condense to form acetoacetylCoA molecules (pathway 1).


Figure 2. Utilization of the acetyl moiety from [(1,2-13C2)acetyl]-l-carnitine for cholesterol synthesis in liver. To facilitate the understanding about the source and the fate of the 13C-13C units from acetyl-l-carnitine, the cholesterol structure was drawn in terms of isopentenyl pyrophosphate precursor molecules (heavy lines). The mechanisms by which a 13C atom can lose its adjacent 13C neighbour, thus originating a single 13C-enriched site (crossed and grey circles) with respect to other carbon positions (black and white circles), are described in the text. Nevertheless, the cleavage of 13C-13C units in the conversion of mevalonate to isopentenyl pyrophosphate (crossed circles), the two methyl migrations, and subsequent loss of three methyls (grey circles), were shown.

Note that the acetoacetylCoA molecules can also be produced in the extramitochondrial compartment by activation of the mitochondria-synthesized acetoacetate or 3-hydroxybutyrate to their CoA esters via acetoacetate-CoA ligase or 3-hydroxybutyrylCoA synthetase (pathway 2). The next step introduces a third molecule of acetylCoA, thereby producing the branched chain compound 3-hydroxy-3-methylglutarylCoA (HMGCoA) via hydroxymethylglutarylCoA synthase (EC4.1.3.5). This compound is then reduced to mevalonate by hydroxymethylglutarylCoA reductase (NADPH) reaction (EC1.1.1.34).

In the cholesterol synthesis pathway, mevalonate molecules first undergo a double phosphorylation to produce 5-pyrophosphomevalonate and are then decarboxylated to form isopentenyl pyrophosphate via diphosphomevalonate decarboxylase reaction (EC4.1.1.33). As a consequence, mevalonate decarboxylation involves a cleavage of the C–C fragment deriving from the acetylCoA molecule added in the hydroxymethylglutarylCoA synthase reaction. Thus, in the cholesterol synthesis by the pathway 2, the 13C enrichment of C1, C2, C3, and C4 carbons of isopentenyl pyrophosphate directly deriving from the 4-carbon units (acetoacetate or 3-hydroxybutyrate), is influenced by the 13C enrichment of the acetylCoA pools located in both the extra- and intramitochondrial compartments. In contrast, the 13C enrichment of the C5 carbon of isopentenyl pyrophosphate reflects exclusively that of the extramitochondrial acetylCoA pool.

Further steps of cholesterol synthesis lead to formation of squalene from six isopentenyl pyrophosphate molecules. Finally, after a series of methyl and proton migrations, and decarboxylation reactions, the synthesis of cholesterol is completed.

Table 4 displays the percentage 13C enrichment in the different carbon positions of hepatic cholesterol resulting from the pathways described in Fig. 2. Data were reported according to three types of 13C labelling as follows. In type I are gathered the carbon positions where the 13C label was found as 13C-13C fragment, so that the resonances of cholesterol carbon positions C2-C3, C5-C6, C9-C11, C10-C19, C12-C13, C16-C17, C20-C21, C23-C24, C25-C27 (displayed in Fig. 2 as black and white circles) appeared in 13C NMR spectra as doublets. These carbons derive from the acetoacetylCoA backbone and in the cholesterol synthesis pathway, the 13C-13C units were maintained as an intact fragment. In type II are found the same 13C-13C fragments of the acetoacetylCoA carbon skeleton, which subsequent to the cyclization of squalene to lanosterol, have lost their 13C neighbours by methyl migrations (one from C14 to C13 and the other from C8 to C14, respectively), resulting in a single 13C enriched site at cholesterol carbons C8, C14, C18 (grey circles of Fig. 2). In type II are also gathered the carbons which in the conversion from lanosterol to cholesterol, have lost three methyl groups (two methyls from C4 and one methyl from C14) resulting in a single 13C enriched site at cholesterol carbon C4 (grey circles of Fig. 2). Being originally part of an intact 13C-13C fragment in isopentenyl pyrophosphate molecules prior to their condensation, these carbons were also identified as singlets from former doublets. In type III are found the carbon atoms deriving from to the third acetylCoA molecule utilized for HMGCoA synthesis, which have successively lost their 13C neighbours in the conversion of mevalonate to isopentenyl pyrophosphate. As a consequence of the fragmentation of the 13C-13C units, unpaired 13C atoms were incorporated in the cholesterol molecules and a single 13C enriched site was found in the carbon positions C1, C7, C15, C22, C26 (crossed circles of Fig. 2). A quantitative analysis of the three different types of 13C labelling as found in the cholesterol carbon skeleton, revealed that the total amount of 13C label incorporated as doublets (black and white circles) and as singlets from former doublets (grey circles) originating from the acetoacetylCoA backbone, was significantly higher (P < 0.001) than the amount of the 13C label incorporated as 13C singlets (unpaired 13C atoms; crossed circles of Fig. 2), the latter directly derived from the extramitochondrial acetylCoA used for HMGCoA synthesis.

Table 4. Percent13C enrichment of the different13C labelling types found in liver cholesterol carbon positions. Rats were infused 4 h with [(1,2-13C2)acetyl]-l-carnitine and liver samples were collected at the end of infusion. Percent 13C enrichment was calculated as described in Materials and methods. Values are means ± SD of three rats. I = doublets, II = singlets from former doublets, III = singlets.
Type of 13C labelling% 13C enrichment
  • *

    P < 0.001 vs singlets.

(I) Doublets0.71 ± 0.14*
(II) Singlets from former doublets0.65 ± 0.13*
(III) Singlets0.36 ± 0.10


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In the present study, we have reported some new aspects of acetyl-l-carnitine metabolism in an in-vivo animal model. Acetyl-l-carnitine was shown to enter both liver biosynthetic and catabolic pathways in conscious, freely moving rat. In particular, the acetyl moiety of acetyl-l-carnitine appears to feed two distinct acetylCoA pools (intra- and extramitochondrial) involved in cholesterol synthesis.

In a previous in-vivo study, it was reported that 1 h following [1-14C]acetyl-l-carnitine injection to fasted mice most of the 14C label was metabolized in mitochondria and was released as expired 14CO2[15]. Moreover, the liposoluble fraction of tissue extracts retained most of 14C label, whereas the aqueous and protein fractions were labelled to a minor extent [15]. In our study, the incorporation of 13C-labelled acetyl units from [(1,2-13C2)acetyl]-l-carnitine into glutamate and glutamine carbons provides evidence that exogenous acetyl-l-carnitine is able to reach the mitochondrial compartment. As a result of the operation of the tricarboxylic acid cycle, 13C-labelled acetylCoA molecules deriving from [(1,2-13C2)acetyl]-l-carnitine were incorporated as 13C-13C units in C4-C5 carbons of 2-oxoglutarate and hence, in the C4-C5 of glutamate and glutamine (Fig. 3). The appearance of 13C label in the C4-C5 carbons of the glutamyl moiety of glutathione showed that 13C-labelled glutamate synthesized in mitochondria was used for γ-glutamylcysteine synthesis in the cytosol.


Figure 3. Metabolism of [(1,2-13C2)acetyl]-l-carnitine in the liver. The 13C label distribution from acetyl moiety of acetyl-l-carnitine into different metabolites by tricarboxylic acid cycle and ketogenesis.

A mitochondrial carnitine O-acetyltransferase was also involved in converting [(1,2-13C2)acetyl]-l-carnitine to [(1,2-13C2]acetylCoA molecules, the latter being utilized for ketone bodies synthesis (Fig. 3).

Mitochondrial acetylCoA accounts for about 88% of total liver acetylCoA [26], while its isotopic homogeneity is still controversial. Labeling experiments in perfused liver [27] and isolated hepatocytes [28] provided indirect evidence for compartmentation of mitochondrial acetylCoA. Nevertheless other studies based on the comparison of the specific activity of ketone bodies versus citrate [29] or CO2[30] following administration of different [14C]-labelled substrates concluded that the pool of liver acetylCoA was homogeneous. Our data do not allow us to identify unequivocally whether the ketogenic and oxidative acetylCoA pools in mitochondria were metabolically compartmentalized. However, there is strong evidence that the ketogenic acetylCoA pool deriving from acetyl-l-carnitine is homogeneous, in that all carbon positions of 3-hydroxybutyrate were equally labelled. Moreover, following [(1,2-13C2]acetylCoA synthesis by carnitine O-acetyltransferase reaction, these molecules were readily available for 3-hydroxybutyrate synthesis via the HMGCoA pathway, with no passage through either the tricarboxylic acid cycle or the pyruvate recycling pathway, which are known to produce labelling fragmentation leading to monolabelled acetylCoA molecules from doubly and contiguously 13C-labelled precursors [16].

Consistent with this hypothesis, no incorporation of single 13C atoms in 3-hydroxybutyrate carbon positions was observed in 13C NMR spectra.

Cholesterol synthesis is known to occur in microsomes using acetylCoA units that are mostly produced in mitochondria as precursors. The main mechanism of acetyl transfer takes place via the citrate-cleavage pathway, which includes mitochondrial citrate synthesis, transfer of citrate to the cytosol, and cleavage by ATP citrate lyase. Peroxisomes also contain a number of enzymes required for cholesterol biosynthesis [31] and they have been showed to synthesize cholesterol from mevalonate [32,33].

The peculiar 13C labelling pattern observed by us in cholesterol carbon skeleton results from the specific pathways involved in its synthesis (see Fig. 2). In particular, we observed an enhancement of resonance peaks as singlets when they represent one 13C-enriched site, and the appearance of doublets when two 13C atoms in the cholesterol were adjacent neighbours. As shown in Table 4, a quantitative analysis revealed that the percentage 13C enrichments of cholesterol carbons appearing as doublets and singlets from former doublets and deriving from acetoacetylCoA moiety (C1, C2, C3 and C4 of isopentenyl pyrophosphate, see Fig. 2), were significantly higher than those found as singlets (C1, C7, C15, C22, and C26), the latter corresponding to C5 of isopentenyl pyrophosphate (see Fig. 2). It should also be kept in mind that the C5 carbon of isopentenyl pyrophosphate derives from the extramitochondrial acetylCoA pool. These results provide evidence that in the liver of fasted rats, cholesterol was synthesized from two differently 13C-enriched acetylCoA pools. Moreover, according with the pathway 2 described in Fig. 2, these acetylCoA pools were likely located in mitochondrial and extramitochondrial compartments, respectively.

The extramitochondrial acetylCoA pool used in the HMGCoA synthase reaction could be directly fed by acetyl-l-carnitine via carnitine O-acetyltransferase, known to be present in liver microsomes and peroxisomes [5,6]. Regarding the mitochondrial acetylCoA pool, our findings were consistent with the hypothesis that the ketone bodies synthesized in mitochondria from [(1,2-13C2)acetyl]-l-carnitine were directly utilized as an intact 4-carbon units for cholesterol synthesis via HMGCoA pathway, without equilibrating with the extramitochondrial acetylCoA pool via acetoacetylCoA thiolase (EC2.3.1.9) (pathway 2 of Fig. 2).

In agreement with this hypothesis, it has been previously reported that both exogenous acetoacetate and 3-hydroxybutyrate can be used as intact 4-carbon units for cholesterol synthesis in isolated and perfused rat liver [34]. Furthermore, acetoacetate-CoA ligase and 3-hydroxybutyrylCoA synthetase, which catalyse the activation of acetoacetate and 3-hydroxybutyrate to CoA ester, respectively, have been recently found in the rat liver cytosol [35,36].

Note that the contribution of endogenous and newly synthesized ketone bodies to cholesterol synthesis observed by us has been shown under in-vivo conditions, thus avoiding any possible interference deriving from the direct activation of exogenous ketone bodies in the cytosol, or the in-vitro liver preparation.

Previous studies on the acetyl-l-carnitine flux to lipids, assessed in cell cultures by isotopomer spectral analysis [37] have shown that the acetyl moiety of acetyl-l-carnitine contributes 10% to the lipogenic acetylCoA pool used for the de-novo synthesis of palmitate and 6% to the ketogenic acetylCoA pool, respectively [38]. In agreement with our results, these authors have also shown that acetyl-l-carnitine may feed directly the cytosolic acetylCoA pool used for both cholesterol and fatty acid synthesis, thus, bypassing the citrate-cleavage pathway [38].

In a recent study on lipogenesis in HepG2 hepatoma cell lines, it was shown that despite the use of different substrates (i.e. acetate, acetoacetate, and octanoate) as carbon sources, a common cytosolic acetylCoA pool appeared to be the precursor for both cholesterol and fatty acid synthesis [39]. Moreover it was also assumed that the acetoacetylCoA thiolase reaction is the major route for acetoacetate metabolism to acetylCoA, and acetoacetateCoA ligase did not contribute to cholesterol synthesis [39]. In this regard, it should be considered that cholesterogenesis in neoplastic tissue differs from normal tissues in being elevated and resistant to sterol feedback regulatory control [40]. Nevertheless, changes in culture medium composition in terms of substrates and hormones were reported to dramatically affect the membrane lipid metabolism in cultured human T lymphocytes [41], as well as the cholesterol synthesis and/or turnover pattern in HepG2 cell lines [39]. Thus, differences in acetyl-l-carnitine utilization and in cholesterol synthesis pathway between liver cells in vivo and hepatoma cells may be related with changes in the enzymatic expression pattern, the different substrates utilized, and the conditions in vivo (in rat) versus those in vitro (in cultured cells). Interestingly, it should be also pointed out that the direct utilization of mitochondria-synthesized 4-carbon units for cholesterol synthesis (Table 4 and Fig. 2), supports the hypothesis of an enzyme organization, and/or channelling of metabolites within the early cholesterogenic steps [42].

In conclusion, we have reported new insights into acetyl-l-carnitine metabolism allowing a better understanding of its physiological role in the cell biochemical machinery. For the first time in the conscious rat, acetyl-l-carnitine was shown to be involved in feeding both biosynthetic and oxidative acetylCoA pools, and the acetyl moiety appeared to be present in ketone bodies and cholesterol synthesis. Interestingly, these acetylCoA pools appeared to be nonhomogeneous and were probably located in different intracellular compartments, thus suggesting a pivotal role for acetyl-l-carnitine in acetyl group trafficking and compartmentation within cells.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors thank Mrs. L. Mattace for editorial assistance. This work is supported in part by National Research Council (CNR) grant 96.01124.CT03.


  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Bieber, L.L. (1988) Carnitine. Annu. Rev. Biochem. 57, 261283.
  • 2
    Bremer, J. (1983) Carnitine-metabolism and functions. Physiol. Rev. 63, 14201480.
  • 3
    Choi, Y.R., Fogle, P.J., Clarke, P.R. & Bieber, L.L. (1977) Quantitation of water-soluble acylcarnitines and carnitine acyltransferases in rat tissues. J. Biol. Chem. 252, 79307931.
  • 4
    Ohue, M. & Makita, T. (1994) Localization of carnitine acetyltransferase activity in brown adipocytes of the rat. J. Vet. Med. Sci. 5, 329333.
  • 5
    Markwell, M.A.K., Tolbert, N.E. & Bieber, L.L. (1976) Comparison of the carnitine acyltransferase activities from rat liver peroxisomes and microsomes. Arch. Biochem. Biophys. 176, 479488.
  • 6
    Chung, C., Chung, C.D. & Bieber, L.L. (1991) Purification of heart and liver mitochondrial carnitine acetyltransferase. Protein Expr. Purif. 2, 426431.
  • 7
    Aureli, T., Miccheli, A., Ricciolini, R., Di Cocco, M.E., Ramacci, M.T., Angelucci, L., Ghirardi, O. & Conti, F. (1990) Aging brain: effect of acetyl-l-carnitine treatment on rat brain energy and phospholipid metabolism. A study by 31P and 1H NMR spectroscopy. Brain Res. 526, 118112.
  • 8
    Lysiak, W., Lilly, K., DiLisa, F., Toth, P.P. & Bieber, L.L. (1988) Quantitation of the effect of l-carnitine on the levels of acid-soluble short-chain acyl-CoA and CoASH in rat heart and liver mitochondria. J. Biol. Chem. 263, 11511156.
  • 9
    Hagen, T.M., Ingersoll, R.T., Weher, C.M., Lykkesfeldt, J., Vinarsky, V., Bartholomew, J.C., Song, M. & Ames, B.N. (1998) Acetyl-l-carnitine fed to old rats partially restores mitochondrial function and ambulatory activity. Proc. Natl Acad. Sci. USA 95, 95629566.
  • 10
    Sleboda, J., Pourfarzam, M., Bartlett, K. & Osmundsen, H. (1995) Effects of added l-carnitine, acetyl-CoA and CoA on peroxisomal beta-oxidation of [U-14C]hexadecanoate by isolated peroxisomal fractions. Biochim. Biophys. Acta 1258, 309318.
  • 11
    van Roermund, C.W., Elgersma, Y., Singh, N., Wanders, R.J. & Tabak, H.F. (1995) The membrane of peroxisomes in Saccharomyces cerevisiae is impermeable to NAD(H) and acetyl-CoA under in vivo conditions. EMBO J. 14, 34803486.
  • 12
    Bressler, R. & Katz, R.I. (1965) The role of carnitine in acetoacetate production and fatty acid synthesis. J. Clin. Invest. 44, 840848.
  • 13
    Bressler, R. & Katz, R.I. (1965) The effect of carnitine on the rate of incorporation of precursors into fatty acids. J. Biol. Chem. 240, 622627.
  • 14
    Abdel-aleem, S., Sayed-Ahmed, M., Nada, M.A., Hendrickson, S.C., St. Louis, J. & Lowe, J.E. (1995) Stimulation of non-oxidative glucose utilization by l-carnitine in isolated myocytes. J. Mol. Cell. Cardiol. 27, 24652472.
  • 15
    Farrell, S., Vogel, J. & Bieber, L.L. (1986) Entry of acetyl-l-carnitine into biosynthetic pathways. Biochim. Biophys. Acta 876, 175177.
  • 16
    Künnecke, B. (1995) Application of 13C NMR Spectroscopy to metabolic studies on animals. Carbon-13 NMR Spectroscopy of Biological Systems (Beckmann, N., ed.), pp. 226240. Academic Press, San Diego.
  • 17
    Kuipers, F., Dijkstra, T., Havinga, R., van Asselt, W. & Vonk, R.J. (1985) Acute effects of pentobarbital-anaesthesia on bile secretion. Biochem. Pharmacol. 34, 17311736.
  • 18
    Ruckebusch, Y. (1981) Motor functions of the intestine. Adv. Vet. Sci. Comp. Med. 25, 345369.
  • 19
    Miccheli, A., Aureli, T., Delfini, M., Di Cocco, M.E., Viola, P., Gobetto, R. & Conti, F. (1988) Study on influence of inactivation enzyme techniques and extraction procedures on cerebral phosphorylated metabolite levels by 31P and 1H NMR spectroscopy. Cell. Mol. Biol. 34, 591603.
  • 20
    Popjak, G., Edmond, J., Anet, F.A. & Easton, N.R. Jr (1977) Carbon-13 NMR studies on cholesterol biosynthesized from [13C]mevalonates. J. Am. Chem. Soc. 99, 931935.
  • 21
    Badar-Goffer, R.S., Bachelard, H.S. & Morris, P.G. (1990) Cerebral metabolism of acetate and glucose studied by 13C-n.m.r. spectroscopy. A technique for investigating metabolic compartmentation in the brain. Biochem. J. 266, 133139.
  • 22
    Marzo, A., Arrigoni Martelli, E., Urso, R., Rocchetti, M., Rizza, V. & Kelly, J.G. (1989) Metabolism and disposition of intravenously administered acetyl-l-carnitine in healthy volunteers. Eur. J. Clin. Pharmacol. 37, 5963.
  • 23
    Marzo, A., Corsico, N., Cardace, G. & Morabito, E. (1993) Effect of acetyl-L-carnitine treatment on the levels of levocarnitine and its derivatives in streptozotocin-diabetic rats. Arzneimittelforschung 43, 339342.
  • 24
    Molstad, P. (1980) The efflux of l-carnitine from cells in culture (CCL 27). Biochim. Biophys. Acta 597, 166173.
  • 25
    Brass, E.P. (1992) Carnitine transport. L-Carnitine and its Role in Medicine: from Function to Therapy (Ferrari, R., Di Mauro, S. & Sherwood, G., eds), pp. 2136. Academic Press, New York.
  • 26
    Soboll, S., Scholz, R., Freisl, M., Elbers, R. & Heldt, H.W. (1976) Distribution of metabolites between mitochondria and cytosol of perfused liver. Use of Isolated Liver Cells and Kidney Tubules in Metabolic Studies (Tager, J.M., Söling, H.D. & Williamson, J.R., eds), pp. 2940. North-Holland Publishing Co, Amsterdam, the Netherlands.
  • 27
    Müllhofer, G., Schwab, A., Muller, C., Von Stetten, C. & Gruber, E. (1977) Carbon-14 tracer studies in rat-liver perfusion experiments under conditions of gluconeogenesis from lactate and pyruvate. Eur. J. Biochem. 75, 319330.
  • 28
    Baranyai, J. & Blum, J.J. (1989) Quantitative analysis of intermediary metabolism in rat hepatocytes incubated in the presence and absence of ethanol with a substrate mixture including ketoleucine. Biochem. J. 258, 121140.
  • 29
    Lopes-Cardozo, M., Klazinga, W. & van den Bergh, S.G. (1978) Evidence for a homogeneous pool of acetyl-CoA in rat-liver mitochondria. Eur. J. Biochem. 83, 635640.
  • 30
    Des Rosiers, C., David, F., Garneau, M. & Brunengraber, H. (1991) Nonhomogeneous labeling of liver mitochondrial acetyl-CoA. J. Biol. Chem. 266, 15741578.
  • 31
    Ericsson, J., Appelkvist, E.L., Runquist, M. & Dallner, G. (1993) Biosynthesis of dolichol and cholesterol in rat liver peroxisomes. Biochimie 75, 167173.
  • 32
    Biardi, L. & Krisans, S.K. (1996) Compartmentalization of cholesterol biosynthesis. Conversion of mevalonate to farnesyl diphosphate occurs in the peroxisomes. J. Biol. Chem. 271, 17841788.
  • 33
    Thompson, S.L., Burrows, R., Laub, R.J. & Krisans, S.K. (1987) Cholesterol synthesis in rat liver peroxisomes. Conversion of mevalonic acid to cholesterol. J. Biol. Chem. 262, 1742017425.
  • 34
    Endemann, G., Goetz, P.G., Edmond, J. & Brunengraber, H. (1982) Lipogenesis from ketone bodies in the isolated perfused rat liver. Evidence for the cytosolic activation of acetoacetate. J. Biol. Chem. 257, 34343440.
  • 35
    Ito, M., Fukui, T., Kamakari, M., Saito, T. & Tomita, K. (1984) Purification and characterization of acetoacetyl CoA synthetase from rat liver. Biochim. Biophys. Acta 794, 183193.
  • 36
    Hildebrandt, L., Spennetta, T., Ackerman, R., Elson, C. & Shrago, E. (1996) Acyl CoA and lipid synthesis from ketone bodies by the extramitochondrial fraction of hepatoma tissue. Biochem. Biophys. Res. Commun. 225, 307312.
  • 37
    Kelleher, J.K. & Masterson, T.M. (1992) Model equations for condensation biosynthesis using stable isotopes and radioisotopes. Am. J. Physiol. 262, E118E125.
  • 38
    Lligoña-Trulla, L., Arduini, A., Aldaghlas, T.A., Calvani, M. & Kelleher, J.K. (1997) Acetyl-l-carnitine flux to lipids in cells estimated using isotopomer spectral analysis. J. Lipid Res. 38, 14541462.
  • 39
    Kelleher, J.K., Kharroubi, A.T., Aldaghlas, T.A., Shambat, I.B., Kennedy, K.A., Holleran, A.L. & Masterson, T.M. (1994) Isotopomer spectral analysis of cholesterol synthesis: applications in human hepatoma cells. Am. J. Physiol. 266, E384E395.
  • 40
    Coleman, P.S. (1986) Membrane cholesterol and tumor bioenergetics. Ann. NY Acad. Sci. 488, 451467.
  • 41
    Ricciolini, R., Miccheli, A., Peluso, G., Delfini, M. & Conti, F. (1991) 31P and 1H NMR studies of ethanolamine-linked phosphoglycerides metabolism in human T lymphocytes. Cell. Mol. Biol. 37, 705711.
  • 42
    Miziorko, H.M., Laib, F.E. & Behnke, C.E. (1990) Evidence for substrate channeling in the early steps of cholesterogenesis. J. Biol. Chem. 265, 96069609.
  1. Enzymes: carnitine O-acetyltransferase (EC2.3.1.7); hydroxymethylglutarylCoA synthase (EC4.1.3.5); hydroxymethylglutarylCoA reductase NADPH (EC1.1.1.34); diphosphomevalonate decarboxylase (EC4.1.1.33); acetoacetylCoA thiolase (EC2.3.1.9).