Low C18 to C20 Fatty Acid Elongase Activity and Limited Conversion of Stearidonic Acid, 18:4(n-3), to Eicosapentaenoic Acid, 20:5(n-3), in a Cell Line from the Turbot, Scophthalmus Maximus

The TF cell line, derived from a carnivorous marine teleost, the turbot (Scophthalmus maximus), is known to have a reduced rate of polyunsaturated fatty acid (PUFA) biosynthesis. In order to establish the enzymic step responsible, the metabolic conversions of a range of PUFA including the key intermediates of n-3PUFA metabolism, stearidonic acid (18:4n-3) and its elongation product 20:4n-3, were studied in TF cells and compared with the AS cell line derived from Atlantic salmon (Salmo salar). Cells were cultured in the presence of unlabelled (25 µM), [U-14 C] (1 µM/0.25 µCi) or deuterated (d5; 25 µM) fatty acids. The metabolism of stearidonic acid was qualitatively similar in both cell lines, predominantly elongation to 20:4n-3 with subsequent desaturation to eicosapentaenoic acid (20:5n-3), but there were significant quantitative differences. The alternative techniques and mass concentrations of substrates used gave different amounts of overall conversion, but all the results were consistent with an apparent partial deficiency in C 18 to C 20 fatty acyl elongation in TF cells compared to AS cells. In contrast, Δ5 desaturation was apparently greater in TF cells than in AS cells. Only small amounts of docosahexaenoic acid (22:6n-3) were produced by either cell line, although there was significant production of 22:5n-3, especially when 20:4n-3 was supplemented to cultures. From the results obtained in this study it is concluded that there is an apparent partial deficiency of C 18 to C 20 fatty acyl elongation in the turbot cell line and that this, rather than lower fatty acyl Δ5 desaturation, may contribute to the reduced rate of conversion of 18:3n-3 to 20:5n-3 observed in the turbot compared with salmonid fish in vivo.


Introduction
Vertebtrate animals lack the Δ12 and Δ15 fatty acyl desaturase enzymes responsible for the production of 18:2n-6 and 18:3n-3 from 18:1n-9 and so these essential fatty acids (EFA) are required in the diet and serve as the precursors for the longer-chain, more unsaturated polyunsaturated fatty acids (PUFA) characteristic of animal cells [1][2][3][4]. PUFA are essential components of biological membranes that have a role in the gross structural functions of the membrane, such as fluidity, and also have a considerable impact on the activity and function of membrane-associated proteins including receptors and enzymes [1][2][3][4]. PUFA themselves are precursors of a whole range of highly biologically active derivatives such as eicosanoids that are important in various forms of cell signalling [5][6][7][8]. Therefore, PUFA have important roles in both normal physiology and many pathological conditions. The EFA desaturation/elongation pathways in animals are still to be fully characterised as evidenced by the recent work suggesting that the production of 22:6n-3 proceeded without the Δ4 desaturase activity classically thought to be responsible for introduction of the final double bond [9].
Studies of these pathways have been greatly facilitated by the use of cell lines [10].
The EFA requirements of freshwater and marine fish are qualitatively different. For example, in the freshwater rainbow trout (Oncorhynchus mykiss)18:3n-3 alone can satisfy the EFA requirement, with 18:2n-6 only required for optimal growth [11]. However, in most marine species, including turbot (Scophthalmus maximus), the longer chain PUFA 20:5n-3 and 22:6n-3 are required [12]. This suggested a species difference in fatty acid desaturase/elongase activities and it was subsequently shown that this in vivo difference was also present in cultured cell lines [13]. Studies involving supplementation of turbot cells (TF) in culture, compared to both rainbow trout cells (RTG-2) and Atlantic salmon cells (AS), with various n-3 and n-6 PUFA had shown that the deficiency in the desaturase/elongase pathway in turbot 4 was either in the C 18 to C 20 elongase (C 18-20 elongase) or the fatty acyl Δ5 desaturase step [13,14]. Studies performed on turbot in vivo were also inconclusive but more consistent with a deficiency in the Δ5 desaturase activity as the fish appeared unable to produce specifically the Δ5 desaturation products 20:4n-6 and 20:5n-3 [12,15]. In addition, results from in vivo studies with other marine fish species such as gilthead sea bream (Sparus aurata) and golden grey mullet (Liza aurata) were consistent with a deficiency in Δ5 desaturation [16,17]. Reduced flux through desaturase/elongase pathways due to reduced desaturase activities have precedents in terrestrial carnivores such as cats in which Δ6 and Δ5 desaturase activities may both be severely limited [18][19][20][21]. It was hypothesised that this situation may be an evolutionary response to differences in diet. Therefore, whereas the natural diets of the generally more herbivorous/omnivorous freshwater fish, such as the well-studied salmonids, are relatively rich in C 18 PUFA and do not contain much C 20 or C 22 PUFA, marine fish like the highly carnivorous turbot consume diets rich in 20:5n-3 and 22:6n-3 and so do not require to biosynthesise these PUFA [22,23].
As the situation in marine fish was unclear, we aimed to further define the enzymic location limiting the desaturase/elongase pathway in the turbot cell line. We report here the metabolism of various PUFA, including the key intermediates 18:4n-3 and 20:4n-3, in the TF and AS cell lines and that there was an apparent partial deficiency in C 18-20 elongation and not Δ5 desaturation in TF cells.
The details of the fatty acid substrates added to the cell cultures are summarised in Table   1. Preliminary experiments incubating AS cells with 25µM unlabelled 18:4n-3 as fatty acid salt/BSA complex, as methyl ester/BSA complex and as methyl ester in ethanol, showed no difference in the metabolism to 20:4n-3 and 20:5n-3 after 6 days of culture, therefore these three chemical forms of substrate can be considered equivalent for the purpose of the present study. No accumulation of methyl esters was found in the lipid composition of cells incubated with methyl esters, compared to fatty acid salt/BSA complexes.
Lipid was determined gravimetrically and resuspended in chloroform/methanol (as above) at a concentration of 10 or 100 mg/ml and stored at -20°C under argon.
Separation and quantification of lipid classes was performed by single-dimension double-development HPTLC followed by scanning densitometry as described previously [28].
Fatty acid methyl esters (FAME) were prepared by acid-catalysed transmethylation, extracted and purified by HPTLC as described previously [29]. Analysis of FAME was performed by GC in a Fisons GC8000 gas chromatograph (Crawley, UK) equipped with a fused-silica capillary column (30m x 0.32 mm i.d., CP Wax 52 CB, Chrompack, UK) using hydrogen as carrier gas. Temperature programming was from 50 to 150°C at 35°C/min and to 225°C at 2.5°C/min. Individual FAME were identified by comparison with known standards and published data [30,31].
For radioisotope experiments, FAME were prepared as above and separated by argentation chromatography on TLC plates impregnated with 2 g silver nitrate in 20 ml acetonitrile and activated at 110°C for 30 min. Plates were developed with toluene/acetonitrile (95:5, v/v) to separate PUFA [32] and autoradiography was performed using Kodak BioMax MR film for 6 days at room temperature. Bands corresponding to individual FAME, identified 8 in comparison with known standards, were scraped from the plate and assayed for radioactivity by scintillation counting (TRI-CARB 2000CA, United Technologies Packard, Pangbourne, UK). Data were corrected for quench and counting efficiency determined. For [U-14 C]18:4n-3 metabolism, the identity of radioactive fatty acids recovered from cell incubations was also confirmed by capillary radio-GC, by repeating the incubations with a higher concentration of substrate (1µCi/flask).
Total lipid FAME of samples incubated with deuterated fatty acids were analysed by GC as above and the identity of deuterated fatty acids confirmed by GC mass-spectrometry (MS).
Where the concentrations of deuterated fatty acids were too low to be determined in this way, FAME were derivatised to pentafluorobenzyl esters [33] and quantified by negative chemical (100 µE s -1 ). Cells were harvested by centrifugation at 2500 g for 5 min and the algal pellet extracted with chloroform/propan-2-ol (2:1, v/v) [27]. Over 1 mCi of radioactivity was recovered in the total lipid extract (13 mg). The total lipid was applied to a TLC plate, chromatographed in diethyl ether, dried under vacuum and developed a second time in methyl acetate/propan-2-ol/ chloroform /methanol/ 0.25% aqueous KCl (25:25:25:10:9, by vol.) [35].
After desiccation, mono-and di-galactosyldiacylglycerols (MGDG and DGDG) were detected under UV light after spraying with 0.1% 2',7'-dichlorofluorescein (DCF) in 97% methanol, the silica scraped and lipids transmethylated directly on the silica as above. FAME were extracted as described previously [29], quantified gravimetrically and radioactivity determined as described above. 25.1 µCi were recovered in 0.63 mg FAME from MGDG and 11.2 µCi were recovered in 0.29 mg FAME from DGDG. FAME were separated by argentation TLC as above and the individual FAME bands identified after 1 h autoradiography. The silica containing 18:4n-3 was scraped from the plate, 5 ml 2% aqueous NaCl added and the silica extracted 6 times with 10 ml hexane/diethyl ether (1:1, v/v) and once with 20 ml chloroform/ methanol (2:1, v/v) to elute FAME. Pooled fractions were dried and radioactivity determined as above. A total of 5 µCi of [U-14 C]18:4n-3 was obtained, with a specific activity of approximately 12 mCi/mmol. The identity of the 18:4n-3 was confirmed and its purity (>99%) checked by radio-gas chromatography as described by Buzzi et al. [36]. incubated for 6 passages with (d5)18:4n-3 methyl ester. HPLC separation was performed as previously described, but as 20:4n-3 coeluted with 16:2n-9, the sample was rechromatographed using methanol/water (97:3, v/v) at the same flow rate. Identity, purity and adjustment of concentration was performed as described above.

Statistical analysis
Results are presented as means ± SD (n = 3 or 4). The data were checked for homogeneity of the variances by the Bartlett test and, where necessary, the data were arc-sin transformed before further statistical analysis. Differences between mean values were analysed by the Student t-test or one-way analysis of variance (ANOVA), followed when pertinent by a multiple comparison test (Tukey). Differences were reported as statistically significant when P < 0.05 [37].

Effects of 18:4n-3 and 20:4n-3 on cell growth, lipid content and composition
Supplementation of TF and AS cells with 18:4n-3 and 20:4n-3 at final concentrations of 11 25 µM did non influence the proliferation rate of cells, nor their lipid content, as shown in Table 2. The only difference detected between the two cell lines was in the culture density, higher for TF than for AS cells as the latter are bigger cells. The ratio between the lipid and protein content of both cell lines was constant under all conditions (0.2 mg lipid/ mg protein), indicating that no extra lipid deposition occurred in the cells due to supplementation with these fatty acids at the concentration used. Consistent with this, the lipid class compositions, which were very similar in TF and AS cells, were not significantly changed by supplementation with 18:4n-3 and 20:4n-3. The majority of lipids were polar lipids (72%) with the neutral lipids being predominantly cholesterol (23%). Phosphatidylcholine was the most abundant phospholipid (30% of total lipid), followed by phosphatidylethanolamine (19%), phosphatidylinositol (7.5%) and phosphatidylserine (7%).

Effects of 18:4n-3 and 20:4n-3 on fatty acid compositions
The fatty acid profiles of total lipids from TF and AS cells in control cultures (2% FCS) and cultures supplemented with 25 µM 18:4n-3 or 20:4n-3 are shown in Table 3. Total PUFA were only slightly increased in supplemented TF cells, whereas PUFA more than doubled in supplemented AS cells, compared to control cultures. PUFA of the n-3 series were increased in supplemented TF cells (up to 19-21% from 8%) and supplemented AS cells (up to 27-29% from 6%), compared to controls. Supplementation with 18:4n-3 increased18:4n-3 (from about zero to 8% in TF and 12% in AS) and 20:5n-3 (from 0.5% to 3% in TF and from 0.8% to 5% in AS) in both cell lines, whereas 20:4n-3 was only increased (zero to 5%) in AS cells.
The increased percentages of n-3PUFA were balanced by decreased percentages of monoenes, predominantly 18:1n-9, in AS cells whereas in TF cells, although monoenes were decreased, the main balancing reduction was in n-9PUFA, predominantly 18:2n-9, which were reduced in supplemented cells (from 20% to 12-13%) ( Table 3). PUFA of the n-6 series were only slightly decreased in 20:4n-3 supplemented cells in both lines. Saturated fatty acids were generally increased in supplemented cells, particularly TF cells, primarily due to increased 18:0.

Metabolism of deuterated 18:3n-3, 18:4n-3 and 20:4n-3
The results for the metabolism of deuterated fatty acids are presented in a similar way to the metabolism of 14 C-labelled fatty acids, i.e. data are percentages (by weight) of total deuterated fatty acids recovered in total lipids (Table 6).
Chain shortened products were generally more prominent in AS cells than TF cells (Table 6). For (d5)18:4n-3, twice as much deuterated 16:4n-3 was detected in AS cells compared to TF. The same fatty acid was also present in cells incubated with (d5)20:4n-3, again at higher levels in AS (11%) than in TF (7%), although comparable percentages of deuterated 18:4n-3 were detected.

Individual metabolic steps in desaturation/elongation pathway
A summary of the results obtained in the isotopic experiments, derived from the data contained in Tables 4-6, is presented in Table 7. The amounts of the products of each individual step in the desaturation/elongation pathway were summed taking into account all 14 the fatty acids that are a result of that step, irrespective of subsequent conversions. The results obtained for each enzymatic process when incubating the cells with the fatty acid which is the direct substrate (i.e. 18:4n-3 for C 18-20 elongase and 20:4n-3 for Δ5 desaturase) are of particular interest as they are an indication of the potential activity of each enzyme independent of the effects of enzymes preceding it in the pathway. Table 7 shows C 18-20 elongase products are always lower in TF cells than in AS cells for both n-6 and n-3 substrates, and this is particularly evident with the direct substrate, 18:4n-3, both radioactive (26% in TF and 81% in AS) and deuterated (10% in TF and 19% in AS). In contrast, Δ5 products are higher in TF cells compared to AS cells (65% v. 40%, 71% v. 31% and 10% v. 6%, respectively) when a direct substrate is given to the cells ([U- 14

Discussion
The present study attempted to establish the enzyme responsible for reduced flux in the fatty acid desaturation/elongation pathway in a cell line, TF, derived from turbot. In order to do this, the direct substrates for all the component enzymes in the pathway were required so that each step could be assayed and determined in isolation from the influence of preceding enzymes in the pathway. As isotopes for two crucial fatty substrates, 18:4n-3 (the direct substrate for C 18-20 elongase) and 20:4n-3 (the direct substrate for Δ5 desaturase) were not available commercially they had to be synthesised before supplementation to TF and AS cells and analysis of the metabolic products. Therefore, the investigation focused on the metabolism of these two key intermediate fatty acids and in so doing suggested a likely biochemical cause for the lower rates of overall conversion in the marine fish, the turbot, when compared with the salmonid, the Atlantic salmon.
All the results, using unlabelled, [1-14 C]-, [U-14 C]-and (d5)-labelled fatty acids, are consistent with the fact that the TF cell line has a partial deficiency or reduced rate of C [18][19][20] elongation, while Δ5 desaturase activity is apparently greater than in the AS cell line. In an earlier study on the TF cell line, "an apparent deficiency in the C 18 to C 20 elongase multienzyme complex" was hypothesised as the reason for the low conversion of 18:3n-3 to 20:5n-3 [13]. However, studies in vivo [12,15], together with the general belief that desaturase activities were more likely to be controlled and, therefore, the rate limiting steps in desaturation/elongation pathways [38] together with evidence that terrestrial carnivores appeared to lack, or express relatively low Δ6 and, possibly, Δ5 desaturases [18] led to the proposition that the Δ5 desaturase activity would be deficient in carnivorous marine fish in general, including turbot and sea bream (Sparus aurata L.) [16]. Only the use of isotopically labelled fatty acid intermediates could attempt to discriminate between the two hypotheses of reduced Δ5 desaturation or reduced C 18-20 elongation.
The present results, suggesting that TF cells have a partial deficiency in C 18-20 elongation would be consistent with the known EFA requirements of turbot. However, Maeda et al. [39] found that three out of six mammalian cell lines had lost Δ6 desaturase activity, Δ5 desaturase activity appeared to be absent in four cell lines and only one cell line expressed any desaturase activity beyond Δ5. There is also evidence for loss of stearoyl-CoA Δ9 desaturase in murine T lymphocytes in culture [40]. Therefore, loss or down -regulated activities in the desaturase/ elongase pathway may be a common feature of cell lines, and so it is prudent to be cautious in translating the results in cell culture to in vivo. In consequence, it is important to determine that this situation in TF cells is the one in vivo, and that TF cells are a model for the whole animal, via experiments with deuterated substrates similar to those performed by other authors in felines [41,42] and humans [43,44].
Cell culture studies have proved very useful in determining the potential effects of fatty acid supplementation on cellular fatty acid compositions as a more rapid and less expensive alternative to dietary trials with whole animals [13]. Therefore, the incubations with unlabelled 18:4n-3 and 20:4n-3 at 25 µM showed the effects of these fatty acids on the fatty acid compositions of TF and AS cells. Previously, we established that a fatty acid concentration of 25 µM in the medium alters the membrane phospholipid fatty acid composition of the cells without affecting the lipid class composition or precipitating the appearance of cytoplasmic lipid droplets [13]. The (d5) isotopes were also used at 25 µM enabling the metabolism of the supplemented fatty acid to be determined under conditions known to affect fatty acid compositions. This is in contrast to the radiotracer studies where concentrations of fatty acid rarely exceed 1 µM, a concentration which has little impact on cellular fatty acid compositions. It is therefore noteworthy that the data obtained from all the experiments were generally consistent and qualitatively similar although there were quantitative differences between tracer experiments and those using a significant mass of fatty acid substrate. In the latter case, the effects of cellular fatty acid composition in altering enzyme substrate concentrations and inhibitory processes such as product inhibition are evident.
There are few data in the literature describing the metabolism via desaturation/elongation of either 18:4n-3 or 20:4n-3 by other cell lines. However, incubation of NIH-3T3 cells with approximately 70µM unlabelled 18:4n-3 resulted in the accumulation in phospholipids of primarily 20:5n-3 (15%) and 20:4n-3 (11%) followed by 18:4n-3 (10%) and 22:5n-3 (4%), whereas the percentage of 22:6n-3 was only slightly increased [45]. This pattern was similar to AS cells except the percentages were higher in NIH-3T3 cells possibly due to a mass effect, as discussed above, as the level of 18:4n-3 supplemented to the 3T3 cells was almost 3-fold greater. This also resulted in significant deposition of fatty acids, including 20:4n-3, 18:4n-3 and 20:5n-3, in triacylglycerols which were increased from trace amounts to almost 16% of total cellular lipid [45], whereas there was no accumulation of fatty acid in neutral lipid and no accumulation of any neutral lipid class, including triacylglycerols, in the present study. In a separate feeding study, rats were fed a lipid-free diet supplemented with either 1% 18:3n-3-ethyl ester or 1% 18:4n-3-ethyl ester and the effects on fatty acid compositions of plasma and liver lipid fatty acid compositions determined [46]. The molar ratio of 20:5n-3 in most lipid fractions was approximately 2-fold higher in the rats fed 18:4n-3, and 18:4n-3 was found in the liver lipids in only very small amounts suggesting that it was rapidly metabolised to 20:5n-3 and that Δ6 desaturase was the limiting step in the production of 20:5n-3 from 18:3n-3 [46].
Apparently higher levels of Δ6 desaturation were observed in the TF cell line compared to the AS cell line. Higher Δ6 desaturase activity in TF cells, compared to another cell line, RTG-2 from trout, had been reported previously [13], but is more interesting with the observation that TF also showed apparently higher Δ5 desaturase activity than AS cells.
However, increased cellular levels of 18:2n-6 may inhibit Δ6 desaturation [47] and, as the level of 18:2n-6 is slightly higher in AS cells than in TF cells, this this may be a contributing factor. The use of sequential substrates clearly demonstrated how the previous enzyme in the desaturation/elongation pathway can affect the apparent activity of subsequent enzymes. For instance, with both [1-14 C]18:3n-3 and [U-14 C]18:4n-3, the results could suggest that AS cells had higher Δ5 desaturase activity than TF cells but this is shown to be highly dependent on the preceding activities with reduced C 18-20 elongase activity in TF cells.
In conclusion, the present results provide evidence that TF cells have a partial deficiency in C 18-20 elongation which would be consistent with the known EFA requirements of turbot.
However, it is important to determine that this situation is the one in vivo, and that TF cells are a model for the whole animal, so appropriate in vivo experiments perhaps similar to those performed in felines and humans are required [41][42][43][44].