The Utilization of the Acyl-CoA and the Involvement PDAT and DGAT in the Biosynthesis of Erucic Acid-Rich Triacylglycerols in Crambe Seed Oil
The triacylglycerol of Crambe abyssinica seeds consist of 95 % very long chain (>18 carbon) fatty acids (86 % erucic acid; 22:1∆13) in the sn-1 and sn-3 positions. This would suggest that C. abyssinica triacylglycerols are not formed by the action of the phospholipid:diacylglycerol acyltransferase (PDAT), but are rather the results of acyl-CoA:diacylglycerol acyltransferase (DGAT) activity. However, measurements of PDAT and DGAT activities in microsomal membranes showed that C. abyssinica has significant PDAT activity, corresponding to about 10 % of the DGAT activity during periods of rapid seed oil accumulation. The specific activity of DGAT for erucoyl-CoA had doubled at 19 days after flowering compared to earlier developmental stages, and was, at that stage, the preferred acyl donor, whereas the activities for 16:0-CoA and 18:1-CoA remained constant. This indicates that an expression of an isoform of DGAT with high specificity for erucoyl-CoA is induced at the onset of rapid erucic acid and oil accumulation in the C. abyssinica seeds. Analysis of the composition of the acyl-CoA pool during different stages of seed development showed that the percentage of erucoyl groups in acyl-CoA was much higher than in complex lipids at all stages of seed development except in the desiccation phase. These results are in accordance with published results showing that the rate limiting step in erucic acid accumulation in C. abyssinica oil is the utilization of erucoyl-CoA by the acyltransferases in the glycerol-3-phosphate pathway.
KeywordsCrambe PDAT DGAT Microsomal preparation Triacylglycerol Erucic acid Lipids
18:1n-9, Oleic acid
18:2n-6, Linoleic acid
18:3n-3, Alpha-linolenic acid
22:1n-9, Erucic acid
Lysophosphatidic acid acyltransferase
Fatty acid elongase
Days after flowering
Materials and Methods
Seeds of C. abyssinica, cv. Mayer were planted in peat-based soil and transferred to a growth chamber with 60 % relative humidity and a 16 h photoperiod (200 μmol radiation; a day temperature of 21 °C; a night temperature of 18 °C). After a few weeks, the plants started flowering. At specific stages following flowering, the developing fruits were harvested and seeds were separated manually from the surrounding siliqua. Based on the days after flowering (DAF) and morphological criteria, the harvested seeds were classified into one of the six stages of seed development, prior to further analyses (Supplement Table S1).
Radioactive fatty acids were obtained from Biotrend, Cologne, Germany. Di-18:1-DAG, di-6:0-DAG were purchased from Sigma and unlabelled fatty acids supplied by Larodan Fine Chemicals (Malmö, Sweden). The [14C]acyl-CoAs and acyl-CoAs were synthesized according to modified method described by Sanchez et al. . Radioactive DAG (rac-sn-1,2-di-[14C]18:1-DAG) was obtained by partial lipase (Rhizopus arrhizus; Sigma-Aldrich, St. Louis, MO, USA) treatment of tri-[14C]18:1-TAG (Perkin Elmer, Waltham, MA, USA) followed by separation of the obtained lipid products using thin layer chromatography (TLC), elution from the gel and determination of concentration by analyzing the fatty acid content of aliquots as methyl esters on GC with methyl-heptadecanoic acid added as an internal standard as described below.
Microsomal Preparation and Enzyme Assays
Microsomal membranes were prepared from freshly harvested seeds (separated from the surrounding siliqua). The seeds coats were removed manually and microsomes were prepared according to the method previously described  and stored at −80 °C until used for assays. DGAT activity was measured in assays with two different acceptors of fatty acids: di-6:0-DAG (only DGAT assays) and sn-1-18:1-sn-2-[14C]18:1-DAG (PDAT and DGAT + PDAT assays). In assays with di-6:0-DAG, 5 nmol [14C]acyl-CoA ([14C]16:0-CoA, [14C]18:1-CoA or [14C]22:1-CoA) together with 5 nmol di-6:0-DAG were added with incubation buffer to the microsomal membranes (9 μg of microsomal protein, which was equivalent to approximately 2 nmol of microsomal PtdCho) with incubation buffer (0,05 M HEPES, pH 7.2; 5 mM MgCl2; 1 mg BSA/ml) in a final volume 100 μl and incubated for 30 min at 30 °C with shaking (1,250 rpm). In the case of PDAT and DGAT + PDAT assays, [14C]18:1-DAG was dissolved in 19 μl of benzene and added to aliquots of microsomal fractions lyophilized overnight (corresponding to 22 μg of microsomal protein). After immediate evaporation of the solvent, buffer (0.05 M HEPES—pH 7.2; 5 mM MgCl2, 1 mg BSA/ml) was added and, where measuring combined DGAT + PDAT activities 5 nmol acyl-CoA was included. The assays (final volume 100 μl) were incubated for 30 min at 30 °C with shaking (1,250 rpm). In assays with [14C]18:1-DAG + acyl-CoA, formation of [14C]TAG was regarded as resulting from both DGAT and PDAT activity. Formation of [14C]TAG in assays with only [14C]18:1-DAG added was regarded as only PDAT activity. DGAT activity was calculated as amount of [14C]TAG in assays with [14C]DAG + acyl-CoA minus the amount of [14C]TAG in assays with [14C]DAG only. However, the standard deviation between triplicate samples in those assays was too high to achieve reliable calculated DGAT activity. Therefore only results from PDAT activity are presented.
At the end of incubation, lipids were extracted from the reaction mixtures into chloroform according to the method of Bligh and Dyer  and separated on TLC (silica gel 60 plates; Merck, Darmstadt, Germany) in hexane:diethyl ether:acetic acid (70:30:1 by volume). Radioactive TAG (TAG with two 6:0 moieties clearly separated on TLC from TAG with only long chain fatty acids), products of PDAT and DGAT activity, were visualized and quantified on the plate using electronic autoradiography (Instant Imager, Packard instruments). All assays were repeated independently from three to six times and mean values are presented in the figures.
Seeds were homogenized in chloroform:methanol:0.15 M acetic acid (1:2:0.8) using a Potter–Elvehjem homogenizer and the lipids were subsequently extracted into chloroform according to Bligh and Dyer . For total lipids analysis, aliquots of the chloroform phase were evaporated and methylated as described below. Individual lipids in the chloroform phase were separated by TLC in hexane:diethyl ether:acetic acid (70:30:1) for neutral lipids or in chloroform:methanol:acetic acid:water (85:15:10:3.5) for separation of polar lipids. Gel, from areas corresponding to the various lipids (identified by means of authentic standards), was removed and lipids were methylated in situ on the gel with 2 % H2SO4 in dry methanol (60 min at 90 °C). The methyl esters were extracted with hexane and analyzed by GLC equipped with a flame ionization detector and a WCOT fused-silica 50 m × 0.32 mm ID coating CP-Wax 58-CB DF 5 0.2 capillary column (Chrompack International, Middleburg, The Netherlands) with methyl-heptadecanoic acid added as an internal standard.
Freshly harvested seeds from different stages of development were frozen in liquid nitrogen and stored at −80 °C until further analyses. The samples were extracted according to Larson and Graham , and then analyzed using electrospray ionization tandem mass spectrometry (multi reaction monitoring) or LC–MS/MS MRM in positive ion mode. The LC–MS/MS + MRM analysis (AB4000 QTRAP) followed the methods described by Haynes et al.  (Agilent 1200 LC system; Gemini C18 column, 2-mm inner diameter, 150 mm with 5-mm particles). For the purpose of identification and calibration, standard acyl-CoA esters with acyl chain lengths from C14 to C20 were purchased from Sigma as free acids or lithium salts. The experiments were repeated twice independently.
Analyses of the acyl-CoA profile in Crambe seed during seed development revealed that the proportion of 22:1-CoA was much higher than that of the 22:1 acid found in lipids (mainly TAG). This indicates that 22:1-CoA is poorly used by the acylation enzymes compared to other acyl-CoAs. When acyl-CoA profiling was done in developing seed of Cuphea hookeriana, which accumulates 25 % of capric acid (10:0) in its oil, the percentage of this fatty acid in the acyl-CoA fraction was significantly lower than found in the TAG, indicating that Cuphea acyl transferases have high specificity for this acyl group . On the other hand, transgenic rape seed producing 10:0 fatty acids by expression of C. hookeriana medium chain thioesterase had a significantly higher percentage of 10:0 in the acyl-CoA pool than in complex lipids , probably due to the poor affinity for 10:0-CoA by the rape seed acyl transferases. Therefore, the amount of 22:1 in Crambe seed TAG might be limited by a poor capacity to acylate this acyl group into the glycerol backbone, thereby causing a build-up of 22:1-CoA and subsequently a product inhibition of the elongation of 18:1-CoA into 22:1-CoA. The amount of 22:1 is very low in the sn-2 position of TAG  and it has been shown that the Crambe lysosphosphatidic acid acyltransferase (LPAAT), the enzyme responsible for the incorporation of acyl groups at the sn-2 position, has very low activity with 22:1-CoA . Limnanthes douglasii, which accumulate over 90 % of very long chain fatty acids in its seed TAG, has an LPAAT that has been shown to have good activity with 22:1-CoA [10, 17]. Co-expression of a rape seed elongase (FAE1) and a L. douglasii LPAAT in transgenic Crambe seeds led to a significant increase in 22:1 acid in TAG . Expression of the same multigene construct in high 22:1 rape seed did not increase the 22:1 acid in TAG significantly compared to overexpression of FAE1 alone, but rather led to a re-distribution of 22:1 acid between the outer and middle positions of TAG . However, when the transgenic 22:1 rape overexpressing the rape FAE1 and L. douglasii LPAT was crossed with a mutant rape seed low in polyunsaturated fatty acids, a substantial increase in the proportion of 22:1 acid in TAG could be observed in the subsequent generations . This indicates that acyltransferase activities for 22:1-CoA was limiting in Crambe whereas in rape, it was the availability of 18:1-CoA substrate for elongation. However, further increase in 22:1 acid in TAG in Crambe could be achieved if conversion of oleic acid to linoleic acid was inhibited by a FAD2-RNAi combined with the expression of FAE1 and L. douglasii LPAAT , demonstrating that when acyltransferase activity was not limiting, the availability of 18:1-CoA substrate for the FAE became the rate limiting step.
Our assays of DGAT activity in microsomal membranes prepared from seeds at different stages of development suggest that an isoform of DGAT with high specificity for 22:1-CoA is induced at around 19DAF since the specific activity of DGAT with 22:1-CoA at that time point was substantially increased compared to earlier stages of development, whereas the specific activity for 16:0-CoA and 18:1-CoA remained essentially unchanged. This higher activity for 22:1-CoA coincided with a period of rapid TAG and 22:1 acid accumulation.
Despite the fact that the outer positions of TAG in Crambe have over 95 % of very long chain fatty acids , significant PDAT activity could be measured in membrane preparations from developing seeds. Since PDAT transfers fatty acids from mainly the sn-2 position of PtdCho to the sn-3 position of DAG in the formation of TAG [7, 8], it is highly unlikely that PDAT in Crambe transfers any very long chain fatty acids to TAG. No complete stereospecific analysis of Crambe seed TAG has been done, so it is not known how the 5 % of C16 and C18 fatty acids in sn-1 + sn-3 are distributed between these positions. The specific activities of PDAT in microsomal membranes from Crambe reported here was 50 % of that reported in sunflower membranes and about 20 % of that found in safflower membranes . A similar comparison of the highest DGAT activities (achieved in Crambe with 22:1-CoA as the acyl donor from membranes prepared from seed at 19 DAF) was 20 and 40 % of that found in sunflower and safflower membranes, respectively . Although these comparisons suggest that Crambe PDAT plays a minor role in TAG synthesis compared to both sunflower and safflower, the actual contributions of these enzymes in TAG synthesis might be rather different from the in vitro assays. A caveat for in vitro assays of these enzymes regarding the dilution of added substrates with endogenous substrate is discussed in some detail by Banaś et al. . This point is clearly demonstrated in our work with the different use of added and endogenous DAG substrates in DGAT assays done on Crambe membranes prepared at different developmental stages. Nevertheless, our assays suggest that PDAT could be a significant contributor to the maximal 10 % of C16 and C18 fatty acids that could reside in the sn-3 position of TAG in Crambe seeds.
This research was financially supported by the European commission FP7 project ICON, Vinnova, VR and FORMAS.
- 1.Sonntag NOV (1991) Erucic, behenic: feedstocks of 21th century. Informatics 5:449–463Google Scholar
- 2.Leonard C (1994) Sources and commercial applications of high erucic vegetable oil. Lipid Tech 4:79–83Google Scholar
- 3.Pizza GJ, Foglia TA (2001) Rapeseed oil for oleochemical uses. Eur J Lipid Sci Technol 103:405–454Google Scholar
- 8.Banaś W, Sanchez Garcia A, Banaś A, Stymne S (2013) Activities of acyl-CoA:diacylglycerol acyltransferase (DGAT) and phospholipid:diacylglycerol acyltransferase (PDAT) in microsomal preparation of developing sunflower and safflower seeds. Planta 237:1627–1636PubMedCentralPubMedCrossRefGoogle Scholar
- 9.Pan X, Siloto RM, Wickramarathna AD, Mietkiewska E, Weselake RJ (2013) Identification of a pair of phospholipid:diacylglycerol acyltransferases from developing flax (Linum usitatissimum L.) seed catalyzing the selective production of trilinolenin. J Biol Chem 288:24173–24188PubMedCrossRefGoogle Scholar
- 13.Blight EG, Dyer WJ (1959) A rapid method of total lipid extraction and purification. Can J Med Sci 37:911–917Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.