Triacylglycerol synthesis by PDAT1 in the absence of DGAT1 activity is dependent on re-acylation of LPC by LPCAT2

Background The Arabidopsis thaliana dgat1 mutant, AS11, has an oil content which is decreased by 30%, and a strongly increased ratio of 18:3/20:1, compared to wild type. Despite lacking a functional DGAT1, AS11 still manages to make 70% of WT seed oil levels. Recently, it was demonstrated that in the absence of DGAT1, PDAT1 was essential for normal seed development, and is a dominant determinant in Arabidopsis TAG biosynthesis. Methods Biochemical, metabolic and gene expression studies combined with genetic crossing of selected Arabidopsis mutants have been carried out to demonstrate the contribution of Arabidopsis PDAT1 and LPCAT2 in the absence of DGAT1 activity. Results Through microarray and RT-PCR gene expression analyses of AS11 vs. WT mid-developing siliques, we observed consistent trends between the two methods. FAD2 and FAD3 were up-regulated and FAE1 down-regulated, consistent with the AS11 acyl phenotype. PDAT1 expression was up-regulated by ca 65% while PDAT2 expression was up-regulated only 15%, reinforcing the dominant role of PDAT1 in AS11 TAG biosynthesis. The expression of LPCAT2 was up-regulated by 50-75%, while LPCAT1 expression was not significantly affected. In vitro LPCAT activity was enhanced by 75-125% in microsomal protein preparations from mid-developing AS11 seed vs WT. Co-incident homozygous knockout lines of dgat1/lpcat2 exhibited severe penalties on TAG biosynthesis, delayed plant development and seed set, even with a functional PDAT1; the double mutant dgat1/lpcat1 showed only marginally lower oil content than AS11. Conclusions Collectively, the data strongly support that in AS11 it is LPCAT2 up-regulation which is primarily responsible for assisting in PDAT1-catalyzed TAG biosynthesis, maintaining a supply of PC as co-substrate to transfer sn-2 moieties to the sn-3 position of the enlarged AS11 DAG pool.


Background
Triacylglycerols (TAGs) are the major storage lipids which accumulate in developing seeds, flower petals, anthers, pollen grains, and fruit mesocarp of a number of plant species [1,2]. TAGs are thought to be not only the major energy source for seed germination but also essential for pollen development and sexual reproduction in many plants [3,4]. In oilseeds, TAG bioassembly is catalyzed by the membrane-bound enzymes of the Kennedy pathway that operate in the endoplasmic reticulum [4]. The biosynthesis of TAGs is catalyzed by the sequential acylation of the glycerol backbone involving three acyltransferases: glycerol-3-phosphate acyltransferase (GPAT), lyso-phosphatidic acid acyltransferase (LPAAT) and diacylglycerol acyltransferase (DGAT). DGAT catalyses the final acylation of DAG to give TAG, which has been suggested to be the rate-limiting step in plant lipid accumulation.
In the traditional Kennedy pathway DGAT was thought to be the only enzyme that is exclusively committed to TAG biosynthesis using acyl-CoA as its acyl donor. The first DGAT gene was cloned from mouse and is a member of the DGAT1 family, which had high sequence similarity with sterol: acyl-CoA acyltransferase [5].
We had previously characterized an EMS-induced mutant of Arabidopsis, designated AS11, which displayed a decrease in stored seed TAG, delayed seed development, and an altered fatty acid composition [6]. We analyzed WT vs. AS11 lipid pools and Kennedy pathway enzyme activities in fractions isolated from green mid-developing seed, and performed parallel labeling of intact seeds at this developmental stage, with [ 14 C] acetate. We found that compared to WT, there was an increase in all fatty acids in the DAG pool of AS11 seeds at mid-development, and, to a lesser extent, an associated backup of fatty acids in the PC pool. DAG was elevated from 1% in WT to 10-12% in AS11 and PC pools were elevated from about 2% in WT, to 8-12% in AS11. Cell-free fractions from WT and AS11 green seeds at mid-development were compared for their ability to incorporate [ 14 C]-18:1-CoA into glycerolipids in the presence of G-3-P. Proportions of labeled LPA and PA formed during the incubation period were similar in WT vs AS11, indicating that the activities of the Kennedy pathway enzymes GPAT and LPAAT (EC 2.3.1.51) were relatively unaffected in the AS11 mutant. However, the proportion of labeled TAG was much lower and that of DAG was much higher in AS11. The TAG/DAG ratio was therefore consistently 3-5-fold lower in AS11 compared to WT at all developmental stages (early-, mid-and late development). Furthermore, the ratio of 18:3/20:1 dramatically increased about 7-10 fold [6].
Cumulatively, this data suggested a lesion in DGAT1 which was subsequently demonstrated upon cloning the mutated gene from AS11. There is an 81 bp in-frame insertion consisting entirely of exon 2 in the transcript from AS11. The exon 2 in the repeat is properly spliced, thus the alteration of the transcript does not disturb the reading frame. However, this additional exon 2 sequence in the AS11 transcript would result in an altered DGAT protein with a 27 amino acid insertion ( 131 SHAGLFNLC VVVLIAVNSRLIIENLMK 157 ) [7]. It is important to note that two other labs independently and simultaneously cloned the A. thaliana DGAT1 [8,9].
Studies manipulating the expression of DGAT1 followed: We demonstrated that expression of the Arabidopsis DGAT1 cDNA in a seed specific manner in the AS11 mutant restored wild type levels of TAG and VLCFA content. The acyl distribution, specifically, the sn-3 composition of the TAGs, was also restored to WT proportions. Furthermore, overexpression of the Arabidopsis DGAT1 in wild type plants led to an increase in seed oil content and seed weight [10]. Over the past 10 years, DGAT1 expression has been genetically manipulated to produce Brassica napus prototypes containing increased seed oil [11,12].
A second family of DGAT genes (DGAT2), first identified in the oleaginous fungus Morteriella ramanniana, has no sequence similarity with DGAT1 [13]. A human DGAT2 and plant DGAT2s from tung and castor were subsequently identified by Cases et al. [14], Shockey et al. [15] and Kroon et al. [16], respectively. The putative DGAT2 from Arabidopsis has been studied by several labs including ours; functional expression in yeast has not been successful, and therefore whether it is a true functioning DGAT is still in question. Notably, an Arabidopsis dgat2 knockout mutant has a wild-type seed oil content and fatty acid composition [17].
TAG can also be formed by an acyl-CoA-independent enzyme, phosphatidylcholine: diacylglycerol acyltransferase (PDAT), in which the transfer of an acyl group from the sn-2 position of PC to the sn-3 position of DAG yields TAG and sn-1 lyso-PC [18,19]. During the exponential growth phase in yeast, PDAT1 is a major determinant in TAG synthesis. In Arabidopsis, two close homologs to the yeast PDAT gene have been identified: PDAT1 At5g13640 and PDAT2 At3g44830 [20]. Mhaske et al. [21] isolated and characterized a knockout mutant of Arabidopsis which has a T-DNA insertion in the PDAT1 locus At5g13640 (PDAT1, EC 2.3.1.158). Lipid analyses were conducted on this mutant to assess the contribution of PDAT1 to seed lipid biosynthesis; surprisingly, and in contrast to the situation in yeast, the oil content and composition in seeds did not show significant changes in the mutant. At the time, these results were interpreted to indicate that PDAT1 activity as encoded by At5g13640 is not a major determining factor for TAG synthesis in Arabidopsis seeds.
Nonetheless, because the Arabidopsis DGAT1 mutant AS11 shows only a 30-35% decrease in oil content [6,9], it was apparent that other enzymes must contribute to oil synthesis in the developing seed [22]. Thus, an examination of the contribution of DGAT2, PDAT2 or PDAT1 to oil deposition in an AS11 background was studied by performing double mutant crosses with AS11 [17]. While the dgat2-ko line has no oil phenotype, homozygous double mutants from cross of AS11 with dgat2-ko mutant showed an oil fatty acid profile similar to AS11. The same pattern was observed with the pdat2-ko mutant alone and in crosses of the pdat2-ko mutant with AS11. In contrast, while the pdat1-ko has no oil or fatty acid composition phenotype, crosses of the pdat1-ko with AS11 were embryo-lethal in the double homozygous condition; only heterozygous lines produced by having expression of the pdat1 or dgat1 gene only partially inhibited using RNAi, allowed an examination of the double mutants. These detailed studies resulted in the finding that DGAT1 and PDAT1 have overlapping functions in both embryo development and TAG biosynthesis in the developing seed and in pollen. When DGAT1 is compromised in AS11, it is PDAT1, and not DGAT2 or PDAT2, that is responsible for the remaining 65-70% of TAG which is synthesized. This finding suggested a major, perhaps dominant role of PDAT1 in this process [17].
Recently, a castor bean-specific PDAT, PDAT1-2, was cloned and found to be highly expressed in developing seeds and localized in the ER, similar to the castor FAH12 hydroxylase. Transgenic Arabidopsis co-expressing the castor PDAT1-2 and FAH12 showed enhanced ricinoleate accumulation to up to 25% in TAGs (compared to 17% in FAH12 -only transgenics) [23,24]. This study suggests that specialized PDATs may play a significant role in channeling PC-synthesized unusual fatty acids such as ricinoleic (from castor), or epoxy fatty acids (e.g. from Vernonia galamensis), into TAGs.
The discovery of the critical role PDAT1 has in TAG synthesis [17] suggested the importance of re-acylation of lysophospholipids and especially LPC, since it is produced as a result of PDAT1 activity. We hypothesized that in a situation where the responsibility for TAG synthesis is shifted to PDAT1 such as in the AS11 dgat1 mutant, profound changes in activity of re-acylation enzymes would be evident. As candidates of genes coding such enzymes, we proposed the two Arabidopsis genes LPLAT1 (At1g12640) and LPLAT2 (At1g63050), characterized by Ståhl et al. [25] as possessing broad specificity lysophospholipid acyltransferase activity. Although both of the genes to some extent could acylate a range of different lysophospholipids, they were shown to have a strong preference towards LPC. For this reason we chose to name the two genes LPCAT1 (At1g12640) and LPCAT2 (At1g63050), in the current study.
Here we report the further genetic and biochemical characterization of the AS11 dgat1 mutant. During the course of microarray and qRT-PCR studies of AS11 vs WT gene expression in mid-developing siliques, we found that LPCAT2, encoding acyl-CoA:lysophosphatidylcholine acyltransferase 2 (EC 2.3.1.23), was up-regulated while LPCAT1 was not affected. By a series of biochemical studies and key crosses of AS11 with either lpcat1 or lpcat2 knockout mutants, we determined that LPCAT2 (and much less so LPCAT1) is most critical for TAG synthesis in the AS11 mutant, primarily to maintain the PC pool for TAG assembly primarily catalyzed by PDAT1.

Results and discussion
Summary of the AS11 mutant developmental and oil phenotypes The AS11 mutant line was about one week behind WT in bolting and entering the generative phase and thus, under our growing conditions, AS11 seed set was also delayed to four weeks post-anthesis instead of three, as typically observed in WT. For comparison, we studied another DGAT1 mutant, which we designated A7, which is a homozygous SALK line (Salk 039456) with a T-DNA insertion in the last exon of the same DGAT1 gene (At2g19450). A7 shows a developmental delay similar to that exhibited by AS11 (Additional file 1: Figure S1).
Using protein fractions prepared from WT and AS11 mid-developing siliques containing embryos at middevelopment (Stages 6-8 as designated in the "Methods" section) we were able to determine the relative changes of TAG assembly activity in the mutant line. TAG synthesis capacity was measured in WT and AS11 lines with 14 C-labeled diolein and unlabeled oleoyl-CoA as co-substrates; the 14 C-labeled triolein product was measured by radio-HPLC as described previously [26]. As shown in Figure 1, by stages 7-8 there was a 35% decrease in the acylation of radiolabeled DAG in AS11, a finding which was strongly correlated with the ca 30% reduction in oil content in mature AS11 seed [6,10].
Heterologous expression of mutated Arabidopsis DGAT1s from both A7 and AS11 in Yeast Strain H1246MATa The altered fatty acid and low TAG phenotype in AS11 seed raised questions as to how the AS11 mutant still manages to make 65-70% of WT levels of seed oil. Because AS11 has a reduced TAG phenotype, it was essential to determine whether the DGAT1 in AS11 was merely mutated and exhibited reduced activity as we initially suggested [6], or whether it is, in fact, non-functional. This has not heretofore been confirmed. The importance of doing so will become apparent in the PCR assessment of the dgat1 transcript in both AS11 and in some genetic crosses performed in the current study, and discussed below. The cDNA from the AS11 (with a 81 bp in-frame repeat insertion of the second exon) was cloned into a yeast expression vector pYES2.1 under the control of the galactose-inducible GAL1 promoter, and the construct was used to transform a yeast mutant strain H1246MATa, which lacks all four genes, ARE1, ARE2, DGAT1 and LRO1, which were found to contribute to TAG synthesis [27]. This yeast mutant has been successfully used in studies of DGAT function wherein WT and mutated forms of plant DGAT1s have been expressed and thereby function or lack thereof in the plant inferred [28][29][30][31][32]. H1246MATa yeast cells harboring an empty pYES2.1 vector plasmid or transformed with WT DGAT1 cDNA were used as a negative and positive controls, respectively ( Figure 2). A western blot of the microsomal membrane fractions isolated from the induced yeast H1246MATa cells (Figure 2A) showed that both the mutated AS11 and A7 and WT DGAT1 proteins (calculated M r of 62.2 kD, 53.9 kD and 58.9 kD, respectively) were stably expressed. In the pYES plasmid-only transformant, there was neither detectable DGAT1 protein, nor non-specific cross-reactivity of the antibody with non-DGAT1 proteins of a similar size. The western showed that the AS11 DGAT1 was slightly larger (by 27 amino acids, probably due to the exon 2 repeat) than that of WT, while for the A7 DGAT1, the largest possible stable transcript (truncated at the start of T-DNA insertion) encoded a protein that was significantly smaller than WT. However, as shown in Figure 2B, the AS11 DGAT1 could not compensate for the inability to produce TAG in this yeast quadruple mutant. Equally, when we assayed the transformed yeast microsomal protein fractions in vitro for DGAT activity using unlabeled oleoyl-CoA (18:1) as an acyl donor, and 14 C-sn-1,2 diolein (18:1) as acceptor, enzyme activity was not detected in the yeast strain harboring the mutated DGAT1 cDNA from the AS11 mutant and empty control pYES2.1 vector, but was found in the positive WT control. This indicated, perhaps not unexpectedly, but for the first time, that the mutated dgat1 from AS11 is nonfunctional. Equally, the truncated DGAT1 from the A7 mutant (SALK-039456) was shown to be non-functional when expressed in H1246MATa cells ( Figure 2B).
Thus our results strongly support that DGAT1 is not functioning in the developing seed of the AS11 mutant line. Clearly, the (radiolabeled) TAG formation observed in protein fractions from AS11 was coming from another path and not via reduced DGAT1 catalysis.
Given the importance of PDAT1 in oil biosynthesis in Arabidopsis as we earlier defined [17], and combining this new information with the 14

AS11 microarray and qRT-PCR analyses
These cumulative findings prompted us to examine a broader inventory of gene transcripts/encoded proteins that may be involved in regulating lipid biosynthesis in AS11 when DGAT1 activity is compromised, compared to their corresponding expression pattern in WT. Thus we performed a microarray analysis of gene expression in mid-developing seeds of mutant AS11 and WT Arabidopsis. Based on selected probable lipid assemblyrelated transcripts [33] differentially expressed through the microarray study, we complemented this with a semi-quantitative qRT-PCR analysis and compared the % change in AS11 transcript relative to WT, for each. While neither method is truly quantitative, the qualitative trends in each study were highly consistent ( Figure  3). Some general observations from these combined gene expression studies follow: In AS11, FAE1 is down-regulated, FADs 2 &3 are up-regulated; this is consistent with the AS11 acyl composition profile (reduced 20:1 and elevated 18:3). DGAT1 expression was not significantly affected. Interestingly, LPCAT2 was up-regulated by an average of 65%, while LPCAT1 expression was indifferent. PDAT2 was only marginally affected, but PDAT1 expression was up-regulated by an average of 62% compared to WT.
Given the critical role of PDAT1 in embryo development and TAG deposition in both pollen and seeds [17], we were interested in the relative differences in LPCAT1 and LPCAT2 expression between the AS11 mutant and WT. We hypothesized that an acyl-CoAdependent LPCAT may be critical to maintain the PC pool as one of the co-substrates for PDAT1-catalyzed TAG synthesis, particularly in the absence of a functional DGAT1, and performed a series of genetic and metabolic studies to examine these relationships. RT-PCR Characterization of the lpcat1, lpcat2 and dgat1 (AS11) mutants In addition to the WT and AS11 lines, we selected accession lines with putative mutations in LPCAT1 and LPCAT2, and characterized each of the individual mutants using RT-PCR to confirm the transcript phenotype of each line. As shown in Figure 4, in the putative lpcat1 mutant the ca 1100 bp transcript was absent, but it was amplified in the control (WT) sample, indicating that the mutant At1g12640 is indeed a knockout for the LPCAT1 gene. Similarly as shown in Figure 5B, lane 2, the band amplified for the LPCAT2 transcript is absent in the putative lpcat2 mutant, but present in the WT and AS11 mutant (lanes 1 and 3, respectively) confirming that line At1g63050 is a knockout for the LPCAT2 gene. As expected, both the WT and the lpcat2 mutant samples are positive for the WT DGAT1 amplicon ( Figure 5A, lanes 1 and 3, respectively). As shown in Figure 5A, lane 3, the AS11 mutant dgat1 transcript is present, encoding the inframe exon 2, 27-amino acid repeat, the amplified fragment being 81 bp larger than that found in the WT and in the lpcat2 mutant. However, as demonstrated in Figure 2 and discussed above, though present, we have very strong evidence from yeast expression studies that the mutant dgat1 encoded by this transcript is non-functional.

Metabolic studies
We performed in vitro LPCAT assays in a time course incubation of protein fractions from AS11 and WT middeveloping seed, in the presence of sn-1 palmitoyl-3 H LPC + 14 C-18:1-CoA and followed the 3 H and 14 C labeling patterns in PC. As shown in Figure 6A the 3 H accumulation pattern indicated that tritiated PC was rapidly synthesized from 3 H LPC at a rate that was 3.5fold higher in AS11 than WT within 20 min. Based on the proportion of sn-2 14 C oleoyl moieties incorporated into PC, the LPCAT activity was consistently 40-60% higher in AS11 at all time points ( Figure 6B). Equally, the proportion of 3 H in LPC concomitantly decreased in AS11 relative to WT over this period (data not shown). This indicated that in AS11, the LPCAT activity was strongly enhanced relative to WT. We also performed an in vivo feeding study wherein we supplied 14 C acetate to bolted stems containing pods with mid-developing seeds of both AS11 and WT plants, and then analyzed the label patterns in various lipid fractions in the mid-developing seeds after a chase period of 7, 24 and 30 h. Two major differences were immediately obvious (Figure 7): AS11 showed a higher relative incorporation of 14 C into PC and a lower relative incorporation of 14 C into TAG over the time course. These trends were entirely consistent with an elevated LPCAT activity and reduced TAG synthesis in AS11 as shown in the in vitro LPCAT ( Figure 6) and DGAT1 (Figure 1) assays.
Based on these trends we were confident that in addition to DGAT1 and PDAT1, LPCAT plays a significant role in TAG biosynthesis, and postulated that it is important not only for membrane development and acyl turnover therein, but also to replenish the supply of PC for PDAT1-catalyzed TAG biosynthesis in the developing seed. To resolve this question we needed to study what occurs when both DGAT1 and LPCAT expression are co-disrupted.

Genetic crosses
It had previously been shown, following expression in a yeast lca1Δ mutant background, that both Arabidopsis LPCAT1 (At1g12640) and LPCAT2 (At1g63050) are functional LPCATs. In the presence of sn-1 [ 14 C] 16:0-LPC and 18:1-CoA (the same combination of substrates used in the current study), expressed LPCAT1 and LPCAT2 proteins had activities of 12 nmol/min/mg/mg protein and 18 nmol/min/mg protein, respectively [34].
As confirmed by RT-PCR (discussed above), lpact1 (At1g12640) and lpcat2 (At1g63050) are T-DNA insertion knockouts for their respective genes, while AS11 encodes a mutated and non-functional ( Figure 2) dgat1. To examine the relationship between these two LPCATs and DGAT1 in TAG biosynthesis, we performed crosses of the AS11 dgat1 with the lpcat1 or the lpcat2 T-DNA insertion knock-out mutants and characterized the hemizygous/homozygous and double knockout seed oil profiles.

dgat1 × lpcat1
The lpcat1 mutant is devoid of any significant oil phenotype compared to its null segregant or to WT (Figure   [17] showed that without DGAT1, the PDAT1 route contributed 75% oil in AS11; thus even without combined contributions from [DGAT1 + LPCAT1], the PDAT1 route could still provide up to 70% oil in the developing AS11 seeds. In other words, the co-incident loss of LPCAT1 with DGAT1 reduced the capacity for PDAT1-catalyzed oil synthesis by only 5%. While oil content in the dgat1/lpcat1 H/H line is reduced, it is interesting to note that with respect to seed weight, there is a significant increase (by ca 6%) relative to the dgat1/lpcat1 null/null control ( Figure 9). Thus, while there was no penalty in this regard, further study is needed to ascertain which components of the seed e.g. protein or cellulosic biomass of the seed coat) account for this observation.
The fatty acid profile from the lpcat1 mutant is identical to that of its null segregant ( Figure 10). The AS11 (dgat1) x lpcat1 double knockout lines show an AS11like profile (low 18:1 and 20:1; high 18:3); the 18:3 proportion is about 2-5% lower than in AS11 alone. The latter is not inconsistent with the fact that LPCATs are involved in the shuttling of 18:1 moieties to the PC backbone for desaturation by FADs 2 and 3.

dgat1 × lpcat2
The lpcat2 mutant does not have a significant oil phenotype compared to its null segregant-i.e. it is similar to WT ( Figure 11). The fatty acid composition is essentially identical to WT and to that exhibited by the LPCAT2 null segregant (Figure 12). When we performed the AS11(dgat1) x lpcat2 crosses, the dgat1/lpcat2 He/H shows only a small reduction in oil content which suggests that when DGAT1 is partially expressed (He), and LPCAT2 is knocked out (H), there is a reduced capacity for compensation in TAG biosynthesis, probably  because PDAT is insufficiently "fueled" with its co-substrate, PC. In these lines there is a severe penalty on seed development; there are many gaps in developing siliques due to non-fertilized ovules (Additional file 1: Figure S2A,B). Relative to the dgat1/lpcat2 double null, in the lpcat2 H and dgat1/lpcat2 He/H lines, seed weight is increased by about 10% (Figure 13). However, the dgat1/lpcat2 H/ H lines show an strong decrease in seed weight even beyond that observed in the dgat1 mutant alone or when the double mutant genotype has dgat1 in the homozygous condition, Taken together with the seed weight increases cited above for the lpcat1 mutant studies, these findings make it very clear that both the lpcat1 or lpcat2 mutations have a complicated effect on seed weight which may or may not be related to changes in oil content.
The acyl composition of the dgat1/lpcat2 He/H lines was again essentially like the LPCAT2 NS ( Figure 12). In contrast, the dgat1/lpcat2 H/He lines have an AS11-like reduction in oil content which is about 70% of that found in the WT and the LPCAT2 null segregant (Figure 11). Interestingly, there is no developmental penalty in this heterozygous combination and seed set is normal as in the AS11 background, provided expression of LPCAT2 is in the heterozygous state (Additional file 1: Figure S3). Not unexpectedly, the acyl composition of the oil in these lines is consistent with the AS11 profile ( Figure 12). The results were quite different in the AS11 (dgat1) x lpcat2 double homozygous mutant. Shown in Figure 14 is a PCR confirmation of the genotype of putative H/H lines # 6-3-10-19, # 6-3-20-1 compared to WT, AS11 and lpcat2 lines. AS11, and dgat1/lpcat2 H/H lines # 6-3-10-19 and # 6-3-20-1 have the expected 147 bp insert indicating they are homozygous for the insertion mutation resulting in a non-functional DGAT1 (as shown in Figure 2, above), while the WT and lpcat2 samples do not. T-DNA screening was done on the samples of the same lines to test for the presence of LPCAT2. Samples were amplified using [left primer + right primer] or with [right primer + left border 2] primer. The lpcat2 and the dgat1/lpcat2 H/H lines # 6-3-10-19 and # 6-3-20-1 have a homozygous knockout of the LPCAT2 gene. qRT-PCR also confirmed the transcript profile in these two lines for both dgat1 ( Figure 5A, lanes 4 and 5) and lpcat2, (Figure 5B, lanes 4 and 5).). Biochemically, in the double homozygous mutant (dgat1/lpcat2 H/H) there is a severe penalty on TAG synthesis; there is a 65% reduction in oil content relative to WT: Mature seeds have only about 30% oil content relative to the WT and the LPCAT2 null segregant proportions ( Figure 11). Without DGAT1 (AS11), recent studies suggested that PDAT1 catalyzes the synthesis of essentially all the oil which accumulates at 70-75% of that found in WT [17]. Without both LPCAT2 and DGAT1, PDAT1 is able to synthesize only about 35% of WT oil levels. Thus, the loss of LPCAT2 strongly reduces the capacity of PDAT1 to synthesize oil. This suggests a strong link between PDAT1 and its supply of PC as a co-substrate in TAG synthesis; clearly LPCAT2 plays a strong role in supplying this PC. When the DGAT1 mutation is homozygous, the AS11 acyl profile dominates when the LPCAT2 mutation is in the He condition ( Figure 12). However, in both the lpcat2 homozygous mutant and in the double mutant dgat1/lpcat2 H/H lines, the 20:1 and 18:1 are slightly increased while 18:2 is decreased (Figure 12). It is significant to note that Shen et al. [34] demonstrated that an Arabisdopsis line transformed with an LPCAT1-LPCAT2 double RNAi construct, partially repressing the expression of both genes, also exhibited a 3% increase in the proportion of 20:1, a 1.5% increase in 18:1 and a concomitant 3% decrease in 18:2 in seed TAGs compared to WT. While the current phenotype is difficult to interpret without further study, these findings will be important in future analyses of LPCATs in terms of substrate selectivity and potential channeling of fatty acids into oil. Plant growth and development is also delayed in the double homozygous mutant (dgat1/lpcat2 H/H), as it is in AS11: the seedlings are small and slow to develop ( Figure 15A). Once transferred to soil, bolting to enter the generative phase is delayed in dgat1/lpcat2 H/H ( Figure 15B), but is normal in dgat1/lpcat2 H/He (Figure 15C). Seed development is poor, as in the dgat1/ lpcat2 He/H lines; there are many non-fertilized ovules in developing siliques resulting in seed which is visibly reduced compared to that normally observed in WT or LPCAT2 null segregant (NS). Thus, it is clear that some level of LPCAT2 expression is necessary for normal seed development, especially in an AS11 background. However, future investigations of this preliminary observation, including scoring of developmental phenotypes and a microscopic study of pollen viability, will be required to confirm a disturbed pollen functionality as a result of the double mutation in the dgat1/lpcat2 H/H lines.
Summarizing the mutant crossing experiments, in lpcat2 × dgat1, oil synthesis, vegetative growth and seed set are severely penalized. Compared to the relative oil content in dgat1 alone (75% of WT), there is only a 5% further relative decrease in dgat1/lpcat1 oil (70% of WT) which suggests that LPCAT2 is the major determinant and can almost fully compensate for the loss of LPCAT1. In contrast, in the co-incident absence of LPCAT2 and DGAT1 (lpcat2 × dgat1), relative oil content drops to 30% of WT, which suggests that LPCAT1 can only partially compensate for the loss of LPCAT2 with respect to oil. Metabolic and microarray studies in AS11 alone support this hypothesis as LPCAT2 expression and LPCAT activity are up-regulated in the absence of DGAT1. In addition, the acyl profiles indicate that any altered oil composition phenotype observed in the lpcat2 x dgat1 is predominantly influenced by the dgat1 lesion. Collectively, the data from the crosses of AS11 with lpcat1 or lpcat2 mutants strongly suggest that in an AS11 background, while both LPCAT2 and LPCAT1 can contribute to PDAT1-catalyzed TAG synthesis, it is LPCAT2 up-regulation (and not LPCAT1), which is primarily responsible for assisting in PDAT1-catalyzed TAG biosynthesis by supplying an enhanced pool of PC as co-substrate to transfer sn-2 moieties to the sn-3 position of the enlarged AS11 DAG pool. Nonetheless, while LPCAT2 is the most important for maintaining PDAT1-catalyzed TAG synthesis in the AS11 background, the partial over-lapping function of LPCATs 1 and 2 identified in this study require further investigation to determine their relative roles in oil synthesis when DGAT1 is functional in a WT background. The recent discovery of a phosphatidylcholine: diacylglycerol cholinephosphotransferase (PDCT), encoded by the Arabidopsis ROD1 gene, indicates that the issue of DAG-PC interconversion during TAG biosynthesis is considerably more complex than previously thought [35]. The PDCT enzyme catalyzes transfer of the phosphocholine headgroup from PC to DAG and is thought to be a major reaction for the transfer of 18:1 into PC for desaturation by FAD2 and FAD3, as well as enabling the reverse transfer of 18:2 and 18:3 into the TAG synthesis pathway. In the rod1 mutant, the accumulation of PUFAs in TAGs is reduced by 40%. However, in AS11 where 18:3 is strongly increased, and notably, shows a strong stereospecific shift to the sn-3 position of TAGs [6], it is unclear whether there is a strong involvement of PDCT; were it so, one would expect a more even distribution of 18:3 at the sn-2 and sn-3 positions. However, the strong re-distribution of 18 dgat1/lpcat1 N/N Control)
proportions to the sn-3 and away from the sn-2 position in AS11 suggests that the major shift in positional composition has to do with PDAT1 catalysis. Thus there are still many questions to address with respect to the roles ROD1 and other PC-related metabolic steps such as CPTase might play in oil synthesis in Arabidopsis and in AS11 in particular.

Conclusions
Based on the cumulative results of these studies, in Figure 16 we summarize our current hypothesis regarding the TAG biosynthesis pathway in WT vs. that observed when DGAT1 is compromised as in AS11. When both DGAT1 and PDAT1 are operating they both contribute to TAG synthesis, but the consensus of recent literature supports PDAT1 as a major route in Arabidopsis. When DGAT1 is eliminated, the PDAT1 pathway produces up to 75% of WT oil levels [17]. PDAT1 acylates the sn-3 position of DAGs which are accumulating (in the absence of DGAT1), to give TAG. The acyl composition of AS11 with highly enhanced PUFAs and reduced VLCFAs at the sn-3 position [6] supports the donation of acyl groups, especially 18:3, from the sn-2 position of PC to the sn-3 position of DAG as the major route for TAG synthesis in Arabidopsis. This adjustment in the TAG assembly route in AS11 is aided by enhanced LPCAT2 activity to supply the additional PC required by PDAT1. Given these findings and the implicit plasticity of Arabidopsis TAG assembly mechanisms to overcome critical bottlenecks, it will be interesting to determine the relative roles played by DGAT1 and [PDAT1 + LPCAT2] during TAG assembly in higher oilseeds (e.g. canola, soybean, sunflower, safflower, flax).  Table S3, with genotypes designated on the × axis. dgat1/lpcat1 null for both mutations # 3-3-9; lpcat1 homozygous mutant N123480-6; AS11 dgat1 homozygous mutant CS3861; dgat1/lpcat1 homozygous for both mutations #3-3-12. The values are reported as % (wt/wt) and are the average of triplicate determinations ± S.D.

Lipid analyses
Preparation of total lipid extracts (TLEs) and study of lipid classes, determination of oil content and acyl composition in seeds of WT and the AS11, lpcat1 and lpcat 2 mutant lines and progeny from crosses were performed as described previously [26,36]. In all cases, the data represent the averages of three-five determinations.

Preparation of Arabidopsis protein fractions
In general, enzyme preparations were made from 200 Arabidopsis siliques of AS11 and WT containing midgreen developing seeds (stages 7-8 as defined above), and immediately powdered with liquid nitrogen in a mortar and pestle. Grinding medium (100 mM HEPES-KOH, pH 7.4 containing 0.32 M Sucrose, 1 mM EDTA, and 1 mM dithiothreitol; 8 mL/50 siliques) along with 65 mg polyvinylpolypyrrolidone was immediately added, and grinding continued on ice for 5 min. The slurried cell-free homogenate was filtered through two layers of Miracloth (Calbiochem, La Jolla, CA), centrifuged at 3,000 × g for 5 min, the pellet discarded and the supernatant re-centrifuged at 15,000 × g for 30 min. The supernatant was re-centrifuged at 100,000 × g for 1 h and the resultant pellet was resuspended in 2 mL of grinding medium, probe-sonicated on ice for 30 s and protein concentrations were determined using BioRad™ reagent based on the method of Bradford [37]. Protein concentrations were normalized to the same value for WT and AS11 in each experiment.

Immunodetection
The yeast cell lysates were run on a 10% Tris-HCl SDS-PAGE gel, the proteins were then transferred to a nitrocellulose membrane (Nitrobind, Thermo Fisher Scientific, Waltham, MA). The membrane was blocked in PBST (phosphate buffered saline containing 0.5% Tween 20) containing 4% skim milk for 60 min, and then incubated with the primary antibody, Anti-Xpress (epitopetagged) antibody (Invitrogen, San Francisco, CA) diluted to 1:5000 with PBST containing 2% skim milk, for 60 min. The membrane was submitted to three washes with PBST followed by three washes with PBS to remove any unbound antibody. Next, the membrane was incubated with a goat anti-mouse IgG peroxidase antibody (Sigma-Aldrich, A2554, St. Louis, MO), diluted to 1:5000 with PBST containing 2% skim milk, for 60 min. The membrane was washed three times with PBST followed by three times with PBS, then the proteins were detected using the Amersham ECL Plus Western Blotting Detection Kit (GE Healthcare Life Sciences) with a 10 min exposure time.

LPCAT assays
LPCAT assays were conducted at pH 7.4, with shaking at 100 rev/min in a water bath at 30°C in a 20-minute time course. Assay mixtures (500 μL final volume) contained 300 μg protein normalized as described above, 90 mM HEPES-NaOH, 0.5 mM ATP, 0.5 mM CoASH, 1 mM MgCl 2 in the presence of 6 μM L-α-

In vivo feeding experiments
Twenty bolted stems with mid-developing siliques of both WT and AS11 plants were harvested under water and then immediately placed in a solution containing 1 μCi of 14 C sodium acetate in 100 μL of water and incubated at room temperature in a fume hood. Once most of this solution was taken up, the plants were supplied with equal aliquots of distilled water during the chase period. At 0, 7, 24 and 30 h, 5 bolted stems were harvested, siliques counted and weighed and then ground with a polytron and a TLE extraction performed as described by Taylor et al. [26]. The lipid extract was resuspended in 1 mL CHCl 3 :MeOH 2:1. 10 μL of the solution was counted and the remainder spotted on TLC plates, 3D TLC performed, radiolabeled spots resolved on a radio-TLC scanner, silica bands scraped and radioactivity measured with a scintillation counter.
qRT-PCR comparison of AS11 and WT expression of key lipid genes Triplicate biological samples of total RNA were extracted using the method of Wang and Vodkin [39] from Arabidopsis siliques of AS11 and WT containing mid-green developing seeds (stages 7-8 as defined above) and used for qRT-PCR [40]. One microgram of total RNA was reverse-transcribed using SuperScript ® II Reverse Transcriptase (Invitrogen) according to the manufacturer's protocol. The resulting cDNA was then amplified by PCR using Taq DNA polymerase (Invitrogen). PCR conditions comprised an initial cycle of 94°C for 3 min; followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min; then 72°C for 10 min to complete the reaction. Concentrations for the individual samples within each group (AS11 and WT) were normalized using 18S rRNA levels. The gene-specific primers (listed in Additional file 2: Table S1) were designed according to the target gene sequences as annotated in Genbank™ (http://www.ncbi.nlm.nih.gov/ genbank/). Amplified RT-PCR products were resolved by electrophoresis on 1% agarose gels and gel photos taken. The densities of the PCR bands were analyzed using ImageJ image analysis software (National Institutes of Health, USA http://rsbweb.nih.gov/nih-image/) and quantified relative to the "housekeeping gene" 18S rRNA transcript signal. Then the ratio of AS11:WT gene expression level was calculated. Specifically, for each gene of interest, first we calculated the ratio of AS11 density/18S rRNA density (= "AS11 ratio") and WT density/18S rRNA density (= "WT ratio"). The difference in specific gene expression in AS11 vs. WT, was then calculated as ("AS11 ratio")/("WT ratio") = V. If V >1.0, then the gene is up-regulated in AS11 relative to WT; if V <1.0, then the gene is down-regulated in AS11 relative to WT. The % relative increase or decrease is then = ([V × 100]-100).

Affymetrix microarray analysis
The same triplicate biological samples prepared for the qRT-PCR study (above) were sub-sampled for microarray analysis. Affymetrix microarray hybridizations using the Ath1 whole genome array, containing probe sets representing~22,800 genes, were performed using three biological replicate samples for each genotype. Labeling, hybridization, and scanning were performed by the Affymetrix Gene Chip Facility at the University of Toronto (http://www.csb.utoronto.ca/resources/facilities/affymetrix-genechip). Data analysis was performed using Gene-Spring™ software version 7.2. To identify key lipid genes that were differentially expressed between the two genotypes, a per-gene normalization was applied to the values and a parametric test was performed. Genes that exhibited a false discovery rate of p < 0.05 and passed the minimum signal and fold-change threshold were determined to be differentially expressed. Reported candidate genes were selected by comparing those listed by Beisson et al. [33] with our microarray and qRT-PCR data. The "fold change" was converted to % change. For the group of genes detailed in Figure 3, the final changes ranged from -60% to + 150%.

Confirmation of lpcat1 and lpcat2 T-DNA insertion mutation lines
Arabidopsis insertion mutation lines [41] SALK_123480 putative for LPCAT1 (At1g12640) and SAIL_357_H01 putative for LPCAT2 (At1g63050) were identified in the Salk Institute T-DNA insertion library database (http:// signal.salk.edu/cgi-bin/tdnaexpress), and seeds were obtained from the Nottingham Arabidopsis Stock Centre (University of Nottingham, UK). According to annotation in the database SALK_123480 contains a T-DNA insertion in the middle of the 7th exon of the LPCAT1 gene and SAIL_357_H01 a T-DNA insertion in the 8th exon of the LPCAT2 gene. Individual plants homozygous for a T-DNA insertion in each the LPCAT1 or LPCAT2 genes were identified by PCR screening using primers SALK_123480LP, SALK_123480RP and SALK_LBb1 (Additional file 2: Table S2) for LPCAT1 and SAIL_357_H01LP, SAIL_357_H01RP and SAIL_LB2 (Additional file 2: Table S2) for LPCAT2. Individual plants from each mutant line lacking a T-DNA insertion in the LPCAT1 or LPCAT2 genes (null segregants) were also identified in the same PCR primer set. Annotation of lines from each set of crosses is as designated in Additional file 2: Table S3.
Transcript analyses of lpcat1, lpcat2, dgat1 mutant lines and crosses lpcat1 (See Figure 4) Total RNA was isolated from developing siliques from control and SALK_12340 T-DNA insertion mutation plants using CONCERT Plant RNA Reagent (Invitrogen, San Francisco, CA) according to the technical specifications given by the manufacturer. The RNA was then DNAse-treated with TURBO™ DNase (Ambion, Huntingdon, UK)) and used (1 mg) for synthesize first strand cDNA with Maxima First Strand cDNA Synthesis Kit (Fermentas.Thermo Fisher Scientific, St. Leon-Rot, Germany). The cDNA was amplified by PCR using Taq DNA Polymerase (Sigma-Aldrich, St. Louis, MO) and the primers SALK_12640-1 s and SALK_12640-6r as listed in Additional file 2: Table S2. PCR conditions comprised an initial cycle of 94°C for 1 min; followed by 40 cycles of 94°C for 20 s, 64°C for 20 s, and 72°C for 35 s; then 72°C for 8 min to complete the reaction. PCR samples were separated on 1% agarose gel using λ-DNA digested with PstI as size marker.
lpcat2 and AS11 dgat1 (See Figure 5) Total RNA was isolated from developing siliques from WT control, AS11(dgat1), SAIL_357_H01 insertion mutation (lpcat2) and AS11/lpcat2 plants by the method of Wang and Vodkin [39]. One microgram of total RNA was reverse-transcribed using SuperScript ® II Reverse Transcriptase (Invitrogen) according to the manufacturer's protocol. The resulting cDNA was then amplified by PCR using Taq DNA polymerase (Invitrogen). PCR conditions comprised an initial cycle of 94°C for 3 min; followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min; then 72°C for 10 min to complete the reaction. Concentrations for the individual samples within each group (AS11 and WT) were normalized using 18S rRNA levels. Gene specific primers spanning the T-DNA insertional region were used to check for gene transcription. For DGAT1 transcript analysis in lines of WT, the AS11 (dgat1) mutant, lpcat2 mutant and crosses of lpcat2 with dgat1, the primer pairs were TAG1-mut-primerA and TAG1-mut-primerB. For LPCAT2 transcript analysis in these same lines, the primer pairs were LPCAT2-F2 and LPCAT2-R2 as listed in Additional file 2: Table S2.
Creating double mutants of dgat1 × lpcat1 and dgat1 × lpcat2 Crosses between the AS11 mutant and lpcat1 and lpcat2 T-DNA insertion mutation lines, respectively, were made and F 1 plants heterozygous for AS11 and the insertion mutations were identified by PCR using primers listed in Additional file 2: Table S2. F 2 seeds segregating for the mutations were planted and screened by PCR. All PCR screening was done as above for LPCAT1 and LPCAT2 and for the AS11 mutation vs WT using primers dgat1-mut-primerA and dgat1-mut-primerB (Additional file 2: Table S2), as designed by Zou et al. [7]. Identification of individuals homozygous for both the AS11 mutation and lpcat2 insertion mutation was only possible after growing a segregating F 2 seed population (homozygous for AS11 and heterozygous for lpcat2 mutant) on agar media containing a 1/3 strength MS and 1% sucrose (Additional file 2: Table 3).
Additional file 2: Table S1 Primers used for semi-quantitative RT-PCR conducted in this study. Table S2. Primers used to screen the AS11 dgat1 EMS mutant and lpcat1 and lpcat2 T-DNA insertion mutant lines by PCR. Table S3. Details on crosses performed to develop lines homozygous for both dgat1 and lpcat1 mutations or for both dgat1 and lpcat2 mutations.