Strain, seed culture, and reagents

Umbelopsis Isabellina ATCC 42613 was obtained from the American Type Culture Collection (Manassas, VA). Spores were washed by sterile deionized water and collected after cultivation on potato dextrose agar (PDA) medium at 30 °C for two weeks. Seed was cultured in 24 g L⁻¹ potato dextrose broth (PDB) medium with 8 g L⁻¹ yeast extract (YE) (Neogen, Lansing), at room temperature of 25 °C and a shaking speed of 180 rpm for 36 h. Inoculation size was 5% (v/v) if not specified. Chemicals, substrates (except ¹³C-formate from Cambridge Isotope Laboratories, Inc., Tewksbury, MA), and reagents were purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO).

Fermentation

Batch fermentation was conducted using shaking flasks and fermenters. For the shaking flask culture, 2-L Erlenmeyer flasks containing 0.7 L medium were placed on a New Brunswick Shaker at 180 rpm and cultured at 25 °C. For the fermenter culture, 7.5-L fermenters (New Brunswick–Eppendorf, Bioflo 115, Hauppauge, NY) were used to carry out the experiment. The fermenters contained 3 L medium. The airflow rate was 1 vvm, and the agitation speed was 200 rpm. The flask culture of formate medium contained 0.5 g L⁻¹ YE, 2 g L^{−1 13}C-labeled formate, and trace elements [27]. The flask culture of glucose and formate medium contained 0.5 g L⁻¹ YE, 4 g L^{−1 13}C-labeled formate (99% purity), 10 g L⁻¹ glucose, and trace elements. The duration of all flask cultures was 72 h. The medium for the fermenter culture contained 1 g L⁻¹ YE, 10 g L⁻¹ glucose and 4 g L^{−1 13}C-labeled formate (the control experiment excluded formate supplement). The duration of fermenter culture was 48 h. The pH of the flask cultivation was maintained at 6.0 by manually adding 1 M sulfuric acid or sodium hydroxide every 6 h in first 12 h and every 12 h afterwards. The pH of the fermenter cultivation was automatically maintained at 6.0 using 1 M sulfuric acid or sodium hydroxide. Samples were taken periodically for biomass composition and metabolite measurement.

Metabolite and biomass composition analyses

Formate and glucose in the fermentation broth were determined by a High-Performance Liquid Chromatography (HPLC) (Shimadzu Prominence, Japan) equipped with a Bio-rad Aminex HPX-87H analytical column and a refractive index detector [27]. As for NADH measurement, one ml broth was centrifuged at 4000 rpm for 1 min and the precipitated biomass was washed by cold phosphate buffered saline (PBS) before the NADH was measured by the NAD+/NADH Quantification Colorimetric Kit (BioVision, San Francisco). Standard solutions were prepared and measured on daily basis.

Fungal biomass with the known volume of fermentation broth was collected by vacuum filtration and washed three times by deionized water before drying at 105 °C overnight for dry matter measurement. Lipid extraction of the fungal biomass was conducted using the Bligh and Dyer method [28], and the lipid content was determined gravimetrically. The extracted lipid was then subjected to methanolysis to turn fatty acids into methyl esters [29]. The resulting methyl esters were further analyzed by gas chromatography–mass spectrometry (GC–MS) to obtain the fatty acid profile [27]. F.A.M.E. Mix (C4–C24) with serial dilutions (100, 50, 25, 12.5, 6.25, 3.125, 1.5625, 0.78125 μg mL⁻¹) and chloroform was applied as external standard. Samples were diluted ten times prior to analysis to reach reasonable detection levels. All samples and standards were spiked with 50 μg mL⁻¹ methyl nonadecanoate (C19:0) as internal standard.

Carbon isotopomer analysis

Biomass samples were harvested by centrifuge, and the cell pellets were hydrolyzed with 6 mol L⁻¹ HCl at 100 °C overnight. The resulting amino acid mixtures were subsequently air dried and derivatized with N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide (TBDMS) prior to GC–MS analysis. A published software was used to analyze and correct amino acid MS data, as described in the previous report [30]. The fragments [M-57] or [M-15], containing the entire carbon backbone of the amino acids, were used to demonstrate the incorporation of ¹³C-formate. Results are described using m + 0, m + 1, and m + 2 for unlabeled, singly labeled, doubly labeled amino acids, and so on.

Flux balance analysis (FBA) of fungal formate co-utilization

To understand the influence of fungal formate metabolism on biomass growth, a genome-scale yeast model, iMM904 (bigg.ucsd.edu), was adapted to provide stoichiometric insights into formate–glucose co-metabolism [31]. The model was modified by adding a pathway from formate to methylene-THF. The growth rates were predicted based on different glucose/formate co-utilization influxes and the objective function of maximal biomass production. The FBA simulation using COBRA tool box can then determine the carbon flows from formate to biomass synthesis (formate assimilation) and to CO₂ release for NAD(P)H generations (formate oxidation).

Formate as an energy and carbon source to enhance the growth of U. isabellina

The isotopomer analysis [32] provided the labeling information of critical precursor metabolites during formate assimilation such as pyruvate, acetyl-CoA, methylene-THF, formyl-THF, glycine, and serine (Fig. 1). The fungal culture on ¹³C-formate supplemented with yeast extract shows that eight amino acids (alanine, aspartate, glutamate, glycine, leucine, methionine, phenylalanine, and serine) were detected with significant labeling extent (Fig. 2a). Among them, methionine and serine are the amino acids with a large amount of m + 2 (e.g., over 10% of methionine was labeled with two carbons). Besides, methionine labeling enrichment is significantly (P < 0.05) higher than its precursor aspartate, indicating the labeling of Methyl-THF unit (Asp + Methyl-THF → methionine). This result demonstrates that U. isabellina assimilates ¹³C-formate via folate-mediated C1 pathways to generate Methyl-THF (Fig. 1). On the other hand, the TCA cycle-derived amino acids (aspartate and glutamate) show enriched ¹³C (15~30%, mostly m + 1), which indicates that the strain also uses anaplerotic pathway (PEP + CO₂ → Oxaloacetate) to fix the CO₂ from formate oxidation (Fig. 1). Similar labeling distributions were observed when adding formate in glucose medium (Fig. 2b). The ¹³C-fingerprinting provided the evidence of formate contribution to fungal growth through both formate assimilation and oxidation pathways. However, the experimental data also demonstrate that formate is not an efficient carbon source to support high-yield biomass growth (Table 1). The biomass accumulation on the formate medium was much lower than the culture on the glucose medium. Mixing formate with other carbon sources could be a good solution to synergistically utilize carbon and energy contents in formate and enhance carbon utilization efficiency of the fermentation. Therefore, the effects of formate as a supplemental carbon source on fungal metabolism was further studied.

	Biomass concentration (g L−1)
Culture time (h)	0	48	96
Formate medium	0.31 ± 0.03	0.58 ± 0.04	0.58 ± 0.08
Glucose medium	0.21 ± 0.02	3.67 ± 0.43	7.58 ± 0.28

Data are average of three replicates with standard deviation

Umbelopsis isabellina growth on glucose media with/without ¹³C-formate demonstrates that formate as a supplemental carbon and energy source significantly (P < 0.05) improved the fungal biomass accumulation (Fig. 3a). At the end of the culture (48 h), the cultivations with and without formate reached 4.3 and 3.8 g L⁻¹ biomass, respectively. The data also demonstrate that the formate and glucose were simultaneously consumed without glucose catabolite repression. Formate also influences intracellular NADH concentrations (Fig. 3b). NADH in the culture with formate peaked at 352 pmol mg⁻¹ dry biomass at 24 h (in the middle-log phase growth). In contrast, the control culture without formate had a much lower NADH at the same growth phase. When glucose and formate were depleted at the end of fermentations, NADH contents in both cultures were decreased to ~110 pmol mg⁻¹ dry biomass. This phenomenon confirms the role of formate as an auxiliary energy source to alleviate glucose catabolic burdens for NADH production and save glycolysis metabolites for biosynthesis [26]. Since glucose shows no catabolite repression on formate metabolism, the use of formate to relieve energy burden could be a preferable technique to improve performance of microbial cultivation [33].

Effects of formate on lipid synthesis

Lipid accumulation during the fermentation was also influenced by formate (Fig. 4). The lipid contents in the fungal biomass at the end of cultivation reached 43 and 38 g lipid/100 g dry fungal biomass for the media with and without formate, respectively. Compared to the control culture (glucose medium only), the culture on the glucose plus formate medium significantly (P < 0.05) increased biomass yield (up to 13%), biomass productivity (up to 20%), lipid yield (up to 30%), and lipid productivity (up to 70%) (Fig. 4). As discussed in the previous section, formate co-metabolism alleviates the glucose dissimilation burden, and subsequently allows more glucose carbon to flow into lipid.

Interestingly, adding formate also changed the fatty acids profile (Fig. 5). The fatty acid biosynthesis includes recurring reactions that require a large amount of reducing factors and ATP [36]. The energy from formate oxidation may re-program fatty acid profiles. Short- and medium-chain fatty acids (C6:0 to C16:0) were all upregulated, while most of long-chain fatty acids (C17:0 to C22:1) were downregulated with the addition of formate (Fig. 5). Medium-chain fatty acids are widely used in detergents and lubricants due to their lathering and low-viscosity properties [37]. However, yeast species usually did not produce short- and medium-chain fatty acids except for engineered strains or cultures growing in synthetic media with high C/N ratios and low culture temperature [37]. Therefore, using formate to control fatty acid profile can be applicable in industrial fermentation processes.

Flux balance analysis (FBA) of fungal growth on glucose and formate

Glucose/formate uptake rates and biomass accumulation from the experiments were used to carry out FBA (Table 3). The FBA results verify the positive effects of formate on fungal biomass accumulation and ATP generations (Fig. 6). First, formate oxidation significantly contributes to ATP generation through respirations (formate → NADH → ATP), particularly at lower glucose uptake rate (Fig. 6a). ATPs were released from mitochondria when cell metabolism aerobically consumes both glucose and formate (Fig. 6a). Since this process requires a large amount of dissolved oxygen, high aeration is needed to maintain the formate oxidation of ATP generation. Second, the amount of formate carbon used for the biomass accumulation (through folate metabolism) depends on both glucose and formate uptake rates. Lower glucose and higher formate uptake rates are beneficial for formate assimilation (Fig. 6b, c, d). For instance, at high glucose uptake rates of 0.2–0.8 mmol glucose g⁻¹ dry cell weight (DCW) h⁻¹, formate was mainly used as the energy source and contributed little to biomass production  (Fig. 6b). Once the glucose uptake rate dropped to 0.15 mmol glucose g⁻¹ DCW h⁻¹, increasing formate utilization from 0 mmol glucose g⁻¹ DCW h⁻¹ to 5 mmol glucose g⁻¹ DCW h⁻¹ promoted biomass production (Fig. 6d). It is apparent that formate becomes both energy and carbon sources under glucose-limiting conditions (Fig. 6c). Third, the FBA also demonstrates in silico growth with formate in the absence of glucose (Fig. 6d). Under a high formate uptake rate of 5 mmol formate g⁻¹ DCW h⁻¹, the C1 assimilation pathways that U. isabellina possesses led to incorporation of 8% formate carbon into fungal biomass, while the remaining formate was oxidized as the energy source. The simulation results theoretically conclude the potential of formate as a carbon and energy source to support fungal growth. Although genetically modified organisms have been developed to use C1 feedstock, these engineered platform shows poor C1 utilizations (consumes less than 0.5 g L⁻¹ C1 substrate) and slow fermentation processes [18–20]. U. isabellina offers an alternative non-model platform, and future metabolic engineering efforts should focus on development of genetic tools to re-program its metabolisms to produce diverse products by co-utilization of formate with other organic substrates.

Fermentation medium	Glucose uptake (mmol g−1 DCW h−1)		Formate uptake (mmol g−1 DCW h−1)		Biomass accumulation (mmol g−1 DCW h−1)
	Glucose & formate medium	Glucose medium	Glucose & formate medium	Glucose medium	Glucose & formate medium	Glucose medium
3-L fermenter	0.609	0.576	0.499	–	0.054	0.031
3-L fermenter	0.505	0.577	0.816	–	0.027	0.037
0.7-L flask	0.484	0.397	0.345	–	0.012	0.010
0.7-L flask	0.026	–	0.117	–	0.017	–