Myceliophthora thermophila ATCC 42464 was obtained from the ATCC. All strains used in this study are listed in Supplementary Table 1. The strains were grown on Vogel’s minimal medium supplemented with 2% (w/v) sucrose (MM medium) at 35°C for 10–15 days to obtain mature conidia. Antibiotics were added when needed to screen for transformants. For flask cultures, conidia of M. thermophila were inoculated into 100 mL fermentation medium to a final concentration of 2 × 10⁵ conidia/mL in a 250-mL Erlenmeyer flask. The fermentation medium (per liter) consisted of 75 g glucose, 10 g yeast extract, 0.15 g KH₂PO₄, 0.15 g K₂HPO₄, 0.15 g MgSO₄⋅7H₂O, 0.1 g CaCl₂⋅2H₂O, 1 mL biotin (0.1 g/L), and 1 mL trace elements (Vogel’s salts). Escherichia coli strains DH5α and BL21(DE3) were used for plasmid amplification and protein expression, respectively, and were cultured in Luria–Bertani (LB) medium supplemented with antibiotics as necessary.

The primers used in this study are listed in Supplementary Table 2.

For constructing the vector-carrying donor DNA for disrupting pfk2, the 5′- and 3′-flanking fragments of pfk2 were amplified from M. thermophila genome. PtrpC-bar from the p0380-bar plasmid and PtrpC-neo from the p0380-neo plasmid were each cloned using primers and fused with pfk2-up and pfk2-down, respectively. The fragments 5′-pfk2-a-up-Bar-down-3′ and 5′-pfk2-b/c-up-Neo-down-3′ were created by overlapping PCR and cloned into the pJET1.2/blunt cloning vector.

To select for specific single-guided RNAs (sgRNAs) targeting pfk2-a, pfk2-b, and pfk2-c, we used the sgRNA-Cas9 tool to identify sgRNA target sites with high scores (Shengsong et al., 2014). Then, a target-directed M. thermophila U6 promoter-driven sgRNA was created by overlapping PCR and cloned into the pJET1.2/blunt cloning vector, giving the corresponding plasmids U6-pfk2-a-sgRNA, U6- pfk2-b-sgRNA, and U6- pfk2-c-sgRNA. The Cas9-expression PCR cassette Ptef1-Cas9-TtprC was amplified using Ptef-cas-F/TtprC-cas-R from the plasmid p0380-bar-Ptef1-Cas9-TtprC (Liu et al., 2017).

For overexpression of pfk2-a/b/c, fragments of pfk2-a/b/c were amplified from M. thermophila DNA using the primers OEpfk2-a/b/c-F and OEpfk2-a/b/c-R and assembled into BcuI/BamHI-digested pAN52-MtPgpdA-TtprC-Neo using the NEB Gibson assembly kit (New England Biolabs, Ipswich, MA, USA). The open reading frame of pfk1 of Neurospora crassa (NCU202198) was amplified from cDNA of N. crassa using the primers Ncpfk1-F/Ncpfk1-R and assembled into BcuI/BamHI-digested pAN52-MtPgpdA-TtprC using the NEB Gibson assembly kit. The pAN52-PtrpC-hph-Ptef-Aomae-PAngpdA-Aopyc plasmid construction was the same as that described previously (Li et al., 2019). GsalaD from Geobacillus stearothermophilus and Scgpd1 from S. cerevisiae were cloned into pAN52-PgpdA-Neo to generate the plasmids pAN52-PgpdA-GsalaD-Neo and pAN52-PgpdA-Scgpd1-Neo, respectively, using the NEB Gibson assembly kit.

All vectors were constructed using E. coli DH5α and all constructed plasmids were verified by DNA sequencing.

Transformation of M. thermophila protoplasts was performed using a previously described procedure (Liu et al., 2017). For target gene expression, 10 μg plasmid was added to the fungal protoplasts (Wang et al., 2015). For gene disruption, sgRNA, the donor expression cassette for disruption, and the Cas9-expression PCR cassette were mixed and co-transformed into the WT strain. Colonies were screened for bar or neo gene resistance using phosphinothricin (150 μg/mL) or geneticin (80 μg/mL), followed by sequential identification via PCR with paired primers. Gene disruption by CRISPR/Cas9 was performed as described previously (Liu et al., 2017).

To compare growth of WT and other strains in solid medium, conidia of M. thermophila were collected from each strain, filtered and adjusted to 2.5 × 10⁷ conidia/mL. A 2-μL aliquot of the suspension was plated on fermentation medium supplemented with 7.5% glucose at 45°C for 3 days. To evaluate glucose consumption in each strain, 100 mL of medium was inoculated with mature spores (2.5 × 10⁵ spores/mL) and batch-cultured in a 250-mL Erlenmeyer flask. The culture was incubated at 45°C on a rotary shaker at 150 rpm. Samples (5 mL) were taken at different intervals. The fungal biomass dry weights were estimated at 8 days after inoculation of the Erlenmeyer flask. Fungal mycelial biomass was then harvested, dried, and weighed.

The cultures were centrifuged and the resulting supernatants were filtered through a 0.22-μm filter (Corning, NY, USA). HPLC was used to detect and quantify the glucose, glycerin, and malate acid using an e2695 instrument (Waters, Manchester, UK) and an Aminex HPX-87H column (Bio-Rad, Hercules, CA, USA) at 45°C. The mobile phase (5 mM H₂SO₄) was set at a flow rate of 0.5 mL min^–1 and the compounds detected with a Waters 410 refractive index detector (Waters Corporation, Milford, MA, USA).

Escherichia coli Rosetta (DE3) transformed with PFK2A-pET32a was grown in LB at 37°C to an optical density at 600 nm of 0.6, and then protein expression was induced with 1.4 mM isopropyl β-D-thiogalactoside (IPTG) at 16°C for 15 h. Expression of PFK2B/C-pET28a was induced with 0.4 mM IPTG. Cells were lysed at 4°C using Scientz-IID ultrasonic processors (SCIENTZ, Zhejiang, China) in phosphate-buffered saline (PBS), and the cell debris was cleared by centrifugation at 4°C. Cleared lysates were applied to a 1-mL His-Trap column (GE Healthcare, Pittsburgh, USA) and purified by an AKTA avant150 (GE Healthcare) with a 50–500 mM gradient of imidazole buffer. The purified protein was detected following sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Protein concentrations were determined using a Bio-Rad Protein Assay kit (Bio-Rad, Hercules, CA, USA).

PFK2 activity was measured as described previously except that the protein concentration used was 15 μg (Kurland et al., 2000; Manes and El-Maghrabi, 2005). The fru-2,6-P₂ produced was then quantified by measuring potato inorganic pyrophosphate-PFK activation spectrophotometrically as described previously, except that the reaction was scaled down for use on a 96-well plate (Schaftingen et al., 1982). FBPase-2 activity was measured by incubating 15 μg protein for 60 min at 30°C in assay buffer at pH 7 with 1 mM ATP, 2 mM MgCl₂, 5 mM KPi, and 50 μM fru-2,6-P₂. Reactions were halted by adding NaOH to a final concentration of 100 mM (Kurland et al., 2000; Manes and El-Maghrabi, 2005). Concentrations of the fru-2,6-P₂ product were determined using the liquid chromatography-tandem mass spectrometry (LC–MS/MS) platform.

Activity of PFK1 in the M. thermophila mycelial cultures was determined using the PFK1 Assay Kit (Solarbio, Beijing, China).

Myceliophthora thermophila strains were used to inoculate fermentation medium containing a 7.5% carbon source and cultured at 45°C. Mycelia were collected at 2 and 4 days of culture, immediately homogenized in liquid nitrogen, and then stored at −80°C. Total RNA was extracted as described previously (Xu et al., 2018). Total RNA was extracted from mycelia through mechanical disruption with mortar and pestle, digested with DNase1, and purified using the Qiagen RNeasy Mini kit (Qiagen, Hilden, Germany). A cDNA library was then synthesized and sequenced using the DNBSEQ platform at BGI (Shenzhen, China). All data in this study were generated by sequencing two independent biological replicates.

The qRT-PCR analysis was performed using the iScript cDNA Synthesis Kit, the SYBR Green Realtime PCR Master Mix (TOYOBO, Osaka, Japan), and the CFX96 Real-Time PCR Detection System (Bio-Rad). The qRT-PCR was performed as follows: 95°C for 30 s, 40 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 30 s. A melt curve analysis was performed at the end of each run from 55 to 95°C with a ramp speed of 0.5°C to ensure specific sequence amplification of all primers and only one melting temperature on the melting curve. Each reaction was performed in quadruplicate. The actin gene (Mycth_2314852) was used as an endogenous control for all experiments. The transcription level of each gene was estimated using the 2^–ΔΔCt method (Livak and Schmittgen, 2013). The ratio of the transcription level in the mutant to that in the WT control was calculated as the relative transcription level.

The samples for intracellular metabolite profiling were prepared and analyzed using the standardized and improved LC–MS/MS metabolomics methodology (Xiaomei et al., 2019). Briefly, 20 mL (±0.5 mL) of M. thermophila mycelial culture was fast filtered through a –20°C precooled vacuum filter. After washing with 25 mL precooled PBS buffer, the mycelial samples were flash frozen into liquid nitrogen. After pulverization by pestle and mortar in liquid nitrogen, about 0.1 g (±0.01 g) of sample was suspended in 1 mL precooled 50% methanol and incubated at 4°C for 15 min. Exact weight was determined by weighing each tube before and after adding the samples. After placing on ice for 5 min, the mixture was centrifuged at 12,000 g at 4°C for 15 min to remove cell debris, and the supernatant was collected into a precooled tube (Zhang et al., 2020). The supernatant was then allowed to evaporate under vacuum using a centrifugal vacuum concentrator. The dried metabolites were dissolved in 20 μL Milli-Q water and then analyzed by the LC–MS/MS platform comprising an ultra-performance liquid chromatography (UPLC) 30A system (Shimadzu, Kyoto, Japan) and a TripleTOF™ 6600 mass spectrometer (Applied Biosystem Sciex, Framingham, MA, USA), as described previously. The LC–MS/MS data were normalized by sample weight, and then metabolites involved in central metabolism were identified using the protocol described (Xiaomei et al., 2019).

One-tailed homoscedastic (equal variance) t-test was used for adjusting statistical significance. n.s., no statistical significance; *p < 0.05; **p < 0.01; and ***p < 0.001.

In Ascomycetes, most species have three isoforms of PFK2/FBPase-2 (Sellem et al., 2019), as does M. thermophila. These three isoforms are encoded by Mycth_71484 (pfk2-a), Mycth_2301421 (pfk2-b), and Mycth_2312639 (pfk2-c). Sequence analysis of the corresponding proteins, as shown in Supplementary Figure 1, revealed that PFK2-A, PFK2-B and PFK2-C each contain a PFK2 domain and a FBPase-2 domain. Next, we analyzed the expression patterns of the three genes under multiple carbon source conditions. Generally, the expression level of pfk2-a was the highest among the three genes, while those of pfk2-b and pfk2-c were comparable, consistent with previous transcriptome results (Li et al., 2019, 2020). Comparing different carbon sources, the gene expression levels were similar under glucose, xylose and sucrose, and lower than those under cellobiose, avicel, and starch (Figure 1A).

To gain further insights into the functions of PFK2/FBPase-2 in this thermophilic fungus, cDNAs encoding the proteins were expressed in the E. coli strain BL21(DE3) (Figure 1B). The results of enzyme activity determination are summarized in Table 1. Genes pfk2-a and pfk2-c were shown to encode monofunctional FBPase-2 and monofunctional PFK2 activities, respectively. However, for pfk2-b, neither FBPase-2 nor PFK2 activity was detected, even at high concentration. These results indicated that PFK2/FBPase-2 is encoded by monofunctional genes in filamentous fungi, which is different from mammals and plants. Separated FBPase-2 and PFK2 activities were also found in S. cerevisiae, which has been attributed to crucial substitutions or major deletions in one of the two protein domains that occurred during evolution, rendering them monofunctional.

To learn more about the function of PFK2/FBPase-2, single, double, and triple gene deletion knockouts of pfk2-a/b/c were performed using the CRISPR/Cas9 system in WT. Each of the single-knockout mutants were able to survive, indicating that all three genes were non-essential genes, but the phenotypes of these knockout strains varied. On solid media, the Δpfk2-a and Δpfk-b mutants grew as fast as WT, but the growth of Δpfk2-c was slower than WT which was consistent with its PFK2 function. All of the double-knockout mutants grew slower than the WT strain (Figure 2A), indicating that their combined effect is to promote growth. Despite the observation that pfk-b did not encode a protein with detectable enzymatic activity, the strains carrying a double knockout of pfk-a and pfk-b grew more slowly than the Δpfk-a strain, suggesting that PFK2-B may indeed have PFK2 activity. Complete deletion of all three genes also resulted in slower growth compared with the WT strain. To further analyze the effect of each gene on growth, the same comparisons were made in liquid media. Interestingly, we observed that Δpfk2-a had a lower biomass, but a higher glucose consumption rate, than WT (Figures 2B,C). The biomass of Δpfk2-a decreased from 18.17 to 16.30 g/L of dry cell weight, 10.3% lower than that of WT, and glucose consumption increased to a rate 22.15% higher than that of WT. Furthermore, the glucose consumption rate/dry cell weight was 36.18% higher than that of WT. Taken together, the deletion of pfk2-a appeared to be beneficial for improving the rate of glucose metabolism. Another notable observation of the Δpfk2-abc mutant was that, compared with the WT strain, the biomass was reduced and the glucose consumption rate was 12.11% lower (Figures 2B,C). This was consistent with the phenotype on solid media, indicating that simultaneous deletion of all three genes affects the growth and metabolism of M. thermophila.

We further investigated the functions of all three genes using an overexpression system wherein pfk2-a, pfk2-b and pfk2-c were placed under the strong constitutive promoter of gpdA in WT. No obvious growth differences were seen between the pfk2-b and pfk2-c overexpression (OEpfk2-b and OEpfk2-c) and WT strains (Figure 2D). However, the pfk2-a overexpression strain (OEpfk2-a) grew more slowly than WT on solid media plates, and had 12.88% less biomass than WT in liquid media (Figures 2D,E). This was likely due to overexpression of FBPase-2 activity, which catalyzes the degradation of fru-2,6-P₂, thereby reducing the rate of glycolysis, this might have contribution to the decreasing of biomass production. By contrast, the pfk2-c overexpression strain showed no difference in biomass compared to WT, but the glucose consumption rate/unit biomass was 15.5% higher than that of WT in liquid medium (Figure 2F). This was consistent with its PFK2 activity, which catalyzes the synthesis of fru-2,6-P₂, thereby enhancing the rate of glycolysis. Unlike the other two genes, the pfk2-b overexpression strain showed no significant differences from the WT strain in either liquid or solid media, possibly because it does not possess substantial PFK2 or FBPase-2 activity.

Among the various pfk2-a/b/c mutants, the Δpfk2-a and Δpfk2-abc strains showed significant phenotypic changes. Compared with WT, Δpfk2-a exhibited a significantly accelerated rate of glucose consumption, while Δpfk2-abc showed a lower rate. This intriguing result prompted us to undertake a quantitative examination of the changes in intracellular metabolism in the mutants. The pfk2-a gene, which encodes FBPase-2, plays an essential role in fru-2,6-P₂ degradation; indeed, the triple-knockout mutant completely lost the ability to synthesize and degrade fru-2,6-P₂. Samples were taken for metabolomics analysis during fermentation at days 2–4, and the relative level of each metabolite in the Δpfk2-a and Δpfk2-abc strains was calculated by dividing its normalized peak area by that in the WT strain.

On day 2, there were no significant differences in intracellular metabolite content among the three strains (except for some pentose in the pentose phosphate pathway [PPP] of Δpfk2-a) (Figure 3A). On day 3, the concentrations of metabolites in the Embden Meyerhof Parnas (EMP) pathway in Δpfk2-a were higher than those in WT. These metabolites included fructose-1,6-diphosphate (FBP), 2/3-phosphoglycerate (2PG/3PG), and especially phosphoenolpyruvate (PEP), whose concentration was 3.45-fold higher in Δpfk2-a than in WT. Among the tricarboxylic acid (TCA) cycle intermediates, citrate/isocitrate, 2-oxoglutarate, succinate, fumarate, and malate were 3. 34-, 2. 18-, 2. 86-, 1. 72-, and 2.32-fold higher in Δpfk2-a than in WT, respectively. Intermediates of the PPP were also higher in Δpfk2-a than in WT (Figure 3B). The content of most amino acids was also higher in Δpfk2-a compared with WT (Supplementary Figure 2). The above results showed that, in the absence of FBPase-2 activity, the degradation of fru-2,6-P₂ is reduced, thereby promoting glycolytic metabolism, which in turn promotes the metabolism of TCA. We suspect that the increase in amino acid content may have been due to the vigorous glycolysis and metabolism of TCA, which requires more protein synthesis in cells. The synthesis of amino acids promotes the consumption of intermediate metabolites and NADPH oxidation in the PPP and, coupled with the enhancement in TCA metabolism, promotes the metabolism of the PPP. Similar to day 3, increased metabolism of the above pathways was also observed in Δpfk2-a on day 4 (Figure 3C).

By contrast, the levels of most intracellular intermediates of the PPP, EMP pathway and TCA cycle were lower in Δpfk2-abc mutants than in WT (Figures 3A–C), consistent with the rate of glucose consumption. In the triple knockout, the strain lost the ability to synthesize fru-2,6-P₂, and without the activation of fru-2,6-P₂, the rate of glycolysis was reduced and the metabolism of other downstream pathways also slowed down. These results indicated that PFK2/FBPase-2 plays an important role in the metabolism and growth of M. thermophila, helping fungi better compete for nutrients and grow faster in nutrient-rich environments. Complete deletion of its coding genes reduces the metabolism of carbon and growth rate.

To determine whether PFK2 in filamentous fungi promotes glycolysis by activating PFK1, as previous research has shown in mammals and yeast (Bolesm et al., 2010; Ros and Schulze, 2013), we measured the enzymatic activities of PFK1 in the WT and mutant strains. We found that the PFK1 activity in Δpfk2-a was significantly higher than that in WT, increasing by 119, 46, and 60% on days 2, 3, and 4, respectively (Figure 4A). We speculate that the deletion of pfk2-a promotes glycolytic metabolism in response to the loss of FBPase-2 activity; with only PFK2 activity remaining, PFK1 is activated.

To verify the above speculation, we overexpressed pfk1 from N. crassa in WT (OENcpfk), and found that the overexpression strain had a faster glucose consumption rate than the WT strain (Figure 4B). These results indicated that increasing PFK1 activity promoted glucose consumption in M. thermophila, further confirming that the promotional effect of Δpfk2-a is exerted via activation of PFK1.

One response to rapid metabolism is a significant increase in the expression of alternative oxidase (AOX) in the mitochondrial electron transport chain (Figure 4C). This relaxes the tension in the highly coupled electron transport process in mitochondria, thus providing and maintaining metabolic homeostasis (Saha et al., 2016). We further estimated the amount of AOX from growth resistance of mycelial cultures to salicylhydroxamic acid (SHAM), a specific inhibitor of the alternative respiratory pathway. The correlation between a strong constitutive induction of the alternative pathway and the suppressor effect was also assessed at the protein level showing an increased resistance to SHAM. The growth on solid media shown in Figure 4D demonstrated that knockout of pfk2-a (encoding FBPase-2) led to higher expression of aox.

The above results showed that knockout of pfk2-a promoted the metabolic rate. To verify whether knocking out pfk2-a could improve metabolites production, the production of metabolites was increased in the Δpfk2-a strain, the glycerol and malate were randomly selected for validation (Figure 5A). We created a pfk2-a knockout in a malate-producing strain that was engineered via heterologous overexpression of the malic acid exporter (mae) and pyruvate carboxylase (pyc) gene from Aspergillus oryzae (strain MAL1). The resulting strain, MAL2, produced 52.834 g/L malate acid, a 36.2% increase over that of its parental strain (Figure 5B). Similarly, we introduced glyceraldehyde-3-phosphate dehydrogenase (gpd1) from S. cerevisiae (strain GLY1) driven by the strong constitutive promoter of gpdA into WT, to increase glycerol production. We then created a pfk2-a knockout in the overexpression strain (GLY2) using the CRISPR/Cas9 system. As shown in Figure 5C, the titer of glycerol increased from 0.027 g/L in GLY1 strain to 0.056 g/L in the pfk2-a deletion mutant (strain GLY2), a 107% difference. These data suggested that knockout of pfk2-a may be developed as a strategy to engineer increases in the yield of metabolites and rate of metabolism in M. thermophila.