Production of pomegranate-derived punicic acid

Abstract

Engineering the oleaginous yeast Rhodosporidium toruloides to produce pomegranate-derived punicic acid and pomegranate-derived punicic acid accumulation in Saccharomyces cerevisiae by Ty retrotransposon-targeted random gene shuffling.

Description

PRODUCTION OF POMEGRANATE-DERIVED PUNICIC ACID FIELD

[0001] The present disclosure relates generally engineering the oleaginous yeast Rhodosporidium toruloides to produce pomegranate-derived punicic acid and increasing 5 punicic acid accumulation in Saccharomyces cerevisiae by Ty retrotransposon-targeted random gene shuffling. BACKGROUND

[0002] Plant lipids have significant nutritional, industrial, and pharmaceutical 10 applications. Specific plant species have the remarkable ability to synthesize unusual fatty acids (UFAs), which possess distinct chemical structures compared to common fatty acids. These UFA underlined the biochemical diversity among plant lipid synthetic pathways and offer novel lipid products for industrial and commercial applications. Punicic acid (PuA), also known as conjugated linolenic acid, is an unusual polyunsaturated fatty acid (PUFA) with three 15 conjugated double bonds. The primary natural source of punicic acid is pomegranate (Punica granatum), which contains 80% punicic acid in its seed oil. Studies have indicated that punicic acid offers a wide range of health benefits, including anti-diabetic, anti-obesity, and anti- carcinogenic activities. Moreover, similar to eleostearic acid, which also features conjugated double bonds and provides drying properties to tung oil, PuA holds promise as a cross-linking 20 agent in the production of polymeric materials, such as high-quality resins and drying oils.

[0003] Punicic acid (PuA), a conjugated unusual fatty acid found in pomegranate (Punica granatum) seed oil, has been extensively studied for its health benefits. The three conjugated double bonds in the PuA molecule confer a wide range of bioactive properties, including anti-inflammatory, antioxidant, anti-carcinogenic, and anti-diabetic effects (Machado 25 et al., 2022; Tsuzuki et al., 2006; Vroegrijk et al., 2011). Studies have shown that PuA can improve insulin sensitivity and inhibit the growth of certain cancer cells by the modulation of lipid metabolism and gene expression related to cell proliferation (Costantini et al., 2014; Mele et al., 2013). Due to these properties, PuA and pomegranate seed oil are considered promising dietary supplements in nutraceutical applications (Aruna et al., 2016; Shabbir et al., 2017). 30

[0004] The cultivation of pomegranate on a large scale for PuA extraction presents significant challenges, including a low seed-to-fruit ratio and limited production of seed oil. Although initial efforts have been made to engineer PuA-producing oilseed plants, the content of PuA in seed triacylglycerol (TAG) is still quite low so far. Considering plants with PuA - 1 -accumulation potential face limitations such as climate-dependency and space requirements, microorganisms offer a promising alternative due to their advantages in growth rate, the ability to use renewable feedstock, and minimal space needs.

[0005] For this reason, previously we have engineered baker's yeast (Saccharomyces 5 cerevisiae), a Generally Recognized as Safe (GRAS) microorganism, to produce PuA. In pomegranate, bifunctional acyl lipid desaturase and conjugase (PgFADX) converts the common fatty acid linoleic acid (LA) on the sn-2 position of phosphatidylcholine (PC) to PuA. However, the coexpression of this key enzyme with 12 acyl lipid desaturase (PgFAD2) from pomegranate only led to 0.3% PuA in S. cerevisiae. Since S. cerevisiae is a conventional yeast 10 with low lipid content, the initial low level of PuA could be attributed to the inefficiency in lipid synthesis. However, PuA production remained fairly low in Yarrowia lipolytica, a well-studied oleaginous yeast. Even in its obese form, Y. lipolytica expressing PgFADX only results in the accumulation of 0.5% PuA of total esterified fatty acids. This implies that the efficiency of lipid synthesis alone may not be the sole factor contributing to the low ratio of PuA in yeast cells. 15

[0006] Oleaginous yeasts, such as Trichosporon fermentans, Yarrowia lipolytica and Rhodosporidium toruloides, can accumulate more than 20% of lipids in their biomass. For instance, T. fermentans was able to convert rice straw hydrolysate or waste molasse to 40% - 60% of single cell oil (Huang et al., 2009; Zhu et al., 2008). Y. lipolytica, a well-studied model oleaginous yeast, has the ability to grow on various carbon sources including fatty acids, 20 glucose, fructose, or glycerol, and accumulating 36% of its biomass as lipids (Zhang et al., 2014). R. toruloides, an unconventional yeast capable of concomitant synthesis of lipids and carotenoids, is typically regarded as a promising candidate for biofuel production due to its excellent lipid accumulation capability, which often - 4 - exceeds 60% of dry cell weight (Ratledge and Wynn, 2002). Furthermore, R. toruloides' capacity to use a wide range of 25 substrates, including glucose and xylose, makes it a promising option for sustainable bioproduction processes (Boviatsi et al., 2020; Qi et al., 2020).

[0007] In addition to the key enzymes that directly catalyze the reaction, pathways upstream and downstream of the PuA synthesis in the transgenic host may not effectively synergize with each other to achieve optimal production. Since S. cerevisiae has a relatively 30 simple lipid profile and contains no LA, an efficient supply of fatty acid precursors is crucial for this host. Previous studies have also shown the enrichment of UFA in plant oil involves a series of enzymatic reactions that occur in developing seeds. These reactions form the acyl-editing and TAG assembly network, which transfers nascent UFA from phospholipids to the storage - 2 -lipid (TAG). Because this network dominates the flow of UFA into storage lipids and enhances their accumulation in plant seeds, the UFA levels in the seed TAG of natural producers are significantly higher compared to the engineered transgenic host. Pomegranate seeds, for instance, can accumulate 60% of TAG as PuA, but engineered Brassica napus and 5 Arabidopsis thaliana only accumulate 6.6% and 10.6% of PuA in seed TAG fractions, respectively.

[0008] Given the complexity of the pathways involved in plant-derived UFA synthesis, it is time-consuming to test each gene separately to achieve pathway functionality. Thus, a more effective process is required to facilitate the study of plant UFA metabolism and the 10 heterologous synthesis of plant-derived lipids in microorganisms. SUMMARY

[0009] In one of more embodiments of the present disclosure, there is provided:

[0010] .1. An isolated recombinant Rhodosporidium toruloides (R. toruloides) 15 cell, comprising a polynucleotide encoding Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX).

[0011] 2. The isolated recombinant R. toruloides cell of embodiment 1, further comprising a polynucleotide encoding Punica granatum delta-12 acyl lipid desaturase (PgFAD2), diacylglycerol acyltransferase 2 (PgDGAT2), or phosphatidylcholine: 20 diacylglycerol cholinephosphotransferase (PgPDCT).

[0012] 3. The isolated recombinant R. toruloides cell of embodiment 1 or 2, wherein said polynucleotide encoding PgFADX is a codon-optimized polynucleotide encoding PgFADX.

[0013] 4. Use of the isolated recombinant R. toruloides cell of any one of 25 embodiments 1 – 3 for producing punicic acid (PuA).

[0014] 5. A method of producing recombinant punicic acid (PuA) comprising, providing a recombinant Rhodosporidium toruloides (R. toruloides) cell of any one of embodiments 1 to 3, and culturing said recombinant R. toruloides cell in a culture medium under culture conditions that allow for the production of PuA. 30

[0015] 6. The method of embodiment 5, wherein the culture conditions comprise a temperature of about 30°C for about 1-10 days.

[0016] 7. The method of embodiment 5 or 6, wherein the culture medium is a nitrogen-limited medium. - 3 -

[0017] 8. The method of any one of embodiments 5-7, wherein the culture medium comprises wood hydrolysate.

[0018] 9. The method of any one of embodiments 5 to 8, wherein about 5.8% to about 11.98% of total fatty acids is PuA. 5

[0019] 10. A recombinant Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX) obtained by the method of any one of embodiments 5 to 9.

[0020] 11. An isolated recombinant Saccharomyces cerevisiae cell comprising a polynucleotide encoding a fusion protein comprising: a substrate-binding / cosubstrate- providing / inhibitor-removing protein, a linker, and a Punica granatum bifunctional acyl lipid 10 desaturase and conjugase (PgFADX), wherein the N-terminal amino acid of the linker is attached the C-terminal amino acid of the substrate-binding / cosubstrate-providing / inhibitor- removing protein, and the C-terminal amino acid of the linker is attached to the N terminal amino acid of the PgFADX.

[0021] 12. The isolated recombinant S. cerevisiae cell of embodiment 1, wherein15 the substrate binding protein is phosphatidylcholine-binding proteins (SCP2), oxygen carrier- proteins (Vhb), soluble domain of electron transporter (CB5SD) or catalase (CAT).

[0022] 13. The isolated recombinant S. cerevisiae cell of embodiment 11 or 12, wherein the linker comprises or consists of the amino acid sequence ASGAGGSEGGGSEGGTSGAT (SEQ ID NO: 20) 20

[0023] 14. The isolated recombinant S. cerevisiae cell of any one of embodiments 11 to 13, comprising a snf2 mutant strain or snf2 snf1 double mutant strain.

[0024] 15. The isolated recombinant S. cerevisiae cell of any one of embodiments 11 to 14, wherein the recombinant Saccharomyces cerevisiae yeast strain is a snf2 snf1double knockout strain. 25

[0025] 16. An isolated recombinant Saccharomyces cerevisiae yeast strain comprising a polynucleotide encoding AtCB5SD-PgFADX, wherein the yeast strain is a BY4741 snf2 snf1 double knockout mutant strain.

[0026] 17. The isolated recombinant S. cerevisiae cell of embodiment 16, further comprises a polynucleotide encoding PDCT, a polynucleotide encoding LPCAT and a 30 polynucleotide encoding DGAT2.

[0027] 18. The isolated recombinant S. cerevisiae cell of embodiment 17, further comprising a polynucleotide encoding FAD2 and ELO2. - 4 -

[0028] 19. A method of producing recombinant punicic acid (PuA) comprising, providing a recombinant Saccharomyces cerevisae cell of any one of embodiments 11 – 18, and culturing said recombinant S. cerevisae cell in a culture medium under culture conditions that allow for the production of PuA 5

[0029] 20. The method of embodiment 19, wherein the culture medium comprises about 6-15 % glucose.

[0030] 21. The method of embodiment 19 or 20, wherein the isolated recombinant S. cerevisiae cell is cultured in the presence of 0-0.03% linoleic acid.

[0031] 22. Use of the isolated recombinant S. cerevisiae cell of any one of 10 embodiments 11 – 18 for producing punicic acid.

[0032] 23. A recombinant Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX) obtained by the method of any one of embodiments 19 to 21.

[0033] 24. Use of a Ty retrotransposon-targeted random gene shuffling approach for plant-derived punicic acid production in yeast. 15

[0034] 25. Use of a Ty retrotransposon-targeted random gene shuffling approach for unusual fatty acid synthesis in yeast cell.

[0035] 26. The use of embodiment 22, or the method of any one of embodiments 19 – 21, wherein the PuA content producted is greater than about 20% TFA. 20 BRIEF DESCRIPTION OF THE FIGURES

[0036] Embodiments of the present disclosure will now be described, by way of example only, with reference to the attached Figures.

[0037] Fig.1. Metabolic engineering of PuA-producing R. toruloides. A) Illustration of the metabolic pathway involved in PuA biosynthesis in R. toruloides. Abbreviations: PgDGAT2, 25 P. granatum acyl-CoA: diacylglycerol acyltransferase 2; PgFAD2, P. granatum fatty acid desaturase 2; PgFADX, P. granatum fatty acid desaturase and conjugase; PgPDCT, P. granatum phosphatidylcholine: diacylglycerol cholinephosphotransferase; GA-3-P, glyceraldehyde 3-phosphate; DHAP, dihydroxyacetone phosphate; G-3-P, glycerol-3- phosphate; LPA, lysophosphatidic acid; PA, phosphatidic acid; DAG, diacylglycerol; TAG, 30 triacylglycerol; PC, phosphatidylcholine. B) Donor DNA structures and genomic integrations of codon-optimized PgFADX, PgFAD2, PgDGAT2, and PgPDCT into R. toruloides. C) Colony colour of R. toruloides wild-type strain as well as engineered RX1, RX2, RX3 and RX4 strains. D) Detection of positive transformants. Lane 1, GeneRuler 1 kb Plus DNA ladder; Lane 2 / 4 / 6 / 8, - 5 -R. toruloides wild type strain; Lane 3 / 5 / 7 / 9, detection of integrated cassettes in strain RX1, RX2, RX3 and RX4.

[0038] Fig.2. Genomic integration of pomegranate PgFADX expression cassette led to PuA-producing R. toruloides strain RX1. A) Fatty acid composition of wild-type R. toruloides 5 cultured in chemically defined medium. B) Integration of PgFADX led to PuA product. C) Fatty acid composition and PuA titer of RX1 strain from 48h to 240h. Data represent mean ± SD of triplicates.

[0039] Fig.3. Coexpression of pomegranate PgFAD2, PgDGAT2, and PgPDCT respectively with PgFADX led to increased PuA production. A-C) Fatty acid composition of R. 10 toruloides strains coexpressing PgFAD2-PgFADX, PgDGAT2-PgFADX, and PgPDCT- PgFADX. D-F) Fatty acid composition and PuA titer of R. toruloides strains RX2, RX3 and RX4 from 48h to 240h. Data represent mean ± SD of triplicates.

[0040] Fig.4. Analysis of PgFADX expression levels and PuA distribution in strain RX2. A) Relative expression of PgFADX gene in recombinant R. toruloides strain RX1, RX2, RX3 15 and RX4. R. toruloides GPD1 gene was used as the internal reference gene. B) TLC separation of lipids extracted from wild-type strain cultured in 2*YPD medium, as well as strain RX1, RX2, RX3, and RX4 cultured in nitrogen-limited medium. C) Fatty acid composition of TAG in strain RX2. D) Fatty acid composition of polar lipids in strain RX2. E) Fatty acid composition at the sn-2 position of TAG in strain RX2. F) Fatty acid composition at the sn-1 / sn- 20 3 positions of TAG in strain RX2. Data represent mean ± SD of triplicates.

[0041] Fig.5. Converting wood hydrolysate into PuA-containing single-cell oil by strain RX2. A) Examining the influence of wood hydrolysate concentration and initial OD600 on PuA production. B) Fatty acid composition and PuA titer of R. toruloides strains RX2 cultured in 100% wood hydrolysate from 48h to 240h. Data represent mean ± SD of triplicates. 25

[0042] Fig.6. TAG and UFA synthetic pathways. (A) Rewiring lipid synthetic pathway in S. cerevisiae to produce PuA. DGAT / DGA1, acyl-CoA: diacylglycerol acyltransferase; GPAT / SCT1 / GPT2, glycerol-3-phosphate acyltransferase; LACS / FAA, long-chain acyl-CoA synthetase; LPAT / SLC1, lysophosphatidic acid acyltransferase; LPC, lysophosphatidylcholine; LPCAT, lysophosphatidylcholine acyltransferase; PAP / APP1 / PAH1, 30 phosphatidate phosphatase; PC, phosphatidylcholine; PDAT / LRO1, phospholipid: diacylglycerol acyltransferase; PDCT, phosphatidylcholine: diacylglycerol cholinephosphotransferase; PLA2, phospholipase A2; ACC1, acetyl-CoA carboxylase; FAS, fatty acid synthase; OLE1, acyl-CoA desaturase; ELO, fatty acid elongase; LOA1, - 6 -lysophosphatidic acid: oleoyl-CoA acyltransferase; DGK1, diacylglycerol kinase; CDS1, phosphatidate cytidylyltransferase; EPT1, choline / ethanolamine phosphotransferase; PIS1, CDP-diacylglycerol--inositol 3-phosphatidyltransferase; PSD, phosphatidylserine decarboxylase; CPT, cholinephosphotransferase; PLC, phospholipase C; ALE1, 5 lysophospholipid acyltransferase; FADX, bifunctional fatty acid conjugase / Delta(12)-oleate desaturase; FAD2, delta(12)-fatty-acid desaturase; PI, phosphatidylinositol; PE, phosphatidylethanolamine; CHO1, CDP-diacylglycerol-serine O-phosphatidyltransferase; PS, phosphatidylserine. (B, C) Functional complementation assay in yeast strain H1246 without / with pomegranate oil hydrolysate. Lane 1-9, pomegranate oil, empty vector control, 10 PgDGAT1, PgDGAT2a, PgDGAT2b, PgDGAT2c, PgPDAT.a, PgPDAT.b, PgPDAT.c respectively. (D) Fatty acid profiles of recombinant H1246 strains expressing different TAG assembly genes in the absence or presence of pomegranate oil hydrolysate. (E) PuA level in yeast strains transformed with various PgFADX protein fusions.

[0043] Fig.7. Workflow of the Ty retrotransposon-targeted random gene shuffling. (A) 15 Schematic depicting the screening procedure for Ty retrotransposon-targeted random gene shuffling by CRISPR-Cas9. (B) Schematic of donor DNA design for integration into yeast Ty retrotransposon.

[0044] Fig.8. PuA production in S. cerevisiae strains constructed by Ty retrotransposon-targeted random gene shuffling. (A, C, E) 3D scatter map of the results from 20 round 1, round 2, and round 3 CRISPR-Cas9 assisted random integration. Each red dot represents an individual yeast transformant. Three-dimensional scatter plot visualizing each yeast transformant's PuA content (represented by the percentage of PuA in total fatty acids, Z-axis), PuA amount (represented by the integrated area of PuA's peak, Y-axis), and neutral lipid level (represented by the Nile red fluorescence, X-axis). (B, D, F) Fatty acid composition 25 and PuA production of CARIA266, CARIB650 and CARIC568 at different incubation time under different culturing temperatures.

[0045] Fig.9. Growth condition optimization and the relative content of PuA in TAG and polar lipids. (A, B) Three-dimensional surface plot of PuA content to carbon source level, initial pH, and initial OD. (C) Fatty acid composition of TAG. (D) Fatty acid composition of polar lipids. 30 (E) Fatty acid composition at the sn-2 position of TAG. (F) Fatty acid composition at the sn-1 position of polar lipids. (G) The content of fatty acid at the sn-1 / sn-3 positions of TAG. (H) The content of fatty acid at the sn-2 position of polar lipids. Data represent mean ± SD of triplicates. - 7 -

[0046] Fig.10. Yeast lipidomes containing PuA with or without LA feeding. (A) Heatmap of top-ranking SFA-containing TAG. (B) Heatmap of top-ranking MFA-containing TAG. (C) Heatmap of top-ranking PUFA-containing TAG. (D) Heatmap of top-ranking polar lipid species. Data are normalized by the mean value of each lipid group. (E) Volcano plots 5 showing log2 (fold change) and -log10 (p values) of molecular lipid species comparing CARIA266+LA vs CARIA266. (F) CARIB650+LA vs CARIA266+LA. (G) CARIC568 vs CARIB650+LA. (H) CARIC568 vs CARIA266.

[0047] Fig.11. Three-dimensional surface plot of PuA titer to carbon source level and initial OD. 10

[0048] Fig.12. Three-dimensional surface plot of PuA titer to carbon source level and initial pH.

[0049] Fig.13. Two-dimensional plot of principal component analysis. PCA was performed including all the lipid data in 16 yeast cultures under investigation. 15 DETAILED DESCRIPTION

[0050] In some aspects, the present disclosure provides engineering the oleaginous yeast Rhodosporidium toruloides to produce pomegranate-derived punicic acid. In some aspects, the present disclosure provides optimization of pomegranate-derived punicic acid accumulation in Saccharomyces cerevisiae by Ty retrotransposon-targeted random gene 20 shuffling

[0051] Method of the invention are conveniently practiced by providing the compounds and / or compositions used in such method in the form of a kit. Such kit preferably contains the composition. Such a kit preferably contains instructions for the use thereof.

[0052] To gain a better understanding of the invention described herein, the following 25 examples are set forth. It should be understood that these examples are for illustrative purposes only. Therefore, they should not limit the scope of this invention in anyway.

[0053] EXAMPLES

[0054] EXAMPLE 1

[0055] The accumulation of PuA-rich lipids in P. granatum seed oil begins with the 30 desaturation of oleic acid esterified to phosphatidylcholine (PC). Pomegranate delta-12 acyl lipid desaturase (PgFAD2) catalyzes the desaturation of oleic acid at the sn-2 position of PC, resulting in linoleic acid (LA). Pomegranate fatty acid desaturase and conjugase PgFADX then further converts LA to PuA (Wang et al., 2021). The PuA acyl-chain on PC can be released - 8 -into the cytosol in the form of acyl-CoA, and then transferred to nascent glycerol-3-phosphate to form triacylglycerol (TAG) by various acyltransferases, in which the diacylglycerol acyltransferase (DGAT) is often considered as a key enzyme (Fig.1A). DGAT catalyzes the addition of fatty acid acyl groups to the sn-3 position of diacylglycerol, which is the rate-limiting 5 step for the synthesis of TAG, and a major step responsible for the accumulation of unusual fatty acids (Aymé et al., 2014). DAG containing different degrees of unusual fatty acids can also be interchanged with PC through the catalytic reaction of phosphatidylcholine: diacylglycerol cholinephosphotransferase (PDCT) (Fig. 1A), enhancing the accumulation of TAG with higher unusual fatty acid content (Demski et al., 2022). 10

[0056] Despite the high concentration of PuA in pomegranate seed oil, fruit genotypes, harvesting time, and meteorological circumstances all influence oil quality, making direct extraction of PuA from pomegranate seed oil rather costly (Xu et al., 2020). Although a few studies have successfully produced PuA in plants such as Arabidopsis thaliana and Brassica napus (Mietkiewska et al., 2014; Xu et al., 2020), transgenic plants obtained by these studies 15 still require large arable space and long-term cultivation. In contrast, microbial production provides a sustainable, effective, and scalable approach with significant benefits of improved yields, cost-effectiveness, and shorter production cycles. In our previous study, we modified baker's yeast (Saccharomyces cerevisiae) to produce 3.7% of total fatty acids as PuA using LA precursor feeding (Wang et al., 2021). In another study, the metabolic engineering of 20 Schizosaccharomyces pombe, a yeast with high oleic acid content, led to 25.1% of total fatty acid as PuA (Garaiova et al., 2017). However, since S. cerevisiae and S. pombe have very limited ability to accumulate lipids, the PuA titer was only 7.2 mg / L and 38.7 mg / L, respectively.

[0057] Many yeast species are well-recognized as biologically safe microorganisms with great potential to produce single-cell oil. Oleaginous yeasts, such as Trichosporon 25 fermentans, Yarrowia lipolytica and Rhodosporidium toruloides, can accumulate more than 20% of lipids in their biomass. For instance, T. fermentans was able to convert rice straw hydrolysate or waste molasse to 40% - 60% of single cell oil (Huang et al., 2009; Zhu et al., 2008). Y. lipolytica, a well-studied model oleaginous yeast, has the ability to grow on various carbon sources including fatty acids, glucose, fructose, or glycerol, and accumulating 36% of 30 its biomass as lipids (Zhang et al., 2014). R. toruloides, an unconventional yeast capable of concomitant synthesis of lipids and carotenoids, is typically regarded as a promising candidate for biofuel production due to its excellent lipid accumulation capability, which often exceeds 60% of dry cell weight (Ratledge and Wynn, 2002). Furthermore, R. toruloides' capacity to use - 9 -a wide range of substrates, including glucose and xylose, makes it a promising option for sustainable bioproduction processes (Boviatsi et al., 2020; Qi et al., 2020).

[0058] In this study, we evaluated the engineering the nonconventional yeast R. toruloides to accumulate PuA, a plant-derived value-added unusual fatty acid. R. toruloides 5 transformed with PgFADX was able to accumulate PuA to 3.66% of total fatty acids. By further integrating codon-optimized PgFAD2 and PgDGAT2 into R. toruloides’ genome, a significant improvement in PuA content was observed. Engineered R. toruloides strain RX2 containing both PgFADX and PgFAD2 accumulated 11.97% of its lipid as PuA using glucose as substrate, with a PuA titer of 451.6 mg / L. When wood hydrolysate was used as feedstock, 10 the strain accumulated 6.39% of its lipid as PuA, demonstrating a good potential in converting low-value agricultural waste into value-added PuA.

[0059] 2. Materials and methods

[0060] 2.1 Strains, plasmids and culture conditions

[0061] Strains and plasmids used in this study are listed in Table 1 and Table 2.15 Escherichia coli DH5 was used for routine plasmid construction and preparation. The wildtype of R. toruloides ATCC 204091 (formerly known as Rhodotorula glutinis) was obtained from ATCC (US). To construct mutants with various gene deletions, the 1 kb upstream and downstream sequences of the CAR2 gene were amplified from R. toruloides genomic DNA and linked with expression cassettes containing the zeocin-resistance gene and the target 20 gene as the flanking sequences. EcoRV restriction sites were introduced to both ends of the flanking sequences in the primer design. The zeocin-resistance gene was under the control of GPD1 promoter and 35S terminator. The target genes, including codon-optimized PgFADX, PgFAD2, PgDGAT2, and PgPDCT, were placed under the control of the TEF1 promoter and NOS terminator. To release donor DNAs from the plasmid backbone, either PCR or double 25 digestion of EcoRV sites flanking the donor DNA was conducted. Culture conditions consisting of Luria-Bertani (LB) medium at 37 °C and constant shaking at 200 rpm were used to cultivate E. coli. Kanamycin (50 g / L) was added to maintain a stable inheritance of plasmids for E. coli. R. toruloides was routinely maintained with yeast extract peptone dextrose (YPD) medium containing 10 g / L yeast extract, 20 g / L peptone, and 20 g / L glucose under 30 °C. To induce 30 lipid production, seed culture of R. toruloides was first grown in a 50 mL falcon tube containing 5mL YPD medium and then inoculated into 200 mL of nutrient-limited media (González-García et al., 2017) containing 100 g / L glucose; 0.1 g / L NaNO3, 4.5 g / L KH2PO4, 0.2 g / L MgSO47H2O, - 10 -and 0.11 g / L CaCl22H2O in a 1 L shake flask with an initial OD600of 0.8. The cells were grown under 30 °C with constant shaking at 250 rpm.

[0062] Table 1 Strains used in this study Strain Relevant genotype / property Source F– endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG E. coli DH5apurB20 80dlacZ M15 (lacZYA-argF) U169, hsdR17(rK–Invitrogen mK+), – Wild type R. R. toruloides ATCC 204091 ATCC toruloides R. toruloides ATCC 204091 transformed with codon RX1 This study optimized PgFADX R. toruloides ATCC 204091 transformed with codon RX2 This study optimized PgFAD2 and PgFADX R. toruloides ATCC 204091 transformed with codon RX3 This study optimized PgDGAT2 and PgFADX R. toruloides ATCC 204091 transformed with codon RX4 This study optimized PgPDCT and PgFADX 5

[0063] Table 2 Plasmids and qPCR primers used in this study Name Description Source Plasmids KanR / RtCAR2-left arm / PGPD1-codon-optimized R This pRT-FADX BLE - T35s / PTEF1- PgFADX-TNOS / RtCAR2-right study arm - 11 -KanR / RtCAR2-left arm / PGPD1-codon-optimized R This pRT-FAD2-FADX BLE - T35s / PTEF1- PgFAD2-TNOS / PTEF1- study PgFADX-TNOS / RtCAR2-right arm KanR / RtCAR2-left arm / P -codon-o timized y yqPCR primers GTACGACCAGATCAAGCAGAC (SEQ ID NO: This qRt-GPD1-F 1) study ACAAAGTCGGTTGAGACGAG (SEQ ID NO: This qRt-GPD1-R 2) study This qRT-FADX-F1 ACGGCTTTGCATCACACTAC (SEQ ID NO: 3) study GAGGCGGTAGAGGATGTAGTAG (SEQ ID This qRT-FADX-R1 NO: 4) study

[0064] 2.2 Gene expression analysis

[0065] Total RNA was extracted from recombinant R. toruloides strains R1, R2, R3 and R4 cultured in the nitrogen-limited medium using the Spectrum Total RNA Kit (Sigma- 5 Aldrich). First-strand cDNA synthesis was performed using these RNA as templates and the SuperScript IV first-strand cDNA synthesis kit (Invitrogen) according to the manufacturer’s instructions. Quantification of PgFADX transcripts in various strains was performed using a 1 / 10 dilution of cDNA as the template. R. toruloides GPD1 was used as the reference. Three biological replicates for each strain were analyzed using quantitative RT-PCR (qPCR) on a 10 StepOnePlus Real-Time PCR System (Applied Biosystems, USA) using the GB-Amp™ Sybr Green qPCR Mix (GeneBio) according to the manufacturer’s instructions. Results were analyzed using the comparative Ct method (2- Ctmethod).

[0066] 2.3 Chemical transformation of R. toruloides - 12 -

[0067] Yeast transformations were performed using the lithium acetate and PEG method as described previously (Nora et al., 2019; Tsai et al., 2017). Briefly, the wild-type R. toruloides strain was first cultured in YPD medium at 30°C and 200 rpm overnight. The next day, the cell was diluted and grown to OD600of 0.8. Biomass was recovered, washed and then 5 resuspended in 1 mL 100 mM lithium acetate. After centrifugation, 10 L single-stranded salmon sperm DNA (10 mg / mL), 240 L PEG4000, 36 L lithium acetate (1.0 M), and 5 gDNA of interest was added to the cells, which were then incubated at 30 °C for 40min. Before heat shocked at 42 °C for 15 min, 34 L dimethylsulfoxide (DMSO) was added to the mixture. Next, transformed yeast cells were allowed to recover overnight at 30 °C in the YPD medium 10 with shaking. The cells were plated on YPD agar plates containing 150 g / mL of Zeocin and incubated at 30 °C.

[0068] 2.4 Preparation of wood hydrolysate

[0069] Enzymatic hydrolysis of a wood pulp (Northern Bleached Hardwood Kraft (NBHK) pulp was conducted at a solid concentration of 10 % (w / v) in 50 mM sodium acetate 15 buffer (pH 4.8) at 50 with 200 rpm agitation using cellulase enzyme (20 FPU / g) for 24 h. The liquid enzyme hydrolysate was analyzed for glucose and xylose sugar yields through High- Performance Liquid Chromatography (HPLC, Agilent 1200, Santa Clara, CA, USA) coupled with a Refractive Index Detector (RID, Agilent 1100, Santa Clara, CA, USA). Sugars in the sample (20 μL injection volume) were separated on HPX-87h column (Bio-Rad Aminex, 20 Hercules, CA, USA) with 50 mM of sulfuric acid as the mobile phase at a flow rate of 0.5 mL / min at 60 for 40 min.

[0070] 2.5 Separation of lipid class using thin-layer chromatography (TLC)

[0071] For the detailed analysis of lipid classes within each sample, extracted lipids were separated using one-dimensional thin-layer chromatography (TLC). Silica gel-coated 25 plates (0.25 mm Silica gel, DCFertigplatten, Macherey-Nagel, Germany) were used to resolve lipids into different fractions in a solvent mixture consisting of hexane, diethyl ether, and acetic acid in a 70:30:1 volume ratio. Once the TLC plates were developed, lipid fractions were visualized using a 0.05% primulin dissolved in acetone: water (8:2 v / v), which binds to lipids and fluoresces under UV light, allowing for the non-destructive identification of lipid bands. 30 Following the identification of lipid fractions, the bands corresponding to the target lipid classes were carefully scraped from the plates. These collected fractions were then subjected to a lipid extraction and base-catalyzed lipid derivatization. - 13 -

[0072] 2.6 Positional analysis of TAG using enzymatic hydrolysis

[0073] The detailed examination of the distribution of fatty acids within the sn-2 and sn-1 / 3 positions of TAG was conducted as previously described using enzymatic hydrolysis (Luddy et al., 1964; Xu et al., 2020). Briefly, TAG fractions recovered from TLC separation 5 were moved to clean tubes and dried under a stream of nitrogen. For enzymatic hydrolysis, lipids were treated with a mixture comprising 1 mL of Tris-HCl buffer (1 mM, pH 8.0), 100 L of 2.2% calcium chloride (CaCl2) to provide necessary ions for enzyme activity, and 250 L of 0.1% deoxycholate to help emulsify the lipids. This mixture was then vortexed and sonicated briefly. After prewarming the emulsified lipids for 30 seconds at 40°C, 20 mg of pancreatic 10 lipase (pancreatic lipase type II, Sigma) was introduced to the mixture. This enzyme specifically targets the acyl chain on the sn-1 / 3 positions in TAGs. The reaction was continued for 3 minutes at 40°C and quenched with 500 L of 6 M hydrochloric acid (HCl). Lipid products were then extracted twice with diethyl ether, followed by a secondary TLC separation. The band corresponding to the sn-2 MAG fraction was recovered and prepared for lipid composition 15 analysis.

[0074] 2.7 Lipid extraction, transmethylation and analysis by gas chromatography

[0075] Before lipid extraction, yeast biomass was harvested from liquid culture via centrifugation. The supernatant was then removed and 800 μL of a precooled lipid extraction 20 mixture containing chloroform and isopropanol (2:1, v / v) was added, along with glass beads (0.5mm). Butylated hydroxytoluene (BHT) at a final concentration of 0.01% was also included as an antioxidant to protect the PuA product. Subsequently, the disruption of R. toruloides biomass was achieved through three cycles of bead beating (1-minute duration each) using a Biospec bead beater (Bartlesville, OK), with a 2-minute cooling on ice between each cycle. 25 The extraction was repeated twice for each sample. The collective organic phase was dried under nitrogen, resuspended in HPLC-grade chloroform and stored under -20 °C. For preparing fatty acid methyl ester, transmethylation was done via a base-catalyzed method using 1 mL of 5% sodium methoxide dissolved in methanol (Mietkiewska et al., 2014). After incubation in dark at 30 °C for 1 hour, the reaction was stopped by adding 1.5 mL of 0.9% (w / v)30 sodium chloride solution. Fatty acid methyl esters were then extracted with 1 mL of HPLC- grade hexane. The FAMEs were resuspended in hexane and analyzed on an Agilent 6890N Gas Chromatograph equipped with a 5975 inert XL Mass Selective Detector (Agilent Technologies) and flame ionization detector using the method described in our previous study - 14 -(Mietkiewska et al., 2014). Briefly, the FAMEs were separated on a capillary column DB23 (30 222 m×0.25 mm×0.25 μm, Agilent Technologies, Wilmington, DE, USA) using the following program: 5:1 split ratio, 1 μL injection.4 min at 165 °C, then increased to 180 °C (10 °C / min) and held for 5 min, and increased to 230 °C and held for 5 min. 5

[0076] 3. Results and discussion

[0077] 3.1 Reconstitution of punicic acid synthesis in R. toruloides via the heterologous expression of PgFADX

[0078] When the oleaginous microorganism is exposed to nutritional stress, it accumulates significant amounts of lipids. When nitrogen is depleted, microorganisms shift 10 carbon flow to lipid storage for energy reserves (Feng et al., 2011). According to previous studies, nitrogen deficit in the culture medium reduces the activity of isocitrate dehydrogenase, resulting in excessive citric acid buildup. The excessive citric acid is subsequently transferred and transformed into cytosolic acetyl-CoA, which serves as a precursor for fatty acid production (Ratledge, 2004). Thus, controlling the carbon / nitrogen ratio in the culture medium significantly 15 impacts lipid metabolism in oleaginous microorganisms. Consistency, in the optimized culture mediums with nitrogen limitation, R. toruloides ATCC 204091 was capable of accumulating lipids up to 72% of its dry cell weight (González-García et al., 2017). Therefore, to maximize the PuA production and PuA content in this study, we cultured the wild-type and recombinant R. toruloides under the nitrogen-limited medium designed to maximize lipid accumulation 20 (González-García et al., 2017). As shown in Fig. 2A, after 96 hours of cultivation under nitrogen-limited conditions, the wild-type yeast cell accumulated C18:1 and C18:2 to 47% and 10% of total fatty acids, respectively. In contrast, there’s only a limited amount of C18:3 fatty acids (<2.7%) and no conjugated fatty acid was detected in the wild-type strain. The combined C18 fatty acid occupied over 75% of total fatty acids, which is conducive to PuA accumulation 25 since C18:1 and C18:2 are the important fatty acid precursors to PuA synthesis.

[0079] For the reconstitution of PuA synthetic pathway in R. toruloides, a donor DNA cassette was first assembled (Fig.1B). Since R. toruloides genes have very high (~60%) GC contents and (Bonturi et al., 2022; Hu and Ji, 2016), gene encoding pomegranate-derived PgFADX was first codon optimized to accommodate R. toruloides’ specific codon usage bias 30 (Table S2). The codon-optimized PgFADX was placed under the control of the TEF1 (translation elongation factor 1) promoter, which is a strong constitutive promoter showing robust performance under various growth conditions (Nora et al., 2019). A zeocin resistance gene was linked to PgFADX expression cassette and the whole construct was flanked by 1000 - 15 -bp homologous arms, targeting R. toruloides CAR2 gene. CAR2 encodes lycopene cyclase, which is responsible for carotenoid pigment biosynthesis (Qi et al., 2020). A previous study showed that targeting the CAR2 region could increase gene editing efficiency compared to other commonly used loci, such as URA3 (Otoupal et al., 2019). Replacing the CAR2 gene by 5 homologous recombination leads to white colonies due to the disruption of carotenoid biosynthesis, which serves as a visual marker besides antibiotic selection for identifying transformants with successful integration. As shown in Fig. 1C, after transformation of PgFADX, strains with white colonies were obtained on the zeocin agar plate. Subsequent analysis detected the PgFADX cassette on the genomic DNA recovered from positive 10 transformants (Fig. 1D). In order to examine the PuA-accumulating ability of engineered strains, seven single white colonies were cultured in nitrogen-limited medium. As shown in Fig.2B, the transformation of PgFADX led to 1.6%-2.6% of total fatty acids as PuA, with an average content of 2.28%.

[0080] Table 3 Codon optimized sequences used in this study Name SequenceTCTCGATCTACGGCGTCCCGGTCCTCATCCTCAACGCCTTCGTCGTCCAAGGCCATGTGGCGCGAGGCCCGCGAGTGTCTCTACGTCGAGCCGCGCTCGCTCCACCGAGTCCTATCGTCGTCGTCGGAGCTCAACACCAT ATCGAGCCGTCACTGACTTCGTCCTCGTCCTCCCTCCTCTACTCTCTTG- 20 -BLE- ATGGCCAAGCTCACCTCGGCCGTCCCGGTCCTCACCGCCCGAGACG[ ] a pe ous su y e eop g u -po uc g a sge c a opss, a uant line lacking fatty acid desaturase 3 was used to limit the native linolenic acid synthesis, allowing more LA precursor to be diverted to PuA production (Mietkiewska et al., 2014). In oleaginous - 21 -yeast, such as R. toruloides and Lipomyces starkeyi, bifunctional 12 / 15 fatty aciddesaturases were suggested to carry out further desaturation converting LA into linolenic acid (Y. Liu et al., 2021; Matsuzawa et al., 2018). Notably, when PgFADX was expressed, the native linolenic acid content in R. toruloides decreased significantly to less than 0.3% (Fig.2B). The 5 spontaneous reduction of linolenic acid levels implies that PgFADX may effectively compete with R. toruloides' endogenous desaturase for LA substrate. As a result, removing the activities of native bifunctional enzymes may not be necessary. To further characterize the influence of PgFADX overexpression, the transformant with the highest PuA content, designated as RX1, was tested in a 10-day shake flask cultivation. As shown in Fig.2C, the level of LA and PuA in 10 strain RX1 gradually increased in the 10-day period. At the end of cultivation, PuA accounted for 3.66% of total fatty acids in RX1. A concomitant increase in LA and PuA was observed, which possibly owing to PgFADX's bifunctional activity that can produce LA as well as PuA (Garaiova et al., 2017). The increase of polyunsaturated fatty acid (PUFA) was at the expense of oleic acid level, which reduced from 58% to 50% between day 2 and day 10. The PuA titer 15 at the end of the cultivation reached 88.8 mg / L, which is higher than the level previously achieved in S. cerevisiae and Y. lipolytica with single PgFADX expression (Urbanikova et al., 2023; Wang et al., 2021), demonstrating the potential of R. toruloides in plant-derived conjugated fatty acid production.

[0082] 3.2 Coexpression of PgFAD2 or PgDGAT2 with PgFADX significantly 20 improved PuA content in recombinant R. toruloides

[0083] Considering the importance of LA precursor to PuA synthesis, we further constructed the PgFAD2 and PgFADX coexpression cassettes, aiming to provide more LA precursor to PgFADX on top of R. toruloides’ native LA synthesis. In addition to the upstream precursor supply, the enrichment of unusual fatty acids in plants also requires a series of 25 downstream enzymes, in which DGAT2 and PDCT play important roles. Apart from DGAT2's vital function in the conventional TAG synthesis pathway, previous studies showed that many plant-derived DGAT2s have distinct acyl-CoA and DAG substrate preferences, which considerably contribute to the accumulation of unusual fatty acids (Burgal et al., 2008; Shockey et al., 2006). A recent study also showed that the DAG pool used by DGAT2 is strongly related 30 to the activity of PDCT in unusual fatty acid-producing plants. The PDCT in castor bean, a perennial flowering plant that accumulates high levels of ricinoleic acid in seed oil, preferably converts DAG with one ricinoleoyl group into PC. By doing so, the ricinoleoyl group was reused in the synthesis of DAG or TAG with all positions occupied by ricinoleic acid (Demski et al., - 22 -2022). In light of this, strains coexpressing PgFADX with either PgDGAT2 or PgPDCT were also constructed. Only a few transformants were recovered from the zeocin selection plate after transforming wild-type R. toruloides with PgFADX-PgFAD2, PgFADX-PgDGAT2, and PgFADX-PgPDCT coexpression cassettes, and all transformants displayed an orange 5 pigmentation (Fig.1C). The incomplete CAR2-targeted insertion and disruption may be attributed to the significantly longer length of the co-expression cassettes, in comparison to the single expression of PgFADX (Fig.1B), or the preference for nonhomologous end joining (NHEJ) over homologous recombination (HR) for DNA repair in the R. toruloides strain used in this study (Schultz et al., 2019). Indeed, earlier studies indicated that high CAR2 deletion 10 efficiency up to 75.3% was only obtained in the KU70-deficient R. toruloides strain, which has a mutation of Ku70 / 80 regulatory DNA-binding subunits that is critical to the NHEJ system (Koh et al., 2014). In the wild-type R. toruloides strain, the targeted deletion frequency of CAR2 was only 10.5%. The donor DNA cassettes generated for coexpression appear to be randomly integrated into the R. toruloides genome by NHEJ, in line with what was found by prior studies 15 using a similar transformation method (Tsai et al., 2017). As a result, only zeocin resistance was conferred, without a complete change in colony color. To take into account the position effect, in which surrounding genetic elements may enhance or inhibit gene expression, different transformants were examined for their varying abilities to accumulate PuA.

[0084] As shown in Fig. 3A, the resulting transformants of PgFADX-PgFAD2 20 coexpression led to 5.8%-9.6% of total fatty acids as PuA, with an average of 7.4%, representing a 2.3-fold increase compared to PgFADX single expression. In addition to an increase in PuA levels, the average LA content increased by 30% at the expense of oleic acid content. Similarly, the coexpression of PgDGAT2 with PgFADX resulted in a 2-fold rise in PuA content, whereas LA content decreased by 14% (Fig.3B). In contrast, the coexpression of 25 PgPDCT with PgFADX only increased the PuA content by merely 17% (Fig.3C). The transformants with highest PuA content for PgFADX-PgFAD2, PgFADX-PgDGAT2, and PgFADX-PgPDCT coexpression, designated as strains RX2, RX3, and RX4, were further investigated in shake flask culture. As shown in Fig. 4A, the relative expression levels of PgFADX in RX2, RX3, and RX4 were slightly lower compared to the level in strain RX1 when 30 cultured under the nitrogen-limited condition. However, by the end of the 10-day cultivation, LA and PuA accounted for 13.8% and 11.98% of total fatty acid in RX2, respectively (Fig.3D). The final lipid content was 66% of dry cell weight, and PuA titer of strain RX2 reached 451.6 mg / L, which was five-fold that of RX1. As shown in Figs.3E and 3F, at 240 hours, the PuA - 23 -levels in strain RX3 and RX4 reached 8% and 4.03% of total fatty acids, respectively. The PgDGAT2 coexpression in strain RX3 led to a final PuA titer of 285.4 mg / L, also demonstrating a substantial improvement above RX1.

[0085] In our previous study, when coexpressing PgFAD2 with PgFADX in S. 5 cerevisiae BY4741, the LA content was increased to over 6%, whereas the PuA content was only 0.3% of total fatty acids (Wang et al., 2021). The much greater levels of LA relative to PuA indicated that LA was not effectively converted to PuA in S. cerevisiae due to unknown reasons. In the metabolic-engineered obese Y. lipolytica with PgFADX overexpression, the level of LA (2.9%) was also much higher than the PuA level (0.5%) (Urbanikova et al., 2023). 10 In comparison, the level of LA and PuA obtained in the recombinant R. toruloides strain was relatively more balanced. When combined with R. toruloides' high oleagincity and enhanced PuA content, the PuA titer achieved in strain RX2 outperformed other microorganisms we have studied previously (Garaiova et al., 2017; Wang et al., 2021).

[0086] 3.3 PuA-producing recombinant R. toruloides accumulates high level of 15 PuA in TAG fraction

[0087] Although the conversion of fatty acid precursors into unusual fatty acids often takes place at the sn-2 position of PC, the majority of the unusual fatty acids in the natural producer are eventually integrated into TAG through highly efficient and specific enzymatic pathways. As a result, PuA made up 60% of the fatty acids in TAG and just 0.8% of the fatty 20 acids in PC in the seed lipid that was isolated from P. granatum (Mietkiewska et al., 2014). In contrast, the distribution of unusual fatty acids between polar lipids and TAGs significantly varies in the transgenic hosts. Thus, in order to examine the PuA distribution in recombinant R. toruloides, total lipids extracted from strain RX1, RX2, RX3 and RX4 cultured in nitrogen- limited conditions were separated by TLC. To serve as a comparison, the lipid extracted from 25 wild-type R. toruloides cultured in the 2*YPD medium was also loaded onto the TLC plate. As shown in Fig.4B, under the nitrogen-limited condition, recombinant R. toruloides produced a higher amount of neutral lipid compared to polar lipid (PL). Notably, 4.9% PuA, 17.2% LA, and 47.2% oleic acid were found in the TAG isolated from strain RX1 (Fig.4C), while PuA made up just 0.8% of the total fatty acids in the PL fraction of RX1 lipid (Fig.4D). When PgFAD2 was 30 co-expressed with PgFADX, strain RX2 accumulated 13.63% PuA in TAG, representing a 1.78-fold increase compared to strain RX1. Meanwhile, the levels of LA and oleic acid were reduced by 9% and 15%, respectively. The PL fraction separated from strain RX2 lipid mainly comprises 1.37% PuA, 32.1% LA, and 46.5% oleic acid. Since both PgFAD2 and PgFADX - 24 -recognize 1-acyl-2-olyeol PC as substrate, the amount of oleic acid in RX2's PL was considerably reduced by 46% compared to RX1. Albeit at a lower level relative to RX2, strains RX3 and RX4 also accumulated 10.7% and 5.5% PuA in TAG, respectively, indicating a 1.2- and 0.13-fold increase compared to RX1. Positional analysis of TAG showed that in all four 5 strains, oleic acid was the predominant fatty acid at the sn-2 position of TAG (Fig.4E). The sn- 2 position of TAG in strain RX2 contained higher content of PuA (2.38%) and LA (17.8%) compared to the other recombinant strains. In contrast, in strain RX2, PuA accounted for 19.3% of the fatty acids at the sn-1 / 3 TAG, suggesting a preference for enriching PuA at the sn-1 / 3 position in this recombinant R. toruloides strain (Fig.4F). A similar result was found in a prior 10 study, in which metabolic engineered eicosapentaenoic acid (EPA)-producing Y. lipolytica preferably concentrate EPA at the sn-1 / 3 positions of TAG (Xue et al., 2013). The rearrangement of EPA across the glycerol backbone suggests significant lipid remodelling between the PL and TAG fractions in Y. lipolytica (Xue et al., 2013), a process that might similarly occur in R. toruloides. 15

[0088] Although some naturally occurring polar lipids are being investigated for their potential in nutraceutical and pharmaceutical applications (Da Costa et al., 2021; Venkat et al., 2024), the presence of phospholipids in crude oils is often considered to negatively affect the oil's appearance and flavour (Liu et al., 2023). Since phospholipids are more susceptible to oxidation than neutral lipids, their oxidation has a substantial influence on the stability, shelf 20 life, and quality of food oils. In food lipid production, such as the manufacturing of soybean and rapeseed oil, degumming is often required to remove phospholipids from crude oils (Li et al., 2023). Acids, enzymes, or water are used to hydrolyze and hydrate the phospholipids, separating them from the crude oil to produce high-quality cooking oil. In the case of PuA- containing single-cell oils, removing polar lipids may result in product loss, especially if a larger25 ratio of PuA is retained in the polar lipid fraction rather than the neutral lipid fraction. In PuA- producing transgenic Arabidopsis, the PuA content of TAG was only 6.6%, whereas PC contains up to 12.5% of total fatty acids as PuA (Mietkiewska et al., 2014). In our previous study, the TAG and polar lipids isolated from the recombinant PuA-accumulating S. cerevisiae cells also showed a similar distribution. The heterologous synthesis of PuA led to 4.8% of total 30 fatty acids as PuA in the PL fraction, 1.28-fold higher compared to its content in TAG (Wang et al., 2021). In this study, under nitrogen-limited conditions, the PuA content in the TAG fraction generated by recombinant R. toruloides was significantly higher compared to the PuA content in its PL fraction. A possible explanation is that under nitrogen-limited conditions, TAG - 25 -synthesis in R. toruloides underwent a significant upregulation, channelling more fatty acid precursors towards TAG assembly instead of being utilized for membrane expansion and PL synthesis (Fig. 4B). A similar observation has been made previously in EPA-producing Y. lipolytica, which has greater EPA content in the TAG fraction than in the phospholipid fraction, 5 indicating that Y. lipolytica may have a mechanism for modulating EPA distribution across different lipid species (Xue et al., 2013). These results demonstrated that oleaginous yeasts, such as R. toruloides and Y. lipolytica, have high potential in producing unusual fatty acid- enriched TAG with improved purity and stability.

[0089] 3.4 Converting wood hydrolysate into PuA-containing single-cell oil 10

[0090] Lignocellulose feedstock, along with many other renewable feedstocks, is frequently regarded as a suitable substrate for microbial cell factories due to its availability and renewable nature. Lignocellulosic biomass is the most abundant raw material derived from agriculture and forestry waste, thus it is affordable and readily available (Gao et al., 2023). Unlike starch, sugar and lipid feedstock, the abundance of lignocellulose feedstock ensures a 15 long-term and secure supply of substrates for industrial bioprocesses without competing with food sources. To transform lignocellulosic substrates into bioproducts, a microorganism must be capable of utilizing both pentose and hexose sugars. However, not all industrially significant strains have this capability. In the yeast S. cerevisiae, the introduction of heterologous xylose reductase / xylitol dehydrogenase pathway or xylose isomerase pathway was needed to enable 20 the use of xylose as the carbon source (Gao et al., 2023). Although oleaginous yeast Y. lipolytica has emerged as a promising host for lipid and bioproduct synthesis, it is also unable to grow with xylose as the sole carbon source (Ledesma-Amaro et al., 2016; Zhao et al., 2015). The co-expression of xylose reductase and xylitol dehydrogenase from Scheffersomyces stipitis coupled with overexpression of the endogenous xylulokinase are necessary to permit 25 normal growth of engineered Y. lipolytica on xylose (Ledesma-Amaro et al., 2016). Conversely, R. toruloides has a broader substrate range and is notable for its inherent ability to metabolize pentose sugars, attracting interest for its potential in converting lignocellulosic biomass into valuable bioproducts (Boviatsi et al., 2020; Qi et al., 2020). Moreover, R. toruloides has shown strong growth in the presence of various inhibitors that are typically found in pretreated 30 agricultural wastes, demonstrating its tolerance to stressful conditions compared to other yeasts (Jiao et al., 2021; Zhang et al., 2022).

[0091] In this regard, we sought to examine the capability of converting lignocellulose feedstock into PuA-containing single-cell oil using the engineered R. toruloides in this study. - 26 -Wood pulp hydrolysate was used as the feedstock, which is generally composed of mixed pentose and hexose sugars derived from the hydrolysis of lignocellulosic biomass. Enzymatic hydrolysis of wood pulp composed of 79.1 ± 1.0% cellulose, 21.2 ± 0.6% hemicellulose and 4.0 ± 0.1% lignin yielded a liquid wood hydrolysate containing 62 mg / mL glucose and 16 5 mg / mL xylose. The differentially diluted wood hydrolysates were used directly to replace the glucose in the nitrogen-limited medium (González-García et al., 2017). Strain RX2 with the highest PuA content was inoculated into wood hydrolysate medium to different initial OD600. As shown in Fig.5A, RX2 successfully converted wood hydrolysate to PuA containing single- cell oil. When cultivated in 80% wood hydrolysate, strain RX2 accumulated 5.45% of total fatty 10 acid as PuA. A slight increase in PuA content and titer was observed when using a higher concentration of wood hydrolysate and a higher initial OD600. Subsequently, a 10-day cultivation in 100% wood hydrolysate was conducted (Fig.5B). By the end of the cultivation, strain RX2 accumulated 6.39% of its fatty acids as PuA, and the PuA titer reached 310 mg / L. Although a lower level of PuA was obtained in the wood hydrolysate medium compared to 15 what was achieved in the glucose medium (Fig.3D), considering the renewable and abundant nature of wood hydrolysate, the result still points out the potential of R. toruloides in producing value-added lipid from low-cost feedstock. While this study did not measure carotenoid levels in strains RX2, RX3, and RX4, since carotenoid biosynthesis was not completely disrupted, the engineered RX2, RX3 and RX4 might be capable of concomitant production of carotenoid 20 and PuA. The carotenoid produced by R. toruloides has high radical scavenging efficiency and plays an important role in inhibiting oxidative damage (Li et al., 2019; Singh et al., 2016). Combined with the health benefits of PuA, engineered PuA-producing R. toruloides may serve as a potential source for nutritional supplements (Fraser and Bramley, 2004).

[0092] Previous studies indicated that R. toruloides' ability to convert lignocellulosic 25 feedstock into bioproducts could be improved through the enhancement of xylose assimilation and the adaptive evolution of the engineered strain. R. toruloides has an unusual xylose metabolism featuring the reduction to D-arabitol, oxidation to D-ribulose, and phosphorylation to ribulose 5-phosphate (Adamczyk et al., 2023). By overexpressing a putative transcription factor (RTO4_12978, Pnt1) that acts as a major regulator of pentose metabolism, the 30 expression of enzymes involved in xylose catabolism was increased and the specific growth rate was improved significantly in cultures on xylose (Coradetti et al., 2023). The adaptation of R. toruloides through evolutionary strategies has also led to the development of strains with increased tolerance to the major inhibitors present in lignocellulosic hydrolysates (Díaz et al., - 27 -2018; Z. Liu et al., 2021). Furthermore, previous studies have demonstrated that S. cerevisiae strains producing PUFA were initially more susceptible to oxidative stress. However, throughout adaptation, yeast strain developed increased resistance to oxidative stress induced by PUFA, attributed to a rise in catalase activity (Cipak et al., 2008). Through additional 5 adaptation of PuA-producing R. toruloides in lignocellulosic medium and metabolic engineering of the xylose assimilation pathway, both the cell growth and the production of the PuA could be further enhanced.

[0093] 4. Conclusion

[0094] In this study, the unconventional oleaginous yeast R. toruloides was engineered 10 for the first time to produce PuA, a unique and valuable unusual fatty acid derived from pomegranate. By integrating codon-optimized PgFADX, strain RX1 successfully synthesized 3.66% PuA. Further coexpression of PgFAD2 or PgDGAT2 led to a notable increase in PuA levels. The engineered strain RX2 achieved 11.97% of its lipid content as PuA, with a production titer of 451.6 mg / L using glucose. Utilizing wood hydrolysate as the feedstock, a 15 PuA content of 6.39% was achieved, highlighting R. toruloides’ capability to transform agricultural waste into valuable lipid products.

[0095] Reference

[0096] Adamczyk, P.A., Coradetti, S.T., Gladden, J.M., 2023. Non-canonical d-xylose and l-arabinose metabolism via d-arabitol in the oleaginous yeast Rhodosporidium toruloides. 20 Microb Cell Fact 22, 145. https: / / doi.org / 10.1186 / s12934-023-02126-x

[0097] Aruna, P., Venkataramanamma, D., Singh, A.K., Singh, R.P., 2016. Health Benefits of Punicic Acid: A Review: Health benefits of punicic acid. . . COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY 15, 16–27. https: / / doi.org / 10.1111 / 1541- 4337.12171 25

[0098] Aymé, L., Baud, S., Dubreucq, B., Joffre, F., Chardot, T., 2014. Function and Localization of the Arabidopsis thaliana Diacylglycerol Acyltransferase DGAT2 Expressed in Yeast. PLoS ONE 9, e92237. https: / / doi.org / 10.1371 / journal.pone.0092237

[0099] Beopoulos, A., Verbeke, J., Bordes, F., Guicherd, M., Bressy, M., Marty, A., Nicaud, J.-M., 2014. Metabolic engineering for ricinoleic acid production in the oleaginous 30 yeast Yarrowia lipolytica. Appl Microbiol Biotechnol 98, 251–262. https: / / doi.org / 10.1007 / s00253-013-5295-x

[0100] Bonturi, N., Pinheiro, M.J., de Oliveira, P.M., Rusadze, E., Eichinger, T., Liudži t , G., De Biaggi, J.S., Brauer, A., Remm, M., Miranda, E.A., Ledesma-Amaro, R.,- 28 -Lahtvee, P.-J., 2022. Development of a dedicated Golden Gate Assembly Platform (RtGGA) for Rhodotorula toruloides. Metabolic Engineering Communications 15, e00200. https: / / doi.org / 10.1016 / j.mec.2022.e00200

[0101] Boviatsi, E., Papadaki, A., Efthymiou, M., Nychas, G.E., Papanikolaou, S., Da 5 Silva, J.A.C., Freire, D.M.G., Koutinas, A., 2020. Valorisation of sugarcane molasses for the production of microbial lipids via fermentation of two Rhodosporidium strains for enzymatic synthesis of polyol esters. J of Chemical Tech & Biotech 95, 402–407. https: / / doi.org / 10.1002 / jctb.5985

[0102] Burgal, J., Shockey, J., Lu, C., Dyer, J., Larson, T., Graham, I., Browse, J., 10 2008. Metabolic engineering of hydroxy fatty acid production in plants: RcDGAT2 drives dramatic increases in ricinoleate levels in seed oil. Plant Biotechnology Journal 6, 819–831. https: / / doi.org / 10.1111 / j.1467-7652.2008.00361.x

[0103] Cipak, A., Jaganjac, M., Tehlivets, O., Kohlwein, S.D., Zarkovic, N., 2008. Adaptation to oxidative stress induced by polyunsaturated fatty acids in yeast. Biochimica et 15 Biophysica Acta (BBA) - Molecular and Cell Biology of Lipids 1781, 283–287. https: / / doi.org / 10.1016 / j.bbalip.2008.03.010

[0104] Coradetti, S.T., Adamczyk, P.A., Liu, D., Gao, Y., Otoupal, P.B., Geiselman, G.M., Webb-Robertson, B.-J.M., Burnet, M.C., Kim, Y.-M., Burnum-Johnson, K.E., Magnuson, J., Gladden, J.M., 2023. Engineering transcriptional regulation of pentose metabolism in 20 Rhodosporidium toruloides for improved conversion of xylose to bioproducts. Microb Cell Fact 22, 144. https: / / doi.org / 10.1186 / s12934-023-02148-5

[0105] Costantini, S., Rusolo, F., De Vito, V., Moccia, S., Picariello, G., Capone, F., Guerriero, E., Castello, G., Volpe, M.G., 2014. Potential anti-inflammatory effects of the hydrophilic fraction of pomegranate (Punica granatum L.) seed oil on breast cancer cell lines. 25 Molecules 19, 8644–8660. https: / / doi.org / 10.3390 / molecules19068644

[0106] Da Costa, E., Melo, T., Reis, M., Domingues, P., Calado, R., Abreu, M.H., Domingues, M.R., 2021. Polar Lipids Composition, Antioxidant and Anti-Inflammatory Activities of the Atlantic Red Seaweed Grateloupia turuturu. Marine Drugs 19, 414. https: / / doi.org / 10.3390 / md19080414 30

[0107] Demski, K., Jeppson, S., Stymne, S., Lager, I., 2022. Phosphatidylcholine:diacylglycerol cholinephosphotransferase’s unique regulation of castor bean oil quality. Plant Physiol 189, 2001–2014. https: / / doi.org / 10.1093 / plphys / kiac209 - 29 -

[0108] Díaz, T., Fillet, S., Campoy, S., Vázquez, R., Viña, J., Murillo, J., Adrio, J.L., 2018. Combining evolutionary and metabolic engineering in Rhodosporidium toruloides for lipid production with non-detoxified wheat straw hydrolysates. Appl Microbiol Biotechnol 102, 3287–3300. https: / / doi.org / 10.1007 / s00253-018-8810-2 5

[0109] Feng, D., Chen, Z., Xue, S., Zhang, W., 2011. Increased lipid production of the marine oleaginous microalgae Isochrysis zhangjiangensis (Chrysophyta) by nitrogen supplement. Bioresource Technology 102, 6710–6716. https: / / doi.org / 10.1016 / j.biortech.2011.04.006

[0110] Fraser, P.D., Bramley, P.M., 2004. The biosynthesis and nutritional uses of 10 carotenoids. Prog Lipid Res 43, 228–265. https: / / doi.org / 10.1016 / j.plipres.2003.10.002

[0111] Gao, J., Yu, W., Li, Y., Jin, M., Yao, L., Zhou, Y.J., 2023. Engineering co- utilization of glucose and xylose for chemical overproduction from lignocellulose. Nat Chem Biol 19, 1524–1531. https: / / doi.org / 10.1038 / s41589-023-01402-6

[0112] Garaiova, M., Mietkiewska, E., Weselake, R.J., Holic, R., 2017. Metabolic 15 engineering of Schizosaccharomyces pombe to produce punicic acid, a conjugated fatty acid with nutraceutic properties. Appl Microbiol Biotechnol 101, 7913–7922. https: / / doi.org / 10.1007 / s00253-017-8498-8

[0113] González-García, Y., Rábago-Panduro, L.M., French, T., Camacho-Córdova, D.I., Gutiérrez-González, P., Córdova, J., 2017. High lipids accumulation in Rhodosporidium 20 toruloides by applying single and multiple nutrients limitation in a simple chemically defined medium. Ann Microbiol 67, 519–527. https: / / doi.org / 10.1007 / s13213-017-1282-2

[0114] Hu, J., Ji, L., 2016. Draft Genome Sequences of Rhodosporidium toruloides Strains ATCC 10788 and ATCC 10657 with Compatible Mating Types. Genome Announc 4, e00098-16. https: / / doi.org / 10.1128 / genomeA.00098-16 25

[0115] Huang, C., Zong, M., Wu, H., Liu, Q., 2009. Microbial oil production from rice straw hydrolysate by Trichosporon fermentans. Bioresource Technology 100, 4535–4538. https: / / doi.org / 10.1016 / j.biortech.2009.04.022

[0116] Jiao, X., Lyu, L., Zhang, Y., Huang, Q., Zhou, R., Wang, Shian, Wang, Shuang, Zhang, S., Zhao, Z.K., 2021. Reduction of lipid-accumulation of oleaginous yeast 30 Rhodosporidium toruloides through CRISPR / Cas9-mediated inactivation of lipid droplet structural proteins. FEMS Microbiology Letters 368, fnab111. https: / / doi.org / 10.1093 / femsle / fnab111 - 30 -

[0117] Koh, C.M.J., Liu, Y., Moehninsi, Du, M., Ji, L., 2014. Molecular characterization of KU70 and KU80 homologues and exploitation of a KU70-deficient mutant for improving gene deletion frequency in Rhodosporidium toruloides. BMC Microbiol 14, 50. https: / / doi.org / 10.1186 / 1471-2180-14-50 5

[0118] Ledesma-Amaro, R., Lazar, Z., Rakicka, M., Guo, Z., Fouchard, F., Coq, A.- M.C.-L., Nicaud, J.-M., 2016. Metabolic engineering of Yarrowia lipolytica to produce chemicals and fuels from xylose. Metabolic Engineering 38, 115–124. https: / / doi.org / 10.1016 / j.ymben.2016.07.001

[0119] Li, X., He, Y., Zhang, L., Xu, Z., Ben, H., Gaffrey, M.J., Yang, Y., Yang, S., 10 Yuan, J.S., Qian, W.-J., Yang, B., 2019. Discovery of potential pathways for biological conversion of poplar wood into lipids by co-fermentation of Rhodococci strains. Biotechnol Biofuels 12, 60. https: / / doi.org / 10.1186 / s13068-019-1395-x

[0120] Li, Z., Wang, W., Liu, X., Qi, S., Lan, D., Wang, Y., 2023. Effect of different degumming processes on the retention of bioactive components, acylglycerol and 15 phospholipid composition of rapeseed oil. Process Biochemistry 133, 190–199. https: / / doi.org / 10.1016 / j.procbio.2023.08.019

[0121] Liu, X., Wang, W., Zhao, Z., Xu, L., Yang, B., Lan, D., Wang, Y., 2023. Monoacylglycerol lipase from marine Geobacillus sp. showing lysophospholipase activity and its application in efficient soybean oil degumming. Food Chemistry 406, 134506. 20 https: / / doi.org / 10.1016 / j.foodchem.2022.134506

[0122] Liu, Y., Koh, C.M.J., Yap, S.A., Cai, L., Ji, L., 2021. Understanding and exploiting the fatty acid desaturation system in Rhodotorula toruloides. Biotechnol Biofuels 14, 73. https: / / doi.org / 10.1186 / s13068-021-01924-y

[0123] Liu, Z., Radi, M., Mohamed, E.T.T., Feist, A.M., Dragone, G., Mussatto, S.I., 25 2021. Adaptive laboratory evolution of Rhodosporidium toruloides to inhibitors derived from lignocellulosic biomass and genetic variations behind evolution. Bioresource Technology 333, 125171. https: / / doi.org / 10.1016 / j.biortech.2021.125171

[0124] Luddy, F.E., Barford, R.A., Herb, S.F., Magidman, P., Riemenschneider, R.W., 1964. Pancreatic lipase hydrolysis of triglycerides by a semimicro technique. J Am Oil Chem 30 Soc 41, 693–696. https: / / doi.org / 10.1007 / BF02661412

[0125] Machado, M., Costa, E.M., Silva, S., Rodriguez-Alcalá, L.M., Gomes, A.M., Pintado, M., 2022. Pomegranate Oil’s Potential as an Anti-Obesity Ingredient. Molecules 27, 4958. https: / / doi.org / 10.3390 / molecules27154958 - 31 -

[0126] Matsuzawa, T., Maehara, T., Kamisaka, Y., Ara, S., Takaku, H., Yaoi, K., 2018. Identification and characterization of 12 and 12 / 15 bifunctional fatty acid desaturases inthe oleaginous yeast Lipomyces starkeyi. Appl Microbiol Biotechnol 102, 8817–8826. https: / / doi.org / 10.1007 / s00253-018-9345-2 5

[0127] Mele, M.C., Cannelli, G., Carta, G., Cordeddu, L., Melis, M.P., Murru, E., Stanton, C., Banni, S., 2013. Metabolism of c9,t11-conjugated linoleic acid (CLA) in humans. Prostaglandins, Leukotrienes and Essential Fatty Acids 89, 115–119. https: / / doi.org / 10.1016 / j.plefa.2013.05.005

[0128] Mietkiewska, E., Miles, R., Wickramarathna, A., Sahibollah, A.F., Greer, M.S., 10 Chen, G., Weselake, R.J., 2014. Combined transgenic expression of Punica granatum conjugase (FADX) and FAD2 desaturase in high linoleic acid Arabidopsis thaliana mutant leads to increased accumulation of punicic acid. Planta 240, 575–583. https: / / doi.org / 10.1007 / s00425-014-2109-z

[0129] Nora, L.C., Wehrs, M., Kim, J., Cheng, J.-F., Tarver, A., Simmons, B.A., 15 Magnuson, J., Harmon-Smith, M., Silva-Rocha, R., Gladden, J.M., Mukhopadhyay, A., Skerker, J.M., Kirby, J., 2019. A toolset of constitutive promoters for metabolic engineering of Rhodosporidium toruloides. Microb Cell Fact 18, 117. https: / / doi.org / 10.1186 / s12934-019- 1167-0

[0130] Otoupal, P.B., Ito, M., Arkin, A.P., Magnuson, J.K., Gladden, J.M., Skerker, 20 J.M., 2019. Multiplexed CRISPR-Cas9-Based Genome Editing of Rhodosporidium toruloides 4, 13.

[0131] Qi, F., Shen, P., Hu, R., Xue, T., Jiang, X., Qin, L., Chen, Y., Huang, J., 2020. Carotenoids and lipid production from Rhodosporidium toruloides cultured in tea waste hydrolysate. Biotechnol Biofuels 13, 74. https: / / doi.org / 10.1186 / s13068-020-01712-0 25

[0132] Ratledge, C., 2004. Fatty acid biosynthesis in microorganisms being used for Single Cell Oil production. Biochimie 86, 807–815. https: / / doi.org / 10.1016 / j.biochi.2004.09.017

[0133] Ratledge, C., Wynn, J.P., 2002. The Biochemistry and Molecular Biology of Lipid Accumulation in Oleaginous Microorganisms, in: Advances in Applied Microbiology. Elsevier, pp.1–52. https: / / doi.org / 10.1016 / S0065-2164(02)51000-5 30

[0134] Schultz, J.C., Cao, M., Zhao, H., 2019. Development of a CRISPR / Cas9 system for high efficiency multiplexed gene deletion in Rhodosporidium toruloides. Biotechnology and Bioengineering 116, 2103–2109. https: / / doi.org / 10.1002 / bit.27001 - 32 -

[0135] Shabbir, M.A., Khan, M.R., Saeed, M., Pasha, I., Khalil, A.A., Siraj, N., 2017. Punicic acid: A striking health substance to combat metabolic syndromes in humans. Lipids Health Dis 16, 99. https: / / doi.org / 10.1186 / s12944-017-0489-3

[0136] Shockey, J.M., Gidda, S.K., Chapital, D.C., Kuan, J.-C., Dhanoa, P.K., Bland, 5 J.M., Rothstein, S.J., Mullen, R.T., Dyer, J.M., 2006. Tung Tree DGAT1 and DGAT2 Have Nonredundant Functions in Triacylglycerol Biosynthesis and Are Localized to Different Subdomains of the Endoplasmic Reticulum. The Plant Cell 18, 2294–2313. https: / / doi.org / 10.1105 / tpc.106.043695

[0137] Singh, G., Jawed, A., Paul, D., Bandyopadhyay, K.K., Kumari, A., Haque, S., 10 2016. Concomitant Production of Lipids and Carotenoids in Rhodosporidium toruloides under Osmotic Stress Using Response Surface Methodology. Front. Microbiol. 7. https: / / doi.org / 10.3389 / fmicb.2016.01686

[0138] Tsai, Y.-Y., Ohashi, T., Kanazawa, T., Polburee, P., Misaki, R., Limtong, S., Fujiyama, K., 2017. Development of a sufficient and effective procedure for transformation of 15 an oleaginous yeast, Rhodosporidium toruloides DMKU3-TK16. Curr Genet 63, 359–371. https: / / doi.org / 10.1007 / s00294-016-0629-8

[0139] Tsuzuki, T., Kawakami, Y., Abe, R., Nakagawa, K., Koba, K., Imamura, J., Iwata, T., Ikeda, I., Miyazawa, T., 2006. Conjugated linolenic acid is slowly absorbed in rat intestine, but quickly converted to conjugated linoleic acid. J Nutr 136, 2153–2159. 20 https: / / doi.org / 10.1093 / jn / 136.8.2153

[0140] Urbanikova, V., Park, Y.-K., Krajciova, D., Tachekort, M., Certik, M., Grigoras, I., Holic, R., Nicaud, J.-M., Gajdos, P., 2023. Yarrowia lipolytica as a Platform for Punicic Acid Production. International Journal of Molecular Sciences 24, 8823. https: / / doi.org / 10.3390 / ijms24108823 25

[0141] Venkat, M., Chia, L.W., Lambers, T.T., 2024. Milk polar lipids composition and functionality: a systematic review. Critical Reviews in Food Science and Nutrition 64, 31–75. https: / / doi.org / 10.1080 / 10408398.2022.2104211

[0142] Vroegrijk, I.O.C.M., van Diepen, J.A., van den Berg, S., Westbroek, I., Keizer, H., Gambelli, L., Hontecillas, R., Bassaganya-Riera, J., Zondag, G.C.M., Romijn, J.A., 30 Havekes, L.M., Voshol, P.J., 2011. Pomegranate seed oil, a rich source of punicic acid, prevents diet-induced obesity and insulin resistance in mice. Food and Chemical Toxicology 49, 1426–1430. https: / / doi.org / 10.1016 / j.fct.2011.03.037 - 33 -

[0143] Wang, J., Xu, Y., Holic, R., Yu, X., Singer, S.D., Chen, G., 2021. Improving the Production of Punicic Acid in Baker’s Yeast by Engineering Genes in Acyl Channeling Processes and Adjusting Precursor Supply. J. Agric. Food Chem. 69, 9616–9624. https: / / doi.org / 10.1021 / acs.jafc.1c03256 5

[0144] Xu, Y., Mietkiewska, E., Shah, S., Weselake, R.J., Chen, G., 2020. Punicic acid production in Brassica napus. Metabolic Engineering 62, 20–29. https: / / doi.org / 10.1016 / j.ymben.2020.08.011

[0145] 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, 10 D.J., Hollerbach, D.H., Zhang, H., Arcilla, D.M., Bledsoe, S.A., Croker, K., McCord, E.F., Tyreus, B.D., Jackson, E.N., Zhu, Q., 2013. Production of omega-3 eicosapentaenoic acid by metabolic engineering of Yarrowia lipolytica. Nat Biotechnol 31, 734–740. https: / / doi.org / 10.1038 / nbt.2622

[0146] Zhang, H., Zhang, L., Chen, H., Chen, Y.Q., Chen, W., Song, Y., Ratledge, C., 15 2014. Enhanced lipid accumulation in the yeast Yarrowia lipolytica by over-expression of ATP:citrate lyase from Mus musculus. Journal of Biotechnology 192, 78–84. https: / / doi.org / 10.1016 / j.jbiotec.2014.10.004

[0147] Zhang, Y., Zhang, S., Chu, Y., Zhang, Q., Zhou, R., Yu, D., Wang, S., Lyu, L., Xu, G., Zhao, Z.K., 2022. Genetic manipulation of the interconversion between diacylglycerols 20 and triacylglycerols in Rhodosporidium toruloides. Frontiers in Bioengineering and Biotechnology 10.

[0148] Zhao, C., Gu, D., Nambou, K., Wei, L., Chen, J., Imanaka, T., Hua, Q., 2015. Metabolome analysis and pathway abundance profiling of Yarrowia lipolytica cultivated on different carbon sources. Journal of Biotechnology 206, 42–51. 25 https: / / doi.org / 10.1016 / j.jbiotec.2015.04.005

[0149] Zhu, L.Y., Zong, M.H., Wu, H., 2008. Efficient lipid production with Trichosporonfermentans and its use for biodiesel preparation. Bioresource Technology 99, 7881–7885. https: / / doi.org / 10.1016 / j.biortech.2008.02.033 30

[0150] EXAMPLE 2

[0151] By targeting the locations of yeast Ty retrotransposon, the current study directly shuffled genes that potentially contribute to PuA synthesis on yeast genome and generated a recombinant yeast library with various contents of PuA. The workflow developed - 34 -by this study significantly accelerated the development of a PuA-enriched yeast strain. Through rigorous screening of 1752 strains, a recombinant S. cerevisiae capable of accumulating 26.66% of total fatty acids as PuA was obtained without the need for LA precursor feeding. In shake flask cultivation, the PuA titer reached 424.57 mg / L. The 5 following analysis identified the presence of PgFADX-containing expression cassette as well as pomegranate phosphatidylcholine: diacylglycerol cholinephosphotransferase (PgPDCT), lysophospholipid acyltransferase (PgLPCAT), acyl-CoA: diacylglycerol acyltransferase 2 (PgDGAT2), PgFAD2 and fatty acid elongase from Rattus norvegicus (RnELO2). Moreover, PuA comprised over 22% of total fatty acids in the TAG fraction of yeast single-cell oil, 10 demonstrating a considerable increase compared to PuA levels in the TAG fraction of transgenic A. thaliana and B. napus (6.6% and 10.6%). By applying tandem mass spectrometry, substantial changes in the yeast lipidome, including TAG and polar lipid fractions were found as the result of PuA synthesis.

[0152] Strains, genes, and plasmids 15

[0153] All strains used in this study are listed in Table 11. In brief, Escherichia coli DH5 was used for routine plasmid construction and preparation. To obtain the template forcloning TAG assembly and acyl-editing genes from P. granatum, total RNA was isolated from P. granatum tissues using Spectrum Plant Total RNA Kit (Sigma-Aldrich, Oakville, ON, Canada) and cDNA was synthesized using the SuperScript IV first-strand cDNA synthesis kit 20 (Invitrogen). Putative sequences coding for pomegranate acyl-editing and TAG assembly enzymes were obtained by performing a blast search against the online draft genome of P. granatum (51). Genes were inserted into the multiple cloning sites of expression plasmids using the conventional restriction cloning method or ClonExpress One Step Cloning Kit (Vazyme Biotech). S. cerevisiae BY4741- snf2 obtained from the Euroscarf collection was25 used as the starting strain in this study. S. cerevisiae H1246 quadruple mutant lacking lipid synthetic ability was kindly provided by K. Athenstaedt (Graz University of Technology, Austria) and used to test the performance of pomegranate genes. Yeast transformations were performed using the traditional lithium acetate and PEG3350 method as described previously (52). 30

[0154] The target Ty retrotransposon region was analyzed based on the S288C reference genome from the Saccharomyces Genome Database (SGD) (http: / / www.yeastgenome.org / ). Similarity and identity search for target sequences in Ty retrotransposon were conducted using NCBI Basic Local Alignment Search Tool (BLAST). The - 35 -plasmids providing the donor DNA were derived from pUC19. Four components, including upstream homology sequence, gene expression cassette, selection marker with truncated promoter region (leu2, his3, and ura3 for round 1, 2 and 3 integrations, respectively) downstream homology sequence were inserted. To release donor DNAs from the pUC19 5 backbone, either PCR or double digestion of SmaI sites flanking the donor DNA was conducted. For different genes, 500 fmol of each donor DNA were pooled together. To facilitate the integration, CRISPR / Cas9 expression vectors (pCRCT, obtained from Addgene, plasmid #60621) harbouring Cas9 protein, tracrRNA and crRNA expression cassettes were constructed and co-transformed into yeast cells with donor DNA (53). Three 20bp sequences 10 adjacent to the PAM sites (Table 12) that occur frequently in the respective target regions were selected as the spacer sequences for delivering crRNA. These sequences were ordered as single-stranded, complementary oligos and annealed to obtain double-stranded DNA. By adding homology arms, these sequences were inserted in between two Eco31I restriction sites on the pCRCT plasmid. 15

[0155] Culture conditions and optimization

[0156] For plasmid construction and preparation, E. coli was cultured at 37 °C in Luria- Bertani (LB) medium with continual shaking at 225 rpm. To sustain the plasmids in E. coli, 100 mg / L ampicillin was applied. For testing PuA accumulation in shake flask cultivation, individual colonies of the transformed yeast cells were first grown in 5 mL yeast nitrogen base (YNB) 20 supplemented with appropriate amino acid drop-out mix and 2% glucose for 24 h at 30 °C with shaking. Seed cultures were then inoculated into 1 L shake flasks containing 60 mL synthetic complete medium (6.7 g / L YNB, 2 g / L synthetic complete supplement mixture of amino acids, and 6% glucose). Inoculum for PuA production experiments was grown for 72 hours at 20 °C for 30 °C, using an incubated shaker with a shaking speed of 250 rpm. For culture condition 25 optimization, Box–Behnken Design was employed for response surface methodology. The factors considered for optimization include carbon level (X1), initial pH (X2), and initial OD. They were examined at three levels (6%, 10.5%, and 15% for glucose; 5.3, 6.5, and 7.6 for initial pH; 0.2, 0.6, and 1 for initial OD). The two responses considered for analysis are PuA content (% of total fatty acids) and PuA titer (mg / L). A 17 run randomized experiments were 30 conducted. Based on the result, second-order polynomial equations were fitted to the experimental results to predict the optimal points within experimental constraints (Table 10).

[0157] Preparation of plant oil hydrolysate and fatty acid feeding - 36 -

[0158] Preparation of free fatty acids from pomegranate seed oil was conducted using chemical hydrolysis (54). In brief, 50 g of pomegranate seed oil was mixed with 12 g potassium hydroxide, 117 mL pure ethanol, and 35 mL H2O in a shake flask flushed with nitrogen.1.65 mL of butylated hydroxytoluene (BHT) solution (50 mg / mL) was added to the mixture to protect 5 the sensitive punicic acid. The reactor was then sealed and maintained at 50 °C with constant shaking. After 1 hour of incubation, the removal of the unsaponifiable matter was performed three times by mixing 100 mL of distilled water and 100 mL of hexane. The free fatty acid was extracted with 100 mL hexane and dried over anhydrous sodium sulphate. The solvent was then removed under vacuum to obtain pomegranate oil hydrolysate that contains free PuA. 10 When fatty acid feeding is necessary, 0.03% v / v ethanol dissolved linoleic acid or pomegranate oil hydrolysate was supplemented to the culture medium along with 0.2% non-ionic surfactant NP-40 (TERGITOL™ solution) for even distribution of fatty acid in the aqueous medium.

[0159] Nile red staining of neutral lipids in yeast

[0160] The Nile Red fluorescence detection was conducted as described previously15 (14). Briefly, 100 μl aliquots of the yeast cell suspension were transferred to a 96-well dark flat- bottom plate (Corning Inc.). The background fluorescence was measured using a Synergy H4 Hydrid multimode microplate reader (Biotek Instrument, Inc.) with emission and excitation filters set to 485 and 538 nm, respectively.5 μl of newly prepared methanolic Nile red solution (0.1 mg / ml) was added and the second fluorescence intensity was measured. The Nile Red20 values were calculated based on the change in fluorescence over OD600 ( F / OD600).

[0161] Lipid extraction and separation of lipid class using thin-layer chromatography (TLC)

[0162] Before lipid extraction, yeast biomass was harvested from liquid culture via centrifugation. The supernatant was then removed and 800 μL of a cold lipid extraction mixture 25 comprising chloroform and isopropanol (2:1, v / v) was added, along with glass beads (0.5mm) and BHT at a final concentration of 0.01%. Subsequently, cellular disruption was achieved through three cycles of bead beating (1-minute duration each) using a Biospec bead beater (Bartlesville, OK), with a 2-minute cooling on ice between each cycle. The extraction procedure was repeated twice for each sample. The collective organic phase, containing both polar lipids 30 and TAG, was dried under nitrogen and resuspended in 200 μl chloroform. For preparing fatty acid methyl ester, transmethylation was done via a base-catalyzed method using 1 mL of 5% sodium methoxide dissolved in methanol (10). After incubation at 30 °C for 1 hour, the reaction was stopped by adding 1.5 mL of 0.9% (w / v) sodium chloride solution. Fatty acid methyl esters - 37 -were then extracted with 1 mL of HPLC-grade hexane. For lipid class analysis, the lipid extracted from each sample was separated using one-dimensional TLC with silica gel-coated plates (0.25 mm Silica gel, DCFertigplatten, Macherey-Nagel, Germany). The TLC plates were developed using a solvent mixture comprising hexane / diethyl ether / acetic acid (in a 70:30:1 5 ratio). Lipid fractions on the TLC plate were visualized via 0.05% primulin staining under UV light. Bands corresponding to target lipid fractions were scraped, extracted, derivatized via a base-catalyzed method, and subsequently analyzed.

[0163] Positional analysis of triacylglycerol and polar lipids

[0164] The fatty acid distribution between sn-2 and sn-1 / 3 TAG was analyzed by 10 previously described enzymatic reaction (11, 55). Briefly, after TLC separation, TAG was recovered from silica gel, and transferred to a new screw cap tube and dried under nitrogen. Subsequently, 1 mL Tris-HCl buffer (1 mM, pH 8.0), 100 L 2.2% CaCl2 and 250 L 0.1% deoxycholate were added. Each mixture was vortexed for 2 minutes and sonicated for 60 seconds to emulsify the lipid. The mixture was pre-warmed in a water bath at 40 °C for 30 s, 15 and then 20 mg pancreatic lipase (pancreatic lipase type II, Sigma) was added to initiate hydrolysis. The mixture was further incubated for 3 min at 40 °C, and the reaction was terminated by adding 500 L of 6 M HCl. The resulting lipids, containing unreacted TAG, DAG, 2-MAG, and free fatty acids, were extracted twice with 3 mL of diethyl ether. The lipids samples were separated using TLC and the sn-2 MAG was scraped into screw cap tubes for 20 transmethylation.

[0165] Positional analysis of polar lipids was performed by cleaving the fatty acids at the sn-1 position of polar lipids using phospholipase A1 (11, 56). In brief, phospholipase A1 was first mixed with water in a 1:4 (v / v) ratio. Polar lipids were recovered from the silica gel and dissolved in 2 mL of diethyl ether.1 mL of phospholipase A1 (Sigma) solution was added 25 to initiate hydrolysis. The mixture was then vortexed at maximum speed for 5 min and the reaction was terminated by evaporation of diethyl ether under nitrogen. The hydrolyzed lipids were extracted and separated by TLC. The cleaved fatty acids were scraped into screw cap tubes for transmethylation.

[0166] Gas chromatography-flame ionization detection (GC-FID) and lipidomic 30 profiling

[0167] The FAMEs were resuspended in hexane and analyzed on an Agilent 6890N Gas Chromatograph equipped with a 5975 inert XL Mass Selective Detector (Agilent Technologies) and flame ionization detector using the method described in our previous study - 38 -(6). Briefly, the FAMEs were separated on a capillary column DB23 (30222 m×0.25 mm×0.25 μm, Agilent Technologies, Wilmington, DE, USA) using the following program: 2:1 split ratio, 1 μL injection.4 min at 165 °C, then increased to 180 °C (10 °C / min) and held for 5 min, and increased to 230 °C and held for 5 min. The lipidomic analyses described in this work were 5 performed at the Kansas lipidomics research center analytical laboratory. The lipidome of three chosen yeast strains under two culture conditions (totally four lipidomes) was investigated. In the analysis, four cultures of each yeast strain were used as replicates. All biological replicates were cultured under the same conditions, and samples were collected at the same time after 96 hours. 10

[0168] Results and Discussion

[0169] Functional validation and complementation assay of pomegranate TAG- assembly genes and PgFADX variants

[0170] The acyl-editing and TAG assembly pathways are important biochemical processes involved in the modification of fatty acids in plants, particularly in those that produce 15 UFA (Fig. 6A). Within these pathways, specialized enzymes collaborate to modify and exchange specific acyl groups, redistributing them across various lipid pools and enriching the target fatty acid within the TAG fraction. Since most pomegranate enzymes involved in TAG biosynthesis have not been verified functionally. We first conducted the identification and isolation of genes encoding acyl-editing and TAG-synthesizing enzymes from pomegranate. 20 As shown in Table 5, enzymes from tung tree (Vernicia fordii), castor bean (Ricinus communis), and flax (Linum usitatissimum), which are involved in the synthesis of UFA or PUFA were used as queries. Results showed three pomegranate DGAT2s, three PDATs, and one DGAT1, PDCT, LPCAT, GPAT9, LPAT2, PLA2, PLC, and LACS8 were detected. In plant lipid studies, PDAT and DGAT have often been the subject of investigation owing to their direct contribution 25 to the final step of TAG synthesis (12, 13). Given their critical roles, we first conducted a functional validation assay of these proteins. The genes encoding PgDGAT1, PgDGAT2.a, PgDGAT2.b, PgDGAT2.c, PgPDAT.a, PgPDAT.b, and PgPDAT.c were cloned from pomegranate cDNA and transformed into S. cerevisiae mutant strain H1246. H1246 is a yeast quadruple mutant that lacks genes which encode yeast native DGAT, PDAT and acyl-CoA: 30 sterol acyltransferases, thus devoid of TAG synthesizing ability (14). Expressing heterologous TAG assembly enzymes in this yeast mutant will result in the reconstitution of TAG biosynthesis, given the enzyme is fully functional. As shown in Fig.6B, after 72h culture, yeast cells expressing PgDGAT1, PgDGAT2.a, and PgDGAT2.b produced strong TAG bands, - 39 -suggesting their strong complementary activity in restoring TAG synthesis using yeast native fatty acids. Moreover, since pomegranate contains high levels of PuA, it is plausible to assume that pomegranate-derived TAG assembly enzymes may evolved the preference for PuA- containing substrate. Therefore, to further evaluate their abilities, recombinant yeast strains 5 were cultured in the presence of exogenously added pomegranate oil hydrolysate which contains free PuA.

[0171] Interestingly, pomegranate PgDGAT2.c and PgPDAT.a produced weak TAG bands when pomegranate oil hydrolysate was provided, suggesting fatty acids commonly present in yeast may not be the preferred substrates for these enzymes (Fig.6C). It is also 10 worth noting that PgDGAT2.a, PgDGAT2.b, and PgDGAT2.c produced TAG with higher unsaturation levels, which have a slower migration rate on the TLC plate. Subsequent GC-FID analysis of the TAG bands confirmed a reduction of over 60% in saturated fatty acid content in cells expressing PgDGAT2.a and PgDGAT2.b compared to those expressing PgDGAT1 (Fig.6D). When pomegranate oil hydrolysate was added to the culture, PuA accounts for over 15 20% in the TAG fraction of strain expressing PgDGAT2.c, which suggested that this enzyme might prefer substrates containing PuA. While yeast has been extensively used as a model organism to study TAG assembly genes in plant species, it has been shown that some plant- derived enzymes have poor expression levels in yeast cells due to unknown reasons (14). This could potentially account for the failure of PgPDAT.b and c to complement TAG synthesis in 20 H1246, and the lower activity of PgDGAT2.c and PgPDAT.a compared to PgDGAT1 when free PuA was supplied.

[0172] In this study, the substrates for PgFADX desaturase activity are phosphatidylcholines with precursor fatty acids (oleic acid or LA) esterified to the sn-2 position (Fig.6A). Molecular oxygen and electrons are also required as the co-substrates. The 25 desaturase activity may also be inhibited by H2O2(15), which arises from the normal metabolic processes in vivo. Our previous study found that fusing the substrate-binding protein with the catalyzing enzyme works better than co-expressing them separately, particularly in the context of yeast lipid accumulation (16). A kinetic improvement was achieved by fusing algal DGAT with an acyl-CoA binding protein, which attracts the substrates for DGAT-catalyzed reaction. 30 In addition, protein fusion also has the potential to effectively remove harmful byproducts. For example, a previous study found that combining aldehyde deformylating oxygenase with catalase could alleviate the inhibition caused by the accumulation of the byproduct H2O2(17, 18). Therefore, in our study, selected sequences encoding phosphatidylcholine-binding - 40 -proteins (SCP2), oxygen carrier-proteins (Vhb), soluble domain of electron transporter (CB5SD) or catalase (CAT) were cloned and linked with PgFADX at the N-terminus by a flexible peptide linker (ASGAGGSEGGGSEGGTSGAT (SEQ ID NO: 20)). The expression cassettes containing protein fusions were transformed into BY4741 snf2 mutant, which5 exhibits an elevated level of fatty acid synthesis (19–21). As shown in Fig.1E, except for the fusion with catalase, all other PgFADX variants showed a slight increase in the PuA level. Given the significantly larger size of catalase (84.2 kDa) compared to other proteins, it is possible that combining this enzyme with FADX potentially disrupted the structure of PgFADX. In addition, we also constructed a BY4741 snf2 snf1 double knockout mutant by replacing10 the snf1 gene with AtSCP2-PgFADX expression cassette. Yeast SNF1 (Sucrose Non- Fermenting 1) is a protein kinase and master regulator that plays a critical role in regulating yeast energy metabolism and glucose homeostasis (22, 23). SNF1 is activated when the ATP level in vivo is low, leading to the degradation of storage lipids and the inhibition of lipid synthesis (24). As shown in Fig. 6E, when AtCB5SD-PgFADX was expressed in this strain 15 background, PuA accounts for 6.17% of total fatty acids. In addition to the proposed function, modifications to the N-terminus of PgFADX may also influence the turnover rate of this key enzyme. For instance, a previous study showed the N-terminal sequences of Brassica and tung tree FAD3 proteins confer a rapid degradation of the fluorescent reporter protein in both plant and yeast cells (25, 26). For FAD2 expression, the N-terminus also plays an important 20 role in controlling the protein half-life (27). Given the structural similarities between various UFA-producing enzymes and FAD2, further characterization of their protein half-life in transgenic plant and yeast cells could aid in the design of a more stable enzyme, thereby improving UFA production.

[0173] Improved bioconversion from LA to PuA in yeast cells constructed by Ty 25 retrotransposon-targeted random gene shuffling

[0174] The above result was obtained using the plasmid-based expression system, which is less stable and can be easily lost without selective pressure, leading to lower expression levels and a decrease in enzyme concentration in vivo (28). In addition, the plasmid copy number varies depending on the genes it contains (29). Therefore, following the 30 characterization of pomegranate genes, we developed a workflow aiming for the effective integration and testing of pathway genes responsible for UFA accumulation directly on the yeast genome (Fig.7). To deliver this objective, yeast Ty retrotransposon elements were chosen as the locus for integration and CRISPR / Cas9 systems were used to facilitate the - 41 -process. There are four flanking long terminal repeats associated with S. cerevisiae Ty retrotransposon elements, in which the delta sequences are the most abundant (30–33). Since retrotransposon regions do not participate in yeast metabolism, integration of pathway genes into this region will lead to greater copy numbers with lower interference with cellular fitness. 5 Therefore, sequence analysis of yeast delta sequence was conducted first. In total, 299 sequences were pulled from the yeast genome based on the genomic sequence of S288C. After sequence alignment, YERCdelta20 (Table 6), YDRWdelta23 (Table 7), and TyA Gag gene (Table 8) were chosen as the targets for gene shuffling due to their high similarity with other analyzed sequences. 10

[0175] A brief description of the workflow is shown in Fig.7A. First, the expression cassettes for providing the donor DNA were constructed. Homology arms targeting yeast Ty retrotransposon were added as the flanking sequences. Since previous studies have shown the selection marker with a defective promoter causes a selective pressure for cells to increase the copy number of the plasmid (34), auxotrophic markers with truncated promoter regions 15 (down to 20-50bp) were linked with candidate genes to enhance the probability of obtaining positive transformations (Fig.7B). Next, the cassette was linearized and co-transformed into yeast cells as donor DNA with Ty retrotransposon sequence targeting CRISPR / Cas9 system. After obtaining the transformants, single colonies were cultured in liquid mediums.0.03% LA was added to the liquid culture in rounds 1 and 2 but was omitted in round 3 screening, which 20 aimed for PuA neosynthesis. After culturing under 20°C for 72 hours, the cell suspension was subjected to OD600 measurements and Nile red staining to determine intracellular lipid level. The rest of the biomass was harvested for detecting PuA levels.

[0176] Given the critical role of PgFADX, the first round of yeast Ty retrotransposon- targeted integration was conducted to solely integrate the AtCB5SD-PgFADX expression25 cassette into BY4741 snf2 snf1 double knockout mutant. As shown in Fig.3A, 360 strainswere screened in total. With LA feeding, PuA accumulation was detected in all tested transformants. Merely 10.3% of strain candidates accumulated PuA less than 6% of total fatty acids. The majority of the strain candidates (83.7%) accumulated 6%-10% PuA, and 6.1% of the strain candidates accumulated higher than 10% of total fatty acids as PuA. The PuA 30 content in the best strain, designated as CARIA266, reaches nearly 11%. To further probe the performance of CARIA266, we conducted a growth analysis of this strain in shake flasks under different culturing temperatures. As shown in Fig.8B, CARIA266 maintained a high PuA level of up to 12.59% at 96 hours of cultivation under 20 °C. Compared to 30 °C, CARIA266 - 42 -accumulates more monounsaturated fatty acid (MFA) under 20 °C, and the level of PuA gradually increased during the 4-day periods. When cultured under 30 °C, the level of PuA was relatively stable and lower than the level obtained under 20 °C. The highest titer of PuA in the liquid culture on day 5 corresponded to 68.649 mg / L. 5

[0177] Instead of being used as a fatty acid precursor, CARIA266 enriched LA in yeast lipids by nearly 50%. Even though the PuA level obtained by Ty retrotransposon-targeted integration is significantly higher than the plasmid-based approach, the unexpected accumulation of LA indicates that this fatty acid was not efficiently edited, and yeast native TAG synthesis may not favour PuA product. One possible cause is the lack of specialized 10 enzymes from pomegranate acyl-editing and TAG assembly network (Fig.6A). Therefore, a round 2 Ty retrotransposon-targeted random gene shuffling was conducted, focusing on the genomic integration of those enzymes. Since PgDGAT1, PgDGAT2a, PgDGAT2b, PgDGAT2c, and PgPDAT.a showed successful complementation of TAG synthesis in yeast H1246 (Fig.6B, 6C, 6D), the encoding genes along with PgPDCT, PgLPCAT, PgGPAT9, 15 PgLPAT2, PgPLA2, PgPLC, and PgLACS8 were selected as the candidates for round 2 integration. All genes were under the control of the TEF1 promoter and CYC1 terminator. Donor DNAs were pooled together and co-transformed into CARIA266 with YDRWdelta23- targeting CRISPR / Cas9. The resulting transformants were again cultured in the presence of 0.03% LA. As shown in Fig. 8C, 792 strains were screened in total for round 2 integration. 20 22.6% produced PuA below 8% of total fatty acids, 69.8% produced 8%-12% PuA, and 7.6% produced PuA higher than 12% of total fatty acids. Compared to round 1 integration (Fig.8A), the average signal intensity of the PuA from round 2 samples was significantly higher (Fig. 8C), indicating the enhanced production of PuA. The vast majority of the population (96%) of round 2 transformants has higher Nile red fluorescence, suggesting the improvement in neutral 25 lipid accumulation caused by the pomegranate acyl-editing and TAG assembly genes. The PuA content in the best strain, designated as CARIB650, reaches 16.071% in tube culture. As shown in Fig.8D, when CARIB650 is cultured in the shake flask under 20 °C, the highest PuA content further increased to over 17.7% of total fatty acids, and the highest PuA titer reached 93.615mg / L. Compared to CARIA266, the content of LA on day 4 significantly dropped by 30 25%, indicating a higher conversion rate of this fatty acid precursor.

[0178] High-level neosynthesis of PuA in yeast cells using yeast Ty retrotransposon-targeted random gene shuffling and culture condition optimization. - 43 -

[0179] The above results were obtained by bioconversion of manually fed LA precursor to PuA, which enters the yeast metabolism as a free fatty acid and then linoleoyl-CoA (Fig.6A). Before being recruited to the sn-2 position of PC and edited by PgFADX, it is very likely that other lipid biosynthetic pathways may compete for this substrate. Given the additional cost and 5 the requirement for constant precursor quality associated with the bioconversion method, neosynthesis may be a better option. Therefore, a round 3 integration was conducted to generate LA precursor in vivo, directly on the sn-2 position of PC. The codon-optimized donor DNA encoding AtCB5SD-FADX, PgFAD2, Puccinia graminis acyl-CoA desaturase (PgOLE1) and R. norvegicus fatty acid elongase (RnELO2) (Table 9) were transformed into CARIB650 10 strain to obtain the library for round 3 screening. In total, 600 strains were screened without LA feeding (Fig.8E). Via neosynthesis, 15.5% of the strain candidates produced 2%-10% PuA, 3.3% produced 10%-14% PuA and 0.8% produced PuA higher than 14%.80.4% of the strain candidates only produced 1%-2% PuA, which is possibly caused by the lack of PgFAD2 integration. The best strain, designated as CARIC568, accumulated 15.998% PuA in tube 15 culture. The shake flask cultivation of CARIC568 over 6 days showed it is capable of accumulating 18.38% PuA (101.84mg / L) under 20 °C on day 4 without LA feeding (Fig.8F). In total, 1752 yeast transformants were screened using the workflow proposed by this study. The PuA content and product titer have been improved significantly, and the content of total unsaturated fatty acid was increased from 81% of total fatty acids (with LA feeding) in 20 CARIA266 to 90% in CARIC568 (without LA feeding). As shown in Fig.9, the PuA titer was higher at 20 °C compared to 30 °C. Previous studies have shown that PUFA concentration in plant cells rises with decreasing temperature due to the post-transcriptional regulation of fatty acid desaturase (35). Culturing temperature also influences the abundance of plant FAD2 and FAD3 proteins expressed in yeast cells (26, 27). Consistently, a lower temperature may extend 25 the half-life of pomegranate-derived desaturase in yeast cells thus improving the production.

[0180] To maximize PuA content and titer in shake flask cultivation of strain CARIC568, a response surface methodology using a multifactorial Box-Behnken design was employed. The chosen variables including carbon (X1), initial pH (X2), and initial OD (X3), were varied at three levels (-1, 0, and +1), as detailed in Table 10. Results showed carbon source level and 30 initial pH are the most significant factors affecting PuA content in CARIC568 (Fig.9A and 9B). Higher pH, carbon and initial OD levels promote PuA content and production (Fig.11 and Fig. 12). Second-order polynomial equations were fitted to the experimental results to predict the optimal points within experimental constraints (Table 10), which led to optimized conditions at - 44 -12% glucose, pH 7.04, and an initial OD of 0.721. A verification experiment in the optimized condition was then carried out to determine the accuracy of the prediction. Following 120 hours of incubation, CARIC568 produced 424.57 mg / L PuA (Table 4). The lipid content reached 15.5% of dry cell weight, with PuA accounting for 26.66% of total fatty acids. The results 5 validated the predictions provided by the statistically based experimental designs used in this study. To the best of our knowledge, this is the highest content and titer reported to date for PuA production in S. cerevisiae.

[0181] Table 4. Fatty acid profile and PuA production of CARIC568 over a 6-day growth period in the optimized medium. Fatty acid content (% of total fatty acids)PuA titerC16:0 C16:1 C18:0 C18:1 C18:2 PuA (mg / L) 6.522 ± 15.352 ± 13.336 ± 31.011 ± 26.449 ± 7.334 ± 25.556 ± 24h 0.064 1.159 0.383 1.017 0.174 0.431 10.1962.764 ± 9.823 ± 13.048 ± 27.159 ± 31.41 ± 15.798 ± 180.378 ± 48h 0.169 0.42 0.186 0.234 0.511 0.742 48.256 1.415 ± 7.583 ± 11.003 ± 27.391 ± 31.067 ± 21.544 ± 271.683 ± 72h 0.062 0.327 0.493 0.344 0.392 0.094 35.745 0.915 ± 7.022 ± 8.832 ± 26.376 ± 31.777 ± 25.08 ± 330.908 ± 96h 0.036 0.475 0.598 0.923 0.705 0.419 32.034 0.774 ± 6.995 ± 7.343 ± 26.517 ± 31.711 ± 26.663 ± 424.572 ± 120h 0.078 0.88 0.591 0.917 0.34 1.01 32.591 0.634 ± 6.804 ± 6.813 ± 26.869 ± 32.276 ± 26.607 ± 379.995 ± 144h 0.035 0.844 0.238 0.912 0.383 1.245 59.084 10

[0182] So far, the exact roles of enzymes involved in plant UFA synthesis are still under investigation. Determining the optimal ratio of these proteins in plant or microbial hosts adds to the challenge of efficient heterologous production of UFA. Instead of using traditional gene stacking strategies, where the exact gene candidates, combinations, and ratio are determined 15 before the experiment, this study randomly combined and integrated potentially necessary pathway genes into the yeast genome using CRISPR / Cas9 at a variable copy number, resulting in a pool of transformants. The candidates were then screened, yielding a strain with a gene combination that is best suited for UFA synthesis under specific culture conditions. - 45 -

[0183] Accordingly, subsequent analysis detected the presence of AtCB5SD-PgFADX as well as PgPDCT, PgLPCAT, PgDGAT2.c, PgFAD2 and RnELO2 in CARIC568. Normally when BY4741 was cultured under 30 °C, C16 fatty acids accounted for over 60% of its total fatty acid composition (36). In comparison, total C16 fatty acid accounts for only 9% at 72 hours 5 in BY4741-derived CARIC568 grown at 20 °C (Table 4), which is 85% lower than the literature value. In addition to PgFAD2 and PgFADX, which stimulated lipid synthesis to produce more LA and PuA, the presence of RnELO2 also greatly contributed to the increase in C18 fatty acid. This result is consistent with previous studies, in which the expression of the rat elongase 2 gene (rELO2) reversed the relative quantities of C16:1 and C18:1, and introducing the 10 elongase rather than the desaturase was more efficient in raising the quantity of C18:1 (37, 38). PDCT, LPCAT, and DGAT2 hold special positions in the plant TAG assembly and acyl- editing pathway. Arabidopsis PDCT, encoded by the ROD1 gene, plays a critical role in sending C18:1 to PC for desaturation and transferring PUFA products into TAG synthesis (39). A recent study also suggests that PDCT plays an even more important role in UFA-producing15 plants. Castor PDCT and DGAT2, which have been well characterized to prefer UFA- containing acyl-CoA, work together to increase tri-ricinoleoyl glycerol levels in castor seeds while also generating nonricinoleate lipids for membrane biosynthesis (9). LPCAT was also involved in this process. As illustrated in Fig.6A, plant LPCAT enzymes play an important role in regulating acyl-CoA composition by catalyzing both the forward reaction to synthesize PC 20 and the reverse reaction to release acyl-CoA. LPCAT has low activities with UFA groups in the forward reaction, but higher activity on common unsaturated C18 fatty acids (40). When coupled with flax DGAT in yeast cells, flax LPCAT reverse reaction specifically incorporates PUFA into TAG (41), which is consistent with what has been found in the synthesis of PuA in yeast (6, 42). 25

[0184] It should be noted that, given S. cerevisiae's limited PuA accumulating capabilities (0.3–3.7%) observed in our previous study (6), at the beginning of the round 1 and round 2 integration, we chose a bioconversion strategy to directly convert exogenously added LA to PuA. As a result, in future, it would be interesting to conduct another round of Ty retrotransposon-targeted random gene shuffling of pomegranate acyl-editing and TAG 30 assembly genes on CARIC568. Given the presence of a complete pathway for PuA neosynthesis, a better combination of pomegranate genes in terms of PuA neosynthesis could be found. In addition, considering the potential contribution of yeast native lipid anabolism to PuA accumulation, the genes encoding yeast native fatty acid synthesis and TAG assembly - 46 -enzymes in CARIC568 were kept intact. However, these enzymes may compete with incorporated heterologous enzymes and preferentially integrate fatty acids other than PuA into the glycerol backbone of the TAG. Replacing them with equivalent pomegranate genes may further enhance PuA levels in yeast lipids. 5

[0185] Lipidomic analysis revealed substantial changes in yeast lipidomes

[0186] After three rounds of genomic integration, three strains with medium to high PuA content were developed, making them excellent for investigating the impact of heterologous PuA synthesis on the yeast native lipidome. To this end, lipids were extracted from CARIA266 and CARIC568 without LA feeding, as well as CARIA266 and CARIB650 with 10 0.03% LA feeding. All strains were cultivated in the optimized medium for 96 hours. Extracted lipids were then separated by TLC for the distribution analysis of PuA. As shown in Fig. 9C and 9D, when LA was omitted in the culture medium, the PuA level in CARIA266 was relatively low. When LA was available, CARIA266 accumulated 7.8% and 51% of TAG as PuA and LA respectively. CARIB650 accumulated 14.99% PuA in TAG and 17.63% PuA in polar lipids after 15 adding pomegranate acyl-editing and TAG assembly genes, representing a 92% and 25% increase over CARIA266, respectively. Meanwhile, the content of LA in both TAG and polar lipid fraction was lowered in CARIB650. By synthesizing LA directly on PC, CARIC568 accumulated the highest PuA (22.37%) and the lowest LA (34.36%) in TAG fraction compared to other strains. The level of PuA in the TAG fraction is a substantial improvement over our 20 prior study, in which two genetically engineered oilseed plants A. thaliana and B. napus only accumulated 6.6% and 10.6% of the fatty acid in TAG as PuA. Positional analysis of TAG showed LA was the dominant fatty acid at the sn-2 position of TAG, followed by oleic acid and PuA (Fig.9E). In CARIA266, the sn-2 position of TAG contains 44.3% LA and 6.33% PuA. After round 2 integration, LA at the sn-2 position of TAG decreased by around 19% whereas PuA 25 was increased by 81.7%. In terms of CARIC568, PuA accounted for 14.5% of the fatty acids in the sn-2 TAG and 26.3% of the fatty acids at the sn-1 / 3 TAG, suggesting a slight preference for enriching PuA at the sn-1 / 3 position in this strain (Fig.9G). Positional analysis of polar lipids indicated that the sn-1 position contained relatively low PuA (Fig.9F). In contrast, PuA accounted for 21.68%, 25.25%, and 47.13% of the fatty acids in the sn-2 position of polar 30 lipids, which is consistent with the widely held opinion that the sn-2 position is the major location for acyl-editing (Fig.4H).

[0187] To better study the influence of PuA production on the yeast lipidome in detail, we conducted a lipidomic analysis of yeast lipids. A principal components analysis (PCA) was - 47 -first conducted to determine the difference between the lipidomes of CARIA266, CARIA266 with LA addition, CARIB650 with LA addition, and CARIC568 (Fig.13). Analysis of the entire lipid dataset including different polar lipids, TAG and DAG revealed a clear distinction between the four lipidomes. The first component clearly distinguishes between the CARIA266 and the 5 CARIC568, accounting for 59.3% of the differences. The second major component accounted for approximately 25% of the changes and shows the difference between groups with LA addition and without LA addition. When LA was added exogenously, CARIA266+LA clustered firmly with CARIC568+LA and away from groups without LA feeding, indicating that fatty acid precursor uptake dominated the shift in yeast lipidome. The distance between the final strain 10 CARIC568 and starting strain CARIA266 on the PCA plot indicates significant alterations in the yeast lipidome as a result of genetic modifications made by Ty retrotransposon-targeted integration.

[0188] The TAG species in yeast lipidome were separated into saturated fatty acid (SFA)-containing TAG, monounsaturated fatty acid (MFA)-containing TAG, and 15 polyunsaturated fatty acid (PUFA)-containing TAG. Heatmaps displaying the relative abundance of different TAG species are shown separately in Fig.10A-C. The most abundant SFA-containing TAG in CARIA266, CARIA266+LA, CARIB650+LA, and CARIC568 are TAG(54:2_18:0), TAG(54:4_18:0), TAG(54:4_18:0), and TAG(54:3_18:0), respectively. In CARIA266, the overall content of TAG containing SFA was noticeably greater. Compared with 20 the CARIC568, CARIA266 accumulates more TAG with shorter chain SFA, such as TAG(50:2_16:0) and TAG(52:2_16:0). In terms of MFA-containing TAG, CARIA266 accumulates a higher content of C16:1-containing TAG, and the overall saturation of CARIA266 TAG is higher. As seen in Fig.10C, the content of C18:2-containing TAG was relatively greater in CARIA266+LA and CARIB650+LA as a result of exogenously added LA. 25 In this regard, CARIA266+LA had a higher level of C18:2-containing TAG than CARIB650, indicating the introduction of pomegranate TAG assembly and acyl-editing genes suppressed the level of TAG species containing only LA. Since S. cerevisiae does not have native C18:3 fatty acids, the C18:3 detected in the TAG fraction solely reflects the content of PuA. CARIC568 has the highest content of PuA-containing TAG species, led by TAG(54:6_18:3) 30 and TAG(54:5_18:3). Compared to the other three lipidomes, CARIC568 has a significantly higher level of TAG species with high unsaturation degree (5-8 double bonds in total). However, in contrast to pomegranate seed oil, where the majority of PuA incorporated into TAG occupies all three positions of the glycerol backbone, the absence of TAG(54:9) in the - 48 -top-ranking TAG species in all lipidomes suggests the PuA is still primarily found at a single or double position of the glycerol backbone of TAG.

[0189] The comparison of differentially expressed lipids between various strains is shown in the volcano maps (Fig.10E-H). As indicated in Fig.5E, TAG accounts for 60% of the 5 up-regulated lipids with C18:2-containing lipids being the majority when exogenous LA was made accessible to CARIA266. After round 2 integration, 82 lipid species were up-regulated, with TAG accounting for 50% of them. Meanwhile, 45 lipid species were down-regulated, with TAG accounting for 78% of them (Fig.10F). By switching from bioconversion to neosynthesis, 34 TAG and 25 polar lipid species were up-regulated, whereas 47 TAG and 20 polar lipid 10 species with saturated acyl chains were down-regulated (Fig.10G). When comparing the finial strain CARIC568 with CARIA266, significant changes in 205 lipid species were observed. TAG accounts for 55% of both up- and down-regulated lipids (Fig.10H). Most up-regulated lipids have more than four double bonds, whereas the down-regulated lipids consist of shorter acyl chains with fewer double bonds, such as PC(34:1) (decreased by 94.7%) (Fig.10H). 15

[0190] The tested yeast samples contain a variety of phospholipids, including PC, lysophosphatidylcholine (LPC), lysophosphatidylethanolamine (LPE), phosphatidylserine (PS), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI), and phosphatidic acid (PA) (Fig. 10D). In CARIA266, CARIA266+LA, CARIB650+LA, and CARIC568, the most abundant polar lipid species are PE(34:2), PI(36:2), PC(36:4), and PC 20 (36:5). Previous study found the parental strain of BY4741 (S288c) lack polar lipids containing fatty acids with more than two double bonds (43). This remains the case for strain CARIA266 without LA feeding, caused by the weak ability to accumulate PuA. S. cerevisiae's physiology and cellular processes rely heavily on phospholipids. For example, PC, the most abundant phospholipid in S. cerevisiae membranes, helps to maintain membrane integrity, fluidity, and 25 stability (43). Therefore, the optimal setting for PuA formation in S. cerevisiae is to cause as little disturbance as possible to the yeast's native polar lipid species while enriching the TAG fraction with the majority of the PuA. However, as shown in Fig.10D, the content of polar lipids with more than two double bonds rises in accordance with PuA accumulation capability, and this up-regulation was observed across major polar lipid species, such as PC(36:4), PC(36:5), 30 PI(36:3), PE(36:5), PE(36:5), and PS(36:3). Certain important branching points exist in phospholipid biosynthesis that may control the acyl-chain species. For example, yeast EPT1 showed the highest activity toward di-unsaturated DAG species as lipid substrates, while yeast CPT1 favours C32:2 substrates (44). A previous study on human CDP-diacylglycerol - 49 -synthases (CDS) revealed that this key enzyme, catalyzing the formation of CDP- diacylglycerol, also exhibits acyl-chain specificities for its lipid substrate (45). Although little information is known concerning yeast native lipid synthetic enzymes' selectivity on conjugated PUFA, the polar lipidomic profile obtained in this study suggested a crosstalk between PuA 5 and major polar lipid synthesis.

[0191] On the other hand, the presence of PuA in yeast polar lipid and its accompanied influence on cellular fitness may contribute to the design of a novel rapid screening method, targeting cells with higher PuA. Since screening by gas chromatography is relatively time- consuming and labour-intensive, previous studies have used UV spectral scans to detect the 10 amount of conjugated dienes and trienes in bacterial isolates or pomegranate oils (46, 47). However, such methods still require the prior extraction of conjugated fatty acids. Alternatively, the membrane lipid composition of S. cerevisiae is known to affect strain survival under stress conditions (48). For instance, when Arabidopsis FAD2 was expressed in S. cerevisiae, yeast cells acquired greater resistance to 15% (v / v) ethanol (49). Expressing FAD2 from the rubber 15 tree (Hevea brasiliensis) increased the S. cerevisiae's susceptibility to oxidative stress, triggered by paraquat, tertbutyl hydroperoxide, and hydrogen peroxide-induced lipid peroxidation (50). From a future perspective, it would be intriguing to examine how CARIC568 responds to these stress factors. These stress conditions could serve as selective or counter- selective measures to identify the strain with the highest PuA level among the transformants 20 produced by Ty retrotransposon-targeted random gene shuffling.

[0192] Conclusion

[0193] In this study, genes potentially contributing to plant-derived UFA synthesis were directly integrated and shuffled on the S. cerevisiae genome. Screening 1752 strains led to the identification of a recombinant capable of accumulating 26.66% of total fatty acids as PuA 25 without LA precursor feeding. In shake flask cultivation, the PuA titer reached 424.57 mg / L. PuA comprised over 22% of total fatty acids in the TAG fraction of yeast single-cell oil, a significant increase compared to transgenic A. thaliana and B. napus. Subsequent analysis revealed significant lipidome changes due to PuA synthesis. Based on the result, the PuA level in this PuA-enriched yeast platform could be further enhanced by iterative integration using the 30 same workflow.

[0194] A results-driven random gene shuffling approach was developed to investigate the biosynthesis of plant-derived punicic acid (PuA) in yeast without step-by-step testing of each gene or gene combination. Genes potentially contributing to PuA synthesis were directly - 50 -shuffled into the yeast genome. The identified strain had 26.7% of total fatty acids and 22% of that of the triacylglycerol fraction as PuA through neosynthesis. Further lipidomic analysis revealed both the prospects and potential bottlenecks for PuA production in microorganisms. PuA production can be further enhanced through iterative integration using the same 5 approach. This strategy could facilitate the production of other high-value, plant-derived unusual fatty acids in transgenic hosts, and be easily adopted for the production of other value added bioproducts.

[0195] Table 5 Pomegranate acyl-editing and TAG assembly genes Enzym Query Query Annotation Subject No Designation Subject protein Subject CDS e species protein species .- 51 -PDCT Linum >AHE8067 Flax PDCT, natural Punica 12 PgPDCT >XP_03138944 >XM_031533587.1:34 usitatissi 9.1 producer of 18:3 PUFA granatum 7.1 9-1212[ ] a e eque ce s a y a e y seac o e a us g asc Local Alignment Search Tool (BLAST) - 52 -sic Local Alignment Search Tool (BLAST) - 57 -Basic Local Alignment Search Tool (BLAST) - 61 -

[0199] Table 9 codon optimization of AtCB5SD-FADX, PgFAD2, PgOLE1 and RnELO2 - 63 -

[0200] Table 10 Optimization of culture condition using response surface methodology. 5

[0201] Table 11 Strains and plasmids used in this study - 64 -

[0202] Table 12 Guide sequences and target region used in this study Target Sequence- 68 -TyA Gag ATGGAATCCCAACAATTATCTCAACATTCACCCATTTCTCATGG- 69 -GAACCGATTCAATTGAACAATAAGCACGACCTTCACCTTAGGC

[0203] References 1. I. O. C. M. Vroegrijk, et al., Pomegranate seed oil, a rich source of punicic acid, 5 prevents diet-induced obesity and insulin resistance in mice. Food Chem. Toxicol.49, 1426– 1430 (2011). 2. R. Holic, et al., Bioactivity and biotechnological production of punicic acid. Appl. Microbiol. Biotechnol.102, 3537–3549 (2018). 3. P. Aruna, D. Venkataramanamma, A. K. Singh, R. P. Singh, Health Benefits of 10 Punicic Acid: A Review: Health benefits of punicic acid... Compr. Rev. Food Sci. Food Saf. 15, 16–27 (2016). 4. K. F. Adekunle, A Review of Vegetable Oil-Based Polymers: Synthesis and Applications. Open J. Polym. Chem.05, 34–40 (2015). 5. H. S. Lee, et al., Tung Oil-Based Production of High 3-Hydroxyhexanoate-15 Containing Terpolymer Poly(3-Hydroxybutyrate-co-3-Hydroxyvalerate-co-3- Hydroxyhexanoate) Using Engineered Ralstonia eutropha. Polymers 13, 1084 (2021). - 70 -6. J. Wang, et al., Improving the Production of Punicic Acid in Baker’s Yeast by Engineering Genes in Acyl Channeling Processes and Adjusting Precursor Supply. J. Agric. Food Chem.69, 9616–9624 (2021). 7. V. Urbanikova, et al., Yarrowia lipolytica as a Platform for Punicic Acid 5 Production. Int. J. Mol. Sci.24, 8823 (2023). 8. A. D. Wickramarathna, et al., Heterologous expression of flax PHOSPHOLIPID:DIACYLGLYCEROL CHOLINEPHOSPHOTRANSFERASE (PDCT) increases polyunsaturated fatty acid content in yeast and Arabidopsis seeds. BMC Biotechnol. 15, 63 (2015). 10 9. K. Demski, S. Jeppson, S. Stymne, I. Lager, Phosphatidylcholine:diacylglycerol cholinephosphotransferase’s unique regulation of castor bean oil quality. Plant Physiol.189, 2001–2014 (2022). 10. E. Mietkiewska, et al., Combined transgenic expression of Punica granatum conjugase (FADX) and FAD2 desaturase in high linoleic acid Arabidopsis thaliana mutant 15 leads to increased accumulation of punicic acid. Planta 240, 575–583 (2014). 11. Y. Xu, E. Mietkiewska, S. Shah, R. J. Weselake, G. Chen, Punicic acid production in Brassica napus. Metab. Eng.62, 20–29 (2020). 12. E. B. Cahoon, Y. Li-Beisson, Plant unusual fatty acids: learning from the less common. Curr. Opin. Plant Biol.55, 66–73 (2020). 20 13. J. A. Napier, The Production of Unusual Fatty Acids in Transgenic Plants. Annu. Rev. Plant Biol.58, 295–319 (2007). 14. X. Pan, R. M. P. Siloto, A. D. Wickramarathna, E. Mietkiewska, R. J. Weselake, Identification of a Pair of Phospholipid:Diacylglycerol Acyltransferases from Developing Flax (Linum usitatissimum L.) Seed Catalyzing the Selective Production of Trilinolenin. J. Biol. 25 Chem.288, 24173–24188 (2013). 15. J. A. Browse, C. R. Slack, Catalase stimulates linoleate desaturase activity in microsomes from developing linseed cotyledons. FEBS Lett.131, 111–114 (1981). 16. Y. Xu, K. M. P. Caldo, L. Falarz, K. Jayawardhane, G. Chen, Kinetic improvement of an algal diacylglycerol acyltransferase 1 via fusion with an acyl-CoA binding 30 protein. Plant J.102, 856–871 (2020). - 71 -17. C. Andre, S. W. Kim, X.-H. Yu, J. Shanklin, Fusing catalase to an alkane- producing enzyme maintains enzymatic activity by converting the inhibitory byproduct H2O2to the cosubstrate O2. Proc. Natl. Acad. Sci.110, 3191–3196 (2013). 18. C. Jia, et al., Structural insights into the catalytic mechanism of aldehyde- 5 deformylating oxygenases. Protein Cell 6, 55–67 (2015). 19. Y. Kamisaka, N. Tomita, K. Kimura, K. Kainou, H. Uemura, DGA1 (diacylglycerol acyltransferase gene) overexpression and leucine biosynthesis significantly increase lipid accumulation in the snf2 disruptant of Saccharomyces cerevisiae. Biochem. J. 408, 61–68 (2007). 10 20. P. Chumnanpuen, J. Zhang, I. Nookaew, J. Nielsen, Integrated analysis of transcriptome and lipid profiling reveals the co-influences of inositol–choline and Snf1 in controlling lipid biosynthesis in yeast. Mol. Genet. Genomics 287, 541–554 (2012). 21. Y. Kamisaka, K. Kimura, H. Uemura, M. Shibakami, Activation of diacylglycerol acyltransferase expressed in Saccharomyces cerevisiae: overexpression of Dga1p lacking the 15 N-terminal region in the snf2 disruptant produces a significant increase in its enzyme activity. Appl. Microbiol. Biotechnol.88, 105–115 (2010). 22. J. Seip, R. Jackson, H. He, Q. Zhu, S.-P. Hong, Snf1 Is a Regulator of Lipid Accumulation in Yarrowia lipolytica. Appl. Environ. Microbiol.79, 7360–7370 (2013). 23. R. Usaite, et al., Reconstruction of the yeast Snf1 kinase regulatory network 20 reveals its role as a global energy regulator. Mol. Syst. Biol.5, 319 (2009). 24. F. V. Mayer, et al., ADP Regulates SNF1, the Saccharomyces cerevisiae Homolog of AMP-Activated Protein Kinase. Cell Metab.14, 707–714 (2011). 25. N. Khuu, S. Gidda, J. M. Shockey, J. M. Dyer, R. T. Mullen, The N termini of Brassica and tung omega-3 fatty acid desaturases mediate proteasome-dependent protein 25 degradation in plant cells. Plant Signal. Behav.6, 422–425 (2011). 26. J. B. O’Quin, et al., Temperature-sensitive Post-translational Regulation of Plant Omega-3 Fatty-acid Desaturases Is Mediated by the Endoplasmic Reticulum-associated Degradation Pathway. J. Biol. Chem.285, 21781–21796 (2010). 27. G.-Q. Tang, W. P. Novitzky, H. Carol Griffin, S. C. Huber, R. E. Dewey, Oleate 30 desaturase enzymes of soybean: evidence of regulation through differential stability and phosphorylation. Plant J. Cell Mol. Biol.44, 433–446 (2005). 28. J. Lian, R. Jin, H. Zhao, Construction of plasmids with tunable copy numbers in Saccharomyces cerevisiae and their applications in pathway optimization and multiplex - 72 -genome integration: Plasmid Copy Number Engineering. Biotechnol. Bioeng.113, 2462–2473 (2016). 29. A. S. Karim, K. A. Curran, H. S. Alper, Characterization of plasmid burden and copy number in Saccharomyces cerevisiae for optimization of metabolic engineering 5 applications. FEMS Yeast Res.13, 107–116 (2013). 30. M. J. Curcio, S. Lutz, P. Lesage, The Ty1 LTR-Retrotransposon of Budding Yeast, Saccharomyces cerevisiae. Microbiol. Spectr.3, 3.2.19 (2015). 31. M. Hanasaki, H. Masumoto, CRISPR / Transposon gene integration (CRITGI) can manage gene expression in a retrotransposon-dependent manner. Sci. Rep. 9, 15300 10 (2019). 32. K. Malc , L. E. Walls, L. Rios-Solis, Multiplex Genome Engineering Methods forYeast Cell Factory Development. Front. Bioeng. Biotechnol.8, 589468 (2020). 33. N. S. McCarty, A. E. Graham, L. Studená, R. Ledesma-Amaro, Multiplexed CRISPR technologies for gene editing and transcriptional regulation. Nat. Commun.11, 1281 15 (2020). 34. H. Alper, C. Fischer, E. Nevoigt, G. Stephanopoulos, Tuning genetic control through promoter engineering. Proc. Natl. Acad. Sci.102, 12678–12683 (2005). 35. Y. Lou, J. Schwender, J. Shanklin, FAD2 and FAD3 Desaturases Form Heterodimers That Facilitate Metabolic Channeling in Vivo. J. Biol. Chem.289, 17996–18007 20 (2014). 36. Q. He, et al., Oleaginicity of the yeast strain Saccharomyces cerevisiae D5A. Biotechnol. Biofuels 11, 258 (2018). 37. K. Inagaki, et al., Identification and expression of a rat fatty acid elongase involved in the biosynthesis of C18 fatty acids. Biosci. Biotechnol. Biochem. 66, 613–621 25 (2002). 38. H. Yazawa, Y. Kamisaka, K. Kimura, M. Yamaoka, H. Uemura, 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. Biotechnol.91, 1593–1600 (2011). 30 39. C. Lu, Z. Xin, Z. Ren, M. Miquel, J. Browse, An enzyme regulating triacylglycerol composition is encoded by the ROD1 gene of Arabidopsis. Proc. Natl. Acad. Sci.106, 18837– 18842 (2009). - 73 -40. I. Lager, et al., Plant Acyl-CoA:Lysophosphatidylcholine Acyltransferases (LPCATs) Have Different Specificities in Their Forward and Reverse Reactions. J. Biol. Chem. 288, 36902–36914 (2013). 41. X. Pan, et al., In Vivo and in Vitro Evidence for Biochemical Coupling of 5 Reactions Catalyzed by Lysophosphatidylcholine Acyltransferase and Diacylglycerol Acyltransferase. J. Biol. Chem.290, 18068–18078 (2015). 42. K. Wang, et al., Engineering Yarrowia lipolytica for Sustainable Production of the Pomegranate Seed Oil-Derived Punicic Acid. J. Agric. Food Chem. (2024) https: / doi.org / 10.1021 / acs.jafc.3c08718 (February 3, 2024). 10 43. E.-M. Hein, H. Hayen, Comparative Lipidomic Profiling of S. cerevisiae and Four Other Hemiascomycetous Yeasts. Metabolites 2, 254–267 (2012). 44. H. A. Boumann, B. de Kruijff, A. J. R. Heck, A. I. P. M. de Kroon, The selective utilization of substrates in vivo by the phosphatidylethanolamine and phosphatidylcholine biosynthetic enzymes Ept1p and Cpt1p in yeast. FEBS Lett.569, 173–177 (2004). 15 45. K. D’Souza, Y. J. Kim, T. Balla, R. M. Epand, Distinct Properties of the Two Isoforms of CDP-Diacylglycerol Synthase. Biochemistry 53, 7358–7367 (2014). 46. X. Liu, H. Li, Y. Chen, Y. Cao, Method for Screening of Bacterial Strains Biosynthesizing Specific Conjugated Linoleic Acid Isomers. J. Agric. Food Chem.60, 9705– 9710 (2012). 20 47. S. Ülker, et al., New lipase assay using Pomegranate oil coating in microtiter plates. Biochimie 120, 110–118 (2016). 48. C. M. Henderson, D. E. Block, Examining the Role of Membrane Lipid Composition in Determining the Ethanol Tolerance of Saccharomyces cerevisiae. Appl. Environ. Microbiol.80, 2966–2972 (2014). 25 49. S. Kajiwara, et al., Polyunsaturated fatty acid biosynthesis in Saccharomyces cerevisiae: expression of ethanol tolerance and the FAD2 gene from Arabidopsis thaliana. Appl. Environ. Microbiol.62, 4309–4313 (1996). 50. A. Cipak, et al., Saccharomyces cerevisiae strain expressing a plant fatty acid desaturase produces polyunsaturated fatty acids and is susceptible to oxidative stress induced 30 by lipid peroxidation. Free Radic. Biol. Med.40, 897–906 (2006). - 74 -51. X. Luo, et al., The pomegranate ( Punica granatum L.) draft genome dissects genetic divergence between soft- and hard-seeded cultivars. Plant Biotechnol. J.18, 955–968 (2020). 52. R. D. Gietz, R. H. Schiestl, High-efficiency yeast transformation using the 5 LiAc / SS carrier DNA / PEG method. Nat. Protoc.2, 31–34 (2007). 53. Z. Bao, et al., Homology-Integrated CRISPR–Cas (HI-CRISPR) System for One-Step Multigene Disruption in Saccharomyces cerevisiae. ACS Synth. Biol. 4, 585–594 (2015). 54. J. Salimon, B. M. Abdullah, N. Salih, Hydrolysis optimization and 10 characterization study of preparing fatty acids from Jatropha curcasseed oil. Chem. Cent. J.5, 67 (2011). 55. F. E. Luddy, R. A. Barford, S. F. Herb, P. Magidman, R. W. Riemenschneider, Pancreatic lipase hydrolysis of triglycerides by a semimicro technique. J. Am. Oil Chem. Soc. 41, 693–696 (1964). 15 56. A. F. Vikbjerg, H. Mu, X. Xu, Elucidation of acyl migration during lipase- catalyzed production of structured phospholipids. J. Am. Oil Chem. Soc.83, 609–614 (2006). 1. C. Baker Brachmann, et al., Designer deletion strains derived from Saccharomyces cerevisiae S288C: A useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 14, 115–132 (1998). 20 The embodiments described herein are intended to be examples only. Alterations, modifications and variations can be effected to the particular embodiments by those of skill in the art. The scope of the claims should not be limited by the particular embodiments set forth herein, but should be construed in a manner consistent with the specification as a whole.

[0204] The embodiments described herein are intended to be examples only. 25 Alterations, modifications and variations can be effected to the particular embodiments by those of skill in the art. The scope of the claims should not be limited by the particular embodiments set forth herein, but should be construed in a manner consistent with the specification as a whole.

[0205] All publications, patents and patent applications mentioned in this Specification 30 are indicative of the level of skill those skilled in the art to which this invention pertains and are herein incorporated by reference to the same extent as if each individual publication patent, or patent application was specifically and individually indicated to be incorporated by reference. - 75 -

[0206] The invention being thus described, it will be obvious that the same may be varied in many ways. Such variations are not to be regarded as a departure from the spirit and scope of the invention, and all such modification as would be obvious to one skilled in the art are intended to be included within the scope of the following claims. 5 - 76 -

Claims

WHAT IS CLAIMED IS:

1. An isolated recombinant Rhodosporidium toruloides (R. toruloides) cell, comprising a polynucleotide encoding Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX). 2.The isolated recombinant R. toruloides cell of claim 1, further comprising a polynucleotide encoding Punica granatum delta-12 acyl lipid desaturase (PgFAD2), diacylglycerol acyltransferase 2 (PgDGAT2), or phosphatidylcholine: diacylglycerol cholinephosphotransferase (PgPDCT).

3. The isolated recombinant R. toruloides cell of claim 1 or 2, wherein said polynucleotide encoding PgFADX is a codon-optimized polynucleotide encoding PgFADX.

4. Use of the isolated recombinant R. toruloides cell of any one of claims 1 – 3 for producing punicic acid (PuA).

5. A method of producing recombinant punicic acid (PuA) comprising, providing a recombinant Rhodosporidium toruloides (R. toruloides) cell of any one of claims 1 to 3, and culturing said recombinant R. toruloides cell in a culture medium under culture conditions that allow for the production of PuA.

6. The method of claim 5, wherein the culture conditions comprise a temperature of about 30°C for about 1-10 days. 7.The method of claim 5 or 6, wherein the culture medium is a nitrogen-limited medium.

8. The method of any one of claims 5-7, wherein the culture medium comprises wood hydrolysate.

9. The method of any one of claims 5 to 8, wherein about 5.8% to about 11.98% of total fatty acids is PuA. - 77 -10. A recombinant Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX) obtained by the method of any one of claims 5 to 9.

11. An isolated recombinant Saccharomyces cerevisiae cell comprising a polynucleotide encoding a fusion protein comprising: a substrate-binding / cosubstrate-providing / inhibitor- removing protein, a linker, and a Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX), wherein the N-terminal amino acid of the linker is attached the C- terminal amino acid of the substrate-binding / cosubstrate-providing / inhibitor-removing protein, and the C-terminal amino acid of the linker is attached to the N terminal amino acid of the PgFADX.

12. The isolated recombinant S. cerevisiae cell of claim 1, wherein the substrate binding protein is phosphatidylcholine-binding proteins (SCP2), oxygen carrier-proteins (Vhb), soluble domain of electron transporter (CB5SD) or catalase (CAT).

13. The isolated recombinant S. cerevisiae cell of claim 11 or 12, wherein the linker comprises or consists of the amino acid sequence ASGAGGSEGGGSEGGTSGAT (SEQ ID NO: 20) 14. The isolated recombinant S. cerevisiae cell of any one of claims 11 to 13, comprisinga snf2 mutant strain or snf2 snf1 double mutant strain.

15. The isolated recombinant S. cerevisiae cell of any one of claims 11 to 14, wherein therecombinant Saccharomyces cerevisiae yeast strain is a snf2 snf1 double knockout strain.

16. An isolated recombinant Saccharomyces cerevisiae yeast strain comprising apolynucleotide encoding AtCB5SD-PgFADX, wherein the yeast strain is a BY4741 snf2 snf1double knockout mutant strain.

17. The isolated recombinant S. cerevisiae cell of claim 16, further comprises a polynucleotide encoding PDCT, a polynucleotide encoding LPCAT and a polynucleotide encoding DGAT2. - 78 -18. The isolated recombinant S. cerevisiae cell of claim 17, further comprising a polynucleotide encoding FAD2 and ELO2.

19. A method of producing recombinant punicic acid (PuA) comprising, providing a recombinant Saccharomyces cerevisae cell of any one of claims 11 – 18, and culturing said recombinant S. cerevisae cell in a culture medium under culture conditions that allow for the production of PuA 20. The method of claim 19, wherein the culture medium comprises about 6-15 % glucose.

21. The method of claim 19 or 20, wherein the isolated recombinant S. cerevisiae cell is cultured in the presence of 0-0.03% linoleic acid.

22. Use of the isolated recombinant S. cerevisiae cell of any one of claims 11 – 18 for producing punicic acid.

23. A recombinant Punica granatum bifunctional acyl lipid desaturase and conjugase (PgFADX) obtained by the method of any one of claims 19 to 21.

24. Use of a Ty retrotransposon-targeted random gene shuffling approach for plant-derived punicic acid production in yeast.

25. Use of a Ty retrotransposon-targeted random gene shuffling approach for unusual fatty acid synthesis in yeast cell.

26. The use of claim 22, or the method of any one of claims 19 – 21, wherein the PuA content producted is greater than about 20% TFA. - 79 -

Images