In the present paper, we describe the current state of knowledge regarding the enzymology of the phytanic acid α-oxidation pathway. The product of phytanic acid α-oxidation, i.e. pristanic acid, undergoes three cycles of β-oxidation in peroxisomes after which the products, including 4,8-dimethylnonanoyl-CoA, propionyl-CoA and acetyl-CoA, are exported from the peroxisome via one of two routes, including (i) the carnitine-dependent route, mediated by CRAT (carnitine acetyltransferase) and CROT (carnitine O-octanoyltransferase), and (ii) the free acid route, mediated by one or more of the peroxisomal ACOTs (acyl-CoA thioesterases). We also describe our recent data on the ω-oxidation of phytanic acid, especially since pharmacological up-regulation of this pathway may form the basis of a new treatment strategy for ARD (adult Refsum's disease). In patients suffering from ARD, phytanic acid accumulates in tissues and body fluids due to a defect in the α-oxidation system.
- acyl-CoA thioesterase
- fatty acid oxidation
- phytanic acid
- Refsum's disease
ARD (adult Refsum's disease) was first described in the 1940s by Sigvald Refsum and is characterized by early-onset retinitis pigmentosa and anosmia as universal abnormalities found in all ARD patients, with variable combinations of neuropathy, deafness, ataxia, ichthyosis and cardiac manifestations. In fact, cardiac arrhythmias and heart failure caused by cardiomyopathy are frequent causes of death. It should be noted that the full constellation of signs and symptoms is rarely seen in affected individuals and that most features develop with age .
The biochemical abnormality in ARD was discovered by Klenk and Kahlke  in 1964, who reported the accumulation of phytanic acid (3,7,11,15-tetramethylhexadecanoic acid) in plasma and tissues of an ARD patient. This discovery was rapidly followed by studies to identify the underlying defect by administering radiolabelled phytol, a precursor of phytanic acid, and phytanic acid itself to ARD patients, which revealed that [U-14C20]phytanic acid is rapidly oxidized to [14C1]CO2 in control subjects but not in ARD patients.
It was soon figured out that the most likely mechanism of phytanic acid oxidation would involve an initial oxidative decarboxylation step in which the terminal carboxy group would be released as CO2. The details of this mechanism remained fully unclear, however, until the late 1990s, when the enzyme phytanoyl-CoA 2-hydroxylase was discovered, which turned out to be the enzyme deficient in ARD  followed by resolution of the rest of the α-oxidation pathway.
In the present paper, we will review the enzymology of the phytanic acid α-oxidation pathway and its importance for ARD. Furthermore, we will describe our latest results, which deal with another mechanism to oxidize phytanic acid, i.e. via ω-oxidation.
Enzymology of the phytanic acid α-oxidation pathway
Phytanic acid as derived from dietary sources occurs in two stereoisoforms, including (3R)- and (3S)-phytanic acid, which can both be handled by the α-oxidation system yielding (2R)- and (2S)-pristanic acid. Since the peroxisomal β-oxidation system only accepts (2S)-acyl-CoAs, an additional peroxisomal enzyme called AMACR (α-methylacyl-CoA racemase) is required to oxidize (2R)-pristanic acid, but not (2S)-pristanic acid (Figure 1).
The enzymology of the α-oxidation pathway has been worked out in some detail and involves four subsequent enzymatic reactions in order to convert phytanic acid into pristanic acid (2,6,10,14-tetramethylpentadecanoic acid) (Figure 2A). Since pristanic acid now has the first methyl group at position 2 rather than position 3 as in phytanic acid, pristanic acid can undergo β-oxidation. Pristanic acid undergoes three cycles of β-oxidation in peroxisomes, with 4,8-dimethylnonanoyl-CoA as the end product, which is subsequently transported to mitochondria where full oxidation occurs . The phytanic acid α-oxidation pathway involves four subsequent steps, including the following.
Activation of phytanic acid to phytanoyl-CoA
Initial studies by Muralidharan and Muralidharan  had suggested that the activation of phytanic acid and palmitic acid might involve different enzymes, at least in rat liver microsomes. Subsequent work by Pahan et al.  led to the conclusion that peroxisomes contain a distinct acyl-CoA synthetase, specific for phytanic acid. Vanhooren et al.  studied the activation of a phytanic acid analogue (3-methyl margaric acid), a pristanic acid analogue (2-methyl palmitic acid), palmitic acid and lignoceric acid in rat liver peroxisomes, and concluded that different enzymes were involved. On the other hand, Watkins et al.  showed that the long-chain acyl-CoA synthetase cloned and sequenced by Suzuki et al. , now called ACSL1 (long-chain acyl-CoA synthase 1) , also has phytanoyl-CoA synthetase activity. Earlier work by Miyazawa et al.  had shown that this synthetase was present in the mitochondria, peroxisomes as well as in microsomes. Careful studies by Coleman and co-workers  have clearly shown that these results are erroneous, the reason being that the antibody as used by Miyazawa et al.  was raised against the complete ACSL1 protein. This antibody also detects other ACSL family members since they share highly conserved amino acid sequences giving rise to similar epitopes. So far, five distinct ACSL isoforms have been identified. This family is subclassified into two subfamilies with ACSL1, ACSL5 and ACSL6 constituting one family and ACSL3 and 4 constituting the other. Using specific peptide antibodies raised against ACSL1, ACSL4 and ACSL5, Lewin et al.  showed that ACSL4 is the only ASCL in peroxisomes. In contrast with ACSL5, which is localized in the mitochondria, and ASCL1, localized in MAM (mitochondria-associated membrane), ER (endoplasmic reticulum) and the cytosol, but not in the mitochondria and peroxisomes, ACSL4 is a peripheral membrane protein, whereas ACSL1 and ACSL5 are integral membrane proteins. The role of each of these ACSLs in the activation of phytanic acid as well as that of members of the FATP (fatty acid transport protein)/ACSVL (very-long-chain acyl-CoA synthase) family remain to be established (Figure 2B).
Hydroxylation of phytanoyl-CoA to 2-hydroxyphytanoyl-CoA
Phytanoyl-CoA 2-hydroxylase belongs to the family of the 2-oxoglutarate-dependent oxygenases, the largest known family of non-haem metal-dependent oxidizing enzymes, and was first purified from rat liver peroxisomes by Jansen et al. . The rat liver protein has a typical PTS2 (peroxisome-targeting signal 2) sequence (RLQIVLGHL), which directs the protein to peroxisomes. Similar PTS2 sequences were later found in other phytanoyl-CoA 2-hydroxylases. Interestingly, a putative Caenorhabditis elegans phytanoyl-CoA 2-hydroxylase lacks the N-terminal extension, containing the PTS2 sequence. Instead, the protein contains a typical PTS1 sequence (-RSNL) in line with the notion that in C. elegans only the PTS1 pathway is operational.
Cleavage of 2-hydroxyphytanoyl-CoA to pristanal
The subsequent enzyme in the pathway, i.e. 2-HPCL (2-hydroxyphytanoyl-CoA lyase)/HACL (hydroxyacyl-CoA lyase), was purified by Foulon et al.  and turned out to be a TPP (thiamin pyrophosphate)-dependent enzyme equipped with an atypical PTS1 signal. Recent work has shown that this enzyme accepts a range of 2-hydroxyacyl-CoAs, which implies that this enzyme probably plays a key role in the oxidation of a number of 2-hydroxy fatty acids .
Oxidation of pristanal to pristanic acid
The fourth step is catalysed by an as yet unidentified aldehyde dehydrogenase. Early studies  had suggested that the oxidation of pristanal to pristanic acid would take place at the ER membrane as catalysed by the microsomal FALDH (fatty aldehyde dehydrogenase) as encoded by ALDH3A2. This was concluded from experiments with fibroblasts from Sjogren–Larsson patients in which FALDH is deficient, due to mutations in the structural gene (ALDH3A2) encoding FALDH. Subsequent studies, notably by Jansen et al.  and Croes et al. , however, led to the identification of a distinct pristanal dehydrogenase in peroxisomes, which implies that α-oxidation of phytanoyl-CoA can proceed entirely within peroxisomes (see Figure 2B). The nature of this peroxisomal enzyme still remains to be established with certainty, although very recent studies have brought FALDH back into the arena. Indeed, according to Ashibe et al. , one of the splice variants of ALDH3A2, named FALDH-V, is specifically targeted to peroxisomes and may catalyse the oxidation of aldehydes including pristanal to the corresponding acids. The authors propose that the FALDH-V, which has a C-terminal extension of 27 amino acids as compared with the FALDH-N, which is strictly localized in the ER membrane, is targeted to peroxisomes by virtue of the C-terminal tripeptide SKH, which differs from the consensus PTS1 sequence S/A/C-K/R/H-L/M, however.
Another point of concern is the fact that according to Ashibe et al.  FALDH-V is an integral membrane protein in peroxisomes as well as in the ER membrane, whereas the pristanal dehydrogenase activity as identified by Jansen et al.  in peroxisomes is catalysed by a soluble peroxisomal dehydrogenase as concluded from experiments in which peroxisomes were sonicated followed by high-speed centrifugation. Taken together, the question of which peroxisomal enzyme actually catalyses the oxidation of pristanal to pristanic acid is definitely not yet resolved.
The next step in the pathway, i.e. the activation of pristanic acid to pristanoyl-CoA, is also not completely resolved. Early studies by us had already shown that the mitochondria, peroxisomes and microsomes contain pristanoyl-CoA synthetase activity . Unfortunately, the question whether the catalytic side is exposed to the cytosol or to the peroxisomal matrix was not addressed in these studies. More recently, Steinberg et al.  cloned the human orthologue of the gene coding for the rat liver VLCS (very-long-chain acyl-CoA synthetase) as purified and cloned by Hashimoto and co-workers [21,22] and showed that this enzyme is localized in peroxisomes and ER. Importantly, this enzyme, now called FATP2/ACSVL1, is topographically oriented towards the matrix, and reacts with a range of fatty acids, including pristanic acid and phytanic acid.
The fate of pristanoyl-CoA has been relatively well established thanks to the work of Verhoeven et al.  which showed that pristanoyl-CoA undergoes three cycles of β-oxidation in peroxisomes, after which the products, notably 2,6-dimethylnonanoyl-CoA, propionyl-CoA and acetyl-CoA, are shuttled to the mitochondria for full oxidation to CO2 and water.
In principle, there are at least two ways in which the CoA esters may be shuttled to the mitochondria. The first pathway involves the conversion of 4,8-dimethylnonanoyl-CoA, propionyl-CoA and acetyl-CoA into the corresponding carnitine esters as catalysed by CROT (carnitine O-octanoyltransferase) and CRAT (carnitine acetyltransferase) respectively followed by transport to the mitochondria, which possess a specific carrier, called the CACT (carnitine/acylcarnitine translocase), which allows the uptake of acylcarnitines from the cytosol into mitochondria in exchange for free carnitine. The other pathway involves hydrolytic cleavage of the different CoA esters by one of the ACOTs (acyl-CoA thioesterases) to produce the corresponding fatty acids , which may be further oxidized in the mitochondria, but may also undergo different fates. Early studies by Leighton et al. , for instance, have shown that in rat liver hepatocytes, acetate is the predominant product of dicarboxylic acid β-oxidation in peroxisomes and is rapidly excreted out of the cell. In vivo, the acetate released in this way may be oxidized to CO2 and H2O in other tissues, including the brain, and may thus be considered to be the third ketone body after acetoacetate and 3-hydroxybutyrate. It may well be that the contribution of the two different pathways varies between tissues and between cell types. In fibroblasts, for instance, it has been established that the carnitine-dependent route is the preferred route for the further oxidation of 4,8-dimethylnonanoyl-CoA, whereas in hepatocytes, at least for acetyl-CoA, the thioesterase-dependent pathway appears to predominate (Figure 2C).
The question of whether a certain acyl-CoA species, as generated in peroxisomes, is removed from the intraperoxisomal lumen via the carnitine-dependent route or the thioesterase-dependent route, may well have important physiological implications. This is especially true for acetyl-CoA. Indeed, if acetyl-CoA follows the carnitine-dependent pathway, acetyl-carnitine is produced in peroxisomes, followed by its transport into the cytosol. Cytosolic acetylcarnitine can only go one way, i.e. uptake into mitochondria, followed by retroconversion into acetyl-CoA via mitochondrial CRAT, and oxidation to CO2 and H2O. However, when the acetyl-CoA, as generated in peroxisomes, follows the thioesterase-dependent pathway, acetate is produced, which probably leaves the peroxisome as acetic acid to end up in the cytosol, where it may take several metabolic routes, including: (i) release from the cell, as described in Leighton et al. ; (ii) uptake into mitochondria, followed by activation to acetyl-CoA within mitochondria, followed by oxidation to CO2 and H2O; and (iii) activation to acetyl-CoA in the cytosol, after which the acetyl-CoA may be used for different purposes, including cholesterol biosynthesis. After carboxylation to malonyl-CoA by malonyl-CoA carboxylase, the acetyl-CoA units ultimately derived from fatty acid oxidation in peroxisomes may also be used for fatty acid synthesis and (ether)phospholipid biosynthesis.
Phytanic acid α-oxidation and the transport of metabolites across the peroxisomal membrane
One of the other aspects of the α-oxidation system, which has remained unresolved, at least in part, concerns the transport of substrates and products across the peroxisomal membrane. PMP34 (34 kDa peroxisomal membrane protein) is a member of the mitochondrial family of solute transporters, is exclusively peroxisomal and has been shown to be able to catalyse the transport of ATP , probably in exchange for AMP as shown for Ant1p, the yeast orthologue of PMP34 [26,27]. Furthermore, we have recently described the presence of a peroxisomal transport activity for 2-oxoglutarate, which is required in the hydroxylase reaction . Since peroxisomes do not contain the enzymatic machinery to convert succinate back into 2-oxoglutarate, a logical mechanism would involve the 1:1 exchange between 2-oxoglutarate and succinate (Figure 2B). With respect to the reoxidation of intraperoxisomal NADH, studies in yeast have clearly shown the involvement of a redox shuttle, mediated by the cytosolic and peroxisomal isoforms of malate dehydrogenase . Mammalian peroxisomes do not contain a specific peroxisomal malate dehydrogenase, however, which renders a malate dehydrogenase-based redox shuttle in mammalian peroxisomes not very likely. Baumgart et al.  have proposed the existence of a redox shuttle involving the cytosolic and peroxisomal forms of lactate dehydrogenase.
Phytanic acid ω-oxidation, an alternative mechanism to degrade phytanic acid
Early studies had shown that phytanic acid may also be degraded via another mechanism, i.e. ω-oxidation. Studies by Greter et al.  and Wierzbicki et al.  have shown that this pathway may also be operative under in vivo conditions. We have recently begun to try to resolve the mechanism involved and have first concentrated on the identification of the first step in the pathway, which is the hydroxylation of phytanic acid at the ω-end of the molecule. Studies in rat and human liver microsomes have shown that phytanic acid is readily ω-hydroxylated to ω-hydroxyphytanic acid and that one or more members of the CYP (cytochrome P450) enzyme superfamily catalyse this reaction [33,34]. More recent work in which use was made of specific inhibitors has shown that members of the CYP family 4 are responsible for phytanic acid ω-hydroxylation. Incubations with microsomes containing human recombinant CYPs (supersomes) revealed that multiple CYP enzymes of the family 4 class are able to ω-hydroxylate phytanic acid with the following order of efficiency: CYP4F3A>CYP4F3B>CYP4F2>CYP4A11 .
The ultimate idea is to find ways in which the capacity of the ω-oxidation system for phytanic acid can be induced. In this respect, it is important to mention that the expression of CYP4A11 is under the control of PPARα (peroxisome-proliferator-activated receptor α), which implies that the expression of the CYP4A11 may be induced by known PPARα ligands including fibrates such as phenofibrate and bezafibrate. Interestingly, in microsomes isolated from mice fed in a diet supplemented with 0.1% (w/w) WY14643, a known PPARα ligand, phytanic ω-hydroxylation was markedly (8-fold) induced .
Very recent work by Hsu et al.  has shown that the expression of CYP4F2 is controlled by the SREBP (sterol-regulatory-element-binding protein) pathway. This finding opens up another line of future research in which use will be made of statins, which are known activators of the SREBP pathway. The availability of a mouse model for ARD, which we have recently generated, allows these possibilities to be tested.
This work was financially supported by grants from the European Union (grant no. LSHG-CT-2004-512018, entitled ‘Peroxisomes in health and disease’, and grant no. QLG3-CT-2002-00696, entitled ‘Refsum disease’) and a grant from the Princes Beatrix Fonds (project number MAR04-116, entitled: ‘Adult Refsum disease and the search for pathological mechanisms, induced by phytanic acid, using a newly generated ARD mouse modeL'). We thank Mrs Maddy Festen for expert preparation of this paper and Mr Jos Ruiter for preparation of the Figures.
Cardiovascular Bioscience: A Focus Topic at Life Sciences 2007, held at SECC Glasgow, U.K., 9–12 July 2007. Edited by S. Kennedy (Strathclyde, Glasgow, U.K.), M. Lloyd (Bath, U.K.) and C. Wainwright (Robert Gordon University, Aberdeen, U.K.).
Abbreviations: ACSL, long-chain acyl-CoA synthase; ACSVL, very-long-chain acyl-CoA synthase; ARD, adult Refsum's disease; CRAT, carnitine acetyltransferase; CYP, cytochrome P450; ER, endoplasmic reticulum; FALDH, fatty aldehyde dehydrogenase; FATP, fatty acid transport protein; PPARα, peroxisome-proliferator-activated receptor α; PTS2, peroxisome-targeting signal 2; SREBP, sterol-regulatory-element-binding protein
- © The Authors Journal compilation © 2007 Biochemical Society