• Open Access

Engineering plastid fatty acid biosynthesis to improve food quality and biofuel production in higher plants


(Tel +55 19 3429 4344 ext. 21; fax +55 19 3422 2644; email hecarrer@esalq.usp.br)


The ability to manipulate plant fatty acid biosynthesis by using new biotechnological approaches has allowed the production of transgenic plants with unusual fatty acid profile and increased oil content. This review focuses on the production of very long chain polyunsaturated fatty acids (VLCPUFAs) and the increase in oil content in plants using molecular biology tools. Evidences suggest that regular consumption of food rich in VLCPUFAs has multiple positive health benefits. Alternative sources of these nutritional fatty acids are found in cold-water fishes. However, fish stocks are in severe decline because of decades of overfishing, and also fish oils can be contaminated by the accumulation of toxic compounds. Recently, there is also an increase in oilseed use for the production of biofuels. This tendency is partly associated with the rapidly rising costs of petroleum, increased concern about the environmental impact of fossil oil and the attractive need to develop renewable sources of fuel. In contrast to this scenario, oil derived from crop plants is normally contaminant free and less environmentally aggressive. Genetic engineering of the plastid genome (plastome) offers a number of attractive advantages, including high-level foreign protein expression, marker-gene excision and transgene containment because of maternal inheritance of plastid genome in most crops. Here, we describe the possibility to improve fatty acid biosynthesis in plastids, production of new fatty acids and increase their content in plants by genetic engineering of plastid fatty acid biosynthesis via plastid transformation.


Fatty acids are essential constituents of all plant cells. The fatty acid biosynthesis pathway is a primary metabolic pathway, because they are found in every plant cell and are essential for cell division, growth and development. No mutations that block fatty acid synthesis have been isolated to date (Ohlrogge and Browse, 1995; Kode et al., 2005). Fatty acids are a class of compounds that are insoluble in water and extremely diverse in structure. They are the major components of membranes, which delineate the cell and its compartments. The de novo biosynthesis of fatty acids is essential for cell survival (Kode et al., 2005) and has important functions in many other fundamental processes (Raffaele et al., 2009; Takami et al., 2010; Wallis and Browse, 2010; Wasternack and Kombrink, 2010; Christensen and Kolomiets, 2011).

Despite the presence of fatty acids in every cell, some plant cells produce and/or store much more lipids than leaf mesophyll cells do. Lipids are the major form of carbon storage in the seeds of many plant species, reaching up to 60% of their dry weight of such seeds. Epidermal cells are another example of cells that can produce a high amount of lipids. They can produce cuticular lipids that coat the surface of plants, providing the crucial hydrophobic barrier that prevents water loss and also forms a protection against pathogens and other environmental stresses. In addition, fatty acids are important as precursors for hormone production (e.g. jasmonates) and act in the acylation of certain membrane proteins (Ohlrogge and Browse, 1995; Ohlrogge and Jaworski, 1997; Wasternack and Kombrink, 2010).

The major fatty acids found in plants (and most other organisms) have chain lengths of 16 or 18 carbons and contain 1–3 cis double bonds (Ohlrogge and Jaworski, 1997). Five fatty acids (18 : 1, 18 : 2, 18 : 3, 16 : 0 and in some species, 16 : 3) make up over 90% of the acyl chains of the structural glycerolipids in most plant membranes (Ohlrogge and Browse, 1995; Thelen and Ohlrogge, 2002). Fatty acids in cells are predominantly found with their carboxyl group esterified or otherwise modified, and they are almost never found in the form of ‘free’ fatty acids. The predominant form of fatty acids found in membranes is the form esterified to glycerol, which is termed glycerolipid. Membrane glycerolipids may have three fatty acids attached to the sn-1, sn-2 and sn-3 positions of the glycerol backbone. The combination of nonpolar fatty acyl chains (sn-1 and sn-2 positions) and polar head groups (sn-3 position) leads to the amphipathic physical properties of glycerolipids, which are essential for the formation of membrane bilayers. If all three positions on glycerol are esterified with fatty acids, a triacylglycerol structure results, which is less suitable for membranes but instead constitutes the major form of lipid storage in seeds (Ohlrogge and Browse, 1995; Ohlrogge and Jaworski, 1997; Napier, 2007).

In recent years, many biotechnological approaches have been focused on genetically engineered oilseed plants to incorporate additional fatty acids of nutritional importance (e.g. arachidonic acid—AA, eicosapentaenoic acid—EPA and docosahexaenoic acid—DHA) (Wu et al., 2005; Napier, 2007; Damude and Kinney, 2008; Napier and Graham, 2010) that are not currently obtained from domesticated plant crops (Singh et al., 2005; Cahoon et al., 2007; Napier, 2007). Likewise, these approaches have also been used to optimize and/or enhance oil content in plants for biodiesel (Durrett et al., 2008; Graef et al., 2009; Lu et al., 2010).

Over the last decades, the ability to manipulate fatty acids in plants was stimulated by the understanding of the metabolic pathways, the increasing availability of genes that encode enzymes related to fatty acids biosynthesis and the establishment of biotechnologies to metabolically engineer plants expressing single genes or multistep biosynthetic pathways (Wu et al., 2005; Napier, 2007; Damude and Kinney, 2008; Graef et al., 2009; Lu et al., 2010; Napier and Graham, 2010). The feasibility of the manipulation of fatty acids in transgenic plants open up the possibility of plant fatty acid engineering via the transformation of another genome, the plastid genome (Bock and Khan, 2004; Bock, 2007; Craig et al., 2008). In this review, we examine attempts to produce long-chain polyunsaturated fatty acids similar to those found in fish oils and to increase fatty acid content by engineering the metabolic pathway using plastid-located nuclear-encoded genes by expressing foreign enzymes via plastid transformation.

Why to genetically engineer fatty acids in plants

The main sources of oils and fats in the human diet are oilseed crop plants (Gunstone, 2001). The abundant fatty acids produced in major commercial oilseeds comprise just four oil species (linoleic acid, palmitic acid, lauric acid and oleic acid) and they completely lack VLCPUFAs (Lands, 2005). Mammals, including humans, have an absolute dietary requirement for two fatty acids, which are ω-6 linoleic acid (LA) and ω-3 α-linoleic acid (ALA). Therefore, these two fatty acids are classified as essential fatty acids (EFAs) (Burdge and Calder, 2005; Graham et al., 2007). Most oil crops contain high levels of LA and ALA. However, the percentage of LA and ALA is highly dependent on the plant species and vary intensely within the different oil crops (Gunstone et al., 2006).

The optimal balance of ω-3 and ω-6 VLCPUFAs must be maintained for good human health. The recommended ratio of dietary ω-6 and ω-3 fatty acids ranges from 2 : 1 to 6 : 1 (Sargent, 1997; Simopoulos, 2000).The consumption of fatty acids has shifted heavily towards the ω-6 fatty acids in the current diet and by some estimates, it is up to 30-fold too high (Simopoulos, 1999), attributable in part to the large increase in the consumption of vegetable oils that are rich in LA. Because ω-6 and ω-3 fatty acids are not interconvertible in the human body, the ratio of LA/ALA in our diet influences the ratio of ω-6/ω-3 VLCPUFAs synthesis. To correct this imbalance, the needed ω-3 VLCPUFAs in our diet can be obtained from fish sources (Napier, 2007; Durrett et al., 2008; Dyer et al., 2008).

VLCPUFAs containing at least 20 carbons such as ω-3 EPA and DHA are valuable commodities that provide important human health benefits in various physiological and pathophysiological processes (Spector, 1999; Funk, 2001; Hong et al., 2003). They are essential components of the cell membrane and constitute over 30% of the fatty acids in the brain (Crawford et al., 1997). In the retina, DHA accounts for more than 60% of the total fatty acids (Giusto et al., 2000). Clinical studies show that DHA is essential for growth and development of the brain in infants and for the maintenance of normal brain function in adults (Martinetz, 1992). Cell growth and division, platelet aggregation, inflammatory responses, haemorrhage, vasoconstriction and vasodilatation, and immune functions are associated with unbalanced VLCPUFAs in the diet (Dnyaneshwar et al., 2006). Studies have also shown their role in the prevention and treatment of coronary heart diseases, hypertension, type 2 diabetes, arthritis, cancer and other inflammatory and autoimmune disorders (Simopoulos, 1999; Shapira, 2007).

The primary sources of VLCPUFAs are marine organisms (microbes and algae) that are consumed by fishes, which in turn accumulate these fatty acids (Williams and Burdge, 2006). However, marine sources of ω-3 VLCPUFAs for human health and nutrition are becoming problematic because the natural fish stocks are not managed in a sustainable way and are in severe decline attributed to decades of overfishing (Domergue et al., 2005; Heinz, 2006; Napier, 2007). Another problem is the pollution of the oceans, resulting in the accumulation of toxic compounds such as heavy metals, dioxins and plasticizers in the fish oils (Jacobs et al., 2004; Napier and Sayanova, 2005; Robert, 2006; Graham et al., 2007). A current alternative to obtain a source of ω-6 VLCPUFAs is via fermentative cultures of filamentous fungi, but these production systems are expensive to maintain and have limited increasing capacity (Yamada et al., 1992). Because of the importance of VLCPUFAs as components of human diet, their demand is constantly increasing (e.g. being now standard components of infant formula milks) (Calder, 2004; Napier, 2007; Shapira, 2009; Agostoni, 2010; Makrides et al., 2010).

Plant oils represent an important renewable resource from nature, and oilseeds provide a unique platform of low costs for the production of high-value fatty acids that can replace nonsustainable oceanic sources (Damude and Kinney, 2007, 2008; Napier, 2007; Napier and Graham, 2010). Two examples show the possibility to produce long-chain polyunsaturated fatty acids in plants. Wu et al. (2005) demonstrated the feasibility of making VLCPUFAs in transgenic Brassica juncea, which resulted in the accumulation of arachidonic acid (ARA) (4%), EPA (8%) and DHA (0.2%), while Kinney (2004) reached 3.1% of DHA and 6.1% of EPA in transgenic soybean somatic embryos.

In addition to the nutritional importance of fatty acids, the utilization of plant fatty acids as a source for biodiesel has grown extensively in the last years. The tendency for biodiesel use is induced by the rapidly increasing costs of petroleum, reduction in petroleum reserves, concern about the adverse environmental impact of burned fossil oil and the need to develop renewable sources of fuel (Durrett et al., 2008; Dyer et al., 2008; Yuan et al., 2008; Carlsson, 2009).

Plant oils have a higher energy content, reaching 90% of the heat content of petroleum-derived diesel. The production of biodiesel from plant oils can reach a favourable energy input/output ratio of about 1 : 2 to 1 : 4 in nonirrigated fields, which means a positive balance between energy required to produce the crop compared with the energy obtained from the crop (Agarwal, 2007). In comparison with conventional diesel, biodiesel presents advantages such as its oxygenated state, which leads to lower carbon monoxide (CO) production and emission (Graboski and McCormick, 1998). Biodiesel also contains little to no sulphur or aromatic compounds reducing sulphur oxide, sulphuric acid formation and aromatic compounds considered carcinogens (Knothe et al., 2005). However, the greatest advantage of biodiesel is that it is essentially neutral with respect to the production of carbon dioxide. This is because the energy contained within the triacylglycerols is derived from the sun and captured by plants through photosynthesis (Hill et al., 2006).

Many oil crop species are being used for biodiesel production (e.g. soybean, oil palm, canola, Jatropha), and the oil composition determines several proprieties of their biodiesel, such as cold-temperature flow, oxidation, cetane number and NOx emissions. As the properties cited previously are affected positively and negatively by the degree of unsaturation (Durrett et al., 2008; Dyer et al., 2008), there is no fatty acid profile that gives optimal conditions for the desired biodiesel proprieties. However, a very good profile can be achieved with a fuel containing high amounts of mono-unsaturated fatty acids, such as oleate (18 : 1 Δ9) or palmitoleate (16 : 1 Δ9), and low in both saturated and polyunsaturated fatty acids. The single double bond can improve the cold-temperature flow properties. There are evidences from several studies that high levels of methyl oleate would have optimal characteristics regarding ignition quality, NOx emissions and fuel stability (Stournas et al., 1995; Serdari et al., 1999; McCormick et al., 2001; Knothe and Dunn, 2003; Knothe, 2005).

Although high oleic and palmitic plant oil would contain the desired fatty acid profile for biodiesel production, no domesticated oil crop plant can have high content of oleic acid and palmitic acid. For example, the levels of these fatty acids in soybean are influenced by temperature indicating that any environmental change can affect the profile of fatty acids (Rebetzke et al.,1998). The implementation of biotechnological tools can be an alternative to solve this problem by manipulating the enzymes of the fatty acids pathway. Thereby, Buhr et al. (2002) described the development of transgenic soybean plants with down-regulation of the genes FAD2-1 and FatB. The oil produced by these plants presented a reduced level of palmitic acid content (<5%) and significantly increased oleic acid content (>85%). The oil obtained from them produced a biodiesel with improved cold flow and enhanced oxidative stability, two critical fuel parameters that can limit the utility of biodiesel (Graef et al., 2009).

Another exciting approach is to improve the oil content in the seeds or the amount of seeds per plant and, consequently, the production per hectare. However, most of this research is at an early stage. Evidences suggest that oil synthesis may be limited by the production of fatty acids (Ohlrogge and Jaworski, 1997; Bao and Ohlrogge, 1999; Sasaki and Nagano, 2004), which can be regulated by the acetyl-CoA carboxylase (ACCase) activity. Targeted expression of ACCase in plastids by transit sequence resulted in a 5% increase in seed oil (Roesler et al., 1997). On the other hand, reduction in ACCase activity lowered the fatty acid content in transgenic seeds (Thelen and Ohlrogge, 2002). Overexpression of the plastid-encoded acetyl-CoA carboxylase β-subunit (AccD), by plastid transformation, induced an increase of 5%–10% of leaf oil content and the number of seeds per plant increased about 2-fold over the control and wild-type tobacco plants (Madoka et al., 2002). These examples, related to the increase in fatty acid content and the seed number, demonstrate the potential of the use of genetic engineering to enhance the quality of fatty acids and the content of oil per hectare.

Advantages of plastid transformation to manipulate fatty acids

Over the last decades, chloroplast genetic engineering has attracted the interest of biotechnologists by offering several exceptional features and advantages when compared with nuclear transformation. They include high transgene expression levels, foreign protein accumulation of up to >70% of the total soluble cellular protein (100 times higher than nuclear transformation because of polyploidy of the plastid genome and/or the high stability of foreign proteins) (Oey et al., 2009; Ruhlman et al., 2010), capacity for multigene (operons) engineering in a single transformation event because of efficient translation of polycistronic mRNAs in plastids (De Cosa et al., 2001; Ruiz et al., 2003; Loessl et al., 2005; Quesada-Vargas et al., 2005; Krichevsky et al., 2010), absence of position effects in plastids because of lack of a compact chromatin structure (Bock, 2001), efficient and precise transgene integration by homologous recombination via a RecA system that was inherited from their cyanobacterial ancestors (Cerutti et al., 1992), absence of epigenetic effects or gene silencing (Bock, 2001) and absence of pollen transmission of transgenes because of transgene containment via maternal inheritance of plastids in most higher plants (Daniell, 2007; Ruf et al., 2007; Svab and Maliga, 2007).

Plastid engineering also offers another advantage. The selectable markers employed for plastid transformation can be successfully eliminated from the genome by using the following methods: cre-lox site-specific recombination (Corneille et al., 2001; Hajdukiewicz et al., 2001; Lutz et al., 2006), phiC31 phage site-specific integrase (Kittiwongwattana et al., 2007) and use of direct repeats for gene excision via homologous recombination (Klaus et al., 2004b; Dufourmantel et al., 2007). The elimination of the antibiotic resistance gene after the generation of transgenic plants is desirable to prevent the risk of the antibiotic resistance gene flow (public concern) and to recycle the selectable marker genes that are relatively restricted in number for plastid transformation (Carrer et al., 1993; Svab and Maliga, 1993; Corneille et al., 2001; Barone et al., 2009; Li et al., 2010a).

The plastid transformation methodology is based on the following steps: insertion of the transforming DNA into the target organelle (preferentially performed by particle gun); integration of transforming DNA into the plastid genome (homologous recombination); selection of cells containing transformed plastome; and regeneration of stable transplastomic plants from a single cell (Figure 1). The chloroplast transformation starts from a single or few plastid genome molecules and after several events of plastid and cell division in the presence of the selective agent, the transformants reach the homoplasmic state (Figure 1). The homoplasmy is generated when all plastid genome copies are transformed and it is a prerequisite for stable transformation (Bock, 2001; Bock and Khan, 2004; Maliga, 2004; Verma et al., 2008).

Figure 1.

 Biolistic chloroplast transformation methodology, integration of transgenes into the plastid genome and isolation of homoplasmic transplastomic cell lines. Leaves are shot with gold (or tungsten) particles coated with the transformation vector and then cut in small pieces and exposed to a regeneration medium containing the selection agent (e.g. spectinomycin). The transgene is targeted in a noncoding intergenic region of the plastid genome (ptDNA). The aadA and the gene of interest are integrated into the plastid genome by two homologous recombination events in the flanking regions (dashed lines). The selection of transplastomic chloroplasts starts by sorting a single (or at most a few) plastid genome copy. As a leaf cell contains several thousand plastid genome copies, a subsequent selection trough cell and organelle divisions, in the presence of high concentrations of the selecting antibiotic, is needed. Antibiotic-sensitive plastids are efficiently sorted out by the inhibition of plastid protein biosynthesis. Transplastomic spectinomycin-resistant plants can regenerate because the transformation vector contains a spectinomycin resistance gene (aadA). Primary transformants appear after 4–6 weeks and are usually heteroplasmic. The homoplasmic state is reached by passing explants through additional rounds (normally 2–4) of regeneration under antibiotic selection. Lastly, homoplasmic shoots are proliferated and rooted in phytohormone-free medium. The additional rounds of regeneration on selective medium eliminate the residual wild-type genomes and produce cells containing a population of plastid genomes totally transformed termed as homoplasmic state or homoplasmy (Figure modified from Bock, 2001; and Bock and Khan, 2004).

Because of these advantages, the chloroplast genome has been engineered to confer several useful agronomic traits, including herbicide resistance (Ye et al., 2001, 2003; Kang et al., 2003; Dufourmantel et al., 2007), insect resistance (De Cosa et al., 2001; Dufourmantel et al., 2005; Chakrabarti et al., 2006), drought tolerance (Zhang et al., 2008), metabolic engineering (Wurbs et al., 2007; Apel and Bock, 2009; Krichevsky et al., 2010), enzymes for biofuel production (Gray et al., 2009; Verma et al., 2010a), salt tolerance (Zhang et al., 2008) and phytoremediation (Ruiz et al., 2003; Ruiz and Daniell, 2009). Moreover, the plastid transformation has great potential for the production of enzymes of industrial interest (Viitanen et al., 2004), biopharmaceutical compounds (Boyhan and Daniell, 2010; Lee et al., 2011) and vaccines (Daniell et al., 2005, 2009; Daniell, 2006; Bock, 2007; Singh et al., 2009).

Nowadays, plastid transformation is available in many plant species, including tobacco (Svab et al., 1990), tomato (Ruf et al., 2001), potato (Valkov et al., 2010), Lesquerella (Skarjinskaia et al., 2003), Brassica napus (Hou et al., 2003), petunia (Zubkot et al., 2004), soybean (Dufourmantel et al., 2004), carrot (Kumar et al., 2004a), cotton (Kumar et al., 2004b), lettuce (Lelivelt et al., 2005), poplar (Okumura et al., 2006) and cabbage (Liu et al., 2007). However, the ability of each species to regenerate in vitro (i.e. generation of a whole plant from a single transformed cell) determines the rate of plastid transformation efficiency.

Although during the past few years plastid transformation was more explored for the expression of genes related to insect/herbicide resistance and molecular farming—biopharmaceuticals and vaccines (Daniell, 2006; Bock, 2007; Daniell et al., 2009; Singh et al., 2009; Davoodi-Semiromi et al., 2010; Verma et al., 2010b)—the plastid harbours a huge number of metabolic pathways that can be engineered by genetic transformation of the plastome. Recently, a study showed the feasibility to engineer a nutritionally important metabolic pathway in nongreen plastids. Apel and Bock (2009) produced transgenic tomato plants that resulted in an increase up to 50% in provitamin A (an important antioxidant and essential vitamin for human nutrition) content in tomato fruits (Figure 2a). Another pigment of human health interest, astaxanthin, was produced via plastid transformation at very high levels (Hasunuma et al., 2008). The pigment accumulated at concentrations up to 5.44 mg/g dry weight, which corresponds to 74% of the total carotenoid content in the transplastomic plants (Hasunuma et al., 2008). Newly, the expression of LUX operon containing six genes in plastids yielded plants that are capable of autonomous light emission (Krichevsky et al., 2010). This work demonstrates that transplastomic plants can emit light that is visible by naked eye (Krichevsky et al., 2010) and, moreover, that complex metabolic pathways of prokaryotes can be successfully reconstructed in higher plant chloroplasts.

Figure 2.

 Genetic engineering of metabolic pathways in transplastomic plants and simplified overview of plant fatty acids biosynthesis. (a) Examples of different metabolic pathways successfully manipulated by plastid transformation. (i) Metabolic engineering of carotenoid biosynthesis by expression of the lycopene β-cyclase from daffodil (Narcissus pseudonarcissus). The tomato transplastomic plants showed an increase of 50% in total carotenoid content. The provitamin A levels reached 1 mg/g dry weight, while wild-type fruits have <200 ng provitamin A per g (Apel and Bock, 2009). (ii) Expression of Δ9-desaturase genes in tobacco chloroplasts from wild potato species (Solanum commersonii) or cyanobacterium (Anacystis nidulans) resulted in altered fatty acid profiles and an increase in their unsaturation level both in leaves and seeds (Craig et al., 2008). (iii) Genetic manipulation by overexpression of the plastid accD gene (beta-carboxyl transferase subunit of acetyl-CoA carboxylase) increased leaf fatty acid content by approximately 5%–10%. Fatty acid content of the seeds and the fatty acid composition in the wild-type and transplastomic plants did not present significant differences. However, the number of seeds per plant in the transplastomic plants increased about 2-fold over the control plants, which doubled the fatty acid production per plant (Madoka et al., 2002). (b) Suggested target enzymes and metabolic pathways for genetic engineering of plastid fatty acid biosynthesis. Several enzymes are shown, which provide substrates for the fatty acids biosynthesis. The production of long-chain polyunsaturated fatty acids is driven using acyl carrier protein (ACP) substrates via the polyketide synthase pathway. The scheme indicates a combination of nuclear and plastid genome engineering to manipulate and increase VLCPUFAs in plants. PEP, phosphoenolpyruvate; Pyr, pyruvate; VLCPUFAs, very long chain polyunsaturated fatty acids; EPA, Eicosapentaenoic acid; ARA, arachidonic acid; DHA, docosahexaenoic acid; TE, thioesterases; ACS, acyl-CoA synthetase.

Regarding the alteration of the metabolic pathway of fatty acids in plastids, only two attempts were made by plastid transformation up to now. The first one aimed to induce the production of fatty acids by overexpression of the accD gene (coding for the β subunit of acetyl-CoA carboxylase) and the second one was directed to increase the unsaturation degree through the expression of Δ9-desaturases (Figure 2a). Madoka et al. (2002) overexpressed the accD gene by changing the natural promoter to the strong constitutive rRNA operon promoter. Plants showed an increase in fatty acid content in leaves; they also displayed retarded leaf senescence and increased seed production (Figure 2a). Craig et al. (2008) produced transplastomic tobacco plants individually expressing Δ9-desaturase genes from Solanum commersonii and from Anacystis nidulans. In comparison with control plants, transplastomic plants containing either Δ9-desaturase gene showed an increase in the unsaturation degree of fatty acids in leaves and seeds. The increased unsaturation degree in transplastomic plants also increased the cold tolerance (Craig et al., 2008). These examples demonstrate the potential of plastid transformation to engineer plastid fatty acid biosynthesis in both vegetative and reproductive tissues (Figure 2a).

Several studies have focused on the development of expression cassettes (containing regulatory elements, such as promoters, 5′ UTRs, ribosome binding sites and 3′ UTRs compatible with the plastid gene expression machinery) in transformation vectors for expression in plastids (Kuroda and Maliga, 2001; Lutz et al., 2007; Verma et al., 2008; Sinagawa-García et al., 2009; Ruhlman et al., 2010). The efforts to optimize signals for gene expression in chloroplasts have shown that this expression can be successfully optimized according to the purpose and/or metabolic pathway (Dufourmantel et al., 2007; Apel and Bock, 2009; Oey et al., 2009; Ruhlman et al., 2010; Verhouning et al., 2010). This optimized expression cassettes allowed very high accumulation of foreign proteins (Oey et al., 2009) and pharmaceutical substances (Ruhlman et al., 2010), and production of compounds of nutritional importance in plastids (Hasunuma et al., 2008; Apel and Bock, 2009). Lately, the analysis of transcription and translation of plastid genes during tomato fruit development (differentiation from chloroplast to chromoplast) was shown (Kahlau and Bock, 2008). This study showed that the expression pattern is strongly altered, and the only highly expressed gene in tomato chromoplast is accD. Interestingly, this is the only gene present in the plastids involved in fatty acid biosynthesis for which the expression machinery is maintained (Kahlau and Bock, 2008). The results of this study will facilitate to design novel expression cassettes containing strong promoters and 5′ untranslated regions, maximizing the accumulation of foreign protein and nutritional compounds in chromoplast-containing fruits (Kahlau and Bock, 2008). Also, a gene expression characterization was determined in potato tuber amyloplasts (Valkov et al., 2009). This study identified potential candidate regulatory sequences, which could improve plastid transgene expression in amyloplasts. Some of these potential regulatory sequences were tested in potato via plastid transformation (Valkov et al., 2010). The authors showed that the protein accumulation was much lower in tuber amyloplasts than in chloroplasts. At the same time, the potential use of novel regulatory sequences to express transgenes in amyloplasts and other nongreen plastids is suggested (Valkov et al., 2010).

Improving plant oil quality by genetic engineering of VLCPUFAs

De novo fatty acid synthesis in plants occurs in plastids and is catalysed by multisubunit fatty acid synthase (FAS) complexes. The final products of these enzymatic complexes are 16 : 0- and 18 : 0-acyl carrier protein (ACP). Normally, 18 : 0-ACP undergoes desaturation by a soluble stearoyl ACP desaturase to form 18 : 1Δ9-ACP (Ohlrogge and Jaworski, 1997). Most fatty acids are hydrolysed from ACP by acyl-ACP thioesterase, then they leave the plastid and are esterified to coenzyme A (CoA) to form acyl-CoA (Ohlrogge and Jaworski, 1997). Through a series of reactions, the acyl moieties can become esterified to phosphatidylcholine (PC) and can then undergo desaturation by Δ12- and Δ15-desaturases to form LA and ALA (Ohlrogge and Browse, 1995; Ohlrogge and Jaworski, 1997; Sasaki and Nagano, 2004).

All higher plants have the enzymatic complex to synthesize the C18 polyunsaturated fatty acids (C18-PUFAs) LA and ALA, while some of them can also synthesize γ-linolenic acid (GLA) and stearidonic acid (SDA). However, higher plants do not possess the enzymes to follow the elongation and desaturation steps to convert C18-PUFAs into VLCPUFAs (Napier, 2007; Damude and Kinney, 2008; Napier and Graham, 2010).

Most animals can synthesize EPA and DHA de novo from the essential dietary fatty acid ALA by using a common metabolic pathway (Cahoon et al., 2007; Graham et al., 2007; Napier, 2007; Damude and Kinney, 2008). VLCPUFAs are present in the marine food chain (e.g. microorganisms such as diatoms, golden-brown algae, green algae, blue-green algae, microbial fungi and dinoflagellates), all of which are rich in VLCPUFAs and are food sources mainly for fishes. These microorganisms synthesize EPA and DHA via two different classes of biochemical pathways: i) aerobic fatty acid desaturation/elongation pathway (Sayanova and Napier, 2004) and ii) the anaerobic polyketide synthase (PKS) pathways (Metz et al., 2001). Most marine organisms utilize the aerobic fatty acid desaturation/elongation pathway, in which EPA is generated by a Δ6-desaturase pathway in fish and as well as in humans (i.e. Δ6-desaturation, Δ6-elongation and Δ5-desaturation; Sayanova and Napier, 2004). The PKS pathway synthesizes from acetate, in a similar way as that for fatty acid biosynthesis, by using a single enzyme complex (Bentley and Bennett, 1999; Metz et al., 2001). Some marine organisms use PKS-type complexes to synthesize EPA and/or DHA directly from malonyl-CoA (Metz et al., 2001; Napier, 2007). These enzyme complexes can introduce double bonds by using a dehydrase–isomerase mechanism similar as found in Escherichia coli, which do not require oxygen as do fatty acid desaturases (Qiu et al., 2001; Metz et al., 2006, 2009).

Owing to the attractiveness of producing a plant-derived VLCPUFA, much efforts have been invested in the cloning of several VLCPUFA-related genes (desaturases and elongases), the characterization of the metabolic pathway and the expression of the genes in suitable hosts such as microorganisms and plants (Meyer et al., 2004; Sayanova and Napier, 2004; Domergue et al., 2005; Damude and Kinney, 2007).

Initial experiments were focused on the expression of individual genes (predominantly encoding desaturases) in transgenic plants for VLCPUFA biosynthesis (Sayanova et al., 1997). These studies indicated the relatively high accumulation of non-native C18 fatty acids, such as GLA and SDA, resulting from the expression of heterologous Δ6-desaturases. Likewise, the expression of Δ5-desaturases resulted in the synthesis of C18 Δ5-desaturated fatty acids (Knutzon et al., 1998). Qi et al. (2004) demonstrated the first examples of the production of C20 VLCPUFAs, ARA and EPA, to ∼ 7% and 3% of total fatty acids in transgenic Arabidopsis leaves, respectively. Wu et al. (2005) demonstrated the feasibility of making VLCPUFAs in transgenic Brassica juncea, and then adopted a step-wise approach to first optimize the accumulation of ω-3 fatty acids (i.e. EPA rather than ARA) and then to convert this to DHA, resulting in the accumulation of ARA (4%), EPA (8%) and DHA (0.2%). A very similar approach was described in a patent application, in which the primary genes for the conversion of oleate to DHA were co-expressed in transgenic soybean somatic embryos, resulting in 3.1% DHA and 6.1% EPA (Kinney, 2004). Collectively, these studies demonstrate the clear potential for the synthesis of VLCPUFAs to meaningful levels (10%–25%) by using the aerobic fatty acid desaturation/elongation pathway.

The attempt to alter the fatty acids profile by the expression of PKS genes in plants has been reported by Metz et al. (2006). The authors used the three subunits (ORFA, B, C) of PKS pathway from Schizochytrium and coexpressed the enzyme phosphopantetheinyl transferase (PPT) from Nostoc, which is essential to activate the ACP domains of the DHA synthase PKS. The expression of these four genes enabled the plants to generate DHA from malonyl-CoA. In this study, each of the four genes (three PKS ORF and one PPT) was expressed individually under control of a linin seed-specific promoter. The targeted expression in the plastid was achieved by the use of an acyl-ACP thioesterase plastid-targeting signal. Arabidopsis seeds accumulated up to 0.8% DHA with an additional 1.7% docosapentaenoic acid (DPA) ω-6 (Metz et al., 2006).

Although many studies were carried out focusing on the production of VLCPUFA in oil plants by the insertion of genes into nuclear genome (independently of the pathway), no attempt was made by genetic engineering of VLCPUFA-related genes into the plastome. As plastid transformation has the advantage to engineer multiple genes in operons (De Cosa et al., 2001; Ruiz et al., 2003; Loessl et al., 2005; Quesada-Vargas et al., 2005; Krichevsky et al., 2010), this tool permits the expression of the different genes (e.g. the three subunits ORFA, B, C and the enzyme phosphopantetheinyl transferase) required for the production of VLCPUFAs in plastids by the polyketide synthase pathway (Metz et al., 2006). Furthermore, plastids contain the substrates (e.g. ACP) required for the production of VLCPUFAs via polyketide synthase system (Ohlrogge and Jaworski, 1997; Metz et al., 2001).

The expression of PUFA synthase system from Schizochytrium in E. coli accumulated free fatty acids as products of the enzymes. This suggests that the thioesterase activity required for fatty acid release is an integral part of the PUFA synthase and they may release their products through a different, as yet not understood, mechanism (Metz et al., 2001). In higher plants, the release of the products of Schizochytrium PUFA synthase would possibly need the use of a VLCPUFA-specific thioesterase. Plants synthesize fatty acids in the plastid and those fatty acids intended for export are cleaved from their cognate ACPs by thioesterases. The free fatty acids are then exported from the plastid by a mechanism that is not well defined and are converted to acyl-CoAs by acyl-CoA synthetases (Koo et al., 2004). However, plant fatty acid synthetases do not produce VLCPUFAs, and it remains to be determined whether there will be problems associated with the export of those novel free fatty acids from plastids (Metz et al., 2009).

Improving plant oil production

Most plant oils produced by agricultural activities are consumed as food and/or animal feed. Although plant oils that are a renewable energy source possess similar properties as compared with petroleum diesel and can be used as logical substitute for the conventional diesel, most oilseed crops have been domesticated focusing on productivity and nutritional properties (Durrett et al., 2008; Dyer et al., 2008). However, their widespread use leads to many challenges. The greatest one is perhaps the limited supply of biodiesel feedstocks. This problem can be solved by increasing the hectares of oil crop farming, increase in oil per plant/hectare and/or the search for new oil-producing platforms (Durrett et al., 2008; Hu et al., 2008; Azócar et al., 2010; Lu et al., 2010). Increasing seed and/or oil yield per plant and, consequently, per hectare seems to be the most promising approach to provide more oil for biodiesel and avoid land competition. The understanding of the different steps of the fatty acids pathway, cloning and characterization of genes and new techniques of genetic engineering open up the door to increase oil yield (Figure 2b). There are evidences that suggest that fatty acid biosynthesis is limited by the early steps (e.g. substrates and enzyme activity) of the pathway (Roesler et al., 1997; Bao and Ohlrogge, 1999; Klaus et al., 2004a). These evidences suggest that the regulation of the biosynthesis can be determined by ACCase activity and the presence of substrates.

As most higher plants have the heteromeric prokaryotic type of ACCase in plastids, any manipulation of the enzymatic complex is quite difficult (Sasaki and Nagano, 2004). The enzyme subunits are regulated by feedback-inhibition, and, additionally, the enzymes are encoded by multiple genes located in the nucleus and in the plastid genome (Thelen and Ohlrogge, 2002; Sasaki and Nagano, 2004). However, the overexpression of the plastid-encoded acetyl-CoA carboxylase β-subunit increased the leaf oil content by 5%–10%, and the number of seeds per plant was about 2-fold higher in comparison with wild-type plants (Figure 2a; Madoka et al., 2002). Interestingly, the fatty acid production per seed was not altered in transgenic plants but plants showed enhanced seed yield, which can serve the basis for developing new strategies to improve commercially important oil-producing crops (Madoka et al., 2002). Genetic manipulation of the homomeric eukaryotic type of the ACCase seems to be a more rational strategy. Roesler et al. (1997) expressed the cytosolic homomeric type of the enzyme into the plastids. The presence of the enzyme in the plastids resulted in a 5% increase in seed oil content. Targeting overexpression of ACCase in amyloplasts led to an increase in fatty acid synthesis and a more than 5-fold increase in the amount of triacylglycerol (Klaus et al., 2004a). This study demonstrated that potato amyloplasts have the capacity for storage-lipid synthesis and that malonyl-CoA, the substrate for elongation during fatty acid synthesis, is a limiting factor for oil accumulation (Klaus et al., 2004a). Many reactions of fatty acid biosynthesis are localized in the plastids and they can be targeted via plastid transformation to increase fatty acids production in plants (Figure 2b). Although the examples of manipulation of some of these enzymes were not carried out only by genetic engineering of plastid genome, they demonstrate the potential for agricultural and industrial utilities of such genetic modifications.

Conclusions and perspectives

Several studies have demonstrated the potential of the expression of fatty acid-related genes in plastids to improve fatty acid biosynthesis or to alter the fatty acid profile, either through nuclear (Roesler et al., 1997; Klaus et al., 2004a) or plastid transformation (Madoka et al., 2002; Craig et al., 2008). These approaches have shown an increase in oil content of about 5%–10% in leaves (Madoka et al., 2002) and in seeds (Roesler et al., 1997). The increase in fatty acid content is of great value when we consider the huge size of planting areas.

In addition to the genetic manipulation of early steps of the fatty acid biosynthesis in plastids, plastome genetic engineering can also be potentially used to produce unusual fatty acids such as VLCPUFAs (Figure 2b). The expression of the different ORFs related to the polyketide synthase system can be made using operons, and the plastid compartment can provide the substrates for this metabolic pathway (Figure 2b). However, the question that remains is about the export of unusual free fatty acids from plastids or the accumulation of them in plastids, because plastids normally do not produce VLCPUFAs (Metz et al., 2001, 2009).

Plastid genetic engineering offers several attractive features and advantages including high transgene expression levels, marker-gene excision (an interesting feature of public concern) and absence of pollen transmission of transgenes because of transgene containment via maternal inheritance of chloroplasts in most higher plants (Maliga, 2004; Bock, 2007; Daniell, 2007; Ruf et al., 2007; Svab and Maliga, 2007). Owing to these interesting features and the localization of fatty acid biosynthesis, plastid transformation has a great potential for the manipulation of genes of this pathway and contribute to the increase in oil content and oil quality for industry or food purposes.

Plastid transformation gives the opportunity to express several genes of different steps of the plastid pathway and can also be used in combination with nuclear transformation aiming to the manipulation of different steps of plant fatty acid biosynthesis (Figure 2b). The characteristic of plastid gene expression opens up the possibility to express, simultaneously, in a single plastid transformation event, genes inducing fatty acid biosynthesis (e.g. acetyl-CoA carboxylase) (Madoka et al., 2002) and genes that alter the fatty acid profile (e.g. Δ9-desaturase) (Craig et al., 2008). Different strategies could also be created as well as the expression of genes acting during early steps of fatty acid biosynthesis via plastid transformation (e.g. acetyl-CoA carboxylase) (Madoka et al., 2002) and later steps of fatty acid biosynthesis via nuclear transformation (e.g. desaturases and elongases) (Damude and Kinney, 2008; Li et al., 2010b; Napier and Graham, 2010; Nguyen et al., 2010).

Finally, the availability of new gene sequence coding for active enzymes (Napier, 2007; Damude and Kinney, 2008; Napier and Graham, 2010), the knowledge of the critical steps of plant fatty acid biosynthesis (Durrett et al., 2008; Dyer et al., 2008; Nguyen et al., 2010) and new techniques of genetic engineering are valuable tools to increase oil content and produce new fatty acids not previously present in plants (Wu et al., 2005; Lu et al., 2010). Characterization of (trans)gene expression in edible plant organs (leaves, fruits and tubers) opens up the possibility to use plastid transformation for the production of healthy nutritional compounds and for biofuel use (Kahlau and Bock, 2008; Valkov et al., 2009). Plastid transformation is available in edible crops (Kumar et al., 2004a; Kanamoto et al., 2006; Apel and Bock, 2009; Valkov et al., 2010) and oil crops (Hou et al., 2003; Skarjinskaia et al., 2003; Dufourmantel et al., 2007; Cheng et al., 2010), which potentially allows yields of a new healthy and environment-friendly product of commercial crops. Results from a number of different studies have demonstrated significant progress towards the goal of improving plant fatty acids. In the most successful examples, this has been achieved through manipulation of enzymes involved in different steps of the fatty acid biosynthesis. Increased understanding of the complexities of the metabolic pathway will be essential to allow a rational design of a metabolism for high-level accumulation of a desired fatty acid. For this, the use of valuable tools offers new possibilities to analyse different steps of the metabolic pathway and to engineer fatty acids with increased industrial and health properties.