Membrane transport of fatty acylcarnitine and free L -carnitine by rat liver microsomes Jason M. Gooding, Majid Shayeghi and E. David Saggerson Department of Biochemistry and Molecular Biology, University College London, UK Recent studies have suggested that parts of the hepatic activities of diacylglycerol acyltransferase and acyl choles- terol acyltransferase are expressed in the lumen of the endoplasmic reticulum (ER). However the ER membrane is impermeable to the long-chain fatty acyl-CoA substrates of these enzymes. Liver microsomal vesicles that were shown to be at least 95% impermeable to palmitoyl-CoA were used to demonstrate the membrane transport of palmitoylcarnitine and free L -carnitine – processes that are necessary for an indirect route of provision of ER luminal fatty acyl-CoA through a luminal carnitine acyltransferase (CAT). Experi- mental conditions and precautions were established to per- mit measurement of the transport of [ 14 C]palmitoylcarnitine into microsomes through the use of the luminal CAT and acyl-CoA:ethanol acyltransferase as a reporter system to detect formation of luminal [ 14 C]palmitoyl-CoA. Rapid, unidirectional transport of free L-[ 3 H]carnitine by micro- somes was measured directly. This process, mediated either by a channel or a carrier, was inhibited by mersalyl but not by N-ethylmaleimide or sulfobetaine – properties that differentiate it from the mitochondrial inner membrane carnitine/acylcarnitine exchange carrier. These findings are relevant to the understanding of processes for the reassembly of triacylglycerols that lipidate very low density lipoprotein particles as part of a hepatic triacylglycerol lipolysis/ re-esterification cycle. Keywords: acylcarnitine; carnitine; microsomes; liver; trans- port. Although mammalian intracellular membranes are imper- meable to the Coenzyme A thioesters of long-chain fatty acids, these activated derivatives, which are synthesized from nonesterified fatty acids by fatty acyl-CoA synthetase on the cytosolic aspect of organelle membranes, are the substrates for metabolic processes within at least three cellular organelles. In mitochondria, it is well established that CPT 1 , a carnitine acyltransferase associated with the outer membrane, generates fatty acylcarnitine derivatives which can then traverse the inner membrane via a carnitine/acylcarnitine exchange carrier (CAC). Within the mitochondrial interior, the latent carnitine acyltrans- ferase CPT 2 then facilitates the re-formation of fatty acyl- CoAs which are then substrates for b-oxidation within the mitochondrial matrix [1]. Fatty acyl-CoAs are also sub- strates for chain-shortening by b-oxidation within the matrix of peroxisomes. As peroxisomes also contain overt and latent carnitine acyltransferase activities [2,3] and express the CAC protein [4], it has been concluded that activated fatty acids access the peroxisomal matrix through a system that is closely analogous to the mitochondrial one. Enzymes within the lumen of the endoplasmic reticulum (ER) also require fatty acyl-CoA thioesters as substrate. A latent form of diacylglycerol acyltransferase (DGAT), assigned to the luminal surface of the ER membrane and which can be differentiated from a cytosolically oriented DGAT, has been described [5–7]. This latent DGAT may be involved in the reassembly of triacylglycerols which lipidate very low density lipoprotein (VLDL) particles as part of a hepatic triacylglycerol lipolysis/re-esterification cycle [2,3,8–12]. Two forms of acyl cholesterol acyltrans- ferase (ACAT) are also known and one of these is suggested to be oriented from the ER membrane towards the lumen and to contribute to provision of cholesteryl esters for lipidation of VLDL particles [13–16]. Finally, acyl-CoA:eth- anol acyltransferase (AEAT) is an enzyme activity 1 that appears to be exclusively localized to the ER lumen [17]. As they do not readily penetrate the ER membrane [17], it has been proposed that the fatty acyl-CoA substrates for luminal enzymes such as DGAT, ACAT and AEAT are generated by a malonyl-CoA-insensitive carnitine acyltransferase (CAT) that is localized in the ER lumen [18,19]. It has been envisaged that the substrate for this luminal CAT is fatty acylcarnitine, which is trans- ported from the cytosol to the ER lumen [2,3,12] (Fig. 1) and which is generated by CPT 1 located at sites distinct from the ER [20] or also by an ER-targeted form of CPT 1 [12,21–23]. For the luminal CAT to function in the Correspondence to E. D. Saggerson, Department of Biochemistry and Molecular Biology, University College London, Gower Street, London, WC1E 6BT, UK. Fax: + 44 20 76797193, Tel.: + 44 20 76797320, E-mail: Saggerson@biochem.ucl.ac.uk Abbreviations: ACAT, acyl cholesterol acyltransferase; AEAT, acyl- CoA:ethanol acyltransferase; CAC, carnitine/acylcarnitine exchange carrier; CAT, carnitine acyltransferase; CPT 1 , the overt carnitine palmitoyltransferase of mitochondria; DGAT, diacylglycerol acyltransferase; ER, endoplasmic reticulum; etomoxir, 2-[6-(4-chlorophenoxy)hexyl]oxirane carboxylic acid; [(Np-O) 2 P i ], bis-(4-nitrophenyl)phosphate; VLDL, very low density lipoprotein. (Received 20 October 2003, revised 8 January 2004, accepted 16 January 2004) Eur. J. Biochem. 271, 954–961 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.03997.x way envisaged in Fig. 1 there must additionally be a means whereby the L -carnitine product can escape from the ER and there must be a supply of ER luminal free CoASH. The CAC protein is not expressed at detectable levels in liver microsomes [4], so at present it is not clear how inward and outward transport of fatty acylcarnitine and L -carnitine, respectively, across the ER membrane are facilitated. In the present study, we report direct evidence for the passage of radiolabelled L -carnitine across micro- somal membranes. We also report on studies of the uptake of [ 14 C]palmitoylcarnitine into the lumen of ÔsealedÕ microsomal vesicles. Direct measurement of this is technically very difficult or impossible as these lipid substrates bind nonspecifically to membrane proteins and lipids. Therefore indirect approaches using Ôreporter systemsÕ must be used. Broadway et al. [12] used the ER luminal coupled system of CAT and AEAT as a ÔreporterÕ in preliminary studies to show that ÔsealedÕ liver microsomes could generate ethyl palmitate from palmi- toylcarnitine provided exogenous free CoASH was pre- sent, and concluded that these findings were consistent with a trans-ER membrane transport of palmitoylcarni- tine together with an ER transport process for CoASH, the cosubstrate for microsomal CAT [12]. In support of the notion of fatty acylcarnitine transport in the ER, Abo-Hashema et al. [24] used the latent DGAT activity as a Ôreporter systemÕ to demonstrate a carnitine-dependent conversion of oleoyl-CoA into luminal triacylglycerol. In those experiments, microsomes were fused with liposomes encapsulating a supply of CoASH for the luminal CAT. In the present study, we have re-evaluated and further developed the experimental approach of Broadway et al. [12] through the use of bis-(4-nitrophe- nyl)phosphate [(Np-O) 2 P i ] which is used to decrease interference from carboxyesterase and thioesterase activ- ities in microsomes [25]. From these studies, we confirm the presence of a system that facilitates the entry of palmitoylcarnitine into ÔsealedÕ microsomes but discount the previous notion of a concomitant transport system for CoASH [12]. Materials and methods Chemicals Routinely used chemicals were from BDH Ltd. (Poole, Dorset, UK) or from Sigma Chemical Co. Ltd. (Poole, Dorset, UK). The sodium salt of etomoxir{2-[6-(4-chloro- phenoxy)hexyl]oxirane carboxylic acid} was from H. P. O. Wolf, (Projekt-Entwicklung GmbH, Allensbach, Germany). [ 3 H]Carnitine, [ 14 C]inulin, [ 14 C]palmitoylcarni- tine and [ 14 C]palmitoyl-CoA were from Amersham Inter- national (Little Chalfont, Bucks, UK). The radiolabelled palmitoylcarnitine was supplied in a 50 : 50 (v/v) water/ ethanol solution which was removed by evaporation in a stream of nitrogen before use. This ensured that no ethanol was present in ethanol-independent acylation assays. Ethyl[ 14 C]palmitate was synthesized as described by Diczfalusy et al. [25]. Isolation of microsomes from rat liver 2,3 Fed male Sprague–Dawley rats (180–220 g) were killed 2,3 by an approved schedule/method [schedule/method approved by the UK Home Office under the Act of Parliament ÔAnimals (Scientific Procedures) Act 1986Õ and the local UCL Animal Experimentation Ethics Committee]. Livers were immedi- ately removed and washed in ice-cold isolation medium (5 m M Tris/HCl buffer, pH 7.4, containing 220 m M mann- itol, 80 m M sucrose, 1 m M EDTA, 5 lgÆmL )1 bestatin, 5 lgÆmL )1 leupeptin and 1 m M phenylmethanesulfonyl fluoride) and then homogenized in 4 vols of the same buffer with four strokes of a motor-driven Potter-type homogeniser. The homogenate was centrifuged at 500 g for 10 min; the supernatant was then centrifuged at 10 000 g for 20 min followed by centrifugation of the 10 000 g super- natant at 20 000 g for 20 min. The resulting 20 000 g supernatant was then ultra-centrifuged at 100 000 g for 60 min. The pellet was resuspended in fresh isolation medium and recentrifuged at 100 000 g for 60 min. The microsomal pellet was then resuspended (80–100 mgÆmL )1 ) and stored at )70 °Cas100 lL aliquots in HS buffer (5 m M Hepes buffer, pH 7.4 containing 250 m M sucrose). The specific activity of cytocrome c oxidase (a mitochondrial inner membrane marker) was less than 1% of that found in mitochondrial fractions (results not shown). Similar micro- somal fractions have previously been shown to contain very little activity of monoamine oxidase (a mitochondrial outer membrane marker) [19]. In some instances 4 m M CoASH with 4 m M dithiothre- itol were included in isolation medium and in HS buffer. Microsomal vesicles described as ÔsealedÕ were obtained by thawing stored aliquots and suspending at the required concentrationinHSbuffer. The intactness of microsomal preparations was routinely assessed by the latency of mannose 6-phosphatase [26] which exceeded 90%. Further evidence of the intactness of micro- somal vesicles was obtained from measurements of AEAT activity (see below). Microsomal vesicles described as ÔpermeabilisedÕ were similarly suspended in HS buffer to which the pore-forming antibiotic alamethicin [27–29] had been added from a stock solution of 20 mgÆmL )1 dissolved in dimethyl sulfoxide. At the concentration used (15 lgÆmL )1 ) Fig. 1. A scheme summarizing processes to provide the fatty acyl-CoA substrate for AEAT, DGAT and ACAT in the ER lumen. Ó FEBS 2004 Transport of carnitine and acylcarnitine (Eur. J. Biochem. 271) 955 we found that alamethicin abolished latency of mannose 6-phosphatase essentially immediately (results not shown). Treatment of microsomes with etomoxir Microsomes were diluted to approximately 10 mgÆmL )1 in HS buffer containing 5 m M ATP, 10 m M MgCl 2 ,2m M CoASH, 50 l M sodium etomoxir and BSA (1 mgÆmL )1 ) and incubated at 25 °C for 30 min followed by addition of 4 vols of ice-cold HS buffer. The diluted microsomes were then centrifuged at 200 000 g for 30 min, resuspended in HS buffer (approximately 50 mgÆmL )1 ) and stored at )70 °C. Assay of ethanol-dependent and -independent acyltransferase activities Assays were performed either with sealed microsomes or in the permeabilized state (with alamethicin [15 lgÆmL )1 ]) at 30 °C in a final volume of 1.4 mL in 10 m M Tris/HCl buffer (pH 7.4) containing 300 m M sucrose, 10 m M MgCl 2 , 0.8 m M EDTA, bovine albumin (1 mgÆmL )1 )and40l M [ 14 C]palmitoyl-CoA (1.8 lCiÆlmol )1 ) with or without 10 lL of ethanol (123 m M final concentration). Where indicated, 250 l M (Np-O) 2 P i [25] was included in assays. Assays were initiated by the addition of 700 lg of microsomal protein. At zero, 1, 2, 5, 10, 15 and 20 min 200 lL samples of the assay mixture were removed and mixed with 1.5 mL propan-2-ol/heptane/water (80 : 20 : 2; v/v/v). After mixing with a further 1 mL of heptane and 0.5 mL of water the upper (heptane) layer was removed and washed with 2 mL of 50 m M NaOH dissolved in 50% (v/v) ethanol. Washed heptane layer (650 lL) was taken for liquid scintillation counting. AEAT (ethanol-dependent) activity was calcula- ted by subtracting 14 C-labelled product formation in the absence of ethanol from that in its presence. Measurement of AEAT activity in this way was found in preliminary experiments to agree with the formation of ethyl[ 14 C]pal- mitate as detected by TLC [25]. In experiments in which AEAT was used as a reporter enzyme to detect entry of palmitoylcarnitine into sealed microsomal vesicles, [ 14 C]palmitoyl-CoA was replaced by 40 l M [ 14 C]palmitoyl- carnitine (1.8 lCiÆlmol )1 ) with or without 0.5 m M CoASH and 1 m M dithiothreitol. Measurement of uptake of L -[ 3 H]carnitine into microsomal vesicles Assays were performed at 10 °C and were commenced by the addition of 500 lL of microsomes in HS buffer (60–80 mgÆmL )1 protein, pre-equilibrated at 10 °C) to 500 lL of pre-equilibrated HS buffer containing various concentrations of unlabelled carnitine, 5 lCi [ 3 H]carnitine and 2 lCi [ 14 C]inulin. At each time-point, two 50 lL samples were removed quickly and transferred to tubes containing 2 mL of ice-cold HS buffer containing polyethy- lene glycol 8000 (5%, w/v). One of each pair of tubes additionally contained 200 lg of alamethicin to permeablise the microsomal vesicles. Both tubes were immediately centrifuged for 60 s at 6000 g to sediment the microsomal material. The supernatants were removed by aspiration and the walls of the tubes wiped with tissue to remove adhering fluid. The pellets were dissolved in 500 lL of Triton X-100 (10%, v/v), transferred to scintillation fluid and 3 Hand 14 C measured by liquid scintillation counting. The amount of [ 3 H]carnitine within microsomal vesicles at each time-point was calculated by subtracting the 3 H-associated with the alamethicin-treated pellet from the 3 H-associated with the untreated pellet after correcting for adhering medium, which was determined from the amount of associated [ 14 C]inulin. Measurement of protein content This was by a bicinchoninic acid assay kit (Sigma). Statistical methods Values are shown in figures as means of the number of separate measurements (n) ± SD. Where SD bars are not shown in figures, these are within the symbol. Results Transport of palmitoylcarnitine into sealed microsomal vesicles TheexperimentsshowninFig.2wereperformedinthe absence of (Np-O) 2 P i . Figure 2A confirms the previously reported high degree of latency of AEAT [5,17] in that minimal formation of ethanol-dependent product (ethyl palmitate) from [ 14 C]palmitoyl-CoA was seen with sealed microsomes whereas permeabilization by alamethicin allowed ethyl palmitate formation at an initial rate of 1020 ± 30 pmolÆmin )1 Æmg )1 .When[ 14 C]palmitoylcarnitine was provided alone, there was no appreciable formation of ethyl palmitate by sealed microsomes (Fig. 2B). A similar lack of ethyl palmitate formation from palmitoylcarnitine was seen with alamethicin-permeabilized microsomes, con- firming that palmitoylcarnitine is not a substrate for AEAT (data not shown). However, when CoASH was also present (together with dithiothreitol to keep coenzyme A in the reduced form) ethyl palmitate formation from [ 14 C]palmi- toylcarnitine by sealed microsomes was observed at a steady Fig. 2. Radiolabelled product formation by microsomes in the absence of (Np-O) 2 P i . All values are means ± SD of four independent meas- urements. (A) Ethanol-dependent product formation from 40 l M [ 14 C]palmitoyl-CoA. h, sealed microsomes; j, microsomes permea- bilized by alamethicin. (B) Product formation from 40 l M [ 14 C]palmitoylcarnitine by sealed microsomes. s, ethanol-dependent without CoASH and dithiothreitol; j, ethanol-dependent with 0.5 m M CoASH + 1 m M dithiothreitol. 956 J. M. Gooding et al. (Eur. J. Biochem. 271) Ó FEBS 2004 rate of 180 ± 20 pmolÆmin )1 Æmg )1 over 10 min (Fig. 2B). From experiments such as these, Broadway et al. [12] arrived at the conclusion that although palmitoyl-CoA could not gain access to the interior of sealed microsomes, both palmitoylcarnitine and CoASH could do so, thereby becoming substrates for the coupled microsomal CAT/ AEAT system – it is reasonable to expect that diffusion across the microsomal membrane of ethanol, the cosub- strate for AEAT, is not rate-limiting [30]. Figure 2A also shows that ethyl palmitate formation from [ 14 C]palmitoyl-CoA by alamethicin-permeabilized microsomes plateaued and then declined after 10 min. As only  10% of the available 14 C was in ethyl palmitate or in ethanol-independent products (not shown) at this 10 min time interval, extensive destruction of the product(s) and/or of the substrate must have occurred. To a lesser extent, the same is suggested by Fig. 2B where the timecourse of product formation from palmitoylcarnitine became nonlin- ear after only  5% of the available 14 C could be detected in ethanol-dependent or -independent products. Diczfalusy et al. [25] have shown that the serine esterase inhibitor (Np-O) 2 P i at 250 l M enhanced AEAT activity in rat liver microsomes by approximately fivefold. This effect was primarily explained through inhibition of carboxyesterase ES-4, which has both ethyl esterase and thioesterase activities. We therefore reinvestigated the formation of ethyl palmitate from [ 14 C]palmitoyl-CoA and [ 14 C]palmi- toylcarnitine in the presence of 250 l M (Np-O) 2 P i (Figs 3 and 4). Ethyl palmitate formation from [ 14 C]palmitoyl-CoA was considerably enhanced by (Np-O) 2 P i in alamethicin- permeabilized microsomes (compare Fig. 2A and 3A) and also facilitated the detection of a small but significant amount of this conversion by sealed microsomes (Fig. 3A). However, even though  40% of the available [ 14 C]palmi- toyl-group now appeared in ethyl palmitate after 5 min with permeabilized microsomes, (Np-O) 2 P i did not totally abol- ish the destruction of this product (Fig. 3A). Reasonable estimates of initial rates of ethyl palmitate formation (AEAT activity) can be made from Fig. 3A. These were 480 ± 52 pmolÆmin )1 Æmg )1 for sealed microsomes and 18.4 ± 1.2 nmolÆmin )1 Æmg )1 for those permeabilized by alamethicin, i.e. 97.5% of AEAT activity was latent, indicating that these experiments were conducted with vesicles that had a high degree of ÔintactnessÕ.Itwas important to establish this because palmitoyl-CoA, albeit at higher concentrations in the absence of BSA, permeabilizes rat liver microsomes [31]. AEAT is inhibited by high concentrations of its fatty acyl-CoA substrate [25]. Figure 3B shows that our experimental conditions employed a palmitoyl-CoA concentration (40 l M )that was not inhibitory. Also in general (not shown), we observed that ethanol-independent product formation from [ 14 C]palmitoyl-CoA or [ 14 C]palmitoylcarnitine was hardly increased by permeabilization of the microsomes, suggesting that these unspecified products are largely made by enzymes oriented towards the cytosolic face of the ER. In the presence of 250 l M (Np-O) 2 P i ,ratesofformation of ethanol-dependent and -independent products from 40 l M [ 14 C]palmitoyl-CoA or [ 14 C]palmitoylcarnitine by sealed microsomes were consistently linear from 1 to 5 min and usually were linear from 1 to 10 min. Figure 4 shows values obtained within this linear range. Under these experimental conditions, rates of ethanol-dependent and -independent product formation of 548 ± 34 and 148 ± 25 pmolÆmin )1 Æmg )1 , respectively, were seen when thesubstratewas40l M [ 14 C]palmitoylcarnitine with CoASH/dithiothreitol (essentially zero product formation was seen in the absence of CoASH/dithiothreitol – results not shown). CoASH/dithiothreitol had minimal effect on ethanol-dependent product formation from [ 14 C]palmitoyl- CoA by sealed (Fig. 4) or permeabilized (not shown) microsomes. When microsomes were pretreated with etom- oxir, ATP and CoASH in order to generate etomoxiryl- CoA (an irreversible inhibitor of CPT 1 in microsomal fractions [19]) AEAT activity measured in alamethicin- permeabilized microsomes was not inactivated (results not Fig. 3. Radiolabelled product formation from palmitoyl-CoA by microsomes in the presence of 250 l M (Np-O) 2 P i . All values are means ± SD of four independent measurements. (A) shows ethanol- dependent product formation from 40 l M [ 14 C]palmitoyl-CoA. h, Sealed microsomes; j, microsomes permeabilized by alamethicin. (B) Dependence of product formation by alamethicin-permeabilized microsomes on the concentration of [ 14 C]palmitoyl-CoA. Values are calculated from the difference between zero and 2 min time-points. j, Ethanol-dependent products; h, ethanol-independent products. Fig. 4. Radiolabelled product formation by sealed microsomes in the presence of 250 l M (Np-O) 2 P i . All values are means ± SD of four independent measurements calculated from the difference between 1 and 5 min time-points. Product formation was measured from 40 l M [ 14 C]palmitoyl-CoA or [ 14 C]palmitoylcarnitine with additions of 0.5 m M CoASH + 1 m M dithiothreitol (DTT) as indicated. In some instances, microsomes were pretreated with etomoxir to inactivate CPT 1 (Materials and methods). Open bars, ethanol-independent product formation; solid bars, ethanol-dependent product formation; cross-hatched bar, ethanol-dependent product formation by CoASH- loaded microsomes (Materials and methods). Ó FEBS 2004 Transport of carnitine and acylcarnitine (Eur. J. Biochem. 271) 957 shown) and there was only a small increase in overt AEAT activity, suggesting that the pretreatment caused little increase in leakiness of the microsomes (Fig. 4). Inactivation of CPT 1 by etomoxiryl-CoA totally inhibited the formation of ethanol-dependent and -independent products from [ 14 C]palmitoylcarnitine. This suggested that the effect of CoASH/dithiothreitol to facilitate formation of ethanol- dependent product from [ 14 C]palmitoylcarnitine, in contra- diction of [12], cannot be explained by CoASH being transported into the ER lumen to provide a cosubstrate for the ER luminal CAT. Rather, it strongly suggested that this product formation was due to conversion of [ 14 C]palmi- toylcarnitine by CPT 1 (either an integral component of the ER membrane [21] or a contaminant arising from mito- chondrial contact sites [20]) to external [ 14 C]palmitoyl-CoA, which is then converted into ethyl palmitate by the small amount of AEAT activity that is overt because of incom- plete sealing of the vesicles. Free CoASH is not detectable in rat liver microsomal fractions [32] implying that any luminal pool of CoASH is lost during tissue extraction and/or fractionation. In order to attempt to make CoASH available to the interior of the microsomal vesicles, some preparations were isolated with CoASH present (Materials and methods). After pretreat- ment with etomoxiryl-CoA, these microsomes were still relatively sealed, as evidenced by an overt AEAT activity of only 972 ± 88 pmolÆmin )1 Æmg )1 (indicating 95% latency of AEAT) and had complete inactivation of CPT 1 ,as evidenced by the lack of any ethanol-independent product formation from [ 14 C]palmitoylcarnitine (Fig. 4). However, [ 14 C]palmitoylcarnitine was converted into ethanol-depend- ent product at a rate of 939 ± 188 pmolÆmin )1 Æmg )1 (Fig. 4). Overall, these results support the notion that palmitoylcarnitine can be transported to the ER lumen where it can become a substrate for the microsomal CAT/ AEAT. However, externally derived CoASH cannot play any significant role in the CAT reaction, which must rely on an internal, luminal pool of CoASH. Transport of L -carnitine into sealed microsomal vesicles Transport of L -[ 3 H]carnitine was studied by measurement of uptake into microsomal vesicles at 10 °C. Figure 5 shows initial experiments in which 2 m M carnitine was used. Because of the time needed to separate microsomes from the incubation medium, it was not feasible to measure uptake at times earlier than 2 min. With 2 m M carnitine at 10 °C the uptake reached  60% of the equilibrium value at 2 min allowing only a crude minimal estimate that the initial rate of unidirectional uptake was at least 0.62 nmolÆmin )1 Æmg )1 .MersalylandN-ethyl maleimide are known to inactivate the mitochondrial CAC [33,34]. Mersalyl at 0.5 m M caused  40% inhibition of carnitine uptake at 2 min and 5 m M mersalyl almost abolished uptake (Fig. 5). However, we found that neither N-ethyl maleimide nor sulfobetaine (which is a competitive inhibitor of the mitochondrial CAC [35]) at concentrations up to 5 m M had any effect on carnitine uptake by microsomes (results not shown). We attempted to study the concentration dependence of L -[ 3 H]carnitine uptake (Fig. 6). These experiments appeared to show that the rate of uptake increased linearly with carnitine up to 10 m M , i.e. there was no indication of saturation of the process. However, as indicated above, initial rates determined from the first time-point can only be regarded as crude minimum estimates of the true initial rates of unidirectional uptake. These were 0.034 ± 0.003 and 2.9 ± 0.6 nmolÆmin )1 Æmg )1 at 0.1 and 10 m M carnitine, respectively. In other experiments (results not shown), microsomes were preloaded with 2 m M unlabelled L -carni- tine. This had no effect on the time profile of subsequently measured uptake of [ 3 H]carnitine, suggesting that an exchange carrier identical or similar to the mitochondrial CAC was not involved. Data showing unidirectional import of L -carnitine by the purified mitochondrial CAC in a reconstituted system suggest a V max forthisprocessat10°Cof 2nmolÆ min )1 Æmg )1 with a K m of 0.53 m M (the rate of exchange transport was much faster) [36]. Even relative to total Fig. 5. Effect of mersalyl on uptake of L -[ 3 H]carnitine by sealed microsomes. Sealed microsomes were incubated with 2 m M L-[ 3 H]carnitine and uptake measured as described under Materials and methods. Values are means ± SD of four independent experiments. h, No mersalyl; j,0.5m M mersalyl; s,5m M mersalyl. Fig. 6. Effect of L -carnitine concentration on L -[ 3 H]carnitine uptake by sealed microsomes. Sealed microsomes were incubated with L -[ 3 H]carnitine and uptake measured as described under Materials and methods. Values are means ± SD of four independent observations. L -carnitine concentrations were: h,0.1m M ; j,0.25m M ; s,0.5m M ; d,1.0m M ; n,2.0m M ; m,10m M . 958 J. M. Gooding et al. (Eur. J. Biochem. 271) Ó FEBS 2004 microsomal protein we have observed comparable or higher rates of unidirectional transport than 2 nmolÆmin )1 Æmg )1 . This further differentiates the microsomal process from the mitochondrial CAC. Figure 7 shows a linear plot of [ 3 H]carnitine uptake at equilibrium (60 min values) vs. the L -carnitine concentra- tion. The slope of this line, which represents the Ôcarnitine spaceÕ of the microsomal vesicle preparation, was 1.03 lLÆmg )1 . In other experiments (not shown), we found the internal H 2 O space to be 2.4 ± 0.2 lLÆmg )1 [3]. 4 Therefore, only 43% of the vesicles present in the micro- somal fraction appeared to contain the capability of L -carnitine transport. Discussion The findings of this study provide a framework whereby, acting in concert with the ER CAT, enzymes in the ER lumen (e.g. DGAT, ACAT or AEAT) are supplied with their fatty acyl-CoA substrate via a fatty acylcarnitine intermediate and the resulting free L -carnitine product can be disposed of (Fig. 1). As AEAT activity appears to be totally localized to the interior of the ER in cells, it, together with the luminal CAT, provides an easily quantifiable (via assay) and unambiguous 5 reporter of the entry of fatty acylcarnitine into the ER lumen provided certain experimental conditions are met, i.e., inactivation of CPT 1 , minimization of interference from esterases and provision of a source of luminal CoASH either as described here or by delivery from liposomes [24]. As far as we are aware, this study presents the first direct demonstration of fatty acylcarnitine transport across microsomal membranes. After taking precautions to minimize the ambiguity inherent in using DGAT as the reporter enzyme, Abo- Hashema et al. observed carnitine-dependent incorporation of [ 14 C]oleoyl-CoA into liver microsomal luminal triacyl- glycerol to the extent of 11.58 nmolÆmg )1 over an incubation period of 40 min at 37 °C (290 pmolÆmin )1 Æ mg )1 ) [24]. Initial rates were not determined [24] and so it is not known to what extent this product formation was limited by esterase activity or by the necessity to involve CPT 1 as an additional enzyme in the system prior to the membrane transport of the fatty acylcarnitine. With 40 l M [ 14 C]palmi- toylcarnitine, we observed formation of ethyl palmitate that was linear with time at a rate of 939 pmolÆmin )1 Æmg )1 at 30 °C, provided microsomes were isolated previously with CoASH present (Fig. 4). Activities of rat liver microsomal luminal CAT (2.7–3.6 nmolÆmin )1 Æmg )1 at 25 °C[19])and AEAT (18.4 nmolÆmin )1 Æmg )1 at 30 °C; see Fig. 3A) exceed this value of 939 pmolÆmin )1 Æmg )1 suggesting that it is likely to be a reasonable estimate of the rate of palmitoylcarnitine transport. It is of note that measurements of rat liver microsomal latent DGAT activity (0.47–1.9 nmolÆmin )1 Æ mg )1 at 37 °C [5,7]) are not dissimilar from our measure- ment of the rate of palmitoylcarnitine transport. As we discuss below, it is highly unlikely that the CAC protein plays any role in the ER membrane. At present, this microsomal transport process for fatty acylcarnitine awaits characterization in future studies which should also investi- gate whether it makes a significant contribution to the control of metabolic flux to luminal enzymes such as DGAT and ACAT. Our demonstration that L -carnitine can move across the membrane of microsomal vesicles that have a high degree of ÔintactnessÕ as judged by enzyme latency is a significant finding as a key feature of the scheme in Fig. 1 is that free carnitine should be able to escape from the ER after its generation by the luminal CAT. The observation that microsomal carnitine transport is sensitive to mersalyl (Fig. 5) suggests that the process is protein-mediated rather than being a simple diffusion process (the extreme hydro- philicity of carnitine also makes this highly unlikely). The lack of sensitivity to N-ethyl maleimide and sulf- obetaine and the apparent lack of requirement for a counter-transport partner for carnitine discriminates this microsomal process from carnitine/acylcarnitine transport in mitochondria – findings that are not at variance with the report of Fraser & Zammit that the CAC protein is not detectable in liver microsomes [4]. Our attempts at a kinetic analysis (Fig. 6) could not differentiate between a carrier or a channel that facilitates the rapid equilibrium of carnitine across the microsomal membrane. In this regard, it is of note that there have been reports [37–39] of a microsomal membrane channel that permits the passage of certain small molecules, some similar to carnitine (e.g. choline [39]). The microsomal isotope space for choline of 1.05 lLÆmg )1 reported by Meissner & Allen [39] is remarkably similar to the space of 1.03 lLÆmg )1 that we found for L -carnitine (Fig. 7). Further studies are needed to characterize the microsomal L -carnitine process. Finally, the need for an ER luminal pool of CoASH (Fig. 1) is demonstrated by this study. However, questions regarding the source of this pool and how it is maintained in vivo remain totally unanswered. Acknowledgements We are grateful to the Medical Research Council and to the Wellcome Trust for support. Fig. 7. Content of L -[ 3 H]carnitine by sealed microsomes at equilibrium. Equilibrium was reached by incubation with L -[ 3 H]carnitine for 60 min (Fig. 6). Values are means ± SD of four independent experi- ments. Symbols indicating L -carnitine concentrations are the same as those in Fig. 6 except that values with 10 m ML -carnitine, which also lie on the line, are omitted for reasons of scale. An intramicrosomal space for L -carnitine (1.03 lLÆmg )1 ) is obtained from the slope of the line. Ó FEBS 2004 Transport of carnitine and acylcarnitine (Eur. J. Biochem. 271) 959 References 1. McGarry, J.D. & Brown, N.F. (1997) The mitochondrial carnitine palmitoyltransferase system. From concept to molecular analysis. Eur. J. Biochem. 244, 1–14. 2. Zammit, V.A. (1999) Carnitine acyltransferases: functional signi- ficance of subcellular distribution and membrane topology. Prog. Lipid Res. 38, 199–224. 3. Zammit, V.A. (1999) The malonyl-CoA-long-chain acyl-CoA axis in the maintenance of mammalian cell function. Biochem. J. 343, 505–515. 4. Fraser, F. & Zammit, V.A. (1999) Submitochondrial and sub- cellular distributions of the carnitine-acylcarnitine carrier. FEBS Lett. 445, 41–44. 5. Owen, M.R., Corstorphine, C.C. & Zammit, V.A. (1997) Overt and latent activities of diacylglycerol acytransferase in rat liver microsomes: possible roles in very-low-density lipoprotein tri- acylglycerol secretion. Biochem. J. 323, 17–21. 6. Waterman, I.J. & Zammit, V.A. (2002) Activities of overt and latent diacylglycerol acyltransferases (DGATs I and II) in liver microsomes of ob/ob mice. Int. J. Obes. Relat. Metab. Disord. 26, 742–743. 7. Waterman, I.J. & Zammit, V.A. (2002) Differential effects of fenofibrate or simvastatin treatment of rats on hepatic microsomal overt and latent diacylglycerol acyltransferase activities. Diabetes 51, 1708–1713. 8. Gibbons, G.F., Bartlett, S.M., Sparks, C.E. & Sparks, J.D. (1992) Extracellular fatty acids are not utilized directly for the synthesis of very-low-density lipoprotein in primary cultures of rat hepato- cytes. Biochem. J. 287, 749–753. 9. Gibbons, G.F., Islam, K. & Pease, R.J. (2000) Mobilisation of triacylglycerol stores. Biochim. Biophys. Acta 1483, 37–57. 10. Gibbons, G.F. & Wiggins, D. (1994) Intracellular triacylglycerol lipase: its role in the assembly of hepatic very-low-density lipo- protein (VLDL). Adv. Enzyme Regul. 35, 179–198. 11. Yang, L.Y., Kuksis, A., Myher, J.J. & Steiner, G. (1995) Origin of triacylglycerol moiety of plasma very low density lipoproteins in the rat: structural studies. J. Lipid Res. 36, 125–136. 12. Broadway, N.M., Gooding, J.M. & Saggerson, E.D. (1999) Car- nitine acyltransferases and associated transport processes in the endoplasmic reticulum. Missing links in the VLDL story? Adv. Exp Medical Biol. 466, 59–67. 13. Joyce, C., Skinner, K., Anderson, R.A. & Rudel, L.L. (1999) Acyl- coenzyme A: cholesteryl acyltransferase 2. Curr. Opin. Lipidol. 10, 89–95. 14. Rudel, L.L., Lee, R.G. & Cockman, T.L. (2001) Acyl coenzyme A: cholesterol acyltransferase types 1 and 2: structure and function in atherosclerosis. Curr. Opin. Lipidol. 12, 121–127. 15. Buhman, K.F., Accad, M. & Farese, R.V. (2000) Mammalian acyl-CoA: cholesterol acyltransferases. Biochim. Biophys. Acta 1529, 142–154. 16. Joyce, C.W., Shelness, G.S., Davis, M.A., Lee, R.G., Skinner, K., Anderson, R.A. & Rudel, L.L. (2000) ACAT1 and ACAT2 membrane topology segregates a serine residue essential for activity to opposite sides of the endoplasmic reticulum membrane. Mol. Biol. Cell. 11, 3675–3687. 17. Polokoff, M.A. & Bell, R.M. (1978) Limited palmitoyl- CoA penetration into microsomal vesicles as evidenced by a highly latent ethanol acyltransferase activity. J. Biol. Chem. 253, 7173–7178. 18. Murthy, M.S. & Pande, S.V. (1994) Malonyl-CoA-sensitive and -insensitive carnitine palmitoyltransferase activities of micro- somes are due to different proteins. J. Biol. Chem. 269, 18283– 18286. 19. Broadway, N.M. & Saggerson, E.D. (1995) Solubilization and separation of two distinct carnitine acyltransferases from hepatic microsomes: characterization of the malonyl-CoA-sensitive enzyme. Biochem. J. 310, 989–995. 20. Broadway, N.M., Pease, R.J., Birdsey, G., Shayeghi, M., Turner, N.A. & David Saggerson, E. (2003) The liver isoform of carnitine palmitoyltransferase 1 is not targeted to the endoplasmic reticulum. Biochem. J. 370, 223–231. 21. Fraser, F., Corstorphine, C.G., Price, N.T. & Zammit, V.A. (1999) Evidence that carnitine palmitoyltransferase I (CPT I) is expressed in microsomes and peroxisomes of rat liver. Distinct immunoreactivity of the N-terminal domain of the microsomal protein. FEBS Lett. 446, 69–74. 22. Cohen, I., Guillerault, F., Girard, J. & Prip-Buus, C. (2001) The N-terminal domain of rat liver carnitine palmitoyltransferase 1 contains an internal mitochondrial import signal and residues essential for folding of its C-terminal catalytic domain. J. Biol. Chem. 276, 5403–5411. 23. Prip-Buus,C.,Cohen,I.,Kohl,C.,Esser,V.,McGarry,J.D.& Girard, J. (1998) Topological and functional analysis of the rat liver carnitine palmitoyltransferase 1 expressed in Saccharomyces cerevisiae. FEBS Lett. 429, 173–178. 24. Abo-Hashema, K.A., Cake, M.H., Power, G.W. & Clarke, D. (1999) Evidence for triacylglycerol synthesis in the lumen of microsomes via a lipolysis-esterification pathway involving carni- tine acyltransferases. J. Biol. Chem. 274, 35577–35582. 25. Diczfalusy, M.A., Bjorkhem, I., Einarsson, C. & Alexson, S.E. (1999) Formation of fatty acid ethyl esters in rat liver microsomes. Evidence for a key role for acyl-CoA: ethanol O-acyltransferase. Eur. J. Biochem. 259, 404–411. 26. Jones, L.R., Maddock, S.W. & Besch, H.R. Jr (1980) Unmasking effect of alamethicin on the (Na+,K+) -ATPase, beta-adrenergic receptor-coupled adenylate cyclase, and cAMP-dependent protein kinase activities of cardiac sarcolemmal vesicles. J. Biol. Chem. 255, 9971–9980. 27. Nagaraj, R. & Balaram, P. (1981) Alamethicin, a transmembrane channel. Acc. Chem. Res. 14, 356–362. 28. Ojcius, D.M. & Young, J.D. (1991) Cytolytic pore-forming pro- teins and peptides: is there a common structural motif? Trends Biochem. Sci. 16, 225–229. 29. Fulceri, R., Banhegyi, G., Gamberucci, A., Giunti, R., Mandl, J. & Benedetti, A. (1994) Evidence for the intraluminal positioning of p-nitrophenol UDP-glucuronosyltransferase activity in rat liver microsomal vesicles. Arch. Biochem. Biophys. 309, 43–46. 30. Pederson, B.A., Foster, J.D. & Nordlie, R.C. (1998) Low-Km mannose-6-phosphatase as a criterion for microsomal integrity. Biochem. Cell Biol. 76, 115–124. 31. Banhegyi, G., Csala, M., Mandl, J., Burchell, A., Burchell, B., Marcolongo, P., Fulceri, R. & Benedetti, A. (1996) Fatty acyl- CoA esters and the permeability of rat liver microsomal vesicles. Biochem. J. 320, 343–344. 32. Garland, P.B., Shepherd, D. & Yates, D.W. (1965) Steady-state concentrations of coenzyme A, acetyl-coenzyme A and long-chain fatty acyl-Coenzyme A in rat-liver mitochondria oxidizing pal- mitate. Biochem. J. 97, 587–594. 33. Pande, S.V. & Parvin, R. (1976) Characterization of carnitine acylcarnitine translocase system of heart mitochondria. J. Biol. Chem. 251, 6683–6691. 34. Noel, H., Goswami, T. & Pande, S.V. (1985) Solubilization and reconstitution of rat liver mitochondrial carnitine acylcarnitine translocase. Biochemistry 24, 4504–4509. 35. Murthy, M.S. & Pande, S.V. (1984) Mechanism of carnitine acylcarnitine translocase-catalyzed import of acylcarnitines into mitochondria. J. Biol. Chem. 259, 9082–9089. 36. Indiveri, C., Tonazzi, A. & Palmieri, F. (1991) Characterization of the unidirectional transport of carnitine catalyzed by the recon- stituted carnitine carrier from rat liver mitochondria. Biochim. Biophys. Acta 1069, 110–116. 960 J. M. Gooding et al. (Eur. J. Biochem. 271) Ó FEBS 2004 37. Nilsson, R., Peterson, E. & Dallner, G. (1973) Permeability of microsomal membranes isolated from rat liver. J. Cell Biol. 56, 762–776. 38. Ballas, L.M. & Arion, W.J. (1977) Measurement of glucose 6-phosphate penetration into liver microsomes. Confirmation of substrate transport in the glucose-6-phosphatase system, J. Biol. Chem. 252, 8512–8518. 39. Meissner, G. & Allen, R. (1981) Evidence for two types of rat liver microsomes with differing permeability to glucose and other small molecules. J. Biol. Chem. 256, 6413–6422. Ó FEBS 2004 Transport of carnitine and acylcarnitine (Eur. J. Biochem. 271) 961 . Membrane transport of fatty acylcarnitine and free L -carnitine by rat liver microsomes Jason M. Gooding, Majid Shayeghi and E. David Saggerson Department. to demonstrate the membrane transport of palmitoylcarnitine and free L -carnitine – processes that are necessary for an indirect route of provision of ER luminal