The bacterium Escherichia coli Top10 was used for plasmids construction, preservation, and propagation. LB medium (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, and 20 g/L agar) were used to grow E. coli and supplemented with ampicillin (100 μg/mL) when necessary. The uridine auxotroph pleu-MU402 derived from wild-type strain CBS 277.49 (Yang et al., 2020) was used as the parent strain in genetic modification experiments (Yang et al., 2019). The cDNA library of WJ11 (CCTCC No. M 2014424) was used as the source of the genes mt and sodit. The fungus was maintained on complete media YPG (20 g/L glucose, 10 g/L peptone, 3 g/L yeast extract, and 20 g/L agar; Bartnicki-Garcia and Nickerson, 1962) or uridine-less media MMC (20 g/L glucose, 10 g/L casaminoacid, 0.5 g/L yeast nitrogen base without amino acids and ammonium sulfate, and 20 g/L agar; Nicolas et al., 2007), which were adjusted to pH 4.5 or 3.2 for the mycelia and colonial growth, respectively, and the culture temperature was set at 28∘C. Transformation was carried out as previously described (Gutierrez et al., 2011; Garre et al., 2015), and the transformants were grown at 28∘C in MMC medium supplemented with uracil at 200 mg/mL when required.

Transformants Mc-mt (mt-overexpressing), Mc-sodit (sodit-overexpressing), and Mc-1552 [as the control, wherein orotidine 5′-phosphate decarboxylase gene (pyrG) stored in pMAT1552 was transformed into MU402] were inoculated in 1 L baffled flasks, contained 150 mL K&R medium, in which 30 g/L glucose was added as the carbon source, and 3.3 g/L diammonium tartrate and 1.5 g/L yeast extract were added as the nitrogen sources, 1.5 g/L MgSO₄⋅7H₂O, 8 mg/L FeC1₃⋅6H₂O, 1 mg/L ZnSO₄⋅7H₂O, 0.1 mg/L CuSO₄⋅5H₂O, 0.1 mg/L Co(NO₃)₂⋅6H₂O and 0.1 mg/L MnSO₄⋅5H₂O, 7.0 g/L KH₂PO₄, 2.0 g/L Na₂HPO₄, and 0.1 g/L CaCl₂⋅2H₂O (Kendrick and Ratledge, 1992) were included. The culture was kept for 24 h at 28∘C, shaking at 130 rpm and then incubated into 2 L fermenter with 1.5 L of the modified K&R medium that contained glucose (80 g/L) and diammonium tartrate (2 g/L). Fermenters were monitored during the growth process at 28∘C, stirred at 700 rpm with 1.0 v/v min^–1 aeration and the pH automatically adjusted at around 6.0 by 2 mol/L NaOH.

Plasmid pMAT1552 (Zhao et al., 2015), contained the pyrG of CBS 277.49 was used as a selectable maker and the 1 kb fragments of up- and downstream of CarRP sequences, was adopted to construct the mt-overexpressing and sodit-overexpressing plasmids. Mt and sodit genes were amplified from the cDNA of WJ11 with corresponding primers mt-F/R, sodit-F/R (Supplementary Table 1). The PCR fragments were inserted into plasmid pMAT1552 digested by XhoI after a strong promoter pzrt1 to generate gene overexpressing plasmids pMAT1552-mt and pMAT1552-sodit (Figure 1A) using the One Step cloning kit (Takara).

For DNA extraction and RNA isolation, M. lusitanicus transformants were grown for 4 days in fermenter with the modified K&R medium at 28∘C. After fermentation period, the samples were harvested by filtration, and then washed three times with distilled water. Genomic DNA was extracted from mycelium after crumbling under liquid nitrogen using the DNA Quick Plant System kit (Tiangen Biotech Co., Ltd). Mycelia, harvested at 24 h, were fully grounded with Trizol in liquid nitrogen to obtain the total RNA of the transformants. Then the total RNA was used to reverse transcribed to cDNA via ReverTra Ace qPCR RT Kit (Roche) as described in the instructions. RT-qPCR was carried out in LightCycler 96 (Roche) using the SYBR Green Realtime PCR Master Mix in accordance with the manufacturer’s instructions, which was based on the method of 2^–ΔΔCt. Primers for RT-qPCR are listed on the Supplementary Table 1.

The biomass was collected on a weighed and dried filter paper by filtered via a Buchner funnel, purged three or four times with distilled water, frozen overnight at −80∘C and freeze-dried to constant weight, the weight of which was determined gravimetrically. Cellular lipids were extracted from ∼10 mg freeze-dried samples by chloroform/methanol (2:1, v/v; Folch et al., 1956), pentadecanoic acid (C15:0) was used as the internal standard, and then 10% HCl/methanol (w/w) was added for transesterification. The fatty acid methyl esters were separated with n-hexane (HPLC grade) and analyzed by gas chromatography (GC) which is equipped with a 30 m × 0.32 mm DB-Waxetr column. The GC program was set at 120∘C for 3 min and ramped up to 200∘C at the speed of 5∘C/min, then increased to 220∘C and hold for 2 min.

In the case of M. lusitanicus transformants cultivations, the concentration of glucose and ammonium ion in media were determined by glucose oxidase Perid-test kit (Shanghai Rongsheng Biotech Co., Ltd.) and indophenol method as described (Chaney and Marbach, 1962), respectively. Meanwhile, the concentration of extracellular citrate was measured by using a citrate kit (Suoqiao Biotec.).

The extracellular malate concentration in the medium for 72 h was measured by HPLC equipped with a 4.6 mm × 250 mm Diamonsil C18 column (Dikma, China; Yang J. et al., 2021). The mixture of methanol/phosphate solution (Na₂HPO₄ 7.01 g/L; 2:98, v/v) served as mobile phase mobile phase. The initial column temperature was set at 25∘C, flow rate was 0.5 mL/min and the running time was 26 min.

All the experiments were performed in three replicates and the results were presented as mean ± SD. Student’s t test of SPSS 16.0 was used for statistical analysis of the data and p < 0.05 was considered as significant different.

The gene sequences of mt and sodit were found based on the genomic data of WJ11 and sequence alignment analysis in the National Center for Biotechnology Information (Yang et al., 2020). To investigate the involvement of mt and sodit in lipid accumulation in M. lusitanicus CBS 277.49, overexpression strains of these two genes were generated. The expression vector, pMAT 1552, contains the pyrG gene as a selectable maker, a strong promoter pzrt1 for regulating the expression of the target genes, and the up-and down-sequences of carRP. The control plasmid pMAT1552 and the target gene overexpressing plasmids pMAT1552-mt and pMAT1552-sodit were transformed into defective strain MU402 to obtain the transformants Mc-1552, Mc-mt and Mc-sodit, respectively (Figure 1A). The implementation of transformants selection was performed as described in a previous study (Rodriguez-Frometa et al., 2013). For each overexpression plasmid, three independent transformants, named Mc-mt-1, Mc-mt-2 and Mc-mt-3 for pMAT1552-mt, Mc-sodit-1, Mc-sodit-2 and Mc-sodit-3 for pMAT1552-sodit, and one Mc-1552 as the control strain were selected, and the obtained recombinant transformants were verified by PCR analysis. The target genes, mt and sodit, and the 585 bp sequences of plasmid pMAT1552 were amplified by a primer pair 1552-F/R (Supplementary Table 1). As shown in Figures 1B,C, the PCR products for each transformant were 585 bp (Mc-1552), 1,705 bp (Mc-mt-1, Mc-mt-2, and Mc-mt-3), and,2292 bp (Mc-sodit-1, Mc-sodit-2, and Mc-sodit-3), respectively, which were consistent with the control strain and the corresponding transformants. The PCR amplification results verified that the target genes, mt and sodit, were transferred into the fungus. The transformants were cultured in baffled flask with K&R medium for 4 days. Since there was little difference in lipid content among transformants (data not show), the highest lipid content of the obtained transformants, was selected for further experiments. Notably, in our preliminary findings of homologous overexpression of target strains, the lipid content was not changed significantly if compared with the heterologous overexpression. Thus, the heterologous overexpression was performed in further experiments, and the compared data in their lipid content between the homologous and heterologous overexpression were shown in Figure 2 (homologous overexpression related details were showed in Supplementary Figures 1, 2).

To analyze the mRNA levels of mt and sodit genes in transformants, Mc-mt, Mc-sodit and Mc-1552 were grown for 24 h, and were analyzed by RT-qPCR compared to the control Mc-1552. In the transformants Mc-mt and Mc-sodit, there are homology genes with mt and sodit, respectively, therefore, two pairs of primers (Supplementary Table 1) were designed. As shown in Figure 3, the expression levels of the genes from WJ11, mt and sodit, in the recombinant transformants were significantly higher than that of the endogenous genes, which indicates that these genes were successfully overexpressed. However, the expression levels of endogenous genes were not affected.

The biomass and lipid content of mt- and sodit-overexpressing strains were measured when cultured in 2 L fermenters for 96 h (Figure 4). The growth patterns of the transformants were similar with the control strain. The biomass of the mt-overexpressing strain was lower than that of control strain throughout the fermentation, whereas the sodit-overexpressing strain had a higher biomass than the control strain in the early stages of fermentation but decreased after 72 h (Figure 4A). Therefore, the lipid content of mt- and sodit-overexpressing strains was significantly increased (Figure 4B). The lipid accumulation patterns of three strains were almost similar, lipid accumulation increased rapidly during in the first 48 h, afterward entered a stable period, and then ceased after 72 h. The amounts of lipid accumulated in the mt-overexpressing strain reached 16.2%, an increase of 24.6% compared to the control strain, which was coincides with the results of the previous study (Zhao et al., 2016). While the lipid content of sodit-overexpressing strain was increased by 33.8% compared to Mc-1552 (from 13.0% in control to 17.4% in Mc-sodit), nevertheless, the sodit gene deletion in WJ11 experiments, showed lipid accumulation was increased, which indicated this gene had a different function in regulating malate transportation in oleaginous and non-oleaginous fungi (Yang J. et al., 2021). The fatty acid profiles of these transformants illustrated that GLA content in total fatty acids (TFAs) of mt-overexpressing strain (18.41% at 72 h) and sodit-overexpressing strain (18.66% at 72 h) were significantly lower than that of Mc-1552 (27.35% at 72 h; Figure 5). However, the content of hexadecanoic acid (C16:0) in TFAs of mt-overexpressing strain (23.80% at 72 h) and sodit-overexpressing (23.23% at 72 h) strain were significantly higher than that of strain Mc-1552 (17.10% at 72 h).

The consumption rate of glucose and nitrogen sources was nearly similar in all studied strains (Figures 6A,B). Notably, when the nitrogen was depleted at 12 h, the consumption glucose rate, was rapidly increased in transformants and the control strain (Figure 6A). Interestingly, the amounts of extracellular citrate in the transformants were greatly increased during the whole cultivation period compared to the control. There were no significant differences between the amount of extracellular citrate in gene overexpressed transformants, Mc-sodit and Mc-mt during the stage of cell growth and rapid lipid accumulation (Figure 6C). Whereas, the concentration of extracellular citrate of Mc-sodit was much higher than that of Mc-mt after 48 h during the period of slow lipid accumulation, while the biomass decreased relatively as shown in Figure 6C. These results indicated that the intracellular glycolysis, lipid biosynthesis and citrate metabolism had been affected by the heterologous expression of mt and sodit genes and led to the increased accumulation of lipid in M. lusitanicus.

To investigate whether overexpression of two genes (mt and sodit) from WJ11 could affect the malate metabolism, the extracellular malate concentration of samples cultured for 72 h in 2 L fermenter with K&R medium was analyzed by HPLC. As shown in Figure 7, the concentration of extracellular malate in sodit-overexpressing strain was higher compared to the control, whereas, the malate secretion in mt-overexpressing strain was the lowest, which illustrated that these two MT could have different regulation to malate transportation.

The expression levels of key genes related to fatty acid synthesis at 24 h revealed the underlying mechanisms of lipid accumulation in mt- and sodit-overexpressing strains (Figure 8). There are five genes encoding malic enzymes which could convert malate to pyruvate and produce reducing power NADPH for fatty acid synthesis, three localized in mitochondria (mme1, gene ID: 78524; mme2, gene ID: 166127; and mme3, gene ID: 11639) and two localized in cytosol (cme1, gene ID:182779; cme2, gene ID: 186772; Vongsangnak et al., 2012). All of these genes were significantly up-regulated in both gene overexpressing strains, the transcription levels of mme1, mme2, and mme3 were higher in mt-overexpressing strain, while the transcription levels of cme1 and cme2 were higher in sodit-overexpressing strain. There are two genes (fas1, gene ID: 72770; fas2, gene ID: 183529) encoding fatty acid synthase which could catalyze de novo fatty acid synthesis. As shown in Figures 8F,G, the transcription levels of fas1 were slightly down-regulated whereas the fas2 was significantly up-regulated in both strains. In addition, ACL (encoded by acl, gene ID: 110808), which could catalyze the cleavage of citrate to acetyl-CoA for fatty acid synthesis, was also analyzed. The transcription levels of acl were significantly up-regulated in mt- and sodit-overexpressing strains. The mitochondrial citrate transporter (encoded by ct gene, gene ID: 155787) and tricarboxylate transporter (encoded by tct gene, gene ID: 141738) expression levels are investigated, since they are involved in mitochondrial citrate transportation which could be affected by the malate metabolism. The transcription levels of ct and tct were significantly up-regulated in both strains, however, the expression levels of these two genes were higher in sodit-overexpressing strain, which was in agreement with the extracellular citrate concentration.