Patatin-related phospholipase pPLAIIIδ increases seed oil content with long-chain fatty acids in Arabidopsis.

The release of fatty acids from membrane lipids has been implicated in various metabolic and physiological processes, but in many cases, the enzymes involved and their functions in plants remain unclear. Patatin-related phospholipase As (pPLAs) constitute a major family of acyl-hydrolyzing enzymes in plants. Here, we show that pPLAIIIδ promotes the production of triacylglycerols with 20- and 22-carbon fatty acids in Arabidopsis (Arabidopsis thaliana). Of the four pPLAIIIs (α, β, γ, δ), only pPLAIIIδ gene knockout results in a decrease in seed oil content, and pPLAIIIδ is most highly expressed in developing embryos. The overexpression of pPLAIIIδ increases the content of triacylglycerol and 20- and 22-carbon fatty acids in seeds with a corresponding decrease in 18-carbon fatty acids. Several genes in the glycerolipid biosynthetic pathways are up-regulated in pPLAIIIδ-overexpressing siliques. pPLAIIIδ hydrolyzes phosphatidylcholine and also acyl-coenzyme A to release fatty acids. pPLAIIIδ-overexpressing plants have a lower level, whereas pPLAIIIδ knockout plants have a higher level, of acyl-coenzyme A than the wild type. Whereas seed yield decreases in transgenic plants that ubiquitously overexpress pPLAIIIδ, seed-specific overexpression of pPLAIIIδ increases seed oil content without any detrimental effect on overall seed yield. These results indicate that pPLAIIIδ-mediated phospholipid turnover plays a role in fatty acid remodeling and glycerolipid production.

Lipids play essential structural, metabolic, and regulatory roles in plant growth, development, and stress responses. In addition, plant lipids are a major source of food and renewable materials for various industrial and energy applications (Dyer et al., 2008;Hayden et al., 2011;Rogalski and Carrer, 2011;. Substantial progress has been made toward a basic understanding of the biochemical reactions of lipid biosynthesis in plants, but many fundamental questions about lipid metabolism remain unanswered (Weselake et al., 2009;Chapman and Ohlrogge, 2012). Recent results suggest that the metabolism of phosphatidylcholine (PC) plays multiple important roles in glycerolipid production. An increasing line of research shows that storage lipid triacylglycerols (TAGs) are not synthesized primarily via the Kennedy pathway but are derived from PC through acyl editing (Bates et al., 2009Tjellström et al., 2012). PC is also hypothesized to be involved in the trafficking of fatty acids from the plastid to the endoplasmic reticulum (ER), where glycerolipids, including TAG, are assembled (Wang and Benning, 2012). It is proposed that plastidial fatty acids are transferred to lysophosphatidylcholine (LPC) to form PC, which serves as a substrate for fatty acid desaturation and modification. While the importance of PC metabolism in TAG production is clear, the specific enzymes involved in PC turnover are not well elucidated Chapman and Ohlrogge, 2012), and the impact of PC turnover on TAG accumulation remains to be determined.
Phospholipase A (PLA) hydrolyzes PC to produce LPC and a free fatty acid (FFA). This reaction has been implicated in various cellular functions, including the production of lipid mediators, carbon partitioning, and cell elongation. Patatin-containing PLA (pPLA) is a major family of intracellular acyl-hydrolyzing enzymes in plants (Scherer et al., 2010;Murakami et al., 2011). The 10-gene pPLA family in Arabidopsis (Arabidopsis thaliana) is grouped into three subfamilies, pPLAI, pPLAII (a, b, g, d, «), and pPLAIII (a, b, g, d). pPLAI has been shown to contribute to resistance to Botrytis cinerea, possibly by mediating the basal levels of jasmonate production (Yang et al., 2007), whereas pPLAIIa negatively modulates both plant response to bacterial pathogens (La Camera et al., 2005) and oxylipin production (Yang et al., 2012). pPLAIIb impacts root elongation during phosphate deficiency, and pPLAIIg and pPLAIId have been implicated in involvement in auxin responses (Rietz et al., 2004(Rietz et al., , 2010. Activation tagging of pPLAIIId and overexpression (OE) of pPLAIIIb resulted in decreased cell elongation and stunted growth (Huang et al., 2001;Li et al., 2011). These results indicate that the pPLA family plays important, diverse roles in plant growth and stress responses, but their role in seed oil production is not known.
One enigma from recent genomic analysis of Arabidopsis has been that there are as many genes annotated as being involved in lipid catabolism as there are in lipid synthesis (Li-Beisson et al., 2010). While the functions for many genes involved in lipid biosynthesis have been documented, little is known about the role of lipid-hydrolyzing enzymes in lipid metabolism Figure 1. Alterations of pPLAIIId change Arabidopsis seed oil content. A, Transcript levels of pPLAIIIa, pPLAIIIg, and pPLAIIId in 2-week-old rosettes. The RNA levels were determined by realtime PCR and normalized to the level of the wild type (WT). Values are means 6 SE (n = 3). B, Seed oil content in T-DNA insertion mutants of pPLAIIIa-KO, pPLAIIIb-KO, pPLAIIIg-KO, and pPLAIIId-KO. Values are means 6 SE (n = 3). L Significantly lower at P , 0.05 compared with the wild type, based on Student's t test. C, Transcript levels of pPLAIIId in wildtype, pPLAIIId KO, OE, and COM plants. pPLAIIId-OE1, pPLAIIId-OE2, pPLAIIId-COM1, and pPLAIIId-COM2 are independent lines of the T3 generation of pPLAIIId-OE and pPLAIIId-COM. The RNA levels were determined by real-time PCR and normalized in comparison with UBQ10. Values are means 6 SE (n = 3). D, Seed oil content in pPLAIIId-OE and pPLAIIId-COM T3 seeds. pPLAIIId expression in OE lines was under the control of the cauliflower mosaic virus 35S promoter, while in the COM lines, it was under the control of its own promoter. Values are means 6 SE (n = 3). H Significantly higher and L significantly lower, each at P , 0.05 compared with the wild type, based on Student's t test. E, Immunoblotting of GFP-tagged pPLAIIId in Arabidopsis. Leaf proteins extracted from plants were separated by 8% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and visualized with alkaline phosphatase conjugated to a secondary anti-mouse antibody after blotting with GFP antibody. Five independent T3 lines of pPLAIIId-OE mutants were examined. and oil production. A recent study compared the transcriptomes of mesocarp from oil palm (Elaeis guineensis 'Jacq') and date palm (Phoenix dactylifera) that accumulate approximately 90% and 1% oil, respectively (Bourgis et al., 2011). The mRNA level of key genes in fatty acid synthesis in oil palm mesocarp is 2-to 44-fold higher than in date palm. The mRNA level of palm pPLAIIIb is 22-fold higher in oil palm compared with date palm mesocarp (Bourgis et al., 2011), but the role for pPLAIII in oil accumulation remains to be determined. Patatin-related enzymes typically contain a catalytic center with the esterase box GXSXG and other specific motifs including a catalytic dyad motif, which typically contains Asp-Gly-Gly (Scherer et al., 2010). The pPLAIII subfamily differs from pPLAI and pPLAII in that it does not contain the canonical esterase GXSXG motif but instead has the sequence GXGXG (Scherer et al., 2010). Our recent analysis of pPLAIIIb shows that pPLAIIIb hydrolyzes PC to produce LPC and FFAs (Li et al., 2011). Moreover, OE of pPLAIIIb increases membrane glycerolipid content in vegetative tissues, whereas its gene knockout (KO) has the opposite effect. These observations prompted us to determine the role of pPLAIIIs in seed oil production. Here, we show that pPLAIIId promotes TAG production with increased accumulation of long-chain fatty acids in Arabidopsis seeds.

RESULTS pPLAIIId Increases Seed Oil Content
To investigate the function of pPLAIIIs in seed oil production, we isolated transfer DNA (T-DNA) insertional KO mutants for all four pPLAIIIs (Supplemental Fig. S1). The T-DNA insertion sites of pPLAIIIa and pPLAIIIb are in the first exon, while the insertion sites of pPLAIIIg and pPLAIIId locate in the 59 untranslated region (Supplemental Fig. S1B). All of these insertional mutants have a negligible level of transcript as measured by real-time PCR of pPLAIIIa, pPLAIIIg, and pPLAIIId (Fig. 1A). The loss of pPLAIIIb expression in pPLAIIIb-KO was described previously (Li et al., 2011). However, only the pPLAIIId-KO seeds, not the other pPLAIIIs, displayed a significant change in oil content compared with wild-type seeds; the oil contents of pPLAIIId-KO and wild-type seeds were 33% and 35.5% of the seed weight, respectively (Fig. 1B). To confirm the effect of pPLAIIId on seed oil production, we genetically complemented the KO by transferring pPLAIIId with its native promoter and terminator sequences into the KO mutant (designated as COM; Supplemental Fig. S1C). Expression of pPLAIIId in the COM lines was restored to the wild-type level (Fig. 1C), and the oil content in COM seeds was the same as that of the wild type (Fig. 1D).  Analysis of mRNA accumulation patterns for pPLAIIIs in seeds indicates that pPLAIIIa, pPLAIIIb, and pPLAIIIg were expressed in tissues that do not accumulate large amounts of TAG in developing seeds (Supplemental Fig. S2, A-E). In mature green seeds, pPLAIIIg was expressed mostly in seed coat, pPLAIIIb mostly in chalazal seed coat, and pPLAIIIa mostly in seed coats and peripheral endosperm (Supplemental Fig. S2, C-E). In contrast, pPLAIIId was expressed in developing radicle and in cotyledons, the major storage tissue for seed oil in Arabidopsis (Supplemental Fig. S2F). The mRNA accumulation pattern of the pPLAIII genes is consistent with a pPLAIIId-specific effect on seed oil content; thus, further analysis was focused on pPLAIIId.
To further investigate pPLAIIId function, we produced multiple OE Arabidopsis lines by placing pPLAIIId under the control of cauliflower mosaic virus 35S promoter (35S::pPLAIIId-OE; Supplemental Fig. S1C). The mRNA level of pPLAIIId was increased substantially in OE over wild-type plants (Fig. 1C). The presence of the introduced GFP-tagged pPLAIIId was detected by immunoblotting with a GFP antibody (Fig. 1E). Seed oil content in two OE lines was approximately 40.5%, which was 5% higher than that of the wild type (35.5%; Fig. 1D). Taken together, these data indicate that pPLAIIId plays a positive role in seed oil accumulation.
pPLAIIId Increases 20-Carbon Fatty Acid Content at the Expense of 18-Carbon Fatty Acids The fatty acid composition was significantly altered in pPLAIIId-KO and 35S::pPLAIIId-OE seeds ( Fig. 2A). The levels of 18-carbon fatty acids tended to increase in KO and decrease in OE seeds compared with the wild type. For example, 18:1 was increased by 10% in KO but decreased 6% in OE1. Conversely, the amounts of 20-carbon fatty acids 20:0 and 20:2 were decreased by Figure 3. pPLAIIId promotes the accumulation of 20C-containing TAG species. A, Normalized mass spectra (as a percentage of the total) from TAG species with the indicated fatty acyl combinations in wild-type, KO, and OE seeds. The fatty acids making up each molecular species are indicated, but no positional specificity is implied. The TAG species shown on the left side of the dashed lines contain only 16-and 18-carbon chains, while those on the right include one or more 20-carbon chains. Values are means 6 SE (n = 5). H Significantly higher and L significantly lower, each at P , 0.05 compared with the wild type, based on Student's t test. B, Normalized mass spectra (as a percentage of the wild type) from TAGs, grouped by total acyl carbons, were classified as C50, C52, C54, C56, C58, and C60. The major components in C50, C52, and C54 are 18C fatty acylcontaining TAGs, while C56, C58, and C60 are TAGs containing one or more 20C fatty acyl-containing TAGs. Values are means 6 SE (n = 5). H Significantly higher and L significantly lower, each at P , 0.05 compared with the wild type, based on Student's t test.
12% and 12% in KO, while 20:0 and 20:1 were increased by 12% and 15% in OE lines, compared with the wild type. The 22-carbon species, 22:1, showed a trend similar to the 20-carbon species. The ratio of 20to 18-carbon fatty acids was decreased by 10% in KO and increased by 19% in OE compared with the wild type (Fig. 2B). The fatty acid composition in COM seeds was similar to that of wild-type seeds (Fig. 2B). Thus, the increased mRNA level of pPLAIIId promoted the accumulation of longer chain fatty acids at the expense of 18-carbon fatty acids, 18:1 and 18:2, whereas pPLAIIId-KO decreased the production of longer chains with increased accumulation of 18-carbon fatty acids.
Fatty acids in Arabidopsis seeds occur primarily in esterified form in TAGs. TAGs include many different molecular species with varied carbon chain lengths and degrees of unsaturation in the three acyl chains. Three acyl chains in TAG are not randomly distributed. Since pPLAIIId affects 18:1 and 20:1 accumulation in TAG, we wondered if pPLAIIId alters the distribution of three acyl chains and thus produces some unique TAG molecule species. Therefore, we analyzed the TAG species in wild-type, KO, and OE seeds by electrospray ionization-tandem mass spectrometry (ESI-MS/MS). The major fatty acyl chain carbon numbers (C) in seed TAGs are 16C, 18C, and 20C, and the major TAG species have total C of C50 (e.g. 16-16-18), C52 (e.g. 16-18-18), C54 (e.g. 18-18-18), C56 (e.g. 18-18-20), C58 (e.g. 18-20-20), and C60 (e.g. 20-20-20; Fig. 3). The percentages of C50, C52, and C54 TAG species in total TAGs, as indicated by their relative mass spectral signals, tended to be higher in KO and lower in OE mutants when compared with the wild type, while the levels of C56, C58, and C60 TAG species were changed in the opposite manner in KO and OE lines of pPLAIIId ( Fig. 3; Supplemental Fig. S3). For example, the percentages of some 16C-and 18Ccontaining TAGs (16:0-16:0-18:3, 16:0-18:1-18:3, 18:2-18:2-18:3) were significantly lower in OE mutants than in the wild type (Fig. 3A). While certain TAG species could not be quantified individually, and thus their compositional percentages were expressed in combination, the percentages of 20C-containing TAGs and TAG groups tended to be or were significantly lower in KO and higher in OE mutants compared with the wild type (Fig. 3A). Overall, the relative amounts of C50, C52, and C54 TAGs tended to be lower, while C56, C58, and C60 TAGs tended to be higher in OE mutant seeds compared with the wild type (Fig. 3B). Measurement of 113 additional TAG species and eight TAG species groups confirmed the trend for the percentages of 18C-containing TAGs to be lower and the 20C-containing TAGs to be higher in OE lines compared with the wild type (Supplemental Fig. S3, A-E). Taken together, these data indicate that pPLAIIId promotes the accumulation of 20C-containing TAG species.

pPLAIIId-OE Increases the Transcript Levels of Genes in TAG and PC Synthesis
To gain insight into how pPLAIIId facilitates TAG accumulation and modification, we measured the mRNA levels of selected genes in TAG and PC synthesis and metabolism in developing Arabidopsis siliques (Fig. 4). In the Kennedy pathway of TAG biosynthesis, glycerol-3-P is sequentially acylated by glycerol phosphate acyltransferase (GPAT) and lysophosphatidic acid acyltransferase (LPAT), followed by phosphatidic acid phosphohydrolase (PAH) and diacylglycerol acyltransferase (DGAT). The transcript levels for the genes in the Kennedy pathway, including GPAT, LPAT2, LPAT3, PAH, and DGAT1, were increased 2-to 5-fold in pPLAIIId-OE lines (Fig. 4A). By comparison, mRNA levels of both DGAT2 and LPAT5 were the same in wild-type, pPLAIIId-KO, and 35S:: pPLAIIId-OE siliques (Fig. 4). Phospholipid:diacylglycerol acyltransferase (PDAT) catalyzes the transfer of a fatty acid from PC to diacylglycerol (DAG) to produce TAG. The mRNA level of PDAT1 was increased by almost 3-fold in OE lines compared with the wild type (Fig. 4B).
In the Kennedy pathway of PC biosynthesis, choline phosphate:CTP cytidylyltransferase (CCT) synthesizes CDP-choline using CTP and phosphocholine, and aminoalcohol-phosphotransferase (AAPT) catalyzes the last step of PC synthesis by transferring phosphocholine to DAG from CDP-choline. There are two CCTs and AAPTs in Arabidopsis. Compared with the wild type, the mRNA levels of CCT2 and AAPT1 were increased almost by 10-fold, whereas the increase in CCT1 and AAPT2 was about 2-fold in pPLAIIId-OE siliques (Fig.  4B). The mRNA abundance of Lysophosphatidylcholine: acyl-coenzyme A acyltransferase1 (LPCAT1) was also increased 3-fold in OE lines (Fig. 4B). These data demonstrate that OE of pPLAIIId increases the mRNA levels of genes involved in TAG and PC synthesis. On the other hand, in KO siliques, the mRNA levels for the lipid-metabolizing genes were not significantly different from that of the wild type, even though the mRNA levels for several of these genes tended to be lower than that of the wild type (Fig. 4). These results suggest that the loss of pPLAIIId may be partially compensated for by other pPLAs.
pPLAIIId Hydrolyzes PC and Acyl-CoA and Affects Acyl-CoA Levels in Arabidopsis pPLAIIId is more distantly related to the other three pPLAIIIs than they are to each other (Supplemental Fig. S1A). pPLAIIId has an Asp in the Asp-Gly-Gly catalytic dyad motif, similar to pPLAs in the other groups, whereas in pPLAIIIb and pPLAIIIg, the Asp is replaced by Gly (Li et al., 2011). To determine the enzymatic function of pPLAIIId, we expressed 63Histagged pPLAIIId in Escherichia coli and purified it to near homogeneity (Fig. 5A). The PC-hydrolyzing activity of pPLAIIId was examined because PC is the most abundant phospholipid and serves as a key intermediate for TAG synthesis. Incubation of pPLAIIId with 16:0-18:2-PC resulted in the production of FFA and LPC. pPLAIIId hydrolysis at the sn-1 position produces 16:0-FFA and 18:2-LPC ( Fig. 5B), whereas pPLAIIId hydrolysis at the sn-2 position produces 18:2-FFA and 16:0-LPC (Fig. 5C). The production of 18:2-FFA was approximately 5-fold more than that of 16:0-FFA, and correspondingly, much more 16:0-LPC was formed than 18:2-LPC. These data indicate that pPLAIIId hydrolyzes PC at both the sn-1 and sn-2 positions and that pPLAIIId preferentially releases 18:2 from the sn-2 position.
In addition, we determined whether pPLAIIId could hydrolyze acyl-CoA, because our previous study showed that another pPLAIII member, pPLAIIIb, has thioesterase activity (Li et al., 2011). Incubation of pPLAIIId with 18:3-CoA resulted in the steady production of 18:3-FFA with increasing reaction time (Fig. 6A), indicating that pPLAIIId possesses a thioesterase activity. We then determined whether the alterations of pPLAIIId expression impacted the acyl-CoA content in Arabidopsis. In siliques that included developing seeds with active storage lipid biosynthesis, the level of total acyl-CoA was 19% higher in KO and 18% lower in OE mutants than in the wild type (Fig. 6B). The major acyl-CoA species are 18:3-CoA and 18:2-CoA, followed by 16:0-CoA. The levels of 18:1-CoA, 18:2-CoA, and 18:3-CoA were significantly higher in KO, and the levels of 16:0-CoA and 18:2-CoA were significantly lower in OE, than in wild-type siliques (Fig.  6C). These data are consistent with pPLAIIId functioning as an acyl-CoA thioesterase activity in vivo. To determine its subcellular association, a GFPtagged pPLAIIId was expressed in Arabidopsis, and the green fluorescence signal of pPLAIIId-GFP was mostly detected on the inner cell boundary of leaf epidermal cells (Fig. 7A). Plasmolysis by applying saline solution to the roots showed that the GFP signal in root epidermal cells was shrinking along with the plasma membrane (Fig. 7B). To further analyze the intracellular association, total leaf proteins were fractionated into cytosolic and microsomal fractions. All pPLAIIId-GFP was associated with the microsomal membranes but not with cytosol (Fig. 7C). The microsomal proteins were further partitioned into the plasma membrane and intracellular membrane fractions. Approximately 80% of pPLAIIId-GFP was associated with the plasma membrane, whereas 20% was associated with intracellular membranes based on the intensity of the protein bands (Fig. 7C). These data indicate that pPLAIIId is associated with both the plasma and intracellular membranes.

Seed-Specific OE of pPLAIIId Increases Oil Content
The increased oil content in seeds raises the question of whether increased pPLAIIId expression can be used to increase seed oil production. However, constitutive OE of pPLAIIId resulted in a decrease in plant height and overall seed yield (Fig. 8, A and B). The seed yield per 35S::pPLAIIId-OE plants was approximately 50% of that of wild-type plants (Fig. 8B). To explore whether the improved oil content could be uncoupled from decreased seed production, we placed pPLAIIId under the control of the seed-specific promoter of soybean (Glycine max) b-conglycinin (CON::pPLAIIId; Supplemental Fig. S4A). The level of pPLAIIId expression in developing siliques was 25-fold higher in CON::pPLAIIId than that in the wild type (Fig. 8C). The presence of the pPLAIIId-GFP protein was detectable by visualizing the GFP fluorescence (Supplemental Fig. S4B). CON::pPLAIIId plant height and seed yield were comparable with the wild type (Fig. 8, A and B). In three CON::pPLAIIId lines tested, seed oil content was increased over wild-type seeds (39% versus 35%; Fig. 8D). While oil content per CON::pPLAIIId seed weight was lower than that per 35S::pPLAIIId seed weight (Figs. 1D and 8D), the overall seed oil production per CON::pPLAIIId plant was significantly higher than that per 35S::pPLAIIId, due to the higher seed yield per plant (Fig. 8B), and per wild-type plant, due to the increased oil content without change in seed yield (Fig. 8, B and D).
The seed-specific overexpression of pPLAIIId resulted in changes in fatty acid composition, and the changes in CON::pPLAIIId were similar to those in 35S::pPLAIIId seeds. The percentages of 18:1 and 18:2 were lower, while those of 20:0, 20:1, 20:1, and 22:1 were higher, in CON::pPLAIIId lines than in the wild type (Fig. 8E). The ratio of 20:1 to 18:1 was 30% higher in CON::pPLAIIId lines than in the wild type (Supplemental Fig. S4C), and the same pattern was observed when total 20-carbon fatty acids were compared with total 18-carbon fatty acids (Supplemental Fig. S4D). These results indicate that pPLAIIId affects TAG metabolism in the same manner regardless of the promoter used and that the use of seed-specific expression of pPLAIIId has the potential to be applied for increased seed oil production.

DISCUSSION
These data show that pPLAIIId positively impacts seed oil content. Whereas pPLAIIId-KO decreases seed oil content, pPLAIIId-OE, driven either by a constitutive or a seed-specific promoter, increases seed oil content. pPLAIIId hydrolyzes PC to generate FFA and LPC. pPLAIIId may accelerate acyl flux from the plastid to the ER and, therefore, enhance glycerolipid synthesis. Fatty acids in higher plants are synthesized exclusively in the plastid and have to be exported to the ER, where glycerolipids are synthesized (Fig. 9). Lipid trafficking between organelles is a fundamental, yet poorly understood, process in plants. In recent years, excellent progress has been made toward understanding lipid transport from the ER to the plastid for the synthesis of galactolipids (Wang et al., 2012b). Phosphatidic acid (PA) is imported into the plastid through a protein complex (Wang and Benning, 2012). In contrast, the metabolic and regulatory mechanisms by which fatty acids in the plastid are trafficked to the ER are unknown.
16:0 and 18:1 are two major fatty acids exported from the plastid in Arabidopsis (Pidkowich et al., 2007;Li-Beisson et al., 2010). FFAs are thought to be able to cross membrane bilayers through diffusion and possibly protein-mediated translocation (Wang and Benning, 2012). After reaching the plastid outer envelope, long-chain acyl-CoA synthetases convert these fatty acids to acyl-CoA. In the conventional Kennedy pathway, acyl-CoA is used for the sequential acylation of glycerol-3-P → LPA → PA → DAG → TAG (Fig. 9). However, kinetic labeling data indicate that fatty acids exported from the plastid are first incorporated into PC and then channeled to TAG in soybean embryos (Bates et al., 2009Bates and Browse, 2011). The presence of highly active LPCAT on the Arabidopsis plastid outer envelope membrane is consistent with the formation of PC using fatty acids from the plastids (Tjellström et al., 2012;Wang et al., 2012a). Recent data indicate that LPCAT1 and LPCAT2 catalyze the incorporation of fatty acids into PC in Arabidopsis seeds Wang et al., 2012a). However, knowledge is lacking about what enzyme produces LPC that impacts TAG synthesis. PDAT can transfer a fatty acid from PC to DAG to produce TAG and LPC, but its role in TAG production in seeds remains unclear (Chapman and Ohlrogge, 2012). pPLAIIId could be one of the enzymes hydrolyzing PC to produce an FFA and LPC that LPCAT uses to accept fatty acids from the plastid (Fig. 9). The combined activity of pPLAIII and LPCATs may modulate the rate of fatty acid trafficking from the plastid to the ER in Arabidopsis seeds.
Fatty acids, such as 18:1, released from PC by pPLAIIId may enter the acyl-CoA pool for elongation (Fig. 9). KO and OE of the pPLAIIId gene displayed opposite effects on the levels of 18:1 and 20:1 fatty acids in seed oil. Detailed profiling of TAG molecules also showed the opposite effects on the levels of 18:1containing and 20:1-containing TAGs by KO and OE of the pPLAIIId gene. In Arabidopsis, the major fatty acids exported from plastids to the ER are 16:0, 18:0, and 18:1. In the ER, 18:1 on PC is desaturated to 18:2 and 18:3 (Li-Beisson et al., 2010), whereas acyl-CoA is used for fatty acid elongation to form longer chain fatty acids, such as 20:1 (Joubès et al., 2008;Li-Beisson et al., 2010). The effect of pPLAIIId on fatty acid composition is distinctively different from that of the recently described PC:DAG cholinephosphotransferase that transfers phosphocholine from PC to DAG, and a mutation of PC:DAG cholinephosphotransferase decreases the 18:2 and 18:3 level in Arabidopsis seed TAG by 40% (Lu et al., 2009). Thus, the increased pPLAIIId expression may facilitate the release of 18:1 from PC for 20:1 production ( Fig. 9).
Compared with the wild type, OE of pPLAIIId had a lower acyl-CoA pool size in developing silique and higher seed oil content. The decrease in the acyl-CoA pool size could result from the thioesterase activity of pPLAIIId and/or increased PC turnover and TAG synthesis. The exchange of modified acyl groups between PC and the acyl-CoA pool requires extensive acyl editing cycles (Harwood, 1996). Through the acyl editing cycles, modified fatty acids enter the acyl-CoA pool to be utilized for glycerolipid synthesis, and acyl-CoA can be channeled into PC for further modification and directly for TAG production (Stymne and Stobart, 1984;Bafor et al., 1991;Bates et al., 2007Bates et al., , 2009. The inverse association between acyl-CoA pool and TAG contents could mean that the pPLAIIIdcatalyzed turnover of acyl-CoA and PC promotes seed oil accumulation. The enhanced mRNA level of genes, such as AAPT and CCT, in PC biosynthesis in pPLAIIId-OE plants indicates that increased pPLAIIId-mediated PC hydrolysis leads to an increase in PC biosynthesis and, thus, increased PC turnover. Meanwhile, RNA levels are higher for genes in the Kennedy pathway, such as GPAT, LPAT, PA phosphatase, and DGAT, in developing pPLAIIId-OE siliques. The increased transcript levels of glycerolipid-producing genes may be a feedforward stimulation by enhanced substrate supplies, as the increased pPLAIIId expression leads to elevated levels of FFAs and LPC. How the metabolic changes in FFAs and LPC are connected to the altered mRNA levels and potentially gene expression requires further investigation. In budding yeast (Saccharomyces cerevisiae), it has been shown that the transcriptional factor directly binds to PA, senses cellular PA levels, and regulates the expression of many genes involved in membrane lipid synthesis (Loewen et al., 2004). In addition, there is an increase in the mRNA level of LPCAT, which catalyzes the acylation of LPC using fatty acids from the plastid. This could mean an increase in fatty acid trafficking from the plastid to the ER, where glycerolipids are synthesized. Further studies are needed to determine the mechanism by which increased pPLAIIId expression promotes TAG production. Such investigation of how a lipidhydrolyzing enzyme, such as pPLAIIId, promotes lipid accumulation has the potential to better our understanding of lipid metabolism and accumulation.
In summary, our study shows that pPLAIIId hydrolyzes PC to generate FFA and LPC and that genetic alterations of pPLAIIId expression change seed oil content and fatty acid composition in Arabidopsis seeds. Our large-scale TAG species analysis reveals that pPLAIIId promotes the production of 20:1-TAG. We propose that pPLAIIId plays a role in fatty acyl flux from the plastid to the ER and/or PC fatty acyl remodeling for TAG synthesis. Furthermore, these results indicate that the use of seed-specific expression of pPLAIIId has the potential to improve seed oil production in crops.

Generation of pPLAIII KO, OE, and Complementation Plants
Arabidopsis (Arabidopsis thaliana) T-DNA insertional mutants for pPLAIIIa (Salk_040363), pPLAIIIb (Salk_057212), pPLAIIIg (Salk_088404), and pPLAIIId (Salk_029470) were identified from the Salk Arabidopsis T-DNA KO collection obtained from the Arabidopsis Biological Resource Center. The homozygous T-DNA insertion mutant for individual pPLAIIIs was verified by PCR-based screening using a T-DNA left border primer and gene-specific primers as listed in Supplemental Table S1. The isolation of pPLAIIIb-KO was reported previously (Li et al., 2011). The loss of gene transcripts in pPLAIII-KO was confirmed by real-time PCR. To generate the complementation lines (pPLAIIId-COM ), the genomic DNA sequence of pPLAIIId from the promoter region to the terminator region was cloned using two primers as listed in Supplemental Table S1 and fused into binary vector pEC291 for plant transformation.
To overexpress pPLAIIId, the genomic sequence of pPLAIIId was obtained by PCR using ecotype Columbia Arabidopsis genomic DNA as a template and primers listed in Supplemental Table S1. The genomic DNA was cloned into the pMDC83 vector before the GFP-His coding sequence. The expression of pPLAIIId was under the control of the 35S cauliflower mosaic virus promoter or the promoter of soybean (Glycine max) b-conglycinin. The sequences of the fusion constructs were verified by sequencing before they were introduced into the Agrobacterium tumefaciens strain C58C1. Ecotype Columbia Arabidopsis plants were transformed, and transgenic plants were screened and confirmed by antibiotic selection and PCR. Over 15 independent transgenic lines were obtained (pPLAIIId-OE) with similar plant stature. Five independent lines of pPLAIIId-OE were further verified by immunoblotting with anti-GFP antibody.

RNA Extraction and Real-Time PCR
Real-time PCR was performed as described previously (Li et al., 2006(Li et al., , 2011. Briefly, total RNA was extracted from different tissues using the cetyltrimethylammonium bromide method (Stewart and Via, 1993). DNA contamination in RNA samples was removed with RNase-free DNase. An iScript kit (Bio-Rad) was used to synthesize complementary DNA (cDNA) from isolated RNA template by reverse transcription. The MyiQ sequence detection system (Bio-Rad) was used to detect products during quantitative real-time PCR by monitoring SYBR Green fluorescent labeling of double-stranded DNA. Efficiency was normalized to a control gene, UBQ10. The real-time PCR primers are listed in Supplemental Table S2. The data were expressed as means 6 SE (n = 3 replicates). PCR conditions were as follows: one cycle of 95°C for 1 min; 40 cycles of DNA melting at 95°C for 30 s, DNA annealing at 55°C for 30 s, and DNA extension at 72°C for 30 s; and final extension of DNA at 72°C for 10 min.

Analysis of Fatty Acid Composition and Oil Content
Ten milligrams of Arabidopsis seeds was placed in glass tubes with Teflonlined screw caps, and 1.5 mL of 5% (v/v) H 2 SO 4 in methanol with 0.2% butylated hydroxytoluene was added. The samples were incubated for 1 h at 90°C for oil extraction and transmethylation. Fatty acid methyl esters (FAMEs) were extracted with hexane. FAMEs were quantified using gas chromatography on a SUPELCOWAX-10 (0.25 mm 3 30 m) column with helium as a carrier gas at 20 mL min 21 and detection by flame ionization. The oven temperature was maintained at 170°C for 1 min and then ramped to 210°C at 3°C min 21 . FAMEs from TAG were identified by comparing their retention times with FAMEs in a standard mixture. Heptadecanoic acid (17:0) was used as the internal standard to quantify the amounts of individual fatty acids. Fatty acid composition is expressed in weight percentage.
pPLAIIId Cloning and Protein Purification from Escherichia coli The full-length cDNA of pPLAIIId was obtained by PCR using an Arabidopsis cDNA library as a template and a pair of primers as listed in Supplemental Table S1. The cDNA was cloned into the pET28a vector before the 63His coding sequence. The 63His fusion construct was sequenced and confirmed to be error free before it was introduced into E. coli strain Rosetta (DE3; Amersham Biosciences). The bacteria were grown to an optical density at 600 nm of 0.7 and induced with 0.1 mM isopropyl 1-thio-b-D-galactopyranoside for 16 h at 16°C. The pPLAIIId-63His fusion protein was purified as described previously (Pappan et al., 2004). Briefly, the bacterial pellet was resuspended in STE buffer containing 1 mg mL 21 lysozyme (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 1 mM EDTA). The samples were kept on ice for 30 min. Dithiothreitol and N-laurylsarcosine (Sarkosyl) were added to a final concentration of 5 mM and 1.5% (w/v), respectively. The suspension was vortexed and sonicated on ice for 5 min. After centrifugation at 10,000g for 20 min, the supernatant was transferred to a new tube. Triton X-100 was added to a final concentration of 4% (v/v), and 63His agarose beads were added (10%, w/v). The solution was gently rotated at 25°C for 1 h. The fusion proteins bound to agarose beads were washed with 20 volumes of STE buffer. The amount of purified protein was measured with a protein assay kit (Bio-Rad).

Enzyme Assays
Phospholipids and acyl-CoAs were purchased from Avanti Polar Lipids. PC or 18:3-CoA in chloroform was dried under a nitrogen stream and emulsified in reaction buffer (25 mM HEPES, pH 7.5, 10 mM CaCl 2 , and 10 mM MgCl 2 ) by vortexing, followed by 5 min of sonication on ice. Acyl-hydrolyzing activities were assayed in a reaction mixture containing 25 mM HEPES, pH 7.5, 10 mM CaCl 2 , 10 mM MgCl 2 , and 60 mmol PC as substrate. Ten micrograms of purified protein was added to the mixture in a final volume of 500 mL. The reaction was incubated at 30°C for the indicated times and stopped by adding 2 mL of chloroform:methanol (2:1, v/v) and 500 mL of 25 mM LiCl. After vortexing and separation by centrifugation, the lower phase was transferred to a new glass tube. The upper phase was extracted twice more by adding 1 mL of chloroform each time, and the three lower phases were combined. Lipid internal standards were added, and lipid quantification was performed by mass spectrometry as described below.

Lipid Quantification
In in vitro enzyme assays, lipids were extracted for analysis as described previously (Li et al., 2011). Twenty microliters of lipid sample was combined with 340 mL of chloroform and 840 mL of chloroform:methanol:300 mM ammonium acetate in water (300:665:35). FFAs were determined by ESI-MS on an electrospray ionization triple quadruple mass spectrometer (API4000; Applied Biosystems), using the deuterated internal standard (7,7,8,8-d4-16:0 fatty acid; Sigma-Aldrich), by scanning in negative ion mode from mass-to-charge ratio 200 to 350 (Li et al., 2011). LPC was determined with the same instrument as described previously (Li et al., 2011). Plants for acyl-CoA measurement were grown in growth chambers with a 12-h-light/12-h-dark cycle, at 23°C/21°C, 50% humidity, and 200 mmol m 22 s 21 light intensity, and watered with fertilizer once per week. Acyl-CoAs were extracted and analyzed by liquid chromatography-ESI-MS/MS as described previously (Magnes et al., 2005;Han et al., 2010). TAG molecular species were analyzed by ESI-MS/MS using neutral loss scan modes (Lee et al., 2011(Lee et al., , 2012. The TAG analysis is described in detail in Supplemental Materials and Methods S1.

Microscopy Imaging and Subcellular Fractionation
The subcellular location of GFP-tagged protein was determined using a Zeiss LSM 510 confocal microscope equipped with a 340 differential interference contrast, 1.2-numerical aperture water-immersion lens, with excitation using the 488-nm line of an argon gas laser and a 500-to 550-nm band-pass emission filter. Plasmolysis in primary root cells was induced by immersing roots in 0.5 M NaCl for 1, 3, and 5 min. Developing seeds from Arabidopsis siliques were imaged using a Nikon Eclipse 800 wide-field microscope and a 360 differential interference contrast, 1.2-numerical aperture objective, with mercury lamp excitation and a 492/18 band pass excitation filter and a 535/40 band emission filter. For subcellular fractionation, proteins were extracted from leaves of 4-week-old plants using buffer (30 mM HEPES, pH 7.5, 400 mM NaCl, and 1 mM phenylmethanesulfonyl fluoride), followed by centrifugation at 6,000g for 10 min. The supernatant was centrifuged at 100,000g for 60 min. The resulting supernatant is referred to as the soluble cytosol fraction, and the pellet is referred to as the microsomal fraction. The microsomal fraction was separated further into the plasma and intracellular membrane fractions, using two-phase partitioning as described previously (Fan et al., 1999).

SDS-PAGE and Immunoblotting
Leaf samples, each weighing approximately 1 g, were harvested and ground in 3 mL of buffer of 30 mM HEPES, pH 7.5, 400 mM NaCl, 1.0 mM phenylmethanesulfonyl fluoride, and 1 mM dithiothreitol. Proteins were separated by 8% SDS-PAGE and transferred to a polyvinylidene difluoride membrane. The membrane was visualized with alkaline phosphatase conjugated to a secondary anti-mouse antibody after blotting with GFP antibody.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Generation of KO, OE, and complementation mutants of pPLAIIIs.
Supplemental Figure S2. RNA accumulation patterns of four pPLAIII genes in developing Arabidopsis seeds.
Supplemental Figure S3. pPLAIIId promotes increased levels of 20C fatty acyl-containing TAG over 18C fatty acyl-containing TAG in Arabidopsis seeds, as determined by mass spectral analysis.
Supplemental Figure S4. Seed-specific OE of pPLAIIId in Arabidopsis.
Supplemental Table S1. PCR primers for mutant screening and molecular cloning.
Supplemental Table S2. Real-time PCR primers for quantitative measurement of transcript levels.
Supplemental Materials and Methods S1. Mass spectral analysis of TAG. (C) Constructs for generating overexpression and complementation mutants of pPLAIIIδ. In the pPLAIIIδ overexpressors (pPLAIIIδ-OE), the Arabidopsis pPLAIIIδ gene was driven by the cauliflower mosaic virus 35S promoter and tagged on the C-terminus with green fluorescence protein and 6xHistidine. In the pPLAIIIδ complementation lines (pPLAIIIδ-COM), Arabidopsis pPLAIIIδ genomic DNA sequence, cloned from promoter to terminator, was transferred into T-DNA of the T-DNA insertional knockout mutant of pPLAIIIδ. The empty boxes and filled boxes denote exons and untranslated regions, respectively. Ten independent lines of pPLAIIIδ-COM were generated and they were indistinguishable from wild-type. Fifteen independent lines of pPLAIIIδ-OE were generated; plans of these OE lines were  (% of total TAG mas ) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )  (A) Constructs for generating seed specific expression mutants of pPLAIIIδ. Arabidopsis pPLAIIIδ genomic DNA cloned from start codon to stop codon (stop codon removed) was driven by soybean β-conglycinin promoter and tagged on C-terminus by green fluorescence protein and 6xHistidine. The resulting transgenic Arabidopsis were designated as CON::pPLAIIIδ. Ten independent lines of T3 generation mutants were obtained The growth of the CON::pPLAIIIδ mutants was comparable with wild type obtained. The growth of the CON::pPLAIIIδ mutants was comparable with wild-type. (B) Detection of the green fluorescence signal in developing seeds of CON::pPLAIIIδ mutants. Developing seeds from Arabidopsis siliques were imaged using a Nikon Eclipse 800 widefield microscope and a X60 differential interference contrast, 1.2-numerical aperture objective, with mercury lamp excitation and a 492/18 BP excitation filter and a 535/40 B emission filter.  (A) Constructs for generating seed specific expression mutants of pPLAIIIδ. Arabidopsis pPLAIIIδ genomic DNA cloned from start codon to stop codon (stop codon removed) was driven by soybean β-conglycinin promoter and tagged on C-terminus by green fluorescence protein and 6xHistidine. The resulting transgenic Arabidopsis were designated as CON::pPLAIIIδ. Ten independent lines of T3 generation mutants were obtained. The growth of the CON::pPLAIIIδ mutants was comparable with wild-type.

Supplemental
(B) Detection of the green fluorescence signal in developing seeds of CON::pPLAIIIδ mutants.
Developing seeds from Arabidopsis siliques were imaged using a Nikon Eclipse 800 widefield microscope and a X60 differential interference contrast, 1.2-numerical aperture objective, with mercury lamp excitation and a 492/18 BP excitation filter and a 535/40 B emission filter.
3 (C) Ratio of the level of fatty acids with 20 carbons over 18 carbons in seeds of WT and

Supplemental Method 1. Mass Spectral Analysis of TAG
An automated, direct-infusion electrospray ionization-tandem mass spectrometry approach was used for TAG analysis. A precise amount of internal standard (0.5 nmol tri17:1-TAG, Avanti Polar Lipids) was added to around 25 mg of dry Arabidopsis seeds, which were ground with mortar and pestle in 1.0 mL of chloroform/methanol (2:1). The mixture were extracted with shaking for 1 h at room temperature and centrifuged to pellet the debris. nitrogen in the collision cell, was +20 V, declustering potential was +100 V, entrance potential 4 was +14 V, and exit potential was +14 V. Sixty continuum scans were averaged in MCA mode (multiple channel analyzers).
For all analyses the collision gas pressure was set on "low", and the mass analyzers were adjusted to a resolution of 0.7 u full width at half height. The source temperature (heated nebulizer) was 100°C, the interface heater was on, +5.5 kV was applied to the electrospray capillary, the curtain gas was set at 20 (arbitrary units), and the two ion source gases were set at 45 (arbitrary units).
For TAG analyses, the background of each spectrum was subtracted, the data were smoothed, and peak areas integrated using a custom script and Applied Biosystems Analyst software. Peaks corresponding to the target lipids in these spectra were identified, and the data were corrected for A+2 isotopic overlap (based on the m/z of the charged fragments) within each spectra. Signals were also corrected for isotopic overlap across spectra, based on the A+2 overlaps and masses of the neutral fragments. All signals for each sample were normalized to the signal of the internal standard. A sample containing internal standard alone, run through the same series of scans, was used to correct for chemical or instrumental noise: amounts of each target lipid detected in the ''internal standards-only''sample were subtracted from the molar amounts of each target lipid calculated from the plant lipid spectra. The ''internal standards-only'' spectra were used to correct the data from the following five samples run on the instrument.
The corrected data from all fatty acyl (NL) scans for each TAG species, as defined by m/z, which corresponds to total acyl carbons: total double bonds (e.g. 52:3), were used to calculate the amount of each individual TAG species. As described by Han and Gross, formulas were developed to assign particular signals from the NL scans to particular TAGs (Han and Gross, 2001). Once values for all TAGs were calculated, the amount of each TAG was expressed as a percentage of the total values for all TAG species. Because there is variation in ionization efficiency among acyl glycerol species with different fatty acyl groups (Han and Gross, 2001) and, here, no response factors for individual species were determined, the values are not directly proportional to the TAG content of each species. However, the amounts of particular TAG species can be meaningfully compared across samples.