Results
1. Construction of an engineered Yarrowia lipolytica to produce Eicosapentaenoic acid (EPA)
The fatty acid profile of the Yarrowia lipolytica wild-type strain Po1f indicated that it can produce palmitic acid (C16:0, 14.9%), palmitoleic acid (C16:1, 6.9%), stearic acid (C18:0, 5.0%), oleic acid (C18:1, 53.8%) and linoleic acid (LA, C18:2,19.3%)，which account for more than 99 percentage in total fatty acid. However, omega-3 polyunsaturated fatty acids cannot be produced by Po1f (Fig. 1). In theory, the introduction of either the Δ-6 pathway or the Δ-9 pathway into the wild-type Y. lipolytica Po1f could produce Eicosapentaenoic acid (EPA) in the presence of desaturases and elongases (Fig. 2). However, the elongation step from γ-linolenic acid (GLA; C18:3n-6) to dihomo-γ-linolenic acid (DGLA; C20:3n-6) was rate limiting, contributing to a large accumulation of GLA (Xue, Z. et al., 2013). Thus, Δ-9 pathway was selected for the Y. lipolytica wild- type strain Po1f (Ura-, Leu-) to produce EPA.
Fig.1 (A) Fatty acid profile of the Y . lipolytica wild-type strain Po1f determined by Gas Chromatography analysis. (B) Different fatty acid percentage of TFAs.
Fig. 2 Schematic diagram of the aerobic pathways for EPA biosynthesis. The native pathway in Y. lipolytica is indicated with the green arrow, and the engineered ∆-9 pathway for EPA biosynthesis is indicated with the yellow arrow.
In Δ-9 pathway, four enzymes are required for the biotransformation of LA to EPA. Therefore, one copy of codon-optimized ∆-9 elongase from Euglena gracilis (EgElo9) and one copy of ∆-8 desaturase from Euglena gracilis(EgDes8) under the control of strong promoter PTEF were integrated into the genome of Y. lipolytica wild-type strain Po1f by non-homologous End Joining (NHEJ), yielding engineered strain Po1f-1. And the percentage of eicosadienoic acid (EDA; C20: 2n-6) and DGLA were 20.8% and 8.5%, respectively (Fig. 3A-B). Furthermore, one copy of ∆-5 desaturase from Euglena gracilis (EgDes5) and one copy of ∆-17 desaturase from Pythium aphanidermatum(PaDes17) continued to be integrated into the genome of the engineered strain Po1f-1, yielding engineered strain Po1f-2. Fig. 3A-B showed that the engineered strain Po1f-2 was able to produce 16.3% of arachidonic acid (ARA; C20:4 n-6) and 2.4% of EPA in the TFAs, which enabled the heterologous synthesis of EPA in the Po1f.
Fig.3 (A) Different fatty acid percentage of TFAs in Po1f-1 and Po1f-2. (B) Fatty acid profiles of the Po1f-1 and Po1f-2 determined by Gas Chromatography analysis.
Furthermore, free fatty acids can be converted to short-chain fatty acyl-coenzyme A via the β-oxidation pathway in the peroxisome, and deletion of the peroxisome generator peroxin-10 (PEX10) would disrupt the β-oxidation pathway, thereby preventing the degradation of free fatty acids. Therefore, homologous recombination was used to knock out PEX10 of the genome of the engineered strain Po1f-2 to construct the engineered strain Po1f-3. The GC analysis illustrated that EPA content in the TFAs was further enhanced to 4.5%, which is about 1.88 times more than that of the engineered strain Po1f-2 (Fig. 4).
Fig.4 Different fatty acid percentage of TFAs in Po1f-3.
2. Screening of different sources of elongases/desaturases
Fig. 1 revealed that C18:1 and C18:2 of the engineered strain Po1f accounted for a high percentage of TFAs, among which C18:1 accounting for approximately more than 50% of the TFAs. Besides, Δ-12 desaturase is able to convert C18:1 to C18:2 by adding a double bond to the carbon chain through dehydrogenation, which suggests that the activity of endogenous Δ-12 desaturase of Y. lipolytica Po1f was low (Fig. 1). Moreover, studies have demonstrated that elongases and desaturases form different sources have different catalytic activities and substrate specificities for fatty acid substrates, so the selection of elongases/desaturases with high substrate specificity is a significant strategy to improve the synthesis of EPA. For example, heterologous expression of a highly substrate-specific Δ6 desaturase in Mortierella alpina contributed to a remarkable 26-fold increase in EPA content (Shi, H. et al., 2016). Po1f was used as the starting strain and the ∆-12 desaturase from Fusarium moniliforme (FmDes12), Coprinus cinereus (CcDes12), and Parietichytrium sp. (PaDes12) were integrated at the Intc site of its genome (Damude, H. G. et al., 2006; Ishibashi, Y. et al., 2021a; Zhang, S. et al., 2007), the engineered strains were named Po1f-4, Po1f-5 and Po1f-6 respectively. Besides, GC analysis was used to illustrate the conversion of endogenous fatty acids catalyzed by different ∆-12 desaturase. Fig. 6 indicated that the enzyme activity of FmDes12 was the best, which results in C18:1 decreasing from 53.8% to 9.5%, while C18:2 reached 62.7% of TFAs. And the ratio of C18:2/C18:1 went from 0.378 to 6.62, with a conversion rate of 81.1% for C18:1. While the conversion of CcDes12 from C18:1 to C18:2 was 62.7% and PaDes12 was the least effective catalyst for only 37% conversion, so Po1f-4 was used to perform subsequent validation of elongases/desaturases effects (Fig. 5).
Fig.5 Different fatty acid percentage of TFAs produced by different sources of ∆-12 desaturase.
In addition to the high proportion of C18:1, C18:2 also accounts for high percentage of the TFAs, especially up to 62.7% in Po1f-4. To reduce the percentage of C18:2, six different sources of ∆-9 elongase gene from Blattella germanica (BgElo9), Emiliania huxleyi (EhElo9), Isochrysis galbana (IgASE1), Isochrysis galbana (IgASE2), Pavlova pinguis (PpElo9), and Pavlova salina (PsElo9) were selected (Juárez, M. P., 2004; Li, M. et al., 2012; Petrie, J. R. et al., 2010; Qi, B. et al., 2002), which were reported as elongases with a preference for C18:2 in the previous study. Similarly, Po1f-4 was used as the starting strain and the ∆-9 desaturase above were integrated at the Scp2 site of its genome, the engineered strains were named Po1f-7, Po1f-8, Po1f-9, Po1f-10, Po1f-11, and Po1f-12 respectively. Fig. 6 indicated that the enzyme activities of IgASE2 and EhElo9 were the best, with 40.0% and 35.2% of C20:2 generated, respectively. Besides, C18:2 produced by them decreased from 62.7% to 22.2% and 32.8%, and the conversion of C18:2 to C20:2 reached 63.8% and 56.1%, respectively. IgASE1 and BgElo9 showed the worst enzyme activities towards C18:2. The conversion of C18:2 to C20:2 under the catalytic action of IgASE1 was around 26.3%, with only 16.5% of C20:2 produced. However, Po1f-7 rarely produced C20:2, so BgElo9 was thought to have no preference towards C18:2 (Fig. 6). Taken together, IgASE2 and EhElo9 had a strong preference for C18:2 and could significantly promote the synthesis of C20:2.
Fig.6 Different fatty acid percentage of TFAs produced by different sources of ∆-9 elongase.
In addition to exploring the substrate preferences of the two main fatty acid in Po1f-3, it was found that the content of C16:0 accounted for about 15% of the TFAs of strain Po1f-3. Thus, C16/18 elongase from Mortierella alpina (MaElo2), Misgurnus anguillicaudatus (MaElo16), Parietichytrium sp. (PaElo16), and rat (rElo2) were integrated at the Scp2 site of its genome (Chen, J. et al., 2018; Ishibashi, Y. et al., 2021b; Wang, F. et al., 2019; Yazawa, H. et al., 2011), and the engineered strains were named Po1f-13, Po1f-14, Po1f-15, and Po1f-16 respectively. It was indicated that MaElo2 and rElo2 were the most effective catalysts for C16:0, with the C16:0 content reduced to 10.5% and 12.4% of TFAs, respectively. It is a remarkable fact that the four C16/18 elongases also have catalytic activity for C18:1 as well because the percentage of C18:2 they produced were almost doubled, which suggested that the C16/18 elongases may have both elongase and desaturase activities (Fig. 7).
Fig.7 Different fatty acid percentage of TFAs produced by different sources of C16/18 elongase.
3. Optimization of EPA yield using push-pull strategy
3.1 Increasing gene copy number
In order to reduce the amounts of intermediates and increase the amount of EPA, elongases/desaturases above were randomly integrated in the genome. Besides, the optimal copy numbers of the elongases/desaturases were detected by QPCR. Firstly, MaElo2 was randomly integrated into the Po1f-3 genome because the content of C16:0 accounted for about 15% of the TFAs of strain Po1f-3. And it was found that MaElo2 with three copies (Po1f-17) showed the best result and EPA percentage of total fatty acids it produced was 7.3% (Fig. 8A-B). Similarly, Po1f-18 was achieved by integrating two copies of FmDes12 and the proportion of EPA it obtained was 9.6% (Fig. 8A-B).
Fig.8 (A) Different fatty acid percentage of TFAs produced by random integration of the best strains of elongases/desaturases. (B) Optimal copy numbers detected by QPCR. (C) Fatty acid profiles of the Po1f-3 and Po1f-20 determined by Gas Chromatography analysis.
Furthermore, EhElo9, IgASE2 and EgDes8 were randomly integrated in the genome of Po1f-18 to obtain Po1f-19. It can be found that EPA Po1f-19 produced made up 16.3% of TFAs, and the integration of 3 copies of the EhElo9, 3 copies of the IgASE2 and 2 copies of the EgDes8 gene in this strain was demonstrated by QPCR (Fig. 8A-B). Po1f-20 was further randomly integrated with 2 copies of EgDes5 and 2 copies of PaDes17 in the genome of the engineered strain Po1f-19, and EPA proportion reached 22.4% of the TFAs (Fig. 8A-B).
3.2 Optimizing promoters
In addition to increasing copy numbers, regulating the strength of promoters is also a potential strategy for adjusting EPA content. Therefore, Po1f-3 was used as the starting strain, EhElo9 was ultilized to verify whether the strength of promoter had an effect on the EPA ratio. And different strong constitutive promoters including PGAP, PEXP, PYAT, PFBA in and PTEF in was used to instead of PTEF, which was integrated at the Scp2 site of Po1f-3 genome respectively. However, it was found that there was no significant change in fatty acid content under different promoters (Fig. 9), which demonstrates that the promoter optimization did not have a significant effect on EPA yield.
Fig.9 Different fatty acid percentage of TFAs produced by engineered strains transferred of EhElo9 controlled by different promoters.
4. Modification of Yarrowia lipolytica for co-production of EPA and DHA
In order to obtain an engineered strain for co-production of DHA and EPA production, Po1f-20 was used as the starting strain. Besides, ∆-5 elongase and Δ-4 desaturase were identified to convert EPA to DHA, so ∆-5 elongase from Thraustochytrium sp. and Δ-4 desaturase from Schizochytrium sp. were randomly integrated in the genome of Po1f-20 to produce Po1f-21. The GC analysis indicated that the percentage of EPA and DHA the engineered strain produced were 2.5% and 18.9% of TFAs respectively, indicating successful heterologous synthesis of DHA and EPA in the Yarrowia lipolytica (Fig. 10).
Fig.10 (A) Fatty acid profiles of the Po1f-20 and Po1f-21 determined by Gas Chromatography analysis. (B) Different fatty acid percentage of TFAs in Po1f-20 and Po1f-21.
References
Chen, J., Cui, Y., Yan, J., Jiang, J., Cao, X., and Gao, J., 2018. Molecular characterization of elongase of very long-chain fatty acids 6 (elovl6) genes in Misgurnus anguillicaudatus and their potential roles in adaptation to cold temperature. Gene. 666, 134-144.
Damude, H. G., Zhang, H., Farrall, L., Ripp, K. G., Tomb, J. F., Hollerbach, D., and Yadav, N. S., 2006. Identification of bifunctional Δ12/ω3 fatty acid desaturases for improving the ratio of ω3 to ω6 fatty acids in microbes and plants. Proceedings of the National Academy of Sciences. 103, 9446-9451.
Ishibashi, Y., Goda, H., Hamaguchi, R., Sakaguchi, K., Sekiguchi, T., Ishiwata, Y., Okita, Y., Mochinaga, S., Ikeuchi, S., Mizobuchi, T., Takao, Y., Mori, K., Tashiro, K., Okino, N., Honda, D., Hayashi, M., and Ito, M., 2021a. PUFA synthase-independent DHA synthesis pathway in Parietichytrium sp. and its modification to produce EPA and n-3DPA. Commun Biol. 4, 1378.
Ishibashi, Y., Goda, H., Hamaguchi, R., Sakaguchi, K., Sekiguchi, T., Ishiwata, Y., Okita, Y., Mochinaga, S., Ikeuchi, S., Mizobuchi, T., Takao, Y., Mori, K., Tashiro, K., Okino, N., Honda, D., Hayashi, M., and Ito, M., 2021b. PUFA synthase-independent DHA synthesis pathway in Parietichytrium sp. and its modification to produce EPA and n-3DPA. Communications Biology. 4, 1378.
Juárez, M. P., 2004. Fatty acyl-CoA elongation in Blatella germanica integumental microsomes. Archives of Insect Biochemistry and Physiology. 56, 170-178.
Li, M., Ou, X., Yang, X., Guo, D., Qian, X., Xing, L., and Li, M., 2012. Cloning and identification of a novel C18-Δ9 polyunsaturated fatty acid specific elongase gene from DHA-producing Isochrysis galbana H29. Biotechnology and Bioprocess Engineering. 17, 22-32.
Petrie, J. R., Mackenzie, A. M., Shrestha, P., Liu, Q., Frampton, D. F., Robert, S. S., and Singh, S. P., 2010. Islation of three novel long-chain polyunsaturated fatty acid Δ9-Elongases and the transgrntic assembly of the the entire palova salina docosahexaenoic acid pathway in Nicotiana benthamiana1. Journal of Phycology. 46, 917-925.
Qi, B., Beaudoin, F., Fraser, T., Stobart, A. K., Napier, J. A., and Lazarus, C. M., 2002. Identification of a cDNA encoding a novel C18-Δ9 polyunsaturated fatty acid-specific elongating activity from the docosahexaenoic acid (DHA)-producing microalga, Isochrysis galbana 1. FEBS Letters. 510, 159-165.
Shi, H., Chen, H., Gu, Z., Zhang, H., Chen, W., and Chen, Y. Q., 2016. Application of a delta-6 desaturase with alpha-linolenic acid preference on eicosapentaenoic acid production in Mortierella alpina. Microb Cell Fact. 15, 117.
Wang, F., Bi, Y., Diao, J., Lv, M., Cui, J., Chen, L., and Zhang, W., 2019. Metabolic engineering to enhance biosynthesis of both docosahexaenoic acid and odd-chain fatty acids in Schizochytrium sp. S31. Biotechnol Biofuels. 12, 141.
Xue, Z., Sharpe, P. L., Hong, S. P., Yadav, N. S., Xie, D., Short, D. R., Damude, H. G., Rupert, R. A., Seip, J. E., Wang, J., Pollak, D. W., Bostick, M. W., Bosak, M. D., Macool, D. J., Hollerbach, D. H., Zhang, H., Arcilla, D. M., Bledsoe, S. A., Croker, K., McCord, E. F., Tyreus, B. D., Jackson, E. N., and Zhu, Q., 2013. Production of omega-3 eicosapentaenoic acid by metabolic engineering of Yarrowia lipolytica. Nat Biotechnol. 31, 734-40.
Yazawa, H., Kamisaka, Y., Kimura, K., Yamaoka, M., and Uemura, H., 2011. Efficient accumulation of oleic acid in Saccharomyces cerevisiae caused by expression of rat elongase 2 gene (rELO2) and its contribution to tolerance to alcohols. Appl Microbiol Biot. 91, 1593-1600.
Zhang, S., Sakuradani, E., Ito, K., and Shimizu, S., 2007. Identification of a novel bifunctional Δ12/Δ15 fatty acid desaturase from a basidiomycete, Coprinus cinereus TD#822-2. FEBS Letters. 581, 315-319.