The plant lipidome in human and environmental health

See allHide authors and affiliations

Science  16 Sep 2016:
Vol. 353, Issue 6305, pp. 1228-1232
DOI: 10.1126/science.aaf6206


Lipids and oils derived from plant and algal photosynthesis constitute much of human daily caloric intake and provide the basis for high-energy bioproducts, chemical feedstocks for countless applications, and even fossil fuels over geological time scales. Sustainable production of high-energy compounds from plants is essential to preserving fossil fuel sources and ensuring the well-being of future generations. As a result of progress in basic research on plant and algal lipid metabolism, in combination with advances in synthetic biology, we can now tailor plant lipids for desirable biological, physical, and chemical properties. We highlight recent advances in plant lipid translational biology and discuss untapped areas of research that might expand the application of plant lipids.

Glycerolipids consist of a glycerol backbone with various combinations of fatty acids and head groups, thus generating a vast array of molecular species (Fig. 1). Glycerolipids constitute the largest fraction of the plant lipidome, or the total collection of plant lipids, and their chemical diversity is associated with many cellular functions. Plants build photosynthetic and cell membranes from polar lipids (Fig. 1). During the evolution of land plants, developmental adaptations led to the sequestration of carbon fixed by photosynthesis into high-energy compounds in the form of neutral lipids (Fig. 1), such as triacylglycerols in seeds. Algae accumulate triacylglycerols to survive adverse conditions—for instance, in response to nutrient deprivation. Glycerolipids also serve as mobile signals in cellular regulation and communication, and they function as components in photosynthetic and other enzymatic complexes. Basic research in model plants such as Arabidopsis thaliana and algae such as Chlamydomonas reinhardtii has generated insights into the regulation, synthesis, assembly, storage, and turnover of the plant and algal lipidomes. This intellectual framework leads to a design toolbox that enables translational research, construction of novel biotechnological processes, and generation of bioproducts.

Fig. 1 The plant glycerolipidome.

Plants are capable of combinatorial lipid chemistry. Glycerolipids consist of a combination of one to three fatty acids (inner dial) connected to a glycerol backbone. Positions on the glycerol backbone are designated as sn-1, sn-2, or sn-3. The glycerolipid subclass (outer dial, arranged in order of decreasing polarity) is specified by different polar head groups at the sn-3 position, or a hydroxyl (–OH) or acyl group in the case of neutral lipids.

Although most plants share a common set of reactions in lipid metabolism, the unique, specialized metabolism of select plants draws interest as a starting point for novel bio-based industrial processes and for improving human nutrition. Rapid technological advances in synthetic biology aided by complementary “omics” approaches have enabled the translation of these basic design principles through genetic modification (GM) of crops to synthesize valuable lipids. Examples of successful translational plant lipid research abound. Plant oils can be produced with tailored compositions for downstream production of desirable varnishes, soaps, or specialized lubricants (16); vegetable oil in nonseed plant tissues, such as leaves, to increase the biomass energy content (7, 8); healthy fish oil–like vegetable oils in field-grown GM crops without depleting the oceans (9, 10); and a potentially leaner seed oil with lower caloric content that may aid in combating obesity (11). In more futuristic efforts, basic research on the formation of lipid droplets, which store neutral lipids in plants and algae, may lead to practical synthetic biology platforms for the safe sequestering and easy harvesting of bioactive and hydrophobic lipid-based molecules. Finally, uncoupling the intertwined nutritional state and regulation of cell division of algal cells governing oil accumulation looks promising as a model system for potentially addressing instances when this analogous relationship goes awry in human cells, which may lead to cancer.

“Unusual fatty acids provide chemical feedstocks serving as precursors for lubricants, nutraceuticals, plastics, paints, natural insecticides, biodiesel, and jet fuels.”

Plants as chemical feedstock factories

A number of nondomesticated plant species (Table 1) produce less common fatty acids that are of interest for industrial or human health applications. These include fatty acids with short (<8 carbons), medium (8 to 14 carbons), or very long chain lengths (>20 carbons), with distinct patterns of unsaturation such as conjugated (separated by 1 carbon bond) ω3 or ω7 double bonds (positioned 3 or 7 carbons from the methyl end of the fatty acid, respectively), and with additional chemical modifications such as hydroxyl, epoxy, or cyclic groups (Fig. 1). Unusual fatty acids provide chemical feedstocks serving as precursors for lubricants, nutraceuticals, plastics, paints, natural insecticides, biodiesel, and jet fuels. In most cases, considerable constraint remains on cultivating nondomesticated plant species because of their poor agronomic performance or their adaptation to climates or ecological niches not permissive for widespread cultivation. Nevertheless, exploring the synthetic capacity of these nondomesticated plants has been essential for identifying specialized fatty acid synthesis and modification enzymes. These enzymes can be introduced into model plants for proof-of-concept demonstrations and ultimately into high-performing GM crops (Table 1).

Table 1 Examples of engineering GM crops and model plants to accumulate desirable fatty acids.
View this table:

The introduction of a diacylglycerol acyltransferase from the ornamental burning bush (Euonymus alatus) into the promising niche crop relative of canola, Camelina sativa, resulted in seeds with altered oil composition in which up to 85% of all triacylglycerols occur in the form of acetyl-triacylglycerols (11). Camelina is not a high-volume commodity crop, which should enable GM varieties to be more readily commercialized because of its easier identity preservation and stewardship. Acetyl-triacylglycerols contain a two-carbon acetyl chain at the sn-3 position of the glycerol backbone (Fig. 1). Development of this GM crop would address concerns about sustainable energy sources and the environment because acetyl-rich oils have reduced viscosity and freezing points and are therefore useful as replacements for fossil fuel–derived diesel #4 in heavy train and ship engines. More relevant to the average consumer, acetyl-triacylglycerols have a lower caloric value, providing a potentially leaner and healthier vegetable oil for human consumption.

Medium-chain fatty acids are also key ingredients for personal health care products such as lotions, shampoos, and soaps (1, 6) and are often sourced from plants with a limited global distribution, such as coconut. For the purpose of generating GM crops for the production of medium-chain fatty acids, cDNAs encoding specialized fatty acyl–acyl carrier protein thioesterases with altered substrate specificities have been overexpressed in canola (6), tobacco (1), or Camelina (12); this prevents the conventional medium-chain fatty acid elongation to 16- or 18-carbon fatty acids in the plastid, resulting in altered acyl compositions of lipids. For example, after co-producing these thioesterases and medium-chain coenzyme A (CoA)–specific lysophosphatidate acyltransferases, which promote sn-2 incorporation into glycerolipids, from Cuphea sp. in Camelina, 14:0 (carbons:double bonds) accumulated up to 37% of total fatty acids in seeds (12). In vegetative tissues that are often even less accommodating to unusual fatty acids, medium-chain fatty acids were generated in tobacco leaves by transient co-production of selected thioesterases with the coconut lysophosphatidate acyltransferase and the Arabidopsis WRINKLED1 transcription factor, which governs fatty acid biosynthetic gene expression in all oil-accumulating tissues tested, such that 12:0, 14:0, and 16:0 accumulated up to 10%, 19%, and 38%, respectively (1). Incorporation of unusual fatty acids at the sn-2 glycerol position is limited by endogenous fatty acid–editing mechanisms (13) that still need to be overcome to achieve higher productivity.

Engineering of very-long-chain fatty acids in two distinct but related species showed that selection of the engineered host plant is absolutely crucial. The expression of a cDNA encoding β-ketoacyl–CoA synthetase, which elongates long-chain fatty acids from oilseeds enriched in nervonic acid (24:1), produced only 13% 24:1 in Camelina versus up to 30% of total fatty acids in the close relative Brassica carinata (4, 14). Presumably, the higher erucic (22:1) content in some Brassica varieties presents a more readily available substrate for 24:1. Crambe abyssinica, naturally high in erucic fatty acid, was further engineered to enhance the production of 22:1 and to reduce competing lipid pathways, such that levels were increased from 60% to 73% of all fatty acids (3). These very-long-chain fatty acids find uses as lubricants, synthetic rubber, and cosmetics additives; as precursors for nylon and plasticizers; in the treatment of neurological diseases such as multiple sclerosis; and in infant nutrient supplementation (4).

Many of the useful chemical modifications in unusual fatty acids arise from catalysis by specialized, evolutionarily diverged fatty acid desaturases. Introduction of hydroxyl groups into fatty acids allows these molecules to be directly bonded as lubricants to surfaces of high-performance engine parts. Castor bean, the main source of hydroxy fatty acids, produces more than 90% hydroxy fatty acid (relative to total fatty acid content) but is also a natural source of the infamous poison ricin. Introducing the hydroxy fatty acid trait into GM crops has frequently led to deleterious side effects—for instance, a 20% reduction in total seed oil and poor seed germination (15). Combinatorial studies of hydroxylases introduced into Arabidopsis and GM plants have led to steady increases in hydroxy fatty acid production over time (Fig. 2). The highest levels of hydroxy fatty acids in Arabidopsis and Camelina were reported at 29% and 21% of total fatty acid content, respectively (5, 16). A comparison of global gene expression profiles from closely related Camelina to hydroxy fatty acid–accumulating developing seeds of Physaria fendleri suggested coevolution of many lipid biosynthetic genes and others associated with the production of hydroxy fatty acids (17). Furthermore, multiple differences in developmental timing of gene expression, substrate promiscuity by enzymes, allelic variation, ploidy, etc., likely contribute to differences in lipid composition among host systems. Consequently, a strategy for further increasing unusual fatty acid levels might require engineered plants with coexpression of a larger number of genes with targeted precision than previously assumed. In synthetic biology terms, the appropriate chassis makes a difference.

Fig. 2 Twenty years of engineering hydroxy fatty acids in oilseeds.

Data points represent individual reports of GM crops or Arabidopsis accumulating hydroxy fatty acids. Data are adapted from table S1 of Horn et al. (17).

Up to 65% of the total fatty acids in Camelina seeds have been converted to ω7 monounsaturated fatty acids, a sustainable precursor for 1-octene used in the production of polyethylene (2). Punicic acid, an 18:3 conjugated fatty acid from pomegranate (Punica granatum) thought to contribute to its antiatherogenic health benefits, was produced in amounts up to 21% of total fatty acids when introduced into a high linoleic (18:2) Arabidopsis background (18). By co-producing a cyclopropane fatty acid synthase and a cyclopropane-specific lysophosphatidate acyltransferase, up to 35% cyclopropane fatty acids, which serve as natural insecticides, of total fatty acids were produced in Arabidopsis seeds, although they showed reduced germination rates (19).

Increasing energy density and nutritional value

Diversity among land plants is also evident in the wide range of lipid content in storage and specialized tissues. Many domesticated crops such as palm, soy, and canola produce lipid-rich seeds with extractable seed oils, the primary source of commercial vegetable oils. Even modest increases in the total seed oil content will increase the crop’s value. Because yield increases by conventional breeding of high oil-producing natural varieties are expected to reach a limit, attempts have been made to improve performance of these crops through targeted engineering. For example, a 16% relative increase in soybean seed oil was achieved by introducing specialized amino acid residues into the main soybean diacylglycerol acyltransferase, DGAT1 (20). In canola, constitutive overexpression of the Arabidopsis transcription factor gene LEAFY COTYLEDON1 acting upstream of the WRINKLED1 transcription factor produced a range of relative seed oil increases up to 16% (21). Alternatively to targeting anabolism, the suppression of SUGAR-DEPENDENT1, a triacylglycerol lipase, by RNA interference (RNAi) increased relative canola seed oil content up to 8% (22).

Energy density and nutritional value of vegetative plant tissues can also be increased by diverting photosynthetically fixed carbon into triacylglycerols. An integrated strategy for increasing triacylglycerol levels in vegetative leaves has been applied through simultaneous alterations that increase the production of fatty acids by overexpressing WRINKLED1, in addition to driving the incorporation of fatty acids through increasing diacylglycerol acyltransferase abundance, and by protecting accumulated triacylglycerols by reducing lipase-mediated breakdown, as well as by directly altering carbon partitioning (7, 8, 23). Applying this approach to Arabidopsis roots, stems, and leaves resulted in accumulated triacylglycerol levels from 5 to 8% dry weight (23). In crop systems, tobacco leaf triacylglycerol levels increased up to 15% of dry weight (7), whereas triacylglycerol levels in sugarcane leaves and stems increased up to 1.9% of dry weight (8). Ectopic expression of a green algal diacylglycerol acyltransferase type II encoding cDNA in Arabidopsis increased leaf triacylglycerol content by a factor of 10, such that caterpillar larvae feeding on transgenic tissue gained 45% more weight, thereby demonstrating the nutritional enhancement of this GM plant (24). However, there remain challenges for further increasing oil in some vegetative tissues, as demonstrated by overexpressing WRINKLED1 in the grass Brachypodium distachyon, where local cell death is likely due to a doubling of free fatty acid levels (25).

One of the most striking successes for the nutritional enhancement of seed oil, while at the same time providing a solution for the protection of natural resources, is the production of fish oil–like vegetable oils in GM crops. The daily intake of ω3 long-chain polyunsaturated fatty acids by most humans—especially consumers of the Western diet, rich in saturated animal fats—falls short of recommendations despite well-documented health benefits, in particular those of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) (26, 27). The current global supply of naturally produced EPA and DHA in the form of marine fish oils is insufficient to meet increasing demand. Although many plants accumulate another ω3 fatty acid that is readily available in the human diet, α-linolenic acid (18:3), its conversion to EPA and DHA in humans is inefficient. Therefore, introducing these fish oil–like vegetable oils into GM crops provides a feasible, sustainable avenue for the production of ω3 long-chain polyunsaturated fatty acids for human consumption (9, 10). This synthetic biology–based engineering feat was achieved in Camelina by redirecting the endogenous production of 18:3 through a series of unusual desaturases and elongases derived from different microalgae and fungi. In seeds, up to 12% EPA and 14% DHA of total fatty acids accumulated when targeted for cosynthesis, or up to 31% when only EPA was targeted (28). In tests of the bioavailability of plant EPA and DHA, Camelina ω3s fed as supplements to mice or fish in aquaculture led to the improvement of key health markers, such as 36 to 38% lower blood glucose concentrations, indicating that these GM crop–derived vegetable oils could effectively replace marine fish oils (29, 30) used in aquaculture and the human diet.

Lipid droplets as synthetic biology platforms

The accumulation of high amounts of engineered plant oils is enabled by a plant’s innate capacity to dynamically generate lipid droplets capable of safely sequestering excess fatty acids as triacylglycerols. Lipid droplets are spherical organelles with a neutral lipid triacylglycerol-rich core enveloped by a polar lipid monolayer into which proteins are embedded. Plant lipid droplets associate with other compartments such as the endoplasmic reticulum and, particularly in microalgae, plastid envelope membranes. These membranes assemble triacylglycerols destined for lipid droplets. More than merely storage organelles, lipid droplets participate in maintaining cellular energy homeostasis, membrane remodeling during stress responses, and postgerminative growth (31, 32). The complex multicompartment nature of fatty acid synthesis, fatty acid modification, and triacylglycerol assembly in plants and algae complicates strategies to engineer total oil and unusual fatty acid accumulation. Synthesis and trafficking of lipids destined for lipid droplets currently represent a major knowledge gap and an opportunity for discoveries.

Being able to manipulate the formation, composition, and turnover of lipid droplets to broaden the engineering capacity of plants should enable us to reengineer lipid droplets, such that multiple enzymatic reactions constituting orthogonal pathways could be directly assembled on the lipid droplet surface (Fig. 3). One approach might be to use proteins known to reside at the lipid droplet surface, such as oleosins (33, 34), or algal lipid droplet proteins that recruit various proteins to the lipid droplet surface (32, 35). Protein(s) of interest could be fused to these lipid droplet surface proteins for targeting or scaffold engineering such that they associate at the surface through protein-protein interactions. For example, if one could produce hydroxy fatty acids at the lipid droplet surface through fatty acid hydroxylase supplied with its requisite cofactor provided by cytochrome b5 reductase and then channel these fatty acids to triacylglycerols by a lipid droplet–localized acyltransferase, one might be able to bypass many of the deleterious side effects these chemical modifications exert on membranes. As an alternative approach, existing divergent fatty acid desaturases targeted to lipid droplets could be engineered to directly act on triacylglycerol. This lipid droplet catalysis approach could also extend beyond lipid metabolism to generate and sequester bioactive compounds, such as drugs, or to introduce enzymes for nonplant purposes. As an example, human fibroblast growth factor 9 fused to oleosin and produced in Arabidopsis cells has been targeted to lipid droplets (36). Challenges include the short half-life of some lipid droplets, substrate availability and competition within competing pathways, and incompatibility of proteins targeted to the same membrane.

Fig. 3 Redesigning lipid droplets (LDs) as synthetic biology platform.

The proteome of plant and algal LDs is dominated by a set of “major proteins,” MP1 or MP2. These contain specific LD-targeting information and are inserted into the monolayer membrane at the periphery of the LD. Enzymes of interest (E) may be targeted to the LD using protein fusions (E1 fused to MP1) or through scaffold associations with other already LD-anchored proteins (E4). These enzymes would then catalyze available or directed substrates (S) on the LD surface, sequestering their lipophilic products (P4) within the LD or releasing hydrophilic products (P1, P2, P3) into cytosol. The LD shown is forming at the endoplasmic reticulum (ER) but could also be associated with chloroplast envelopes.

Unlinking triacylglycerol biosynthesis and cell division

Interest in microalgae has surged because of their high capacity for high-energy compound production in culture systems not competing with food crops. However, one of the biggest conundrums has been that microalgae accumulate triacylglycerols in lipid droplets when they are nutrient-deprived and cease growth. That is, triacylglycerol accumulation and biomass production are inversely correlated. Conceptually, cells must progress from the cell division cycle to the quiescent state, when cells are metabolically active but do not divide, to produce high amounts of triacylglycerol. How this transition is controlled is not yet known, but one possible regulatory factor, COMPROMISED IN THE HYDROLYSIS OF TAG 7 (CHT7), has been identified in Chlamydomonas (37). This protein is a component of a presumed large nuclear transcriptional complex not unlike the retinoblastoma tumor suppressor protein complex governing the cell division cycle in Chlamydomonas (38). Understanding how the nutritional state of cells affects quiescence and cell division will provide a solution to the conundrum hampering algal feedstock engineering. Although many studies suggest that nutrition determines the metabolic state of human cells and can thereby affect the outlook for the development of cancer, mechanistic insights are scarce. Chlamydomonas provides a single-celled, easily manipulated model system that has the same basic machinery governing cell division and cellular quiescence as found in human cells. Hence, basic research in algae on cell cycle regulation and triacylglycerol formation during nutrient deprivation–induced quiescence, and research into mechanisms by which the nutritional state of human cells affects cell division for better or worse, may converge in the near future. As such, fundamental insights gained by studying the regulation of algal lipid metabolism promise to allow us to address health and environmental issues of concern to current and future generations.


Measurable progress in engineering the plant lipidome has been made, but additional advances will require a holistic approach. Versatile “omics” approaches aided by bioinformatics will be needed to address several areas of lipid metabolism, including complexities within acyl exchange and lipid remodeling, characterization of orthologous enzymes for lipid modification, availability and localization of multiple lipid substrate pools within subcellular compartments, and trafficking of these lipid constituents between organelles. Addressing these areas of basic research, a renewed focus on characterizing the subtleties of lipid metabolism among natural and GM plants, and improved gene stacking and gene editing technologies should enable engineers of the next generation of GM plants to make better use of the lipidome for the benefit of all.

References and Notes

Acknowledgments: The Benning lab is supported by the Chemical Sciences, Geosciences, and Biosciences Division, Office of Basic Energy Sciences, Office of Science, U.S. Department of Energy grants DE-FG02-91ER20021 and DE-FG02-98ER20305; NSF grants MCB-1157231 and MCB-1515169; U.S. Department of Energy–Great Lakes Bioenergy Research Center Cooperative Agreement DE-FC02-07ER64494; and Michigan State University AgBioResearch.
View Abstract

Navigate This Article