Metabolic Engineering of Mortierella alpina for Enhanced Arachidonic Acid Production through the NADPH-Supplying Strategy

ABSTRACT

NADPH is known to be a key cofactor required for fatty acid synthesis and desaturation. Various enzymatic reactions can generate NADPH. To determine the effect of NADPH sources on lipogenesis, glucose-6-phosphate dehydrogenase (G6PD), 6-phosphogluconate dehydrogenase (PGD), isocitrate dehydrogenase (IDH), and malic enzyme (ME) were overexpressed in Mortierella alpina. Our results showed that G6PD2 had the most significant effect on fatty acid synthesis, with a 1.7-fold increase in total fatty acid, whereas ME2 was more effective in desaturation, with a 1.5-fold increase in arachidonic acid (AA) content over control. Co-overexpression of G6PD2 and ME2 improved both fatty acid synthesis and desaturation. Within 96 h of fermentation using the fed-batch method, the co-overexpressing strain accumulated AA at a productivity of 1.9 ± 0.2 g/(liter · day), which was 7.2-fold higher than that in the M. alpina control that was cultured in a flask.

IMPORTANCE This study proved that the pentose phosphate pathway is the major NADPH contributor during fatty acid synthesis in M. alpina. The NADPH sources may be differently responsible for fatty acid synthesis or desaturation. Co-overexpression of G6PD2 and ME2 significantly increases AA production.

INTRODUCTION

NADPH is a critical factor for many biological processes in oleaginous microbes (1, 2). However, the function of the various sources of NADPH in fatty acid synthesis is still not fully understood. NADPH can be generated by several enzymatic reactions. Malic enzyme (ME; EC 1.1.1.40) was considered the sole NADPH provider for fatty acid synthesis in oleaginous microbes (3). Some evidence, however, indicated that ME is only one of several NADPH sources (4, 5). We and others have recently demonstrated that the pentose phosphate pathway (PPP), specifically the glucose-6-phosphate dehydrogenase (G6PD; EC 1.1.1.49) and the 6-phosphogluconate dehydrogenase (PGD; EC 1.1.1.44) genes, plays a major role during microbial fatty acid synthesis (1, 2, 6). Isocitrate dehydrogenase (IDH; EC 1.1.1.42) is another potential NADPH source and was proven to play a key role in fatty acid metabolism in adipocytes (7).

NADPH also plays an important role during fatty acid desaturation, which could be significantly improved by overexpressing a mitochondrial ME2 in Mortierella alpina (8) and Mucor circinelloides (9). These results implied that NADPH generated by cytosolic and membrane-bound enzymes may affect fatty acid synthesis and desaturation, respectively. Therefore, the role of the various sources of NADPH needs to be systematically evaluated and explored as an effective approach to increase the production of total fatty acids (TFA) and especially of unsaturated fatty acids.

Mortierella alpina is an oleaginous fungus used for the commercial production of arachidonic acid (AA); this organism can accumulate fatty acids of up to 50% of dry weight (10, 11). An expression system using the uracil auxotroph strain CCFM 501 was recently established and reliably applied for gene manipulation (1, 5, 8, 12, 13). In this study, the M. alpina genes encoding G6PD, PGD, and IDH were overexpressed in the same organism, and their effects on fatty acid synthesis were evaluated and compared with the effect of ME (5, 8). Subsequently, M. alpina was metabolically engineered for enhanced AA production using a strategy based on improving NADPH supply.

MATERIALS AND METHODS

Strains and culture media.

M. alpina strains were maintained on GY medium, which consists of 30 g/liter glucose, 5 g/liter yeast extract, 2 g/liter KNO₃, 1 g/liter NaH₂PO₄, and 0.3 g/liter MgSO₄·7H₂O; when culturing a uracil auxotroph, 5-fluoroorotic acid (5-FOA; 0.5 mg/ml) and uracil (0.05 mg/ml) were added. Agrobacterium tumefaciens C58C1 was cultured in YEP medium, which consists of 10 g/liter tryptone, 10 g/liter yeast extract, and 5 g/liter NaCl. The compositions of synthetic complete (SC) medium, minimal medium (MM), and induction medium (IM) were described previously (8, 14, 15).

Growth conditions.

Escherichia coli strain TOP10 was cultivated at 37°C on LB agar medium. Agrobacterium tumefaciens C58C1 was grown at 28°C in YEP medium. M. alpina strains were cultured at 28°C in broth medium at 200 rpm for 168 h. For batch fermentation, the proliferative-phase cultures of M. alpina were inoculated at 10% (vol/vol) into a 7.5-liter fermentor containing 4 liters broth medium, which consisted of 50 g/liter glucose, 5 g/liter yeast extract, 1.0 g/liter KH₂PO₄, 0.25 g/liter MgSO₄·7H₂O, and 10 g/liter KNO₃. Fermentors were held at 28°C and stirred at 500 rpm with an aeration rate of 0.5 vol/vol/min (vvm), and pH was maintained at 6.0. Samples were harvested prior to (sample A, −12 h; sample B, −2 h; sample E, −30 min) and after (sample K, +1 h; sample L, +12 h; sample M, +48 h) nitrogen exhaustion as previously described (1).

Construction of the transfer DNA binary vector.

The glucose-6-phosphate dehydrogenase genes (g6pd1, g6pd2, and g6pd3), the 6-phosphogluconate dehydrogenase gene (pgd), and the isocitrate dehydrogenase 1 gene (idh1) were amplified from the M. alpina cDNA with the primer pairs listed in Table 1. Genes were ligated into the pGEM-T Easy vector (Promega, Madison, WI, USA), and sequences were analyzed with an ABI Prism 3730 DNA analyzer. After being digested with the appropriate restriction enzymes, as indicated in Table 1, the genes were ligated into the binary vector pBIG2-ura5s-Its (5). For co-overexpression, the ME2 expression cassette was amplified with primer pair InFusF/InFusR and ligated into the XbaI-digested g6pd2 expression vector using the In-Fusion HD cloning kit (Clontech Laboratories, Mountain View, CA, USA).

TABLE 1.

Primers used in this study

Primer	Sequence (5′–3′)a	Description
Primers used for PCR amplification
G6PD1F	GCACGGGCTAGCATGACCTCCACCACCACCAC	g6pd1 amplification
G6PD1R	GCTCCCCCCGGGTCAAAGCTTGCTGTCTGCGT
G6PD2F	GCACGGGGTACCATGTCTGAGAAGAAGAAGCATCTTT	g6pd2 amplification
G6PD2R	GCTCCCCCCGGGTTAATGGTCAGTCCTTGTGTCCT
G6PD3F	GCACGGGGTACCATGTCCGCTGCCAAAACCG	g6pd3 amplification
G6PD3R	GCTCCCCCCGGGTTATGCCTTGTCAACCTTTTGGTC
PGDF	GCACGGGGTACCATGAACGACAATGGCTACACC	pgd amplification
PGDR	GCTCCCCCCGGGTTAAGCAAGGTAGGTGGTCGAG
IDH1F	ATACCCAAGCTTGAATGCTTGCCAACAAAATCAACG	idh1 amplification
IDH1R	ATACCCGAGCTCTTAAACGGTGCGCTTCTTCTGC
InFusF	CTCTCCTATGAGTCGTTTACCCAGAATGCACAGGTACACTTGTTT	PCR amplification
	AGAGGTCTAGATTTAGTTGATGTGAGAGTTGTGAGATTCGTG
InFusR	AAACGACAATCTGATCATGAGCGGAGAATTAAGGGAGTCACGTT
	ATGACCTCTAGACCTCTAAACAAGTGTACCTGTGCATTCTGGG	For in-fusion clone
HisproF1b	CACACACAAACCTCTCTCCCACT	T-DNA insert detection
TrpCR1b	CAAATGAACGTATCTTATCGAGATCC
HisproF2b	GTGTTCACTCGCATCCCGC	T-DNA insert detection
TrpCR2b	AGGCACTCTTTGCTGCTTGG
Primers used for RT-qPCR
G6PD1RTF	AGAGCTGGATCTGTCCTACCAT	G6PD1 RT-qPCR
G6PD1RTR	TCCAACTCGTCGCTTCTCAC
G6PD2RTF	GCGTACAAAGATGGATCGG	G6PD2 RT-qPCR
G6PD2RTR	TGAAAGCCGTCGTCTGTG
G6PD3RTF	CTCTTATCAGAGCGGGCAGTAC	G6PD3 RT-qPCR
G6PD3RTR	CACAGGAATAAAGACCGTGGG
PGDRTF	ACAATGGCTACACCGTCTGC	PGD RT-qPCR
PGDRTR	GACCTTACGAGGGCGCTTC
IDH1RTF	TTGCCAACAAAATCAACGGAG	IDH1 RT-qPCR
IDH1RTR	GATCGACATAGGGGAGAATGAGC
ME1RTF	GGCTGTTGCCGAAGGGACT	ME1 RT-qPCR
ME1RTR	GGCAAAGGTGGTGCTGATTTC
18SRTFc	CGTACTACCGATTGAATGGCTTAG	Internal control for RT-qPCR
18SRTRc	CCTACGGAAACCTTGTTACGACT

^aUnderlined sequences indicate the additional restriction sites.

^bAs described by Hao et al. (5).

^cAs described by Hao et al. (8).

ATMT.

Agrobacterium tumefaciens-mediated transformation (ATMT) was performed essentially as described previously (5, 8). The spores of M. alpina uracil-auxotrophic strain CCFM 501 were harvested from 2-week cultures grown on GY agar medium containing 5-FOA and uracil, followed by centrifugation at 12,000 × g for 20 min and dilution with fresh liquid IM to 10⁸ spores/ml. A. tumefaciens C58C1 was electrotransformed, and the transformants were identified by PCR. A. tumefaciens transformants were cultured at 28°C for 48 h with shaking at 200 rpm in 20 ml of liquid MM, which contained 100 μg/ml kanamycin and 100 μg/ml rifampin. A. tumefaciens cultures were centrifuged at 4,000 × g for 5 min and diluted to an optical density at 600 nm (OD₆₀₀) of 0.3 with fresh IM. The cells were incubated for 8 to 12 h at 28°C with shaking at 200 rpm until they reached an OD₆₀₀ of 1.2. Equal volumes of cell suspension and spore suspension were mixed and spread onto a cellophane membrane that was placed on solid IM (containing 0.9 g/liter glucose). The plates were incubated for 48 h in a dark incubator at 23°C. Incubated membranes were transferred onto uracil-free SC medium, containing 50 μg/ml cefotaxime and 50 μg/ml spectinomycin, which was followed by incubation at 28°C until colonies appeared. Positive transformants were transferred onto uracil-free SC agar plates (containing 50 μg/ml cefotaxime and 50 μg/ml spectinomycin) and were subcultured three times to obtain stable transformants. All experiments were carried out in triplicate.

Genomic DNA preparation.

M. alpina strains were cultivated in GY liquid medium for 4 days at 28°C with shaking at 200 rpm. Mycelia were harvested and washed twice with sterile water and then immediately frozen in liquid nitrogen. M. alpina genomic DNA was extracted as described previously (10).

Dry cell weight and glucose concentration assay.

Fungal mycelia were harvested and washed twice with distilled water and then frozen in liquid nitrogen. After lyophilization, the dry cell weight (DCW) was determined gravimetrically. The concentration of medium glucose was determined using a glucose oxidase test kit (Rongsheng Biotech, Shanghai, China) according to the manufacturer's instructions.

RT-qPCR analysis.

Reverse transcriptase quantitative PCR (RT-qPCR) was performed essentially as described previously (1). Total RNA of the M. alpina strains was extracted with the TRIzol reagent (Life Technologies, Grand Island, NY, USA) and reverse transcribed using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Otsu, Shiga, Japan) according to the manufacturer's instructions. The primer pairs used for RT-qPCR are shown in Table 1. The ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA) was used to perform RT-qPCR with the Power SYBR green PCR master mix (Applied Biosystems, Foster City, CA) in accordance with the manufacturers' instructions. The reaction mixture consisted of 10 μl of SYBR green PCR master mix, 0.5 μl of each primer pair, 8 μl of distilled water, and 1 μl of DNA template or distilled water as a no-template control. The PCR cycling conditions were 50°C for 2 min and 95°C for 10 min, followed by 40 cycles of amplification at 95°C for 15 s and 60°C for 30 s. The expression of the internal control gene (18S rRNA) was used as the normalization standard for gene expression.

Determination of enzymatic activity.

Mycelia were collected by filtration, immediately frozen and ground in liquid nitrogen, and then suspended in extraction buffer containing 20% (wt/vol) glycerol, 100 mM KH₂PO₄/KOH (pH 7.5), 1 mM benzamidine, and 1 mM dithiothreitol (1). The suspension was centrifuged at 10,000 × g for 10 min at 4°C. The supernatant was used for the determination of protein concentration using the Bradford method. The activities of ME2 and G6PD1 were analyzed essentially as described previously, after centrifugation at 100,000 × g for 1 h at 4°C (8). The enzymatic activities were detected using a continuous spectrophotometric assay at 340 nm, which was performed at 30°C as described previously (16–19). One unit of enzyme activity was defined as the amount of enzyme required to produce 1 nmol NADPH per min.

NADP and NADPH quantification.

The mycelia of each M. alpina strain were rapidly collected and frozen in liquid nitrogen. Samples were lyophilized and ground in liquid nitrogen. The NADP and NADPH levels were determined using the NADP/NADPH quantification colorimetric kit (BioVision, CA, USA) according to the manufacturer's instructions.

FAME analysis.

Approximately 20 mg of mycelia (dry weight) was used for fatty acid extraction and methyl esterification as described previously (10). Fatty acid profiles were analyzed as their methyl esters by gas chromatography-mass spectrometry (GC-MS) (GC-2010 Plus and GCMS-QP2010 Ultra; Shimadzu Co., Kyoto, Japan) with a 30-m by 0.25-mm Rtx-WAXetr column (film thickness, 0.25 μm). The temperature program was as follows: 40°C for 5 min, ramp to 120°C at 20°C per min, ramp to 190°C at 5°C per min, hold for 5 min, ramp to 220°C at 5°C per min, hold for 17 min. Helium was the carrier gas. Fatty acid quantification was carried out using peak height area integrals. The internal standard to quantify fatty acid methyl esters (FAME) with an aliphatic chain of ≤18 was pentadecanoic acid (C_15:0), and the internal standard to quantify fatty acid methyl esters with an aliphatic chain of >18 was heneicosanoic acid (C_21:0).

Statistical analysis.

All experiments were performed in triplicate, and the means and standard errors were calculated. SPSS 20.0 was used for one-way analysis and canonical correlation analysis, and significant differences (P < 0.05) were determined by the least significant difference test.

RESULTS

Expression levels of the endogenous g6pd, pgd, idh, and malE1 genes during fatty acid synthesis in M. alpina.

During fatty synthesis in oleaginous fungi, cytosolic NADPH is thought to be a rate-limiting factor that is mainly generated by PPP, IDH, and ME (1, 20). The expression levels of these genes were analyzed by RT-qPCR during fatty acid synthesis in M. alpina. Based on the results of analyses with the SubLoc v1.0 program (http://www.bioinfo.tsinghua.edu.cn/SubLoc/eu_predict.htm) and the subCELlular LOcalization predictor (CELLO) v2.5 (21), the G6PD1, IDH2, and ME2 were predicted to be localized in mitochondria, while the G6PD2, G6PD3, PGD, IDH1, and ME1 were cytosolic. The expression levels of G6PDs and PGD were significantly upregulated at the onset of the fatty acid accumulation stage (time point E), especially in the case of G6PD2, which showed a 30-fold increase (Fig. 1). The expression levels of IDH1 and ME1 were significantly elevated at the beginning of the fatty acid accumulation period, followed by a continuous downregulation during the rest of the culture period (Fig. 1). In contrast, ME2 remained at its original expression level, and IDH2 was significantly downregulated when M. alpina stopped proliferating to accumulate fatty acids (Fig. 1). This transcriptional downregulation of the mitochondrial IDH2 may result from the suppression of the tricarboxylic acid (TCA) cycle when cells stop proliferating after nitrogen exhaustion. All of the expression patterns of these genes were essentially consistent with the results of our previously performed transcriptome analysis (1). These results indicated that the expression of G6PDs, PGD, IDH1, and MEs, but not IDH2, were related to fatty acid synthesis in M. alpina.

FIG 1.

Expression levels of G6PDs, PGD, IDHs, and MEs in M. alpina during fatty acid synthesis. M. alpina was cultured in a 7.5-liter fermentor and sampled at various time points prior to and after nitrogen exhaustion (sample A, −12 h; sample B, −2 h; sample E, −30 min; sample K, +1 h; sample L, +12 h; and sample M, +48 h), and transcript levels were analyzed by RT-qPCR. G-6-P, glucose-6-phosphate; Gn-6-P, gluconate-6-phosphate; Ru-5-P, ribulose-5-phosphate; 2-KG, 2-ketoglutarate; CoA, coenzyme A; TCA, tricarboxylic acid. Three independent experiments were performed, and the error bars represent standard deviations.

Homologous overexpression of G6PDs, PGD, and IDH.

The genes g6pd1, g6pd2, g6pd3, pgd, and idh1 from M. alpina were overexpressed in the same organism. Two pairs of previously described specific primers (HisproF1/TrpCR1 and HisproF2/TrpCR2) (Table 1) were applied to confirm the presence of the integrated transfer DNA (T-DNA) in the genomic DNA of transformants (1, 5, 8). Expression levels, enzymatic activities, NADPH, and fatty acid levels were analyzed in three replicates of each transformed strain, including ME1 (malE1)- and ME2 (malE2)-overexpressing strains that were previously constructed (5). The expression levels of each analyzed gene in the corresponding overexpressing strains were significantly increased by 2- to 6-fold (Fig. 2A) without affecting the transcription of the other genes (see Fig. S1 in the supplemental material). As a result, the cytosol- or mitochondrion-specific enzymatic activities in the supernatant or pellet fractions were significantly increased in the corresponding overexpressing strains (Fig. 2B). Consequently, NADPH levels were significantly improved in the strains overexpressing cytosolic enzymes but not in the strains overexpressing mitochondrial enzymes (g6pd1 or malE2) (Fig. 2C).

FIG 2.

The transcription level (A), enzymatic activity (B), NADPH level (C), total fatty acid content (D), and AA content (E) of G6PD-, PGD-, IDH-, and ME-overexpressing M. alpina (MA) strains. Strains were cultured at 28°C in 500-ml shaking flasks containing 200 ml broth medium for 168 h at 200 rpm. Three independent experiments were performed, and the error bars represent standard deviations. *, P < 0.05 compared to wild type.

Overexpression of the g6pd2 gene led to the most significant increase in fatty acid content compared to that in the wild-type control (Fig. 2D). In strain MA-g6pd2-2, the fatty acid production reached 8.1 ± 0.5 g/liter, including 2.9 ± 0.2 g/liter AA (Table 2; see also Table S1 in the supplemental material). During the 168-h cultivation, the production of total fatty acids (TFA) increased by 1.7-fold and that of AA improved by 1.5-fold (Table 3). The improvement in TFA in the g6pd3-, idh1-, and malE1-overexpressing strains was not as high as that in the g6pd2-overexpressing strains (Fig. 2D). Interestingly, overexpression of G6PD1 significantly improved the fatty acid desaturation level (Fig. 2E). However, the improved AA productivity was not as significant as the >1.4-fold increase seen in the malE2-overexpressing strains (Table 3). Therefore, we reasoned that co-overexpressing the g6pd2 and malE2 genes would improve fatty acid accumulation and desaturation in M. alpina.

TABLE 2.

Fatty acid production in different M. alpina strains grown in broth medium for 168 h

M. alpina straina	Biomass (g/liter)b	Fatty acids (g/liter)b	AA (g/liter)b
Wild type	13.7 ± 1.1 AB	4.7 ± 0.3 AB	1.8 ± 0.1 A
MA-g6pd1-1	14.1 ± 0.5 AB	4.8 ± 0.2 ABC	2.2 ± 0.2 ABC
MA-g6pd1-2	13.6 ± 0.6 B	4.6 ± 0.2 C	2.1 ± 0.2 AC
MA-g6pd1-3	14.5 ± 0.8 ABC	4.6 ± 0.1 C	2 ± 0.2 A
MA-g6pd2-1	15.9 ± 0.7 AC	7.7 ± 0.4 D	2.8 ± 0.2 D
MA-g6pd2-2	16.4 ± 1.2 C	8.1 ± 0.5 D	2.9 ± 0.2 D
MA-g6pd2-3	15.3 ± 0.9 ABC	7.4 ± 0.4 D	2.8 ± 0.3 D
MA-g6pd3-1	14.2 ± 0.7 ABC	5.5 ± 0.6 CE	2 ± 0.2 A
MA-g6pd3-2	14.3 ± 0.6 ABC	5.6 ± 0.4 EF	2.1 ± 0.1 AC
MA-g6pd3-3	14.5 ± 1.1 ABC	5.4 ± 0.4 ACE	2.1 ± 0.1 AC
MA-pgd-1	13.8 ± 0.9 AB	5.6 ± 0.6 EF	2.1 ± 0.2 AC
MA-pgd-2	14.6 ± 0.7 ABC	6 ± 0.3 EFG	2.2 ± 0.2 ABC
MA-pgd-3	14.5 ± 0.7 ABC	5.9 ± 0.4 EFG	2.1 ± 0.1 AC
MA-idh-1	14.7 ± 0.6 ABC	5.8 ± 0.4 EFG	2.2 ± 0.2 ABC
MA-idh-2	14.9 ± 0.8 ABC	5.6 ± 0.3 EF	2.2 ± 0.2 ABC
MA-idh-3	14.3 ± 0.7 ABC	5.6 ± 0.6 EF	2.2 ± 0.1 ABC
MA-malE1-1	15.1 ± 0.9 ABC	6.5 ± 0.3 G	2.5 ± 0.1 BCD
MA-malE1-2	14.9 ± 1.3 ABC	6.3 ± 0.4 EG	2.5 ± 0.2 BCD
MA-malE1-3	14.8 ± 0.6 ABC	6.4 ± 0.5 G	2.6 ± 0.1 BD
MA-malE2-1	13.9 ± 1.2 AB	4.5 ± 0.3 AB	2.7 ± 0.2 BD
MA-malE2-2	14.2 ± 1.1 ABC	4.6 ± 0.4 ABC	2.6 ± 0.1 BD
MA-malE2-3	13.8 ± 0.7 AB	4.4 ± 0.2 AB	2.6 ± 0.2 BD

^aMA-g6pd1-1, MA-g6pd1-2, and MA-g6pd1-3, G6PD1-overexpressing M. alpina strains; MA-g6pd2-1, MA-g6pd2-2, and MA-g6pd2-3, G6PD2-overexpressing M. alpina strains; MA-g6pd3-1, MA-g6pd3-2, and MA-g6pd3-3, G6PD3-overexpressing M. alpina strains; MA-pgd-1, MA-pgd-2, and MA-pgd-3, PGD-overexpressing M. alpina strains; MA-idh-1, MA-idh-2, and MA-idh-3, IDH-overexpressing M. alpina strains; MA-malE1-1, MA-malE1-2, and MA-malE1-3, ME1-overexpressing M. alpina strains; MA-malE2-1, MA-malE2-2, and MA-malE2-3, ME2-overexpressing M. alpina strains.

^bValues within a row followed by different capital letters were significantly different (P < 0.05) as analyzed by analysis of variance (ANOVA).

TABLE 3.

Fatty acid production of MA-g6pd2-malE2-3 in fermentation cultures

M. alpina straina	Culture method	Yield (% glucose)b		Productivity, g/(liter · day)b
		TFA	AA	TFA	AA
Wild type	Flask	11.72 ± 0.61 A	4.74 ± 0.23 A	0.67 ± 0.11 A	0.27 ± 0.03 A
MA-g6pd2-2	Flask	19.38 ± 1.62 B	6.94 ± 0.34 B	1.16 ± 0.24 BC	0.41 ± 0.06 A
MA-malE2-1	Flask	10.90 ± 0.43 A	6.54 ± 0.53 B	0.64 ± 0.08 A	0.39 ± 0.04 A
MA-g6pd2-malE2-3	Flask	16.74 ± 1.72 C	9.41 ± 0.91 C	1.08 ± 0.15 C	0.60 ± 0.09 B
Wild type	Batch	20.20 ± 1.38 B	8.80 ± 1.03 C	1.44 ± 0.21 BD	0.63 ± 0.07 BC
MA-g6pd2-2	Batch	24.20 ± 1.34 DE	12.20 ± 0.32 D	1.73 ± 0.12 D	0.87 ± 0.12 DE
MA-malE2-1	Batch	19.12 ± 1.51 BC	11.43 ± 0.57 D	1.35 ± 0.16 BC	0.81 ± 0.08 CE
MA-g6pd2-malE2-3	Batch	23.80 ± 2.25 DE	13.63 ± 1.62 E	1.70 ± 0.09 D	0.97 ± 0.11 DE
Wild type	Fed batch	21.06 ± 0.91 BE	9.12 ± 0.53 C	2.31 ± 0.24 E	1.03 ± 0.06 DE
MA-g6pd2-2	Fed batch	24.41 ± 1.57 D	12.23 ± 0.81 D	3.05 ± 0.21 F	1.51 ± 0.16 F
MA-malE2-1	Fed batch	19.60 ± 1.81 B	11.23 ± 0.47 D	2.17 ± 0.22 E	1.22 ± 0.14 G
MA-g6pd2-malE2-3	Fed batch	24.60 ± 1.64 A	14.80 ± 0.95 E	3.14 ± 0.28 F	1.94 ± 0.22 H

^aMA-g6pd2-2, G6PD2-overexpressing strain; MA-malE2-1, ME2-overexpressing strain; MA-g6pd2-malE2-3, G6PD2- and ME2-cooverexpressing strain.

^bValues within a row followed by different capital letters were significantly different (P < 0.05) as analyzed by ANOVA.

Co-overexpression of g6pd2 and malE2.

Based on the results above, g6pd2 and malE2 were co-overexpressed to optimize AA production in M. alpina by the combined action of the two enzymes (Fig. 3). After 168 h of cultivation in shaken flasks, the transcription levels and enzymatic activities of co-overexpressing strains were significantly improved by 3- to 5-fold and 1.5- to 3-fold, respectively (Fig. 3A and B). Consequently, the NADPH level was also increased to the same level as that in the g6pd2-overexpressing strains (Fig. 3C). As a result, the TFA in co-overexpressing strains increased up to a relatively high level of approximately 46% of dry cell weight (Fig. 3D), which may be primarily due to the overexpression of G6PD2. Moreover, the AA contents of the g6pd2- and malE2-co-overexpressing strains were significantly increased up to the same level as those in the malE2-overexpressing strains (Fig. 3E), which indicates that the overexpressed ME2 was concurrently active with the overexpressed G6PD2. The improved fatty acid accumulation and desaturation resulted in a significantly increased AA productivity of over 2.2-fold compared with that of the wild-type control (Table 3). The co-overexpressing strain MA-g6pd2-malE2-3 was selected for further analysis in batch experiments.

FIG 3.

Co-overexpression of G6PD2 and ME2 in M. alpina. The expression level (A), enzymatic activity (B), NADPH level (C), TFA content (D), and AA content (E) in M. alpina strains were analyzed after co-overexpression of G6PD2 and ME2. Strains were cultured at 28°C in 500-ml shaking flasks containing 200 ml broth medium for 168 h at 200 rpm. Three independent experiments were performed, and the error bars represent standard deviations. *, P < 0.05 compared to control.

Batch fermentation of MA-g6pd2-malE2.

Batch fermentation was carried out on strain MA-g6pd2-malE2-3 in a 7.5-liter fermentor for 168 h with wild-type M. alpina as the control (Fig. 4; see also Table S2 in the supplemental material). Strain MA-g6pd2-malE2-3 consumed glucose more rapidly than the wild-type strain (Fig. 4A). The DCW therefore increased faster than in the wild-type control, which may reflect an improved fatty acid accumulation in MA-g6pd2-malE2-3 (Fig. 4B). The difference in TFA and AA production between MA-g6pd2-malE2-3 and wild-type M. alpina was approximately 2-fold at 120 h (Fig. 4C). After 168 h of fermentation, MA-g6pd2-malE2-3 accumulated TFA to approximately 12 g/liter (Fig. 4C). The AA productivity reached 0.97 ± 0.11 g/(liter · day) and was 3.6-fold higher than the AA productivity of the wild-type control in the flask culture, which was 0.27 ± 0.03 g/(liter · day) (Table 3). However, the improvement was not as significant as that in the flask fermentation experiments. Therefore, the fed-batch fermentation method may further enhance fatty acid production (22).

FIG 4.

Time course of residual glucose (A), DCW (B), and TFA and AA levels (C) in batch fermentation of MA-g6pd2-malE2-3. The 7.5-liter fermentors containing 4 liters of broth medium supplemented with 50 g/liter glucose were used for the batch experiment. Fermentors were held at 28°C and stirred at 500 rpm with an aeration rate of 0.5 vvm, and pH was maintained at 6.0. TFA, total fatty acids; AA, arachidonic acid; DCW, dry cell weight. Three independent experiments were performed, and the error bars represent standard deviations.

Glucose fed-batch fermentation of MA-g6pd2-malE2.

The fed-batch method was applied to maintain the glucose concentration at a relatively low level (approximately 10 g/liter). Glucose was started at a concentration of 25 g/liter and was subsequently supplied at 5, 10, 5, and 5 g/liter at time points 36, 48, 60, and 72 h, respectively (Fig. 5; see also Table S3 in the supplemental material). The total glucose supplied (50 g/liter) was depleted within 96 h, which saved approximately one more day of culture compared to the batch fermentation of MA-g6pd2-malE2-3. The DCW and TFA of MA-g6pd2-malE2-3 reached the same levels as those in the former fermentation experiment (Fig. 4A and B and 5). Consequently, the AA productivity of MA-g6pd2-malE2-3 increased by 7.2-fold and reached 1.9 ± 0.2 g/(liter · day), with a yield of 14.80% ± 0.95% of glucose (Table 3). The relative high yield of fatty acids may also be due to the presence of considerable amounts of yeast extract in the culture medium (Table 3).

FIG 5.

Time course of residual glucose concentration, DCW, TFA, and AA levels in fed-batch fermentation of MA-g6pd2-malE2-3. The 7.5-liter fermentors containing 4 liters of broth medium were used for the batch experiment. Fermentors were held at 28°C and stirred at 500 rpm with an aeration rate of 0.5 vvm, and pH was maintained at 6.0. Glucose was started at the concentration of 25 g/liter and was subsequently supplied at 5, 10, 5, and 5 g/liter at time points 36, 48, 60, and 72 h, respectively. Three independent experiments were performed, and the error bars represent standard deviations.

DISCUSSION

NADPH is a critical cofactor required for fatty acid synthesis in all organisms, including the oleaginous microbes. Many enzymes can generate NADPH; however, their relative contributions to lipogenesis are still unclear. We have previously shown that ME1 overexpression increases total fatty acid production whereas ME2 facilitates fatty acid desaturation (5, 8). This finding suggests that various sources of NADPH may play different roles during fatty acid synthesis and that the simultaneous overexpression of cytosol- and membrane-type NADPH-producing enzymes would augment the levels of total fatty acids as well as their desaturation. We have recently demonstrated that the pentose phosphate pathway is the major NADPH contributor during lipogenesis in M. alpina (1). In this study, homologous overexpression revealed that G6PD2 and ME2 play more significant roles than G6PD1, G6PD3, PGD, IDH1, and ME1 as NADPH suppliers during fatty acid synthesis and desaturation, respectively. Therefore, co-overexpression of G6PD2 and ME2 may be the method of choice to increase unsaturated fatty acid production.

IDH is a key enzyme in the TCA cycle and may play an important role in mammalian energy metabolism and tumorigenesis (23–25). In the oleaginous fungus Mucor circinelloides, IDH was shown to be a minor NADPH contributor compared to G6PD and ME (6). In M. alpina, overexpression of IDH1 slightly increased the level of NADPH but did not enhance fatty acid synthesis (Fig. 2C and D). These results suggest that, unlike in mammalian cells, IDH may not play a significant role in fungal lipogenesis, possibly due to the insufficient supply of cytosolic isocitrate (26, 27).

Overexpression of G6PD2 and ME2 improved AA productivity by 52% and 44%, respectively, in M. alpina. Thus, it is possible that co-overexpression of G6PD2 and ME2 has an additive effect, which would result in a nearly 2-fold increase in AA production. Interestingly, the co-overexpression resulted in a 2.2-fold increase (Fig. 6), suggesting that G6PD2 and ME2 have a synergistic effect on AA production (28).

FIG 6.

Relative AA productivity levels of different overexpressing strains in flask, batch, and fed-batch cultivation.

We noted accelerated cell growth in MA-g6pd2-malE2-3 compared to that in the wild-type strain. This may due primarily to the increased G6PD activity. G6PD is essential for cell survival and plays an important role in reductive biosynthesis for lipogenesis as well other roles, such as protection from oxidative stress and reactive oxygen species (ROS)-mediated cell death (29–32). Moreover, the cell antioxidant system, composed of the glutathione system, catalase, and superoxide dismutase, is NADPH dependent (33). NADPH is also the substrate of the family of NADPH oxidases that plays an important role during normal cell growth (34).

Although a synergistic effect of G6PD2 and ME2 on AA production was observed in flask cultures, the levels of total fatty acids and AA were relatively low. In order to increase their production, a fermentor was used. As a result, total fatty acid and AA production were significantly enhanced. However, the synergistic effect of G6PD2 and ME2 waned. This may be due, in part, to the differential growth rate of M. alpina under flask and fermentor conditions. High glucose concentrations change mycelium morphology from feather-like to a tight pellet in fermentors, which disrupts the transfer of nutrients and oxygen and thereby reduces fatty acid desaturation (35). In addition, fermentor cultures have high basal levels of fatty acids, which make further increases in AA production more difficult. To increase AA production, a glucose fed-batch method was applied to maintain a low yet sufficient concentration of this carbon source. This approach restored the synergistic effect of G6PD2 and ME2 on AA production (Fig. 6).

In conclusion, our results indicate that cytosolic and membrane sources of NADPH can differentially affect fatty acid synthesis. Simultaneous overexpression of G6PD2 and ME2 has been identified as an effective approach to improve AA production in M. alpina. This work represents a significant step forward in polyunsaturated fatty acid (PUFA) production, and a similar approach may also be applicable in engineering other oleaginous microbes.