Diacylglycerol acyltransferases (DGATs) catalyze the final and only committed step of triacylglycerol synthesis. DGAT activity is rate limiting for triacylglycerol accumulation in mammals, plants and microbes. DGATs belong to three different evolutionary classes. In Arabidopsis thaliana, DGAT1, encoded by At2g19450, is the major DGAT enzyme involved in triacylglycerol accumulation in seeds. Until recently, the function of DGAT2 (At3g51520) has remained elusive. Previous attempts to characterize its enzymatic function by heterologous expression in yeast were unsuccessful. In the present report we demonstrate that expression of a codon-optimized version of the DGAT2 gene is able to restore neutral lipid accumulation in the Saccharomyces cerevisiae mutant strain (H1246), which is defective in triacylglycerol biosynthesis. Heterologous expression of codon-optimized DGAT2 and DGAT1 induced the biogenesis of subcellular lipid droplets containing triacylglycerols and squalene. Both DGAT proteins were found to be associated with these lipid droplets. The fatty acid composition was affected by the nature of the acyltransferase expressed. DGAT2 preferentially incorporated C16:1 fatty acids whereas DGAT1 displayed preference for C16:0, strongly suggesting that these enzymes have contrasting substrate specificities.
Citation: Aymé L, Baud S, Dubreucq B, Joffre F, Chardot T (2014) Function and Localization of the Arabidopsis thaliana Diacylglycerol Acyltransferase DGAT2 Expressed in Yeast. PLoS ONE 9(3): e92237. https://doi.org/10.1371/journal.pone.0092237
Editor: Monika Oberer, University of Graz, Austria
Received: October 25, 2013; Accepted: February 19, 2014; Published: March 24, 2014
Copyright: © 2014 Aymé et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was done in the context of the program CAER supported by DGA (Direction Générale de l’Armement)/DGAC (Direction Générale de l’Aviation Civile) under the convention n°2012 93 0805. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Triacylglycerols (TAGs) are the main form of fatty acid (FA) storage in most eukaryotes . They accumulate in dedicated organelles called lipid droplets (LDs). LDs are composed of a neutral lipid core surrounded by a monolayer of phospholipids (PLs), in which various types of specialized proteins are embedded. LDs serve as an energy reservoir, used for example during seed germination in plants , and are the source of signaling molecules and substrates for membrane biogenesis .
Three enzymes catalyze TAG synthesis by the acylation of sn-1,2-diacylglycerol (DAG) . Two acyltransferases catalyze acyl-CoA independent reactions. Phospholipid: diacylglycerol acyltransferase (PDAT, EC 126.96.36.199) uses PLs as acyl donors. Its activity has been reported in yeast and plants and corresponding genes have been identified , . DAG transacylase uses two DAGs as substrates to produce one TAG and one monoacylglycerol (MAG). This activity has been reported in various organisms ,  but to date, genes encoding this enzyme remain unknown. In contrast to PDAT and DAG transacylase, acyl-CoA: diacylglycerol acyltransferases (DGAT, EC 188.8.131.52) are acyl-CoA dependent enzymes, which catalyze the final and only committed step of the Kennedy pathway , , . In seeds, the absence of DGAT activity leads to a significant reduction in TAG accumulation , .
DGAT genes fall within three distinct families that share no sequence homology and probably result from convergent evolution , . DGAT1 belongs to the superfamily of membrane bound O-acyltransferases , which includes acyl-CoA:cholesterol acyltransferase-1 and -2 (ACAT, EC 184.108.40.206). DGAT2 belongs to the same family as acyl-CoA:monoacylglycerol acyltransferases (EC 220.127.116.11) and wax monoester synthases (18.104.22.168) . The existence of a cytosolic form of DGAT, belonging to a third family was reported in developing peanut cotyledons  and in Arabidopsis thaliana seedlings .
DGAT1 and 2 are found in mammals. These endoplasmic reticulum (ER) integral proteins ,  are both responsible for TAG production ,  and presumably have different physiological functions , , . The most significant contribution to TAG synthesis in the yeast Saccharomyces cerevisiae is from Dga1p, a member of the DGAT2 family , . In the oleaginous yeast Yarrowia lipolytica Dga2p, a member of the DGAT1 family makes a major contribution to TAG synthesis .
In A. thaliana DGAT1 encoded by At2g19450 plays a major role in seed lipid accumulation. Its activity and function have been extensively characterized in planta , . DGAT1 inactivation results in a 20 to 40% decrease in seed oil content , , . The observation that oil deposition was not completely abolished in this mutant background strongly suggested that other enzymes also contribute to TAG accumulation in seeds. Accordingly, a concomitant down regulation of PDAT1 and DGAT1 further decreased the seed oil content (by 70 to 80%), thus establishing that PDAT1 and DGAT1 play partially redundant functions in seeds .
DGAT2 coding sequences are found in the genome of several plants. This enzyme can incorporate unusual FAs into TAGs , such as eleostearic acid, a polyunsaturated FA found in Tung tree (Vernicia. fordii) , or ricinoleic acid, a hydroxy FA from Ricinus communis . The A. thaliana genome encodes a putative DGAT2 (At3g51520), the functionality of which was recently demonstrated by transient expression in Nicotiana benthamiana leaves . However, seeds from A. thaliana dgat2 mutants do not show altered TAG accumulation and introduction of the dgat2 mutation in a dgat1 mutant background did not aggravate the dgat1 seed phenotype, suggesting that DGAT2 is not involved in seed oil accumulation .
The current model of LD biogenesis proposes that LDs occur in specialized subdomains of the ER, gathering enzymes of the neutral lipid biosynthetic pathway . Accumulation of lipids between the two membrane leaflets would lead to budding of LDs covered with a PL monolayer , . The presence of proteins normally associated with the ER membrane at the surface of LDs is in favor of the budding model. This has been verified for A. thaliana DGAT1  and DGAT2 .
Here we report the functional characterization of A. thaliana DGAT2 in the yeast mutant strain H1246 which is unable to accumulate neutral lipids . We demonstrate that yeast codon usage limits protein expression in this system and that expression of codon-optimized DGAT2 can restore TAG synthesis by functional complementation. A. thaliana DGAT1 was used as a positive control. We then show that the two DGAT proteins are associated with LDs. Both enzymes exhibit contrasted substrate specificity and induce squalene accumulation in LDs.
Materials and Methods
Cloning of DGAT sequences for expression in yeast
The A. thaliana DGAT1 (At2g19450) and DGAT2 (At3g51520) cDNAs (Fig. S1) were amplified with proofreading Pfu Ultra DNA polymerase (STRATAGENE, La Jolla, CA, USA) from a mixture of seed cDNAs (of the Ws ecotype) using the primers: 5′-attB1-ATGGCGATTTTGGATTC-3′ and 5′-attB2-TCATGACATCGATCCTTTTC-3′ (DGAT1); 5′-attB1-ATGGGTGGTTCCAGAG-3′ and 5′-attB2-TCAAAGAATTTTCAGCTCAAG-3′ (DGAT2). AttB1 and attB2 refer to the corresponding Gateway recombination sequences. BP recombination was used to introduce the PCR products into the pDONR207 entry vector (INVITROGEN, Carlsbad, CA, USA) for sequencing. Codon sequences were optimized for expression in S. cerevisiae (Fig. S1) by Eurofin MWG Operon (Ebersberg, Germany). These sequences were amplified using Phusion Hot Start polymerase (Thermo Scientific, Illkirch, France) and introduced into the BamHI/SmaI restriction sites of the pRT21 vector (also known as pNBT29)  to generate proteins fused with a C-terminal GFP. Corresponding constructs were made with a stop codon introduced for expression of DGAT alone without the GFP fusion. Details of the oligonucleotides used are shown in Table 1. All constructs were sequenced to confirm the absence of mutations in amplified sequences (Genoscreen, Lille, France).
Expression of DGATs in yeast
Details of the recombinant plasmids as well as an empty plasmid control used to transform the S. cerevisiae quadruple mutant (are1, are2, dga1, lro1) strain H1246 or the SCY62 wild-type (WT) strain  are given in Table 2. Cells were grown in Ura deficient medium as described in  and protein expression was induced with 2% galactose in the presence of 0.02% glucose.
Lipid droplet purification
Yeast LDs were separated by sucrose density gradients as previously described by Yu et al. . Cells corresponding to 400 UA600nm were harvested by centrifugation, washed with water and resuspended in Fat Body Buffer (FBB, 10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, pH 7.5) supplemented with protease inhibitors (cOmplete, Mini, EDTA-free, Roche, Indianapolis, USA). Cells were disrupted with a One Shot cell disruptor (Constant Systems Ltd, Daventry, UK) at a pressure of 2.97 kbar. Lysates were centrifuged for 10 min at 12 000 g. The supernatant volume was adjusted to 5.4 ml with FBB containing sucrose at a final concentration of 0.54 M. It was overlaid sequentially with FBB containing 0.27 M, 0.135 M and 0 M sucrose (3×1.8 mL). The gradients were subjected to ultracentrifugation for 90 min at 150 000 g and 4°C in a SW41 Ti swing-out rotor (Beckman Coulter, Villepinte, France). The floating lipid layer, or the corresponding volume for the strain transformed with the empty plasmid, was collected and stored at −80°C. The hydrodynamic diameter of purified LDs was assessed by dynamic light scattering on a Malvern HPPS (Malvern Instruments Ltd, Worcestershire, UK) according to Vindigni et al. .
SDS-PAGE and immunoblotting
LD proteins were separated using 10% ready-to-use NuPAGE gels and LDS sample buffer (Invitrogen, Cergy Pontoise, France). Proteins were transferred to PVDF membrane (Immobilon-P, 0,45 μm pore size, Merck-Millipore, Billerica, USA) and proteins fused to GFP were probed with a mouse anti-GFP IgG (dilution 1∶1000) (Roche, Indianapolis, USA). Primary antibody was detected using horseradish peroxidase-conjugated anti-mouse IgG secondary antibody (dilution 1∶2500) (Sigma-Aldrich, L’Isle d’Abeau, France), and revealed using SuperSignal West Dura Extended Duration Substrate (Thermo Scientific, Illkirch, France). Luminescence was recorded using the LAS-3000 imaging system and MultiGauge software (Fujifilm, Saint Quentin en Yvelines, France).
Lipids from yeast cells or LDs were extracted as described by Folch et al. . Cells were harvested by centrifugation and washed twice with a solution containing 0.5% BSA and 0.9% NaCl. After freeze-drying, 100 mg of dry cells or LDs purified from 1000 UA600nm were resuspended in 5 ml of chloroform/methanol 2/1 (v/v). After 1h incubation with shaking, the extract was centrifuged for 5 min at 900 g and the supernatant was recovered and mixed with 2.5 ml of 0.9% NaCl. The organic phase was collected after centrifugation at 900 g for 5 min and washed three times with chloroform/methanol/water 3/48/47 (v/v/v) . Solvents were evaporated under a stream of nitrogen.
Lipid classes were separated by thin-layer chromatography (TLC). Lipids were solubilized in chloroform/methanol 2/1 (v/v) before separation on silica-coated aluminum plates (TLC Silica gel 60 F254, Merck, Fontenay Sous Bois, France) pre-washed in chloroform/methanol 1/1 (v/v). Plates were developed twice with petroleum ether/diethyl ether/acetic acid 50/50/2 (v/v/v) and petroleum ether/diethyl ether 49/1 (v/v) to 5 and 12 cm , . Lipids were stained with 5% phosphomolybdic acid for 30 min at 100°C. Lipids were identified based on the migration of lipid standards (Sigma-Aldrich, L’Isle d’Abeau, France). Plates were scanned using an Epson expression 1680 Pro Scanner. When necessary, lipids were quantified by densitometry using MultiGauge software (Fujifilm), according to Spanova et al. .
Fatty acid determination
The FA content and composition of yeast strains were determined according to Browse et al. . After freeze-drying, 20 mg of dry yeast were incubated in 2 ml of MeOH/H2SO4 (100/2.5, v/v) at 80°C for 90 min after addition of heptadecanoic acid as an internal standard. After addition of 900 μl of hexane and 3 ml of 0.9% NaCl, FA Methyl Esters (FAMEs) were extracted into the organic phase and analyzed by gas chromatography with flame ionization detection (GC-FID) according to Froissard et al. .
Neutral lipids were obtained upon fractionation of total lipids using an Isolute SPE Aminopropyl column (ALLTECH France Sarl, Epernon, France) according to . After transmethylation, FAMEs were identified and quantified by GC-FID.
Compositional analysis and quantification of the neutral lipid fraction
Neutral lipids were derivatized to TriMethylSylil-lipids (TMS-lipids) prior to their analyses by GC-FID using 100 μL of reagent solution (n-methyl imidazole/N-methyl-N-trimethylsilyl-heptafluorobutyramide, 1/20, v/v). Analysis of TMS-lipid derivatives were performed on a Trace GC Ultra gas Chromatograph (Thermo Scientific, Illkirch, France) equipped with a fused-silica ZB5 HT capillary column (15 m×0.25 mm I.D., 0.1μm film thickness; Phenomenex, Le Pecq, France). Samples were injected into an on-column detector and the FID system was set at 400°C. The oven temperature program was 100°C, increased to 370°C at 10°C/min and isothermal for 5 min at this final temperature (total run time 32 min). The carrier gas (H2) flow was maintained constant at 1.5 mL/min. A mixture containing the following compounds: FAs (C16, C18 and C18:1), MAG (C16, C18 and C18:1), DAG (equivalent to 34 and 36 atoms of carbon), TAG (equivalent to 48, 50, 52 and 54 atoms of carbon), squalene, cholesterol and phytosterols (campesterol, stigmasterol and sitosterol) was used as the standard for further product identification .
After extraction, 100 μg of lipids were resuspended in 300 μl hexane and dried over Na2SO4. Silylation was performed on 10 μl of the sample with 50 μl of BSTFA and 5 μl of pyridine (Sigma-Aldrich, L'Isle d'Abeau, France). Lipids were separated on a DB1 supelco capillary column of 15 m×0.32 mm (carrier gas: helium at a constant flow rate of 1.5 ml.min−1) with a first increase from 45°C to 180°C (at 30°C.min−1) and a second rise from 180°C to 280°C (at 3°C.min−1). Detection was performed by mass spectrometry on a Varian 4000 ion trap operating in the electron impact mode (70 eV) with ions detected on a range of 50 to 800 m/z. The presence of squalene in the sample was assessed by comparing the mass spectra of the component having an identical retention time to squalene with the mass spectra of squalene (≥ 98% squalene, Sigma-Aldrich)
Confocal Image acquisition
Yeast were incubated for 20 min in a Nile Red solution (1 g/l from a 1 g/ml stock solution in acetone), and washed with PBS solution containing 10% v/v glycerol. Images were acquired using an inverted LEICA SP2-AOBS spectral confocal laser microscope (LEICA Microsystems, Mannheim, Germany) using an HCX PL APO 63 X water (long distance) 1.2 objective. GFP and Nile Red fluorescence were observed with a 488 nm light wavelength generated by an argon laser and an emission band of 500−520 nm and 600−650 nm respectively. Transmitted light was captured with a dedicated photomultiplier (PMT trans). Each image is the average of eight scans at a resolution of 512×512 pixels with 8x numerical zoom.
Codon usage affects expression of DGAT2 and DGAT1 in S. cerevisiae
The DGAT2 (At3g51520) sequence is highly similar to several active DGAT2 genes from plants , . However, inactivation of the DGAT2 gene did not significantly reduce seed oil content  and expression of the gene in S. cerevisiae did not lead to lipid accumulation , , , . Taken together, these findings firstly suggested that DGAT2 does not encode a functional DGAT protein . Nevertheless, a very recent report has shown that DGAT2 can be actively expressed in N. benthamiana cells  and exhibits a higher specificity for C18:3 FA.
The function of A. thaliana DGAT2 was examined by heterologous expression in the H1246 S. cerevisiae strain. This strain lacks four acyltransferases and is completely defective in neutral lipid biosynthesis . Codon usage can affect expression of heterologous proteins in S. cerevisiae , , thus leading to the false conclusion that the expressed proteins are not active. Liu et al.  suggested that the lack of DGAT2 activity observed in S. cerevisiae could be due to differences in codon usage between this yeast and A. thaliana. To test this hypothesis and allow detection of the expressed proteins, we used WT or optimized versions of sequences encoding A. thaliana DGAT1 and 2 and harboring GFP tags for expression in S. cerevisiae. Accumulation of the two optimized DGATs in total yeast extracts was confirmed by Western blot analysis with anti-GFP antibodies (Fig. 1). When codon bias was not considered only slight expression of DGAT1 was observed, confirming previous results  and demonstrating the importance of codon optimization for plant DGAT expression in yeast. High molecular mass proteins were identified by immunoblotting which may reflect DGAT oligomerization as previously suggested for DGATs from other organisms , , , .
H1246 strains expressing GFP alone (empty vector), wild-type (WT) or an optimized (opt) version of DGAT-GFP were induced for 18 h. Yeast protein extracts corresponding to 0.06 UA600nm were separated by SDS-PAGE and transferred to a PVDF membrane. The membrane was incubated with anti-GFP antibody, and revealed with an anti-IgG-HRP secondary antibody. Respective molecular masses of DGAT1opt-GFP and DGAT2opt-GFP are 86 and 63 kDa. The marker used was MagicMark (Invitrogen).
DGAT2 restores neutral lipid accumulation in H1246
To study the effect of DGAT2 expression on the H1246 lipid profile, the quadruple mutant was transformed with an optimized version of the protein with no GFP tag (DGAT2opt). As a positive control, we used an H1246 strain transformed with an optimized version of DGAT1 (DGAT1opt). Total FAs from the different strains were quantified following transmethylation and GC-FID analysis. Expression of both DGAT1opt and DGAT2opt significantly increased total FA content in H1246 (Fig. 2.A). Cells expressing DGAT2opt and DGAT1opt contained 1.3 and 2.9 times more FAs, respectively, than H1246 transformed with an empty vector (negative control). Total lipids from the recombinant yeast were extracted and separated on TLC plates (Fig. 2.B). Extracts purified from both DGAT expressing strains contained TAGs. Densitometry showed that the DGAT2opt expressing strain contained 3.3 times less TAGs than the DGAT1opt expressing strain (Fig. 2.C). In addition, a lipid which accumulated in both these DGAT expressing lines had a retention factor identical to that of squalene (Fig. 2.B). Subsequent GC-MS analysis confirmed the identity of the product (Fig. S2). The highest squalene accumulation, determined by densitometric comparison of various amounts of squalene, reached approximately 2.7 g/kg in the DGAT1opt expressing strain.
H1246 strains transformed with the empty vector or expressing DGAT1opt or DGAT2opt were induced for 18 h. (A) FA content of yeast lipids. FAs from 20 mg of dry yeast were transmethylated. The resulting FAMEs were identified and quantified by GC-FID. Values are expressed in % of dry weight. Asterisks indicate statistically significant differences according to t-test (**P<0.01; ***P<0.001). (B) Separation of yeast lipids by TLC. Lipids were extracted from 100 mg of dry yeast. Lipids from 2 mg dry yeast were separated on a silica plate. This experiment is representative of three independent cultures. Vertical arrows indicate the direction of migration. (C) TAG and squalene contents. Relative amounts of TAGs and squalene to PLs were evaluated by densitometry after separation of yeast lipids by TLC (Fig. 2B). Values were calculated from three independent cultures. Asterisks indicate statistically significant differences according to a t-test (**P<0.01; ***P<0.001). (D) FA composition of yeast TAGs. Lipids were extracted from 100 mg of dry yeast. Neutral lipids were separated by SPE and transmethylated. The resulting FAMEs were identified and quantified by GC-FID. Values are expressed in % of total FAs in TAGs. (E) Separation of LD lipids by TLC. Lipids extracted from purified LDs of 30 UA600 nm from recombinant H1246 strains or from the WT strain (SCY62) transformed with the empty vector were separated on silica plates. Vertical arrows indicate the direction of migration. O: origin of migration; S: standard; PLs: phospholipids, St: sterols; FFAs: free fatty acids; TAGs: triacylglycerols; FAMEs: fatty acid methyl esters; SEs: sterol esters, Squal: squalene.
FA composition analyses of the neutral lipid fraction (TAGs) by GC-FID revealed differences depending on expression of DGAT1opt or DGAT2opt. The DGAT2opt expressing strain (Fig. 2.D) accumulated mostly C16:1 whereas the DGAT1opt expressing strain preferentially accumulated C16:0. However, no significant differences were found for the TAG fraction (Fig. S3). TAG profiles of both DGAT1opt and DGAT2opt expressing cells showed a normal distribution centered on C50, close to 35% with also a high proportion of C48 (18%) and C52 (23%). A small proportion of C54 was also noted.
Expression of DGAT2 and DGAT1 in S. cerevisiae H1246 restores LD biogenesis
No LDs form in the H1246 mutant strain . Neutral lipid production was re-established in yeast expressing optimized sequences encoding DGAT2 or DGAT1 as well as the non-optimized sequence coding for DGAT1 (data not shown). We attempted to purify LDs from H1246 transformed with the empty pRT21 vector or pRT21-DGAT1opt or pRT21-DGAT2opt. No floating layer was recovered from the H1246 strain transformed with the empty vector. Using dynamic light scattering we determined that the floating layers recovered from DGAT1opt expressing cells contained objects with diameters similar to LDs from a WT strain (SCY62, 330nm). In cells expressing DGAT2opt we observed smaller objects (220 nm).
Lipids contained in the floating layers were extracted and separated by TLC (Fig. 2.E). Lipids from DGAT1opt and DGAT2opt expressing cells were mainly TAGs, squalene and sterols, but lacked sterol esters in contrast to lipids extracted from the WT strain. No lipids could be extracted from the top fraction of the gradients for the H1246 strain transformed with the empty vector.
DGAT2 and DGAT1 colocalize with intracellular neutral lipids
Confocal fluorescence microscopy of DGAT1opt-GFP and DGAT2opt-GFP revealed the presence of the proteins in typical lipid particle structures after 8 h induction (Fig. 3.A). The same staining pattern was observed using the lipophilic fluorescent dye Nile Red which selectively accumulated in LDs (Fig. 3.A), thus confirming the localization of both DGAT proteins in this compartment. Similar results were obtained after 18 h induction, but neither GFP nor Nile Red were found in punctuate structures (Fig. 3.B).
H1246 strains expressing GFP alone, DGAT1opt-GFP or DGAT2opt-GFP were induced for 8 h (A) and 18 h (B) and observed with a confocal microscope, after incubation with Nile Red. GFP and Nile Red fluorescence was excited with a 488 nm light wavelength generated by an argon laser. Emitted light was collected at 500−520 nm (GFP) and 600−650 nm (Nile Red).
In A. thaliana, DGAT2 inactivation did not modify seed oil content. Furthermore, mutation of DGAT2 in the DGAT1 mutant background did not lead to any additional phenotypes, leading Zhang et al.  to conclude that DGAT2 does not play a substantial role in seed lipid accumulation. The possibility that DGAT2 could be expressed and active in vegetative tissues was very recently demonstrated by Zhou . Despite the fact that A. thaliana DGAT2 shares 71% and 58% identity with active Tung tree  and R. communis  DGAT2, previous heterologous expression experiments did not lead to TAG accumulation in the yeast mutant strain H1246 , . Liu et al.  suggested that the A. thaliana DGAT2 gene may exhibit a strong codon bias affecting its expression in yeast. In this report, we demonstrated that using yeast codon usage is mandatory for DGAT2 expression (Fig. 1) and to restore TAG accumulation (Fig. 2.B) in H1246. Optimization of A. thaliana DGAT1 codon usage also significantly increased the rate of protein synthesis and enhanced TAG accumulation. Heterologous DGAT expression is thus controlled by subtle usage of tRNA via codon bias.
H1246 expressing DGAT1 accumulated FA levels of up to 7% of dry weight (Fig. 2.A). We also found a significant amount of squalene in this strain (Fig. 2.B). Squalene is a triterpene precursor in sterol biosynthesis with potential uses in nutrition, cosmetics, and medicine . Up to 2.7 g of squalene per kg of dry biomass were produced in this strain. This value, obtained in non-optimized conditions, is comparable to the reported yields of promising bioindustrial strains . Spanova et al.  showed that the LD is a neutral lipid depository able to accommodate squalene. The WT yeast strain (SCY62) naturally accumulates squalene in LDs in addition to TAGs and sterols esters (Fig. 2.E). We found small amounts of squalene in the H1246 strain (Fig. 2.B) even though it lacks LDs  (Fig. 2.E). These results are consistent with those of Spanova et al.  who demonstrated that in the absence of LDs, in a dga1Δlro1Δare1Δare2Δ mutant strain, squalene can be stored in microsomal and mitochondrial membranes. In contrast to DGAT1, however, LD biosynthesis was probably not high enough in DGAT2-expressing cells for significant amounts of squalene to be stored. The amount of sterols found in all cells and LDs (Fig. 2.B and 2.E) did not vary, strongly suggesting that sterol biosynthesis is uncoupled from the level of its precursor squalene.
C16 and C18 FAs were the major FA species acylated in TAGs from both DGAT expressing strains (Fig. 2D), DGAT2 preferred C16:1 whereas DGAT1 exhibited specificity toward C16:0. The method used for TAG profiling does not differentiate TAG containing saturated from unsaturated FAs , thus the results could not highlight differences between neutral lipids from strains expressing DGATs (Fig S3). Depending on the plant species, DGAT1 or DGAT2 is a major enzyme responsible for the accumulation of seed TAG. A. thaliana DGAT1 is involved in acylation of the sn-3 position of the glycerol. Stereochemical analysis showed that the sn-1 and sn-3 positions are usually saturated, while unsaturated species occur mainly at sn-2 . In castor bean and Tung tree, DGAT2 is specific for unsaturated FAs , . The specialization of DGAT2 (C16:1) observed in the present study is consistent with these last findings. Similarly, transient heterologous expression of DGAT2 in tobacco leaves leads to the accumulation of TAG containing unsaturated FA  revealing possible expression under the subtle control of tissue specific factors. Thus, very long chain FA or polyunsaturated FA are potential acyl donors found in A. thaliana seeds and N. benthamiana leaves, but absent in S. cerevisiae. It is therefore difficult to compare enzyme specificity and velocity as reflected by lipid accumulation for DGATs expressed in heterologous systems. Overall, the results strongly suggest that these enzymes show specificity toward different acyl donors.
Our confocal microscopy investigations confirmed the co-localization of DGAT-GFP and neutral lipids. The punctuate structures observed after a short induction (8 h, Fig. 3.A) were typical of LDs, however Nile Red staining was more diffuse after longer induction times (18 h, Fig. 3.B). The lipids which accumulated in DGAT expressing yeast were different from the WT strain and lacked sterol esters. As previously suggested by Czabany , lipid composition affects LD protein content and as a consequence, the interactions between LDs and their environment could be modified. Nevertheless the objects purified from the yeast which had been induced for 18 h still had a typical LD hydrodynamic diameter. Thus it could be that close association of numerous LDs rendered individual droplets undistinguishable in vivo using light microscopy, explaining the shape of structures observed after 18 h induction (Fig. 3.B)
The association of DGAT1 with LDs and the ER was previously demonstrated by Bouvier-Navé et al. . DGAT2 localization to microsomes was deduced from in vitro assays by Zhou et al. , however its presence in other compartments cannot be excluded. Recent results by Kwiatkowska et al. tend to suggest multiple localizations of DGAT2 in germinated A. thaliana seeds . These last experiments were based on the use of antibodies directed against human DGAT2, and the cross-reactivity between plant proteins and human antibodies needs to be confirmed. Taken together, these reports and our present findings appear to support the general model proposing that TAG synthesis occurs at the ER with LDs budding from the ER .
We have described the enzymatic function and localization of A. thaliana DGAT2 and DGAT1 in the yeast quadruple mutant H1246. Functional complementation confirmed that DGAT2 has diacylglycerol acyltransferase activity. Differences observed at the level of FA composition suggest different substrate specificities for DGATs. Our study, together with previous reports from other teams, highlights the importance of tissue-specific factors and tRNA pools in regulating DGAT expression. Optimization of codon usage appears to be a valuable tool for increasing DGAT expression and subsequent lipid accumulation in yeast. Accumulation of squalene, a terpene with potential biotechnological applications, increased significantly by expression of Arabidopsis DGAT1. Overall, our results suggest that the nature of the DGAT expressed in yeast, is not only important for TAG accumulation, but also for the production of other lipids of interest.
Native and optimized gene sequences used in this study. Sequences were aligned using BioEdit .
Identification of squalene in yeast total lipid fractions. (A) Chromatogram of authentic squalene (upper panel, retention time 23.5 min) and of total lipids from DGAT1opt expressing strain (lower panel) (B) Fragmentation spectra of authentic squalene (upper panel) and of the molecule eluting at 23.5 min (lower panel).
TAG profiling of strains expressing DGAT. Partial GC-FID chromatogram (ZB5-HT column) focused on separation of TAG compounds of DGAT expressing strains and H1246 transformed with the empty vector (control) lipid extracts.
We wish to thank Marine Froissard for precious help, advice and fruitful discussions. We wish to thank Catherine Lapierre for expertise in squalene identification, and Sten Stymne for the generous gift of the H1246 and SCY62 strains.
We wish to thank Jean-Noël Arnaud for his contribution to the lipid analysis by GC-FID and GC-EI-MS.
We thank Leigh Gebbie, LKG SCIENTIFIC EDITING & TRANSLATION for her excellent help in correcting the English version of the manuscript.
Conceived and designed the experiments: LA TC. Performed the experiments: LA SB TC BD FJ. Analyzed the data: LA TC. Contributed reagents/materials/analysis tools: LA TC. Wrote the paper: LA TC. Plasmid and strain constructions, cell cultures, microscopic observations, lipid extraction and analysis, western blots: LA. Plasmid construction: SB. Confocal microscopy: BD. Neutral Lipid profiling by GC-FID: FJ. Critical reading of the manuscript: SB BD.
- 1. Zweytick D, Athenstaedt K, Daum G (2000) Intracellular lipid particles of eukaryotic cells. Biochim Biophys Acta 1469: 101–120.
- 2. Graham IA (2008) Seed Storage Oil Mobilization. Annual Review of Plant Biology 59: 115–142.
- 3. Liu Q, Siloto RM, Lehner R, Stone SJ, Weselake RJ (2012) Acyl-CoA:diacylglycerol acyltransferase: Molecular biology, biochemistry and biotechnology. Prog Lipid Res 51: 350–377.
- 4. Dahlqvist A, Stahl U, Lenman M, Banas A, Lee M, et al. (2000) Phospholipid:diacylglycerol acyltransferase: An enzyme that catalyzes the acyl-CoA-independent formation of triacylglycerol in yeast and plants. Proceedings of the National Academy of Sciences 97: 6487–6492.
- 5. Stahl U, Carlsson AS, Lenman M, Dahlqvist A, Huang B, et al. (2004) Cloning and Functional Characterization of a Phospholipid:Diacylglycerol Acyltransferase from Arabidopsis. Plant Physiology 135: 1324–1335.
- 6. Lehner R, Kuksis A (1993) Triacylglycerol synthesis by an sn-1,2(2,3)-diacylglycerol transacylase from rat intestinal microsomes. Journal of Biological Chemistry 268: 8781–8786.
- 7. Stobart K, Mancha M, Lenman M, Dahlqvist A, Stymne S (1997) Triacylglycerols are synthesised and utilized by transacylation reactions in microsomal preparations of developing safflower (Carthamus tinctorius L) seeds. Planta 203: 58–66.
- 8. Jako C, Kumar A, Wei YD, Zou JT, Barton DL, et al. (2001) Seed-specific over-expression of an Arabidopsis cDNA encoding a diacylglycerol acyltransferase enhances seed oil content and seed weight. Plant Physiology 126: 861–874.
- 9. Beopoulos A, Haddouche R, Kabran P, Dulermo T, Chardot T, et al. (2012) Identification and characterization of DGA2, an acyltransferase of the DGAT1 acyl-CoA:diacylglycerol acyltransferase family in the oleaginous yeast Yarrowia lipolytica. New insights into the storage lipid metabolism of oleaginous yeasts. Applied Microbiology and Biotechnology 93: 1523–1537.
- 10. Grisart B, Farnir F, Karim L, Cambisano N, Kim J-J, et al. (2004) Genetic and functional confirmation of the causality of the DGAT1 K232A quantitative trait nucleotide in affecting milk yield and composition. Proceedings of the National Academy of Sciences of the United States of America 101: 2398–2403.
- 11. Zou JT, Wei YD, Jako C, Kumar A, Selvaraj G, et al. (1999) The Arabidopsis thaliana TAG1 mutant has a mutation in a diacylglycerol acyltransferase gene. Plant Journal 19: 645–653.
- 12. Routaboul JM, Benning C, Bechtold N, Caboche M, Lepiniec L (1999) The TAG1 locus of Arabidopsis encodes for a diacylglycerol acyltransferase. Plant Physiol Biochem 37: 831–840.
- 13. Turchetto-Zolet AC, Maraschin FS, de Morais GL, Cagliari A, Andrade CM, et al. (2011) Evolutionary view of acyl-CoA diacylglycerol acyltransferase (DGAT), a key enzyme in neutral lipid biosynthesis. BMC Evol Biol 11: 263.
- 14. Hofmann K (2000) A superfamily of membrane-bound O-acyltransferases with implications for Wnt signaling. Trends in Biochemical Sciences 25: 111–112.
- 15. Yen CLE, Stone SJ, Koliwad S, Harris C, Farese RV (2008) DGAT enzymes and triacylglycerol biosynthesis. Journal of Lipid Research 49: 2283–2301.
- 16. Saha S, Enugutti B, Rajakumari S, Rajasekharan R (2006) Cytosolic Triacylglycerol Biosynthetic Pathway in Oilseeds. Molecular Cloning and Expression of Peanut Cytosolic Diacylglycerol Acyltransferase. Plant Physiology 141: 1533–1543.
- 17. Hernandez ML, Whitehead L, He Z, Gazda V, Gilday A, et al. (2012) A cytosolic acyltransferase contributes to triacylglycerol synthesis in sucrose-rescued Arabidopsis seed oil catabolism mutants. Plant Physiol 160: 215–225.
- 18. McFie PJ, Stone SL, Banman SL, Stone SJ (2010) Topological orientation of acyl-CoA:diacylglycerol acyltransferase-1 (DGAT1) and identification of a putative active site histidine and the role of the n terminus in dimer/tetramer formation. J Biol Chem 285: 37377–37387.
- 19. McFie PJ, Banman SL, Kary S, Stone SJ (2011) Murine diacylglycerol acyltransferase-2 (DGAT2) can catalyze triacylglycerol synthesis and promote lipid droplet formation independent of its localization to the endoplasmic reticulum. J Biol Chem 286: 28235–28246.
- 20. Cases S, Smith SJ, Zheng Y-W, Myers HM, Lear SR, et al. (1998) Identification of a gene encoding an acyl CoA:diacylglycerol acyltransferase, a key enzyme in triacylglycerol synthesis. Proceedings of the National Academy of Sciences 95: 13018–13023.
- 21. Cases S, Stone SJ, Zhou P, Yen E, Tow B, et al. (2001) Cloning of DGAT2, a second mammalian diacylglycerol acyltransferase, and related family members. J Biol Chem 276: 38870–38876.
- 22. Stone SJ, Myers HM, Watkins SM, Brown BE, Feingold KR, et al. (2004) Lipopenia and skin barrier abnormalities in DGAT2-deficient mice. J Biol Chem 279: 11767–11776.
- 23. Smith SJ, Cases S, Jensen DR, Chen HC, Sande E, et al. (2000) Obesity resistance and multiple mechanisms of triglyceride synthesis in mice lacking Dgat. Nat Genet 25: 87–90.
- 24. Chen HC, Stone SJ, Zhou P, Buhman KK, Farese RV Jr (2002) Dissociation of obesity and impaired glucose disposal in mice overexpressing acyl coenzyme a:diacylglycerol acyltransferase 1 in white adipose tissue. Diabetes 51: 3189–3195.
- 25. Lardizabal KD, Mai JT, Wagner NW, Wyrick A, Voelker T, et al. (2001) DGAT2 Is a New Diacylglycerol Acyltransferase Gene Family. Journal of Biological Chemistry 276: 38862–38869.
- 26. Sandager L, Gustavsson MH, Stahl U, Dahlqvist A, Wiberg E, et al. (2002) Storage lipid synthesis is non-essential in yeast. J Biol Chem 277: 6478–6482.
- 27. Bouvier-Nave P, Benveniste P, Oelkers P, Sturley SL, Schaller H (2000) Expression in yeast and tobacco of plant cDNAs encoding acyl CoA:diacylglycerol acyltransferase. Eur J Biochem 267: 85–96.
- 28. Zhang M, Fan J, Taylor DC, Ohlrogge JB (2009) DGAT1 and PDAT1 acyltransferases have overlapping functions in Arabidopsis triacylglycerol biosynthesis and are essential for normal pollen and seed development. Plant Cell 21: 3885–3901.
- 29. Katavic V, Reed DW, Taylor DC, Giblin EM, Barton DL, et al. (1995) Alteration of Seed Fatty Acid Composition by an Ethyl Methanesulfonate-Induced Mutation in Arabidopsis thaliana Affecting Diacylglycerol Acyltransferase Activity. Plant Physiology 108: 399–409.
- 30. Cahoon EB, Shockey JM, Dietrich CR, Gidda SK, Mullen RT, et al. (2007) Engineering oilseeds for sustainable production of industrial and nutritional feedstocks: solving bottlenecks in fatty acid flux. Current Opinion in Plant Biology 10: 236–244.
- 31. Shockey JM, Gidda SK, Chapital DC, Kuan JC, Dhanoa PK, et al. (2006) Tung tree DGAT1 and DGAT2 have nonredundant functions in triacylglycerol biosynthesis and are localized to different subdomains of the endoplasmic reticulum. Plant Cell 18: 2294–2313.
- 32. Kroon JT, Wei W, Simon WJ, Slabas AR (2006) Identification and functional expression of a type 2 acyl-CoA:diacylglycerol acyltransferase (DGAT2) in developing castor bean seeds which has high homology to the major triglyceride biosynthetic enzyme of fungi and animals. Phytochemistry 67: 2541–2549.
- 33. Zhou XR, Shrestha P, Yin F, Petrie JR, Singh SP (2013) AtDGAT2 is a functional acyl-CoA:diacylglycerol acyltransferase and displays different acyl-CoA substrate preferences than AtDGAT1. FEBS Lett 587: 2371–2376.
- 34. Murphy DJ (2001) The biogenesis and functions of lipid bodies in animals, plants and microorganisms. Prog Lipid Res 40: 325–438.
- 35. Kwiatkowska M, Stepinski D, Poplonska K, Wojtczak A, Polit JT (2012) DGAT2 revealed by the immunogold technique in Arabidopsis thaliana lipid bodies associated with microtubules. Folia Histochemica Et Cytobiologica 50: 427–431.
- 36. Froissard M, Belgareh-Touze N, Buisson N, Desimone M, Frommer WB, et al. (2006) Heterologous expression of a plant uracil transporter in yeast: improvement of plasma membrane targeting in mutants of the Rsp5p ubiquitin protein ligase. Biotechnol J 1: 308–320.
- 37. Froissard M, D'Andrea S, Boulard C, Chardot T (2009) Heterologous expression of AtClo1, a plant oil body protein, induces lipid accumulation in yeast. FEMS Yeast Res 9: 428–438.
- 38. Yu W, Cassara J, Weller PF (2000) Phosphatidylinositide 3-kinase localizes to cytoplasmic lipid bodies in human polymorphonuclear leukocytes and other myeloid-derived cells. Blood 95: 1078–1085.
- 39. Vindigni JD, Wien F, Giuliani A, Erpapazoglou Z, Tache R, et al. (2013) Fold of an oleosin targeted to cellular oil bodies. Biochim Biophys Acta 1828: 1881–1888.
- 40. Folch J, Lees M, Sloane Stanley GH (1957) A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem 226: 497–509.
- 41. Schneiter R, Daum G (2006) Extraction of yeast lipids. Methods Mol Biol 313: 41–45.
- 42. Athenstaedt K, Zweytick D, Jandrositz A, Kohlwein SD, Daum G (1999) Identification and characterization of major lipid particle proteins of the yeast Saccharomyces cerevisiae. J Bacteriol 181: 6441–6448.
- 43. Schneiter R, Daum G (2006) Analysis of yeast lipids. Methods Mol Biol 313: 75–84.
- 44. Spanova M, Zweytick D, Lohner K, Klug L, Leitner E, et al. (2012) Influence of squalene on lipid particle/droplet and membrane organization in the yeast Saccharomyces cerevisiae. Biochim Biophys Acta 1821: 647–653.
- 45. Browse J, McCourt PJ, Somerville CR (1986) Fatty acid composition of leaf lipids determined after combined digestion and fatty acid methyl ester formation from fresh tissue. Anal Biochem 152: 141–145.
- 46. Beopoulos A, Haddouche R, Kabran P, Dulermo T, Chardot T, et al. (2012) Identification and characterization of DGA2, an acyltransferase of the DGAT1 acyl-CoA:diacylglycerol acyltransferase family in the oleaginous yeast Yarrowia lipolytica. New insights into the storage lipid metabolism of oleaginous yeasts. Appl Microbiol Biotechnol 93: 1523–1537.
- 47. Pocklington WD, Hautfenne A (1985) Determination of triglycerides in fats and oils - Results of a collaborative study and the standardized method. Pure and Applied Chemistry 57: 1515–1522.
- 48. Weselake RJ, Taylor DC, Rahman MH, Shah S, Laroche A, et al. (2009) Increasing the flow of carbon into seed oil. Biotechnol Adv 27: 866–878.
- 49. Burgal J, Shockey J, Lu C, Dyer J, Larson T, et al. (2008) Metabolic engineering of hydroxy fatty acid production in plants: RcDGAT2 drives dramatic increases in ricinoleate levels in seed oil. Plant Biotechnol J 6: 819–831.
- 50. Xu J, Carlsson AS, Francis T, Zhang M, Hoffman T, et al. (2012) Triacylglycerol synthesis by PDAT1 in the absence of DGAT1 activity is dependent on re-acylation of LPC by LPCAT2. BMC Plant Biol 12: 4.
- 51. Chiapello H, Lisacek F, Caboche M, Henaut A (1998) Codon usage and gene function are related in sequences of Arabidopsis thaliana. Gene 209: GC1–GC38.
- 52. Weselake RJ, Madhavji M, Szarka SJ, Patterson NA, Wiehler WB, et al. (2006) Acyl-CoA-binding and self-associating properties of a recombinant 13.3 kDa N-terminal fragment of diacylglycerol acyltransferase-1 from oilseed rape. BMC Biochem 7: 24.
- 53. Spanova M, Daum G (2011) Squalene - biochemistry, molecular biology, process biotechnology, and applications. European Journal of Lipid Science and Technology 113: 1299–1320.
- 54. Spanova M, Czabany T, Zellnig G, Leitner E, Hapala I, et al. (2010) Effect of lipid particle biogenesis on the subcellular distribution of squalene in the yeast Saccharomyces cerevisiae. J Biol Chem 285: 6127–6133.
- 55. Miquel M, Browse J (1995) Lipid biosynthesis in developing seeds. In: Kigel J, Galili G, editors. Seed Development and Germination. New-York: Marcel Dekker. pp. 169−193.
- 56. Czabany T, Wagner A, Zweytick D, Lohner K, Leitner E, et al. (2008) Structural and biochemical properties of lipid particles from the yeast Saccharomyces cerevisiae. J Biol Chem 283: 17065–17074.
- 57. Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser: 95−98.