In our project, we plan to synthesize the PUFA synthase gene cluster (around 20kb in total) originally from Schizochytrium sp. (ATCC 20888) and integrate it into the genome of Yarrowia lipolytica strain Po1f, which can achieve a relatively high lipid content, growth rate, and growth density, boosting the production^[1]. The DHA-synthesis machine — PUFA synthase consists of three subunits encoded by 3 open reading frames (ORFs), and requires a phosphopantetheinyl transferase (PPT) gene for its activation. Different engineering strategies including stage control, stopping beta-oxidation, replenishing NADPH and DHA efflux are explored. Please click on the below graph to learn more about each part.

Please click on the orange buttons below to learn more about each part

Graph 1. A schematic graph showing the inputs and outputs of our chosen DHA production pathway with the “machine” PUFA synthase. DPA represents Docosapentaenoic acid (22, n-5) and TAG stands for triacylglycerol.

Yarrowia lipolytica is normally unable to synthesize DHA, so a pathway for DHA production must be introduced. In nature, there are 2 major pathways for synthesizing DHA:

Desaturase/Elongase Pathway

Also known as DES/ELO pathway or the aerobic pathway, as it requires oxygen during the process. This pathway utilizes the existing fatty acids (C16, C18) as main substrates, and each elongation and desaturation steps towards DHA are catalyzed by separate enzymes^[2]. Team NNU iGEM 2022 successfully introduced the pathway into Yarrowia lipolytica. However, the actual conversion rate of each step through this pathway poses a limitation to the final yield of DHA.

PKS Pathway

The pathway, also known as PUFA synthase pathway or the de novo synthesis, involves a single enzyme complex from the polyketide synthase (PKS) family — PUFA synthase. The synthase takes up metabolite acetyl-CoA as the starter unit, and takes up malonyl-CoA as the building blocks. It works iteratively to lengthen and reduce the fatty acid chain without releasing the acyl chain from the complex, until the synthesis of DHA finishes after 10 cycles. PUFA synthase enzymes catalyzing the synthesis of DHA are found both in eukaryotic microalgae and prokaryotic myxobacteria^[3][4], while the exact mechanism of the pathway is still not well-known and under intensive study.

By comparing data from several existing research, we choose to introduce a eukaryotic polyketide synthase pathway originally found in microalgae Schizochytrium sp., which has a relatively high output in comparison to its prokaryotic counterpart, and requires less resources (i.e., NADPH, oxygen) than the aerobic DHA synthesis pathway^[4][5].

Graph 2. Predicted structure of PUFA synthase subunits with Alphafold 3. From left to right: Subunit A, B and C.

Graph 3. A graph showing the chemical reactions and role of different domains. This one is from paper, to be replaced.

The PUFA synthase gene cluster involves 3 ORFs corresponding to different protein subunits, which will assemble together to become one large enzyme in the cytoplasm^[6]. A phosphopantetheinyl transferase (PPTase) from the same organism will also be introduced to activate the enzyme^[7]. Each subunit contains multiple functional domains that play different roles during each iterative cycle to synthesize DHA:

Acyl-carrier protein (ACP)

AACP domains “hold” the acyl chain throughout the entire reaction, “rotating” it to interact with other different domains for certain reactions without releasing fatty acid byproducts from the enzyme cluster until reaching the final step, which characterizes the de novo DHA-synthesis pathway with a high product specificity. Increased copy number of ACP leads to an increased binding probability of the domain with the substrate malonyl-CoA, which can potentially accelerate DHA synthesis[4][8]. Our sequence contains 9 tandem repeats of the ACP domain, enabling an efficient synthesis process.

Malonyl-CoA:ACP transacylase (MAT)

AMAT domain transfers the malonyl group from malonyl-CoA substrate to ACP domains to enter the reaction. It is crucial to the initiation of the elongation steps. In contrast, the acyltransferase domain on subunit B helps release the final product from the ACP domain, corresponding to the termination step of the whole reaction.

Ketoacyl-ACP synthase (KS)

There are 2 Ketoacyl-ACP synthase domains identified on the PUFA synthase cluster. KS domain is mainly in charge of the elongation of acyl chain by 2 carbon at a time, interacting with the acyl-carrier protein (ACP) domains carrying a malonyl group. One CO2 molecule is released during each elongation step. Mutation of only 3 amino acids in this domain can cause the main final product to switch from DHA to EPA[4].

Keto-reductase (KR)

After elongation, the KR domain reduces one ketone group to hydroxyl group by consuming 1 NADPH from other metabolic pathways. This is the first step towards desaturation, which means introducing a double bond. Unlike other domains, the KR domain only exists on ORF_A but not on other subunits.

Dehydratase (DH)

There are also 2 DH domains identified on the PUFA synthase cluster. DH domain is in charge of the second step of desaturation, by catalyzing the formation of double bond while releasing one water molecule. Different DH domains can affect the position of double bond by isomerization and following steps, leading to a different final product. DHC subsequently catalyzes the 2, 2 or 2, 3-isomerization of the double bond, preparing the acyl-chain for a next elongation cycle or release[4].

Enoyl-reductase (ER)

According to the previous description, a double bond is formed after each elongation step, leading to 10 double bonds formed after 10 elongation cycles, which is impossible in chemical structure. For DHA synthesis, 6 double bonds are the right amount, so at certain steps the double bond formed must be further reduced, again consuming one NADPH, catalyzed by the ER domain.

Acyl-transferase (AT)

AT domain is in charge of the release of the final product as free fatty acid from the enzyme complex, in our case mainly DHA, with some omega-6 DPA as byproducts[9]. The release step is considered to be different in prokaryotic and eukaryotic PUFA synthases, where in the prokaryotic system, the product will likely be released as phospholipid by an AGPAT domain and can be further converted into free fatty acid or triacylglycerol (TAG)[3].

To introduce the enzyme into Y. lipolytica, we first put each of the codon-optimized ORFs under control of a common constitutive promoter TEF1. It is reported that the RBS region is contained in the promoter sequence^[10]. A native intron sequence is added to the 3’ terminus of the promoter, which is reported to increase the promoter strength by 17 folds^[11]. Each subunit is His-tagged in order to detect their expression. After each coding sequence, a common LIP2 terminator and a spacer is added.

Graph 4. Design of the expression cassette for separate PUFA synthase subunits

Genomic Integration

There exists several means that have been applied to achieve exogenous gene expression in Y. lipolytica. As an unconventional yeast species, Y. lipolytica prefers non-homologous end-joining (NHEJ) as the main mechanism for genomic integration instead of homologous recombination (HR), which generally occurs at a very low frequency except for in some strains engineered to include a genomic docking platform for HR^[12]. Both NHEJ and HR require linearized fragments to be transformed into Y. lipolytica. For HR, two homologous arms flanking the expression cassette are needed to enable the recombination, while for NHEJ no other parts are involved besides the linear expression cassette, whatever the two ends are. The fragments are automatically integrated into the genome, and the rest, unintegrated ones will soon be degraded. So we chose NHEJ as the major mechanism to randomly integrate our designed expression cassette into the strain, enabling a relatively simple and efficient integration, as pointed out in our interview by Dr. DU Fei from Nanjing Normal University.

NHEJ utilizes the random double-strand break (DBS) repair occurring in Y. lipolytica facilitated by several proteins (e.g., KU70, KU80). It also can lead to some risks, for example, due to its randomness, the introduced gene will sometimes be accidentally integrated into the middle of an essential coding sequence and affect normal cellular functions. But as the genome itself is huge, and the chance of the risk happening is very low, we believe the risk was covered by the advantages brought with NHEJ.

Other methods of expression include transforming plasmids or shuttle vectors into Y. lipolytica, but most such vectors are not replicative nor inheritable, so this mean is mainly used for temporal or transient expressions of proteins in Y. lipolytica, for example, expressing Cre recombinase to release selective markers with the Cre-loxP system strategy, which will be discussed later^[12].

As our genes to be integrated are huge and it is hard to clone and transform such a large linearized fragments with all ORFs at once, in our interview with Prof. Yong Lai from HKUST we also discussed the possibility of using Yeast Artificial Chromosomes (YACs) to integrate the large enzyme with several expression cassettes at once into Y. lipolytica. However, existing YAC uses are mainly restricted to conventional yeast S. cerevisiae, and no evidence was found that it could be applied to Y. lipolytica. Therefore, at our current stage we will stick to NHEJ, which has proved by numerous researches to be a reliable method for Y. lipolytica genomic integration.

All of our integration cassettes include a LEU2 selection marker (BBa_K5159016) that can make the transformed cells selectively grow on auxotrophic plates without leucine. In order to repeatedly integrate multiple genes into Y. lipolytica, we will release the LEU2 marker expression cassette with the Cre-loxP system, as the LEU2 expression cassette is already flanked with 2 loxP sites by its design. When we transform the yeast again with a pSL16-CRE-HPH plasmid temporarily expressing the Cre recombinase, kindly suggested and provided by Prof. Lee Joon Foo from NUS, the loxP sites are located and the LEU2 marker will be released, enabling the next rounds of integration using the same selection marker^[13].

Graph 5: An example of the linear construct to be transformed into Yarrowia lipolytica

As presented in our final design, the expression of PUFA synthase will be controlled by promoters induced with nitrogen starvation conditions, which promotes DHA accumulation after the cells reach a high biomass and enter the stationary phase, or so-called production phase. This will release the burden of cells expressing such a large enzyme and inefficiently synthesizing DHA while it is growing, which may negatively affect the normal growth of the cells^[14]. For detailed design considerations and the choice of promoters, please refer to Engineering Cycle 3.1.

We plan to combine our central pathway with various optimization strategies to increase the ultimate yield, as the same time balancing growth and production, which include:

Strategy 1:

Knocking out genes KU70 and PEX10 by homologous recombination to efficiently prevent PUFA degradation:

PEX10 encodes a key protein for peroxisomal beta-oxidation in Y. lipolytica^[14]. Inspired by https://2022.igem.wiki/nnu-china/ , to prevent the degradation of synthesized DHA through beta-oxidation, we plan to knock out the PEX10 gene in Y. lipolytica by homologous recombination (HR). KU70, a key gene enabling non-homologous end-joining in Y. lipolytica, may first need to be knocked out to increase the chance of HR^[15], which is originally low in Y. lipolytica.

The knockout is carried out by replacing part of the PEX10 gene with the LEU2 gene construct, which is used for nutrient auxotrophic selection. The construct is flanked with 2 loxP sites, so that after the knockout, the LEU2 marker gene can be released using the Cre-loxP system by introducing another plasmid containing the Cre recombinase and a hygromycin resistance gene for selection, as mentioned in the previous paragraphs.

For the design rationale and more details, please refer to Engineering Cycle 1.

Strategy 2

Replenishing reactant NADPH supply through metabolic engineering:

At least 14 NADPH molecules are consumed to provide the redox power for generating a DHA molecule. Various literatures have reported NADPH being a limiting factor in lipid production^[16][17]. We plan to express the GapC gene from Clostridium acetobutylicum (BBa_K5159008) encoding a NADP+-dependent G3P dehydrogenase in Y. lipolytica, which converts NADP+ to NADPH, increasing its availability during DHA synthesis ^[18].

For the design rationale and more details, please refer to Engineering Cycle 3.2.

Strategy 3

Controlling expression level and production stages via promoter designs:

In order to achieve a higher DHA yield in Y. lipolytica, we are also considering dividing the process into growth and production stages ^[13]. During the growth stage, we would maintain normal cell metabolism, while during the production stage, PUFA synthase expression is activated. The activation can be achieved via some inducible promoters in Y. lipolytica, which we will first test with hrGFP and measure the relative fluorescence.

For the design rationale and more details, please refer to Engineering Cycle 3.1.

Future Cycle 1: Inhibit interfering fatty acid synthesis

During our communication with other teams, one of the concerns regarding the current system is that many strategies do not exclusively promote DHA accumulation but all fatty acids (in majority C16 and C18 fatty acids) synthesized by native fatty acid synthase (FAS) enzymes, which might lead to a competition for substrates and result in a lack of DHA purity. To deal with this potential drawback, we came up with the idea to redirect the substrates to DHA by suppressing the native fatty acid synthesis in Yarrowia lipolytica. Before approaching genetic circuit design, we again predicted the effectiveness of the strategy through our Dry Lab model, and showed a favorable result leading to the maximized DHA percentage in total fatty acids.

This repression can be achieved by the CRISPRi system, where several gRNAs are designed to guide the deactivated Cas9 protein to the matching sequence region and block the gene transcription. This strategy is already successfully employed to silence the KU70 gene in Yarrowia lipolytica to increase chance of homologous recombination ^[19].

Graph: Inhibition of the native fatty acid synthase in Yarrowia lipolytica via CRISPRi during production stage

The gRNA sequence is a 20-30 bp short sequence expressed without the ribosome binding site sequence, designed to target the core promoter region such as the TATA box and Transcription Start Sites (TSS) of Yarrowia lipolytica FAS2 gene, preventing it from transcription initiation. Multiple short gRNA sequences with different targets are introduced to enhance the repression efficiency.

Compared to other approaches, when placing the gene of either gRNA or dCas9 under an inducible promoter, this method offered a inheritable but controlled gene silencing without permanently destructing the gene itself, which might affect the normal function of the cells.

Future Cycle 2: DHA efflux

In the project, we not only consider metabolic strategies to increase the yield of DHA. Currently DHA production requires harvesting and lysis of the cell, which is unsustainable. However, DHA is a large molecule that would be difficult to transport out through the cell membrane efficiently. To address this issue, we will also explore possible strategies for DHA secretion, aiming to achieve a continuous production.

DHA efflux is considered to be a very potential strategy to increase the yield of microbial DHA production, as it can bypass the upper limit of cellular lipid accumulation and reach a continuous production without lysing the cells. To gain more insights into this strategy, we consulted Professor Rodrigo Ledesma-Amaro from Imperial College London, who first published the paper proposing this strategy, where all lipids are solely retained in form of or converted to free fatty acids by downregulating all other conversion pathways^[20]. An efflux from the cell possibly through simple diffusion of fatty acid is thus achieved, as the cell membrane is quite permeable for fatty acids.

We also identified a few DHA transporters that can transport DHA-related products across the cell membrane. One of them is the Major Facilitator Superfamily Domain containing 2a (Mfsd2a) DHA-LPC transporter, which is an import transporter driven by sodium ion concentration and exhibits a flippase mechanism found in human blood brain barrier (BBB)^[21]. According to our interview with Professor David Silver from Duke-NUS medical school, the efficiency of the transporter in terms of flux is considered as high enough for industrial purposes, however, it is only for importing DHA-PC into the cell and is not reversible. Professor Silver further pointed out that the outward DHA transporters identified so far have a totally different mechanism, which is mainly classified as ATP-binding cassette (ABC) transporters that are less known for their role in PUFA-specific transportation.

Efflux of DHA as free fatty acid will be not good for the downstream processing due to the instability and the fact that DHA is prone to oxidation, indicating that it might be better just to store DHA in forms of TAG inside the cells. For more details about the transporter efficiency, please refer to the calculation on the Dry lab page.

Also for the purification process, after consulting several researchers including industrial stakeholders (e.g., founders of AlGreen), we found that DHA efflux also might not facilitate the purification, unless a separate layer of lipids the media is observed after efflux, indicating its success.

Future Cycle 3: Biocontainment strategy

For a biomanufacturing project aiming for real world application in the long term, it is crucial to include biocontainment strategies to prevent the accidental release of engineered organisms into the environment and cause potential threats to the natural ecosystem. We identified several feasible means to ensure that the engineered chassis only survive within the designated facilities, and corresponding engineering strategies are proposed as follows:

Current Strategy: Auxotrophy:

Our strain Po1f is a leucine and uracil double auxotrophic strain, so it is unable to grow in environments lacking either constant leucine or uracil supply, preventing it from harming in natural environments when accidentally released. Though the LEU2 gene is used for our selection marker, it will be immediately released by the Cre-loxP system after the engineering, making it again an auxotrophic strain. This is also one of the most mature biocontainment methods currently applied. However, there are still concerns that a tiny proportion strain may survive even with the auxotroph, for example due to unsuccessful marker release, or happened to be supplied with the essential nutrients after being released. Therefore, a more reliable biocontainment strategy is still valued. To tackle this problem, we proposed two future strategies to ensure a safe biocontainment of our strain during the industrial application.

Potential Strategy 1: Toxin-antitoxin system

This first strategy is achieved via a two-part toxin-antitoxin system. Existing examples that can be applied to yeast include the RelE/RelB, Kid/Kis system and epsilon-zeta genes ^[23] . Under normal culture conditions, both toxin and antitoxin expressions are on, and the toxin is repressed. However, when released into the environment, the expression of antitoxin gene is inhibited, leading to the accumulating toxicity and thus cell death. In our chassis, this can be implemented by placing the antitoxin gene under a special compound-inducible promoter (e.g., Yarrowia lipolytica pETK1), and that compound is exclusively added into the culture media, while not found in nature. Based on this idea, a more complicated design can be built to optimize the use of inducers or an inducing condition.

Potential Strategy 2: Direct apoptosis

In comparison, the second strategy is more straightforward, which is to induce the expression of multiple or even a single gene that leads to the cell death. Nuclease A (NucA) from Serratia marcescens is identified as a good candidate for effectively killing the cell^[23]. However the gene must be expressed under a promoter induced by environmental factors, for example, low glucose or high salt level, again highlighting the importance of understanding inducible promoters, as it is now demanding to find such a suitable promoter in Yarrowia lipolytica.

[1] Young-Kyoung Park, Rodrigo Ledesma-Amaro (2023). What makes Yarrowia lipolytica well suited for industry? Trends in Biotechnology. Volume 41, Issue 2, 242-254, https://doi.org/10.1016/j.tibtech.2022.07.006
[2] Ishibashi, Y., Goda, H., Hamaguchi, R. et al (2021). PUFA synthase-independent DHA synthesis pathway in Parietichytrium sp. and its modification to produce EPA and n-3 DPA. Commun Biol 4, 1378. https://doi.org/10.1038/s42003-021-02857-w
[3] Gemperlein, K., Dietrich, D., Kohlstedt, M. et al (2019). Polyunsaturated fatty acid production by Yarrowia lipolytica employing designed myxobacterial PUFA synthases. Nat Commun 10, 4055. https://doi.org/10.1038/s41467-019-12025-8
[4] Guo P, Dong L, Wang F, Chen L and Zhang W (2022), Deciphering and engineering the polyunsaturated fatty acid synthase pathway from eukaryotic microorganisms. Front. Bioeng. Biotechnol. 10:1052785. https://doi.org/10.3389/fbioe.2022.1052785
[5] Lizhen Cao, Mingxue Yin, Tian-Qiong Shi, Lu Lin, Rodrigo Ledesma-Amaro, Xiao-Jun Ji (2022). Engineering Yarrowia lipolytica to produce nutritional fatty acids: Current status and future perspectives. Synthetic and Systems Biotechnology, Volume 7, Issue 4, 1024-1033. https://doi.org/10.1016/j.synbio.2022.06.002
[6] Hauvermale, A., Kuner, J., Rosenzweig, B. et al (2006). Fatty acid production in Schizochytrium sp.: Involvement of a polyunsaturated fatty acid synthase and a type I fatty acid synthase. Lipids 41, 739–747.
[7] Wang, S., Lan, C., Wang, Z. et al (2020). PUFA-synthase-specific PPTase enhanced the polyunsaturated fatty acid biosynthesis via the polyketide synthase pathway in Aurantiochytrium. Biotechnol Biofuels 13, 152. https://doi.org/10.1186/s13068-020-01793-x
[8] Hayashi, S., Satoh, Y., Ujihara, T. et al (2016). Enhanced production of polyunsaturated fatty acids by enzyme engineering of tandem acyl carrier proteins. Sci Rep 6, 35441. https://doi.org/10.1038/srep35441
[9] J.G. Metz et al (2009). Biochemical characterization of polyunsaturated fatty acid synthesis in Schizochytrium: Release of the products as free fatty acids. Plant Physiology and Biochemistry, 47, 472–478. https://doi.org/10.1016/j.plaphy.2009.02.002
[10] Benjamin Ouellet, A.M. Abdel-Mawgoud (2023). Strong expression of Cas9 under a new 3′-truncated TEF1α promoter enhances genome editing in Yarrowia lipolytica. Current Research in Biotechnology, Volume 6, 100147. https://doi.org/10.1016/j.crbiot.2023.100147
[11] Mitchell Tai, Gregory Stephanopoulos (2013). Engineering the push and pull of lipid biosynthesis in oleaginous yeast Yarrowia lipolytica for biofuel production. Metabolic Engineering, Volume 15, 1-9. https://doi.org/10.1016/j.ymben.2012.08.007
[12] M. Larroude, T. Rossignol, J.-M. Nicaud, R. Ledesma-Amaro (2018). Synthetic biology tools for engineering Yarrowia lipolytica. Biotechnology Advances, Volume 36, Issue 8, 2150-2164. https://doi.org/10.1016/j.biotechadv.2018.10.004
[13] Yu, A., Pratomo, N., Ng, T., Ling, H., Cho, H., Leong, S. S. J., Chang, M. W (2016). Genetic Engineering of an Unconventional Yeast for Renewable Biofuel and Biochemical Production. J. Vis. Exp. (115), e54371. https://doi.org/10.3791/54371 .
[14] Seong Gyeong Kim, Myung Hyun Noh, Hyun Gyu Lim, Sungho Jang, Sungyeon Jang, Mattheos A G Koffas, Gyoo Yeol Jung (2018). Molecular parts and genetic circuits for metabolic engineering of microorganisms. FEMS Microbiology Letters, Volume 365, Issue 17. https://doi.org/10.1093/femsle/fny187
[15] Xue, Z., Sharpe, P., Hong, SP. et al. (2013). Production of omega-3 eicosapentaenoic acid by metabolic engineering of Yarrowia lipolytica. Nat. Biotechnol., 31(8), 734–740. https://doi.org/10.1038/nbt.2622
[16] Jang I-S, Yu B.J, Jang J.Y, Jegal J, Lee J.Y. (2018). Improving the efficiency of homologous recombination by chemical and biological approaches in Yarrowia lipolytica. PLoS ONE 13(3): e0194954. https://doi.org/10.1371/journal.pone.0194954
[17] Qin J, Kurt E, L Bassi T, Sa L and Xie D (2023). Biotechnological production of omega-3 fatty acids: current status and future perspectives. Front. Microbiol. 14:1280296. https://doi.org/10.3389/fmicb.2023.1280296
[18] Qiao, K., Wasylenko, T., Zhou, K. et al (2017). Lipid production in Yarrowia lipolytica is maximized by engineering cytosolic redox metabolism. Nat Biotechnol 35, 173–177. https://doi.org/10.1038/nbt.3763
[19] Liu, B.; Sun, X.; Liu, Y.; Yang, M.; Wang, L.; Li, Y.; Wang, J. (2022). Increased NADPH Supply Enhances Glycolysis Metabolic Flux and L-methionine Production in Corynebacterium glutamicum. Foods 2022, 11, 1031. https://doi.org/10.3390/foods11071031
[20] Zhang, J.l., Peng, Y.Z., Liu, D. et al (2018). Gene repression via multiplex gRNA strategy in Y. lipolytica. Microb Cell Fact 17, 62. https://doi.org/10.1186/s12934-018-0909-8
[21] Rodrigo Ledesma-Amaro, Remi Dulermo, Xochitl Niehus, Jean-Marc Nicaud (2016). Combining metabolic engineering and process optimization to improve production and secretion of fatty acids. Metabolic Engineering, Volume 38, 38-46. https://doi.org/10.1016/j.ymben.2016.06.004
[22] Chua, G. L., Tan, B. C., Loke, R. Y., He, M., Chin, C. F., Wong, B. H., ... & Silver, D. L. (2023). Mfsd2a utilizes a flippase mechanism to mediate omega-3 fatty acid lysolipid transport. Proceedings of the National Academy of Sciences, 120(10), e2215290120. https://doi.org/10.1073/pnas.2215290120
[23] Pavão, G.; Sfalcin, I.; Bonatto, D. (2023). Biocontainment Techniques and Applications for Yeast Biotechnology. Fermentation 2023, 9, 341.