Metabolic control analysis reveals an important role for diacylglycerol acyltransferase in olive but not in oil palm lipid accumulation


J. L. Harwood, School of Biosciences, Cardiff University, Cardiff CF10 3US, UK
Fax: +44 29 2087 4117
Tel. +44 29 2087 4108


We applied metabolic control analysis to the Kennedy pathway for triacylglycerol formation in tissue cultures from the important oil crops, olive (Olea europaea L.) and oil palm (Elaeis guineensis Jacq.). When microsomal fractions were incubated at 30 °C rather than 20 °C, there was an increase in triacylglycerol labelling. This increase was accompanied by a build up of diacylglycerol (DAG) radioactivity in olive but not in oil palm, suggesting that the activity of DAG acyltransferase (DAGAT) was becoming limiting in olive. We used 2-bromooctanoate as a specific inhibitor of DAGAT and showed that the enzyme had a flux control coefficient under the experimental conditions of 0.74 in olive but only 0.12 in oil palm. These data revealed important differences in the regulation of lipid biosynthesis in cultures from different plants and suggest that changes in the endogenous activity of DAGAT is unlikely to affect oil accumulation in oil palm crops.


bottom-up control analysis




diacylglycerol acyltransferase


metabolic control analysis


quantitative trait loci



Oil-producing plants are among the more important agricultural crops. Worldwide, ≈ 340 million metric tonnes of oil are produced annually [1] with an emphasis on edible oils, but with an increasing production of chemical feedstocks and speciality fats [2]. Despite considerable knowledge of the biochemistry and molecular biology of the metabolic pathways involved [3], their regulation and control are still largely unknown [4,5].

Recent efforts to find out more about the control of lipid synthesis in oil crops have used different approaches. Quantitative trait loci that control seed oil and fatty acid composition have been examined in several important crops such as Brassica napus[6], as well as in the experimental model plant Arabidopsis[7]. In addition, cDNA microarrays comparing wild-type with the low-lipid wrinkled 1 mutation of Arabidopsis have examined over 3500 genes in order to see which may exert important control [8].

We used metabolic control analysis (MCA) to gain more information about the regulation of lipid accumulation. MCA, pioneered by Kacser and Burns [9] and Heinrich and Rapoport [10] allows a quantitative measure of the control exerted by individual reactions (or groups of reactions) in a metabolic pathway under defined conditions. More details of the method and its application can be found in Fell [11]. We applied top-down MCA to lipid formation in tissue cultures from two important oil crops, olive and oil palm [12,13]. These experiments revealed that lipid synthesis was shared between fatty acid formation and lipid assembly, but with the former exerting greater control (≈ 60%). This conclusion was consistent with indirect data using developing seeds of Cuphea[14] and other maturing embryos [15].

Radiolabelling experiments had previously suggested that, within lipid assembly by the Kennedy pathway [16] (Fig. 1), diacylglycerol acyltransferase (DAGAT) appeared to exert strong flux control in some crops [17]. The data were supported by experiments in Arabidopsis in which mutants with a decreased seed oil content had lowered DAGAT activity [18] and this TAG1 mutant was shown to have a mutation in the DAGAT gene [19]. Furthermore, there was an increase in oil content when an Arabidopsis DAGAT was overexpressed in Arabidopsis seeds [20].

Figure 1.

The Kennedy pathway for TAG biosynthesis with DAGAT catalysing the final step. Two alternative methods of forming TAG are also shown. These are catalysed by DAG/DAG acyltransferase and phospholipid/DAG acyltransferase (PDAT), respectively. Both have been demonstrated in plants [26], but neither seems quantitatively important in oil palm or olive calli (see text).

The availability of an inhibitor specific for DAGAT allowed us to determine the flux control coefficient for this enzyme in olive and oil palm. The results show that there are significant differences between the two tissues, such that DAGAT exerts strong control over flux in the Kennedy pathway in olive but not in oil palm under the experimental conditions.

Results and Discussion

Using temperature to promote triacylglycerol synthesis

In some plants, we had noticed previously that there were indications that DAGAT could exert significant flux control, especially when the rate of triacylglycerol (TAG) synthesis was high. Thus, for example, when lipid accumulation in developing oilseed rape embryos was maximal there was noticeable diacylglycerol (DAG) accumulation [17,21] indicating a constraint in flux at the level of its conversion to TAG. Previous experiments in our laboratory had shown that it was possible to enhance TAG formation several-fold by increasing the incubation temperature from 20 to 30 °C with both types of callus cultures [12]. Therefore, in order to see if there were indications that DAGAT could exert strong flux control in either olive or oil palm cultures, we carried out incubations at the two temperatures.

In Table 1, the results show that, for both tissues, there was an approximate doubling of total incorporation in raising the incubation temperature from 20 to 30 °C. This would be expected given a Q10 of two for most reactions, in the absence of any protein instability/denaturation effects. In addition to the increase in total labelling, there was also an increase in the proportion of radioactivity in TAG for both tissues. Thus, for oil palm the percentage increased from six to 10, meaning that total labelling was increased threefold. For olive microsomes, the percentage labelling of TAG increased from seven to 12, so that the total radiolabel in this lipid went up by 3.8-fold. In short, the synthesis of storage TAG had been enhanced considerably at 30 °C in both microsomal preparations.

Table 1.  Effect of incubation temperature on the incorporation of radioactivity from [14C]oleoyl-CoA into total lipids and Kennedy pathway intermediates in microsomal fractions from oil palm or olive calli. Means ± SD (three experiments in triplicate) are shown. LysoPtdOH, lysophosphatidate; PtdOH, phosphatidate; DAG, diacylglycerol; TAG, triacylglycerol. * P < 0.05.
SpeciesTemp. (°C)Total labelling (nmol·min−1·mg−1 prot.)Lipid class labelling (% total)
Oil palm201.7 ± 0.137 ± 630 ± 527 ± 46 ± 1
303.1 ± 0.7*29 ± 532 ± 229 ± 210 ± 1*
Olive202.4 ± 0.434 ± 516 ± 443 ± 37 ± 2
305.3 ± 0.6*23 ± 1*12 ± 253 ± 3*12 ± 1*

Increased lipid synthesis has variable effects on DAG accumulation

However, when the pattern of products was analysed there was a marked difference in the effect of temperature on olive and oil palm microsomes. For oil palm, the increase in TAG percentage labelling was not accompanied by any significant changes in the proportions of Kennedy pathway intermediates (Table 1). Thus, the enhanced synthesis of storage lipid was not accompanied by the build-up of any intermediate, including DAG. By contrast, the increased flux through the Kennedy pathway in olive was accompanied by a significant decrease in the percentage labelling of 1-monoacylglycerol 3-phosphate (lysophosphatidate) and, interestingly, a build-up of radioactivity in DAG. Thus, under conditions of enhanced flux through the Kennedy pathway, DAGAT activity appeared to exert significant flux control in olive, but not oil palm, microsomes. It is also perhaps noteworthy that, under the labelling conditions (which were linear with time), proportional labelling of DAG was significantly greater in olive than in oil palm. Again, this suggested that DAGAT could exert significant flux control in olive.

Using bromooctanoate as an inhibitor of DAGAT

To evaluate the control exerted by DAGAT we used 2-bromooctanoate as an inhibitor. Mayorek and Bar-Tana [22] first reported the use of this compound as a specific inhibitor of DAGAT in rat hepatocytes and it has also been used to inhibit the enzyme from Brassica napus[23]. However, bromo-fatty acids have been suggested to have more general effects on membrane-located enzymes [24] so we carried out experiments at 30 °C to evaluate 2-bromopalmitate as well as 2-bromooctanoate.

The effect of increasing concentrations of 2-bromooctanoate on the labelling of Kennedy pathway intermediates in oil palm and olive microsomes is shown in Fig. 2. At low 2-bromooctanoate concentrations (up to 0.1 mm in oil palm and 0.5 mm in olive microsomes) the only effect was to reduce the proportion of TAG labelling and increase the proportion of DAG labelling – indicating inhibition of DAGAT. By contrast, 2-bromopalmitate caused a general decrease in lipid labelling without any change in the pattern of products (data not shown). This result was consistent with the reported effect of 2-bromopalmitate on glycerol 3-phosphate acyltransferase [25]. By contrast, 2-bromooctanoate, when used at the low concentrations mentioned above, had no effect on any enzyme in the Kennedy pathway other than DAGAT when assayed directly (data not shown). We also checked for effects on phospholipid/DAGAT or DAG/DAGAT (Fig. 1), because each of these activities could, potentially, give rise to TAG [26]. The activities of these two enzymes were barely detectable in microsomal preparations from oil palm or olive calli (< 5% DAGAT activity) and no effect of bromooctanoate (0.1 mm in oil palm, 0.5 mm in olive) could be detected (data not shown). The data for 2-bromooctanoate also agree with the specific nature of its inhibition (at low concentrations) on DAGAT in intact callus cultures [27].

Figure 2.

Effect of 2-bromooctanoate on the pattern of labelling of Kennedy pathway intermediates from [U-14C]glycerol 3-phosphate by microsomal preparations from oil palm (A) or olive (B) calli. Results are presented as means ± SD (three experiments, each in triplicate). Statistical significance in comparison with no bromooctanoate controls by Student's t-test (*P < 0.05). For abbreviations see Table 1.

Estimation of flux control coefficient for DAGAT

The effect of increasing concentrations of 2-bromooctanoate on DAGAT activity and the total flux through the Kennedy pathway is shown in Fig. 3. These experiments were performed at 30 °C, which is an environmentally relevant temperature for both plant species in vivo. At low concentrations of 2-bromooctanoate, which inhibited only DAGAT, there was no significant effect on total flux in oil palm microsomes even though DAGAT showed linear inhibition in the concentration range used. In contrast, for olive microsomes the progressive inhibition of DAGAT with increasing concentrations of 2-bromooctanoate was accompanied by a similar reduction in total flux (Fig. 3).

Figure 3.

Effect of 2-bromooctanoate on total flux through the Kennedy pathway and on DAGAT activity in microsomal preparations from oil palm (A) or olive (B) calli. Results are presented as means ± SD (three experiments, each in triplicate). Typical control values for DAGAT activity at 30 °C were 5–7 nmol·min−1·mg−1 protein for oil palm and 4–6 nmol·min−1·mg−1 protein for olive microsomes.

Because no effect of 2-bromooctanoate on glycerol 3-phosphate acyltransferase (or enzymes other than DAGAT) was observed at 0.03 mm, we conclude that the reduction in total flux in olive (Fig. 3) was due to the accumulation of DAG. As noted previously (Table 1), the accumulation of label in DAG from [14C]oleoyl-CoA was much higher in olive than oil palm. This difference was even more pronounced when labelling from [14C]glycerol 3-phosphate was measured. From the latter precursor, DAG represented 25% of the total label which was increased to 37% with bromooctanoate. For oil palm calli, DAG was only 11% of the total label from [14C]glycerol, increasing to 19% with the highest concentrations of bromooctanoate (Fig. 2). Therefore, any feedback effect due to DAG accumulation would be more likely in olive than in oil palm.

The slopes of tangents to the initial part of inhibition curves provide sufficient information for the calculation of control coefficients via a method justified previously [28,29]. We used this technique to estimate the numerical values of flux control coefficients for DAGAT over Kennedy pathway flux after fitting curves to the inhibition data and calculating the gradients of the tangents to the initial part of the curves using a simfit computer package [30]. The values (Table 2) for DAGAT in oil palm and in olive microsomes were 0.12 and 0.74, respectively.

Table 2.  Gradient of tangent estimations for the calculation of flux control coefficients of DAGAT over Kennedy pathway flux.
SpeciesGradient of tangent (total flux)Gradient of tangent (DAGAT)Flux control coefficient
Oil palm−36.3−1092.00.12

Thus, for oil palm only ≈ 10% of the control for the four enzymes of the Kennedy pathway lies at the DAGAT step under our experimental conditions. Without measuring individual flux control coefficients for the other enzymes of the pathway it is not possible to say where most of the control lies. However, based on the data with temperature manipulation of [12] or oleate addition to [13] callus cultures, we conclude that glycerol 3-phosphate acyltransferase may be important. This conclusion could also be supported by the data in Table 1, where a doubling of carbon flux to TAG did not produce any significant changes to the proportional labelling of the pathway intermediates. Recently, evidence has shown [31] that the levels of glycerol 3-phosphate and, hence, the activity of glycerol 3-phosphate acyltransferase may limit TAG formation in developing oilseed rape seeds. Thus, for some oil crops, this step in the Kennedy pathway may exert significant control.

In contrast, the value of 0.74 for the flux control coefficient for olive DAGAT suggests that this enzyme exerts strong control over carbon flux through the Kennedy pathway. This conclusion agrees with the build-up of DAG, which is seen whenever flux to TAG is increased in intact olive calli [12,13] or when using microsomal fractions (Table 1).


The notable difference in flux control coefficients for DAGAT from the two tissues highlights the potential difficulty in extrapolating data to other crop species. In addition, callus cultures are not the same as whole plants and our findings may not extrapolate directly to crops. Nevertheless, we have used environmentally relevant temperatures in studying cultures that accumulate TAG at rates and of a composition similar to developing fruits [12]. Therefore, we believe that our data should be considered seriously in relation to the regulation of oil synthesis in olive and oil palm crops. Thus, one could predict that increasing DAGAT activity in olive might be of benefit in facilitating additional TAG accumulation (i.e. oil yields). By analogy with previous data obtained using oilseed rape [17] one could also suggest that similar benefits might also be shown with this crop. Indeed, genetic manipulation of Arabidopsis for altered DAGAT activity has confirmed this prediction [18,20]. By contrast, the low control exerted by DAGAT activity over lipid accumulation in oil palm suggests that breeding or genetic manipulation to increase the activity of the enzyme in this crop might not be worthwhile.

Experimental procedures


Oil palm and olive callus cultures were established and maintained on modified Murashige and Skoog [32] medium as described previously [12]. They were subcultured every 28 days and used for experiments 20–25 days after subculturing.

Sodium [1-14C]acetate (specific radioactivity 2.11 GBq·mmol−1), [U-14C]glycerol (specific radioactivity 5.5 GBq·mmol−1), 1,2 di[1-14C]oleoylphosphatidylcholine (specific radioactivity 3.7 GBq·mmol−1), [1-14C]oleoyl-CoA (specific radioactivity 2.04 GBq·mmol−1) and [U-14C]glycerol 3-phosphate (specific radioactivity 5.44 GBq·mmol−1) were purchased from Amersham International (Amersham, UK). TLC was performed on silica gel G plates (Merck Ltd, Lutterworth, UK). Fatty acid and phospholipid standards were obtained from Nu-Chek (Elysian, MN) and Sigma (Poole, UK), respectively. All other enzymes, chemicals and solvents used in lipid extraction, modification and analysis were from Sigma or BDH (Poole, UK). The best available analytical grades were used.

Radiolabelling studies with cultures

Callus cultures were selected for uniformity of mass and morphological appearance. Fresh masses before incubation were recorded and results normally expressed per gram of fresh weight. Oil palm or olive callus cultures (0.5–1.0 g fresh weight) were preincubated with inhibitor at the stated concentration in 100 mm sorbitol (0.2 mL·g−1 callus) for 1 h. The calli were then rinsed briefly with 100 mm sorbitol and incubated with 1 µCi of [1-14C]acetate or [U-14C]glycerol for up to 8 h [13]. After incubation, the calli were rinsed briefly with 100 mm sorbitol and then inactivated by heating in 1.25 mL propan-2-ol for 30min at 70 °C and quantitative extraction was carried out as described by Garbus et al. [33] and modified for plant tissues [34].

Microsomal preparation and incubation

Microsomal fractions were prepared from calli by homogenizing at 4 °C in a buffer of 50 mm Hepes (pH 7.2), 330 mm sorbitol, 1 mm MgCl2, 3 mm EDTA, 5 mmβ-mercaptoethanol, 0.1% bovine serum albumin, 0.2% ascorbate and 1% polyvinylpyrollidine using a domestic blender and a buffer to tissue ratio (ml/g) of 10. After filtration through four layers of miracloth, the homogenate was centrifuged at 5000 g for 10 min and the supernatant at 18 000 g for 20 min. A final spin at 105 000 g for 60 min yielded a microsomal pellet and the particle-free supernatant. The microsomal pellet was re-suspended in 50 mm Hepes (pH 7.2), 330 mm sorbitol, 1 mm dithiothreitol using a precooled glass homogeniser.

Microsomes (up to 0.1 mg protein) were incubated with [U-14C]glycerol 3-phosphate (200 µm, 0.1 µCi) in a medium containing 35 mm Hepes/NaOH (pH 7.2), 300 mm sorbitol, 0.5% bovine serum albumin, 75 µm palmitoyl-CoA and 100 µm oleoyl-CoA for up to 30 min at 30 °C. For incubations with [1-14C]oleoyl-CoA, the same incubation components were used except that the labelled substrate was replaced. Incorporation of radioactivity from [1-14C]oleoyl-CoA was linear for a minimum of 4 min but most incubations were for 2 min. The concentrations of components had been checked to be optimal and the reactions were proportional to protein and linear with time over the period of incubation. Reactions were terminated and products analysed as described previously [12].

Enzyme assays

DAGAT was assayed as described previously [21]. Reaction conditions were again determined to be optimal for both olive and oil palm callus microsomes.

Phospholipid/DAGAT (PDAT) and DAG/DAGAT activities (Fig. 1) were examined using methodology essentially as described previously [26]. Incubations of microsomal fractions were carried out with di[1-14C]oleoylphosphatidylcholine or with di[1-14C]oleoylglycerol, (generated by phospholipase C (Bacillus cereus) digestion of the former substrate and TLC separation), for the above two activities, respectively.

Other analytical procedures

Radioactive samples were measured in a LKB Wallac 1209 Rackbeta counter (Wallac Oy, Turku, Finland) using Opti-Fluor scintillation fluid (Packard Bioscience, Groningen, the Netherlands). Quench corrections were made by the external standard channels ratio method.

Protein was measured by the Bradford method [35].

Bottom-up control analysis

Application of bottom-up control analysis (BUCA) to DAGAT was made using microsomal fractions from oil palm or olive calli as the model system. DAGAT activity and carbon flux from [U-14C]glycerol 3-phosphate through the Kennedy pathway to TAG in the absence or presence of the specific inhibitor 2-bromooctanoate were measured. From these data we plotted inline image and inline image against inline image and used the simfit package [30] to determine tangents to each of the curves at zero inhibitor as described previously [29]. The ratio of these two tangents gives the flux control coefficient for DAGAT over the Kennedy pathway flux as described by:


where inline image is the flux control coefficient of DAGAT over Kennedy pathway flux (JKen) at zero inhibitor (2-bromooctanoate) concentration (i = 0) and the subscript DAGAT is the activity of DAGAT.

During our experiments, we allowed the system to settle to a new ‘steady-state’ (as used in the widely accepted biochemical sense to refer to the quasi steady-state in a dynamic metabolic system where there was net flux, rates were linear and metabolites time-invariant over the period of study) after each perturbation. The substrates used were provided in excess so that their concentration remained effectively constant during the duration of the experiment. All necessary controls [12,13], were performed.


We are grateful to the Malaysian Palm Oil Board for studentship support for USR.