Metabolic pathway that produces essential fatty acids from polymethylene-interrupted polyunsaturated fatty acids in animal cells Tamotsu Tanaka 1 , Jun-ichi Morishige 1 , Dai Iwawaki 1 , Terumi Fukuhara 1 , Naomi Hamamura 1 , Kaoru Hirano 1 , Takashi Osumi 2 and Kiyoshi Satouchi 1 1 Department of Applied Biological Science, Fukuyama University, Hiroshima, Japan 2 Graduate School of Life Sciences, University of Hyogo, Japan Linoleic acid (18:2 D-9,12) and a-linolenic acid (18:3 D- 9,12,15) are fatty acids that are essential to animals for maintenance of growth, reproduction and development of brain function. Because they lack the ability to introduce double bonds at both D-12 and D-15 positions of a fatty acid, animals have to acquire these poly- unsaturated fatty acids (PUFAs) from their diet. The alignment of the double bonds of these essential fatty acids as well as important metabolites, such as arachidonic acid (20:4 D-5,8,11,14) and eicosapentae- noic acid (EPA, 20:5 D-5,8,11,14,17), is interrupted by one methylene group. On the other hand, gymnosperm plants have PUFAs with the alignment of their double bonds interrupted by two or more methylenes [1–4]. PUFAs with this characteristic alignment of double bonds are categorized as polymethylene-interrupted- PUFAs (PMI-PUFAs). Pinolenic acid (18:3, D-5,9,12), sciadonic acid (20:3 D-5,11,14) and juniperonic acid (20:4 D-5,11,14,17) are typical PMI-PUFAs found in conifer plants such as Pinaceae, Taxodiaceae and Keywords essential fatty acids; peroxisomal b-oxidation; polymethylene-interrupted polyunsaturated fatty acids; polyunsaturated fatty acid remodeling Correspondence T. Tanaka, Department of Applied Biological Science, Fukuyama University, Higashimura, Fukuyama, Hiroshima, 729-0292, Japan Fax: +81 84 9362459 Tel: +81 84 9362111 E-mail: tamot@fubac.fukuyama-u.ac.jp (Received 22 January 2007, revised 9 March 2007, accepted 23 March 2007) doi:10.1111/j.1742-4658.2007.05807.x Sciadonic acid (20:3 D -5,11,14) and juniperonic acid (20:4 D-5,11,14,17) are polyunsaturated fatty acids (PUFAs) that lack the D-8 double bond of arachidonic acid (20:4 D-5,8,11,14) and eicosapentaenoic acid (20:5 D-5,8,11,14,17), respectively. Here, we demonstrate that these conifer oil- derived PUFAs are metabolized to essential fatty acids in animal cells. When Swiss 3T3 cells were cultured with sciadonic acid, linoleic acid (18:2 D-9,12) accumulated in the cells to an extent dependent on the con- centration of sciadonic acid. At the same time, a small amount of 16:2 D-7,10 appeared in the cellular lipids. Both 16:2 D-7,10 and linoleic acid accumulated in sciadonic acid-supplemented CHO cells, but not in peroxisome-deficient CHO cells. We confirmed that 16:2 D-7,10 was effect- ively elongated to linoleic acid in rat liver microsomes. These results indi- cate that sciadonic acid was partially degraded to 16:2 D-7,10 by two cycles of b-oxidation in peroxisomes, then elongated to linoleic acid in micro- somes. Supplementation of Swiss 3T3 cells with juniperonic acid, an n)3 analogue of sciadonic acid, induced accumulation of a-linolenic acid (18:3 D-9,12,15) in cellular lipids, suggesting that juniperonic acid was meta- bolized in a similar manner to sciadonic acid. This PUFA remodeling is thought to be a process that converts unsuitable fatty acids into essential fatty acids required by animals. Abbreviations AgTLC, argentation thin-layer chromatography; EPA, eicosapentaenoic acid; PtdCho, phosphatidylcholine; PtdEtn, phosphatidylethanolamine; PtdIns, phosphatidylinositol; PtdSer, phosphatidylserine; PMI-PUFA, polymethylene-interrupted polyunsaturated fatty acid; PUFA, polyunsaturated fatty acid. 2728 FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS Cupressaceae [1–4]. Some of these plant seeds are used as food, condiments or traditional Chinese medicines [5], and their oils are considered to be an alternative source of edible oil [6]. PMI-PUFAs have been shown to be constituents of membrane phospholipids of animal cells [6–11]. We have demonstrated that sciadonic acid and juniperonic acid are metabolized in a similar manner to arachido- nic acid and EPA, respectively, in the process of acyla- tion to phospholipids in HepG2 cells [12,13]. We also have demonstrated that membrane phosphatidylinosi- tol (PtdIns) containing sciadonate can be converted into PtdIns 4,5-bisphosphate and subsequently to diacylglycerol in response to agonistic stimulation. Despite the resulting diacylglycerol containing an unusual PUFA residue for animals, it can effectively activate protein kinase C [14]. The metabolism of PMI-PUFAs in the processes of oxidation [15] and chain elongation [16] in animal cells has also been investigated. However, the metabolic fate of PMI- PUFAs in animal cells is not fully understood. In this study, we found evidence that sciadonic acid and juniperonic acid are converted into linoleic acid and a-linolenic acid, respectively, in animal cells. This metabolic pathway can be regarded as a way to produce essential fatty acids from gymnosperm PMI-PUFAs. Results Sciadonic acid-dependent formation of linoleic acid and 16:2 D-7,10 in Swiss 3T3 cells The fatty acid compositions of phosphatidylcholine (PtdCho) and triaclglycerol of Swiss 3T3 cells cultured with purified sciadonic acid are shown in Table 1. Consistent with our previous results and other reports [6–14], sciadonic acid is available for acyl residues of glycerolipids of animal cells. In fact, 31.1% and 51.9% of the fatty acid acylated in PtdCho and triacylgly- cerol, respectively, was sciadonic acid under our experimental conditions. The sciadonic acid-supple- mented cells showed increased concentrations of linoleic acid as shown in Table 1 and Fig. 1. Dihomo- c-linolenic acid, a common C 20 PUFA in animals, has double bonds at D-8,11,14, and therefore it is an isomer of sciadonic acid. Supplementation of cells with dihomo-c-linolenic acid did not increase the concen- tration of linoleic acid, as shown in Fig. 1. Another feature of the fatty acid composition of lipids of sciadonic acid-supplemented cells was the appearance of an unknown fatty acid. This fatty acid, identified as 16:2 D-7,10 by the following experimental results, is an important metabolite in the metabolic Table 1. Fatty acid composition of PtdCho and triacylglycerol of Swiss 3T3 cells incubated with sciadonic acid (20:3 D-5,11,14). Swiss 3T3 cells were incubated with or without 50 l M sciadonic acid for 24 h. The cellular lipids were extracted and separated into each lipid class by TLC. The fatty acid composition was analyzed by GC after methanolysis. Values are weight percentages of total fatty acid, given as the mean ± SD (three cell harvests). The peak number corresponds to the number at the top of the peak on GC shown in Fig. 2A. 16:2 (D-7,10), detected as unknown fatty acid. SciA, Sciadonic acid; ND, not detected. Peak no. Fatty acid PtdCho Triacylglycerol Control SciA Control SciA 1 14:0 0.8 ± 0.1 2.2 ± 0.7 2.1 ± 0.3 1.8 ± 1.0 2 16:0 36.7 ± 1.0 35.8 ± 2.2 24.1 ± 0.2 8.4 ± 1.0 3 16:1 (D-9) 5.5 ± 0.3 2.9 ± 0.2 5.9 ± 1.1 2.1 ± 0.7 4 16:2 (D-7,10) ND 1.2 ± 0.2 ND 3.2 ± 0.7 5 18:0 8.7 ± 0.5 2.8 ± 0.8 14.0 ± 0.3 2.7 ± 1.7 6 18:1 (D-9) 33.0 ± 2.2 5.8 ± 0.5 36.5 ± 1.3 6.7 ± 2.3 7 18:1 (D-11) 7.2 ± 0.3 1.8 ± 0.2 7.4 ± 0.4 1.5 ± 0 8 18:2 (D-9,12) 1.0 ± 0.1 8.5 ± 2.9 0.6 ± 0.2 8.2 ± 1.7 9 18:3 (D-6,9,12) ND ND ND 0.7 ± 0.2 10 18:3 (D-9,12,15) ND ND ND 0.9 ± 0.7 20:3 (D-5,8,11) 2.3 ± 0.2 0.5 ± 0.2 2.9 ± 0.5 ND 11 20:3 (D-5,11,14) (SciA) ND 31.1 ± 3.4 ND 51.9 ± 5.5 12 20:4 (D-5,8,11,14) 1.9 ± 0.4 1.7 ± 0.4 0.9 ± 0.1 1.7 ± 1.1 13 20:5 (D-5,8,11,14,17) 0.7 ± 0.2 ND 0.3 ± 0.1 0.3 ± 0.3 14 22:3 (D-7,13,16) (SciA + C2) a ND 3.8 ± 1.0 ND 5.3 ± 0.2 15 22:4 (D-7,10,13,16) 0.7 ± 0 0.4 ± 0.1 0.6 ± 0.1 1.6 ± 0.4 16 22:5 (D-7,10,13,16,19) 1.0 ± 0.1 0.7 ± 0 2.6 ± 0.4 2.5 ± 0.6 17 22:6 (D-4,7,10,13,16,19) ND ND 2.6 ± 0.2 0.8 ± 0.1 a Chain-elongated metabolite. T. Tanaka et al. Polymethylene-interrupted fatty acid metabolism FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS 2729 pathway described in this paper. This unknown fatty acid was detected just after the methyl palmitoleate (16:1 D-9) on GC (Fig. 2A, peak 4), suggesting that it was an unsaturated C 16 fatty acid. We isolated it from cellular lipids of sciadonic acid-supplemented Swiss 3T3 cells by argentation TLC (AgTLC). In our AgTLC system, the methyl ester of the unknown fatty acid (R F 0.41) was located just below the methyl linol- eate (R F 0.45) on the TLC plate, suggesting that it was dienoic acid. The isolated methyl ester of the unknown fatty acid was partially hydrogenated to yield mono- enoates that retain one of the double bonds of the par- ent fatty acid. This treatment gave two species of monoenoate which can be separated by AgTLC (R F 0.36 and 0.31) with another solvent system, and they were separately converted into dimethyl disulfide adducts for GC-MS. The mass spectra and possible structures of fragment ions are shown in Fig. 2B. The set of three intense fragment ions were generated by cleavage between the methylthio-substituted carbons. They clearly showed that the two species of monoeno- ates formed by the partial hydrogenation of the unknown fatty acid were 16:1 D-7 and 16:1 D-10. Thus, the unknown fatty acid detected in PtdCho and the triacylglycerol fraction of sciadonic acid-supple- mented cells (Table 1) was identified as 16:2 D-7,10. It was detected only in sciadonic acid-supplemented cells. In neither the control cells nor the cells cultured with dihomo-c-linolenic acid was it present at a detectable level. The formation of 16:2 D-7,10 and accumulation of linoleic acid in cellular lipids depended on the concen- tration of sciadonic acid in the culture medium (Fig. 3A). Despite the considerable increase in the level 0 5 10 15 18:2 -9,12 concentration (%) PtdCho PtdEtn PtdSer PtdIns TG Lipid class +20:3 -5,11,14 +20:3 -8,11,14 Control Fig. 1. Level of linoleic acid (18:2 D-9,12) in fatty acids acylated in PtdCho, phosphatidylethanolamine (PtdEtn), phosphatidylserine (PtdSer), PtdIns and triacylglycerol of Swiss 3T3 cells. Swiss 3T3 cells were cultured without fatty acid (Control) or with 50 l M sciadonic acid (20:3 D-5,11,14) or dihomo-c-linolenic acid (20:3 D-8,11,14) for 24 h. Each lipid class was isolated from the cellular lipids by TLC, and the fatty acid composition was determined by GC. Values are weight percentages, given as the mean ± SD (three cell harvests). A B Fig. 2. Identification of 16:2 D-7,10 detected in Swiss 3T3 cells cul- tured with sciadonic acid (20:3 D-5,11,14).Swiss 3T3 cells were cultured with 50 l M sciadonic acid. (A) Fatty acid methyl esters pre- pared from the triacylglycerol fraction of the cells were analyzed by GC. The number at the top of the peak on the chromatogram cor- responds to the peak number in Table 1. Peak number 4 corres- ponds to the methyl ester of 16:2 D-7,10. (B) The isolated methyl ester of 16:2 D-7,10 was partially hydrogenated with hydrazine monohydrate. The two resulting species of 16:1 were separately isolated, and converted into dimethyl disulfide adducts for analysis by GC-MS. Polymethylene-interrupted fatty acid metabolism T. Tanaka et al. 2730 FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS of linoleic acid in cellular lipids, formation of arachi- donic acid was not apparent. Possibly, enrichment of sciadonate in cellular lipids inhibited the metabolic conversion of linoleate into arachidonate. This effect is not intrinsic to sciadonic acid because PUFA enrich- ment of cells is known to inhibit the desaturase activity required for PUFA synthesis [17]. 16:2 D-7,10 was detected in the cellular lipids when the cells were incu- bated in >50 lm sciadonic acid. We confirmed that neither purified sciadonic acid nor Torreya nucifera seed oil, the source of the sciadonic acid in this study, contains 16:2 (D-7,10) at a detectable level. From these observations, it was deduced that 16:2 D-7,10 and linoleic acid were formed as a result of the meta- bolism of sciadonate. A possible metabolic route for the synthesis of 16:2 D-7,10 and linoleic acid from sciadonic acid is shown in Fig. 4. In this pathway, 16:2 D-7,10 emerges as a result of two cycles of b-oxidation of sciadonic acid. The linoleic acid is formed by chain elongation of 16:2 D-7,10. The resulting linoleic acid is converted into a higher metabolite, such as arachidonic acid. A significant proportion of the sciadonic acid incorpor- ated into the cells seemed to be metabolized by this pathway. When Swiss 3T3 cells were incubated with 50 lm sciadonic acid, 24 lg sciadonic acid per dish was present in the cellular lipids (data not shown). At that time, increments in the concentration of 16:2 D-7,10, linoleic acid and arachidonic acid in the cellular lipids were calculated to be 0.7 , 3.5 and 0 lg ⁄ dish, respectively (Fig. 3A). The sum of these metabolites of sciadonic acid was calculated to be 4.2 lg ⁄ dish. The ratio between sciadonic acid and its metabolites was  6:1. Juniperonic acid, an n)3 analogue of sciadonic acid, is also a conifer-derived PMI-PUFA (Fig. 4). When Swiss 3T3 cells were cultured with purified juniperonic acid, the amounts of a-linoleate and EPA in the cellu- lar lipids increased depending on the concentration of juniperonic acid (Fig. 3B). When we analyzed the fatty acid of each lipid class in the control cells, a-linolenic acid was not detected in any lipid class. In contrast, it emerged as a fatty acid residue of PtdCho and phos- phatidylethanolamine (PtdEtn) at 2.9% and 1.0%, respectively, in juniperonic acid-supplemented cells. In control cells, the concentrations of EPA in fatty acids of PtdCho, PtdEtn and PtdIns were 0.7%, 2.5% and 1.7%, respectively. On the other hand, the concentra- tions of EPA in PtdCho, PtdEtn, and PtdIns in junipe- ronic acid-supplemented cells were 2.5%, 4.5% and 4.5%, respectively. These results indicate that the metabolic pathway works not only for sciadonic acid but also for juniperonic acid, an n) 3 analogue of scia- donic acid. Involvement of peroxisomal b-oxidation in the chain shortening of C 20 PMI-PUFAs It is known that peroxisomal b-oxidation of long-chain fatty acids in animals results in chain shortening with only a few b-oxidation cycles [18]. Such metabolism of long-chain fatty acids is called retroconversion and is known to be an important process in the biosynthesis of docosahexaenoic acid (22:6 D-4,7,10,13,16,19) from 24:6 D-6,9,12,15,18,21 [19–21]. We investigated the possibility that partial degradation of sciadonic acid to 0 5 10 µ g/dish 0 25 50 75 100 20:3 ( -5,11,14) ( µ M) 16:2( -7,10) 18:2( -9,12) 20:4( -5,8,11,14) A 0 2 4 µ g/dish 0 25 50 75 100 20:4 ( -5,11,14,17) ( µ M) 18:3( -9,12,15) 20 :5( -5,8,11,14,17) B Fig. 3. Supplementation of C 20 PMI-PUFAs results in the accumula- tion of common PUFAs in Swiss 3T3 cells. Swiss 3T3 cells were incubated with an increasing concentration of sciadonic acid (20:3 D-5,11,14) (A) or juniperonic acid (20:4 D-5,11,14,17) (B) for 30 h. After extraction, the cellular lipids were mixed with 15 lg 20:0 as an internal standard and subjected to methanolysis for GC analysis. The amounts of fatty acids were determined from the peak area and expressed as lg ⁄ dish ( 2 · 10 6 cells). Experiments were con- ducted in duplicate, and the mean values are given. Similar results were obtained in another two independent experiments with differ- ent cultures of Swiss 3T3 cells. T. Tanaka et al. Polymethylene-interrupted fatty acid metabolism FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS 2731 16:2 ( D-7,10) occurs in peroxisomes. The cells used for this experiment were CHO-K1 and ZP102, the wild- type and a peroxisome-deficient mutant of CHO cells, respectively [22]. The loss of PEX5, which encodes the peroxisome targeting signal-1 receptor, has been shown to cause deficiency of peroxisomes in ZP102 [22]. When CHO-K1 and ZP102 cells were incubated with sciadonic acid, sciadonic acid was incorporated into the cellular lipids of both (Fig. 5A,C). Sciadonic acid- dependent accumulation of linoleic acid and 16:2 D-7,10 was observed in the cellular lipids of the wild-type cells (Fig. 5B). On the other hand, the Fig. 4. Schematic representation of meta- bolic pathway for synthesis of essential fatty acids from C 20 polymethylene-interrupted polyunsaturated fatty acids. Fig. 5. Requirement of peroxisomes for for- mation of 16:2 D-7,10 and linoleic acid (18:2 D-9,12) from sciadonic acid (20:3 D-5,11,14) in CHO cells. CHO-K1 cells (A, B) and perox- isome-deficient CHO cells (C, D), namely ZP102 cells, were incubated with various concentrations of sciadonic acid for 48 h. The cellular lipids were mixed with 1 lg 20:0 as an internal standard and subjected to methanolysis for GC analysis. The amounts of fatty acids were determined from the peak area, and expressed as lg ⁄ dish ( 1 · 10 7 cells). Experiments were conducted in triplicate, and values are given as the mean ± SD. Polymethylene-interrupted fatty acid metabolism T. Tanaka et al. 2732 FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS cellular lipids of ZP102 cells did not accumulate 16:2 D-7,10 or linoleic acid, even though a high con- centration of sciadonic acid was used (Fig. 5D). These results clearly show that peroxisomes are required for the formation of 16:2 D-7,10 from sciadonic acid, and that 16:2 D-7,10 is essential for the biosynthesis of linoleic acid. The absence of 16:2 D-7,10 in sciadonic acid-supplemented ZP102 cells is an another guarantee of the purity of the sciadonic acid used in this study. In analogy with the results obtained with Swiss 3T3 cells, the peroxisomes of CHO-K1 cells meta- bolized a significant proportion of sciadonic acid. When the cells were cultured with 50 lm sciadonic acid (Fig. 5A,B), 14.6 lg sciadonic acid per dish was pre- sent in the cellular lipids. At that time, increments in the concentration of 16:2 D-7,10, linoleic acid and arachidonic acid in the cellular lipids were calculated to be 0.3, 2.5 and 0 lg ⁄ dish, respectively. Thus, the ratio between sciadonic acid and the sum of its meta- bolites via peroxisomes was  5:1. Conversion of 16:2 D-7,10 into linoleic acid by fatty acid chain-elongation system of microsomes We isolated 16:2 D-7,10 from the cellular lipids of scia- donic acid-supplemented Swiss 3T3 cells, and investi- gated the efficacy of its conversion into linoleic acid by the fatty acid chain-elongation system in microsomes (Fig. 6). When 16:2 D-7,10 was incubated with [2- 14 C]malonyl-CoA in the presence of microsomes, an intensely radiolabeled fatty acid was formed. The labe- led fatty acid migrated at a position corresponding to linoleic acid on AgTLC, indicating that [2- 14 C]malo- nyl-CoA was incorporated into 16:2 D-7,10 to form linoleic acid. On the other hand, little linoleic acid (18:2 D-9,12) was converted into 20:2 D-11,14 under the same experimental conditions. Interestingly, the efficacy of the conversion of 16:2 D-7,10 into linoleic acid was similar to that of c-linolenic acid into dihomo-c-linolenic acid, a common chain-elongation process in PUFA synthesis in animal cells. Discussion Over the past 40 years, a number of studies on the metabolism of PMI-PUFAs in animals have been conducted. In 1965, Takagi [23] reported that feeding sciadonate to rats that had been kept in an essential fatty acid-deficient condition caused an increase in the concentration of arachidonate in the lipids of the liver. From this observation, he suggested the possi- bility that desaturation of sciadonic acid at the D-8 position occurred in the rat liver. However, conclu- sive evidence of the presence of desaturase activity that directly converts sciadonic acid (20:3 D-5,11,14) into arachidonic acid (20:4 D-5,8,11,14) has not yet been reported. On the contrary, evidence of the inab- ility of animals to synthesize arachidonic acid from sciadonic acid has accumulated [24–28]. We also showed that desaturation at the D-8 position of scia- donic acid did not occur in rat liver microsomes [13]. In this report, we present evidence that animals can synthesize arachidonic acid from sciadonic acid inde- pendently of D-8 desaturase activity. The metabolic pathway that synthesizes arachidonic acid from sciadonic acid includes chain shortening and chain elongation of the fatty acid (Fig. 4). Because the for- mer process removes four carbon atoms from the carboxyl terminus of sciadonic acid, arachidonic acid formed from sciadonic acid does not contain the original carboxy group. This is one of the reasons why many investigators, including ourselves, did not observe the formation of arachidonic acid from 0 0.1 0.2 0.3 0.4 0.5 Labeled fatty acid (nmol / min / mg protein) 18:3 ( -6,9,12) 20:3 ( -8,11,14) 18:2 ( -9,12) 20:2 ( -11,14) 16:2 ( -7,10) 18:2 ( -9,12) Fig. 6. Conversion of 16:2 D-7,10 into linoleic acid (18:2 D-9,12) by the fatty acid chain-elongation system in rat liver microsomes. c-linolenic acid (18:3 D-6,9,12), 16:2 D-7,10 or linoleic acid was incu- bated with [2- 14 C]malonyl-CoA in the presence of rat liver micro- somes. The fatty acids formed were analyzed by AgTLC as fatty acid methyl esters. In the experiments with 16:2 D-7,10 and 18:2 D-9,12, the radioactivity located in the dienoate fraction was regar- ded as the chain-elongated metabolite. In the experiments with 18:3 D -6,9,12, the radioactivity located in the trienoate plus tetra- enoate fraction was regarded as the chain-elongated metabolite. T. Tanaka et al. Polymethylene-interrupted fatty acid metabolism FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS 2733 sciadonic acid, because we and others conducted experiments using [1- 14 C]sciadonic acid. The key metabolite in this process is 16:2 D-7,10. This fatty acid was initially detected at a low level as an unknown fatty acid in the cells cultured with sciadonic acid. We determined it that it was a 16:2 fatty acid with double bonds at D-7 and D-10 by GC- MS. It is known that peroxisomal b-oxidation does not go to completion, but results in chain shortening by only a few b-oxidation cycles [18]. Our data obtained from experiments using CHO cells clearly show that peroxisomal b-oxidation was involved in the formation of 16:2 D-7,10 from sciadonic acid. It is not known why peroxisomal b-oxidation of sciadonate stops at C 16 , but it might be explained by the substrate specificity of acyl-CoA thioesterase in peroxisomes. This enzyme in purified peroxisomes has been shown to have high activity towards acyl-CoAs with a chain length of C 16 rather than C 20 [29]. The efficacy of the chain elongation of 16:2 D-7,10 to linoleic acid was comparable to that producing di- homo-c-linolenic acid (20:3 D-8,11,14) from c-linolenic acid (18:3 D-6,9,12), the common process of C 20 PUFA biosynthesis. These results are consistent with in vivo results demonstrating the formation of linoleic acid from 16:2 D-7,10 [30,31]. As not much 16:2 D-7,10 accumulated in cellular lipids compared with linoleic acid, a large proportion of the 16:2 D-7,10, once expor- ted from peroxisomes, must be quickly metabolized to linoleic acid by the fatty acid chain-elongation system in microsomes. The observed accumulation of linoleic acid before 16:2 D-7,10 in cellular lipids may be explained by this rapid conversion. Juniperonic acid is an n )3 analogue of sciadonic acid. When we used juniperonic acid as the supplement, the amounts of both a-linolenic acid (18:3 D-9,12,17) and EPA (20:5 D- 5,8,11,14) increased in cellular lipids to an extent dependent on the juniperonic acid supplementation. This indicates that juniperonic acid is processed by a metabolic pathway similar to that of sciadonic acid (Fig. 4). It is not known why 16:3 D-7,10,13, an expec- ted peroxisomal metabolite of juniperonic acid, was not detected in cellular lipids. The efficiency of the conver- sion of 16:3 D-7,10,13 into a-linolenic acid may be greater than that of the corresponding n)6 analogue. As described in the Results section, a significant pro- portion of the sciadonate seemed to be processed by peroxisomes in both Swiss 3T3 cells and CHO cells. It should be noted that supplementation with dihomo-c- linolenic acid, a common PUFA in animals, did not affect the cellular concentration of linoleic acid (Fig. 1). Dihomo-c-linolenic acid, an isomer of scia- donic acid, might be an unsuitable substrate for this metabolic pathway. It is known that arachidonic acid is converted into linoleic acid by peroxisomal b-oxida- tion, a process known as retroconversion [32]. It is interesting to compare the efficiency of peroxisomal metabolism of PMI-PUFAs with that of common PUFAs to determine substrate preference. As men- tioned above, feeding sciadonate to rats that had been kept in an essential fatty acid-deficient condition increased arachidonate in the liver. This increase was comparable to that in a linoleate-fed group [23]. It is therefore possible that peroxisomal b-oxidation of sciadonate can supply the physiological requirement for essential fatty acids in animals grown in essential fatty acid-deficient conditions. As the data presented here are limited to rodent cells, it is not known whether this metabolic pathway exists in human cells. However, the fact that peroxi- somal retroconversion of long-chain fatty acids such as 24:6 is known to occur in humans cells [33] indicates that the metabolic pathway identified here may also be active in human cells. Gymnosperms are widely distributed. Sciadonic acid and juniperonic acid are known to be present in sev- eral species of coniferous plants [1–4]. In the present times, seeds of gymnosperms are not eaten on a com- mercial scale, but some coniferous plant seeds are used in traditional diets. For example, the roasted seeds of Torreya nucifera, which contain sciadonic acid (10% of total fatty acid), are eaten as a snack food in Japan. The metabolism of PMI-PUFAs shown here is thought to be the process by which unsuitable fatty acids are converted into the essential fatty acids required by ani- mals. It has been proposed that one of the roles of peroxisomal b-oxidation is recycling of PUFAs [34]. The present observations add new insight into the role of peroxisomal b-oxidation of fatty acids. Experimental procedures Materials DMEM, penicillin, streptomycin and fetal bovine serum were obtained from Gibco BRL and Life Technologies, Inc. (Rockville, MD, USA). ATP, CoA, malonyl-CoA and essentially fatty acid-free BSA were from Sigma Chemical Co. (St Louis, MO, USA). [2- 14 C]Malonyl-CoA was from Amersham (Arlington Heights, IL, USA). NADPH was purchased from Oriental Yeast Co. (Tokyo, Japan). Lino- leic acid, c-linolenic acid and dihomo-c-linolenic acid were purchased from Serdary Research Laboratories (London, ON, Canada). Sciadonic acid and juniperonic acid were prepared from seeds of Torreya nucifera and Biota orient- alis, respectively, as described previously [35]. The seeds Polymethylene-interrupted fatty acid metabolism T. Tanaka et al. 2734 FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS were milled in methanol, and lipids were extracted by the method of Folch et al. [36]. The lipids were subjected to methanolysis, and the methyl sciadonate and methyl juni- peronate were purified from the fatty acid methyl esters by AgTLC [35]. They were used after saponification. As judged by GC analysis, the purities of sciadonic acid and juniperonic acid were 99.0% and 100.0%, respectively. All other reagents were of reagent grade. Cell culture and analysis of cellular lipids Swiss 3T3 cells were obtained from American Type Culture Collection (Manassa, VA, USA). They were seeded in 100- mm plastic dishes at 4 · 10 5 cells ⁄ dish and maintained in 10 mL DMEM containing 10% fetal bovine serum in a humidified atmosphere of 10% CO 2 ⁄ 90% air at 37 °C. After they had grown to confluence, 50 lm sciadonic acid was added to the cell cultures as a BSA complex [13]. After 24 h, the cells were harvested, and the lipids of the cells were extracted by the method of Bligh & Dyer [37]. Each phospholipid class and triacylglycerol was isolated by TLC as described previously [13]. After preparation of fatty acid methyl esters, the fatty acid compositions of the lipid clas- ses were analyzed by GC (Shimadzu GC-14A; Shimadzu, Kyoto, Japan) equipped with a capillary column coated with CBP 20 (0.25 lm film, 30 m length; Shimadzu) [35]. The amount of each fatty acid in the cellular lipids was determined by GC using arachidic acid (20:0) as the inter- nal standard. CHO cells and a peroxisome-deficient mutant of CHO cells, named ZP102 [22], were seeded in 100-mm plastic dishes at 1 · 10 5 cells ⁄ dish. After 48 h, sciadonic acid was added to the cell cultures as a BSA complex. Cel- lular lipids were analysed as described above. Structural analysis of 16:2 D-7,10 detected in cellular lipids The 16:2 D-7,10 detected in cellular lipids was analyzed by GC-MS after derivatization. First, fatty acid methyl esters were prepared by methanolysis of the triacylglycerol frac- tion of the Swiss 3T3 cells, and the methyl ester of 16:2 D-7,10 was purified by AgTLC developed with petro- leum ether ⁄ diethyl ether ⁄ acetic acid (80:15:2, v ⁄ v ⁄ v.). Then, the isolated methyl ester of 16:2 D-7,10 was partially hydro- genated with hydrazine monohydrate at 50 °C for 25 min [35]. The two resulting isomers of the methyl ester of 16:1 that retained one of the double bonds in 16:2 D-7,10 were separately isolated by AgTLC (R F 0.36 and 0.31) developed with petroleum ether ⁄ diethyl ether ⁄ acetic acid (90:10:2, v ⁄ v ⁄ v), and converted into dimethyl disulfide adducts for analysis by GC-MS as described in [35]. Analysis of the dimethyl disulfide adducts of the methyl ester of 16:1 was carried out on a Shimadzu QP-2000 quadrupole mass spectrometer equipped with an interface for capillary GC (column coated with a 0.25 lm film of nonpolar CBJ1; 0.25 mm · 30 m; Shimadzu). The column temperature was raised from 215 °C to 265 °C at a rate of 5 °CÆmin )1 . The temperature of the injection port, ion source and interface was set at 250 °C. Helium was used as the carrier gas. The ionization energy was 70 eV. Fatty acid chain-elongation assay The 16:2 D-7,10 used in this assay was prepared from the cellular lipids from 160 (100-mm) dishes of sciadonic acid- supplemented Swiss 3T3 cells grown as described above. The purity of 16:2 D-7,10 was 92.0%. The remaining 8% was linoleic acid. The fatty acid chain-elongation assay was conducted as described previously [16]. In brief, each incu- bation consisted of 0.5 mm nicotinamide, 1.5 mm glutathi- one, 0.15 m KCl, 5 mm MgCl 2 , 0.25 m sucrose, 7.5 mm ATP, 0.4 mm CoA, 1.5 mm NADPH, 0.1 mm [2- 14 C]malo- nyl-CoA (0.05 lCi ⁄ 0.2 lmol), 50 mm potassium phosphate buffer (pH 7.0), 3.0 mg microsomal protein prepared from the liver of a Sprague-Dawley rat, and the desired fatty acid in a total volume of 2.0 mL. Fatty acids were added as a BSA complex, and incubation was conducted at 37 °C for 30 min. After the incubation, lipids in the reaction mix- ture were subjected to alkali hydrolysis, and the released fatty acids were analyzed by AgTLC after conversion into fatty acid methyl esters. When 16:2 D-7,10 and linoleic acid were used as substrates, the radioactivity of the dienoate fraction was determined. When c-linolenic acid was used, the radioactivity of the trienoate plus tetraenoate fraction was determined because of the rapid conversion of dihomo- c-linolenic acid into arachidonic acid. References 1 Smith CR Jr (1970) Occurrence of unusual fatty acids in plants. In Progress in Chemistry of Fats and Other Lipids (Holman RT, ed.), Vol. XI, Part 1, pp. 137–177. Pergamon Press, London. 2 Takagi T & Itabashi Y (1982) cis-5-Olefinic unusual fatty acids in seed lipids of gymnospermae and their distribution in triacylglycerols. Lipids 17, 716–723. 3 Wolff RL (1998) Source of 5,11,14–20:3 (sciadonic) acid, a structural analog of arachidonic acid. J Am Oil Chem Soc 75, 1901–1902. 4 Wolff RL, Deluc LG & Marpeau AM (1996) Conifer seeds: oil content and fatty acid composition. J Am Oil Chem Soc 73, 765–771. 5 Nishitama N, Chu PJ & Saito H (1995) Beneficial effects of biota, a traditional Chinese herbal medicine, on learning impairment induced by basal forebrain- lesion in mice. Biol Pharm Bull 18, 1513–1517. 6 Pasquier E, Ratnayake WMN & Wolff RL (2001) Effects of D5 polyunsaturated fatty acids of maritime pine (Pinus pinaster) seed oil on the fatty acid profile of the developing brain of rats. Lipids 36, 567–574. T. Tanaka et al. Polymethylene-interrupted fatty acid metabolism FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS 2735 7 Sugano M, Ikeda I, Wakamatsu K & Oka T (1994) Influence of Korean pine (Pinus koraiensis)-seed oil containing cis-5,cis-9,cis-12 -octadecatrienoic acid on polyunsaturated fatty acid metabolism, eicosanoid pro- duction and blood pressure of rats. Br J Nutr 72, 775– 783. 8 Matsuo N, Osada K, Kodama T, Lim BO, Nakao A & Sugano M (1996) Effects of c-linolenic acid and its posi- tional isomer pinolenic acid on immune parameters of Brown-Norway rats. Prostaglandins Leukot Essent Fatty Acids 55, 223–229. 9 Ikeda I, Oka T, Koba K, Sugano M & Lie Ken Jie MSF (1992) 5c,11c,14c-Eicosatrienoic acid and 5c,11c,14c,17c-eicosatetraenoic acid of Biota orientalis seed oil affect lipid metabolism in the rat. Lipids 27, 500–504. 10 Berger A, Fenz R & German JB (1993) Incorporation of dietary 5,11,14-icosatrienoate into various mouse phospholipid classes and tissues. J Nutr Biochem 4, 409–419. 11 Berger A & German JB (1991) Extensive incorporation of dietary D-5,11,14-eicosatrienoate into the phosphati- dylinositol pool. Biochim Biophys Acta 1085, 371–376. 12 Tanaka T, Takimoto T, Morishige J, Kikuta Y, Sugiura T & Satouchi K (1999) Non-methylene interrupted poly- unsaturated fatty acids: effective substitute for arachido- nate of phosphatidylinositol. Biochem Biophys Res Commun 264, 683–688. 13 Tanaka T, Morishige J, Takimoto T, Takai Y & Satouchi K (2001) Metabolic characterization of sciado- nic acid (5c,11c,14c-eicosatrienoic acid) as an effective substitute for arachidonate of phosphatidylinositol. Eur J Biochem 268, 4928–4939. 14 Morishige J, Takai Y, Hirano K, Tanaka T & Satouchi K (2005) Production and protein kinase C activation of diacylglycerols containing polymethylene-interrupted PUFA. Lipids 40, 155–162. 15 Ide T, Murata M & Sugano M (1995) Octadecatrienoic acids as the substrates for the key enzymes in glyceroli- pid biosynthesis and fatty acid oxidation in rat liver. Lipids 30, 755–762. 16 Tanaka T, Hattori T, Kouchi M, Hirano K & Satouchi K (1998) Methylene-interrupted double bond in poly- unsaturated fatty acid is an essential structure for meta- bolism by fatty acid chain elongation system of rat liver. Biochim Biophys Acta 1393 , 299–306. 17 Ramanadham S, Zhang S, Ma Z, Wohltmann M, Bohrer A, Hsu F-F & Turk J (2002) D6-, Stearoyl CoA-, and D5-desaturase enzymes are expressed in b -cells and are altered by increases in exogenous PUFA concentrations. Biochim Biophys Acta 1580, 40–56. 18 Kunau W-H, Dommes V & Schulz H (1995) b-oxida- tion of fatty acids in mitochondoria, peroxisomes, and bacteria: a century of continued progress. Prog Lipid Res 34, 267–342. 19 Voss A, Reinhart M, Sankarappa S & Sprecher H (1991) The metabolism of 7,10,13,16,19-docosapentae- noic acid to 4,7,10,13,16,19-docosahexaenoic acid in rat liver is independent of a 4-desaturase. J Biol Chem 266, 19995–20000. 20 Caruso D, Rise P, Galella G, Regazzoni C, Toia A, Galli G & Galli C (1994) Formation of 22 and 24 car- bon 6-desaturated fatty acids from exogenous deuter- ated arachidonic acid is activated in THP-1 cells at high substrate concentrations. FEBS Lett 343, 195–199. 21 Mohammed BS, Sankarappa S, Geiger M & Sprecher H (1995) Reevaluation of the pathway for the metabolism of 7,10,13,16-docosatetraenoic acid to 4,7,10,13,16-doco- sapentaenoic acid in rat liver. Arch Biochem Biophys 317, 179–184. 22 Tsukamoto T, Bogaki A, Okumoto K, Tateishi K, Fujiki Y, Shimozawa N, Suzuki Y, Kondo N & Osumi T (1997) Isolation of a new peroxisome-defi- cient CHO cell mutant defective in peroxisome target- ing signal-1 receptor. Biochem Biophys Res Commun 230, 402–406. 23 Takagi T (1965) The dehydrogenation of all cis-5,11, 14-eicosatrienoic acid to arachidonic acid. Bull Chem Soc Jpn 38, 2055–2057. 24 Dhopeshwarkar GA & Subramanian C (1976) Intracra- nial conversion of linoleic acid to arachidonic acid: evi- dence for lack of D 8 desaturase in the rat brain. J Neurochem 26, 1175–1179. 25 Naval J, Martinez-Lorenzo MJ, Marzo I, Desportes P & Pineiro A (1993) Alternative route for the biosynth- esis of polyunsaturated fatty acids in K562 cells. Bio- chem J 291, 841–845. 26 Dhopeshwarkar GA (1976) Biosynthesis of polyunsatu- rated fatty acids in the developing brain. II. Metabolic transformations of intracranially administrated [3- 14 C] eicosapentanoic acid, evidence for lack of D8 desaturase. Lipids 11, 689–692. 27 Sprecher H & Lee C-J (1975) The absence of an 8-desa- turase in rat liver: a re-evaluation of optional pathways for the metabolism of linoleic and linolenic acids. Bio- chim Biophys Acta 388, 113–125. 28 Chen Q, Yin FQ & Sprecher H (2000) The questionable role of a microsomal D8 acyl-CoA-dependent desaturase in the biosynthesis of polyunsaturated fatty acids. Lipids 35, 871–879. 29 Hunt MC, Solaas K, Kase BF & Alexson SEH (2002) Characterization of an acyl-CoA thioesterase that func- tions as a major regulator of peroxisomal lipid meta- bolism. J Biol Chem 277, 1128–1138. 30 Sprecher H (1968) The total synthesis and metabolism of 4-decenoate, dodeca-3,6-dienoate, tetradeca-5, 8-dienoate and hexadeca-7,10-dienoate in the fat- deficient rat. Lipids 3, 14–20. 31 Cunnane SC, Ryan MA, Craig KS, Brookes S, Koletzko B, Demmelmair H, Singer J & Kyle DJ (1995) Polymethylene-interrupted fatty acid metabolism T. Tanaka et al. 2736 FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS Synthesis of linoleate and a-linolenate by chain elonga- tion in the rat. Lipids 30, 781–783. 32 Hansen HS, Jensen B & von Wettstein-Knowles P (1986) Apparent in vivo retroconversion of dietary ara- chidonic to linoleic acid in essential fatty acid-deficient rats. Biochim Biophys Acta 878, 284–287. 33 Su H-M, Moser AB, Moser HW & Watkins PA (2001) Peroxisomal straight-chain acyl-CoA oxidase and d- bifunctional protein are essential for retroconversion step in docosahexaenoic acid synthesis. J Biol Chem 276, 38115–38120. 34 Sprecher H (2002) The role of anabolic and catabolic reactions in the synthesis and recycling of polyunsatu- rated fatty acids. Prostaglandins Leukot Essent Fatty Acids 67, 79–83. 35 Tanaka T, Shibata K, Hino H, Murashita T, Kayama M & Satouchi K (1997) Purification and gas chromatographic-mass spectrometric characterization of non-methylene interrupted fatty acid incorporated in rat liver. J Chromatogr B 700, 1–8. 36 Folch J, Lees M & Sloane-Stanley GH (1957) A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem 226, 497–509. 37 Bligh EG & Dyer WJ (1959) A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37, 911–917. T. Tanaka et al. Polymethylene-interrupted fatty acid metabolism FEBS Journal 274 (2007) 2728–2737 ª 2007 The Authors Journal compilation ª 2007 FEBS 2737 . Metabolic pathway that produces essential fatty acids from polymethylene-interrupted polyunsaturated fatty acids in animal cells Tamotsu. D-7,10 into linoleic acid was similar to that of c-linolenic acid into dihomo-c-linolenic acid, a common chain-elongation process in PUFA synthesis in animal