Compartmentation of Triacylglycerol Accumulation in Plants

Abstract

Triacylglycerols from plants, familiar to most people as vegetable oils, supply 25% of dietary calories to the developed world and are increasingly a source for renewable biomaterials and fuels. Demand for vegetable oils will double by 2030, which can be met only by increased oil production. Triacylglycerol synthesis is accomplished through the coordinate action of multiple pathways in multiple subcellular compartments. Recent information has revealed an underappreciated complexity in pathways for synthesis and accumulation of this important energy-rich class of molecules.

Regulation of Fatty Acid Supply by Plastids

Plant fatty acid (FA)² synthesis differs from almost all other eukaryotes in two fundamental features. First, unlike the cytosolic location in other kingdoms, FAs for triacylglycerol (TAG) and membrane synthesis are produced in the plastid compartment of plant cells: chloroplasts in green tissues and proplastids (or leucoplasts) in non-green tissues. Second, the plant FA synthase (FAS) is a dissociable complex with separate proteins for the acyl carrier protein (ACP) and each enzyme (1). After assembly of C₁₆ and C₁₈ acyl chains by FAS and desaturation of C_18:0 to C_18:1, FAs destined for TAG assembly are released from ACP in the plastid stroma by chain-terminating acyl-ACP thioesterases. Two classes of thioesterases designated FATA and FATB are responsible for hydrolysis of unsaturated and saturated acyl-ACPs, respectively, and thus determine in large part the chain length and saturated FA content of plant oils (2). The FA products of FATA and FATB are activated to CoA before export to the endoplasmic reticulum (ER). The subsequent reactions of TAG synthesis belong to the so-called “eukaryotic” pathway of glycerolipid synthesis, which occurs outside the plastid (3, 4). An overview of the compartmentalization of FA supply for TAG is shown in Fig. 1.

FIGURE 1.

Overview of pathways and compartmentation of FA supply for TAG synthesis in plants. Different oilseed species utilize different pathways for producing acetyl-CoA precursors and cofactors for FA synthesis. Not all are shown. In most oilseeds that accumulate large amounts of TAG, sucrose is imported into the cytosol and provides the ultimate source of carbohydrate for conversion to acetyl-CoA and then to oil. FAs are synthesized in the plastid stroma from acetyl-CoA and malonyl-CoA by a type II dissociable FAS complex. Hexose phosphates from sucrose metabolism can enter glycolysis in the cytosol or plastid to produce pyruvate (Pyr). PDH provides the direct source for acetyl-CoA via a reaction that also provides enough NADH to supply half of the reductant needed for FA synthesis. Malonyl-CoA is formed from acetyl-CoA, ATP, and HCO₃⁻ by ACCase. HCO₃⁻ is equilibrated with CO₂ in the plastid stroma by carbonic anhydrase (CA). The conversion of acetyl-CoA and malonyl-CoA to FAs occurs through a series of condensation, dehydration, and reduction reactions. Three isoforms of ketoacyl-ACP synthase (KAS) cooperate to elongate acyl chains on an ACP platform to C₁₆ or C₁₈. KASIII initiates FA synthesis by condensation of acetyl-CoA and malonyl-ACP, KASI continues elongation to C_16:0, and KASII provides the final elongation step from C_16:0 to C_18:0. 18:0 ACP is desaturated by a stearoyl-ACP desaturase (SAD) to generate 18:1 ACP, and two acyl-ACP thioesterases (FATA and FATB) release newly synthesized 16:0 FA, 18:0 FA, and 18:1 FA in the plastid stroma. Depending upon the plant species and the presence of thylakoid membranes, requisite NADPH and ATP for FAS can be generated by photosystems (PS). The CO₂ from PDH can be recaptured by Rubisco, which can result in 20% higher yields of acetyl-CoA from sucrose. In non-green seeds, NADPH can also be generated by the oxidative pentose phosphate pathway (OPPP). Ribulose 1,5-bisphosphate (RuBP) is generated from hexose phosphate by the non-oxidative pentose phosphate pathway (PPP). ATP required for acetyl-CoA carboxylase can be produced by pyruvate kinase or imported into plastids from the cytosol (not shown). The products of FAS are activated to CoA by LACS in the chloroplast outer envelope before export to the ER for TAG synthesis. Export mechanisms are uncertain (dashed arrows) and may be via acyl-CoA (possibly involving acyl-CoA-binding proteins (ACBP)). Alternatively, after LPCAT activity at the plastid envelope, phospholipid (PC) may move to the ER through direct membrane contact sites (stromules or plastid-associated ER membranes) or by some as yet unknown mechanism. Glycerolipids are assembled in the ER (see details in Fig. 2), and TAG emerges as lipid droplets (LD) that are stored in the cytosol. PEP, phosphoenolpyruvate; Ru5P, ribulose 5-phosphate; 3PGA, 3-phosphoglycerate.

FA production by plastids can limit TAG accumulation in seeds (5, 6), so increasing flux through FA biosynthesis may perhaps have the single greatest influence on the amount of TAG produced in plant tissues. As in bacteria, fungi, and animals, both in vitro and in vivo evidence indicates that acetyl-CoA carboxylase (ACCase) is a key rate-determining step that controls FA biosynthesis. ACCase activity is under complex regulation by light, phosphorylation, thioredoxin, PII protein, and product feedback control (7, 8). Of course, the flux of carbon to FA synthesis can have multiple regulatory steps, which may explain why efforts to increase seed oil by up-regulating ACCase were only modestly successful (9).

Transcriptional regulation of the production of FA for TAG biosynthesis is most directly controlled by the WRINKLED1 transcription factor (7, 10, 11). Arabidopsis wri1 mutants are reduced by 80% in seed oil, and overexpression of WRI1 can increase oil content in seeds of several plants. Targets of WRI1 include ACCase, many enzymes of FAS, and key enzymes and transporters that provide pyruvate and acetyl-CoA in the plastid. Thus, WRI1 can be considered as a “master regulator” that controls transcription of almost all key enzymes that convert sucrose to FA. The importance of WRI1 was confirmed when seed oil was increased by 30% in field trials of maize that overexpress WRI1 (12), an increase that would be valued at $2 billion if extended to all maize production in the United States. In addition to controlling oil production in seeds, recent evidence indicates that WRI1 is likely a major factor responsible for the extremely high oil content (up to 90% of tissue weight) produced by oil palm mesocarp. Transcriptional profiling of oil palm mesocarp revealed >50-fold higher WRI1 expression levels compared with date palm mesocarp, a closely related species that contains no oil (13). Consistent with data in developing seeds (14), genes encoding machinery for FA biosynthesis and pyruvate supply are up-regulated substantially in oil palm (an average of >13-fold) (15). The direct precursor of carbon for FA synthesis is pyruvate in most oil-synthesizing tissues, and a plastidial pyruvate dehydrogenase (PDH) supplies acetyl-CoA for ACCase. Transcriptional profiling in oil palm identified subunits of PDH with 50-fold higher transcript levels compared with date palm, reinforcing transcriptional regulation as a major means of influencing FA supply for TAG synthesis (13).

Evidence suggests that the energetics and coordination of photosynthesis, respiration, and carbon partitioning also must be taken into consideration to influence the amount of TAG accumulated in plant tissues. A strict light dependence of FA synthesis in leaf chloroplasts has been known for years (16), but the relevance to TAG accumulation in oilseeds was revealed only recently at the metabolic level. Even low levels of light penetrating to developing embryos provide sufficient reductant and ATP to power FA biosynthesis (17). Moreover, the activity of ribulose-bisphosphate carboxylase/oxygenase (Rubisco) in developing seeds functions not for net photosynthetic carbon fixation but rather to recycle CO₂ released by PDH and other decarboxylation reactions. This action of Rubisco ultimately increases the supply of carbon for FA biosynthesis by 20% (18, 19), resulting in higher conversion efficiency of carbohydrate to oil. Other work recently pointed to additional factors that can influence FA supply from plastids for TAG synthesis. Overexpression of an Arabidopsis hemoglobin isoform appeared to improve oxygenation of developing seeds and the energetics of carbohydrate conversion to FAs (20). As a result, these transgenic plants were reported to have a 40% increase in total seed FAs compared with wild-type plants. Still other strategies to generally up-regulate lipid content in vegetative tissues rely on pathway engineering, transcriptional regulation, or altering carbon partitioning (21–23). The prospect of producing substantial quantities of TAG in tissues other than seeds likely will require metabolic changes at many levels but may offer new opportunities to impact vegetable oil production worldwide (22).

Transfer of Acyl Chains to ER for Glycerolipid Assembly

After FAs are released in plastids, it is generally presumed that they are converted to acyl-CoAs by long-chain acyl-CoA synthetases (LACSs) in the chloroplast envelope, and it is the acyl-CoAs that are transferred to the ER for incorporation into TAG (24). In vivo ¹⁸O labeling experiments confirmed that free FAs are released as intermediates (25), and additional kinetic studies suggested channeling to LACS (26).

There are nine LACS genes in Arabidopsis, and at least one isoform is localized to plastids (27). However, a T-DNA disruption of this lacs9 showed no impact on lipid synthesis, suggesting functional redundancy in this protein family (28). Furthermore, a double knock-out of lacs9 and lacs1 (or a triple knock-out of lacs1, lacs9, and lacs8, the closest LACS9 homolog) compromised seed oil biosynthesis only modestly (29). Even in these loss-of-function mutants, a considerable amount of acyl export from plastids supports TAG accumulation, and thus, the mechanistic nature of acyl export to the ER continues to be an intriguing puzzle. It is still unclear if the LACS active site is on the inner or outer surface of the outer chloroplast envelope and how this enzyme activity influences the availability of the acyl-CoA for ER lipid synthesis. Some have proposed direct physical contact with plastid-associated ER membranes (30) or stromules (membrane-bound extensions from plastids) (31), which might make direct membrane contact with the ER and allow for lipid exchange. Others have suggested that lipid transfer proteins (e.g. acyl-CoA-binding proteins) might facilitate transport through the cytosol (32). Alternatively, it is possible that acyl groups are incorporated into phosphatidylcholine (PC) in the plastid envelope and then transported to the ER. Results from rapid radiolabeling studies indicate that PC is the first glycerolipid labeled with newly synthesized FAs (33, 34). This may occur via lyso-PC acyltransferase (LPCAT) activity, which has been localized to the chloroplast envelope (35, 36). The details of how acyl groups from plastids are exported to the ER for incorporation into TAG remain a major unresolved question, and answers in this area may help to engineer increased TAG accumulation.

Alternative Pathways for TAG Synthesis in ER

A number of in vitro and in vivo studies have revealed that TAG synthesis can be more complex than previously recognized. In particular, different plants use alternative pathways or combinations of pathways for the assembly of TAG (Fig. 2).

FIGURE 2.

Schematic representation of acyl-CoA-dependent and acyl-CoA-independent pathways for TAG biosynthesis in plants. The Kennedy pathway involving acyl-CoA-dependent acyltransferases (GPAT, LPAT, and DGAT) includes an intervening PA intermediate that is hydrolyzed by PAP to form DAG. Two DGAT classes operate in plant TAG synthesis, and the relative contribution may be dependent upon the plant species and the specific types of FAs stored in TAG. Labeling studies of soybeans indicate that much of the newly synthesized FA supplied from the plastids rapidly enters an acyl-CoA pool that is incorporated into PC by acyl-editing reactions presumably via LPCAT. Acyl groups are desaturated or modified (e.g. hydroxylation) while esterified to PC and then released back to the acyl-CoA pool, where they may co-mingle with newly synthesized acyl-CoAs and are incorporated into glycerolipids by Kennedy pathway enzymes on a slower time course. The rapidly labeled PC pool is used to generate DAG (and vice versa) via PDCT (or CPT). TAG also can be synthesized from DAG and PC by PDAT, yielding lyso-PC, which can be reacylated to form PC. Net synthesis of PC from DAG is accomplished by CPT. PC can be hydrolyzed to PA by phospholipase D (PLD) or to DAG by phospholipase C (PLC; not shown) for TAG synthesis. This diagram does not account for separate pools of intermediates, which are likely. The relative contribution and combination of these alternative pathways to the overall flux to TAG in plants are tissue- and species-specific. Cho, choline; LPA, lysophosphatidic acid; LPC, lyso-PC; PLA, phospholipase A.

Conventional Kennedy Pathway

The conventional Kennedy pathway for the synthesis of glycerolipids in the ER is believed to play a role in the synthesis of TAGs in most organisms. This pathway involves the sequential acylation of the sn-1- and sn-2-positions of glycerol 3-phosphate (G3P) with acyl-CoA to yield phosphatidic acid (PA). PA is hydrolyzed to form diacylglycerol (DAG), and then the sn-3-position is acylated to yield TAG. This pathway overlaps with the synthesis of membrane glycerolipids because PA and DAG are also precursors for the major membrane lipids in all cells. The acyl substrates for the acyltransferases can be supplied either directly from plastid export and/or from acyl-CoA derived from acyl exchange with PC or other glycerolipids. Of the three acyltransferases, only DAG acyltransferase (DGAT) is unique to the synthesis of TAG, and this step has received the most attention in terms of influencing the accumulation of TAG in plant tissues.

Three acyl-CoA-dependent acyltransferases that cooperate to synthesize TAG have been studied in detail in recent years, with some progress toward understanding their contributions to lipid synthesis. Nine extraplastidial acyl-CoA:G3P acyltransferases (GPATs) have been identified in Arabidopsis. However, thus far, none have been demonstrated to be involved in TAG accumulation. Eight of these GPATs form a gene family that is specific to land plants. Several of these have been characterized at the genetic and biochemical levels. They transfer acyl groups to the sn-2-position and are involved in cutin or suberin biosynthesis (37) rather than membrane/storage lipid synthesis. Perhaps the best candidate for an ER-associated GPAT required for TAG biosynthesis is Arabidopsis thaliana GPAT9. This protein is not related to GPAT1–8 but shares the highest homology with the mammalian GPAT (GPAT3) that is involved directly in the synthesis of TAG in adipose tissues (38). A. thaliana GPAT9 has been confirmed to localize to the ER (39), but its impact on TAG accumulation is not certain.

For the lysophosphatidic acid acyltransferase (LPAT), five genes have been annotated in the Arabidopsis genome, and expression patterns and functions appear quite complex, with three (LPAT1, LPAT2, and LPAT3) being essential to normal plant development. Recently, overexpression of a Brassica napus LPAT isoform with homology to A. thaliana LPAT2 resulted in enhanced TAG accumulation in Arabidopsis seeds (40). Genetic studies with Arabidopsis LPAT2 have been limited because female gametophyte development is disrupted in the loss-of-function mutants (41); its direct role in TAG biosynthesis in Arabidopsis should perhaps be revisited. Both Arabidopsis LPAT4 and LPAT5 give rise to alternative transcripts, mostly with variations in the 5′-UTR (The Arabidopsis Information Resource), indicating a higher order of complexity in PA synthesis than is currently appreciated. Much functional work remains to be done on the LPATs and the broader family of LPAT-like genes to understand their metabolic role(s) in acyl lipid synthesis in general and in TAG accumulation specifically.

Comparatively more information is known about the terminal acyltransferases in TAG biosynthesis in plants, the DGATs. Two classes of genes encode DGATs in plants, DGAT1 and DGAT2, and these are homologous to those found in fungi and mammals (42). Genetic studies with mutants have confirmed that DGAT1 is required for normal TAG accumulation in oil-storing tissues of Arabidopsis (43), and its overexpression in vegetative tissues or seeds can lead to enhanced TAG accumulation (44, 45). However, disruption of DGAT1 gene function results in only a 20–40% reduction in Arabidopsis seed oil, so other mechanisms must cooperate in the accumulation of TAG in plants. The role of Arabidopsis DGAT2 is unclear because mutants have no observable phenotype, even when crossed with dgat1 mutants. On the other hand, the function of DGAT2 appears clearer in the castor bean endosperm system, which accumulates large amounts of TAG containing more than 90% ricinoleic acid (18:1 FA hydroxylated at the Δ¹²-position). Castor seeds express DGAT2 at much higher levels compared with DGAT1. Castor bean DGAT2 prefers ricinoleoyl-DAG acceptors (46), and its coexpression with the castor bean hydroxylase increases ricinoleic accumulation in Arabidopsis seeds (47). DGAT2 also has been studied in the Tung tree (which accumulates conjugated FAs in TAG), and it is localized to a subdomain of the ER that is different from that of DGAT1, suggesting that these two proteins may cooperate to synthesize TAG in Tung seeds but in spatially distinct subcellular locations (48). Although DGAT2 has been examined mostly in plant species with unusual FAs (47–50), it likely contributes to TAG accumulation in plants that do not accumulate unusual FAs based on expression levels in palm, olive, and other plants, even though its relative contribution may vary between species.

The formation of DAG from PA is catalyzed by a PA phosphatase (PAP)/phosphohydrolase. This enzyme in mammalian and yeast systems, also termed lipin, functions to control the flux between the synthesis of membrane lipids and TAG (51, 52). Loss-of-function mutations in both mammals and yeast result in increased membrane lipid proliferation and decreased TAG content (51, 52). In fact, mutations in lipin genes in mammals lead to a marked dysregulation of fat accumulation in adipose and other tissues (51). This results from dual activities of PAP both by transcriptional de-repression of lipid metabolism genes and by regulation of DAG levels. A recent additional role for PAH1 in lipid accumulation has been proposed in yeast: the accumulation of DAG within the ER is thought to nucleate lipid droplet formation (53). Two homologous genes designated PAH1 and PAH2 have been identified in Arabidopsis (54–56). Loss of function of PAH1 and PAH2 in Arabidopsis disrupted membrane remodeling, suggesting a role in the eukaryotic pathway of DAG formation (54). Furthermore, pah1/pah2 double mutants showed increased PC and overall expansion of the ER in leaves, consistent with a function in regulating membrane phospholipid accumulation as in animals and yeast (55), although the mechanism(s) remain to be elucidated. Still, these pah1/pah2 double mutants had only an ∼15% reduction in seed TAG, suggesting that additional unidentified PAP enzymes provide DAG for TAG biosynthesis. One potential candidate, the β-isoform of lipid phosphate phosphatase, was identified as markedly up-regulated in oil-storing tissues of palm (13), but the functional activity of this enzyme in TAG biosynthesis remains to be determined.

Considering the nutritional and economic importance of plant TAG, it is surprising how many questions remain about the precise molecular identity of a GPAT, LPAT, and PAP for TAG synthesis. Even the relative contribution of DGATs is somewhat unclear. Thus, fundamental uncertainties remain about central enzymes of plant TAG synthesis. Clearly, more work needs to be done to characterize the “conventional” Kennedy machinery. Furthermore, as outlined below, the recognition of other pathways that lead to TAG has further expanded the metabolic “networks” that need to be considered.

Acyl-CoA-independent Mechanisms Involved in TAG Synthesis

It is clear from many lines of evidence that the synthesis of TAG in plants is not as simple as the sequential acylation of glycerol with GPAT, LPAT, and DGAT (with an intervening PAP) by the conventional Kennedy pathway. In fact, a direct contribution from this pathway to TAG biosynthesis in many cell types may be far less than has been considered historically. In 2000, Stymne and co-workers (57) reported a new DGAT-independent mechanism for TAG synthesis in yeast and plants and identified a yeast gene that increased TAG when overexpressed. Phospholipid:DAG acyltransferase (PDAT) was identified and characterized as an acyl-CoA-independent transacylase that synthesizes TAG from PC and DAG, also yielding lyso-PC. PDAT was heralded as a mechanism for channeling unusual FAs into TAG in some plant species due to its markedly distinct comparative substrate specificities (57), and PDAT has been used to enhance the accumulation of TAG with oxygenated FAs in transgenic seeds (58, 59). PDAT was shown to participate in TAG biosynthesis in yeast log-phase growth and could partially compensate for loss of DGAT to store TAG during stationary phase (60). Six PDAT-like genes were identified in Arabidopsis, and one, designated PDAT1, accounted for most PDAT activity in plant tissues (61). However, there was no significant impact on TAG accumulation in pdat1 loss-of-function mutants or in plants overexpressing A. thaliana PDAT, so the role in overall TAG accumulation in Arabidopsis was uncertain (61, 62). A clear demonstration of a role for PDAT1 in TAG biosynthesis was provided more recently when its expression was silenced by RNA interference in a dgat1-1 mutant background (63). Seed TAG content was reduced by 70–80%, and normal seed (and pollen) development was disrupted. Thus, PDAT1 appears to compensate for a loss of DGAT1 and vice versa, but loss of both greatly compromises TAG deposition in both pollen and seeds, thereby indicating an overlapping role for PDAT1 and DGAT1 in seed oil accumulation. These results strongly suggest that PDAT and DGAT pathways cooperate in the majority of TAG synthesis in oil-storing tissues of plants, but this needs to be examined and extended to other plant tissues and species.

Recently, another DAG-utilizing enzyme has been identified in plants, and it plays a role in channeling unsaturated FAs into TAG. The enzyme PC:DAG phosphocholine transferase (PDCT) is encoded by the ROD1 (reduced oleate desaturation 1) gene in Arabidopsis. A loss-of-function mutation in this gene results in a 40% decrease in polyunsaturated FAs in seed TAG while leaving overall TAG levels essentially unperturbed (64). The enzyme transfers a phosphocholine headgroup of PC to the sn-3-position of a DAG molecule. The substrate specificity of the enzyme directs 18:1-containing glycerolipids toward further desaturation and 18:2- and 18:3-containing DAG toward TAG biosynthesis, thereby playing a role in modulating the FA composition of TAG. The discovery of this PDCT enzyme now helps explain the considerable flux from PC back into DAG required for TAG synthesis that may complement a reversibility of cholinephosphotransferase (CPT) activity. Another DAG-utilizing enzyme activity, a so-called DAG:DAG acyltransferase that forms TAG and monoacylglycerol, has been detected in vitro (61), but its molecular identity and role, if any, in TAG metabolism remain unknown.

With so many diverse activities in TAG metabolism relying on DAG (and other lipid) intermediates, questions have continued to surface about the potential existence of separate substrate pools and their possible role in modulating pathway flux. The existence of substrate pools also may be important when considering the overlap of the TAG biosynthetic pathways with those of phospholipid biosynthesis. Rapid radiolabeling studies of soybean embryos with [¹⁴C]acetate and [¹⁴C]glycerol support the existence of separate pools of DAG for TAG and PC biosynthesis (34), a conclusion that was suggested by a number of studies with different plant systems in vitro (24). Labeling embryos with acetate over a short time course (minutes) labels newly synthesized FAs exported from plastids, whereas labeling with glycerol over these same time scales reflects the de novo assembly of glycerolipids by acylation of G3P (34). Comprehensive molecular species characterization and kinetic analyses of these short-term radiolabeling data implicate at least two pools of DAG: one utilized for de novo PC synthesis and one utilized for TAG synthesis (34). The utilization of PC by plants as the substrate for ER desaturases (65) and as an intermediate in TAG biosynthesis provides a mechanism to increase the presence of polyunsaturated FA in TAG. Moreover, channeling of FAs into glycerolipids via specific acyl-CoA pools from PC acyl exchange has been suggested from numerous radiolabeling studies (26, 66). Accessibility of acyltransferases to these acyl-CoA pools and their compartmentation also should be incorporated into models describing TAG accumulation.

Acyl Hydrolases/Acyltransferases in Remodeling/Editing for TAG Biosynthesis

One interesting feature of plant glycerolipid metabolism is the extensive acyl-editing or acyl-remodeling reactions that are evident (24) and that result in a rapid exchange of acyl groups between PC and the acyl-CoA pool(s) (Fig. 2). In pea leaves (33), where glycerolipid synthesis is primarily for membranes, and in developing soybean (34), where glycerolipid synthesis is overwhelmingly devoted to TAG accumulation, there is a common theme with respect to acyl exchange from PC. In both systems, rapid (within 2 min) [¹⁴C]acetate labeling demonstrated that most newly synthesized FAs are first incorporated into PC before any other glycerolipid (33, 34). This initial flux of acyl groups into PC bypassed PA and the Kennedy pathway, indicating that the process of rapid acyl exchange occurred independently of net synthesis of phospholipid and TAG in developing embryos. The incorporation of newly synthesized FA (mostly oleate) into PC through acyl editing also results in the release of pre-existing FA from PC into the acyl-CoA pool. This pool is then available for acyl transfer by GPAT, LPAT, and DGAT. Because PC is the site of desaturation of oleate to linoleate and linolenate (65), this acyl exchange mechanism also allows polyunsaturated acyl-CoAs to enter any of the three sn-positions of TAG. It should be noted that acyl editing itself does not influence the amount of net PC or TAG synthesis per se, but it is keenly important to allow rapid acyl flux into PC for FA desaturation/modification reactions and to provide a pool of modified acyl-CoA structures for TAG assembly.

Acyl-editing reactions that exchange acyl groups among phospholipid molecular species with an intervening acyl-CoA pool are well established in eukaryotes and prokaryotes (67). These acyl exchange reactions in plants have been measured in vitro (68) and can be catalyzed either (a) by forward and reverse reactions of LPCAT (to yield acyl-CoA directly) or (b) by a phospholipase A-type activity to yield a free FA intermediate that then is activated to CoA (also referred to as the “Lands cycle”). Which of these two alternatives predominates in vivo is an important and challenging question for future research. Although some preference for sn-2 exchange is observed, the in vivo radiolabeling data implicate exchange at both sn-1- and sn-2-positions of PC. The fact that this set of reactions apparently acts preferentially on the PC pool (in comparison with other membrane lipids) reinforces the special roles played by PC in plant lipid metabolism.

Acyl editing can also be viewed as part of a cooperative system involving activities of the LPCAT (and phospholipase A/LACS?), CPT, DGAT, PDCT, and PDAT enzymes that together provide a means to equilibrate and utilize PC, DAG, and acyl-CoAs as crucial acyl intermediates/carriers in the modification and synthesis of TAG in the ER. It is possible that modifications of this cooperative metabolic network have contributed to the marked diversity of seed FA composition that is observed in nature. As noted, the castor bean accumulates 90% ricinoleic acid in TAG. Perhaps substrate specificity or specialized isoforms of LPCAT, PDCT, PDAT, and DGAT evolved in castor bean to be specific for hydroxy-FAs generated on PC by the diverged FAD2 (hydroxylase), and this system would be optimized for directing ricinoleic acid into TAG (59). This is consistent with the incremental increases in accumulation of ricinoleic acid in TAG by coexpressing DGAT2 or PDAT from castor bean in conjunction with the FAD2-like hydroxylase in transgenic oilseeds (47, 58, 59). It is possible that the activity of PDCT or other enzymes from developing castor seed displays specificity for ricinoleoyl-PC, and engineering of these enzymes might further enhance hydroxy-FA accumulation in transgenic plants. Alternatively, acyl-editing machinery may be responsible for enriching the DAG pool with hydroxy-FAs. The so-called “bottleneck” in the accumulation of unusual FAs in transgenic oilseeds might be a result of these cooperative enzyme systems that lack appropriate specificity for structures desired by oilseed engineers (69). Recently, Bates and Browse (70) concluded that the conversion of hydroxy-FA-containing DAG to PC is a major bottleneck in the production of TAG-containing hydroxyl-FAs in transgenic Arabidopsis seeds. Furthermore, it might be necessary to replace the competing endogenous machinery that is likely specific for the endogenous PC, DAG, and acyl-CoA species to gain maximum accumulation of unusual FAs in TAG in heterologous systems. It is of interest to note that it has often been relatively straightforward to engineer much higher levels of a FA that is already present (e.g. oleic acid) with a single-gene change than to increase FA levels in TAG of a FA that normally does not occur in the crop host (71).

Concluding Comments

What we thought we knew about the pathways and compartmentation of TAG accumulation in plants a decade or so ago has turned out to be oversimplified and, to a large extent, inadequate. Advances in our understanding in recent years have led to new discoveries and revisions of old hypotheses for the synthesis and incorporation of FAs into glycerolipids in plant cells. Many gaps still exist in almost all aspects of TAG accumulation, including the precise identity of enzymatic machinery, the relative contributions of various pathways in different species and tissues, the regulation of pathway fluxes, and the nature of metabolites involved in interorganellar transport. Answers to questions in these areas will be critical to the future rational design of biotechnology strategies needed to greatly enhance our agricultural production systems to more efficiently feed, clothe, house, and energize our growing and more affluent world population.