Two Co-Expression Strategies for Eicosapentaenoic Acid Production in Mortierella alpina

State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi, Jiangsu 214122, China School of Food Science and Technology, Jiangnan University, Wuxi, Jiangsu 214122, China National Engineering Research Center for Functional Food, Jiangnan University, Wuxi, Jiangsu 214122, China Wuxi Translational Medicine Research Center and Jiangsu Translational Medicine Research Institute Wuxi Branch, Wuxi, Jiangsu 214122, China Beijing Innovation Centre of Food Nutrition and Human Health, Beijing Technology and Business University (BTBU), Beijing 100048, China


Introduction
Eicosapentaenoic Acid (EPA, C20:5), an omega-3 long-chain Polyunsaturated Fatty Acid (LC-PUFA), has attracted considerable attention due to its beneficial effects on human health, including the prevention and treatment of diabetes, immune disorders, cardiovascular disease and cancer (Hirahashi et al., 2014;Maehre et al., 2015;Xue et al., 2013). Within the food chain, EPA is mainly accumulated in marine fish. However, the market demand for this food source exceeds the supply because of human population growth, environmental pollution and limitations associated with the harvest season and location (Ji et al., 2015;Tang et al., 2018;Xie et al., 2017).
Various microorganisms, including Mortierella alpina and Yarrowia lipolytica, have been investigated as potential hosts for the sustainable commercial production of EPA (Ando et al., 2009;Xue et al., 2013). M. alpina, an oleaginous filamentous fungus currently used to produce commercial Arachidonic Acid (AA, C20:4), can generate EPA from AA via an omega-3 desaturase (EC 1.14.19.-) (Ge et al., 2017;Ji et al., 2014;Okuda et al., 2015). Our latest research identified a new omega-3 desaturase, oPpFADS17, from Phytophthora parasitica. This enzyme converts AA to EPA at a conversion rate of 70%. When expressed in a recombinant strain of M. alpina, oPpFADS17 increased the proportion of EPA among the Total Fatty Acids (TFA) content to 31.5% while the control strain without the recombinant enzyme didn't accumulate EPA . Although this recombinant strain represents a breakthrough in EPA production, further enhancement of the TFA content could potentially increase the EPA yield further. Another previous study demonstrated that the Pentose Phosphate Pathway (PPP), especially Glucose-6-Phosphate Dehydrogenase (G6PD2, EC1.1.1.49), plays a major role in the accumulation of total lipids (Hao et al., 2016). Accordingly, the co-expression of oPpFADS17 and G6PD2 may represent a good strategy toward improving EPA production in M. alpina.
The double expression cassette strategy can be used to co-express multiple copies of a gene or multiple genes with independent promoters (Bouabe et al., 2008). Another multi-gene construction strategy, the selfsplicing 2A peptide, is one of the most widely used tools for the construction of polygenic vector and has also attracted extensive attention (Park et al., 2009;Szymczak et al., 2004). The 2A peptide strategy exhibits obvious advantages with respect to protein activity and the expression of multiple downstream genes. Accordingly, it appears to be an ideal gene expression strategy (Hu et al., 2009;Trichas et al., 2008). Previous studies revealed that 2A peptide could be applied successfully in a broad range of organisms, including fungi (de Felipe et al., 2003), plants (Halpin et al., 1999;Park et al., 2009), mammalian cells (Hu et al., 2009;Szymczak et al., 2004) and transgenic animals (Trichas et al., 2008). In this study, we cooverexpressed oPpFADS17 and G6PD2 using both the double expression cassette and 2A peptide strategies with the aim of improving EPA production in M. alpina.

Growth Conditions
Escherichia coli DH5α was cultivated at 37°C on LB agar medium containing 10 g/L NaCl, 10 g/L tryptone and 5 g/L yeast extract. A. tumefaciens AGL-1 was grown at 28°C in YEP agar medium. M. alpina strains were cultured at 28°C in broth medium at 200 rpm for 168 h (Wang et al., 2018).

Construction of the Transfer DNA Binary Vector Using the Double Expression Cassette Strategy
The pBIG2-ura5s-G6PD2 plasmid was extracted (Hao et al., 2016) and the entire G6PD2 expression unit (including the promoter and terminator) was obtained using Polymerase Chain Reaction (PCR) with G6PD2-F1/R1 primer sequences. The underlined regions of the sequences indicate the XbaI restriction site (Table 1). Subsequently, the purified G6PD2 expression unit fragment and the extracted plasmid expression vector pBIG2-ura5s-oPpFADS17  were digested using the restriction endonuclease XbaI. Recombinant ligase (Exnase) was then used to construct the binary expression vector pBIG2-ura5s-oPpFADS17-G6PD2 using one step cloning kit (Takara).

A. Tumefaciens-Mediated Transformation (ATMT)
ATMT was performed as described previously (Hao et al., 2014). Briefly, M. alpina uracil-auxotrophic strain spores were harvested and diluted with fresh liquid GY medium to a concentration of 10 8 spores/mL. A. tumefaciens AGL-1 was electro-transformed at 2.5 KV and 5.0 ms, then the transformants were cultured and diluted to an OD600 of 0.3 with fresh IM. Subsequently, the cells were incubated to an OD600 of 0.9, mixed with spore suspension and spread onto on solid IM with uracil of 0.05 g/L and transferred onto uracil-free SC medium with 100 µg/mL cefotaxime and 100 µg/mL spectinomycin. The obtained transformants were sub-cultured on uracil-free SC agar in three times to confirm the stability.

Genomic DNA Preparation
M. alpina strains were cultivated in liquid GY medium and the collected, then washed twice with sterile water. After removing most of the water, the mycelia were snap-frozen in liquid nitrogen. The DNA of M. alpina genomic was extracted as described previously (Wang et al., 2011).

Reverse Transcriptase Quantitative PCR (RT-qPCR) Analysis
RT-qPCR was carried out as described previously . Total RNA was extracted from the M. alpina using TRIzol reagent and reverse transcribed by the PrimeScript RT reagent kit. The primer pairs used for RT-qPCR were shown in Table 1. RT-qPCR was performed using the ABI Prism 7900 sequence detection system and Power SYBR green PCR master mix. The expression levels of all genes were normalised to the expression of the internal control gene (18S rRNA). The date were quantified by the method of 2 -∆∆Ct .

Determination of Enzymatic Activity
Mycelia were collected by filtration, snap-frozen, ground in liquid nitrogen and suspended in extraction buffer . The suspension was centrifuged and collected the supernatant, then the protein concentration was determined by the Bradford method. The enzymatic activities of G6PD were detected as described previously (Tang et al., 2015; and indicated by the NADPH production of per min and per mg protein.

NADPH Quantification
The M. alpina mycelia were rapidly collected, frozen and ground in liquid nitrogen. The NADPH levels were determined using a NADPH quantification colorimetric kit (Biovision, Milpitas, CA, USA) according to the manufacturers' instruction.

Cell Dry Weight (CDW) and Fatty Acid Methyl Esters (FAME) Analysis
Fungal mycelia were harvested, washed twice with distilled water and frozen in liquid nitrogen. After lyophilisation, the CDW was determined gravimetrically. Approximately 20 mg (dry weight) of mycelia were subjected to fatty acid extraction and methyl esterification. The profiles of FAMEs were analysed using gas chromatography-mass spectrometry as previously .

Statistical Analysis
The experiments were carried out in three biological replicates and the mean values and standard errors were calculated. A statistical analysis of the data was performed by SPSS 20.0. One-way analysis of variance (ANOVA) was conducted on the data and P < 0.05 was considered significantly different.

Co-expression oPpFADS17 and G6PD2 in M. Alpina Using the Double Expression Cassette and 2A Peptide Strategies
To generate a high-EPA-producing strain, oPpFADS17 and G6PD2 were amplified and co-overexpressed in M. alpina using the double expression cassette strategy. The recombinant plasmid pBIG2-ura5s-oPpFADS17-2A-G6PD2 was constructed as shown in Fig. 1. The recombinant vectors yielded target gene fragments of 818, 1256 and 1711 bp, indicating that the double expression cassette strategy was used to successfully construct the co-expression vector pBIG2-ura5s-oPpFADS17-G6PD2.
To compare the effectiveness of co-expression between the two strategies, oPpFADS17 and G6PD2 were also amplified and co-overexpressed in M. alpina using the 2A peptide strategy (Fig. 2). The recombinant vector pBIG2-ura5s-oPpFADS17-G6PD2 yielded target gene fragments of 818 and 2879 bp, indicating the successful construction of the coexpression vector.

Cell Growth and Lipid Accumulation in oPpFADS17 and G6PD2 Co-Expressing Strains
Next, the CDW, TFA content and EPA content of each M. alpina recombination strain was measured to investigate the effects of oPpFADS17 and G6PD2 coexpression via both strategies on cell growth and lipid accumulation ( Fig. 4 and 5). As shown in Fig. 4A and 5A, neither recombination strategy appeared to have a significant effect on CDW when the recombinant strains were compared with the control strain M. alpina ATCC 32222. In other words, the inserted genes did not affect cell growth. However, the recombination strains generated using the double expression cassette strategy exhibited clear increases in the TFA and MA-oPpFADS17-G6PD2-1/2 contents to 42.2% and 43.0% (the value of control strain is 34.9%), respectively, whereas the strains generated using the 2A peptide strategy did not differ from the control strain (Fig. 4B). Moreover, EPA was not detected in any of the recombination strains generated using the double expression cassette strategy (Fig.  4C). In other words, the co-expression of oPpFADS17 and G6PD2 via a double expression cassette increased the TFA but had no effect on EPA accumulation. In contrast, the TFA contents did not differ significantly between the recombination strains generated using the 2A peptide strategy and the control strain (Fig. 5B). Meanwhile, the recombination strains generated using the 2A peptide strategy varied widely with respect to the ratio of EPA to the TFA content. Specifically, the EPA contents in the strains MA-oPpFADS17-2A-G6PD2-1/2/3/4 increased to 29.4%, 30.6%, 2.4% and 4.1% (EPA is not detected in control strain), respectively, whereas EPA was not detected in MA-oPpFADS17-2A-G6PD2-5/6 (Fig. 5C). The differences in the TFA and EPA contents of these 2A peptide-generated recombination strains can be attributed to random integration based on the ATMT method (Ando et al., 2009).   The transcription levels of oPpFADS17 and G6PD2 in the MA-oPpFADS17-G6PD2-1/2 (double expression cassette strategy) and MA-oPpFADS17-2A-G6PD2-1/2 strains (2A peptide strategy) grown in broth liquid medium for 168 h were determined using RT-qPCR (Fig. 6). As wild-type M. alpina does not express oPpFADS17, we used a single gene-overexpressing strain (MA-oPpFADS17) from the previous study  as the control strain (Fig. 6A). In the transformants generated using the double expression cassette strategy, the respective level of oPpFADS17 transcription in MA-oPpFADS17-G6PD2-1 and -2 either did not differ significantly or was 16% lower when compared with the control strain. In the transformants using the 2A peptide strategy, the respective level of oPpFADS17 transcription in MA-oPpFADS17-2A-G6PD2-1 and -2 either did not differ significantly or increased by 26% when compared with the control strain. The results showed that the transcription level of the oPpFADS17 occurred in these transformants. The levels of G6PD2 transcription in the MA-oPpFADS17-G6PD2-1/2 transformants were approximately five-and three-fold higher than that in the wild-type strain M. alpina ATCC 32222. However, the levels of G6PD2 transcription in MA-oPpFADS17-2A-G6PD2-1/2 did not increase relative to that of the wild-type strain (Fig. 6B). These results suggest that the double expression cassette strategy led to the successful insertion and transcription of G6PD2, whereas the 2A peptide strategy had no significant effect on G6PD2 transcription.
To investigate whether the enzyme encoded by G6PD2 retained its activity in the transformants generated using the two strategies, the activity of G6PD and the production of NADPH, a product of catalysis supplied directly for lipid synthesis, were determined (Fig. 7). The levels of G6PD activity and NADPH in the transformants generated using the double expression cassette strategy increased significantly when compared with the wild-type strain. In contrast, neither level was affected in the transformants generated using the 2A peptide strategy, compared with the wild-type strain. These results were consistent with the transcription level data.

Discussion
The omega-3 desaturase oPpFADS17 is the key enzyme in the bioconversion of AA to EPA and G6PD2 can provide reducing power NADPH for lipid biosynthesis (Hao et al., 2016;Tang et al., 2017;. In this study, we coexpressed oPpFADS17 and G6PD2 in M. alpina using both the double expression cassette and 2A peptide strategies and studied the effects of co-expression on total lipid and EPA accumulation. In the transformants generated using the double expression cassette strategy, the total lipid content increased but no accumulated EPA was detected. Possibly, the promoters in the double expression cassette strategy interfered with each other and affected the expression of the target gene. Additionally, excessively long target fragments may have increased the burden on the host cells and affected cell metabolism. Several studies showed that the target product expression level correlated with the copy number of the corresponding expression cassette (Fang et al., 2014;Mansur et al., 2005;Per et al., 2011;Wu et al., 2014;Zhu et al., 2009). According to Yu et al. (2003)  different combinations of genes and promoters have important effects on promoter activity and high combinatorial matching could better achieve gene expression. In addition, the enhancer element of one transcription unit may activate the distal promoter of another unit, resulting in cross-interference between two adjacent expression units. Although insulators can be used to prevent this interference, the integration of primitive insulators with internal promoters/enhancers may also reduce the potential activities of adjacent genes. Consistent with this possibility, the first expression unit in our study did not achieve the desired effect and was likely controlled by the enhancer of another expression unit, resulting in abnormal expression. In many organisms, the 2A peptide facilitates the long-term stable co-expression of multiple genes (de Felipe et al., 2006). Indeed, we used the 2A peptide strategy to establish a feasible system for the coexpression of multiple genes in M. alpina. Here, the EPA content increased to as much as 30% of the TFA level and the level of oPpFADS17 transcription was higher than that in the wild-type strain. However, the TFA content did not increase significantly in transformants generated using this strategy, possibly due to the relatively low shear efficiency of the 2A peptide in the recombinant strains. Previous studies indicated that different 2A peptide exhibited various shear efficiencies in different hosts (Donnelly et al., 2001;Doronina et al., 2008;Luke et al., 2008;Szymczak et al., 2004). The shear efficiency of the 2A peptide may be associated with two problems. First, the shear efficiencies of polyproteins translated before and after the 2A peptide may differ. Second, the fusion of 2A peptide residues to the C-terminal of the upstream translation protein may interfere with the upstream protein function. The shear activity of the 2A peptide may be related to the total amount of precursor protein. Moreover, G6PD2 plays an important role in the pentose phosphate pathway and its transcription and enzyme activity levels are high. In the presence of a high total amount of precursor protein, the shear activity of the 2A peptide may decrease and the recombinant bacteria would not achieve the expected increase in total lipids. The polycistron employed a single promoter to drive mRNA transcripts containing multi-coding areas and therefore the strength of promoter is very important for the expression of target mRNA. In the 2A peptide strategy of this study, the promoter successfully drived the transcription of oPpFADS17 but may be unable to drive another G6PD2 co-expression gene. This observation is in accordance with other studies that showed the efficiency of heterologous expression and gene transformation were influenced by different promoter (Cheng et al., 2009;Kilaru et al., 2006). Also, the mRNA stability is another important factor in gene expression (Hargrove and Schmidt, 1989). The longer mRNA fragments were more likely to experience endonucleolytic attacks by RNA endonucleases and/or mechanical damage (Feng and Niu, 2007;Trcek et al., 2011). A nearly 3000-bp length might exceed the capacity to maintain mRNA stability.

Conclusion
In this study, we used double expression cassette and 2A peptide strategies to co-express the omega-3 desaturase oPpFADS17 and G6PD2 in M. alpina and investigated the effects on the accumulation of total lipids and EPA. In summary, recombination strains generated using the double expression cassette strategy exhibited an increase in the total lipids but no change in EPA accumulation, whereas those generated using 2A peptide exhibited an increase in EPA accumulation up to 30% of the total lipid content. These multiple gene coexpressing strategies are expected to further enhance the accumulation of PUFAs in M. alpina.