The Biosynthesis of Erucic Acid in Developing Embryos of Brassica rapa¹

Abstract

Erucic acid (cis-13-docosenoic acid) and its homolog, cis-11-eicosenoic acid, are commonly found in the seed oils of the Cruciferae. The oils and the corresponding fatty acids are produced by high-erucic acid cultivars of Brassica and Crambe species and are oleochemical commodities. The reactions leading to the synthesis of erucic acid are, for the most part, well understood. In the developing embryos of Brassica napus (Downey et al., 1964), Crambe abyssinica (Appleby et al., 1974), Simmondsia chinensis (Ohlrogge et al., 1978), Tropaeolum majus (Pollard and Stumpf, 1980a), and Limnanthes alba (Pollard and Stumpf, 1980b), it was demonstrated that erucic acid is synthesized by the elongation of oleic acid rather than by de novo synthesis. This conclusion was deduced from the distribution of label in the long-chain fatty acids after incubating seed tissue with exogenous [¹⁴C]acetate. The label was preferentially incorporated into the carboxyl-terminal carbons of the long-chain fatty acids rather than the methyl-terminal 18 carbons. Subsequently, the work of Ohlrogge et al. (1979) demonstrated that de novo fatty acid synthesis was almost exclusively located in the chloroplast in spinach leaves. Thus, the hypothesis on the biosynthesis of erucic acid was extended to a description in which 18:1 (oleate), synthesized in the plastid, was exported to the cytosol, where, presumably in the endomembrane system, it was elongated via malonyl-CoA-requiring elongases to C-20 and longer-chain monounsaturated fatty acids. Several reports (e.g. von Wettstein-Knowles, 1993; Imai et al., 1995) have confirmed that the location of the oleoyl elongation system was extraplastidial, being associated with oil bodies or microsomal membranes. Créach et al. (1993) and Fehling and Mukherjee (1991) showed that oleoyl-CoA is readily elongated in vitro and that the intermediates of the elongation reaction are acyl-CoA thioesters.

It is generally accepted that the end product of newly synthesized oleic acid exported from plastids is oleoyl-CoA. This oleoyl moiety can be elongated directly to erucic acid in the ER or in oil body-associated membranes through successive additions of two carbons derived from malonyl-CoA. However, in an oil body fraction from developing rapeseed, Hlousek-Radojcic et al. (1995) observed in vitro that radioactivity from oleoyl-CoA was incorporated into 20:1 (eicosenoate) and 22:1 (erucate) at least 2.5-fold more slowly than from malonyl-CoA. Furthermore, radioactivity from oleoyl-CoA was rapidly diluted upon the formation of eicosenoyl-CoA and the elongation could proceed without the addition of exogenous oleoyl-CoA. Based on these in vitro observations, they concluded that oleoyl-CoA is not the immediate substrate for elongation. Instead, they proposed that the intermediate oleoyl donor for the elongase may be either a lipid or an unesterified acid. Furthermore when intact Brassica (which are high in erucic acid) embryos were incubated with [¹⁴C]acetate, PC was always heavily labeled at early time points, with 18:1 constituting more than 90% of the ¹⁴C-labeled fatty acid esterified to PC (X. Bao and J. Ohlrogge, unpublished data). We considered that this oleoyl-PC might contribute to the synthesis of erucic acid, either via a mechanism of direct acyl transfer, as proposed by Hlousek-Radojcic et al. (1995), or via the acyl exchange between acyl-CoA and PC, as first reported by Stymne and Stobart (1984). In this study we examined whether oleoyl-CoA produced by plastids is directly elongated to erucic acid, or if the oleoyl-CoA enters another intermediate pool before it is elongated. In addition, we have examined the influence of light and of an inhibitor of the homodimeric ACCase on erucic acid biosynthesis. Taken together, our results show an additional level of complexity in the biosynthesis of erucic acid: The supply of oleoyl groups for chain elongation is a combination of the release of 18:1 from a large, intermediate lipid pool, probably PC, and the direct provision of newly synthesized 18:1 from the plastid.

MATERIALS AND METHODS

RESULTS

Accumulation of Fatty Acids during B. rapa Seed Development

As shown in Figure ​Figure1,1, developing B. rapa seeds accumulate fatty acids that are primarily derived from oleic acid. The modifications of oleic acid fall into two mutually exclusive types: further desaturation and further chain elongation. Elongation is the more prevalent modification, because erucic acid is the most abundant fatty acid, reaching a level of 56 mol % in seeds at 43 DAF, whereas 18:2 (linoleate) plus 18:3 (linolenate) total <20% at the same stage. 16:0 (palmitate) and 18:0 (stearate), with a content of <3% in mature seeds, are the only significant fatty acids that do not derive from oleic acid. The studies described below were conducted on seeds at 28 DAF, when total fatty acid and erucic acid were accumulating at maximum rates. At this stage, the rate of total 18:1 production, including its elongated and desaturated derivatives, was 16 nmol h⁻¹ embryo⁻¹.

Open in a separate window

Figure 1

Accumulation of fatty acids during B. rapa development. Lipids were extracted from 20 pooled embryos at the times indicated and fatty acid methyl esters were separated and quantified by GC. The total 18:1 derivatives were obtained by adding 18:1, 18:2, 18:3, 20:1, and 22:1 together. The accumulation rate of 18:1 derivatives (16 nmol h⁻¹ embryo⁻¹) was calculated for the embryos at 28 DAF when labeling experiments were conducted.

Light Alters Relative Proportions of Oleic and Erucic Acid Synthesized

To monitor how light influences the accumulation of 18:1 and 22:1, and TAG versus PC, incubations with [¹⁴C]acetate were carried out either in light or in the dark. As shown in Figure ​Figure2, 2, incubation of embryos in the light increased radioactivity in 18:1 approximately 2-fold, whereas the radioactivity in 22:1 was only fractionally higher in light incubations. The comparatively small effect of light on radioactivity of 22:1, which is highly labeled at the carboxyl end, suggests that the homodimeric acetyl-CoA carboxylase and fatty acid elongation are not strongly influenced by light, whereas the 50% reduction of radioactivity in 18:1 in dark versus light implies that a major site of light regulation is located in the plastids, which are responsible for the de novo fatty acid synthesis. An alternative interpretation of these experiments is that reduction in label of 18:1 in the dark reflects changes in the endogenous pools of acetate. This explanation was ruled out because very similar light dependence was also observed when [¹⁴C]Suc or [³H]water was used as the precursor for fatty acid synthesis (not shown). These observations suggest that an endogenous pool of 18:1 might contribute to the synthesis of cis-11-20:1 and 22:1. If the newly synthesized 18:1 was the predominant source of oleoyl moieties for chain elongation and the synthesis of 18:1 was reduced by one-half in dark, we would expect that the labeling of 22:1 would also be reduced similarly. In fact, very little reduction of 22:1 labeling occurs in the dark.

Open in a separate window

Figure 2

Effects of light on the synthesis of 22:1 (A), 18:1 (B), TAG (C), and PC (D). ▪, Dark; □, light. Radioactivity is expressed per embryo.

Analysis of the Distribution of Label from [1-¹⁴C]Acetate in Long-Chain Fatty Acids with Time

Erucic acid is synthesized from 18:1 by the addition of two 2-carbon units from two molecules of malonyl-CoA. When exogenous [¹⁴C]acetate is used as a substrate for the biosynthesis of 22:1, the 2-carbon units from the chain elongation of 18:1 have a higher specific activity than the 2-carbon units of the methyl-terminal 18 carbons, which are derived from de novo fatty acid synthesis of 18:1. The same applies to cis-11-20:1, except that only one elongation cycle occurs from 18:1. This differential labeling of C-20 and longer fatty acids from exogenous acetate was previously documented for four different oilseed species: Brassica napus (Downey et al., 1964), Simmondsia chinensis (Ohlrogge et al., 1978), Tropaeolum majus (Pollard and Stumpf, 1980a), and Limnanthes alba (Pollard and Stumpf, 1980b), and has been interpreted as reflecting different pools of acetate supplying the de novo fatty acid synthesis and the chain-elongation reaction. What has not been examined, however, is the time dependency of this differential labeling. As we will demonstrate below, this time dependency provides information on the pool of 18:1 supplying chain elongation.

Oxidative cleavage at the double bond of monounsaturated fatty acid methyl esters allows determination of the relative specific activity of ¹⁴C in the acid and the acid-ester fragments of the acyl chain because both of the fragments can be quantitatively recovered. Table ​TableII presents the ratio of radioactivity in each fragment for isolated 18:1, cis-11-20:1, and 22:1 when 4-week-old developing B. rapa embryos were incubated with [1-¹⁴C]acetate. The ratios were measured after various incubation times during a 1-hour period. It is expected that 18:1 will be uniformly labeled, and if the substrate is [1-¹⁴C]acetate, the theoretical distribution between the C-9 acid-ester and the C-9 acid fragments will be 1.25 (5/4). The measured ratios decreased in the range of 1.24 to 1.30, with an average value of 1.259 ± 0.016. As an additional control, we used permanganate-periodate cleavage of commercial [U-¹⁴C]18:1. The theoretical ratios of 1.0 (C9AE/C9A) in 18:1 was obtained exactly (1.00 ± 0.002). For further confirmation that this method is suitable for quantitative analysis, the radioactivity in fatty acids was measured before the oxidation, and was measured again after extraction of cleavage products from the reaction solution. Loss of radioactivity was negligible throughout the procedure. All of these controls indicate the high degree of accuracy and reproducibility of the technique, which is essential, because the variations in the ratios with time are the key experimental data.

Table I

Ratio of radiolabel in oxidative cleavage fragments of ¹⁴C-labeled fatty acids from incubation of B. rapa embryos with [1-¹⁴C]acetate

	5 min	10 min	20 min	30 min	40 min	50 min	60 min
Light
C-13/C-9 of 22:1	9.81	9.68	8.65	8.29	7.92	7.42	6.42
C-11/C-9 of 20:1	5.58	4.77	4.52	4.39	4.38	4.14	3.63
C-9/C-9 of 18:1	1.28	1.25	1.27	1.26	1.24	1.25	1.25
Dark
C-13/C-9 of 22:1	14.42	13.23	10.87	9.73	9.51	9.38	8.93
C-11/C-9 of 20:1	8.01	7.21	5.48	5.44	5.34	5.28	5.02
C-9/C-9 of 18:1	1.30	1.27	1.26	1.25	1.25	1.26	1.24

Open in a separate window

Ratios of radioactivity in the 13-C, 11-C, and 9-C carbon fragments were determined after permanganate-periodate oxidation at the double bond and isolation of the cleavage products.

Table ​TableII presents the direct measurements of the ratio of ¹⁴C label in acid and ester fragments. The results indicate that the ¹⁴C ratio in the oxidative cleavage fragments of 18:1 remained constant with time and was close to the expected value. In contrast, the ratios of ¹⁴C in the C13AE/C9A fragments of 22:1 and the C11AE/C9A fragments of 20:1 both decreased with time in both light- and dark-incubated embryos. The experiment was repeated 10 times, and although the absolute values of the acid-to-acid-ester ratios at each time point showed some degree of variation between experiments, the trend was always the same. Also noted in Table ​TableII and consistent with other experiments described below was the observation that the acid-ester-to-acid ratios for 22:1 and 20:1 were always higher in dark- than in light-incubated embryos at any given time point. Thus, in the dark the relative specific activity of C-2 units used for elongation when compared with C-2 units derived from de novo fatty acid synthesis was increased by a factor of 1.5 to 1.6.

It is also instructive to consider these labeling data in terms of the proportion of total plastid-produced [¹⁴C]18:1 units that appear in erucic acid. As calculated from the fatty acid compositions shown in Figure ​Figure1,1, more than 55% of 18:1 fatty acids synthesized by 28 DAF in developing seeds are elongated to erucic acid. Determination of the ¹⁴C in the oxidative cleavage products of 22:1 allows calculation of the ¹⁴C content of the de novo-synthesized 18 carbons. Comparison of this value with the total 18:1 radioactivity accumulated in the incubation (18:1 plus 18-carbon portion of 20:1 and 22:1) gave the values plotted in Figure ​Figure3A.3A. After 5 min of incubation in the light, only 21% of the total 18:1 produced in the incubation appeared in 22:1, and this value increased to 35% by 60 min. Thus, there was a substantial lag in the appearance of ¹⁴C in the 18:1 portion of 22:1. In dark incubations, a higher proportion of [¹⁴C]18:1 initially appeared in erucic acid, but the increase with time was similar. The labeling of the 20:1 de novo and elongation carbons showed parallel patterns in both light and dark (data not shown).

Open in a separate window

Figure 3

Labeling of fatty acids by [¹⁴C]acetate in developing B. rapa embryos. A, The percentage of [¹⁴C]acetate incorporated into total 18:1 (18:1 plus the derived 18:1 portion of 20:1 and 22:1) that appears in 22:1 is plotted versus time. The incubation was conducted either in the light (□) or in the dark (▪). B, Time course of ¹⁴C accumulation into total 18:1 derivatives (18:1 plus the derived 18:1 portion of 20:1 and 22:1) in the light. All of the fatty acids were derived from total lipid extracts, and the numbers represent radioactivity calculated per embryo.

Two general explanations can be proposed for the change in labeling within the very-long-chain fatty acids over time. The first is that there are different-sized acetate-accessible pools supplying malonyl-CoA for de novo fatty acid synthesis and for chain elongation. The pool supplying chain elongation would be relatively small, rapidly reaching a steady-state contribution from exogenous acetate. The pool supplying de novo fatty acid biosynthesis would be large and equilibrate more slowly. A variant of this hypothesis is that the pools for elongation and de novo fatty acid synthesis that use exogenous acetate directly are both small, but that acetate can also be used to sustain de novo fatty acid synthesis via an indirect metabolic pathway that slowly reaches steady-state labeling. In all cases, 18:1 synthesized at early time points would be of a lower specific radioactivity compared with later time points, and thus the acid-ester-to-acid ratio would decrease with time. A corollary of this hypothesis is that the synthesis of total [¹⁴C]18:1 would show a lag at early time points with respect to elongation, and with a time scale similar to that of the change in differential labeling within the very-long-chain fatty acids. Figure ​Figure3B3B shows the time-dependent accumulation of label in total 18:1 derivatives (18:1 plus 18:1 portion of 20:1 and 22:1; 18:2 and 18:3 contributed less than 2% of the total label in these short-term incubations and were not included). The accumulation of [¹⁴C]18:1 was essentially linear, with no lag phase. Furthermore, the ratio of label in total 18:1 to the label in the elongation portion of 22:1 plus cis-11-20:1 remained approximately constant (data not shown), indicating that the first general mechanism, acetate-accessible pool sizes, was not responsible for the changes in ¹⁴C ratios of de novo to elongation carbons (Table ​(TableII).

The second general explanation for the changing ratios of label in the de novo and elongation carbons considers that another 18:1 source in addition to the newly synthesized [¹⁴C]18:1 was used as a substrate to synthesize the very-long-chain fatty acids. As shown in Figure ​Figure3A,3A, the radioactivity accumulated in the de novo (18:1) portion of 22:1 lagged significantly compared with that of total 18:1. This result supports the concept that there is an endogenous source of 18:1 for the synthesis of 22:1 in addition to the newly synthesized [¹⁴C]18:1. If only newly synthesized 18:1 was used directly for elongation, the ¹⁴C ratios of C13AE/C9A and C11AE/C9A (Table ​(TableI)I) and the percentage of total labeled 18:1 in 22:1 would remain constant with time. The contribution of newly synthesized 18:1 is gauged by the extrapolation of the curves of Figure ​Figure3A3A to 0 time. At 0 time the amount of the total labeled 18:1 produced, which contributes to erucic acid biosynthesis, was only 20% in the light (32% in the dark), whereas the theoretical maximum was 55%. Thus, 36% (light) and 58% (dark) of the 18:1 flux through the chain-elongation system to 22:1 was used directly from newly synthesized (¹⁴C-labeled) 18:1 in the plastid, since at 0 time any large intermediate pools of 18:1 had yet to fill.

Inhibition of the Homodimeric ACCase Blocks Erucic Acid Production

The synthesis of erucic acid from 18:1 has been demonstrated in vitro to require malonyl-CoA, and it has been assumed that this malonyl-CoA is produced by the cytosolic homodimeric ACCase. This assumption has never been tested. To evaluate this hypothesis in vivo we have used the herbicide haloxyfop, which specifically inhibits the homodimeric acetyl-CoA carboxylase (Burton et al., 1987, 1991), and examined how it influences the synthesis of erucic acid and other lipid species. In the experiment shown in Figure ​Figure44 embryos were pretreated with different concentrations of haloxyfop for 30 min and then incubated with 2 mm [¹⁴C]acetate in the light for 1 h. Incorporation of radioactivity into 22:1 decreased with increased concentrations of haloxyfop. In contrast, radioactivity in 18:1 increased with haloxyfop concentrations lower than 100 μm. At a concentration of 50 μm haloxyfop, synthesis of 22:1 was inhibited by 70%, but ¹⁴C accumulation in 18:1 increased almost 2-fold. These results demonstrate that the elongation of 18:1 to 22:1 is dependent on homodimeric ACCase to supply malonyl-CoA, and that haloxyfop can inhibit the elongation without inhibition of de novo synthesis of 18:1. Also plotted in Figure ​Figure44 are the accumulations of ¹⁴C in PC and TAG. In response to the addition of haloxyfop, radioactivities in 22:1 and TAG were inhibited in parallel, whereas the accumulation of radioactivities in PC and 18:1 showed parallel patterns of increase.

Open in a separate window

Figure 4

Effect of increasing concentrations of haloxyfop on fatty acid and lipid synthesis. B. rapa embryos were pretreated with the indicated concentrations of herbicide for 30 min, then incubated for 1 h in the light with the addition of [¹⁴C]acetate. Radioactivity of lipid and fatty acid are quantified on the basis of dpm per embryo.

DISCUSSION

The radioactivity incorporated into 18:1 of embryos incubated in the light was approximately 2-fold higher than that in the dark. These results extend the observations of Browse and Slack (1985), who, in a study using isolated plastids from linseed seeds (green embryos) and safflower seeds (white embryos), concluded that linseed plastids are photosynthetically active and provide a source of ATP and NAD(P)H for fatty acid synthesis. Similarly, Eastmond et al. (1996) and Asokanthan et al. (1997) recently concluded that although photosynthesis is unlikely to provide substantial net carbon for B. napus seed anabolism, light-driven electron transport may provide ATP and reducing equivalents for storage-product synthesis. Taken together with the data in Figure ​Figure2,2, these studies suggest that cofactor supply and/or plastidial acetyl-CoA carboxylase regulation by light rather than carbon precursors may be a major limiting factor in the rate of fatty acid accumulation in Brassica embryos. This hypothesis is consistent with the observation that acetyl-ACP and malonyl-ACP pools are relatively high at all stages of active fatty acid synthesis in developing spinach seeds and castor endosperm (Post-Beittenmiller et al., 1991, 1992a).

The strong inhibition of acetate incorporation into 22:1 by haloxyfop, a known noncompetitive inhibitor of the cytosolic homodimeric acetyl-CoA carboxylase, confirms the hypothesis that the acetate that is used for chain elongation is in the cytosolic compartment and is distinct from the acetate used for de novo fatty acid synthesis. In all dicots examined the cytosolic acetyl-CoA carboxylase is the homodimeric form, whereas the plastid acetyl-CoA carboxylase is the heteromeric form (Konishi and Sasaki, 1994). However, a homodimeric form of ACCase has also recently been reported in B. napus plastids (Roesler et al., 1997; Schulte et al., 1997). Although the function of this ACCase is unclear at this time, it appears to be less abundant than the heteromeric form. Furthermore, if this plastid homomeric ACCase is as sensitive to haloxyfop as the cytosolic form, our observation that a 50 μm concentration of this herbicide caused no inhibition of de novo fatty acid synthesis while strongly inhibiting elongation suggests that the homomeric ACCase in the plastid has a quantitatively minor role in de novo fatty acid synthesis.

As presented in Table ​TableII and Figure ​Figure3,3, we observed differential labeling of the elongation and de novo carbons within the very-long-chain fatty acids with time. We considered two explanations for these changes. The explanation that different-sized acetate-accessible pools supplied malonyl-CoA for de novo fatty acid synthesis and for chain elongation was ruled out based on kinetic and pool-size considerations. This is not surprising if acetate is being used directly for acetyl-CoA synthesis in both plastids and the cytosol. Data are not available for oilseeds, but in leaf tissue the chloroplast pool of acetyl-CoA, the direct substrate for the synthesis of malonyl-CoA, ranges from 10 to 20 μm, and it can be equilibrated with exogenously supplied acetate within several seconds (Post-Beittenmiller et al., 1992b; Roughan, 1997). The chloroplast pool of acetyl-CoA dominates the total leaf acetyl-CoA pool, and, indeed total leaf CoA, so the cytosolic pool of acetyl-CoA is also expected to be small. Extrapolating to seed tissue, even if the cytosolic pool of acetyl-CoA is depleted at a rate of about one-tenth of that of the plastid pool (as calculated from the mass composition of the oil) it is expected that the direct utilization of exogenous acetate by either pool will reach steady state very quickly.

Because endogenous free acetate pools have not been measured in developing seeds, it is not possible to estimate the endogenous contribution of free acetate to fatty acid synthesis in seeds. However, utilization of exogenous acetate clearly reaches a steady-state rate very quickly (Fig. ​(Fig.3B),3B), within 1 min. Acyl-CoA and acyl-ACP pools might also contribute to a lag in labeling. In Brassica embryos, and in developing seeds in general, there is limited information on acyl-ACP levels and a dearth of information on acyl-CoA levels. We can make some extrapolations from the situation in leaves and from the limited seed data. In chloroplasts isolated from leaf tissues (Soll and Roughan, 1982; Roughan and Nishida, 1990), it was estimated that acyl-ACP half-lives were approximately 10 s. In seeds ACP levels of approximately 1 μg/g fresh weight were noted (Hannapel and Ohlrogge, 1988), whereas in spinach seeds and leaves up to 60% of the ACP is in the free form. Similarly, long-chain acyl-CoA pools in Cuphea lutea (Singh et al., 1986) and developing B. napus (V. Eccleston and J. Ohlrogge, unpublished observations) indicate levels of <20 μm. The B. rapa seeds in the present study accumulated about 16 nmol of fatty acid h⁻¹ seed⁻¹. Using these numbers, the acyl-CoA-pool and acyl-ACP-pool turnover time is calculated to be less than 1 min. Clearly, accumulation of 18:1 in the acyl-thioester pools cannot explain any lag in the labeling kinetics of 22:1. This indicates that the second general explanation, that another 18:1 source in addition to the newly synthesized [¹⁴C]18:1 is used as substrate to synthesize the very-long-chain fatty acids, is the correct one.

Three distinct scenarios can be envisaged for the supply of 18:1 for chain elongation to cis-11-20:1 and 22:1 in developing oilseeds. In the first model, oleoyl moieties are exported from the plastid, activated to oleoyl-CoA, and can immediately become substrates for elongation. In this scenario the newly synthesized oleoyl groups can be considered channeled directly to the cytosolic elongation system. In the second model, the newly exported oleoyl-CoA is rapidly equilibrated with the bulk cytosolic pool of oleoyl-CoA. The acyl-exchange mechanism first reported by Stymne and Stobart (1984) can be envisaged to dilute the [¹⁴C]oleoyl-CoA pool with oleoyl groups from the sn-2 position of PC, whereas [¹⁴C]18:1 will enter the sn-2 position of PC and be diluted. The bulk pool of oleoyl-CoA is then available for chain elongation. The third hypothesis is based on the observations of Hlousek-Radojcic et al. (1995) and requires that newly synthesized [¹⁴C]18:1 acylate an acceptor lipid, as yet unidentified, and then be transferred directly or indirectly from this lipid to the elongase. An analysis of the kinetics of differential labeling within the acyl chain of very-long-chain fatty acids suggests that an endogenous 18:1 pool contributes to the biosynthesis of 22:1, although the labeling itself cannot be used to distinguish between models two and three. Our results cannot rule out model one or other combinations of these models as a contributing pathway until the details of the endogenous pool can be demonstrated and quantified. However, extrapolating Figure ​Figure3A3A to 0 time indicates that only 20% of total 18:1 synthesis in light (or 32% in the dark) is directly elongated to 22:1. Comparison of this with the in vivo conversion of >55% of 18:1 to 22:1 (Fig. ​(Fig.1)1) suggests that at least one-half of the 18:1 enters an intermediate pool before elongation to 22:1.

What is the endogenous pool that can accept newly synthesized 18:1 from the plastid and also provide 18:1 as a substrate for elongation? A logical approach to obtaining information on such intermediate pools is to perform pulse-chase experiments. However, with intact B. rapa embryos we found that it was not possible to remove the [¹⁴C]acetate sequestered inside the embryos, which continued to sustain fatty acid synthesis and hence confound interpretation. In the absence of pulse-chase data, we decided to take an indirect approach in which we inhibited the elongation from 18:1 to 22:1 with haloxyfop without reducing the de novo synthesis of 18:1; we then monitored where the extra 18:1 accumulated. As shown in Figure ​Figure4, 4, there was an increasing amount of 18:1 accumulated in PC when 22:1 synthesis was inhibited, indirectly supporting the concept that the 18:1 esterified to PC could be a source of oleoyl moieties for the synthesis of 22:1. Consistent with this observation, Hlousek-Radojcic et al. (1995) found that when [¹⁴C]oleoyl-CoA was incubated with B. napus oil bodies, more than 50% of radioactivity was found in PC.

In a study of petroselinic acid biosynthesis in developing coriander and carrot endosperm, Cahoon and Ohlrogge (1994) concluded that PC was an intermediate in the movement of petroselinic acid from its site of biosynthesis in the plastid into TAG. Similar observations have recently been made for the accumulation of 16:1Δ6 in developing seeds of Thunbergia (D. Shultz, E.B. Cahoon, and J. Ohlrogge, unpublished data). Because neither of these unusual fatty acids is synthesized or further modified on PC, a rationale for their movement through a PC pool before incorporation into TAG is not immediately obvious. The present study on erucic acid biosynthesis adds another example in which PC may be an “intermediate” in the flux of fatty acids into TAG, even though no metabolism of the fatty acid may be directly associated with the PC. The observation of a large flux of fatty acids through PC without modification in three diverse oilseed species may simply indicate that acyl exchange between the acyl-CoA pool and PC is very rapid. The low accumulation of unusual fatty acids in PC may further reflect specific mechanisms for their removal (e.g. phospholipases). However, it is interesting to speculate that PC may play some more general role in TAG assembly, perhaps as a carrier of acyl chains toward a subcellular site of TAG assembly. This role might be analogous to the major flux of acyl chains through PC in leaves, followed by their movement from the ER to the chloroplast.

Because of its commercial value, several attempts have been made to increase erucic acid content in transgenic plants. However, the factors that limit erucic acid content in Cruciferae species are still largely unknown. Expression of a sn-2 acyltransferase from Limnanthes in transgenic B. napus led to the accumulation of erucic acid in the sn-2 position, but no increase in total 22:1 content of the oil (Lassner et al., 1995). Similarly, overexpression of elongases has resulted in increased chain length but not an increase in mol % of very-long-chain fatty acids (Lassner et al., 1996). The recent report that a mutated yeast SLC1 gene expressed in Arabidopsis or B. napus gave increased levels of erucic acid (Zou et al., 1997) is at this time difficult to interpret in light of the results with the Limnanthes enzyme. Reduction of 18:1 desaturation by mutation of the oleoyl desaturase might be expected to increase 18:1 availability for elongation. However, in a mutant of Arabidopsis that is deficient in desaturation of 18:1 (Lemieux et al., 1990), 18:2 and 18:3 content of seeds decreased from 53% to 8.7%, 18:1 content increased from 15.4% to 53.5%, but 20:1 and 22:1 content just slightly increased from 20.2% to 26.7%. Likewise, elimination of the elongation of 18:1 might be expected to increase 18:1 availability for desaturation. However, neither low-erucic-acid lines of Brassica (Daun, 1983) nor the fae1 mutant of Arabidopsis (Kunst et al., 1992) exhibit corresponding increases in 18:1 desaturation. Although several interpretations are possible, one hypothesis consistent with these results is that the pathways for 18:1 elongation and desaturation may draw on different pools of 18:1. In summary, our results, together with those of Hlousek-Radojcic et al. (1995) and those cited above, suggest that the pathway for erucic acid biosynthesis may be more complex than originally thought. Furthermore, the flux of 18:1 through distinct intermediate lipid pools before elongation or desaturation may be one factor that limits the availability of 18:1 for elongation.

ACKNOWLEDGMENT

Abbreviation:

Footnotes

LITERATURE  CITED