Metabolic Engineering of Nonplant Pathways

The potential also exists to engineer seeds to produce oils with fatty acids not normally found in plants. Currently, few examples exist for the transgenic production of fatty acids that do not normally occur in plants. However, genes from sources other than plants have been used to alter the fatty acid composition of seed oils. Early examples include the use of genes for the yeast and mouse Δ⁹-stearoyl-CoA desaturase and the Synechocystis Δ⁶-linoleic acid desaturase to produce small amounts of novel fatty acids in plants (Moon et al., 2000; Reddy and Thomas, 1996). Δ⁶-linoleic acid desaturase genes from the fungi M. alpina and P. irregulare have been used to generate seed oils in Brassica species that contain >40% GLA (Das et al., 2000; Hong et al., 2002). This amount of GLA is comparable to that obtained by transgenic expression of the borage Δ⁶ desaturase in soybean seeds (Sato et al., 2004).

Perhaps one of the greatest challenges for metabolic engineering of seed oil composition will be the transgenic production of high levels of the long-chain o-3 fatty acids eicosapentaenoic acid (EPA; 20:5Δ^5,8,11,14,17) and DHA (22:6Δ^{4,7,10,13,16,19}) (Table 7.3). Though EPA can be found in trace amounts in the seed oil of certain gymnosperm species, this fatty acid and DHA are typically absent from seed oils but are major components of many fish and algal oils. Diets rich in the ω-3 fatty acids EPA and DHA have been linked to enhanced cardiovascular fitness (Hu et al., 2001). In addition, DHA has been shown to improve brain development when supplemented in infant diets (Uauy et al., 2003). DHA is present in mother’s milk, but is absent from infant formula prepared from soybean and other vegetable oils. Highly refined fish oils that contain EPA and DHA can sell for >$10/pound, whereas conventional soybean oil sells for about $0.20/ pound. The ability to engineer EPA and DHA synthesis in oilseeds therefore offers a means for significantly increasing not only nutritional quality, but also economic value of vegetable oils.

A number of pathways might lead to the production of EPA in seeds (Napier, 2007). The most direct biosynthetic route would involve the introduction of three new enzymes: (1) a Δ⁶ desaturase for the conversion of α-linolenic acid to stearidonic acid, (2) an ELO-type fatty acid elongase to initiate the elongation of stearidonic acid to ARA (20:4Δ^8,11,14,17), and (3) a Δ⁵ desaturase for formation of EPA from ARA (Fig. 7.4). In addition, most oilseeds contain relatively low amounts of a-linolenic acid, the initial substrate for this pathway. As a result, the production of EPA in the seeds of most crop plants would require not only the introduction of genes for the three enzymes described above but also enhanced expression of the FAD3 gene to increase Δ¹⁵-linoleic acid desaturase activity. Reports have demonstrated the feasibility of assembling the EPA biosynthetic pathway into leaves and seeds of transgenic plants (Abbadi et al., 2004; Qi et al., 2004; Robert et al., 2005;Wu et al., 2005).

The production of DHA in oilseeds requires the transgenic expression of genes for at least two additional enzymes: (1) an ELO elongase to initiate the elongation of EPA to the C₂₂ fatty acid docosapentaenoic acid (DPA, 22:5Δ^{7,10,13,16,19}) and (2) a Δ⁴ ‘‘front-end’’ desaturase to convert DPA to DHA (Fig. 7.4). Genes for Δ⁴ desaturases capable of catalyzing this conversion of DPA have been isolated from E. gracilis and Thraustochytrium sp (Meyer et al., 2003; Qiu et al., 2001a). Indeed, production of small amounts of DHA in seeds of Brassica juncea was recently achieved by the transgenic expression of genes for an Oncorhynchus mykiss C₁₈/ C₂₀-specific ELO and the Thraustochytrium Δ⁴ desaturase along with genes for a P. irregulare Δ⁶ desaturase, a Phytophthora infestans ω3 desaturase, Thraustochytrium Δ⁵ desaturase, acyltransferase and ELO, a C. officinalis Δ¹² desaturase, and a P. patens ELO (Wu et al., 2005). Although the introduction of this pathway into seeds of a transgenic plant via the overexpression of nine transgenes is a remarkable metabolic engineering feat, the low amounts of DHA produced (0.2% of the total fatty acids) indicate that bottlenecks exist for achieving high levels of this commercially valuable fatty acid. In this regard, biochemical studies have indicated that the desaturase and ELO elongases that are required for EPA and DHA synthesis employ fatty acid substrates esterified to different molecules (Domergue et al., 2003). The Δ⁵ and Δ⁶ desaturases from plant, fungal, and algal species use fatty acids bound to PC as their preferred substrates. In contrast, ELO elongases accept acyl-CoAs as substrates. The need for fatty acid substrates to move between pools of PC and acyl-CoAs may limit flux through the engineered EPA and DHA biosynthetic pathways. Furthermore, polyunsaturated fatty acids (PUFAs) do not appear to be efficiently channeled into TAG and excluded from accumulating in membrane lipids of engineered seeds (Abbadi et al., 2004; Wu et al., 2005). Whether membrane-associated PUFAs will impair the agronomic fitness of seeds has yet to be addressed.

An alternative mechanism for EPA and DHA biosynthesis has been demonstrated in the marine bacterium Shewanella sp. and in the thraustochytrid Schizochytrium sp. These organisms contain polyketide synthases consisting of multifunctional polypeptides capable of synthesizing EPA and DHA in an anaerobic manner (Metz et al., 2001). The possibility of transferring these pathways to seeds appears daunting considering that EPA synthesis in Shewanella, for example, results from 5 open-reading frames that code for at least 11 different protein domains. These domains include polypeptides that are related to 3-ketoacyl synthases, ACP, enoyl reductases, dehydratases, and acyltransferases.