Introduction

During their life cycle, plant pathogenic fungi transit between parasitic and saprophytic stages, necessitating adaptation to changing carbon sources1. Carbon catabolite repression (CCR) prioritizes the utilization of preferred carbon sources like glucose, while non-preferred sources, such as lipids and plant cell wall components, are repressed2,3,4,5. When preferred sources are depleted, carbon catabolite de-repression (CCDR) activates the utilization of non-preferred sources2,5,6,7. These processes are critical for fungal pathogenicity8,9,10,11.

In fungi, CreA (or its homolog Mig1) is a key transcription factor in CCR12,13. In Saccharomyces cerevisiae, Mig1 is dephosphorylated by the Reg1-Glc7 phosphatase complex, allowing nuclear localization to repress glucose-repressed genes under glucose-rich conditions14. Under glucose starvation, Mig1 is phosphorylated by the Snf1 kinase complex, facilitating its nuclear export15,16. However, this mechanism alone does not fully explain glucose de-repression in yeast16. Snf1, an AMP-activated protein kinase (AMPK), is itself activated via phosphorylation by Sak1, Tos3, or Elm1 and deactivated by Reg1-Glc717,18,19. In Aspergillus nidulans, the phosphorylation of CreA is crucial for nuclear localization and DNA binding20. The divergence in phosphorylation-driven nuclear export of Mig1/CreA between yeasts and filamentous fungi underscores the complexity of CCR and CCDR regulation, which remains poorly characterized in filamentous fungi.

Magnaporthe oryzae (synonym Pyricularia oryzae), a devastating rice pathogen, causes 10–30% yield losses annually21. Controlled utilization of lipid droplets, trehalose, glycogen granules, and other carbon sources is critical for its pathogenicity10,22,23,24. CCR and CCDR influence its infection process10,21,25,26. Previous studies revealed that CreA negatively and Crf1 positively regulate lipid and l-arabinose catabolism10,11, with deletions of CREA or CRF1 reducing virulence10,27. Despite its conserved protein structure across fungi, Crf1 (carbon metabolism-regulatory factor 1) has been rarely studied in filamentous fungi. In S. cerevisiae, the closest homologous protein of Crf1 is Rtg3 (E-value is 3e−04), while no protein with the same function has been reported.

The serine/threonine protein phosphatase 4 (Pp4) regulatory subunit Smek1 is essential for pathogenicity in M. oryzae by dephosphorylating CreA and Crf1 during CCR and CCDR11. Phosphorylated CreA represses non-preferred carbon source utilization (such as lipids and l-arabinose), while hypophosphorylated Crf1 promotes it. To the best of our knowledge, this Smek1-mediated regulation is the first such report in filamentous fungi. In mammals, Ppp4r3 of the Pp4 complex (Pp4c, Ppp4r2, and Ppp4r3 subunits) has two isoforms Pp4r3a (Smek1) and Pp4r3b (Smek2)28. In the human liver cell, Smek1, combined with two subunits (Pp4c and Pp4r2), dephosphorylates CreB-regulated transcription coactivator 2 (Crtc2) to regulate glucose metabolism29. In S. cerevisiae, the Pph3-Psy2 phosphatase complex (homologous to Pp4c-Smek1) dephosphorylates Mth1 (an Rgt1-associated corepressor) to repress glucose transporter gene transcription30. The functions of Pp4c (or Pph3) and Smek1 (or Psy2) in DNA damage repair have also been reported in mammals or yeast31,32,33,34. However, the role of Pp4c in M. oryzae remains unexplored.

In M. oryzae, the dephosphorylation of CreA and Crf1 during CCR and CCDR is mediated by Smek111. However, the kinase responsible for the phosphorylation of CreA and Crf1 remains unidentified. In S. cerevisiae, Mig1 is directly phosphorylated by Snf115, but direct regulation of CreA by Snf1 in filamentous fungi has not been established. Previous studies from our group revealed that the Snf1 kinase complex, along with its upstream kinases Sak1 and Tos1, is essential for lipid catabolism and virulence in M. oryzae35. Additionally, Yi et al. demonstrated that Snf1 is critical for utilizing plant cell wall-derived carbon sources in M. oryzae36. Snf1's role in non-preferred carbon source utilization and pathogenicity has been similarly highlighted in other plant pathogenic fungi37,38,39,40. However, whether Snf1 regulates lipid catabolism and other non-preferred carbon source utilization through CreA and Crf1 in pathogenic fungi remains unclear.

Here, we identified a regulatory network in M. oryzae that orchestrates CCR and CCDR, involving two subunits of protein phosphatase 4 (Pp4c and Smek1), Snf1 kinase, the transcriptional repressor CreA, and the activator Crf1. Under CCR conditions, Snf1 and Smek1 directly coordinate the phosphorylation status of CreA and Crf1, resulting in CreA activation and Crf1 inactivation, thereby repressing the catabolism of non-preferred carbon sources. In contrast, under CCDR conditions, Snf1 indirectly facilitates the dephosphorylation of CreA and Crf1 through Pp4c and Smek1, leading to CreA inactivation and Crf1 activation, which promotes non-preferred carbon source utilization. Our findings also establish that Pp4c is indispensable for lipid and carbohydrate metabolism, fungal growth, development, and pathogenicity in M. oryzae. These results provide critical insights into the regulatory mechanisms underlying CCR and CCDR and highlight their pivotal roles in the pathogenicity of M. oryzae and other plant pathogenic fungi.

Results

The protein phosphatase catalytic subunit Pp4c interacts with Smek1, Crf1, and CreA in M. oryzae

Previous studies indicated that Smek1 is involved in carbon metabolism and pathogenicity in plants by mediating the dephosphorylation of the transcriptional activator Crf1 and the repressor CreA11. Crf1 (Supplementary Fig. 1) and CreA are conserved proteins in filamentous fungi5. In mammals, Smek1 (Pp4r3a) functions as a regulatory subunit of the Pp4 complex and typically associates with catalytic subunits Pp4c and Pp4r2 to form holoenzyme complexes. To identify Pp4c in M. oryzae, we employed a yeast two-hybrid (Y2H) assay to screen Smek1-interacting proteins. Among eight tested protein phosphatase catalytic subunits, Smek1 interacted with five proteins: MGG_01528 (Pp4c), MGG_03911 (Ppe1), Ppg1 (MGG_01690), MGG_06099 (Pph2), and Cna1 (MGG_07456) (Supplementary Fig. 2). Phylogenetic analysis revealed that MGG_01528 is closely homologous to human Pp4c and yeast Pph3, distinguishing it from other phosphatase subunits tested (Supplementary Fig. 3). Further analysis indicated that homologs of Pp4c are conserved across fungi (Supplementary Fig. 4).

We confirmed the interaction between Smek1 and Pp4c in vitro and in vivo through Y2H, GST pull-down, and co-immunoprecipitation (Co-IP) assays (Fig. 1a–c). To investigate whether Pp4c interacts with CreA and Crf1, as Smek1 does11, we also performed Y2H, GST pull-down, and Co-IP assays. Y2H results demonstrated a strong interaction between Pp4c and Crf1, while its interaction with CreA was weaker (Fig. 1a). Co-IP assays showed that GFP-Crf1 precipitated Pp4c-Flag, confirming the interaction between Pp4c and Crf1 in vivo (Fig. 1c). GST pull-down assays also indicated a direct interaction between Pp4c and CreA (Fig. 1d). Collectively, these findings suggest that Pp4c and Smek1 form a PP4 protein phosphatase complex that interacts with the transcription factors Crf1 and CreA in M. oryzae.

Fig. 1: Pp4c interacts with Smek1, CreA, and Crf1 in Magnaporthe oryzae.
figure 1

a Yeast two-hybrid assays between Pp4c and Smek1, CreA, and Crf1. Yeast strains carrying prey and bait vectors were cultured on SD-Leu-Trp and SD-Leu/Trp/His/Ade media at 30 °C for 4 days. Strains containing pGADT7-T and pGBKT7-53 were used as positive controls. b GST pull-down assays detecting Pp4c-Smek1 interactions. GST-tagged fusion proteins were expressed in Escherichia coli BL21 and pulled down using anti-GST beads. Eluted proteins were detected with anti-GST and anti-Flag antibodies. c Co-immunoprecipitation assays to detect Pp4c interactions with Smek1 and Crf1. Total proteins extracted from strains expressing PP4C-FLAG and GFP-CRF1 or PP4C-FLAG and GFP-SMEK1 were precipitated with anti-GFP beads. Eluted proteins were detected using anti-GFP and anti-Flag antibodies. d GST pull-down assays detecting Pp4c-CreA interactions.

Full size image

The protein kinase Snf1 interacts with Crf1 and CreA in M. oryzae

Previous work demonstrated that deletion of SNF1 or CRF1 results in defects in lipid utilization and loss of virulence in M. oryzae10,35. Additionally, Crf1 and CreA co-regulate lipid utilization10,41. However, the relationship of Snf1 to Crf1 or CreA in lipid catabolism and pathogenicity has not been determined. Here, we tested their protein interactions using Y2H, GST pull-down, and Co-IP assays. Interactions between Snf1 and CreA were confirmed via all three methods (Fig. 2a–c), while interactions between Snf1 and Crf1 were confirmed by GST pull-down and Co-IP assays (Fig. 2d, e). These results suggest that Snf1 directly regulates Crf1 and CreA to control lipid catabolism and pathogenic processes in M. oryzae.

Fig. 2: Snf1 interacts with CreA and Crf1 in Magnaporthe oryzae.
figure 2

a Yeast two-hybrid assays between Snf1, CreA, and Crf1. GST pull-down (b) and co-immunoprecipitation (c) assays detecting Snf1–CreA interactions. GST pull-down (d) and co-immunoprecipitation (e) assays detecting Snf1–Crf1 interactions. f Relative expression levels at different developmental stages in the wild-type. Spores were induced to form appressoria on hydrophobic membranes for 4, 6, 8, 10, and 24 h, while mycelia were cultured in liquid CM for 2 days. H3 and α-ACTIN were used as reference genes. Error bars represent SD. n = 5 biologically independent samples.

Full size image

To assess the temporal expression of SNF1, PP4C, SMEK1, CRF1, and CREA during fungal development, we conducted reverse transcription-quantitative PCR (RT-qPCR) on samples from appressoria (collected at 4, 6, 8, 10, and 24 h post-inoculation [hpi]) and mycelia. The results showed similar expression patterns for these genes, with peak expression observed in appressoria at 8 and 24 hpi (Fig. 2f). The 8-hpi time point corresponds to appressorium maturation, characterized by carbon source degradation, synthesis of a dense melanin layer in the cell wall, and glycerol accumulation. The 24-hpi time point marks plant cell invasion, during which fungal cells encounter drastic changes in available carbon sources11,42. Combined with the evidence that Snf1, Pp4c, and Smek1 interact with Crf1 and CreA, these findings suggest that these five proteins coordinate the carbon metabolism in M. oryzae, supporting fungal development and pathogenicity in vivo.

Pp4c is required for development and pathogenicity in M. oryzae

To determine whether PP4C is as critical as SMEK1, SNF1, CRF1, and CREA for the development and pathogenicity of M. oryzae10,11,27,35,36, we investigated its biological function via gene knockout and phenotypic assays. The Δpp4c mutant exhibited significantly slower growth (69.52 ± 1.79% of the wild-type) and produced fewer conidiophores and spores (19.00 ± 7.81% of the wild-type) on complete medium (CM) (Fig. 3a–d). Additionally, the conidial germination rate at 4 hpi was reduced in the Δpp4c mutant compared to the wild-type (Fig. 3e), and the differentiation of appressoria from germ tubes was delayed. At 2 hpi, wild-type spores developed appressoria on hydrophobic surfaces, whereas Δpp4c spores did not (Fig. 3f).

Fig. 3: Pp4c is required for development and virulence in Magnaporthe oryzae.
figure 3

Colonies (a) and conidiophores (b) of the wild-type, Δpp4c, and PP4C-complemented strain of Δpp4c (pp4c-c). Scale bar = 10 mm (a) and scale bar = 50 μm (b). c Colony diameters of the wild-type, Δpp4c, and pp4c-c. d Conidiation of the wild-type, Δpp4c, and pp4c-c. Strains were cultured on CM at 25 °C for 8 days (c, d). e Conidial germination rates on hydrophobic surfaces under dark conditions at 22 °C for 4 h. f Spores and developing appressoria of the wild-type, Δpp4c, and pp4c-c strains incubated on hydrophobic membranes at 22 °C for 2, 4, 6, 8, and 24 h. Scale bar = 20 μm. g Virulence assays of the wild-type, Δpp4c, and pp4c-c on barley. Scale bar = 10 mm. Virulence assays on rice: representative images (h), and disease index (%) (i). Scale bar = 10 mm. j Invasive growth of the wild-type and Δpp4c on barley leaves. Scale bar = 10 μm. k Penetration rates on barley leaves at 24 and 48 h post-inoculation (hpi). Error bars represent SD. Tukey's HSD test was used for significance testing: *p < 0.05; **p < 0.001. n = 3 biologically independent samples.

Full size image

The impact of Pp4c on fungal pathogenicity was assessed on barley and rice. On barley leaf explants inoculated with conidial suspensions (5.0 × 104 spores mL−1), the wild-type and complementary strains caused gray, spore-producing lesions, while Δpp4c only produced small brown spots, indicating severely compromised virulence (Fig. 3g). Similarly, on rice seedlings sprayed with conidial suspensions (5.0 × 104 spores mL−1), the Δpp4c mutant caused small, scattered lesions, whereas the wild-type and complementary strains caused extensive, fused lesions (Fig. 3h). The disease index for rice leaves infected by Δpp4c was 39.02 ± 11.9% of the wild-type (Fig. 3i).

To determine the stage of infection affected by Δpp4c, we analyzed appressorial penetration and invasive growth. At 24 and 48 hpi, Δpp4c showed significantly reduced penetration rates on barley cuticles compared to the wild-type (Fig. 3j, k). By 48 hpi, Δpp4c formed shorter invasive hyphae in the first rice cell, whereas the wild-type hyphae had penetrated the second cell (Fig. 3j). These findings suggest that reduced penetration ability and slowed invasive growth underlie the diminished virulence of Δpp4c.

Pp4c is involved in lipid and glycogen utilization in vivo

Efficient utilization of lipid droplets and glycogen granules stored in spores is essential for appressorium formation and generation of the turgor pressure (~ 8.0 MPa) required for plant cell invasion in M. oryzae43. We marked lipid droplets with Cap20-GFP and glycogen granules with I2 + KI, to observe their degradation during appressorium formation. In Δpp4c, glycogen degradation in spores and appressoria was delayed. At 8 hpi, nearly all wild-type spores lacked visible glycogen granules, while 65.74 ± 3.34% of Δpp4c spores retained them (Fig. 4a, b). Similarly, at 12 and 16 hpi, more appressoria of Δpp4c contained glycogen granules compared to the wild-type (Fig. 4a, b).

Fig. 4: Pp4c is involved in glycogen and lipid utilization in Magnaporthe oryzae.
figure 4

Glycogen granule visualization (a) and percentage (b) in conidia or appressoria of the wild-type and Δpp4c strains. Spores on hydrophobic surfaces were stained with KI + I2 and observed at 0, 4, 8, 12, 16, and 24 hpi. c Lipid droplets in conidia (0 hpi) or developing appressoria (4, 8, 12, 16, and 24 hpi) marked by Cap20-GFP. Scale bar = 20 μm. d Relative triglyceride utilization rates in the wild-type, ΔcreA, Δpp4c, and Δsnf1 strains were calculated as triglyceride content before and after 24 h of starvation in MM-C. Error bars represent SD. Tukey's HSD test was used for significance testing: *p < 0.05; **p < 0.001. n = 3 biologically independent samples.

Full size image

Lipid droplet degradation was also delayed in Δpp4c. At 12 hpi, Δpp4c spores retained more lipid droplets than the wild-type (Fig. 4c). To further investigate the role of Pp4c in lipid catabolism, we measured triglyceride content in mycelia of Δpp4c, ΔcreA, and Δsnf1 mutants under carbon starvation. The mutants and the wild-type were cultured in liquid CM for 24 h and then transferred to minimal medium without glucose (MM-C) for 24 h. Triglyceride utilization rates (triglyceride consumed within 24 h/initial triglyceride) differed significantly between strains. The order of triglyceride catabolic capacity was ΔcreA > wild-type > Δpp4c > Δsnf1 (Fig. 4d). This indicates that Pp4c and Snf1 enhance triglyceride catabolism in vivo, whereas CreA inhibits it in M. oryzae.

Pp4c and Snf1 are essential for the utilization of non-preferred carbon sources in vitro

To investigate the roles of Pp4c and Snf1 in lipid and carbohydrate catabolism, and to compare their functions with those of Smek1, CreA, and Crf1, we assessed the growth of the wild-type and six mutant strains (ΔcreA, Δcrf1, ΔcreAΔcrf1, Δsnf1, Δsmek1, and Δpp4c) on MM-C media containing 1% glucose, d-xylose, olive oil, l-arabinose, glycerol, or ethanol as the sole carbon source. With the exception of the wild-type and Δcrf1, the other strains (ΔcreA, ΔcreAΔcrf1, Δsnf1, Δsmek1, and Δpp4c) exhibited significantly slower growth on glucose and d-xylose media, indicating defects in glucose and d-xylose utilization (Supplementary Fig. 5a, b). Compared to the wild-type and ΔcreA, all other mutants (Δcrf1, ΔcreAΔcrf1, Δsnf1, Δsmek1, and Δpp4c) grew substantially more slowly on l-arabinose, olive oil, glycerol, and ethanol media, suggesting deficiencies in the utilization of these non-preferred carbon sources (Supplementary Fig. 5a, b). These findings indicate that Snf1, Pp4c, and Smek1 function similarly to CreA in regulating carbon catabolism when M. oryzae utilizes preferred carbon sources (glucose), and align more closely with Crf1 when the pathogen utilizes non-preferred sources, such as lipids, l-arabinose, glycerol, and ethanol.

CreA functions as a repressor, whereas Crf1 promotes the use of non-preferred carbon sources10,11. To further elucidate the relationship between Pp4c, Snf1, CreA, and Crf1, we compared the growth of Δpp4cΔcrf1, Δpp4cΔcreA, Δsnf1Δcrf1, and Δsnf1ΔcreA on olive oil or l-arabinose media (Fig. 5). Δpp4cΔcrf1 showed slower growth than Δpp4c and Δcrf1 on both olive oil and l-arabinose media, suggesting that Pp4c and Crf1 together positively regulate the utilization of these substrates. In contrast, Δpp4cΔcreA grew faster than both Δpp4c and ΔcreA on olive oil medium, indicating a synergistic negative effect of Pp4c and CreA on olive oil utilization. On l-arabinose medium, Δpp4cΔcreA grew faster than Δpp4c but slower than ΔcreA, suggesting opposing roles for Pp4c (positive regulation) and CreA (negative regulation) in l-arabinose utilization (Fig. 5). Δsnf1Δcrf1 exhibited slower growth than Δsnf1 and Δcrf1 on olive oil medium, and slower growth than Δcrf1 but comparable to Δsnf1 on l-arabinose medium, implying a synergistic effect of Snf1 and Crf1 on the utilization of both carbon sources. On the other hand, Δsnf1ΔcreA grew slower than ΔcreA but faster than Δsnf1 on both olive oil and l-arabinose media, suggesting that Snf1 and CreA exert opposing influences on the utilization of these substrates (with Snf1 positively regulating and CreA negatively regulating) (Fig. 5).

Fig. 5: Pp4c and Snf1 regulate triglyceride and l-arabinose utilization via CreA and Crf1 in Magnaporthe oryzae.
figure 5

a Colony morphology of the wild-type, ΔcreA, Δcrf1, Δpp4c, Δsnf1, Δpp4cΔcrf1, Δsnf1Δcrf1, Δpp4cΔcreA, and Δsnf1ΔcreA strains on MM (glucose) or MM-C with 1% olive oil or 1% l-arabinose at 25 °C for 12 days. Scale bar = 10 mm. b Relative growth rates on glucose, olive oil, and l-arabinose media compared to the wild-type. Error bars represent SD. Different lowercase letters indicate significant differences by Tukey's HSD test (p < 0.05). n = 5 biologically independent samples.

Full size image

These findings, together with the interactions between Pp4c and Smek1, CreA, or Crf1 (Fig. 1), between Snf1 and CreA or Crf1 (Fig. 2), and between Smek1 and CreA or Crf1 in our previous report11, strongly suggest that the kinase Snf1 and phosphatases (Pp4c and Smek1) regulate the utilization of both preferred and non-preferred carbon sources in M. oryzae via CreA and Crf1 transcription factors.

Pp4c, Snf1, and CreA are involved in regulating the expression of genes related to lipolysis, β-oxidation, and gluconeogenesis in M. oryzae

Given that Δpp4c, Δsnf1, and ΔcreA strains are impaired in lipid utilization, and that the roles of Pp4c, Snf1, and CreA in the regulation of lipid-related gene expression are unclear or not yet fully defined, we quantified the transcription levels of seven lipase/esterase genes, 17 genes involved in fatty acid catabolism and β-oxidation, and two key genes in gluconeogenesis (Supplementary Table 1) in Δpp4c, Δsnf1, and ΔcreA using RT-qPCR. Mycelia were first cultured in liquid CM for 48 h and then transferred to MM-C for another 24 h to reduce interference from endogenous lipid metabolism. The cultures were subsequently split and incubated in glucose or olive oil media for 6 h. Mycelia were then collected for gene expression analysis.

In ΔcreA, the relative transcription levels of five lipase/esterase genes (TGL1, TGL3, HDL1, VTL1, and LIH1), nine genes involved in fatty acid catabolism and β-oxidation (FAA2, FAT1, ECL1, MFP1, POT1, ECH2, LCAD1, PXA1, and PAX2), and the key gluconeogenesis genes ICL1 and MCL1 were significantly upregulated in glucose medium, while none of these genes were downregulated in olive oil medium (Fig. 6a–c). This suggests that CreA acts as a transcriptional repressor of genes involved in lipid catabolism and gluconeogenesis.

Fig. 6: Regulation of lipid utilization gene expression by Pp4c, Snf1, and CreA in Magnaporthe oryzae.
figure 6

Relative expression level of lipase genes (a) and genes involved in fatty acid activation and transport, and β-oxidation (b), ICL1, and MCL1 (c) of ΔcreA in glucose and olive oil media relative to the wild-type. Relative expression level of lipase genes (d) and genes involved in fatty acid activation and transport, and β-oxidation (e), ICL1, and MCL1 (f) of Δsnf1 in glucose and olive oil media relative to the wild-type. Relative expression level of lipase genes (g) and genes involved in fatty acid activation and transport, and β-oxidation (h), ICL1, and MCL1 (i) of Δpp4c in glucose and olive oil media relative to the wild-type. H3 and α-ACTIN were used as reference genes. Error bars represent SD. Asterisks “*” indicate significant differences from the wild-type (Tukey's HSD test, p < 0.05). n = 3 biologically independent samples. j Schematic diagram of lipid metabolism regulation by Smek1, Pp4c, Snf1, CreA, and Crf1 in M. oryzae. Data for Smek1 and Crf1 were obtained from previous studies10,11.

Full size image

In Δsnf1, the transcription levels of four lipase/esterase genes (TGL3, HDL1, VTL1, and FGL2), 11 genes in fatty acid catabolism and β-oxidation (FAA3, MFP1, POT1, POT3, ACAD11, ECH2, LCAD1, CRC1, PTH2, PXA1, and PAX2), and two gluconeogenesis-related genes (ICL1 and MCL1) were significantly downregulated in glucose medium, whereas 19 out of 26 tested genes showed no significant upregulation in olive oil medium (Fig. 6d–f), indicating that Snf1 serves as an activator for genes involved in lipid catabolism and gluconeogenesis.

In Δpp4c, four lipase/esterase genes (TGL3, HDL1, DGE1, and FGL2) and ICL1 were significantly downregulated in glucose medium, but none of these genes were upregulated in olive oil medium (Fig. 6g, i), suggesting that Pp4c activates lipolysis and gluconeogenesis. The effects of Pp4c on fatty acid catabolism and β-oxidation were more complex (Fig. 6h). Six genes (FAA2, ECL1, MFP1, LCAD1, ECH2, and CRC1) were upregulated, and six genes (FAA3, POT1, POT3, ACAD11, ECH2, and PTH2) were downregulated in Δpp4c on glucose medium, while five genes (FAA2, FAA3, POT3, LCAD1, and CRAT1) were upregulated and six genes (FAA1, FAT1, POT1, ACAD11, ECH2, and PTH2) were downregulated in olive oil medium (Fig. 6h).

When combined with previous studies on the regulation of lipid catabolism by Smek1 and Crf110,11, Fig. 6j provides a summary of how Pp4c, Smek1, Snf1, CreA, and Crf1 either inhibit or promote lipid catabolic processes in both glucose and lipid media (Fig. 6j).

Pp4c, Snf1, and CreA are involved in the regulation of gene expression related to arabinose metabolism in M. oryzae

Since the Δpp4c, Δsnf1, and ΔcreA mutants exhibited varying abilities to utilize l-arabinose, we quantified the transcription levels of genes involved in the metabolic pathways of l-arabinose catabolism (Supplementary Table 1) in these mutants using RT-qPCR. In ΔcreA, the expression levels of five metabolic enzyme genes (PRD1, LAD1, LXR1, TKL1, and XKI1) and two transcription factors (CRF1 and ARA1)—which regulate l-arabinose utilization—were significantly upregulated in l-arabinose medium compared to the wild-type (Fig. 7a). This suggests that CreA inhibits the transcription of genes involved in l-arabinose catabolism.

Fig. 7: Regulation of l-arabinose metabolism gene expression by Pp4c, Snf1, and CreA in Magnaporthe oryzae.
figure 7

Relative expression levels of l-arabinose metabolism genes in ΔcreA (a), Δsnf1 (b), and Δpp4c (c) strains under glucose and l-arabinose conditions compared to the wild-type. Error bars represent SD. Asterisks “*” indicate significant differences (Tukey's HSD test, p < 0.05). d Schematic diagram of l-arabinose metabolism regulation by Pp4c, Smek1, Snf1, CreA, and Crf1. Data for Smek1 and Crf1 were obtained from previous studies10,11. Relative expression levels of genes in Δpp4c (e), Δsnf1 (f), and Δcrf1 (g) strains compared to the wild-type. Error bars represent SD. Asterisks “*” indicate significant differences (Tukey's HSD test, p < 0.05). n= 3 biologically independent samples.

Full size image

In Δsnf1, the expression levels of ITR, PRD1, LAD1, LXR1, XKI1, and ARA1 were downregulated in l-arabinose medium, and ITR, LAD1, and ARA1 were downregulated in glucose medium, while no genes were significantly upregulated in either medium (Fig. 7b). These results indicate that Snf1 acts as a transcriptional activator of l-arabinose catabolic genes.

In Δpp4c, the expression levels of ITR, PRD1, LAD1, LXR1, TKL1, XKI1, and ARA1 were significantly downregulated in glucose medium, and PRD1 and LXR1 were downregulated in l-arabinose medium, suggesting that Pp4c also acts as a transcriptional activator of l-arabinose catabolism (Fig. 7c, d).

In our previous study, we noted that although Δcrf1 and Δsmek1 exhibited similar phenotypes in terms of carbon source utilization, the expression level of CRF1 in Δsmek1 was increased under CCDR conditions11. To determine if a similar feedback phenomenon occurs between Snf1, Pp4c, Crf1, and CreA, we analyzed how these regulators influence each other's expression. In the l-arabinose medium, the expression levels of CRF1 were significantly increased in both Δpp4c and Δsnf1 (Fig. 7e, f). Despite the elevated expression of CRF1 in these mutants, their phenotypes when utilizing non-preferred carbon sources (e.g., l-arabinose) were similar to those of the CRF1 deletion mutants (Fig. 5 and Supplementary Fig. 6), suggesting that Crf1 was not activated in Δpp4c, Δsnf1, and Δsmek1. In contrast, compared with the wild-type strain, the expression level of SNF1 in Δcrf1 was unchanged in glucose medium but significantly decreased in l-arabinose medium (Fig. 7e). This suggests that l-arabinose activates Crf1, which, in turn, promotes the expression of SNF1.

Pp4c and Snf1 regulate the phosphorylation status of Crf1 and CreA during CCR and CCDR

To further investigate the regulatory mechanisms of CCR and CCDR in M. oryzae, we used l-arabinose as a representative non-preferred carbon source. Our previous research confirmed that the phosphatase regulatory subunit Smek1 is required for the dephosphorylation of Crf1 during l-arabinose utilization (a condition for CCDR) and CreA during glucose utilization (a condition for CCR) in vivo11. To determine whether the phosphatase catalytic subunit Pp4c and the kinase Snf1 also regulate carbon source metabolism by altering the phosphorylation status of the transcription factors Crf1 and CreA, we assessed changes in the phosphorylation levels of Crf1 and CreA in Δsnf1 and Δpp4c by using Phos-tag SDS-PAGE.

In wild-type, CreA and Crf1 were dephosphorylated in l-arabinose medium compared to glucose medium (Fig. 8a). Deletion of SNF1 resulted in the blockage of phosphorylation of both CreA and Crf1, suggesting that Snf1 phosphorylates CreA and Crf1 in glucose medium in vivo (Fig. 8a). In l-arabinose medium, the phosphorylation levels of CreA and Crf1 in Δsnf1 were the same as those in Δpp4c, but higher than those in the wild-type, indicating that the dephosphorylation of CreA and Crf1 was blocked in both Δsnf1 and Δpp4c in l-arabinose medium (Fig. 8a). Similarly, the phosphorylation levels of Crf1 in Δsmek1 in l-arabinose medium were higher than those in the wild-type11. Δsnf1, Δpp4c, Δsmek1, and Δcrf1 mutants showed poor growth in l-arabinose medium (Fig. 5 and Supplementary Fig. 6).

Fig. 8: Snf1 dephosphorylates CreA and Crf1 in vivo in Magnaporthe oryzae.
figure 8

a Mn2+ or Zn2+ phos-tag assays showing the phosphorylation levels of CreA and Crf1 in the wild-type, Δpp4c, and Δsnf1 strains cultured in glucose (MM) and l-arabinose media. Strains were grown in liquid CM for 2 days and subsequently transferred to MM and l-arabinose media for 6 h. Higher bands on phos-tag gels indicate higher phosphorylation levels of CreA-Flag and Crf1-Flag. Normal SDS-PAGE confirmed the positions of target proteins. Proteins were detected using anti-Flag antibodies. b Schematic illustrating phosphorylation levels of CreA and Crf1 in the wild-type and mutant strains (Δpp4c, Δsmek1, and Δsnf1) under glucose and l-arabinose conditions. Data for Smek1 were retrieved from a previous report11. NR, no record.

Full size image

Furthermore, mass spectrometry analysis revealed that Crf1 from M. oryzae cultured in glucose medium exhibited a higher abundance of phosphorylated peptides and more phosphorylated amino acid sites compared to cultures in l-arabinose medium (Supplementary Table 2). This suggests that the dephosphorylation of Crf1 is required for l-arabinose utilization, while the dephosphorylation of CreA also contributes to the utilization of l-arabinose.

A detailed comparison of the phosphorylation levels of CreA and Crf1 in glucose and l-arabinose media revealed two phosphorylation states for CreA and three phosphorylation states for Crf1 in the wild-type, Δpp4c, and Δsnf1 (Fig. 8a). Based on our previous findings, in Δsmek1, CreA was phosphorylated at a higher level in glucose medium compared to the wild-type11, and in Δsnf1, CreA was phosphorylated at a lower level in glucose medium compared to the wild-type (Fig. 8a). This suggests the existence of at least three phosphorylation states across different sites of CreA in M. oryzae (CreA, P-CreA, and PP-CreA: the first to third CreA phosphorylation states) (Fig. 8b). The phosphorylation level of Crf1 in Δsmek1 was higher than that of the wild-type in both glucose and l-arabinose media11, while Crf1 in Δsnf1 was similarly phosphorylated to the wild-type in glucose medium but higher in l-arabinose medium (Fig. 8a). These observations suggest at least four phosphorylation states for Crf1 in M. oryzae (Crf1, P-Crf1, PP-Crf1, and PPP-Crf1: the first to fourth phosphorylation states of Crf1) (Fig. 8b).

Snf1 and Pp4 regulate each other's phosphorylation status

Compared to the wild-type, deletion of the SNF1 kinase gene led to dephosphorylation of CreA and Crf1 in glucose medium but to phosphorylation of CreA and Crf1 in arabinose medium (Fig. 8a). In glucose medium, Snf1 acts as a kinase to phosphorylate CreA and Crf1, while in l-arabinose medium, Snf1 behaves like a phosphatase, dephosphorylating CreA and Crf1 (Fig. 8a), which was unexpected. These results imply that Snf1 directly acts on CreA and Crf1 in glucose medium and indirectly regulates their phosphorylation in l-arabinose medium, likely through an unknown phosphatase.

To determine whether the Snf1 kinase regulates the dephosphorylation of CreA and Crf1 via the phosphatases Pp4c or Smek1 in l-arabinose medium, we analyzed the interactions between Snf1 and Pp4c or Smek1. The Y2H and GST pull-down assays revealed that Snf1 directly interacts with both Pp4c and Smek1 in vitro (Fig. 9a–c), and Co-IP assays confirmed that Snf1 interacts with both Pp4c and Smek1 in vivo when the fungus was cultured in glucose (Fig. 9d, e) and l-arabinose media (Supplementary Fig. 6). This suggests that Snf1 likely phosphorylates Pp4c or Smek1, or that Pp4c and Smek1 dephosphorylate Snf1. We assessed the phosphorylation state of Smek1 and Pp4c in the wild-type and Δsnf1 strains grown in glucose (MM) and l-arabinose media using Phos-tag SDS-PAGE. In the wild-type, Pp4c was more phosphorylated in l-arabinose medium than in glucose medium (Fig. 9f), indicating that phosphorylated Pp4c is in an activated state during l-arabinose utilization. Mass spectrometry data also showed that Pp4c from M. oryzae cultured in l-arabinose medium exhibited a higher abundance of phosphorylated peptides and more phosphorylated amino acid sites compared to glucose medium (Supplementary Table 2). In l-arabinose medium, the phosphorylation levels of Pp4c, but not Smek1, were lower in Δsnf1 than in the wild-type (Fig. 9f), suggesting that Snf1 kinase contributes to the phosphorylation of Pp4c in vivo. Therefore, the dephosphorylated and inactivated Pp4c in Δsnf1 was responsible for the higher phosphorylation levels of CreA and Crf1 in l-arabinose medium.

Fig. 9: Snf1 and Pp4 regulate each other's phosphorylation status in Magnaporthe oryzae.
figure 9

a Yeast two-hybrid analysis of Snf1 with Smek1 and Pp4c. GST pull-down (b) and co-immunoprecipitation (d) assays to confirm Snf1 interaction with Pp4c. GST pull-down (c) and co-immunoprecipitation (e) assays to confirm Snf1 interaction with Smek1. Strains were cultured in glucose medium for 2 days (d, e). f Mn2+ phos-tag assays demonstrating Snf1 phosphorylation of Pp4c but not Smek1 in vivo. g Schematic of Snf1-mediated phosphorylation of Pp4c and Crf1, with subsequent Pp4c-mediated dephosphorylation of CreA and Crf1. Western blot of Snf1 phosphorylation levels in the wild-type, Δpp4c, and Δsmek1 strains cultured in glucose and l-arabinose media: representative images (h), and statistical graphs (i). Error bars represent SD. Asterisks “**” indicate significant differences (Tukey's HSD test, p < 0.01). n = 5 biologically independent samples. j Diagram illustrating Snf1 dephosphorylation by Pp4c and Smek1.

Full size image

Based on these findings, we propose that Snf1 regulates the activities of CreA and Crf1 through two distinct pathways. In glucose medium, Snf1 directly phosphorylates CreA to P-CreA and Crf1 to PP-Crf1 (Fig. 9g). In l-arabinose medium, Snf1 phosphorylates Pp4c, which in turn dephosphorylates P-CreA to CreA and PP-Crf1 to P-Crf1 (Fig. 9g).

To clarify whether the Pp4 phosphatases (Smek1 and Pp4c) also regulate the dephosphorylation of Snf1, we assessed the phosphorylation levels of Snf1 in the wild-type, Δpp4c, and Δsmek1 strains grown in glucose or l-arabinose media. The phosphorylation level of Snf1 was reduced in l-arabinose medium in the wild-type compared to glucose medium (Fig. 9h, i), suggesting that Snf1 involved in l-arabinose utilization is in a dephosphorylated state. Snf1 from M. oryzae cultured in l-arabinose medium also showed a lower abundance of phosphorylated peptides and fewer phosphorylated amino acid sites compared to glucose medium (Supplementary Table 2). This indicates that dephosphorylated Snf1 is active during l-arabinose utilization, whereas phosphorylated Snf1 is active during glucose utilization. In both glucose and l-arabinose media, the phosphorylation levels of Snf1 in Δsmek1 and Δpp4c were higher than in the wild-type, suggesting that Smek1 and Pp4c are responsible for the dephosphorylation of Snf1 (Fig. 9h–j).

Discussion

Carbon catabolite repression is crucial in regulating carbon source utilization and pathogenicity in pathogenic fungi5,42. For example, knockout of CREA results in the loss or attenuation of pathogenicity in plant pathogenic fungi such as M. oryzae strain Guy11, Ustilaginoidea virens, and A. flavus27,44,45. However, the mechanisms of carbon catabolite repression and de-repression in pathogenic fungi remain poorly understood. In this study, we propose a regulatory mechanism for coordinating the utilization of preferred and non-preferred carbon sources through CCR, carbon catabolite de-repression (CCDR), and transcriptional activators in M. oryzae (Fig. 10).

Fig. 10: Proposed regulatory network for carbon catabolite repression (CCR) and de-repression (CCDR) in Magnaporthe oryzae.
figure 10

During CCR (glucose as a carbon source), phosphorylated Snf1 and Smek1 regulate the phosphorylation states of CreA and Crf1, activating CreA and inactivating Crf1, thereby repressing non-preferred carbon source catabolism genes. Under CCDR conditions (l-arabinose as the carbon source), Snf1 phosphorylates Pp4c, which dephosphorylates CreA and Crf1, resulting in CreA inactivation and Crf1 activation to promote expression of non-preferred carbon source catabolism genes. Sak1 and Tos3 are kinases activating Snf1. The catalytic subunit of the protein phosphatase that catalyzes CreA dephosphorylation when glucose is the carbon source has not been identified.

Full size image

This regulatory network of carbon metabolism involves phosphorylation, dephosphorylation, and expression of regulatory proteins, as well as feedback and self-maintenance mechanisms (Fig. 10). In the utilization of non-preferred carbon sources such as l-arabinose, the Pp4 phosphatases (Smek1 and phosphorylated Pp4c) dephosphorylate Snf1, and the dephosphorylated Snf1 then phosphorylates Pp4c. This increase in phosphorylated Pp4c subsequently dephosphorylates CreA, relieving its inhibitory effect on the expression of genes required for the utilization of non-preferred carbon sources. Meanwhile, phosphorylated Pp4c and Smek1 (Pp4) dephosphorylate polyphosphorylated Crf1 to hypophosphorylated Crf1, promoting the expression of SNF1 and non-preferred carbon source utilization genes. The newly expressed Snf1 further promotes the phosphorylation of Pp4c, maintaining the activation of CCDR and non-preferred carbon source utilization (Fig. 10). When a preferred carbon source, such as glucose, is utilized, phosphorylated Snf1 phosphorylates Crf1, converting it into polyphosphorylated Crf1, which loses its ability to promote the expression of genes for non-preferred carbon sources. At the same time, phosphorylated Snf1 hypophosphorylates CreA, while Smek1 reduces the phosphorylation level of polyphosphorylated CreA, and the hypophosphorylated CreA suppresses the expression of non-preferred carbon source utilization genes (Fig. 10).

To date, the C2H2 transcription factor CreA is recognized as a key regulator of fungal CCR5,46,47,48. In M. oryzae, CreA suppresses the utilization of lipids, glycerol, ethanol, acetate, and l-arabinose10,11,41. Similarly, in other plant pathogenic fungi, such as Ustilaginoidea virens and A. flavus, CreA also inhibits the use of non-preferred carbon sources44,45. The activity of CreA is modulated by its phosphorylation status. In S. cerevisiae, the phosphorylation of Mig1 (CreA) by Snf1 induces its translocation from the nucleus, thereby relieving CCR. When Mig1 is dephosphorylated by the Reg1-Glc7 protein phosphatase complex in the cytoplasm, it re-enters the nucleus and represses genes involved in the utilization of non-preferred carbon sources49. In contrast to yeasts, filamentous fungi exhibit higher levels of CreA phosphorylation under CCR. For instance, in Trichoderma reesei, the phosphorylation of Ser241 and Ser388 of Cre1 (CreA) is crucial for its DNA-binding ability in the presence of glucose50,51. In A. nidulans, increased phosphorylation of CreA leads to repression of genes required for the utilization of non-preferred carbon sources, such as xylan20. Phosphorylation of CreA at three sites (Ser262, Ser268, and Thr308) is necessary for DNA binding, while phosphorylation at Ser319 is linked to its degradation20. In M. oryzae, the deletion of SMEK1 results in elevated phosphorylation of CreA in glucose medium11, whereas the deletion of SNF1 results in reduced phosphorylation of CreA, indicating that M. oryzae contains at least three distinct phosphorylation states of CreA. Among these, the intermediate phosphorylation state of CreA suppresses the utilization of non-preferred carbon sources (Fig. 10).

The phosphorylation and dephosphorylation of CreA are regulated by protein kinases and phosphatases. In yeasts, the dephosphorylation of cytoplasmic Mig1 by the Reg1-Glc7 protein phosphatase complex occurs under repressing conditions49. In filamentous fungi, however, no other phosphatases besides Smek1, the regulatory subunit of the Pp4 protein phosphatase, have been identified to dephosphorylate CreA11. In this study, we identified Pp4c, the catalytic subunit of Pp4, as the phosphatase responsible for dephosphorylating CreA. Pp4c is a conserved catalytic subunit of the Pp4 phosphatase, though its function in filamentous fungi had not been previously reported. Together with Smek1, Pp4c forms the catalytic and regulatory subunits of the Pp4 protein phosphatase28. In S. cerevisiae, the homologs of Pp4c (Pph3) and Smek1 (Psy2) make up the Pp4 complex, which regulates DNA damage checkpoint recovery, DNA break repair, and the non-homologous end-joining (NHEJ) pathway. The Pp4 complex also represses genes related to glucose transport and activates Gln3 to alleviate nitrogen catabolite repression30,31,52,53,54. In mammals, Pp4c is involved in diverse cellular processes, including DNA damage response, tumorigenesis, immune response, stem cell development, glucose metabolism, and obesity55,56. In M. oryzae, Pp4c, in conjunction with Smek1, regulates CCR, CCDR, and the activation of non-preferred carbon source utilization. Both Pp4c and Smek1 are indispensable for the pathogenicity of M. oryzae.

In yeasts, Snf1 is phosphorylated under glucose limitation, after which it phosphorylates Mig149. Snf1 is an AMP-activated protein kinase (AMPK), and its complex is conserved across fungi, plants, and animals. In many plant pathogenic fungi, including M. oryzae, Verticillium dahliae, Fusarium oxysporum, Cochliobolus carbonum, and Botrytis cinerea, Snf1 is essential for pathogenicity35,57,58,59,60. However, in some fungi, such as Ustilago maydis, Snf1 is not required for pathogenicity61. This functional discrepancy may be attributed to the diversity in Snf1's involvement in CCR, CCDR, or other metabolic pathways across different fungi. The role of Snf1 in regulating CCR and CCDR pathways in filamentous fungi remains poorly characterized. In T. reesei, Snf1 does not phosphorylate Cre1 (CreA), while casein kinase II is responsible for phosphorylating Cre1 at Ser24162. Similarly, in A. nidulans, SnfA (the homolog of Snf1) does not phosphorylate CreA and is not recognized as an interaction partner of CreA63, although it is necessary for CreA’s proper localization in the cytoplasm64. Additionally, Stk22, a kinase that activates SnfA, is regulated by PkaA and can act on CreA65, suggesting an indirect regulatory mechanism of CreA by SnfA20. In this study, we found that Snf1 in M. oryzae was less phosphorylated in l-arabinose medium compared to glucose medium. Snf1 exerts two distinct effects on CreA and Crf1, regulating both CCR and CCDR in M. oryzae. Under glucose conditions, phosphorylated Snf1 directly phosphorylates CreA and Crf1, inhibiting the utilization of non-preferred carbon sources. Conversely, in l-arabinose medium, dephosphorylated Snf1 phosphorylates Pp4c, which subsequently dephosphorylates CreA and Crf1, thereby activating the use of non-preferred carbon sources (Fig. 10). Unlike the yeast mechanism, this regulatory pathway illustrates a distinct role for Snf1 in coordinating CCR and CCDR in M. oryzae.

Crf1 is a pivotal transcription factor driving the utilization of non-preferred carbon sources in M. oryzae. It plays an essential role in the metabolism of glycerol, ethanol, acetate, arabinose, galactose, lipids, aspartic acid, and leucine10,11,41,66. While Crf1 is ubiquitous in filamentous fungi, no close homologs have been identified in S. cerevisiae. In A. nidulans, the homolog GlcD/XP_050469005.1 of Crf1 is critical for glycerol utilization67. Beyond these fungi, the roles of Crf1 homologs remain unexplored in other species. In M. oryzae, Crf1 and CreA work synergistically yet independently to regulate CCR, CCDR, and the metabolism of non-preferred carbon sources. Previous studies demonstrated that CreA and Crf1 independently modulate lipid and l-arabinose catabolism10,11. This study extends these findings by showing that CreA and Crf1 separately regulate glucose, xylose, glycerol, and ethanol metabolism.

Similar to CreA, Crf1 undergoes phosphorylation at multiple sites, with its phosphorylation patterns varying based on carbon source or developmental stage. Franck et al. identified five phosphorylation sites on Crf1 (Thr114, Ser122, Ser206, Ser335, and Ser343), three of which (Thr114, Ser122, and Ser343) show increased phosphorylation during appressorium formation66. During this process, fungal cells operate in a nutrient-deprived environment, synthesizing large quantities of glycerol using glycogen, lipid droplets, and trehalose stored in spores. Wang et al. later identified an additional phosphorylation site (Ser291) in mycelia cultured in yeast extract glucose medium68. In this study, we identified four new phosphorylation sites (Thr178, Ser185, Thr189, and Thr192) and uncovered at least four distinct Crf1 phosphorylation states in glucose and l-arabinose media. The complex and dynamic phosphorylation patterns of Crf1 and CreA highlight their versatile regulatory roles in multiple carbon utilization pathways.

Consistent with findings for Δsmek1 and Δcrf110,11, Δpp4c, Δsnf1, and ΔcreA mutants exhibited similar defects in metabolizing carbon sources such as glucose, l-arabinose, and lipids. However, Pp4c, CreA, and Snf1 exhibit nuanced differences in regulating metabolic gene expression. In glucose medium, CreA suppresses, while Snf1 promotes, the expression of genes involved in lipolysis, fatty acid activation and transport, β-oxidation, and the glyoxylate cycle gene ICL1. Pp4c positively regulates lipolysis and ICL1 expression but exerts mixed effects on genes related to fatty acid metabolism and β-oxidation. For lipid utilization, Snf1 promotes the expression of catabolic genes, aligning with Crf1's function10. Smek1 negatively regulates ICL1 and lipid metabolism genes in glucose medium but positively influences their expression in olive oil medium10,11. In l-arabinose medium, Pp4c and Snf1 upregulate l-arabinose catabolic genes, consistent with roles reported for Smek1 and Crf111. Conversely, CreA suppresses l-arabinose catabolic genes while enhancing the expression of PRD1 and LAD1 in glucose medium. These functional differences arise from several factors. (1) Pp4c and Smek1 interact directly, with Smek1 capable of binding multiple phosphatase catalytic subunits. Similarly, mammalian Pp4c interacts with various regulatory subunits28. (2) Snf1, Pp4c, and Smek1 differentially influence the downstream transcription factors CreA and Crf1. (3) Metabolic gene expression reflects the combined actions of the repressor CreA and the activator Crf1, with CreA occasionally promoting gene expression69. (4) Crf1 is a key regulator of non-preferred carbon source utilization. Other transcription factors, such as Ara1 for arabinose metabolism and Far1, Far2, Gpf1, and Vrf2 for lipid metabolism, also contribute to regulating metabolic genes42. (5) In CCR, CreA inhibits metabolic pathways by repressing a limited set of metabolic genes44,70. Thus, null mutants of regulatory factors in CCR and CCDR share similar carbon source utilization deficiencies but differ in gene expression phenotypes.

In summary, this study confirms that Pp4c, the catalytic subunit of protein phosphatase Pp4, is essential for carbon metabolism, development, and pathogenicity in M. oryzae. Through Pp4c, we have established a regulatory network governing CCR, CCDR, and non-preferred carbon source metabolism. Plant pathogenic fungi adapt to fluctuating carbon and energy sources during invasion via CCR and CCDR, enhancing their virulence. These findings provide a framework for understanding carbon metabolic regulatory mechanisms in phytopathogenic fungi and offer insights for developing disease prevention strategies targeting fungal carbon metabolism in agricultural production.

Methods

Strains and cultural conditions

M. oryzae strain 70-15 and its transformants were used in this study (Supplementary Table 3). The Δpp4c, Δsnf1, Δpp4cΔcrf1, Δsnf1Δcrf1, Δpp4cΔcreA, and Δsnf1ΔcreA mutants were constructed using a gene knockout method via the plasmids pKO3A or pKO3B71,72. The primers used in this study are listed in Supplementary Table 4. Strains were cultured in complete medium (CM), minimal medium (MM), MM-C (MM without glucose), and MM-C supplemented with 1% olive oil, xylose, l-arabinose, d-arabinose, glycerol, and ethanol at 25 °C under light–dark (16-h/8-h) cycles.

Phenotypic and pathogenicity assays

Phenotypic assays were performed as described in previous reports11,73. For colony growth, strains were cultured on CM plates, and measurements were taken at 8 days post-inoculation (dpi). Fungal conidiation was assessed on CM plates at 8 dpi. For conidial germination and appressorium formation, 40 µL of conidial suspensions (5 × 104 spores mL−1) were placed on hydrophobic coverslips, incubated at 22 °C in the dark, and observed at 6, 8, and 24 hpi. For conidiophore observation, vegetative mycelia were scraped, sectioned, and cultured at 25 °C under constant light for 24 h.

In pathogenicity tests, conidial suspensions (5 × 104 spores mL−1) were sprayed onto 9-day-old rice seedlings (Oryza sativa Co-39). Seedlings were incubated at 22 °C in the dark for 2 days, followed by light–dark cycles (16-h/8-h) at 25 °C for 5 days. The disease index was calculated as the proportion of diseased area on 5-cm-long infected leaves. For appressorial penetration tests on barley, conidial suspensions were inoculated onto barley leaves, incubated at 25 °C, and observed at 24 and 48 hpi.

Lipid droplet and glycogen granule observation

Lipid droplets were labeled with Cap20-GFP following previously described methods10,11. Fluorescent fusion proteins were transformed into the wild-type and Δpp4c strains via Agrobacterium tumefaciens-mediated transformation (ATMT). Spores and developing appressoria of the wild-typeCap20-GFP and Δpp4cCap20-GFP strains were observed using fluorescence microscopy (Nikon Eclipse 80i, Japan). Intracellular glycogen was stained with iodine–potassium iodide (I2/KI), and the percentage of conidia or appressoria containing glycogen granules was calculated.

The relative triglyceride utilization rate of the wild-type, ΔcreA, Δpp4c, and Δsnf1 strains was measured. Mycelia cultured in liquid CM at 25 °C with shaking (150 rpm) for 2 days were harvested, split into two equal parts, and re-cultured in liquid MM-C for 24 h. Triglyceride content was determined using a triglyceride content determination kit (AK070, Beijing Bioss Biotech, China).

Protein–protein interaction assays

Yeast two-hybrid (Y2H) assays were performed as previously described11,74. Full-length cDNA fragments of CREA (MGG_11201), CRF1 (MGG_05709), SMEK1 (MGG_06138), PP4C (MGG_01528), and SNF1 (MGG_00803) were amplified from the M. oryzae 70-15 cDNA library and cloned into pGADT7 (prey vector) or pGBKT7 (bait vector). Bait and prey constructs were transformed into yeast cells and cultured on SD-Leu-Trp medium at 30 °C for 4 days. Yeast cells were diluted into a series of concentrations, spotted onto SD-Leu-Trp-His-Ade medium, and cultured for 3–5 days.

For GST pull-down assays, cDNA fragments of SMEK1, PP4C, CREA, CRF1, and SNF1 were inserted into pGEX-4T or pET21a vectors to construct protein expression vectors fused with GST or 3×Flag tags. Proteins were pulled down using anti-GST affinity beads and detected via western blot with anti-GST (EM80701, Huabio, China) and anti-Flag (M1403-2, Huabio, China) antibodies.

For in vivo Co-IP assays, SMEK1, PP4C, CREA, CRF1, and SNF1 coding sequences were fused with GFP or 3×Flag tags and transformed into wild-type or mutant strains. Proteins were extracted from transformants cultured in liquid CM for 2 days or re-cultured in l-arabinose medium for 6 h, precipitated with anti-GFP beads (SA070001, Smart-Lifesciences, China), and detected by western blot using anti-Flag and anti-GFP antibodies (M1403-2, Huabio, China).

In vivo protein phosphorylation analysis

CreA-3×Flag, Crf1-3×Flag, Smek1-3×Flag, or Pp4C-3×Flag constructs were transferred into wild-type and mutant strains. Total proteins were extracted from transformants cultured in CM for 2 days and transferred to MM or MM-C supplemented with 1% l-arabinose for 6 h. According to our previous work11, samples were resolved on 10% SDS-PAGE containing 50 mM acrylamide phos-tag ligand (F4002, APExBIO, USA) and 100 mM MnCl2 or ZnCl2. Gels were electrophoresed, equilibrated in transfer buffer containing 10 mM EDTA, and transferred to PVDF membranes at 80 V for 4 h at 4 °C. Membranes were analyzed by western blot with anti-Flag antibodies.

Phosphorylation levels of Snf1 in wild-type, Δsmek1, and Δpp4c strains were analyzed. Proteins were resolved on 8% SDS-PAGE and transferred to PVDF membranes. Phosphorylated Snf1 was detected using anti-Phospho-AMPKα1 (Thr183)/AMPKα2 (Thr172) antibodies (AF5908, Beyotime Biotechnology, China). GAPDH was used as a control and detected with anti-GAPDH antibodies (EM1101, Huabio, China).

Reverse transcription-quantitative PCR

The mycelia of ΔcreA, Δcrf1, Δsnf1, and Δpp4c strains were cultured in liquid CM at 25 °C for 2 days and subsequently re-cultured in liquid MM-C medium for 24 h. The mycelia were collected, divided into equal parts, and transferred to liquid MM or MM-C supplemented with 1% olive oil or l-arabinose for re-culture for 6 h. Total RNA was extracted using RNAiso Plus (9108, Takara, Japan). RT-qPCR was performed using a TB Green Premix Ex Taq II kit (RR82LR, Takara, Japan). H3 and α-ACTIN were used as reference genes.

Mass spectrometry

The Crf1-3×Flag fusion gene was transformed into the Δcrf1 strain. Mycelia were cultured in CM for 2 days and then re-cultured in MM or l-arabinose medium for 6 h. Total proteins were extracted and precipitated using anti-Flag beads (SA042001, Smart-Lifesciences, China). Eluted proteins were analyzed via mass spectrometry at Novogene (Beijing, China).

Statistics and reproducibility

Each value shown for each group is the mean ± SD (standard deviation). Sample sizes were chosen based on previous publications11,41. All experiments were repeated independently at least twice. The p value was calculated using Tukey's HSD test (one-way ANOVA) in SPSS. p values < 0.05 were considered significant, while p values > 0.05 were considered non-significant. Each images represent at least three independent biological repeats.