Molecular enzymology of carnitine transfer and transport

Review

Molecular enzymology of carnitine transfer and transport

Rona R. Ramsay a;*, Richard D. Gandour b, Feike R. van der Leij c

a Centre for Biomolecular Sciences, University of St Andrews, North Haugh, St. Andrews KY16 9ST, UKb Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0212, USA

c Department of Pediatrics, University of Groningen, Beatrix Children's Hospital and Groningen University Institute for Drug Exploration,P.O. Box 30001, 9700 RB Groningen, The Netherlands

Received 18 September 2000; received in revised form 4 January 2001; accepted 17 January 2001

Abstract

Carnitine (L-3-hydroxy-4-N-trimethylaminobutyric acid) forms esters with a wide range of acyl groups and functions totransport and excrete these groups. It is found in most cells at millimolar levels after uptake via the sodium-dependentcarrier, OCTN2. The acylation state of the mobile carnitine pool is linked to that of the limited and compartmentalisedcoenzyme A pools by the action of the family of carnitine acyltransferases and the mitochondrial membrane transporter,CACT. The genes and sequences of the carriers and the acyltransferases are reviewed along with mutations that affectactivity. After summarising the accepted enzymatic background, recent molecular studies on the carnitine acyltransferasesare described to provide a picture of the role and function of these freely reversible enzymes. The kinetic and chemicalmechanisms are also discussed in relation to the different inhibitors under study for their potential to control diseases of lipidmetabolism. ß 2001 Elsevier Science B.V. All rights reserved.

Keywords: Carnitine; OCTN2; CACT; Carnitine acyltransferases; Fatty acid oxidation; Transport; Molecular genetics; Enzymology;Inhibitor

1. Introduction

The cell relies on the carnitine system (Fig. 1) toregulate the localised, limited pools of CoA deriva-tives. Acyl-CoA pools provide activated substratesfor many key metabolic pathways such as the TCAcycle and lipid and cholesterol synthesis, for post-translational modi¢cation of proteins and for detox-i¢cation mechanisms. The reversible transfer of acti-vated acyl groups from the limited pools of mem-

brane-impermeable CoA to the plentiful, mobile car-nitine provides transport between compartments, aconsiderable reservoir of activated acyl groups andexcretion of excess acyl moieties. Transport occursin the import of fatty acids for energy productionin mammalian mitochondria (reviewed in [1]) andin yeast peroxisomes [2]. The reservoir function refersto the acetyl-L-carnitine pool in heart and sperm (re-viewed in [1]) and to the long-chain acylcarnitinepool, which cells use for membrane repair whenthey lack energy to activate fatty acids [3]. Excretionof carnitine derivatives occurs via the urine, wherethey provide a marker for clinical measurements, andin bile, where it was recently demonstrated that long-chain acyl derivatives accumulate [4]. As a whole, the

0167-4838 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved.PII: S 0 1 6 7 - 4 8 3 8 ( 0 1 ) 0 0 1 4 7 - 9

* Corresponding author. Fax: +44-1334-462-595;E-mail : [email protected]

BBAPRO 36383 7-3-01

Biochimica et Biophysica Acta 1546 (2001) 21^43www.bba-direct.com

carnitine system both connects the various acyl-CoApools and damps £uctuations in their acylation statethat would be detrimental to cell homeostasis.

The carnitine system consists of carrier proteinsthat transport carnitine across the membranes andenzymes, carnitine acyltransferases, that catalyse thereversible equilibrium:

acyl-CoA� L-carnitine3CoA� acyl-L-carnitine

We review here the molecular genetics and enzy-mology of these proteins that play key roles in acti-vated acyl group homeostasis and transport.

2. Getting it in ^ the transport of carnitine into cells

2.1. Molecular genetics of the mammalian plasmamembrane transporters (OCTN2)

Recent e¡orts [5^8] provide a clear picture of theplasmalemmal carnitine transporter OCTN2 (Fig. 2).The human gene for OCTN2, SLC22a5 (i.e., member5 of solute carrier family 22), contains 10 exons andmaps to chromosome 5q31, a region that researchersidenti¢ed using linkage analysis of families with in-herited systemic carnitine de¢ciency [7,9]. The mouse

gene, known from the juvenile visceral steatosis (jvs)mutation [10], a missense mutation (Table 1), mapsto chromosome 11 [11]. This protein of 557 aminoacids probably contains 12 putative transmembranedomains (Figs. 2 and 3); however, some disagreeabout predictions of these domains [7,12,13]. Identi-¢ed sequences include three putative N-glycosylationsites in the ¢rst extracellular loop, several putativephosphorylation sites in intracellular loops [6,7], asugar transporter protein signature motif [6], andan ATP/GTP binding motif [7]. Unique among or-ganic cation transporters, OCTN2 needs sodium totransport carnitine but not for other organic cations[6,14]. OCTN2 most closely resembles OCTN1 andOCTN3, the latter only known from mouse (Gen-Bank database accession number BAA78343) [211].

An ancestral OCTN gene must have duplicated asa direct repeat, as the genes for OCTN1, a low-a¤n-ity, sodium-independent carnitine transporter [15],and OCTN2 are in tandem [16]. Deletion of bothgenes causes carnitine de¢ciency in mice [16], withall characteristics of the jvs mutation [10,17], includ-ing cardiomyopathy and fatty liver. Complementa-tion with a genomic fragment that encodes humanOCTN2, but not human OCTN1 [16], rescues thephenotype.

OCTN2 sequence analysis of systemic carnitinede¢ciency in humans, ¢rst reported by Nezu et al.[12] and then by others [18^23], represented a break-through for understanding of this disease. Fig. 2 andTable 1 present disease-causing amino acid substitu-tions known in human OCTN2. This compilationincludes mutations from a large-scale analysis of aJapanese subpopulation and mutations in heterozy-gotes that may account for low plasma levels of car-nitine [21].

2.2. Functional aspects of and mutations in the plasmamembrane transporters

Carnitine and acylcarnitines do cross membranes;so, a non-saturable component to uptake in cells ex-ists (see [24]). However, proteins mediate both thecellular accumulation of carnitine across the plasmamembrane (from about 50 WM in plasma to millimo-lar levels in cells) and the rapid £ux across the mi-tochondrial inner membrane required for L-oxidation(reviewed in [1]).

Fig. 1. Locations of the carnitine proteins in a mammalian cell.u, unknown; e.r., endoplasmic reticulum; see text for the pro-tein names.

BBAPRO 36383 7-3-01

R.R. Ramsay et al. / Biochimica et Biophysica Acta 1546 (2001) 21^4322

The discovery of the plasma membrane transportof carnitine dates to the recognition in the 1970s(reviewed in [1]) that liver, the major site of the ¢nalstep of carnitine synthesis, has a low a¤nity(KM = 5.6 mM [25]) and other tissues have a higha¤nity (KM = 4^20 WM [26]) for carnitine. OCTN2presumably is responsible for the high-a¤nity activ-ities measured in this early work. Variations in trans-port rates and in levels of expressed mRNA arefound in di¡erent tissues. Starvation and glucagonincrease the carnitine content of liver but not thatof heart [1,27,28]. Local regulation may also altercarnitine transport because palmitoyl-CoA andATP can alter the number of carnitine binding sites[29].

Mutations in human and mouse OCTN2 thatcause carnitine de¢ciency and site-directed mutagen-esis in rat OCTN2 reveal some crucial residues in the

protein (Table 1). Both individual substitutionsY211F and P478L block carnitine transport butnot organic cation transport, e.g., tetraethylammoni-um (TEA) [14]; these experiments suggest spatiallydi¡erent carnitine and TEA transport sites. Theswapping of di¡erent parts of human and ratOCTN2 [14] reveal residues 123^239 (Fig. 2) as crit-ical for the species-speci¢c transport characteristics.Compared to human OCTN2, rat OCTN2 is lesse¡ective as a carnitine transporter and more e¡ectiveas an organic cation transporter [14,30].

Transporter activity in a given cell should in£uencethe intracellular carnitine content. Rat heart containsabout twice as much carnitine as rat liver [27]. Semi-nal £uid contains the most carnitine with levels ashigh as 60 mM (see [1]). It has been speculatedthat two transporters, one at the basal membraneand one at the apical, may be necessary to achieve

Fig. 2. Amino acid sequence alignment of human, rat and mouse OCTN2 and related proteins from the solute carrier superfamily 22.Shaded residues are identical to a consensus derived from the sequences shown. Twelve putative transmembrane domains are indicatedwith roman numbered bars above the sequence alignment. The sugar transporter signature motif (Prosite PDOC00190) is indicated by+ symbols, the ATP/GTP binding signature motif (Prosite PDOC00017) by U symbols. The region that is critical for the discussedspecies-speci¢c OCTN2 characteristics (123^239) is indicated between arrowheads. Mutations that result in decreased OCTN2 functionare circled and numbered below the sequences (see Table 1 for details).

BBAPRO 36383 7-3-01

R.R. Ramsay et al. / Biochimica et Biophysica Acta 1546 (2001) 21^43 23

this large gradient. However, in the mammary gland,only a basal transporter drives the high £ux of car-nitine to these cells and, hence, into milk [24]. Theexpression of the transporter in the mammary glandincreases dramatically with the onset of pregnancyand lactation. This increase suggests a temporal, hor-monal regulation of expression.

The liver transporter has a high KM for L-carnitine(5.6 mM in hepatocytes) and a lower KM for theprecursor, butyrobetaine [25]; therefore, it is prob-ably not OCTN2. OCTN2 is not expressed at highlevels in liver [6,30]. The liver transporter is sodium-dependent and, therefore, is probably not OCTN1either. The liver produces the majority of endoge-nous L-carnitine in mammals; experiments with per-fused liver suggest a protein-mediated release of car-nitine. Mersalyl inhibits the release; ions or ouabaindo not [31]. The observed Vmax is 2.47 nmol min31

g31 of liver and the KM is 0.27 mM (intracellularcarnitine concentration in liver is about 0.5 mM).Starvation apparently decreases carnitine outputfrom perfused liver, suggesting decreased activity ofthe protein. If this work is con¢rmed, the e¥ux pro-tein would be a target for molecular investigation asit could in£uence carnitine output to other tissues.

Such a protein could be hormonally regulated be-cause glucagon increases the carnitine content of liv-er, as does starvation [27,28].

Identifying carnitine derivatives in urine and bileimplicates carnitine in the export of excess activatedacyl groups [3,4]. In particular, urinary output ofvalproylcarnitine during valproate therapy and ofthe carnitine esters of L-oxidation intermediates inacyl-CoA dehydrogenase de¢ciency patients demon-strates a substantial export of intracellular esters.Presumably, this export, down a concentration gra-dient, could be due to passive di¡usion. The slowexport from mammary gland cells is not stimulatedby reversing the sodium ion gradient; so the uptakesystem at the basal membrane probably does notreverse [1]. Although one group reports a saturablee¥ux from perfused rat liver with a KM of 0.15 mMand a turnover greater than the daily e¥ux, the pas-sive e¥ux measured in the same experiment couldaccount for the daily turnover [31]. A volume-acti-vated amino acid channel may also contribute tocarnitine e¥ux from swollen cells [1]. The recentlycharacterised OCTN1 transporter could also modu-late intracellular carnitine, providing a pH-regulatedpath of e¥ux [15].

Table 1Mutations in the carnitine proteins

A. Mutations in the plasmalemma transport protein, OCTN2 (numbering corresponds to circled residues in Fig. 2)

Substitution Species E¡ect on transport References

1 R169Q human [200]2 M179L human slight decrease in vitro [21]3 V211C human [22]

Y211F rat carnitine, not organic cations [14]4 Y251F rat slight decrease in vitro [14]5 W283R human [21,201]6 L352R mouse jvs mutation [12,13]

L352R rat carnitine+organic cations [14]M352R human carnitine+organic cations [14]

7 Y358F rat carnitine+organic cations [14]8 Y426 rat slight decrease in vitro [14]9 V446F human [201]10 E452K human reduced Vmax [23]

E452A/D/Q human increased K(Na) [199]11 S467C human [21]12 P478L human [19]

P478L rat carnitine, not organic cations [14]13 Y482F rat slight decrease in vitro [14]14 Y486F rat carnitine+organic cations [14]15 Y492F rat slight decrease in vitro [14]

BBAPRO 36383 7-3-01


3. Rapid internal equilibration ^ the intracellulartransporters

3.1. Molecular genetics of the mitochondrial carnitinetransporter (CACT)

Clones for the rat [32] and human [33] mitochon-drial inner membrane carnitine/acylcarnitine trans-porter (CACT, also known as the translocator, trans-locase, or carrier) are known. CACT is a member ofthe mitochondrial carrier family, which includes or-nithine, ADP/ATP, phosphate, 2-oxoglutarate andcitrate carriers as well as the proton carriers (uncou-pling proteins). All these transporters are about 30kDa, including rat CACT at 32.5 kDa [38]. Mamma-

lian CACT proteins contain 301 amino acid residues(Fig. 4). These proteins have a triple repeat struc-ture; each repeat contains two membrane-spanningdomains. CACT proteins occupy the mitochondrialinner membrane with the N-terminus, two small hy-drophilic loops and the C-terminus facing the inter-membrane space; three larger hydrophilic loops facethe matrix side (Fig. 3). CACT proteins appearto have a carnitine-speci¢c binding sequence ofR(A,S)(V,F)PANAA(T,C)F [85] near the C-terminus(see Fig. 4). All analyses of CACT mutations (dele-tions, frameshifts) have revealed drastic e¡ects onprotein structure. To date, no single-amino acid sub-stitution studies have been published.

Table 1 (continued)

B. Mutations in the carnitine acyltransferases and their e¡ect (numbering corresponds to circled residues in Fig. 5)

Substitution Enzyme Decrease in References

1 E3A rat L-CPT-I malonyl-CoA sensitivity [181,184]2 H5A rat L-CPT-I malonyl-CoA sensitivity [129,184]3 H140A rat L-CPT-I malonyl-CoA sensitivity [129]4 P50H human CPT-II activity [62,95]5 S113L human CPT-II activity [62,93,95^98,202]6 R151Q human CPT-II activity [98]7 H131A rat COT malonyl-CoA sensitivity [144]8 E174K human CPT-II activity [99,100]9 M214T human CPT-II activity [203]10 P227L human CPT-II activity [98,204]11 D454G human L-CPT-I level and activity [102]

D353A rat CPT-II level and activity [103]12 H372A/K rat CPT-II activity [103]

H327A rat COT activity [144]13 D376A rat CPT-II activity [103]14 R382K human CPT-II activity [98]15 F383Y human CPT-II activity [100]16 D464A rat CPT-II activity [103]17 Y479F human CPT-II activity [203]18 G549D human CPT-II activity [95]19 D553N human CPT-II activity [62]20 R505A bovine COT carnitine binding [205]21 S543A bovine COT carnitine binding [205]

T544A bovine COT carnitine binding [205]S545A bovine COT carnitine binding increase [205]

22 G710E human L-CPT-I enzyme activity [206]23 V605A rat CPT-II carnitine binding [103]24 G609A rat CPT-II carnitine binding [103]25 G611A rat CPT-II carnitine binding [103]26 Y628S human CPT-II activity [94]27 R631C human CPT-II activity [93,96]

BBAPRO 36383 7-3-01


3.2. Functional aspects of the mitochondrial carnitinetransporter

Identi¢cation and characterisation of mitochon-drial carnitine^acylcarnitine exchange carrier func-tion date to the 1970s [39,40]. The same gene productmay also provide carnitine transport in peroxisomesas both organelles express a protein reactive with thesame CACT peptide antibody [42].

The kinetic parameters of CACT reveal the asym-metric nature of the electrically neutral and pH-in-dependent exchange. In a puri¢ed system [34,41], theVmax is 1.7 Wmol min31 mg31, with KM values of 0.5mM and 8.7 mM on the internal and external faces,respectively. Acyl-L-carnitine homologues (acyl chainlengths of 12^16 carbons) transport as rapidly ascarnitine but with KM values of V5 WM. The ping-pong mechanism for the carnitine carrier [41] con-trasts with other mitochondrial carriers. This mech-anism allows the carrier to function as a uniporter to

adjust the organelle concentrations of carnitine. Theuniporter functions more slowly at 0.5% of the ex-change rate [41,43].

The activity of the carrier requires thiol groups. N-Ethylmaleimide and mersalyl inhibit (IC50 5 WM and0.05 WM, respectively) both the exchange and uniportactivities [41]. However, at high concentrations ofmercurials, passive e¥ux of carnitine occurs; theopening of a non-speci¢c channel, as seen for someother mitochondrial carriers, might be the reason.The thiol groups modi¢ed to give the loss of specif-icity seem to di¡er from those modi¢ed for inhibitionof exchange [44].

Other inhibitors include acyl-L-carnitine analogueswith a charged group at the g end [39]. Acyl-D-car-nitines bind as well as acyl-L-carnitines (the Ki valuefor decanoyl-D-carnitine inhibition of L-carnitine ex-change is 12 WM) but are transported slowly (D-car-nitine) or not at all [207]. A very recent study dem-onstrates convincingly that acyl-D-carnitines targetCACT for inhibition of fatty acid oxidation [208].

4. Equilibration with CoA pools ^ the carnitineacyltransferases

4.1. Molecular genetics of the carnitineacyltransferases

4.1.1. Carnitine acetyltransferases (CAT)CAT activity is found in mammalian mitochon-

dria, peroxisomes and endoplasmic reticulum (ER)[45]. In mammals, at least in rat hearts, CAT is nota cytosolic enzyme [46]. Translated as a precursor of626 amino acids, human mitochondrial CAT con-tains a 28- or 29-residue N-terminal mitochondrialtargeting signal (MTS), which is clipped o¡ duringtranslocation through the mitochondrial inner mem-brane [47,48]. Corti et al. [48] have discussed thepossibility that cleavage of the MTS occurs betweenresidues 28 and 29 of the precursor. Their experimen-tal data suggest, however, that cleavage probablyoccurs between residues 29 and 30, where thesequence more closely resembles the well-conservedcleavage site consensus RXYsS/A [163]. At the C-terminus (AKL), the precursor contains a peroxisom-al sorting signal, which functions predominantlywhen the MTS is not translated. Translation of the

Fig. 3. Schematic representation of the plasmalemmal carnitinetransporter OCTN2, carnitine palmitoyltransferase I (CPT-I)and carnitine acylcarnitine translocase (CACT). Bilayer phos-pholipid membranes are depicted in grey: pm, plasma mem-brane; mom, mitochondrial outer membrane; mim, mitochon-drial inner membrane. Hydrophilic parts of the protein (blacklines), membrane-spanning protein domains (white boxes), theamino-terminus (N), and the carboxy-terminus (C) of each pro-tein are indicated. The orientation of the proteins is in accor-dance with the layout of a cell, i.e., cytosol between pm andmom, and mitochondrial intermembrane space between momand mim. Crossing of the lines indicates the close interactionbetween the CPT-I N-terminus and the much larger C-terminalpart of the protein.

BBAPRO 36383 7-3-01


peroxisomal form of CAT starts at a second startcodon and, for human CAT, produces a protein of605 amino acid residues. The di¡erent initial aminoacid sequences come from di¡erential splicing thatproduces two mRNAs with and without the exonfor the mitochondrial targeting sequence [48]. Thesecond or third exon contains the start of translationof the peroxisomal form (as the starts of transcrip-tion are unknown the exon numbering is prelimi-nary) [63]. The di¡erential splicing might resultfrom the aberrant splice donor dinucleotide GC in-stead of GU in the `mitochondrial' ¢rst intron [48].In yeast, a single gene for CAT [49] also enables thetranslation of both a mitochondrial and a peroxi-somal isoform [50], but the underlying mechanismdoes not involve splicing. In both Saccharomyces ce-revisiae and Candida tropicalis, the CtCAT gene fromthe latter is di¡erentially translated because thetranslational machinery has a di¡erential preferencefor the ¢rst or second start codon [51]. Furthermore,studies in yeast show that CAT has both a C-termi-nal AKL and an internal peroxisomal targeting sig-nal (PTS). When both the mitochondrial targeting

sequence and the C-terminal AKL are deleted, themajority of the gene product is still sorted to theperoxisomes [50]. In addition, or alternatively, thefolded state of the protein may contain speci¢c in-formation for peroxisomal sorting [52].

Yeast have another gene, YAT1, which encodes acytosolic form of CAT that is attached to the mito-chondrial outer membrane [53].

In mammals, it is not known whether the ER formof CAT comes from the same gene that encodes mi-tochondrial and peroxisomal CAT. The sequenceKVEL in CAT (pos. 492^495) could possibly func-tion like the ER retention signal KDEL [48], whichnormally is found at the C-terminus. In pigeonbreast CAT [54], the KVEL sequence is absent(KADL in pigeon CAT, pos. 493^496); conse-quently, the KVEL sequence in human CAT maybe irrelevant for sorting.

4.1.2. Carnitine octanoyltransferases (COT)The human [55], rat [56] and bovine [57] cDNA

sequences of peroxisomal COT encode proteins of612 amino acids (about 70 kDa) that contain di¡er-

Fig. 4. Amino acid sequence alignment of the mitochondrial carnitine transporter (CACT) from various species and the related humanornithine transporter (ORNT). Shaded residues are identical to a consensus derived from the sequences shown. Six putative transmem-brane domains are indicated with roman numbered bars above the sequence alignment. A conserved region implicated in carnitinebinding is boxed. D., Drosophila ; C., Caenorhabditis ; E., Emericella (Aspergillus) ; S., Saccharomyces.

BBAPRO 36383 7-3-01


ent C-terminal sequences (THL, AHL and PHL, re-spectively), thought to serve as a PTS. However, asin CAT (see above), other PTSs may be involved orthe folded state of the protein may contain speci¢cinformation for sorting. Genomic sequences from rat[56] reveal 17 exons with the start codon on the sec-ond exon. Expression of the rat gene for COT (Crot)is subject to trans-splicing, i.e., splicing of di¡erentprimary transcripts to one mRNA [58]. In the case ofrat liver COT, the di¡erent mRNAs that result fromtrans-splicing contain either a direct repeat of exon 2or a direct repeat of exons 2 and 3. Therefore, in-cluding the normal transcript, the three mRNA se-quences result in the expression of two COT pro-teins: one of normal size, i.e., 69 kDa, and one of79 kDa [58]. The transcript that contains the repeatof exon 2 does not produce a COT protein due to aframeshift because the number of nucleotides in exon2 is not a multiple of three. The human gene, CROT,contains an additional intron at the start of the cod-ing region [63].

4.1.3. Carnitine palmitoyltransferases II (CPT-II)CPT-II, a mitochondrial matrix protein associated

with the inner mitochondrial membrane, is translatedas a precursor of 658 amino acids in both human [59]and rat [60]. The mature protein is about 70 kDa, asthe N-terminal MTS of 25 amino acid residues isremoved during mitochondrial import [59,61]. Thegenomic structure of human CPT2 [62] is similar tothe mouse gene [64], which also encodes a precursorof 658 amino acids.

4.1.4. Carnitine palmitoyltransferases I(CPT-I; L-CPT-I and M-CPT-I)

Di¡erent genes, called CPT1A and CPT1B, en-

code the two known mammalian isoforms of CPT-I, liver-type carnitine palmitoyltransferase (L-CPT-I)and muscle-type carnitine palmitoyltransferase (M-CPT-I), respectively (reviewed in [65]). The L-CPT-I isoforms, proteins of 773 amino acids, have pre-dicted and apparent sizes of about 88 kDa. The M-CPT-I isoforms, proteins of 772 amino acids, havepredicted sizes of about 88 kDa, but apparent sizesof about 82 kDa by SDS^PAGE analysis. cDNAsequences of the genes for L-CPT-I are known forrat [66,67], man [68], mouse ([69]; van der Leij, un-published) and sheep ([70]; Price et al., unpublished).Protein sequencing of rat L-CPT-I reveals that themature enzyme retains the N-terminus [71]. A mito-chondrial targeting sequence at residues 123^147 im-mediately behind the second transmembrane regionhas also been shown to act as a stop-transfer se-quence to anchor CPT-I in the outer membrane[210]. After ATP-dependent integration in the mem-brane [72], the majority of the enzyme is on the cy-tosolic face of the mitochondria, anchored by twotransmembrane domains (Figs. 3 and 5). The initialbiochemical studies [73] for rat L-CPT-I and the re-cent immunocytological studies on human M-CPT-Iwith green £uorescent protein [74] support these con-clusions.

The rat M-CPT-I cDNA, obtained by screening abrown adipose cDNA library with a white adiposecDNA probe, encodes a protein that was called CPT-I-like protein [75]. The strong cardiac expression ofthis clone [75,76] and its human orthologue [77] isconsistent with the identi¢cation of the encoded pro-tein as M-CPT-I [76].

In another approach to clone human CPT1B, partof the cDNA sequence was assembled after databasescreening of human expressed sequence tags with

C

Fig. 5. Amino acid sequence alignment of nine carnitine acyltransferases. The alignment is shown with shaded residues at positionswhere these are identical to a consensus sequence derived from an alignment of 15 carnitine acyltransferases (including also rat L-CPT-I, rat M-CPT-I, mouse CAT, pigeon CAT, bovine COT and mouse CPT-II). The two CPT-I transmembrane domains are indi-cated with roman numbered bars above the sequence alignment. Carnitine/choline acyltransferase motifs are indicated by + symbols(Prosite PS00439) and by U symbols (Prosite PS00440). N-terminal signals for mitochondrial targeting and C-terminal signals for per-oxisomal targeting are boxed. The cleavage site of the mitochondrial targeting sequence of yeast CAT has not been determined experi-mentally and is predicted according to [163]. Start methionines of the peroxisomal forms of CAT are underscored. Mutations whichresult in loss of malonyl-CoA sensitivity (CPT-I) or loss of enzyme activity are circled and numbered below the sequences (see Table1 for details). The mutations studied in rat L-CPT-I and bovine COT are indicated in the human L-CPT-I and rat COT sequence, re-spectively. D. melan., Drosophila melanogaster ; S. cerev., Saccharomyces cerevisiae.

BBAPRO 36383 7-3-01


BBAPRO 36383 7-3-01


the rat L-CPT-I cDNA sequence as query [78].This in silico cloning e¡ort permitted the isolationfrom human heart and skeletal muscle of severalcDNAs.

Analyses of these cDNAs show that sometimes thelast intron of human CPT1B is not removed fromthe primary transcript because of competition be-tween splicing and polyadenylation. The ¢rst exonis untranslated and alternative ¢rst exons exist,which was noted [78] based on data of Zhu et al.[79]. Further analyses of the alternative splicing ofthe CPT1B gene in human [80] and rat [81] revealthat the two ¢rst exons (exon 1a and 1b, also calledU and M) do not co-exist in one transcript and thattwo promoters drive the expression of humanCPT1B. Consequently, the expression of the 5P un-translated mRNA for human M-CPT-I di¡ers fromthe alternative splicing of the transcript for L-CPT-Iin rats where the second untranslated region is some-times skipped [67].

Human CPT1B (for genomic information, see[63]), which was sequenced as the middle gene ofseven from a 180-kb BAC clone, is located closelydownstream of a choline/ethanolamine kinase gene[78,82]. CPT1B is even co-expressed with this up-stream gene [83]. The gene synteny is conserved inrats [84] and mice [69]. In Drosophila, only one genefor CPT-I appears [86]. The protein is slightly largerthan mammalian counterparts (782 amino acids, Fig.5), and, when expressed in yeast, shows character-istics typical of a CPT-I enzyme.

4.1.5. Carnitine acyltransferases in the endoplasmicreticulum

A recent proposal, based on the common immu-noreactivity, suggests that all the cytoplasmic-facingCPT enzymes are the same [87]. Although kineticdata support this proposal for mitochondrial andperoxisomal enzymes, the overt microsomal activityhas subtle kinetic di¡erences [88^90]. The onlyknown candidate gene sequence for a microsomalcarnitine acyltransferase, unrelated to the transferasefamily discussed above, is GRP58 [91,92], alsoknown as Erp57, Erp60 or Erp61. This protein isthought to have various functions; whether it isalso a carnitine acyltransferase remains controversial(reviewed in [104]).

4.2. Mutations in the carnitine acyltransferases

Mutations at more than 30 positions in the acyl-transferase protein family have been diagnosed ormade through site-directed mutagenesis. Table 2and Fig. 5 contain 30 of these mutations that a¡ectenzyme function. Of the naturally occurring humanmutations, the most frequent occur in CPT-II. CPT-II de¢ciency, the most common muscular lipid me-tabolism disorder [93], can present with di¡erenttimes of onset and with di¡erent phenotypes. Theadult form with muscular or hepatomuscular pre-sentation and the neonatal form with hepatocardio-muscular presentation represent the most extremephenotypes in this recessive disorder. Infantile andintermediate phenotypes occur as well (recently re-viewed in [65,94,95]).

The S113L mutation, among the ¢rst identi¢ed, isthe most common in CPT-II [93]. Reported in severalstudies of homozygotic or compound heterozygoticCPT-II de¢ciency (e.g., [62,94,96^98]), this mutationhas an allele frequency of 60%. The a¡ected serineresidue at human CPT-II position 113 is conservedwithin CPT-I, CPT-II, and `vertebrate' CAT (hu-man, mouse, pigeon), but not in any COT proteinsnor in yeast CAT (Fig. 5). S113L, a `mild' mutation,is associated only with the adult form of CPT-IIde¢ciency. A strict correlation between severity ofeach mutation and onset and phenotype, however,cannot be made as clear exceptions exist. For exam-ple, a homozygous R631C mutation gives a di¡erentexpression of the disorder in unrelated patients [94].Therefore, genotypic di¡erences outside the CPT2gene may have major in£uences in certain cases.

Polymorphisms in CPT-II as single entities do notsigni¢cantly a¡ect CPT-II enzyme function (reviewedin [65,94]. Like S113L, P50H [95,97] and E174K[99,100], which have mild consequences, are associ-ated with the adult form of CPT-II de¢ciency. P50His a change in the ¢rst fully conserved proline of anLPXLP motif (pos. 49^53 of CPT-II). The glutamateat position 174 is less conserved, although othertransferases share glutamates in the vicinity (Fig.5). Severe consequences for CPT-II function resultfrom F383Y [100] and Y628S [101]. Both a¡ect res-idues that are barely conserved outside CPT-II(Fig. 2).

BBAPRO 36383 7-3-01


In contrast to CPT-II de¢ciency, de¢ciencies in theother transferases are still rare and only known atthe sequence level for CPT1A [102,206]. A recentCPT1A mutation analysis [206] has identi¢edG710E, which is adjacent to a GFG pattern thatis associated with carnitine binding in rat CPT-II[103].

4.3. Functional properties of the carnitineacyltransferases

Recent reviews have summarised the locations andfunctions in the cell [104,105]. The native structuresof CAT and COT are monomers [106,107]. For CPT-II, only aggregates, such as 660 000 kDa [108], areisolated after detergent solubilisation, although ahomotetramer has been suggested [109]. Inactivationof CPT-I by radiation of outer membranes has pro-duced a target molecular weight that is close to themonomer molecular weight, suggesting (but notproving) that the normal form in the membranecould be a monomer [110].

The sigmoidal kinetics often observed for CPT-I(e.g., [111,112]) could indicate multisubunit co-oper-ation. However, artefactual sigmoidicity or non-line-ar time courses caused by interactions of palmitoyl-CoA with albumin or micelles must ¢rst be excluded.Acyl-CoA binding protein (ACBP) prevents thesetechnical problems [90]. Sigmoidicity could still arisevia the malonyl-CoA regulatory site, which couldalso explain the high substrate inhibition observedfor palmitoyl-CoA [113]. Sigmoidicity with palmito-yl-CoA is not changed by malonyl-CoA inhibition[111] but disappears after treatment with the thiolreagent DTNB [112]. Consequently, the sigmoidicityarises from the protein and not from an artefact.This implicates a thiol group in the reversible alter-ation of sensitivity to malonyl-CoA [114]. These ob-servations may be explained by (a) the £exible inter-action between the N-terminal region and the mainpart of the enzyme or (b) the complex responses ofmalonyl-CoA inhibition to membrane environment(see below) or (c) both.

The e¡ect of detergents on activity and membraneattachment generated many early controversiesabout the CPTs. The use of Triton X-100, whichinactivates CPT-I completely but activates and solu-bilises CPT-II, delayed the identi¢cation of CPT-I as

a separate protein [60]. In one laboratory, octylglu-coside solubilised active CPT-I [113], whereas in an-other, it inhibited and failed to solubilise it [60]. In athird, it failed to solubilise either CPT-I or the per-oxisomal malonyl-CoA-sensitive enzyme but, of the¢ve detergents studied, octylglucoside had the leaste¡ect on the activity of COT, CPT-I and CPT-II(A.S. Friend and R.R. Ramsay, unpublished data).Others found that octylglucoside produces complexe¡ects on extracted CPT-II [115].

Despite numerous publications on these acyltrans-ferases, the susceptibility of the assay to artefacts,complicated by both high substrate and strong prod-uct inhibition, has produced data that are only suit-able for internal comparisons. Kinetic studies of car-nitine acyltransferases should take into accountproblem variables: micelle concentrations of thelong-chain acyl derivatives (e.g., [116] and referencestherein), the presence (or absence) of albumin([117,118] and references therein), the other compo-nents (membranes bind palmitoyl-CoA [113,116,119])and concentrations of lipid or detergent or both([60,113,115], all of which in£uence acyl-CoA avail-ability [118]). As mentioned above, the recent use ofrecombinant ACBP overcomes some problems inher-ent in experiments with long-chain acyl-CoA sub-strates [90]. Table 2 gives true kinetic constants forpuri¢ed CAT, COT and CPT-II [120^123] but it isdi¤cult to discuss the level of saturation of theseenzymes with acyl-CoA substrates in the cell. Thecytoplasmic content of CoA is low (less than 10%of the cellular pool in heart [124] and variable in liver[125]). The concentration of free acyl-CoA is essen-tially zero because all the cytoplasmic long-chainacyl-CoA will be bound to ACPB. It has been sug-gested that CPT-I uses directly the pool of acyl-CoAbound to ACBP [119].

From the earliest study on the ¢rst puri¢ed trans-ferase, CAT [126], to the more recent studies, all ¢ndthat V changes little in the range pH 7^8 [1]. Wheredi¡erences in KM are noted, at least 85% saturationwith carnitine could be maintained at normal cellularconcentrations. Kinetic parameters are usually deter-mined within this range. Table 2 gives examples ofvalues obtained for each transferase. The V for acyl-carnitine formation (removal of acyl-CoA, the for-ward direction) is generally higher than that for thereverse direction. However, as noted above, local

BBAPRO 36383 7-3-01


concentrations of the substrates will determine thenet £ux.

Table 2 also shows information on the di¡erentkinetic properties of the isoforms of CPT-I (muscleand liver). They have the same KM for palmitoyl-CoA but the KM for carnitine in the muscle form is10-fold higher than in the liver form. This correlateswith the higher carnitine concentration in heart thanin liver [27,124]. Indeed, McGarry et al. [120,127],who surveyed a range of tissues and species, haveshown an inverse correlation between carnitine KM

and malonyl-CoA IC50. Jackson et al. [181] have nowdemonstrated that the inverse relationship of carni-tine KM and malonyl-CoA sensitivity is caused bymodulation of the catalytic domain by the N-termi-nal residues.

The ¢rst structural information on CPT-I and mu-tation of N-terminal residues make it easier to under-stand some observations. Interaction of the N-termi-nal with the large C-terminal domain helps stabilisethe catalytic site and modulate malonyl-CoA sensi-tivity [73,128,129] (see below). As shown by Zam-mit's group [130] using a set of six chimeric proteinsplus the parental forms, interactions between the twocytosolic parts of the protein (see Fig. 3) contributeto interaction with the ¢rst substrate, palmitoyl-CoA.The pairing of the transmembrane domains in£uen-ces the interaction with carnitine [130] but removal ofthe N-terminal transmembrane domain still leaves anactive protein [181]. From their study of both the Mand L forms of the enzyme, we see that the C-termi-nal portion of the protein gives rise to a 100-folddi¡erence in sensitivity to malonyl-CoA (M moresensitive). What is clear from this work [130,181] isthat the catalytic core contains the malonyl-CoA

binding site. Modulation of the malonyl-CoA sensi-tivity depends, however, on the N-terminal section[128^130].

The same studies [130] imply that the membraneanchor helices could transmit the in£uence of mem-brane changes. Recent observations ex vivo [131] em-phasise microenvironmental in£uences on the ki-netics of this special enzyme. More CPT-I is foundin the contact sites between the outer and inner mem-brane than in the normal outer membrane [131,132],and its kinetic properties are changed dramatically(F. Fraser and V.A. Zammit, personal communica-tion). In the outer membrane, malonyl-CoA inhibi-tion a¡ects V but not KM for palmitoyl-CoA, where-as in the contact sites, V is unchanged but theapparent KM for palmitoyl-CoA is greatly increased.Consequently, the mode of inhibition by malonyl-CoA is competitive in the contact sites, where chan-nelling to L-oxidation seems likely, so that excesslong chain acyl-CoA could overcome the inhibitionand prevent accumulation. In the outer membrane,uncompetitive inhibition ensures that accumulatingacyl-CoA required for other cytosolic functions isnot converted into acylcarnitine.

Although the main target for designing inhibitorsin this enzymic family is CPT-I because it in£uencesthe rate of mitochondrial fatty acid oxidation, thekinetics are the hardest to study because they mustbe determined in the mitochondrial outer membrane.In the native state, problems include the in£uence ofthe membrane £uidity on activity, the sensitivity ofmalonyl-CoA inhibition to the £uidity, the bindingof palmitoyl-CoA and palmitoylcarnitine to themembrane, sensitivity to proteolysis and the di¡er-ences in mitochondria prepared from animals in dif-

Table 2Comparison of the kinetic constants for the various carnitine acyltransferases

Gene MW (kDa) Enzyme KM (WM) V (U/mg) References

Acyl-CoA Carnitine Acyl-Carn CoA Forward Reverse

CPT1A 88 L-CPT-I a 30 40 Rat liver [120]CPT1B 82 M-CPT-I a 500 Rat skeletal muscle

[120]CPT2 68 CPT-II 6 5 1500 46 112 32 15 Beef liver [121]CROT 71 COTP 0.6 108 7.4 16 36 37 Beef liver [121]CRAT 67.5 CAT 34 120 350 37 500 396 Beef liver [122]? 54 CPTER 6 4 600 1000 300 100 13 Rat liver [123]a30^60 WM in the presence of bovine serum albumin.

BBAPRO 36383 7-3-01


ferent physiological states (starvation, diabetes, hor-monal changes, the presence of proliferators, etc.).The reconstitution of the protein into arti¢cial mem-branes [79,133,185] and its expression in yeast freefrom endogenous activity [130] will permit detailedkinetic studies.

5. Acyl group transfer ^ the kinetic and chemicalmechanisms

5.1. Kinetic mechanisms

Kinetic mechanisms for CAT [122], COT andCPT-II [121] are known. From primary plots andproduct inhibition studies, CAT and COT show rap-id equilibrium random order kinetics indicating thatall four substrates bind well to the free enzyme. Incontrast, CPT-II follows a compulsory order ternarycomplex mechanism in which CoA must bind ¢rst.CPT-II has a very low a¤nity for L-carnitine, whichhas a Ki value of 11.7 mM for product inhibition. Incontrast, COT has a 100-fold greater a¤nity for L-carnitine, which has a KM value (0.1 mM) equivalentto the dissociation constant. The observation thatCPT-II binds acylcarnitine analogues whereas COTbinds carnitine analogues (see Sections 5.2.4 and5.2.5 below) re£ects this di¡erence.

The ordered mechanism of CPT-II, in which acyl-CoA binds ¢rst, ensures that this reversible enzymeresponds to the acylation state of the CoA pool inthe mitochondrial matrix. The matrix concentrationsof both CoA and acyl-CoA [124] exceed the mea-sured KM values [121] so that most CPT-II willhave the ¢rst (CoA) substrate bound. CertainlyCPT-II in normal mitochondria has never appearedto limit the rate of fatty acid oxidation. The step thatdoes control the access of fatty acyl-CoA to the ma-trix for L-oxidation is CPT-I on the mitochondrialouter membrane. Unpublished data (cited in [134])indicate an ordered mechanism for CPT-I too.

5.2. Chemical mechanism

In the past decade, our understanding of the basicchemistry of carnitine acyltransferases has progressedconsiderably. Computational studies of acyl transferequilibria explain why acylcarnitines have a large

acyl group transfer potential ; sequence studies andsite-directed mutagenesis provide further informationfor elaborating a chemical mechanism; and measure-ments of KM and Ki values for substrate and putativetransition state analogues provide additional data fordeveloping a picture of the active site. Despite theabsence of a crystal structure of a carnitine acyltrans-ferase or detailed chemical mechanistic studies, onecan de¢ne the mechanistic possibilities for the reac-tion shown below.

�1�

5.2.1. Reaction energeticsAcyl-CoA, a thioester, has a high acyl transfer

potential. Acylcarnitine, an oxyester, must have asimilar potential to avoid coupling the acylcarni-tine-to-CoA transfer to an energy-releasing reaction.The equilibria between acetyl-CoA and various oxy-esters (see Eq. 1) are calculated from the free energiesof hydrolysis (vG³hyd at 25³C) for acetyl-CoA[135,136], acetylcarnitine [137], acetylcholine [138]and ethyl acetate [139] of 334.3, 333.0, 327.1 and324.7 kJ/mol, respectively. Acetylcarnitine has themost favourable equilibrium for transfer to CoA;in fact, acetyl transfer occurs readily in both direc-tions [137]. For acetylcholine and ethyl acetate, ace-tyl transfer to CoA costs energy.

A recent study [209] explains the high acetyl trans-fer potential of acetylcarnitine using solvation ener-gies. Solvation energies, rather than intrinsic instabil-ity, can determine the reactivity of `high-energy'compounds [140^142]. For acetyl oxyesters, the sol-vation of acetic acid (or acetate) drives the hydroly-sis ; however, for charged esters and alcohols, therelative solvation energy of the ester compared tothe alcohol accounts for the di¡erences in acetyltransfer potential. Calculations of vvHsolv betweenthree pairs of charged ester alcohols ^ the zwitter-ionic acetylcarnitine^carnitine pair, the cationic ace-tylcholine^choline pair, and the anionic 3-acetoxy-propanoate^3-hydroxypropanoate pair ^ give 9.8,18.5 and 18.7 kJ/mol, respectively [209]. In all cases,the esters are more solvated than the alcohols. Thezwitterionic acetylcarnitine^carnitine pair in whichboth have large dipole moments, has the smallest

BBAPRO 36383 7-3-01


vvHsolv, which results in vG³hyd for acetylcarnitinebeing the most negative.

5.2.2. Proposed mechanismMost mechanisms for acyl transfer involve a tetra-

hedral intermediate. No direct evidence supports ordisputes this mechanism for carnitine acyltransfer-ases. The role of histidine in the mechanism datesback to Chase and Tubbs [143], who used bothchemical modi¢cation and substrate analogues to im-plicate a histidine. Site-directed mutagenesis identi¢esit as H372 in CPT-II [103] and H327 in COT [144].The loss of activity of D376A and D464A mutants ofCPT-II suggests a role for aspartate, and recentchemical modi¢cation studies suggest that a lysinecontributes to catalysis [103,145]. A catalytic role

for an acyl intermediate with serine is unlikely[57,121]. A direct transfer between carnitine andCoA appears logical. The simplest mechanism hasthe histidine serving as a general base ^ removing aproton from carnitine or CoA depending on the di-rection of transfer ^ to promote formation of a tet-rahedral intermediate. An aspartate may potentiatethe histidine function. The lysine stabilises the puta-tive oxyanion. The protonated histidine donates aproton to the departing group, either carnitine orCoA.

Other residues contribute to binding but not di-rectly to catalysis. An arginine probably forms astrong salt bridge with the carboxylate of carnitine.A cleft similar to the aromatic residue-rich pocket inacetylcholinesterase for recognising the trimethylam-

Fig. 6. Inhibitors of the carnitine acyltransferases. (a) Substrate analogues: CC, cyclohexyl carnitine; HDH, t-butyl analogue of carni-tine. (b) Acylcarnitine analogues: N-PAC, N-palmitoyl aminocarnitine; AM, acidomorpholinium analogues. (c) Irreversible inhibitors(the CoA esters are the inhibitory forms): TDGA, tetradecylglycidic acid; DNP-ET, dinitrophenyl etomoxir. (d) Putative transition-state analogues: HPC, hemipalmitoylcarnitinium; SDZ-CPI-975, acyl phosphinyloxy derivative of carnitine.

BBAPRO 36383 7-3-01


monio group of choline [146] is likely for carnitine. Aserine^threonine^serine triad conserved in all the car-nitine acyltransferases (and in choline acyltransfer-ases) contributes, perhaps by providing an aqueous-like microenvironment, to binding of carnitine [57],which is strongly solvated (see above). Substitutionof all three residues of the STS triad gives a 1000-fold increase in KM for carnitine. In CPT-II, threesubstitutions ^ V605A, G609A and G611A ^ resultedin a higher KM for carnitine [103].

5.2.3. Substrate analoguesA study of a series of cyclohexyl carnitine ana-

logues (Fig. 6a) has probed the preferred conforma-tion for the binding of carnitine. The analogue inFig. 6a (CC) binds selectively to the active site ofCPT-I but is a non-competitive inhibitor of CPT-II[147]. In addition, a study with the tert-butyl ana-logue of carnitine, HDH, reveals the enhancementprovided by a positive charge for carnitine acyltrans-ferase activity [148].

5.2.4. Inhibitors ^ irreversibleIrreversible inhibitors of CPT (Fig. 6b) have a re-

active epoxide with an alkyl chain and a carboxylgroup (e.g., TDGA [149], etomoxir [150] and DNP-ET [151]). DNP-ET selectively inhibits L-CPT-I[152]. In a recent study of 12 analogues, SNU-13b[153] shows better hypoglycaemic activity and lessacute toxicity than etomoxir in streptozotocin-in-duced diabetic rats. Presumably, these compoundsselectively inhibit L-CPT-I.

5.2.5. Inhibitors ^ reversibleReversible inhibitors of CPT (Fig. 6b,d) are either

putative transition state analogues (e.g., HPC [134],SDZ-CPI-975 [154]) or acylcarnitine analogues (e.g.,N-PAC [155], SDZ-269^456 and AM [184]). N-PACforms when CPT reacts with palmitoyl-CoA andL-aminocarnitine (emeriamine) [156].

The two putative transition state analogues havesimilar potency for CPT-I. HPC [134] and SDZ-CPI-975 [157] reversibly inhibit CPT-I activity with Ki

values in the low micromolar range as a competitiveinhibitor with palmitoyl-CoA and non-competitiveinhibitor with carnitine. SDZ-CPI-975 [158] andHPC [121] inhibit partially puri¢ed CPT-II in thesub-micromolar range.

Although not as potent as HPC and SDZ-CPI-975, AM stereoisomeric analogues inhibit CPT-Imore e¡ectively than CPT-II [184]. Assays of twomicrosomal CPTs reveal little discrimination amongthe stereoisomers but rat liver mitochondrial CPT-Iand CPT-II show distinct di¡erences. Compound(2R,6S)-AM (n = 12), which does not inhibit CPT-II, emerges as a potentially useful compound forthe selective inhibition of CPT-I.

6. Regulation

Changes in the activity and expression of CPT-I inresponse to hormones have been reviewed [65]. CPT-I increases in starvation (e.g., [159]) and decreases inhypothyroidism [162^164]. Insulin regulates CPT-Ivia the insulin growth factor I receptor [160,161].The switch from the mixture with L-CPT-I presentshortly after birth to the predominant M-CPT-Ifound in adults has been demonstrated in rat [165]and sheep [70] heart.

CPT-II mRNA and activity is likewise increased instarvation and diabetes and also by peroxisomal pro-liferators. Peroxisomal COT increases in parallelwith CPT-II after induction by peroxisomal prolifer-ators suggesting a parallel mechanism (reviewed in[166]). Fish oil diets also induce peroxisomal COTin rats in keeping with the role of the peroxisomesin shortening very-long-chain fatty acids for transferas medium-chain acylcarnitines to the mitochondria.In yeast, where L-oxidation occurs only in peroxi-somes, COT and the enzymes of fatty acid oxidationare induced during growth on fatty acid.

Now that complete sequence information, speci¢cinhibitors of L and M-CPT-I and antibodies to CPT-I, CPT-II and COT are available, changes in thespeci¢c forms of the enzymes can be investigated.For example, epitopes of L-CPT-I are present notonly in mitochondria but also in peroxisomes andendoplasmic reticulum [87]. A study with speci¢c in-hibitors and antibodies demonstrates [127] that heartcontains both M-CPT-I and L-CPT-I (the L form at2% only). This study suggests that activity measure-ments are no longer adequate. Most likely, somehormonal e¡ects mentioned above should be re-ex-amined at the molecular level.

BBAPRO 36383 7-3-01


6.1. Transcriptional regulation

The tissue-speci¢c and temporal expression ofCPT1A and CPT1B as compared to the body-wideand relatively steady expression of CPT2 has beenreviewed [65]. CPT1A is expressed in liver andmany other tissues, whereas CPT1B is signi¢cantlyexpressed in skeletal and cardiac muscle and in testis.However, adipocytes have species-speci¢c di¡erences(mouse versus man and rat) [167]. Despite this,CPT1B promoter sequences are remarkably con-served among man, mouse, rat and sheep.

The expression of several enzymes needed for fattyacid transport and oxidation is regulated at the tran-scriptional level via the peroxisome proliferator-acti-vated receptor K (PPARK) [168,169]. Like manynuclear membrane receptors, PPARK can formheterodimers with related receptors, e.g., the retinoidX receptor [170], to mediate the signalling betweenligand and target gene. The promoters of these targetgenes contain speci¢c sequences that are binding sitesfor the receptor dimers. The PPARK binding sitesare known as peroxisome proliferator response ele-ments and those responsive to fat have also beencalled fat-activated response elements (FARE). Stud-ies with a knockout mouse model provide direct evi-dence that PPARK participates in transcription ofCPT1B, but is less prominent in that ofCPT1A andCPT2 [169^173].

The CPT1B gene responds to accumulating intra-cellular fatty acid intermediates (presumably long-chain acyl-CoAs) via the presence of a FARE inthe promoter [170,174]. Hence, CPT1B is stronglyactivated when etomoxir inhibits CPT-I [170] orunder fasting conditions [173]. A PPARK-mediatedresponse to fasting might also contribute to activat-ing CPT1B transcription in the liver [81]. Further-more, a mammalian orthologue of a chicken ovalbu-min upstream promoter transcription factor (COUP-TF) counteracts the FARE and competes withPPARK [81]. This has special interest becauseCOUP-TF and other transcription factors like SP1play a role in cardiac development and in hypertro-phic responses of the heart associated with metabolicfoetal gene re-expression programmes [175].

The CPT1A gene is also induced by fasting andetomoxir treatment. The induction is also signi¢-cantly when independent of PPARK knockout mice

[172,173], and the CPT1A promoter contains no se-quences similar to the FARE of CPT1B. CPT1Apromoter studies have revealed roles for transcrip-tion factors like SP1 (see above) and SRY [176],but the molecular mechanisms of the responses tofatty acids and cyclic AMP [177] remain unknown.CPT1A is expressed in the heart when cardiac carni-tine levels are low [165], and the juvenile steatosismouse, which lacks a functional OCTN2, shows in-creased cardiac CPT1A expression [178]. Administer-ing L-carnitine represses and reverses the e¡ect [178],thus bypassing the active step in cardiac carnitineuptake through the action of low-a¤nity carnitinetransporters or through di¡usion. Although the e¡ectof carnitine on cardiac CPT1A expression might sug-gest suppression by carnitine, a direct role for carni-tine on gene transcription has not been shown [177],and the relieving e¡ect of carnitine on long-chainacyl-CoA accumulation may well explain the conse-quences of carnitine supply in the juvenile steatosismouse.

Intriguingly, the CPT2 gene promoter does notcontain a FARE like the known one of CPT1B,but CPT2 does respond to speci¢c ¢brate inductionthrough PPARK, at least in the liver [169]. ThisCPT2 response is only seen with ¢brates and notwith (long-chain) fatty acids, pointing to activator-speci¢c causes of CPT2 transcription induction.

The transcription of CPT1 and CPT2 has beenstudied relatively extensively compared to the expres-sion of COT and CAT [177]; no promoter studies ofthe genes for CAT, CACT and OCTN2 have beenreported yet.

6.2. Physiological regulation by malonyl-CoA

Malonyl-CoA, a substrate analogue, is, as wouldbe expected, a competitive inhibitor of CAT, COTand CPT-II. The Ki value of COT is similar to valuesfor other short-chain CoA esters [179]. This inhibi-tion involves interaction of malonyl-CoA presumablyin the active site and di¡ers from the regulatory sen-sitivity of CPT-I to malonyl-CoA.

The observation of the extreme sensitivity ofL-CPT-I to malonyl-CoA dates back to 1978 [180].Malonyl-CoA, a fatty acid synthesis intermediate,enables the inhibition of the fatty acid oxidationunder conditions where synthesis is required. We

BBAPRO 36383 7-3-01


summarise here, omitting much of the controversy,the molecular basis for inhibition.

b Both the L and M form contain a malonyl-CoAbinding site and a catalytic site in one protein [65].Contrary to earlier conclusions (summarised in[65]), the malonyl-CoA binding site and the activesite are on the same side of the membrane [65,73].Fig. 2 shows the proposed orientation of the do-mains.

b Isoforms L and M di¡er in malonyl-CoA sensitiv-ity [112,120].

b The N-terminus, not essential for catalytic activity,modulates the response to malonyl-CoA, but thelarge C-domain sets the sensitivity [128^130,181,182]. Deleting various portions of the N-terminus both negatively and positively a¡ects thesensitivity [181].

b Malonyl-CoA sensitivity decreases as pH increases(for attenuation of fatty acid oxidation duringacidosis [183]).

b Inhibition by malonyl-CoA is non-competitivewith palmitoyl-CoA in the outer membrane butis purely competitive in the contact sites (F. Fraserand V.A. Zammit, unpublished).

b Membrane insertion contributes to modulation ofmalonyl-CoA sensitivity as demonstrated by re-constitution into liposomes at di¡erent tempera-tures ([185] and older references therein). The al-terations in the kinetics of L- and M-CPT-I wheneach transmembrane domain is exchanged demon-strate how these e¡ects are transmitted [130].

The altered kinetics of malonyl-CoA inhibitionof L-CPT-I in response to the environment (tem-perature, lipids, membrane) suggests £exibility inthe protein. Arguments for an allosteric site formalonyl-CoA have been summarised [65]. However,protection by malonyl-CoA against inactivation bythe covalent suicide substrate etomoxiryl-CoA orthe bisubstrate formed in the active site frombromoacetyl-CoA plus carnitine suggests that eitherthe two sites overlap or they a¡ect each otherstrongly enough to mutually in£uence binding. Therecent revelation [181] that the N-terminus a¡ectsboth malonyl-CoA inhibition and carnitine bindingdemonstrates that malonyl-CoA acts at the activesite.

Recently, Guzman et al. [186] proposed that, inconcert with malonyl-CoA, cytoskeletal componentsregulate CPT-I. This novel model of regulation needsfurther exploration.

6.3. Pharmaceutical regulation ^ drug development

One approach to formulating an e¡ective drug fortype II (non-insulin-dependent) diabetes mellitus(NIDDM) is to design selective inhibitors of L-CPT-I. Anderson [187] points out that L-CPT-I isa better target than M-CPT-I or CPT-II. L-CPT-Icatalyses a key step in supplying fatty acyl groupsto fatty acid oxidation. Controlling fatty acid oxida-tion can regulate blood glucose levels and amelioratesome symptoms of NIDDM, a condition that ac-counts for over 90% of the cases of diabetes [188].Inhibitors of L-CPT-I can decrease fatty acid oxida-tion and, therefore, serve as adjuvant therapeuticagents to help manage NIDDM. [187]. Isozyme-se-lective inhibitors o¡er the potential of minimisingundesirable side e¡ects.

In order to treat NIDDM, these inhibitors mustovercome the challenges described in Anderson'strenchant review [187]. In cynomolgus monkeys,SDZ-CPI-975, a reversible inhibitor, lowers bloodglucose without inducing the cardiac hypertrophycaused by etomoxir, an irreversible inhibitor [189].SDZ-CPI-975 causes `hepatic mitochondrial aberra-tions' and development of this drug has slowed [187].Anderson rightfully concludes, ``major issues wouldneed to be more critically examined before commit-ting to full development'' of an L-CPT-I inhibitor asa drug.

Irreversible inhibitors are being developed cau-tiously because of myocardial hypertrophy. Etomoxirhas been used in two recent studies on humans. In adietary study, etomoxir stimulated appetite in sub-jects who had a high dietary intake of fat [190]. In-hibiting fatty acid oxidation signals a desire to eat inthose subjects who consume a lot of fat. In a limitedstudy with patients who su¡ered from chronic heartfailure, etomoxir improved the clinical status of thepatients and showed no side e¡ects of long-term (3months) administration [191]. In congestive heartfailure, calcium homeostasis is impaired and certaincontractile proteins are altered. Through gene expres-sion, etomoxir enhances the levels of sarcoplasmic

BBAPRO 36383 7-3-01


reticulum Ca2�-ATPase and K-myosin heavy chainprotein. These clinical observations may encourageclinical studies of other CPT-I inhibitors.

Developing isozyme-selective inhibitors of CPT re-mains a viable goal for the following reasons: (1)CPTs have functions beyond the liver and (2) acyl-carnitines can modulate the activity of other en-zymes. For example, L- and M-CPT-I have roles insperm maturation [192], so a selective CPT inhibitormight serve as a male contraceptive agent.

Palmitoylcarnitine inhibits protein kinase C inneuroblastoma NB-2a cells [193]; a palmitoylcarni-tine analogue might inhibit tumour cell proliferation.Inhibition of CPT-I produces palmitoyl-CoA, whichis fuel for de novo ceramide synthesis that in turnleads to apoptosis. Palmitoylcarnitine, a lysophos-pholipase transacylase inhibitor [194], interfereswith Candida adherence to lysophospholipids andthe HEp-2 cell line [195]; a palmitoylcarnitine ana-logue might serve as an antimicrobial agent [196].Consequently, continued development of acylcarni-tine analogues o¡ers bene¢ts beyond controllingNIDDM.

7. Future

The carnitine system is being studied in speciesother than mammals. For example, metabolic studiesin ¢sh indicate that the relationship between fattyacid oxidation and CPT levels and the sensitivity ofCPT-I to malonyl-CoA is similar to mammalian ones[197]. Plants have carnitine and carnitine acyltrans-ferase activities, but the proteins are unidenti¢ed.Although CPT activity exists on either side of thepea leaf chloroplast inner membrane, only a 20-kDa protein cross-reacted with antiserum to beefheart CPT-II [198]. A sequence from the Arabidopsisthaliana genome sequencing project is annotated ashomologous to the transporter, CACT, but phyloge-netic analyses point to a closer relationship with or-nithine translocators (F.R. van der Leij, unpub-lished). Some dispute the physiological signi¢canceof the carnitine system in plants, yet plants, like ani-mals, use limited pools of CoA. Carnitine in plantsmost likely facilitates the transport of activated fattyacids during desaturation, elongation and lipid syn-thesis ^ e.g., during periods of rapid membrane syn-

thesis ^ but also during lipid mobilisation and trans-port to the glyoxysome.

Sequencing of genomes and transcripts of variousorganisms continues to reveal genes and proteinsthat belong to the family of carnitine/choline acyl-transferases. Gene phylogeny studies have nowbeen carried out on about 50 genes of this family(F.R. van der Leij et al., unpublished). The ¢rst re-sults of maximum likelihood and parsimony analysesof human and yeast genes, known to encode activetransferases, are given in [63]. For CPT1A andCPT1B the general picture is that an ancestralCPT1 gene duplicated probably when or even beforevertebrates evolved. CPT1 genes are the closest rela-tives, a conclusion supported by conservation of theirexon junction positions. Human genes for COT,CAT and CPT-II are more closely related to eachother than to the genes for CPT-I. Apparently,CPT-I and CPT-II are the most distant members ofthe family. This distance between CPT-I and CPT-IIis independent of the di¡erences in the C- and N-termini. The relation of the human genes to the yeastCAT2 gene for mitochondrial/peroxisomal CAT(Cat2p) and the YAT1 gene for Yat1p points to sep-arate branches for human CAT and Cat2p. How-ever, these branches are at central positions and thedistance between CAT and Cat2p is less than theintraspecies distances between Cat2p and Yat1p orbetween CPT-I and CPT-II. Therefore, although theroot of the tree has not been de¢ned, it is likely thata common ancestor of the carnitine/choline acyl-transferase family is a CAT-like enzyme. Further-more, YAT1 shares an ancestral gene with CPT2that is not shared by the other genes, suggestingthe possible conservation of protein functions despitetheir di¡erences in subcellular localisation and sub-strate speci¢city.

The sequence and molecular genetics informationgenerated in the last 10 years has opened up themolecular studies of carnitine proteins and facilitateddissection of their multiple roles and intracellularlocations. Still awaiting elucidation are microsomalCPTs and further insight into their role in lipopro-tein synthesis (see [89]). Molecular probing of CPT-Ihas begun to give a picture of CPT-I and its physio-logical regulation by malonyl-CoA for the controlof fatty acid oxidation. The next big advancewill be the crystal structures, both as an aid to

BBAPRO 36383 7-3-01


understanding the mechanism and as a tool in drugdesign.

Acknowledgements

The authors are grateful to Drs V.A. Zammit andN.T. Price for fertile discussions and for supportfrom Sigma-Tau (R.R.R.), the National Institutesof Health GM42016 (R.D.G.) and the NetherlandsHeart Foundation NHS97.093 (F.R.v.d.L.).

References

[1] J. Bremer, Physiol. Rev. 63 (1983) 1420^1479.[2] C.T.W. van Roermund, E.H. Hettema, M. van den Berg,

H.F. Tabak, R.J.A. Wanders, EMBO J. 18 (1999) 5843^5852.

[3] A. Arduini, G. Mancinelli, G.L. Radatti, G. Dottori, F.Molajoni, R.R. Ramsay, J. Biol. Chem. 268 (1992) 12673^12681.

[4] M.S. Rashed, P.T. Ozand, M.J. Bennett, J.J. Barnard, D.R.Govindaraju, P. Rinaldo, Clin. Chem. 41 (1995) 1109^1114.

[5] E. Scho«mig, F. Spitzenberger, M. Engelhardt, F. Martel, N.Ording, D. Grundemann, FEBS Lett. 425 (1998) 79^86.

[6] I. Tamai, R. Ohashi, J. Nezu, H. Yabucchi, A. Oku, M.Shimane, Y. Sai, A. Tsuji, J. Biol. Chem. 273 (1998)20378^20382.

[7] X. Wu, P.D. Prasad, F.H. Leibach, V. Ganapathy, Biochem.Biophys. Res. Commun. 246 (1998) 589^595.

[8] T. Sekine, H. Kusuhara, N. Utsunomiya-Tate, M. Tsuda, Y.Sugiyama, Y. Kanai, H. Endou, Biochem. Biophys. Res.Commun. 251 (1998) 586^591.

[9] Y. Shoji, A. Koizumi, T. Kayo, T. Ohata, T. Takahashi, K.Harada, G. Takada, Am. J. Hum. Genet. 63 (1998) 101^108.

[10] T. Koizumi, H. Nikaido, J. Hayakawa, A. Nonomura, T.Yoneda, Lab. Anim. 22 (1988) 83^87.

[11] H. Nikaido, M. Horiuchi, N. Hasimoto, T. Saheki, J. Hay-akawa, Mamm. Genome 6 (1995) 369^370.

[12] J.I. Nezu, I. Tamai, A. Oku, R. Ohashi, H. Yabuuchi, N.Hashimoto, H. Nikaido, A. Sai, K. Koizumi, Y. Shoji, G.Takada, T. Matsuishi, M. Yoshino, H. Kato, T. Ohura, G.Tsujimoto, J. Hayakawa, M. Shimane, A. Tsuji, NatureGenet. 21 (1999) 91^94.

[13] K.M. Lu, H. Nishimori, Y. Nakamura, K. Shima, M. Ku-wajima, Biochem. Biophys. Res. Commun. 252 (1998) 590^594.

[14] P. Seth, X. Wu, W. Huang, F.H. Leibach, V. Ganapathy,J. Biol. Chem. 274 (1999) 33388^33392.

[15] H. Yabuuchi, I. Tamai, J.I. Nezu, K. Sakamoto, A. Oku, M.Shimane, Y. Sai, A. Tsuji, J. Pharmacol. Exp. Ther. 289(1999) 768^773.

[16] Y.W. Zhu, M.C. Jong, K.A. Frazer, E. Gong, R.M. Krauss,J.F. Cheng, D. Bo¡elli, E.M. Rubin, Proc. Natl. Acad. Sci.USA 97 (2000) 1137^1142.

[17] M. Kuwajima, M. Horiuchi, H. Harashima, K.M. Lu, M.Hayashi, S. Sei, K. Ozaki, T. Kudo, H. Kamido, A. Ono, T.Saheki, K. Shima, FEBS Lett. 443 (1999) 261^266.

[18] A.M. Lamhonwah, I. Tein, Biophys. Res. Commun. 252(1998) 396^401.

[19] N.L.S. Tang, V. Ganapathy, X. Wu, J. Hui, P. Seth, P.M.P.Yuen, T.F. Fok, N.M. Hjelm, Hum. Mol. Genet. 8 (1999)655^660.

[20] T. Meissner, B. Beinbrech, E. Mayatepek, Hum. Mutat. 13(1999) 351^361.

[21] A. Koizumi, J. Nozaki, T. Ohura, T. Kayo, Y. Wada, J.Nezu, R. Ohashi, I. Tamai, Y. Shoji, G. Takada, S. Kibira,T. Matsuishi, A. Tsuji, Hum. Mol. Genet. 8 (1999) 2247^2254.

[22] F.M. Vaz, H.R. Scholte, J. Ruiter, L.M. Hussaarts-Odijk,R.R. Pereira, S. Schweitzer, J.B.C. de Klerk, H.R. Water-ham, R.J.A. Wanders, Hum. Genet. 105 (1999) 157^161.

[23] Y.H. Wang, M.A. Kelly, T.M. Cowan, N. Longo, Hum.Mutat. 15 (2000) 238^245.

[24] D.B. Shennan, A. Grant, R.R. Ramsay, C. Burns, V.A.Zammit, Biochim. Biophys. Acta 1393 (1998) 49^56.

[25] R.Z. Christiansen, J. Bremer, Biochim. Biophys. Acta 448(1976) 562^577.

[26] T. Bohmer, K. Eiklid, J. Jonsen, Biochim. Biophys. Acta 465(1977) 627^633.

[27] D.J. Pearson, P.K. Tubbs, Biochem. J. 105 (1967) 953^963.[28] J.D. McGarry, C. Robles-Valdes, D.W. Foster, J. Biol.

Chem. 253 (1978) 8294^8300.[29] B. Gustafson, L.A. Ransna«s, Biochem. Biophys. Res. Com-

mun. 233 (1997) 752^755.[30] X. Wu, W. Huang, P.D. Prasad, P. Seth, D.P. Rajan, F.H.

Leibach, J.W. Chen, S.J. Conway, V. Ganapathy, J. Phar-macol. Exp. Ther. 290 (1999) 1482^1492.

[31] A. Sandor, G. Kispal, B. Melegh, I. Alkonyi, Biochim. Bio-phys. Acta 835 (1985) 83^91.

[32] C. Indiveri, V. Jacobazzi, N. Giangregorio, F. Palmieri, Bio-chem. J. 321 (1997) 713^719.

[33] M. Huizing, V. Iacobazzi, L. IJlst, P. Savelkoul, W. Ruiten-beek, L. van der Heuvel, C. Indiveri, J. Smeitink, K. Trijbels,R. Wanders, F. Palmieri, Am. J. Hum. Genet. 61 (1997)1239^1245.

[34] C. Indiveri, A. Tonazzi, F. Palmieri, Biochim. Biophys. Acta1020 (1990) 81^86.

[38] L. Palmieri, F.M. Lasorsa, V. Iacobazzi, M.J. Runswick, F.Palmieri, J.E. Walker, FEBS Lett. 462 (1999) 472^476.

[39] R.R. Ramsay, P.K. Tubbs, FEBS Lett. 54 (1975) 21^25.[40] S.V. Pande, Proc. Natl. Acad. Sci. USA 72 (1975) 883^887.[41] C. Indiveri, A. Tonazzi, F. Palmieri, Biochim. Biophys. Acta

1189 (1994) 65^73.[42] F. Fraser, V.A. Zammit, FEBS Lett. 445 (1999) 41^44.[43] S.V. Pande, R. Parvin, J. Biol. Chem. 25 (1980) 2994^3001.[44] C. Indiveri, A. Tonazzi, T. Dierks, R. Kra«mer, F. Palmieri,

Biochim. Biophys. Acta 1140 (1992) 53^58.

BBAPRO 36383 7-3-01


[45] M.A. Markwell, E.J. McGroarty, L.L. Bieber, N.E. Tolbert,J. Biol. Chem. 248 (1973) 3426^3432.

[46] A.S. Abbas, G.X. Wu, H. Schulz, J. Mol. Cell. Cardiol. 30(1998) 1305^1309.

[47] O. Corti, G. Finocchiaro, E. Rossi, O. Zu¡ardi, S. DiDona-to, Genomics 23 (1994) 94^99.

[48] O. Corti, S. DiDonato, G. Finocchiaro, Biochem. J. 303(1994) 3472^3479.

[49] G. Kispal, H. Steiner, D.A. Court, B. Rolinski, R. Lill,J. Biol. Chem. 40 (1996) 24458^24464.

[50] Y. Elgersma, C.W. van Roermund, R.J. Wanders, H.E. Ta-bak, EMBO J. 14 (1995) 3472^3479.

[51] M. Ueda, H. Kawachi, H. Atomi, A. Tanaka, Biochim. Bio-phys. Acta 1397 (1998) 213^222.

[52] J.A. McNew, J.M. Goodman, Trends Biochem. Sci. 21(1996) 54^58.

[53] W. Schmalix, W. Bandlow, J. Biol. Chem. 268 (1993) 27428^27439.

[54] T.M. Johnson, H.P. Kocher, R.C. Anderson, G.M. Neme-cek, Biochem. J. 305 (1995) 439^444.

[55] S. Ferdinadusse, J. Mulders, L. IJlst, S. Dennis, G. Dacre-mont, H.R. Waterham, R.J. Wanders, Biochem. Biophys.Res. Commun. 263 (1999) 213^218.

[56] S.J. Choi, D.H. Oh, C.S. Song, A.K. Roy, B. Chatterjee,Biochim. Biophys. Acta 1264 (1995) 215^222.

[57] C.N. Cronin, Eur. J. Biochem. 247 (1997) 1029^1037.[58] C. Caudevilla, D. Serra, A. Miliar, C. Codony, G. Asins, M.

Bach, F.G. Hegardt, Proc. Natl. Acad. Sci. USA 95 (1998)12185^12190.

[59] G. Finocchiaro, F. Taroni, M. Rocchi, A.L. Martin, I. Co-lombo, G.T. Tarelli, S. DiDanato, Proc. Natl. Acad. Sci.USA 88 (1991) 661^665.

[60] K.F. Woeltje, V. Esser, B.C. Weis, A. Sen, W.F. Cox, M.J.McPhaul, C.A. Slaughter, D.W. Foster, J.D. McGarry,J. Biol. Chem. 265 (1990) 10720^10725.

[61] N.F. Brown, V. Esser, A.D. Gonzalez, C.T. Evans, C.A.Slaughter, D.W. Foster, J.D. McGarry, J. Biol. Chem. 266(1991) 15446^15449.

[62] E. Verderio, P. Cavadini, L. Montermini, H. Wang, E. La-mantea, G. Finocchiaro, S. DiDonato, C. Gellera, F. Taro-ni, Hum. Mol. Genet. 4 (1995) 19^29.

[63] F.R. van der Leij, N.C.A. Huijkman, C. Boomsma, J.R.G.Kuipers, B. Bartelds, Mol. Genet. Metab. 71 (2000) 139^153.

[64] B.D. Gelb, Genomics 18 (1993) 651^655.[65] J.D. McGarry, N.F. Brown, Eur. J. Biochem. 244 (1997) 1^

14.[66] V. Esser, C.H. Britton, B.C. Weis, D.W. Foster, J.D.

McGarry, J. Biol. Chem. 268 (1993) 5817^5822.[67] E.A. Park, M.L. Ste¡en, S. Cook, V.M. Park, G.A. Park,

Biochem. J. 330 (1998) 217^224.[68] C.H. Britton, R.A. Schultz, B. Zhang, V. Esser, D.W. Fos-

ter, J.D. McGarry, Proc. Natl. Acad. Sci. USA 92 (1995)1984^1988.

[69] K.B. Cox, K.R. Johnson, P.A. Wood, Mamm. Genome 9(1998) 608^610.

[70] B. Bartelds, Thesis, University of Groningen, 2000.

[71] M.P. Kolodziej, V.A. Zammit, FEBS Lett. 327 (1993) 294^296.

[72] I. Cohen, C. Kolh, J.D. McGarry, J. Girard, C. Prip-Buus,J. Biol. Chem. 273 (1998) 29896^29904.

[73] F. Fraser, C.G. Cortorphine, V.A. Zammit, Biochem. J. 323(1997) 711^718.

[74] F.R. van der Leij, A.M. Kram, B. Bartelds, H. Roelofsen,G.B. Smid, J. Takens, V.A. Zammit, J.R. Kuipers, Biochem.J. 341 (1999) 777^784.

[75] N. Yamazaki, Y. Shinohara, A. Shima, H. Terada, FEBSLett. 363 (1995) 41^45.

[76] V. Esser, N.F. Brown, A.T. Cowan, D.W. Foster, J.D.McGarry, J. Biol. Chem. 271 (1996) 6972^6977.

[77] N. Yamazaki, Y. Shinohara, A. Shima, Y. Yamanaka, H.Terada, Biochim. Biophys. Acta 1307 (1996) 157^161.

[78] F.R. van der Leij, J. Takens, A.Y. van der Veen, P. Terpstra,J.R. Kuipers, Biochim. Biophys. Acta 1352 (1997) 123^128.

[79] H. Zhu, J. Shi, Y. de Vries, D.N. Arvidson, J.M. Cregg, G.Woldegiorgis, Arch. Biochem. Biophys. 347 (1997) 53^61.

[80] G.S. Yu, Y.C. Lu, T. Gulick, J. Biol. Chem. 273 (1998)32901^32909.

[81] G.S. Yu, Y.C. Lu, T. Gulick, Biochim. Biophys. Acta 1393(1998) 166^172.

[82] N. Yamazaki, Y. Yamanaka, Y. Hashimoto, Y. Shinohara,A. Shima, H. Terada, FEBS Lett. 409 (1997) 401^406.

[83] N. Yamazaki, Y. Shinohara, K. Kajimoto, M. Shindo, H.Terada, J. Biol. Chem. 275 (2000) 31739^31746.

[84] D. Wang, W. Harrison, L.M. Buja, F.F. Elder, J.B. McMil-lin, Genomics 48 (1998) 314^323.

[85] J.R. De Lucas, A.I. Dominguez, Arch. Microbiol. 171 (1999)386^396.

[86] V.N. Jackson, J.M. Cameron, V.A. Zammit, N.T. Price,Biochem. J. 341 (1999) 483^489.

[87] F. Fraser, C.G. Costorphine, N.T. Price, V.A. Zammit,FEBS Lett. 446 (1999) 69^74.

[88] R.R. Ramsay, Life Sci. 12 (1993) 23^29.[89] N.M. Broadway, J.M Gooding, E.D. Saggerson, in: P.A.

Quant, S. Eaton (Eds.), Current Views of Fatty Acid Oxida-tion and Ketogenesis : From Organelles to Point Mutations,Kluwer Academic/Plenum Publishing, New York, 1999, pp.59^67.

[90] K.A.H. Abo-Hashema, M.H. Cake, M.A. Lukas, J. Knud-sen, Biochemistry 38 (1999) 15840^15847.

[91] M.S. Murthy, S.V. Pande, Biochem. J. 304 (1994) 31^34.[92] M.S. Murthy, S.V. Pande, in: C. De Simone, G. Famularo

(Eds.), Carnitine Today, Springer-Verlag, Heidelberg, 1997,pp. 39^70.

[93] F. Taroni, E. Verderio, S. Fiorucci, P. Cavadini, G. Finoc-chiaro, G. Uziel, E. Lamantea, C. Gellera, S. DiDonato,Proc. Natl. Acad. Sci. USA 89 (1992) 8429^8433.

[94] J.P. Bonnefont, E. Demaugre, C. Prip-Buus, J.M. Saudu-bray, M. Brivet, N. Abadi, L. Thuillier, Mol. Genet. Metab.68 (1999) 424^440.

[95] R.T. Taggart, D. Smail, C. Apolito, G.D. Vladutiu, Hum.Mut. 13 (1999) 210^220.

BBAPRO 36383 7-3-01


[96] F. Taroni, E. Verderio, F. Dworzak, P.J. Willems, P. Cav-adini, S. DiDonato, Nature Genet. 4 (1993) 314^320.

[97] I. Handig, E. Dams, F. Taroni, S. Van Laere, T. de Barsy,P.J. Willems, Hum. Genet. 97 (1996) 291^293.

[98] B.Z. Yang, J.H. Ding, T. Dewese, D. Roe, G.C. He, J.Wilkinson, D.W. Day, F. Demaugre, D. Rabier, M. Brivet,C. Roe, Mol. Genet. Metab. 64 (1998) 229^236.

[99] S. Yamamoto, H. Abe, T. Kohgo, A. Ogawa, A. Ohtake,H. Hayashibe, H. Sakuraba, Y. Suzuki, S. Aramaki, M.Takayanagi, S. Hasegawa, H. Niimi, Hum. Genet. 98(1996) 116^118.

[100] K. Wataya, J. Akanuma, P. Cavadini, Y. Aoki, S. Kure, F.Invernizzi, I. Yoshida, J. Kira, R. Taroni, Y. Matsubara,K. Narisawa, Hum. Mut. 11 (1998) 377^386.

[101] J.P. Bonnefont, F. Taroni, P. Cavadini, C. Cepanec, M.Brivet, J.M. Saudubray, J.P. Leroux, F. Demaugre, Am.J. Hum. Genet. 58 (1996) 971^978.

[102] L. IJlst, H. Mandel, W. Oostheim, J.P. Ruiter, A. Gutman,R.J. Wanders, J. Clin. Invest. 102 (1998) 527^531.

[103] N.F. Brown, R.C. Anderson, S.L. Caplan, D.W. Foster,J.D. McGarry, J. Biol. Chem. 269 (1994) 19157^19162.

[104] V.A. Zammit, Prog. Lipid Res. 38 (1999) 199^224.[105] R.R. Ramsay, Am. J. Med. Sci. 318 (1999) 28^35.[106] W.R. Huckle, T.M. Tamblyn, Arch. Biochem. Biophys. 226

(1983) 94^110.[107] R.R. Ramsay, J.P. Derrick, A.S. Friend, P.K. Tubbs, Bio-

chem. J. 244 (1987) 271^278.[108] P.R.H. Clarke, L.L. Bieber, J. Biol. Chem. 256 (1981)

9861^9868.[109] G. Finocchiaro, I. Colombo, S. DiDonato, FEBS Lett. 274

(1990) 163^166.[110] M.P. Kolodziej, V.A. Zammit, Biochem. J. 267 (1990) 85^

90.[111] C.J. Fiol, J. Kerner, L.L. Bieber, Biochim. Biophys. Acta

916 (1987) 482^492.[112] E.D. Saggerson, C.A. Carpenter, FEBS Lett. 137 (1982)

124^128.[113] M.S.R. Murthy, S.V. Pande, Biochem. J. 248 (1987) 727^

733.[114] V.A. Zammit, Biochem. J. 214 (1983) 1027^1030.[115] C.J. Fiol, L.L. Beiber, Lipids 23 (1988) 120^125.[116] F.M. Gon¬i, A.M. Requero, A. Alonso, FEBS Lett. 390

(1996) 1^5.[117] K. Bartlett, P. Bartlett, N. Bartlett, H.S.A. Sherratt, Bio-

chem. J. 229 (1985) 559^560.[118] D.F. Pauly, J.B. McMillin, J. Biol. Chem. 263 (1988)

18160^18167.[119] A.K.M.J. Bhuiyan, S.V. Pande, Mol. Cell. Biochem. 139

(1994) 109^116.[120] J.D. McGarry, S.E. Mills, C.S. Long, D.W. Foster, Bio-

chem. J. 214 (1983) 21^28.[121] N. N|c a'Bhaird, G. Kumaravel, R.D. Gandour, M.J.

Krueger, R.R. Ramsay, Biochem. J. 294 (1993) 645^651.[122] J.F.A. Chase, P.K. Tubbs, Biochem. J. 99 (1966) 32^40.[123] C.D. Chung, L.L. Bieber, J. Biol. Chem. 268 (1993) 1519^

1524.

[124] J.A. Idell-Wenger, L.W. Grotyohann, J.R. Neely, J. Biol.Chem. 259 (1978) 4310^4318.

[125] E.P. Brass, L.J. Ru¡, J. Nutr. 122 (1992) 2094^2100.[126] J.F.A. Chase, Biochem. J. 104 (1967) 503^509.[127] B.C. Weis, V. Esser, D.W. Foster, J.D. McGarry, J. Biol.

Chem. 269 (1994) 18712^18715.[128] J. Shi, H. Zhu, D.N. Arvidson, G. Woldegiorgis, Biochem-

istry 39 (2000) 712^717.[129] S.T. Swanson, D.W. Foster, J.D. McGarry, N.F. Brown,

Biochem. J. 335 (1998) 513^519.[130] V.N. Jackson, J.M. Cameron, F. Fraser, V.A. Zammit,

N.T. Price, J. Biol. Chem. 275 (2000) 19560^19566.[131] F. Fraser, V.A. Zammit, Biochem J. 329 (1998) 225^229.[132] C.L. Hoppel, P. Turkaly, J. Kerner, B. Tandler, FASEB J.

(1996) 11, A445 (Abstr. 2577).[133] H. Zhu, J. Shi, J.M. Cregg, G. Woldegiorgis, Biochem.

Biophys. Res. Commun. 239 (1997) 498^502.[134] R.D. Gandour, O. Leung, A.T. Greway, R.R. Ramsay, N.

N|c a'Bhaird, F.R. Fronczek, B.M. Bellard, G. Kumaravel,J. Med. Chem. 36 (1993) 237^242.

[135] K. Burton, Biochem. J. 59 (1955) 44^46.[136] R.M. Burton, E.R. Stadtman, J. Biol. Chem. 202 (1953)

873^890.[137] I.B. Fritz, S.K. Schultz, P.A. Srere, J. Biol. Chem. 238

(1963) 2509^2517.[138] D.M. Mu«ller, E. Strack, Hoppe Seylers Z. Physiol. Chem.

354 (1973) 1091^1096.[139] W.P. Jencks, S. Cordes, J. Carriuolo, J. Biol. Chem. 235

(1960) 3608^3614.[140] P. George, R.J. Witonsky, M. Trachtman, C. Wu, W. Dor-

wart, L. Richman, W. Richman, F. Churayh, B. Lentz,Biochim. Biophys. Acta 223 (1970) 1^15.

[141] D.M. Hayes, G.L. Kenyon, P.A. Kollman, J. Am. Chem.Soc. 100 (1978) 4331^4340.

[142] R. Wolfenden, Y.L. Liang, Bioorg. Chem. 17 (1989) 486^489.

[143] J.F.A. Chase, P.K. Tubbs, Biochem. J. 116 (1970) 713^720.[144] M. Morillas, J. Clotet, B. Rubi, D. Serra, G. Asins, J.

Arino, F.G. Hegardt, FEBS Lett. 466 (2000) 183^186.[145] N. N|c a'Bhaird, V. Yankovskaya, R.R. Ramsay, Biochem.

J. 330 (1998) 1029^1036.[146] J.L. Sussman, M. Harel, F. Frowlow, C. Oefner, A. Gold-

man, L. Toker, I. Silman, Science 253 (1991) 872^879.[147] T.L. Hutchinson, A. Saeed, P.E. Wolkowicz, J.B. McMil-

lin, W.J. Brouillette, Bioorg. Med. Chem. 7 (1999) 1505.[148] A. Saeed, J.B. McMillin, P.E. Wolkowicz, W.J. Brouillette,

Arch. Biochem. Biophys. 305 (1993) 307^312.[149] G.F. Tutwiler, P. Dellevigne, J. Biol. Chem. 254 (1979)

2935.[150] G.D. Lopaschuk, S.R. Wall, P.M. Olley, N.J. Davis, Circ.

Res. 63 (1988) 1036^1043.[151] T.L. Hutchinson, W.J. Brouillette, Bioorg. Med. Chem. 6

(1998) 2133.[152] B.C. Weis, A.T. Cowan, N. Brown, D.W. Foster, J.D.

McGarry, J. Biol. Chem. 269 (1994) 26443^26448.[153] S.-S. Jew, E.-K. Kim, S.-M. Je, L.-X. Zhao, H.-O. Kim,

BBAPRO 36383 7-3-01


H.G. Park, K.H. Ko, W.-K. Kim, H.-J. Kim, J.H. Cheong,E.-S. Lee, Heterocycles 52 (2000) 1087.

[154] R.C. Anderson, M. Balestra, P.A. Bell, R.O. Deems, W.S.Fillers, J.E. Foley, J.D. Fraser, W.R. Mann, M. Rudin,E.B. Villhauer, J. Med. Chem. 38 (1995) 3448^3450.

[155] D.L. Jenkins, O.W. Gri¤th, Proc. Natl. Acad. Sci. USA 83(1986) 290^294.

[156] S. Shinagawa, T. Kanamaru, S. Harada, M. Asai, H. Oka-zaki, J. Med. Chem. 30 (1987) 26443^26463.

[157] W.R. Mann, C.J. Dragland, D.K. Cohen, P.A. Bell, W.S.Fillers, Diabetes 44 (Suppl. 1) (1995) 72A.

[158] W.S. Fillers, D.K. Cohen, S.C. Weldon, P.A. Bell, Diabetes44 (Suppl. 1) (1995) 131A.

[159] B.D. Grantham, V.A. Zammit, Biochem. J. 243 (1987) 261^265.

[160] E.A. Park, R.L. Mynatt, G.A. Cook, K. Kash¢, Biochem.J. 310 (1995) 853^858.

[161] E.K. Hudson, M.H. Liu, L.M. Buja, J.B. McMillin, J. Mol.Cell. Cardiol. 27 (1995) 599^613.

[162] E.D. Saggerson, C.A. Carpenter, FEBS Lett. 137 (1983)124^128.

[163] Y. Gavel, G. von Heijne, Protein Eng. 4 (1990) 33^37.[164] R.L. Mynatt, E.A. Park, F.E. Thorngate, H.K. Das, G.A.

Cook, Biochem. Biophys. Res. Commun. 201 (1994) 932^937.

[165] N.F. Brown, B.C. Weis, J.E. Husti, D.W. Foster, J.D.McGarry, J. Biol. Chem. 270 (1995) 8952^8957.

[166] P.S. Brady, R.R. Ramsay, L.J. Brady, FASEB J. 7 (1993)1039^1044.

[167] N.F. Brown, J.K. Hill, V. Esser, J.L. Kirkland, B.E. Cor-key, D.W. Foster, J.D. McGarry, Biochem. J. 327 (1997)225^231.

[168] T. Gulick, S. Cresci, T. Caira, D.D. Moore, D.P. Kelly,Proc. Natl. Acad. Sci. USA 91 (1994) 11012^11016.

[169] T. Aoyama, J.M. Peters, N. Iritani, T. Nakajima, K. Fur-ihata, T. Hashimoto, F.J. Gonzalez, J. Biol. Chem. 273(1998) 5678^5684.

[170] J.M. Brandt, F. Djouadi, D.P. Kelly, J. Biol. Chem. 273(1998) 23786^23792.

[171] S.S.T. Lee, T. Pineau, J. Drago, E.J. Lee, J.W. Owens, D.L.Kroetz, P.M. Fernandezsalguero, H. Westphal, F.J. Gon-zalez, Mol. Cell. Biol. 15 (1995) 3012^3022.

[172] S. Kersten, J. Seydoux, J.M. Peters, F.J. Gonzalez, B. Des-vergne, W. Wahli, J. Clin. Invest. 103 (1999) 1489^1498.

[173] T.C. Leone, C.J. Weinheimer, D.P. Kelly, Proc. Natl.Acad. Sci. USA 96 (1999) 7473^7478.

[174] C. Mascaro, E. Acosta, J.A. Ortiz, P.F. Marrero, F.G.Hegardt, D. Haro, J. Biol. Chem. 273 (1998) 8560^8563.

[175] M.N. Sack, D.L. Disch, H.A. Rockman, D.P. Kelly, Proc.Natl. Acad. Sci. USA 94 (1997) 6438^6443.

[176] M.L. Ste¡en, W.R. Harrison, F.F. Elder, G.A. Cook, E.A.Park, Biochem. J. 340 (1999) 425^432.

[177] S. Brunner, K. Kramar, D.T. Denhardt, R. Hofbauer, Bio-chem. J. 322 (1997) 403^410.

[178] R. Uenaka, M. Kuwajima, A. Ono, Y. Matsuzawa, J. Hay-

akawa, N. Inohara, Y. Kagawa, S. Ohta, J. Biochem. 119(1996) 533^540.

[179] N. Nic a'Bhaird, R.R. Ramsay, Biochem. J. 286 (1992)637^640.

[180] J.D. McGarry, G.F. Leatherman, D.W. Foster, J. Biol.Chem. 253 (1978) 4128^4136.

[181] V.N. Jackson, V.A. Zammit, N.T. Price, J. Biol. Chem. 275(2000) 38410^38416.

[182] I. Cohen, C. Kohl, J.D. McGarry, J. Girard, C. Prip-Buus,J. Biol. Chem. 273 (1998) 29896^29904.

[183] T.W. Stephens, G.A. Cook, R.A. Harris, Biochem. J. 212(1983) 521^524.

[184] P.S. Savle, S.V. Pande, T.-S. Lee, R.D. Gandour, Bioorg.Med. Chem. Lett. 9 (1999) 3099^3102.

[185] J.D. McGarry, N.F. Brown, Biochem. J. 349 (2000) 179^187.

[186] M. Guzman, G. Velasco, M.J.H. Geelen, Trends Endocrin.Metab. 11 (2000) 49^53.

[187] R.C. Anderson, Curr. Pharm. Design 4 (1998) 1^15.[188] Diabetes in America, 2nd edn., National Diabetes Data

Group, National Institutes of Health, National Instituteof Diabetes and Digestive and Kidney Diseases, NIH Pub-lication No. 95-1468, Bethesda, MD, 1995.

[189] R.O. Deems, R.C. Anderson, J.E. Foley, Am. J. Physiol.(Regul. Integr. Comp. Physiol. 43) 274 (1998) R524^R528.

[190] A. Kahler, M. Zimmerman, W. Langhans, Nutrition 15(1999) 819^828.

[191] S. Schmidt-Schweda, C. Holubarsch, Clin. Sci. 99 (2000)27^35.

[192] S.H. Adams, V. Esser, N.F. Brown, N.H. Ing, L. Johnson,D.W. Foster, J.D. McGarry, Biol. Reprod. 59 (1998) 1339^1405.

[193] K.A. Nalecz, J.E. Mroczkowska, U. Berent, M.J. Nalecz,Acta Neurobiol. Exp. 57 (1997) 263^274.

[194] R.W. Gross, B.E. Sobel, J. Biol. Chem. 258 (1983) 5221^5226.

[195] A. Prakobphol, H. Le¥er, C.I. Hoover, S.J. Fisher, FEMSMicrobiol. Lett. 151 (1997) 89^94.

[196] P.S. Savle, G.F. Doncel, S.D. Bryant, M.P. Hubieki, R.G.Robinette, R.D. Gandour, Bioorg. Med. Chem. Lett. 9(1999) 2545^2548.

[197] L. FrÖyland, L. Madsen, K.M. Eckho¡, Ò. Lie, R.K.Berge, Lipids 33 (1998) 923^930.

[198] C. Masterson, C. Wood, Can. J. Bot. 78 (2000) 328^335.[199] Y.H. Wang, T.A. Meadows, N. Longo, J. Biol. Chem. 275

(2000) 20782^20786.[200] B. Burwinkel, J. Kreuder, S. Schweitzer, M. Vorgerd, K.

Gempel, K.D. Gerbitz, M.W. Kilimann, Biochem. Biophys.Res. Commun. 261 (1999) 484^487.

[201] E. Mayatepek, J. Nezu, I. Tamai, A. Oku, M. Katsura, M.Shimane, A. Tsuji, Hum. Mutat. 15 (2000) 118.

[202] P. Kaufmann, M. El Schahawi, S. DiMauro, Mol. Cell.Biochem. 174 (1997) 237^239.

[203] T. Wieser, M. Deschauer, S. Zierz, Adv. Exp. Med. Biol.466 (1999) 339^345.

BBAPRO 36383 7-3-01


[204] F. Taroni, C. Gellera, P. Cavadini, S. Baratta, E. Laman-tea, S. Dethlefs, S. DiDonato, R.A. Reik, P.J. Benke, Am.J. Hum. Genet. 55 (1994) A245.

[205] C.N. Cronin, Biochem. Biophys. Res. Commun. 238 (1997)784^789.

[206] N. Abadi, L. Thuillier, C. Prasad, H. Belbachir, F. Demau-gre, C. Prip-Buus, M. Brivet, N. Khadom, J.P. Bonnefont,Proceedings 49th Annual Meeting of the American Societyof Human Genetics, San Francisco, CA, 19^26 October,1999.

[207] R.R. Ramsay, Biochem. Soc. Trans. 6 (1978) 72^76.[208] L. Baillet, R.S. Mullur, V. Esser, J.D. McGarry, J. Biol.

Chem. 275 (2000) 36766^36768.[209] V. Rosas-Garc|a, R.D. Gandour, J. Am. Chem. Soc. 119

(1997) 7587^7588.[210] I. Cohen, F. Guillerault, J. Girard, C. Prip-Buus, J. Biol.

Chem. 275 (2000) in press.[211] I. Tamai, R. Ohashi, J. Nezu, Y. Sai, D. Kobayashi, A.

Oku, M. Shimane, A. Tsuji, J. Biol. Chem. 275 (2000)40064^40072.

BBAPRO 36383 7-3-01


Documents

Molecular enzymology of carnitine transfer and transport