Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Peroxisomal Straight-chain Acyl-CoA Oxidase and D-bifunctional Protein Are Essential for the Retroconversion Step in Docosahexaenoic Acid Synthesis

2001; Elsevier BV; Volume: 276; Issue: 41 Linguagem: Inglês

10.1074/jbc.m106326200

ISSN

1083-351X

Autores

Hui‐Min Su, Ann B. Moser, Hugo W. Moser, Paul A. Watkins,

Tópico(s)

Lipid metabolism and biosynthesis

Resumo

Docosahexaenoic acid (DHA, C22:6n-3) is essential for normal brain and retinal development. The nature and subcellular location of the terminal steps in DHA biosynthesis have been controversial. Rather than direct Δ4-desaturation of C22:5n-3, it has been proposed that this intermediate is elongated to C24:5n-3, desaturated to C24:6n-3, and "retroconverted" to DHA via peroxisomal β-oxidation. However, this hypothesis has recently been challenged. The goal of this study was to determine the mechanism and specific enzymes required for the retroconversion step in human skin fibroblasts. Cells from patients with deficiencies of either acyl-CoA oxidase or D-bifunctional protein, the first two enzymes of the peroxisomal straight-chain fatty acid β-oxidation pathway, exhibited impaired (5–20% of control) conversion of either [1-14C]18:3n-3 or [1-14C]22:5n-3 to DHA as did cells from peroxisome biogenesis disorder patients comprising eight distinct genotypes. In contrast, normal DHA synthesis was observed in cells from patients with rhizomelic chondrodysplasia punctata, Refsum disease, X-linked adrenoleukodystrophy, and deficiency of mitochondrial medium- or very long-chain acyl-CoA dehydrogenase. Acyl-CoA oxidase-deficient cells accumulated 2–5 times more radiolabeled C24:6n-3 than did controls. Our data are consistent with the retroconversion hypothesis and demonstrate that peroxisomal β-oxidation enzymes acyl-CoA oxidase and D-bifunctional protein are essential for this process in human skin fibroblasts. Docosahexaenoic acid (DHA, C22:6n-3) is essential for normal brain and retinal development. The nature and subcellular location of the terminal steps in DHA biosynthesis have been controversial. Rather than direct Δ4-desaturation of C22:5n-3, it has been proposed that this intermediate is elongated to C24:5n-3, desaturated to C24:6n-3, and "retroconverted" to DHA via peroxisomal β-oxidation. However, this hypothesis has recently been challenged. The goal of this study was to determine the mechanism and specific enzymes required for the retroconversion step in human skin fibroblasts. Cells from patients with deficiencies of either acyl-CoA oxidase or D-bifunctional protein, the first two enzymes of the peroxisomal straight-chain fatty acid β-oxidation pathway, exhibited impaired (5–20% of control) conversion of either [1-14C]18:3n-3 or [1-14C]22:5n-3 to DHA as did cells from peroxisome biogenesis disorder patients comprising eight distinct genotypes. In contrast, normal DHA synthesis was observed in cells from patients with rhizomelic chondrodysplasia punctata, Refsum disease, X-linked adrenoleukodystrophy, and deficiency of mitochondrial medium- or very long-chain acyl-CoA dehydrogenase. Acyl-CoA oxidase-deficient cells accumulated 2–5 times more radiolabeled C24:6n-3 than did controls. Our data are consistent with the retroconversion hypothesis and demonstrate that peroxisomal β-oxidation enzymes acyl-CoA oxidase and D-bifunctional protein are essential for this process in human skin fibroblasts. docosahexaenoic acid (C22:6n-3) peroxisome biogenesis disorder peroxisomal straight-chain acyl-CoA oxidase peroxisomal D-bifunctional protein X-linked adrenoleukodystrophy sterol carrier protein X rhizomelic chondrodysplasia punctata very long-chain fatty acid medium- and very long-chain acyl-CoA dehydrogenase, respectively Docosahexaenoic acid (C22:6n-3, DHA1) plays an important role in normal neurological development, especially in brain and retina (1Innis S.M. Prog. Lipid Res. 1991; 30: 39-103Crossref PubMed Scopus (796) Google Scholar,2Uauy R. J. Pediatr. Gastroenterol. Nutr. 1990; 11: 296-302Crossref PubMed Scopus (32) Google Scholar), where it occurs in high concentration (3Fliesler S.J. Anderson R.E. Prog. Lipid Res. 1983; 22: 79-131Crossref PubMed Scopus (847) Google Scholar, 4Sastry P.S. Prog. Lipid Res. 1985; 24: 69-176Crossref PubMed Scopus (614) Google Scholar). A deficiency of brain DHA is associated with reduced learning ability in rats (5Lamptey M.S. Walker B.L. J. Nutr. 1976; 106: 86-93Crossref PubMed Google Scholar, 6Greiner R.S. Moriguchi T. Hutton A. Slotnick B.M. Salem Jr., N. Lipids. 1999; 34: S239-S243Crossref PubMed Google Scholar), impaired visual acuity in infant monkeys (7Neuringer M. Connor W.E. Lin D.S. Barstad L. Luck S. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 4021-4025Crossref PubMed Scopus (776) Google Scholar), and abnormal brain development in human infants with the Zellweger syndrome, a peroxisome biogenesis disorder (PBD) (8Moser H.W. Adv. Hum. Genet. 1993; 21: 1-106PubMed Google Scholar, 9Martinez M. J. Inherit. Metab. Dis. 1995; 18: 61-75Crossref PubMed Scopus (52) Google Scholar). DHA is derived from its parent precursor, the dietary essential fatty acid linolenic acid (C18:3n-3), via a series of alternating desaturation and elongation steps (see Fig. 1) (10Brenner R.R. Lipids. 1971; 6: 567-575Crossref PubMed Scopus (200) Google Scholar, 11Sprecher H. Bracco U. Deckelbaum R.J. Polyunsaturated Fatty Acids in Human Nutrition. Nestec Ltd., Vevey/Raven Press, Ltd., New York1992: 13-24Google Scholar). The primary products of C18:3n-3 are eicosapentaenoic acid (C20:5n-3), docosapentaenoic acid (C22:5n-3), and DHA. Until 1991, it had been postulated that conversion of C18:3n-3 to DHA was carried out entirely in microsomes. However, the last step in the biosynthetic pathway required conversion of C22:5n-3 to DHA via Δ4-desaturase, an enzyme that may not exist in higher animals (12Voss A. Reinhart M. Sankarappa S. Sprecher H. J. Biol. Chem. 1991; 266: 19995-20000Abstract Full Text PDF PubMed Google Scholar). A modified pathway was therefore proposed in which C22:5n-3 is elongated to tetracosapentaenoic acid (C24:5n-3), desaturated to tetracosahexaenoic acid (C24:6n-3) in microsomes, and then retroconverted to DHA in peroxisomes (see Fig. 1) (12Voss A. Reinhart M. Sankarappa S. Sprecher H. J. Biol. Chem. 1991; 266: 19995-20000Abstract Full Text PDF PubMed Google Scholar, 13Sprecher H. Luthria D.L. Mohammed B.S. Baykousheva S.P. J. Lipid Res. 1995; 36: 2471-2477Abstract Full Text PDF PubMed Google Scholar). Evidence for this new revised pathway was obtained when the intermediates C24:5n-3 and C24:6n-3 were detected in cultured cell studies (14Moore S.A. Hurt E. Yoder E. Sprecher H. Spector A.A. J. Lipid Res. 1995; 36: 2433-2443Abstract Full Text PDF PubMed Google Scholar, 15Delton-Vandenbrouke I. Grammas P. Anderson R.E. J. Lipid Res. 1997; 38: 147-159Abstract Full Text PDF PubMed Google Scholar) and tissue homogenates (16Yin F.Q. Chen Q. Sprecher H. Biochim. Biophys. Acta. 1999; 1438: 63-72Crossref PubMed Scopus (12) Google Scholar, 17Retterstol K. Haugen T.B. Christophersen B.O. Biochim. Biophys. Acta. 2000; 1483: 119-131Crossref PubMed Scopus (33) Google Scholar). That the retroconversion step for DHA synthesis, a chain shortening of C24:6n-3 to DHA, takes place only in peroxisomes has been called into question, and a mitochondrial contribution has also been suggested (18Infante J.P. Huszagh V.A. Mol. Genet. Metab. 2001; 72: 1-7Crossref PubMed Scopus (22) Google Scholar, 19Infante J.P. Huszagh V.A. Mol. Genet. Metab. 2001; 72: 185-198Crossref PubMed Scopus (49) Google Scholar, 20Infante J.P. Huszagh V.A. FEBS Lett. 2000; 468: 1-5Crossref PubMed Scopus (80) Google Scholar, 21Infante J.P. Huszagh V.A. FEBS Lett. 1998; 431: 1-6Crossref PubMed Scopus (56) Google Scholar, 22Infante J.P. Huszagh V.A. Mol. Cell. Biochem. 1997; 168: 101-115Crossref PubMed Scopus (51) Google Scholar). The identities of specific enzymes or mechanisms involved in the retroconversion of C24:6n-3 to DHA is not known. Both mitochondria and peroxisomes carry out fatty acid β-oxidation. This cyclic process is similar in both organelles with each cycle containing dehydrogenation/oxidation, hydration, dehydrogenation, and thiolytic cleavage steps. However, the functions of mitochondrial and peroxisomal β-oxidation pathways are significantly different (23Wanders R.J. Tager J.M. Mol. Asp. Med. 1998; 19: 69-154Crossref PubMed Google Scholar,24Wanders R.J.A. Barth P.C. Heymans H.S.A. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3219-3256Google Scholar). In general, one cycle of β-oxidation shortens an acyl chain by two carbons, releasing one molecule of acetyl-CoA, in a process mediated by a sequence of enzymes, each of which is specific for its substrate. In mitochondrial β-oxidation, the preferred substrates are fatty acids with a chain length of less than 20 carbons. These fatty acids enter the organelle by the carnitine transport system and are usually degraded completely to acetyl-CoA via several β-oxidation cycles. Peroxisomal fatty acid β-oxidation is capable of oxidizing longer chain length substrates, the very long-chain fatty acids (VLCFA), e.g. hexacosanoic acid (C26:0) and tetracosanoic acid (C24:0). Entry of these substrates does not require carnitine but may involve ATP-binding cassette transporters such as the ALD protein. Peroxisomal β-oxidation does not proceed to completion but rather only through a few cycles in which the acyl chain is shortened. In this study, the mechanism and enzymes involved in the retroconversion of C24:6n-3 to DHA, a two-carbon shortening process, were investigated. We compared the rate of radiolabeled DHA synthesis from [1-14C]18:3n-3, the parent precursor, and [1-14C]22:5n-3, a more direct precursor, in human skin fibroblasts from normal controls with DHA synthesis rates in cells from patients with disorders of peroxisomal or mitochondrial fatty acid β-oxidation. Our results confirm the existence of the new revised DHA synthetic pathway in which C22:5n-3 is elongated to C24:5n-3, desaturated to C24:6n-3 in microsomes, and then retroconverted to C22:6n-3 in peroxisomes. Furthermore, we demonstrate that the peroxisomal β-oxidation enzymes straight-chain acyl-CoA oxidase (AOx) and D-bifunctional protein (DBP) are essential for the retroconversion process. One or both of the known peroxisomal thiolases, 3-oxoacyl-CoA thiolase and sterol carrier protein X (SCPx) thiolase, may participate in retroconversion, but the former enzyme does not appear to be essential for this process. Thus, we conclude that the retroconversion step in DHA synthesis from C24:6n-3 proceeds via the peroxisomal straight-chain fatty acid β-oxidation pathway. We find no evidence for involvement of mitochondrial enzymes in this process. Radiolabeled [1-14C]18:3n-3 (52 mCi/mmol) and [1-14C]22:5n-3 (55 mCi/mmol) were purchased from PerkinElmer Life Sciences and American Radiolabeled Chemicals (St. Louis, MO), respectively. Standards of fatty acids and fatty acids methyl esters were purchased from Matreya (Pleasant Gap, PA). The free fatty acid of C24:5n-3 and C24:6n-3 were generous gifts from A. Spector and H. Sprecher, respectively. Human skin fibroblasts from normal controls and patients with peroxisomal or mitochondrial fatty acid oxidation disorders were obtained through the Kennedy Krieger Institute's Mental Retardation Research Center. The peroxisomal disorders included Zellweger syndrome patients with different genotypes (25Moser H.W. Mol. Genet. Metab. 1999; 68: 316-327Crossref PubMed Scopus (81) Google Scholar), AOx deficiency (26ten Brink H.J. Stellaard F. van den Heuvel C.M. Kok R.M. Schor D.S. Wanders R.J. Jakobs C. J. Lipid Res. 1992; 33: 41-47Abstract Full Text PDF PubMed Google Scholar), DBP deficiency (27Watkins P.A. Chen W.W. Harris C.J. Hoefler G. Hoefler S. Blake Jr., D.C. Balfe A. Kelley R.I. Moser A.B. Beard M.E. Moser H.W. J. Clin. Invest. 1989; 83: 771-777Crossref PubMed Scopus (172) Google Scholar, 28van Grunsven E.G. van Berkel E. Mooijer P.A. Watkins P.A. Moser H.W. Suzuki Y. Jiang L.L. Hashimoto T. Hoefler G. Adamski J. Wanders R.J. Am. J. Hum. Genet. 1999; 64: 99-107Abstract Full Text Full Text PDF PubMed Scopus (94) Google Scholar), rhizomelic chondrodysplasia punctata (RCDP) (29Purdue P.E. Skoneczny M. Yang X. Zhang J.W. Lazarow P.B. Neurochem. Res. 1999; 24: 581-586Crossref PubMed Scopus (37) Google Scholar), Refsum disease (30Wanders R.J.A. Jakobs C. Skjeldal O.H. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3303-3321Google Scholar), and X-linked adrenoleukodystrophy (X-ALD) (31Moser H.W. Smith K.D. Watkins P.A. Powers J. Moser A.B. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3257-3302Google Scholar). The disorders of mitochondrial fatty acid β-oxidation included deficiencies of very long-chain acyl-CoA dehydrogenase (VLCAD) and medium-chain acyl-CoA dehydrogenase (MCAD) (32Kelley R.I. Biochem. Biophys. Res. Commun. 1992; 182: 1002-1007Crossref PubMed Scopus (11) Google Scholar, 33Roe C.R. Ding J. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 2297-2326Google Scholar). Cells were grown in Eagle's minimum essential medium supplemented with 10% fetal bovine serum,l-glutamine, and penicillin/streptomycin. The radiolabeled substrates [1-14C]18:3n-3 and [1-14C]22:5n-3 were solubilized by preparing a complex of their sodium salts with serum albumin at a molar ratio of 3:1 (34Su H.M. Brenna J.T. Anal. Biochem. 1998; 261: 43-50Crossref PubMed Scopus (33) Google Scholar). When fibroblast cultures reached 90% confluence, they were incubated in 2 ml of Dulbecco's modified Eagle's medium (high glucose) containing 10% fetal bovine serum, penicillin/streptomycin, and 0.05 μCi of [1-14C]18:3n-3 or 0.05 μCi of [1-14C]22:5n-3 at 37 °C in a 5% CO2 incubator. Unless otherwise specified in the figure legends, incubations were for 72 h. After this labeling period, the medium was removed, the cells were rinsed with Hanks' balanced salt solution, and then cells were harvested using 0.25% trypsin. The suspended cells were centrifuged, and the pellets were washed twice with Hanks' balanced salt solution and stored at −20 °C after flushing the tube with nitrogen. Cell pellets were resuspended in 0.5 ml of deionized water and disrupted using a cup sonicator (550 Sonic Dismembrator, Fisher Scientific) operated at 35% maximum power for 2 min. After removal of an aliquot for protein determination, lipids were extracted from the cell sonicate. Internal standards (C18:3n-3, C20:5n-3, C22:5n-3, and DHA) were added, and total fatty acyls were converted to their methyl esters as described previously (35Moser A.B. Jones D.S. Raymond G.V. Moser H.W. Neurochem. Res. 1999; 24: 187-197Crossref PubMed Scopus (80) Google Scholar). Radiolabeled fatty acid methyl esters were separated by reverse phase high performance liquid chromatography by a modification of the method of Moore et al. (14Moore S.A. Hurt E. Yoder E. Sprecher H. Spector A.A. J. Lipid Res. 1995; 36: 2433-2443Abstract Full Text PDF PubMed Google Scholar). Briefly, samples were applied to a 4.6 × 150-mm Luna 3-μm C18 column (Phenomenex, Torrance, CA) and eluted with a mobile phase of water and acetonitrile at 0.75 ml/min. The elution program consisted of 76% acetonitrile for 55 min followed by a linear increase to 100% acetonitrile over 10 min and maintenance at this concentration for an additional 15 min. This program efficiently separated methyl esters of C18:3n-3, C20:5n-3, C22:5n-3, C24:5n-3, C24:6n-3, and DHA. Labeled products were collected using a fraction collector. The mobile phase was evaporated under a stream of nitrogen, and radioactivity was determined by liquid scintillation counting. Labeled products were identified by comparing their retention times with fatty acid methyl ester standards. The purity of the major n-3 fatty acid methyl ester peaks was checked by gas chromatography (35Moser A.B. Jones D.S. Raymond G.V. Moser H.W. Neurochem. Res. 1999; 24: 187-197Crossref PubMed Scopus (80) Google Scholar). Collected fractions identified as C18:3n-3, C20:5n-3, and DHA methyl esters were 99% pure. Collected fractions identified as C22:5n-3, C24:5n-3 and C22:5n-3 methyl esters contained unlabeled n-6 and n-9 fatty acids. The C22:5n-3 fraction contained linoleic acid (C18:2n-6), the C24:5n-3 fraction contained C22:4n-6 and 17:1n-9, and the C24:6n-3 fraction contained C22:5n-6. These associated fatty acids do not contribute to the radioactivity of the fraction. Protein concentration was determined by the method of Lowry et al.(36Lowry O.H. Rosebrough N.J. Farr A.L. Randall R.J. J. Biol. Chem. 1951; 193: 265-275Abstract Full Text PDF PubMed Google Scholar) with bovine serum albumin as standard. Data are presented as mean ± S.D. of the specific radioactivity recovered. Until recently, the terminal steps in the accepted pathway for synthesis of DHA from dietary essential fatty acids required a Δ4-desaturase. Because no mammalian enzyme with this activity had been identified, Voss et al. (12Voss A. Reinhart M. Sankarappa S. Sprecher H. J. Biol. Chem. 1991; 266: 19995-20000Abstract Full Text PDF PubMed Google Scholar) proposed the retroconversion pathway shown in Fig. 1. Furthermore, Moore et al. (14Moore S.A. Hurt E. Yoder E. Sprecher H. Spector A.A. J. Lipid Res. 1995; 36: 2433-2443Abstract Full Text PDF PubMed Google Scholar) showed that retroconversion likely occurred in peroxisomes. Infante and Huszagh (18Infante J.P. Huszagh V.A. Mol. Genet. Metab. 2001; 72: 1-7Crossref PubMed Scopus (22) Google Scholar, 22Infante J.P. Huszagh V.A. Mol. Cell. Biochem. 1997; 168: 101-115Crossref PubMed Scopus (51) Google Scholar) have recently challenged the involvement of peroxisomes in DHA synthesis, primarily on theoretical grounds. To address these issues, we examined the synthesis of DHA in cultured human skin fibroblasts using two precursors, [1-14C]18:3n-3 and [1-14C]22:5n-3. Cells from normal controls incubated with these substrates for 72 h incorporated label into DHA at a rate of 75 ± 10 and 31 ± 6.6 dpm/mg of protein/h, respectively. To confirm the original observation of Moore et al. (14Moore S.A. Hurt E. Yoder E. Sprecher H. Spector A.A. J. Lipid Res. 1995; 36: 2433-2443Abstract Full Text PDF PubMed Google Scholar) that DHA synthesis required functional peroxisomes, we studied cells from patients with Zellweger syndrome, a PBD in which multiple peroxisomal functions are defective. PBDs are caused by mutations in one of 23 known PEX genes (25Moser H.W. Mol. Genet. Metab. 1999; 68: 316-327Crossref PubMed Scopus (81) Google Scholar, 37Brown T.W. Titorenko V.I. Rachubinski R.A. Mol. Biol. Cell. 2000; 11: 141-152Crossref PubMed Scopus (37) Google Scholar), and Moore et al. (14Moore S.A. Hurt E. Yoder E. Sprecher H. Spector A.A. J. Lipid Res. 1995; 36: 2433-2443Abstract Full Text PDF PubMed Google Scholar) studied two patients with the same genotype. The clinical presentation of PBD patients ranges from severe (Zellweger syndrome) to moderate (neonatal adrenoleukodystrophy) to mild (infantile Refsum disease), and there is no genotype-phenotype correlation (38Gould S.J. Raymond G.V. Valle D. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3181-3217Google Scholar). In the experiment shown in Fig.2, cells from Zellweger syndrome patients with mutations in PEX1, -2, -3, -5, -6, -10, -12, and -16 were studied. The rate of radiolabeled DHA synthesis in fibroblasts from all Zellweger syndrome patients, with either [1-14C]18:3n-3 (Fig. 2, top panel) or [1-14C]22:5n-3 (Fig. 2, bottom panel) as substrate, was less than 5% of that in control cells. Thus, the defect is not limited to a single PBD genotype. These data therefore validate and extend the original observation that peroxisomes are essential for DHA synthesis. Moreover, radiolabeled C22:5n-3, C24:5n-3, and C24:6n-3 derived from both substrates were detected in Zellweger syndrome patient cell lines (data not shown). These intermediates precede peroxisomal retroconversion of C24:6n-3 to DHA in the revised pathway. Taken together, these results suggest that peroxisomes are essential for DHA synthesis. In addition to defects in peroxisomal β-oxidation, patients with PBDs such as the Zellweger syndrome have multiple biochemical abnormalities, including a low rate of plasmalogen biosynthesis (39van den Bosch H. Schrakamp G. Hardeman D. Zomer A.W. Wanders R.J. Schutgens R.B. Biochimie (Paris). 1993; 75: 183-189Crossref PubMed Scopus (39) Google Scholar). Infante and Huszagh (18Infante J.P. Huszagh V.A. Mol. Genet. Metab. 2001; 72: 1-7Crossref PubMed Scopus (22) Google Scholar) have argued that the defect in plasmalogen biosynthesis in PBD patients contributes to their low rate of DHA synthesis. Therefore, we examined DHA synthesis from [1-14C]18:3n-3 and [1-14C]22:5n-3 in fibroblasts from patients with deficiency of either AOx or DBP. AOx catalyzes the first step in peroxisomal β-oxidation of straight-chain fatty acids, and DBP catalyzes the second (enoyl-CoA hydratase) and third (d-hydroxyacyl-CoA dehydrogenase) reactions of this pathway (Fig. 1) (24Wanders R.J.A. Barth P.C. Heymans H.S.A. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3219-3256Google Scholar). Fibroblasts from patients with AOx deficiency synthesized DHA from either [1-14C]18:3n-3 (Fig. 3, top panel) or [1-14C]22:5n-3 (Fig. 3, bottom panel) at a rate less than 10% of that of control cell lines. Similarly DHA synthesis in cells from patients with DBP deficiency was less than 5% of control with [1-14C]18:3n-3 as substrate (Fig. 3, top panel) and was 20% of control with [1-14C]22:5n-3 as substrate (Fig. 3,bottom panel). Patients with AOx or DBP deficiency have defective peroxisomal fatty acid β-oxidation of VLCFA but normal plasmalogen synthesis (24Wanders R.J.A. Barth P.C. Heymans H.S.A. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3219-3256Google Scholar). Therefore, these data indicate that both AOx and DBP are involved in DHA synthesis. We also investigated DHA synthesis in fibroblasts from patients with RCDP. Cells from RCDP patients fail to import the final peroxisomal β-oxidation enzyme 3-oxoacyl-CoA thiolase into the organelle and have defective plasmalogen biosynthesis (29Purdue P.E. Skoneczny M. Yang X. Zhang J.W. Lazarow P.B. Neurochem. Res. 1999; 24: 581-586Crossref PubMed Scopus (37) Google Scholar). In contrast to AOx- and DBP-deficient fibroblasts, cells from RCDP patients synthesized DHA at nearly normal rates from either [1-14C]18:3n-3 (Fig. 3, top panel) or [1-14C]22:5n-3 (Fig. 3, bottom panel). Peroxisomes contain a second thiolase derived from SCPx in addition to 3-oxoacyl-CoA thiolase (40Seedorf U. Brysch P. Engel T. Schrage K. Assmann G. J. Biol. Chem. 1994; 269: 21277-21283Abstract Full Text PDF PubMed Google Scholar). Due to the presence of SCPx thiolase, RCDP fibroblasts are able to oxidize VLCFA at normal rates (41Hoefler G. Hoefler S. Watkins P.A. Chen W.W. Moser A. Baldwin V. McGillivary B. Charrow J. Friedman J.M. Rutledge L. Hashimoto T. Moser H.W. J. Pediatr. 1988; 112: 726-733Abstract Full Text PDF PubMed Scopus (108) Google Scholar). These data indicate that peroxisomal 3-oxoacyl-CoA thiolase is not absolutely required for DHA synthesis. Furthermore, the results indicate that SCPx thiolase can catalyze this reaction but do not reveal whether 3-oxoacyl-CoA thiolase can do so. AOx is thought to be the rate-limiting enzyme in peroxisomal straight-chain fatty acid β-oxidation and is the first enzyme unique to the pathway (42Aoyama T. Souri M. Kamijo T. Ushikubo S. Hashimoto T. Biochem. Biophys. Res. Commun. 1994; 201: 1541-1547Crossref PubMed Scopus (27) Google Scholar). Therefore, we examined the accumulation of labeled DHA synthesis intermediates in control and AOx fibroblasts incubated for 8–120 h with either [1-14C]18:3n-3 or [1-14C]22:5n-3 substrates. As shown in Fig.4, only a small amount of the [1-14C]18:3n-3 substrate accumulated in either normal or AOx-deficient cell lines. In control cell lines, the amount of radiolabeled DHA increased with increasing incubation time up to 4 days and remained constant to 5 days. In contrast, DHA synthesis was significantly reduced in AOx-deficient cell lines. In both cell lines, the main radiolabeled intermediates, C20:5n-3, C22:5n-3, C24:5n-3, and C24:6n-3, accumulated during the incubation period, suggesting that AOx is not required for any aspect of the n-3 fatty acid synthetic pathway from C18:3n-3 through to C24:6n-3. During the 5-day incubation period, the two most direct precursors of DHA, C24:5n-3 and C24:6n-3, accumulated to a higher level in AOx-deficient cell lines than in control fibroblasts (Fig. 4), supporting the conclusion that peroxisomal straight-chain fatty acid β-oxidation is essential for DHA synthesis. Similar results were obtained using [1-14C]22:5n-3 as substrate (Fig.5). Cellular accumulation of the substrate in normal cells increased slightly from 8 h to 1 day and then slowly decreased up to 5 days. In AOx-deficient cells, levels of the [1-14C]22:5n-3 substrate were 2–3 times higher than in control fibroblasts, consistent with a downstream metabolic impairment. Labeled DHA accumulated in control cells with increasing time of incubation but was never greater than 5% of control in AOx-deficient cells. In contrast, the level of labeled C24:6n-3, the most direct precursor of DHA prior to retroconversion, increased to more than 5-fold over control levels in AOx-deficient cells with increasing time of incubation. Accumulation of C24:5n-3, the intermediate preceding C24:6n-3 in the DHA biosynthetic pathway, was 2–3 times higher in AOx-deficient cell lines as compared with controls. A similar increase in C24:6n-3 and C24:5n-3 levels was observed in fibroblasts from patients with Zellweger syndrome and DBP deficiency (data not shown). These observations are consistent with a block in the DHA biosynthetic pathway between C24:6n-3 and DHA. We next determined whether DHA synthesis was defective in other human disorders of peroxisomal fatty acid catabolism. Peroxisomal β-oxidation of saturated, straight-chain VLCFA is defective in X-ALD due to mutations in the ABCD1 gene (43Smith K.D. Kemp S. Braiterman L.T. Lu J.F. Wei H.M. Geraghty M. Stetten G. Bergin J.S. Pevsner J. Watkins P.A. Neurochem. Res. 1999; 24: 521-535Crossref PubMed Scopus (157) Google Scholar). This gene encodes a peroxisomal membrane protein belonging to the ATP-binding cassette family of transporters, the precise function of which remains unknown (43Smith K.D. Kemp S. Braiterman L.T. Lu J.F. Wei H.M. Geraghty M. Stetten G. Bergin J.S. Pevsner J. Watkins P.A. Neurochem. Res. 1999; 24: 521-535Crossref PubMed Scopus (157) Google Scholar). Synthesis of DHA from either [1-14C]18:3n-3 or [1-14C]22:5n-3 was not impaired in fibroblasts from X-ALD patients (Fig. 6). This result indicates that, unlike saturated VLCFA, peroxisomal β-oxidation of C24:6n-3 to DHA does not require the participation of the product of the ABCD1 gene. Patients with Refsum disease have a defect in the peroxisomal α-oxidation of phytanic acid, a branched-chain fatty acid (30Wanders R.J.A. Jakobs C. Skjeldal O.H. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3303-3321Google Scholar). Like the situation with X-ALD, fibroblasts from Refsum disease patients synthesized DHA normally from either radiolabeled precursor (Fig. 6). Results presented above imply a role for peroxisomal β-oxidation in the retroconversion of C24:6n-3 to DHA. To verify that defects in mitochondrial β-oxidation do not produce a similar impairment in DHA synthesis, we incubated fibroblasts from patients with deficiency of either MCAD or VLCAD with [1-14C]18:3n-3 or [1-14C]22:5n-3. As shown in Fig.7, neither disorder exhibited defective DHA synthesis from either labeled substrate. Furthermore, there was no significant accumulation of the C24:6n-3 intermediate in cells from MCAD or VLCAD patients. These findings support the hypothesis that peroxisomes, not mitochondria, are the subcellular site of retroconversion in DHA biosynthesis. To our knowledge, this is the first study demonstrating that the peroxisomal β-oxidation enzymes AOx and DBP are essential for the retroconversion step in DHA synthesis that shortens C24:6n-3 by two carbons. The involvement of peroxisomal 3-oxoacyl-CoA thiolase or peroxisomal SCPx thiolase in this process is also likely. Our data confirmed the existence of the new revised DHA synthetic pathway originally proposed by Voss et al. (12Voss A. Reinhart M. Sankarappa S. Sprecher H. J. Biol. Chem. 1991; 266: 19995-20000Abstract Full Text PDF PubMed Google Scholar) in which C22:5n-3 is elongated to C24:5n-3 and desaturated to C24:6n-3 before retroconversion. We conclude that in human skin fibroblasts, retroconversion of C24:6n-3 to DHA proceeds via the peroxisomal straight-chain fatty acid β-oxidation pathway responsible for the catabolism of saturated VLCFA. The use of two radiolabeled substrates, [1-14C]18:3n-3 and [1-14C]22:5n-3, allowed us to examine DHA synthesis from both its natural parent precursor (C18:3n-3) and from an intermediate in the pathway (C22:5n-3) just proximal to the retroconversion step. The latter is the most direct radiolabeled precursor of DHA that is commercially available. The radiolabeled intermediates C24:5n-3 and C24:6n-3, derived from either [1-14C]18:3n-3 or [1-14C]22:5n-3, were detected in normal human skin fibroblasts and in cells from patients with peroxisomal disorders or mitochondrial disorders used in this present study. Over time, higher accumulations of those two metabolites were found in Zellweger syndrome patient cell lines and in patients with either AOx deficiency or DBP deficiency but not in those with mitochondrial β-oxidation defects. Analysis of labeled fatty acids accumulating in cells incubated from 8 to 120 h with labeled substrates allowed us to verify that the proposed intermediates in the pathway were indeed synthesized. Our observations are consistent with the hypothesis that C18:3n-3 is converted to DHA via a series of alternating desaturation and elongation steps up to C24:6n-3, reactions that occur in microsomes (44Sprecher H. Chen Q. Prostaglandins Leukot. Essent. Fatty Acids. 1999; 60: 317-321Abstract Full Text PDF PubMed Scopus (54) Google Scholar). Our studies indicate that mitochondrial fatty acid β-oxidation is not essential for DHA synthesis. The first step unique to mitochondrial fatty acid β-oxidation is catalyzed by the acyl-CoA dehydrogenases, a group of enzymes with distinct but overlapping chain length specificities for acyl-CoAs (33Roe C.R. Ding J. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 2297-2326Google Scholar). Four acyl-CoA dehydrogenases have been described and designated as very long-chain (VLCAD), long-chain (LCAD), medium-chain (MCAD), and short-chain (SCAD) enzymes. Human patients with VLCAD, MCAD, and short-chain acyl-CoA dehydrogenase deficiency are known, but true long-chain acyl-CoA dehydrogenase deficiency has not been documented (33Roe C.R. Ding J. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 2297-2326Google Scholar). Our observation that DHA biosynthesis from either [1-14C]18:3n-3 or [1-14C]22:5n-3 in cells from patients with VLCAD and MCAD deficiency is consistent with the hypothesis that retroconversion of C24:6n-3 to DHA does not occur in mitochondria. In contrast, peroxisomes are essential for DHA biosynthesis. The rate of DHA synthesis from either [1-14C]18:3n-3 or [1-14C]22:5n-3 in fibroblasts from Zellweger syndrome patients was less than 5% of control. Peroxisome assembly is defective in these patients, resulting in multiple biochemical abnormalities (8Moser H.W. Adv. Hum. Genet. 1993; 21: 1-106PubMed Google Scholar, 45Goldfischer S. Moore C.L. Johnson A.B. Spiro A.J. Valsamis M.P. Wisniewski H.K. Ritch R.H. Norton W.T. Rapin I. Gartner L.M. Science. 1973; 182: 62-64Crossref PubMed Scopus (633) Google Scholar). At present, at least 23 PEX genes encoding peroxisomal assembly factors (peroxins) have been identified (37Brown T.W. Titorenko V.I. Rachubinski R.A. Mol. Biol. Cell. 2000; 11: 141-152Crossref PubMed Scopus (37) Google Scholar, 46Distel B. Erdmann R. Gould S.J. Blobel G. Crane D.I. Cregg J.M. Dodt G. Fujiki Y. Goodman J.M. Just W.W. Kiel J.A. Kunau W.H. Lazarow P.B. Mannaerts G.P. Moser H.W. Osumi T. Rachubinski R.A. Roscher A. Subramani S. Tabak H.F. Tsukamoto T. Valle D. van der Klei I. van Veldhoven P.P. Veenhuis M. J. Cell Biol. 1996; 135: 1-3Crossref PubMed Scopus (318) Google Scholar, 47Subramani S. Physiol. Rev. 1998; 78: 171-188Crossref PubMed Scopus (286) Google Scholar). Deficiencies of 10 PEX genes are known to cause PBDs in humans (25Moser H.W. Mol. Genet. Metab. 1999; 68: 316-327Crossref PubMed Scopus (81) Google Scholar). In the present study, the genotypes of the Zellweger syndrome patients included defects of PEX1, -2, -3, -5, -6, -10, -12, and -16. Defective peroxisomal fatty acid β-oxidation of saturated VLCFA is observed in all of these genotypes (38Gould S.J. Raymond G.V. Valle D. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3181-3217Google Scholar). Because the rate of DHA synthesis was <5% of control for nearly all genotypes tested, results from all Zellweger syndrome fibroblasts are reported as a group. These data are consistent with the hypothesis that defective peroxisomal β-oxidation is responsible for DHA deficiency and C24:6n-3 accumulation in PBD patients. Fibroblasts from PBD patients with the above genotypes also have defective biosynthesis of plasmalogens (38Gould S.J. Raymond G.V. Valle D. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3181-3217Google Scholar). Infante and Huszagh (18Infante J.P. Huszagh V.A. Mol. Genet. Metab. 2001; 72: 1-7Crossref PubMed Scopus (22) Google Scholar) proposed that low membrane plasmalogen levels contributed to the low rate of DHA synthesis in PBD patients. Our findings that DHA synthesis was normal in fibroblasts from RCDP patients dispute this hypothesis. Classical RCDP results from mutations in PEX7, the gene encoding the cytoplasmic receptor for proteins targeted to peroxisomes via peroxisome-targeting signal 2 (PTS2) (38Gould S.J. Raymond G.V. Valle D. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3181-3217Google Scholar). PTS2-containing proteins include the β-oxidation enzyme 3-oxoacyl-CoA thiolase and a key enzyme in plasmalogen synthesis, alkyl-dihydroxyacetonephosphate synthase (48de Vet E.C. Ijlst L. Oostheim W. Wanders R.J. van den Bosch H. J. Biol. Chem. 1998; 273: 10296-10301Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar). Although plasmalogen synthesis is profoundly deficient in these cells, saturated VLCFA β-oxidation proceeds normally due to the presence of a second thiolase, SCPx thiolase, which is targeted to peroxisomes by a different signal, PTS1 (23Wanders R.J. Tager J.M. Mol. Asp. Med. 1998; 19: 69-154Crossref PubMed Google Scholar). Data reported in Fig. 3were obtained from PEX7-deficient RCDP type 1 patient fibroblasts. We also found that DHA synthesis from [1-14C]18:3n-3 was normal in fibroblasts from two patients with RCDP type 2 (dihydroxyacetonephosphate acyltransferase deficiency) and one patient with RCDP type 3 (alkyl-dihydroxyacetonephosphate synthase deficiency) (data not shown). RCDP type 2 and 3 patients have defective plasmalogen synthesis, normal VLCFA β-oxidation, and normal targeting of both 3-oxoacyl-CoA thiolase and SCPx thiolase to peroxisomes (24Wanders R.J.A. Barth P.C. Heymans H.S.A. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3219-3256Google Scholar). Therefore, at least in fibroblasts, defective plasmalogen synthesis without a concomitant β-oxidation deficiency does not adversely affect DHA synthesis. While our studies indicate that SCPx thiolase can participate in peroxisomal retroconversion of C24:6n-3 to DHA, they do not rule out the possibility that 3-oxoacyl-CoA thiolase is also capable of catalyzing this last step in DHA synthesis. Human peroxisomes contain two fatty acid β-oxidation pathways, one for straight-chain fatty acids and the other for methyl-branched-chain fatty acids (49Hashimoto T. Neurochem. Res. 1999; 24: 551-563Crossref PubMed Scopus (61) Google Scholar, 50Wanders R.J. Neurochem. Res. 1999; 24: 565-580Crossref PubMed Scopus (72) Google Scholar, 51Wanders R.J. van Grunsven E.G. Jansen G.A. Biochem. Soc. Trans. 2000; 28: 141-149Crossref PubMed Scopus (45) Google Scholar). Until recently, it had been believed that human peroxisomal β-oxidation of saturated unbranched fatty acids such as the VLCFA C26:0 required AOx, L-bifunctional protein, and 3-oxoacyl-CoA thiolase. Similarly it was thought that human β-oxidation of branched-chain fatty acids such as pristanic acid involved branched-chain acyl-CoA oxidase, DBP, and SCPx thiolase. In 2001, Wanders et al. (24Wanders R.J.A. Barth P.C. Heymans H.S.A. Scriver C.R. Beaudet A.L. Sly W.S. Valle D. The Metabolic & Molecular Bases of Inherited Disease. 8th Ed. McGraw-Hill, New York2001: 3219-3256Google Scholar) proposed that the human pathway for β-oxidation of saturated VLCFA be revised to include AOx, DBP (and not L-bifunctional protein), and either 3-oxoacyl-CoA thiolaseor SCPx thiolase; the branched-chain pathway remains unchanged. Our observation that AOx or DBP, both of which are essential for β-oxidation of saturated VLCFA, are also required for the retroconversion step of DHA biosynthesis is consistent with Wanders' proposed pathway. Our observations that DHA synthesis is normal in X-ALD fibroblasts highlight one important distinction between peroxisomal β-oxidation of VLCFA and retroconversion of C24:6n-3 to DHA. The ALD protein, product of the ABCD1 gene defective in X-ALD, is apparently not required for DHA biosynthesis. Although the ALD protein belongs to the large family of ATP-binding cassette transmembrane transporters, its precise function is not yet known. Normal DHA synthesis in Refsum disease patient fibroblasts is consistent with the notion that peroxisomal α-oxidation does not participate in DHA synthesis. In conclusion, our results clearly provide evidence that in human skin fibroblasts the retroconversion step of DHA synthesis, a two-carbon shortening system from C24:6n-3, requires peroxisomal enzymes AOx and DBP as well as either 3-oxoacyl-CoA thiolase or SCPx thiolase. We further conclude that the retroconversion step in DHA synthesis proceeds via peroxisomal straight-chain fatty acid β-oxidation but not via peroxisomal branched-chain fatty acid β-oxidation or mitochondrial fatty acid β-oxidation. Our data support all facets of the new revised DHA synthetic pathway.

Referência(s)