Acyl CoA profiles of transgenic plants that accumulate medium-chain fatty acids indicate inefficient storage lipid synthesis in developing oilseeds

Authors

  • Tony R. Larson,

    1. Centre for Novel Agricultural Products, Department of Biology, University of York, PO Box 373, York YO10 5YW, UK, and
    Search for more papers by this author
  • Teresa Edgell,

    1. Centre for Novel Agricultural Products, Department of Biology, University of York, PO Box 373, York YO10 5YW, UK, and
    Search for more papers by this author
  • James Byrne,

    1. Calgene LLC, 1920 Fifth Street, Davis, CA 95616, USA
    Search for more papers by this author
  • Katayoon Dehesh,

    1. Calgene LLC, 1920 Fifth Street, Davis, CA 95616, USA
    Search for more papers by this author
    • #

      Current address: Section of Plant Biology, University of California–Davis, One Shields Avenue, Davis, CA 95616-5270, USA.

  • Ian A. Graham

    Corresponding author
    1. Centre for Novel Agricultural Products, Department of Biology, University of York, PO Box 373, York YO10 5YW, UK, and
      * For correspondence (Tel. +44 1904 328 750; fax +44 1904 328 762; e-mail: iag1@york.ac.uk).
    Search for more papers by this author

* For correspondence (Tel. +44 1904 328 750; fax +44 1904 328 762; e-mail: iag1@york.ac.uk).

Summary

Several Brassica napus lines transformed with genes responsible for the synthesis of medium- or long-chain fatty acids were examined to determine limiting factor(s) for the subsequent accumulation of these fatty acids in seed lipids. Examination of a decanoic acid (10:0) accumulating line revealed a disproportionately high concentration of 10:0 CoA during seed development compared to long-chain acyl CoAs isolated from the same tissues, suggesting that poor incorporation of 10:0 CoA into seed lipids limits 10:0 fatty acid accumulation. This relationship was also seen for dodecanoyl (12:0) CoA and fatty acid in a high 12:0 line, but not for octadecanoic (18:0) CoA and fatty acid in a high 18:0 line. Comparison of 10:0 CoA and fatty acid proportions from seeds at different developmental stages for transgenic B. napus and Cuphea hookeriana, the source plant for the medium-chain thioesterase and 3-ketoacyl-ACP synthase transgenes, revealed that C. hookeriana incorporates 10:0 CoA into seed lipids more efficiently than transgenic B. napus. Furthermore, β-oxidation and glyoxylate cycle activities were not increased above wild type levels during seed development in the 8:0/10:0 line, suggesting that lipid catabolism was not being induced in response to the elevated 10:0 CoA concentrations. Taken together, these data suggest that transgenic plants that are engineered to synthesize medium-chain fatty acids may lack the necessary mechanisms, such as specific acyltransferases, to incorporate these fatty acids efficiently into seed lipids.

Introduction

Controlling the composition and maximizing the yield of native or designer oils in crop plants are important goals for plant breeders and the biotechnology industry. In order to realize these goals, genetic engineering has been used as a tool in attempts to increase yields and to produce economically viable levels of novel fatty acids in the seed oils of crop species such as Brassica napus. One common approach to manipulate lipid synthesis involves identifying a key fatty acid biosynthetic gene in a non-crop species and then expressing this gene during seed development in a model or crop species. This approach has been used to produce medium-chain fatty acids (MCFA) in the seed oils of B. napus by expressing novel medium-chain acyl–acyl carrier protein (ACP) thioesterases alone or together with the medium-chain specific 3-ketoacyl-ACP synthase (KAS IV) in the embryo plastids during seed development (Dehesh et al., 1996, 1998; Voelker et al., 1992).

However, the accumulation of novel fatty acids in the seed oils of crop species requires the concerted action of additional biosynthetic mechanisms, and possibly the repression of catabolism, to maximize final yields. Following plastidic synthesis by the fatty acid synthesis (FAS) enzymes, acyl ACPs are transformed to free fatty acids and then acyl CoAs at or near the outer plastid envelope so the acyl CoAs will be available for further cytosolic metabolism (Pollard and Ohlrogge, 1999). Acyl CoAs are the key intermediates for ER-mediated elongation, desaturation, and transferase reactions, whereby the acyl chain is modified and then attached to a glycerol backbone. Particularly important for seed lipid synthesis is the activity of the Kennedy pathway acyltransferase enzymes, which are responsible for synthesizing storage triacylglycerols (TAGs) from acyl CoAs and glycerol. As reviewed by Voelker and Kinney (2001), TAG synthesis proceeds in oilseed species by the sequential acylation of glycerol by Kennedy pathway enzymes in the endoplasmic reticulum. The first enzyme of the pathway, glycerol-3-phosphate acyltransferase (GPAT) is not selective so that a variety of acyl chain lengths from acyl CoA substrates can be added to the sn-1 position of glycerol-3-phosphate to form lysophosphatidic acid (LPA). However, LPA acyltransferase (LPAAT), which transfers a fatty acid from acyl CoA to the sn-2 position of LPA to form phosphatidic acid (PA), prefers unsaturated long-chain substrates. Diacylglycerol (DAG), formed from PA by the removal of phosphate by PA phosphatase, then acts as a substrate together with acyl CoA for DAG acyltransferase (DAGAT) to acylate the sn-3 position to form TAG. Rapid acyl exchange between lipids and the acyl CoA pool have also been shown to operate as alternative mechanisms to the Kennedy pathway for acyl transfer into TAG. For example, TAG may be formed via a phosphatidylcholine (PC) intermediate. In this pathway, DAG is esterified at the sn-3 position with choline by the action of choline phosphotransferase (CPT) to form PC, from which the acyl chain at the sn-2 position may be transferred to DAG to form TAG. The enzyme responsible for this process has been termed phospholipid:diacylglycerol acyltansferase (PDAT), and is thought to be particularly important for stabilizing membranes by editing unusual fatty acids out of PC and into TAG (Millar et al., 2000).

There is evidence that the constituent enzymes of the Kennedy pathway and acyl exchange mechanisms are specific for different acyl chain lengths (see review by Voelker and Kinney, 2001). Thus, the modification or addition of one biosynthetic step alone may not lead to increased production of the desired fatty acid in the seed lipids.

Catabolism of unusual fatty acids, or their biosynthetic intermediates, may also occur in transgenic plants. This results in a ‘futile cycle’, whereby newly synthesized acyl chains are intercepted by catabolic processes, such as peroxisomal β-oxidation, before they can be incorporated into seed lipids. Evidence for this comes from transgenic B. napus that produces 12:0 fatty acid, where β-oxidation activity appears to be induced during seed maturation (Eccleston and Ohlrogge, 1998).

In this study, we use a recently described sensitive method (Larson and Graham, 2001) to profile acyl CoAs and lipid fatty acids extracted from several transgenic B. napus lines engineered to produce the novel medium-chain fatty acids, octanoic acid (8:0), decanoic acid (10:0) and dodecanoic acid (12:0), or to overproduce the native octadecanoic acid (18:0). We compare these data with results from the high 8:0 (50%) and 10:0 (25%) acid producing plant, Cuphea hookeriana, which is the source of the medium-chain thioesterase and β-ketoacyl-ACP synthase cDNAs used to transform B. napus. We present evidence to suggest that the yields of medium-chain fatty acids in transgenic plants is limited by poor incorporation of acyl CoAs into seed lipids.

Results

In this study, five different sets of transgenic B. napus seeds accumulating null, medium and high levels of 8:0 and 10:0 acids (designated 8:0/10:0 line), or high levels of 12:0 acid (12:0 line), or high levels of 18:0 acid (18:0 line) were employed. All transgenic lines were classified according to the average fatty acid content determined from pools of dry seeds. To obtain MCFA containing seeds, three transgenic lines with levels varying from null (MC30083-2-DH765-1), to medium (MC30083-2-DH170-2) with 2.2 and 20.4 mol% of 8:0 and 10:0, respectively, and high (MC30083-2-DH47-10) with 7.1 and 31 mol% of 8:0 and 10:0, respectively, were selected. The 12:0 producing seeds were derived from the LA30056-5-DH63-14-1-2 line and contained 57 mol% of 12:0 in their seed oil. The 18:0 producing seeds were from the 5266-SP30021-79-6-DH52-4-9 line and contained 21 mol% of 18:0. Siliques from wild type and transgenic plants were tagged, and developing seeds harvested at 20, 30, and 40 days post-anthesis (DPA).

Fatty acid catabolism in high 8:0/10:0 acid B. napus

The seeds of null, medium and high 8:0/10:0 lines were assayed for ACOX (Figure 1a–c) and ICL activity (data not shown). In all the lines examined short-, medium-, and long-chain ACOX activity increased during seed development, reaching a maximum at 40 DPA. The higher ACOX activity seen with shorter acyl CoA substrates is a generally observed phenomenon in a variety of B. napus and Arabidopsis thaliana tissues (Eccleston and Ohlrogge, 1998; Hooks et al., 1999). Relative to the null transformant, medium-chain ACOX activity was not increased during seed development in the medium or high expressing lines. This was despite there being 16 and 22 mol% 10:0 acid and 24 and 47 mol% 10:0 CoA in the medium and high expressing lines, respectively, by 40 DPA (Figure 2). ICL activity was present at zero or barely detectable levels in all three lines, with slight increases in maturing seeds (data not shown). However, similar to the ACOX results, ICL was not increased in the medium or high expressing lines.

Figure 1.

β-oxidation activity in transgenic 8:0/10:0 B. napus seeds that accumulate 8:0 and 10:0 acids. Acyl CoA oxidase (ACOX) activity was assayed with short (6:0), medium (12:0), and long (16:0) acyl CoA substrates to obtain a measure of β-oxidation activity in (a) null, (b) medium and (c) high expressor mutants at 20, 30 and 40 DPA. Results are shown as means ± SD for two ACOX assays.

Figure 2.

Developmental profile of acyl CoA esters and lipid fatty acids in transgenic 8:0/10:0 B. napus seeds that accumulate 8:0 and 10:0 acid. Acyl CoAs and lipid fatty acids were profiled in 20, 30, and 40 DPA seeds collected from WT (a–c), null transformant (d–f), medium MCFA accumulating (g–i), and high MCFA accumulating (j–l) B. napus lines at 20, 30, and 40 DPA. Results are expressed as molar percents ± SD of the total acyl CoA or lipid fatty acid pool quantified for n = 20 acyl CoA determinations and n = 6 lipid fatty acid determinations.

Together, these data show that the concentrations of medium-chain acyl CoAs or lipid fatty acids in the transgenic B. napus lines did not induce activities of marker enzymes for β-oxidation or the glyoxylate cycle above background levels. Thus, it appears that any 10:0 acid or CoA catabolism occurring in these lines was handled by existing catabolic capacity.

Changes in acyl CoA and lipid fatty acid composition during seed development

Acyl CoA and lipid fatty acid composition was monitored at 20, 30, and 40 DPA seeds in WT and transgenic null, medium, and high 8:0/10:0 B. napus lines (Figure 2). In all lines, octadecadienoic (18:2Δ9,12) acid was the major lipid component early in seed development, whilst octadecenoic (18:1Δ9) became the major lipid component by 40 DPA. In contrast, arachidoyl (20:0) CoA was a major component of the acyl CoA pool early in seed development, with 18:1 CoA becoming a major component by 40 DPA. As expected, the WT and null transformant lines showed no differences in either lipid fatty acid or acyl CoA composition during seed development. However, uniquely for the medium and high 8:0/10:0 lines, 10:0 and, to a lesser extent, octanoyl (8:0) CoAs appeared at the earliest stage of seed development and generally increased to maximum proportions by 40 DPA. Similarly, 8:0 and 10:0 acids appeared in these lines, but only at detectable levels by 30 and 40 DPA.

The appearance of 10:0 CoA and fatty acid in the medium and high expressor lines led to a markedly different acyl CoA and lipid fatty acid profile by 40 DPA compared to the WT and null expressor lines (Figure 2). Altered profiles were also observed for B. napus lines that were transformed to produce 12:0 acid or overproduce 18:0 acid (Figure 3). In the 12:0 acid line, 61 mol% 12:0 CoA and 55 mol% 12:0 acid was present by 40 DPA (Figure 3b). These results were somewhat more extreme than the 47 mol% 10:0 CoA and 22 mol% 10:0 acid measured in the high 8:0/10:0 line at 40 DPA (Figure 2i). In the line that overproduces 18:0 acid, 8.2 mol% 18:0 CoA and 20 mol% 18:0 acid had accumulated by 40 DPA (Figure 3d). Thus, in this line, despite relatively low 18:0 CoA concentrations, 18:0 acid had reached a similar proportion of the lipid fatty acid pool as 10:0 acid had in the high 8:0/10:0 line by 40 DPA.

Figure 3.

Comparison of acyl CoA esters and lipid fatty acids in transgenic B. napus seeds that accumulate 12:0 or 18:0 acids. Acyl CoAs and lipid fatty acids were profiled in 40 DPA seeds collected from WT (a); grown at the same time as a high 12:0 acid accumulating line (b); and a second WT (c); grown at the same time as a high 18:0 acid accumulating line (d). Results are expressed as molar percents ±SD of the total acyl CoA or lipid fatty acid pool quantified for n = 10 acyl CoA and lipid fatty acid determinations.

The total quantitative acyl CoA and lipid fatty acid pool sizes, on a fresh weight basis, for all these lines are shown for 40 DPA seeds in Figure 4. In both the 8:0/10:0 and 12:0 lines, the total acyl CoA pool was significantly higher than the pool size in their respective WT seeds (Figure 4a). However, the total acyl CoA pool size was significantly lower in the 18:0 line than its respective WT. In contrast, whilst the acyl CoA pool sizes in the 8:0/10:0 and 12:0 lines were elevated, the total lipid fatty acid pool size was significantly decreased relative to the WT line by 40 DPA (Figure 4b). There was no difference in fatty acid content between the transgenic and WT 18:0 line seeds. These results suggest that a high proportion of medium-chain CoAs in the cytosol, or medium-chain fatty acid in the lipids, have compromised the overall lipid accumulation process.

Figure 4.

Total acyl CoA ester and lipid fatty acid pool sizes in transgenic B. napus seeds. Total acyl CoA (a) and lipid fatty acid (b) concentrations from high 8:0/10:0, 12:0, and 18:0 lines, and their respective WT lines grown alongside, were calculated for 40 DPA seeds on a fresh weight basis. Significant differences between the WT and transgenic lines, where P < 0.05 in a two-tailed Student's t-test, are indicated by an asterisk. Results are expressed as means ± SE for the n-values given in Figures 2 and 3.

Differences in 10:0 acid accumulation in transgenic B. napus and C. hookeriana

Acyl CoAs and lipid fatty acids were profiled at different developmental stages from individual transgenic B. napus seeds and seeds from the thioesterase (Ch FatB2) and 3-ketoacyl-ACP synthase (Ch KAS IV) transgene donor species, C. hookeriana, a species with up to 50% 8:0 and 25% 10:0 acids in its seed oil. Due to the low levels of 8:0 acid in the high 8:0/10:0 transgenic line (7 mol%) as compared with the levels accumulated in maturing seeds of C. hookeriana (approximately 50 mol%), and similar levels of 10:0 acid (22 and 25 mol%, respectively), the following analyses compared the accumulation of 10:0 CoA and acid only.

The data are plotted as mol% 10:0 CoA versus mol% 10:0 acid, to show the relationship between the acyl CoA and fatty acid content for individual seeds (Figure 5). This figure shows the considerable biological variation in acyl CoA and fatty acid mol% values when measurements are made from single seeds harvested at timed developmental stages. This variation is caused by slight differences in actual developmental stages between individual seeds. Variability was greatest for C. hookeriana, where seeds harvested from 14 DPA capsules (mid-stage seed development; seeds are fully mature at 24 DPA) contained between approximately 0 and 40 mol% 10:0 acid, with the average being 23 mol% for all 149 seeds. This approaches the value of 25 mol% cited as the average value for dry seeds in this species (Dehesh et al., 1996). Variability was also seen for the B. napus seeds, where the spread was between 11 and 30 mol% (with an average of 29 mol%) 10:0 acid in the 10 individual 40 DPA seeds used to generate Figure 5. This spread of the data lies within the average value of 22 mol% 10:0 acid reported in Figure 2 for a different batch of 40 DPA high 8:0/10:0 seeds, and the value of 31 mol% measured in pools of dry seeds during selection of this line.

Figure 5.

Molar ratios of 10:0 CoA and fatty acid in individual C. hookeriana and transgenic B. napus seeds. Acyl CoAs and lipid fatty acids were profiled from 20, 30, and 40 DPA transgenic high 8:0/10:0 acid B. napus seeds and maturing C. hookeriana seeds at 14 DPA. Acyl CoAs and lipid fatty acids were profiled from the same seeds so that molar percent data could be matched. Results are shown as 10:0 acid (molar percent of the total lipid fatty acid pool) as a function of 10:0 CoA (molar percent of the total acyl CoA pool) for each individual seed. In order to obtain a good spread of data over several developmental stages, 10 B. napus seeds were profiled for each stage (total = 30), and 149 C. hookeriana seeds were profiled. The inherent variability of individual seeds taken from one developmental stage, even in tagged plants and especially for C. hookeriana, aided in the construction of the figure.

For C. hookeriana, very low mol% values for 10:0 CoA were consistently paired with high mol% values for 10:0 acid in the seed lipids. In contrast, for the transgenic B. napus seeds, 10:0 CoA mol% values were almost always higher than the corresponding 10:0 acid mol% values from the same seeds. These data show that, although both species have the same thioesterase and 3-ketoacyl-ACP synthase enzymes, and produce 10:0 CoA, it is incorporated much less efficiently into 10:0 acid in the transgenic B. napus line.

Discussion

MCFA synthesis and medium-chain acyl CoA accumulation

B. napus lines examined in this study were engineered to synthesize MCFAs, namely 8:0 and 10:0, by the introduction of medium-chain acyl ACP thioesterase together with the medium-chain specific 3-ketoacyl-ACP synthase (KAS IV) in the embryo plastids during seed development (Dehesh et al., 1998). The 12:0 producing seeds were from plants transformed with California Bay thioesterase (BTE) (Voelker et al., 1992). The 18:0 producing seeds were obtained from B. napus plants overexpressing GarmFatA1 thioesterase (Hawkins and Kridl, 1998). Therefore, all transgenic lines examined in this study were similarly engineered to overproduce specific fatty acids by modifications in plastidial fatty acid synthesis. In this study we measured in vivo the size and composition of the acyl CoA pool in developing WT and transgenic B. napus seeds. Effectively, this is a measure of the net efficiency of acyl intermediate transfer from the plastid to storage lipid bodies in the cytosol during seed development. This is the first time that these important metabolic intermediates have been profiled during seed development and oil deposition for any plant species.

The acyl CoA pool size is effectively the sum of acyl CoAs produced by plastidial fatty acid synthesis, transacylation reactions during lipid remodelling, and peroxisomal acyl CoAs destined for β-oxidation. One mechanism that determines the final amount and composition of the triacylglycerols (TAGs) accumulated during oilseed development is the rate and chain length specificity of plastidial fatty acid synthesis (Bao and Ohlrogge, 1999). However, the relationship between the synthetic rate of an unusual fatty acid and its accumulation in TAG is not necessarily linear. Voelker et al. (1996) reported that, in B. napus transformed with BTE, the relationship between BTE activity and 12:0 acid levels in mature seeds was linear to 35 mol% 12:0 acid, but then further increases in BTE activity did not lead to proportional increases in 12:0 in the lipids. Incorporation of MCFA into TAG can be further increased by mutating the 3-ketoacyl–acyl carrier protein synthase (KASIII), so that it is insensitive to inhibition by its medium-chain acyl ACP products in vitro (Abbadi et al., 2000). Nevertheless, our results rule out the possibility that plastidial fatty acid synthesis is limiting MCFA accumulation in TAG. This is because we demonstrated an accumulation of 10:0 and 12:0 CoAs over and above the concentrations of the usual long-chain acyl CoA complement in transgenic B. napus lines expressing Ch FatB2 and Ch KASIV, or BTE, respectively. This indicates that although MCFA accumulation can be induced in the seed lipids by the introduction of novel thioesterase and 3-ketoacyl-ACP synthase (KAS IV) activity, the accumulation in TAG is constricted downstream from fatty acid synthesis.

TAG synthesis from acyl CoA substrates

In B. napus, DAGAT has the lowest activity of the Kennedy pathway enzymes, and the DAG pool increases during lipid accumulation in developing seeds (Perry et al., 1999). Together with results from recent DAGAT knockout complementation work in A. thaliana (Jako et al., 2001), this suggests that DAGAT is an important flux control point during TAG accumulation. However, an in vitro labelling study where labelled 12:0 acid was fed to seed homogenates, showed a transient accumulation of 12:0 CoA in species that do not accumulate MCFAs in their seed lipids, although sufficient DAG appeared to be present for acylation (Battey and Ohlrogge, 1989). Furthermore, the ready appearance of MCFAs in the sn-3 position of TAGs reported for the high 8:0/10:0 B. napus line in a previous study (Wiberg et al., 2000) suggests that DAGAT activity is not limiting TAG synthesis. Rather, rapid editing out of MCFAs from the PC pool without transfer to TAGs during seed maturation for high 8:0/10:0 or high 12:0 B. napus has been reported (Wiberg et al., 1997, 2000). This suggests that acyl exchange mechanisms, which would function to transfer MCFAs from phospholipids and DAG to TAG, may not be operating efficiently in transgenic B. napus. A similar case has been described for transgenic A. thaliana, where poor PDAT activity is responsible for the accumulation of engineered acetylenic, epoxy, and hydroxy fatty acids in PC rather than TAG, whereas the wild species that produce these unusual fatty acids use PDAT to transfer the fatty acids from PC to TAG (Thomaeus et al., 2001).

Knutzon et al. (1999) found that both BTE or BTE + LPAAT transformed plants showed slightly lower oil yields (on a developing seed dry weight basis) at total seed 12:0 acid concentrations above 50 mol%. This is in agreement with our results for the high 12:0 B. napus lines, where 55 mol% 12:0 acid accumulated by 40 DPA but the total amount of lipid fatty acids was decreased on a fresh weight basis. Interestingly, we also noted a decrease in total lipid fatty acid yield in the 8:0/10:0 line, although only 22 mol% 10:0 acid accumulated by 40 DPA (Figure 4b). However, both the 8:0/10:0 and 12:0 lines had elevated 10:0 and 12:0 CoA proportions (47 and 61 mol%, respectively) by 40 DPA, and similarly sized total acyl CoA pools, increased above their respective wild type levels to values of approximately 11–12 pmol mg−1 fresh weight (Figure 4a). In contrast, although the 18:0 CoA and acid proportion increased in the 18:0 line, there was a slight decrease in the total acyl CoA pool relative to the wild type (Figure 4a) with no apparent decrease in total lipid fatty acid yield (Figure 4b). These observations suggest that an increase of medium-chain acyl CoAs in the total cytosolic acyl CoA pool, and/or an increase in the acyl CoA pool size, may initiate feedback mechanisms or catabolism that limit the transfer of acyl CoAs into TAG. Alternatively, if a particular acyl CoA is inefficiently incorporated into TAG or other lipids, the acyl CoA will accumulate at the expense of other chain lengths available for TAG synthesis, with consequent decreases in yield. The results from the high 18:0 line and C. hookeriana both indicate that where there is endogenous ability to incorporate all acyl CoAs into TAG, acyl CoAs do not accumulate and TAG synthesis proceeds efficiently.

β-oxidation as a mechanism to remove excess acyl CoAs

It is likely that medium-chain acyl CoAs accumulated in the cytosol, if not incorporated into TAG, will eventually be catabolized. Using a bacterial polyhydroxyalkanoate (PHA) synthase transformed into A. thaliana and targeted to peroxisomes to monitor carbon flux into β-oxidation, Poirier et al. (1999) found that plants co-expressing a plastidial acyl carrier protein thioesterase from Cuphea lanceolata accumulated 10:0 monomers in the PHA pool, indicating β-oxidation of 10:0 CoA. These authors also showed that when the bacterial PHA synthase is expressed in an A. thaliana mutant deficient in DAGAT, PHAs accumulated to higher levels than in the wild type controls, suggesting that β-oxidation degrades substrates not incorporated into TAG. Therefore, although 10:0 CoA accumulated to 47 mol% in our B. napus lines transformed with Ch FatB2 and Ch KAS IV, the lack of observed β-oxidation induction does not mean that catabolism was not occurring. Endogenous β-oxidation activity was likely to be sufficient to degrade any medium-chain CoAs, as has been demonstrated for transgenic A. thaliana leaves expressing an MCTE (Hooks et al., 1999). In contrast, higher levels of medium-chain acyl CoA or fatty acid accumulation may induce β-oxidation. In the BTE-transformed lines, where we showed that 12:0 CoA and fatty acid reached approximately 60 mol%, Eccleston and Ohlrogge (1998) demonstrated an increase in the medium-chain acyl ACP pool, an induction of β-oxidation, and the flux of fed acetate into sucrose and malate rather than lipids.

Other, as yet uncharacterized catabolic pathways may also be induced, such as the ER-bound acyl CoA thioesterases recently described in A. thaliana (Tilton et al., 2000). The formation of cytosolic acyl CoA binding protein (ACBP)–acyl CoA complexes may also regulate the transport and/or entry of acyl CoAs to either lipid synthesis or β-oxidation. In rat heptoma cells overexpressing ACBP, hexadecanoic (16:0) acid was preferentially diverted away from β-oxidation and towards lipid synthesis. (Yang et al., 2001). Similarly, in developing B. napus seeds, 18:1 CoA was found to stimulate GPAT activity when incubated in vitro with low concentrations of recombinant ACBP (Brown et al., 1998), and recombinant A. thaliana ACBP was found to protect acyl CoAs against hydrolysis in safflower microsomes (Engeseth et al., 1996). However, ACBP has a poor affinity for acyl CoAs with chain lengths shorter than C14 (Rosendal et al., 1993). This may further predispose medium-chain acyl CoAs to catabolism in transgenic plants unless the pool size can be kept very small by efficient lipid synthesis.

Together, these data suggest that an ongoing background catabolism of medium-chain fatty acids may be occurring, with induction of β-oxidation and futile cycling when individual acyl CoAs accumulate above a threshold concentration. To avoid a futile cycle of fatty acids, it is then desirable that the acyl CoA pool size is decreased by increasing the efficiency of incorporation into lipid.

Conclusion

In summary, regardless of the amounts of 10:0 and 12:0 fatty acids accumulating in TAG during seed development in the transgenic lines, the results show that 10:0 and 12:0 acyl CoAs are not efficiently utilized for lipid synthesis. If this were so, their concentrations would be low and similar to the native long-chain acyl CoAs at all stages of development. This was not the case for 10:0 CoA in C. hookeriana, or the high 18:0 line, where a 12-fold increase in 18:0 fatty acid in the lipids was achieved by only a two-fold increase in 18:0 CoA. Therefore, it appears that medium-chain CoAs are being excluded from lipids during seed development in transgenic B. napus. Efforts to further increase yields of medium-chain fatty acids in transgenic plants should then centre on the efficient delivery and utilization of the acyl CoAs by the Kennedy pathway enzymes and/or lipid exchange mechanisms. Monitoring changes in the acyl CoA pool composition and concentration will be important tools in genetic engineering programmes to assess and optimize the metabolic regulation of novel oil synthesis in plants.

Experimental procedures

Plant material

In this study, five different sets of transgenic B. napus seeds accumulating null, medium and high levels of 8:0 and 8:0/10:0, or high levels of 12:0, or high levels of 18:0, as the result of seed-specific expression of different cDNAs under the control of napin promoter (Kridl et al., 1991), were employed. B. napus cv. Quantum (a low erucic acid variety) plants were used as the wild-type (WT) background throughout the study. To obtain seeds containing different levels of 8:0 and 10:0 acids, an F2 population of the dihaploid lines originally generated from crosses between pCGN 4804, a B. napus plant overexpressing Ch FatB2, a C. hookeriana 8:0 and 10:0 specific thioesterase (Dehesh, 1996), and pCGN 5401-LP004-9, a B. napus plant overexpressing Ch KAS IV, a medium-chain specific 3-ketoacyl-ACP synthase cloned from C. hookeriana (Dehesh et al., 1998) were employed. Three lines were taken forward and classified according to the average fatty acid content in pools of dry seeds. These were designated as null (MC30083-2-DH765-1), medium (MC30083-2-DH170-2), with 2.2 and 20.36 mol% of 8:0 and 10:0, respectively, and high (MC30083-2-DH47-10), with 7.1 and 31 mol% 8:0 and 10:0, respectively. The 12:0 acid producing line (LA30056-5-DH63-14-1-2), with 57 mol% 12:0 in the seed oil, was selected from plants (3828) transformed with BTE, a 12:0 ACP thioesterase from California Bay (Voelker et al., 1992). The18:0 acid producing line (5266-SP30021-79–6-DH52-4-9), with 21 mol% 18:0,was obtained from B. napus plants overexpressing GarmFatA1, an 18:0/18:1 thioesterase cloned from MangoSteen (Hawkins and Kridl, 1998).

Plants were grown in the greenhouse and tagged at anthesis. Seeds were harvested at 20, 30, and 40 days post-anthesis (DPA), and stored at −80°C until further analysis. C. hookeriana plants were also grown in the greenhouse as described previously (Dehesh et al., 1996); capsules were tagged and the seeds were harvested at 14 DPA.

Enzyme assays

Acyl CoA oxidase (ACOX) assays were performed on seed material according to Froman et al. (2000). ACOX assays were performed using 6:0, 12:0, and 16:0 CoA substrates to determine short-, medium-, and long-chain activities, respectively. Isocitrate lyase (ICL) assays were performed according to Cooper and Beevers (1969).

Acyl CoA analysis

Acyl CoAs were extracted and measured from single seeds, using the same equipment and methods described in Larson and Graham (2001). Where C. hookeriana and B. napus were compared (Figure 5), data was obtained using a longer HPLC runtime with modified mobile phase composition to better resolve peaks from contaminants present in C. hookeriana. The mobile phases were as follows: A, 0.25% (v/v) triethylamine (Aldrich, UK) in water; B, 90% (v/v) acetonitrile (Rathburn, UK) in water; C, 90% acetonitrile (v/v) + 1% (v/v) acetic acid (Sigma, UK); D, 1% acetic acid (v/v) in water. A linear quaternary gradient program was run as follows: 0 min flow = 0.75 ml min−1, 10% C and 90% D; 0–5 min, to 80% C and 20% D; 5–5.1 min; to 80% A and 20% D; 5.1–7 min, to 97% A and 3% B; 7–10 min, to 95% A and 5% B; 10–10.1 min, flow decreased to 0.5 ml min−1; 10.1–50 min, to 60% A and 40% B; 50.1–52 min, to 100% B; 52–52.1 min, flow increased to 0.75 ml min−1; 52.1–55 min, maintain 100% B; 55–55.1 min, to 10% C and 90% D; 55.1–60 min, maintain 10% C and 90% D. Sample injection volumes of 20 µl were made on a LUNA phenyl-hexyl, 150 mm × 2.00 mm 5 µm particle size column (Phenomenex, UK) maintained at 40°C.

Lipid fatty acid analysis

Single seeds were extracted and the lipid fatty acids transmethylated to fatty acid methyl esters (FAMEs) for gas chromatography (GC) analysis in the one-step procedure described by Larson and Graham (2001). To reduce the losses of volatile medium-chain FAMEs during sample preparation, no sample drying or concentration step was used following transmethylation and prior to injection on the GC. Where the relationship between molar acyl CoA and fatty acid ratios were compared for single seeds (i.e. in Figure 5), lipids were transmethylated and analyzed from the petroleum ether phase of the acyl CoA extracts. Lipid fatty acid data obtained in this way gave identical results to data obtained by one-step extraction and transmethylation of single seeds (data not shown).

Acknowledgments

Financial support for this study was provided to the Centre for Novel Agricultural Products by a grant from the Department for Environment, Food, and Rural Affairs (UK).

Ancillary