The TOR signaling pathway regulates vegetative development and virulence in Fusarium graminearum



  • The target of rapamycin (TOR) signaling pathway plays critical roles in controlling cell growth in a variety of eukaryotes. However, the contribution of this pathway in regulating virulence of plant pathogenic fungi is unknown.
  • We identified and characterized nine genes encoding components of the TOR pathway in Fusarium graminearum. Biological, genetic and biochemical functions of each component were investigated.
  • The FgFkbp12-rapamycin complex binds to the FgTor kinase. The type 2A phosphatases FgPp2A, FgSit4 and FgPpg1 were found to interact with FgTap42, a downstream component of FgTor. Among these, we determined that FgPp2A is likely to be essential for F. graminearum survival, and FgSit4 and FgPpg1 play important roles in cell wall integrity by positively regulating the phosphorylation of FgMgv1, a key MAP kinase in the cell wall integrity pathway. In addition, the FgPpg1 interacting protein, FgTip41, is involved in regulating mycelial growth and virulence. Notably, FgTip41 does not interact with FgTap42 but with FgPpg1, suggesting the existence of FgTap42:FgPpg1:FgTip41 heterotrimer in F. graminearum, a complex not observed in the yeast model.
  • Collectively, we defined a genetic regulatory framework that elucidates how the TOR pathway regulates virulence and vegetative development in F. graminearum.


All living organisms interact actively with their surrounding environments and modulate their physiology to maintain cellular homeostasis. This adaptation process is highly coordinated via diverse signaling pathways. The target of rapamycin (TOR) signaling pathway plays a pivotal role in nutrient signal transduction in eukaryotes (Wang & Proud, 2009). Rapamycin was discovered in the early 1970s as an antifungal agent against the pathogenic yeast Candida albicans (Sehgal et al., 1975). Later, it was found to inhibit proliferation of mammalian cells and to possess immunosuppressive properties. These intriguing observations attracted researchers to investigate the mode of action of this compound. In the early 1990s, Tor kinase was first discovered in the budding yeast Saccharomyces cerevisiae (Heitman et al., 1991). In S. cerevisiae, rapamycin forms a complex with the peptidyl-prolyl cis/trans isomerase Fkbp12 (FK506 binding protein of 12 kDa) (also named Fpr1), and this complex then binds to and inhibits Tor kinases (Heitman et al., 1991). Subsequently, Tor kinases have been identified in various eukaryotes ranging from yeasts to mammals.

In eukaryotic cells, the TOR pathway responds to nutrients and growth factors to orchestrate cell growth and proliferation. Most organisms, including mammals, have a single Tor kinase. Saccharomyces cerevisiae and Schizosaccharomyces pombe, however, contain two Tor homologs (Shertz et al., 2010), and the protozoans Leishmania major and Trypanosoma brucei possess three and four TOR-like genes, respectively (Barquilla et al., 2008; Madeira da Silva & Beverley, 2010). The Tor kinases interact with other proteins to form two complexes known as TORC1 and TORC2, which are essential regulators of cell growth in response to nutrients, hormones or stresses in S. cerevisiae. TORC1 represents the rapamycin-sensitive signaling branch that mediates temporal control of cell growth by activating anabolic processes such as ribosome biogenesis, protein synthesis, transcription and nutrient uptake, and by inhibiting catabolic processes such as autophagy and ubiquitin-dependent proteolysis. TORC2 is rapamycin insensitive and is required for the actin cytoskeleton organization (reviewed in Wullschleger et al., 2006).

In S. cerevisiae, the Tap42-phosphatase complexes are major targets of the Tor kinases in the rapamycin-sensitive signaling pathway (Yan et al., 2006). In these complexes, Tap42 interacts with the catalytic subunits of the type 2A and 2A-like phosphatases, such as Pph3, Pph21, Pph22 and Sit4 (Di Como & Arndt, 1996; Wang et al., 2003). Under starvation or rapamycin treatment, Tor kinase is inactive, which triggers the disassembly of the Tap42-phosphatase complex from TORC1. Subsequently, Tap42 dissociates from Sit4 (Di Como & Arndt, 1996). Sit4 is active under conditions in which Tor kinase is inactive (Beck & Hall, 1999). Tap42-dissociated Sit4 may dephosphorylate the GATA transcription factor Gln3. The dephosphorylated form of Gln3 is dissociated from Ure2, a cytoplasmic anchor protein, and imported into the nucleus (Crespo & Hall, 2002; Inoki et al., 2005; Di Como & Jiang, 2006). In the nucleus, Gln3 binds to GATA-containing promoters and activates transcription of GAP1, MEP1, GLN1 or GDH1 that are required for the adaptation of S. cerevisiae to less preferred nitrogen (N) sources (Cooper, 2002).

Interestingly, genome-wide searches for Tor kinases in filamentous fungi – for example, Aspergillus species, Fusarium graminearum, Neurospora crassa and Magnaporthe oryzae – revealed that all contain only a single Tor ortholog (Shertz et al., 2010). Thus far, little is known about the functions of fungal Tor kinase and the genetic pathways associated with TOR in plant pathogenic fungi (Teichert et al., 2006; Shertz & Cardenas, 2011). Rapamycin is effective against fungi including Cryptococcus neoformans, Aspergillus species and Podospora anserine, as well as the phytopathogenic fungus Botrytis cinerea (Cruz et al., 1999; Dementhon et al., 2003; Muthuvijayan & Marten, 2004; Melendez et al., 2009), suggesting that functions of the TOR pathway in plant pathogenic fungi may be similar to that in S. cerevisiae. However, questions remain unanswered on how this pathway regulates and impacts various cellular processes including virulence in plant pathogenuic fungi.

Fusarium graminearum (teleomorph: Gibberella zeae) is an economically important plant pathogen that causes Fusarium head blight (FHB) on various cereal crops (Goswami & Kistler, 2004). In addition to yield reduction, mycotoxins produced by F. graminearum in infested grains pose a serious threat to human and animal health (McMullen et al., 1997; Pestka & Smolinski, 2005). In a preliminary study, we found that rapamycin is a very strong inhibitor of F. graminearum growth, prompting us to select this fungus as a model to investigate the functional role of the TOR pathway in plant pathogenic fungi. Thus, we identified putative components of the TOR pathway in F. graminearum (Supporting Information Table S1), and characterized deletion mutants of these genes. Our study shows that the TOR signaling pathway plays important roles in various cellular processes including cell wall integrity, secondary metabolisms and virulence in F. graminearum. Through this study, we established a genetic framework that provides an explanation to how the TOR signaling pathway regulates vegetative differentiation and virulence in F. graminearum.

Materials and Methods

Fungal strains, media and culture conditions

Fusarium graminearum strain PH-1 was used as the wild-type (WT) strain for constructing various gene deletion mutants in this study. The WT strain, resulting mutants and complemented strains were routinely cultured on potato dextrose agar (PDA) (200 g potato, 20 g dextrose, 20 g agar and 1 l water) at 25°C with a 12 h : 12 h, light : dark cycle. The WT strain and its derived mutants were grown on carrot agar for induction of sexual development (Klittich & Leslie, 1988) near-UV light (wavelength, 365 nm; HKiv Co., Ltd, Xiamen, China), and in mung bean broth (MBB) for conidiation assays under continuous light. Assays for virulence were performed as described previously (Jiang et al., 2011a). Each experiment was repeated three times.

Isolation of a rapamycin-resistant (RR) mutant of F. graminearum

In order to induce RR mutants, 1 × 105 conidia of PH-1 were spread on PDA plates amended with 10 μg ml−1 rapamycin, and then irradiated under UV for 90 s. After the plates were incubated at 25°C in the dark for 4 d, a RR colony growing on the plate was obtained. A single-spore strain from the resistant colony was further used for DNA sequence analyses of FgKPB12 and FgTOR genes.

Construction of vectors for gene deletion and complementation

The primer pairs used to amplify the flanking sequences or full sequence of each gene are listed in Table S2. Constructs for gene deletion and complementation of F. graminearum were carried out as described previously (Jiang et al., 2011a). PCR products were transformed into PH-1 protoplasts by employing polyethyleneglycol (PEG)-mediated protoplast transformation (Proctor et al., 1995). Putative gene deletion mutants were identified by PCR assays with primer pairs from 45 to 66 as listed in Table S2, and were further confirmed by Southern assays (Fig. S1). All mutants generated in this study were preserved in 15% glycerol at −80°C.

Construction of 3 × FLAG and GFP fusion cassettes

In order to create the FgSIT4-3 × FLAG fusion construct, the FgSIT4 region was amplified with the primers 126 and 127 (Table S2). The resulting PCR products were co-transformed with XhoI-digested pHZ126 (Zhou et al., 2011a) into XK1-25 (Bruno et al., 2004). The FgSIT4-3 × FLAG fusion vector was recovered from yeast transformants and subsequently transformed into the WT strain PH-1. FgPPG1-3 × FLAG, FgTAP42-3 × FLAG and FgTAP42-GFP fusion constructs were constructed using the same strategy.

Microscopic examinations of hyphal and conidial morphology, and GFP fluorescence

The hyphal morphology of each mutant was examined with a Leica TCS SP5 imaging system using fresh mycelia harvested from 3-d-old colonies of each strain growing on PDA plates amended with rapamycin at 0, 0.025, 2.5 or 250 ng ml−1. In addition, hyphal septation was examined after mycelia of each strain were stained with calcofluor white. For histochemical analysis of lipid droplets in the hyphae, mycelia were mounted in Nile Red staining solution consisting of 20 mg ml−1 polyvinylpyrrolidone and 2.5 mg ml−1 Nile Red Oxazone (9-diethylamino-5H-benzo[α]phenoxazine-5-one, Sigma) in 50 mM Tris-maleate buffer (pH 7.5) (Jiang et al., 2011b). Following the treatment, lipid droplets fluoresce within a few seconds and can be viewed under a microscope with episcopic fluorescence attachment. To observe nuclei, fresh conidia or mycelia were washed with sterilized water and stained with 10 μg ml−1 4′6-diamidino-2-phenylindole (DAPI, Sigma). Calcofluor white staining was carried out as described previously (Rui & Hahn, 2007). For observation of GFP signals, fresh mycelia and conidia were examined with the Zeiss LSM780 confocal microscope (Carl Zeiss AG, Germany).

For conidiation assays, three mycelial plugs of each strain were inoculated in 30 ml MBB supplemented with 0, 0.025, 2.5 or 250 ng ml−1 rapamycin. After incubation in a shaker with 180 rpm at 25°C for 1 wk, conidia of each sample were collected by centrifugation and calculated by a hemocytometer. The experiments were repeated three times.

Determination of deoxynivalenol (DON) production

A 50-g aliquot of healthy wheat kernels was autoclaved and then inoculated with five mycelial plugs of each strain. Wheat kernel aliquots inoculated with five agar plugs were used as a negative control. After incubation at 25°C for 20 d, DON and fungal ergosterol were extracted using previously described protocols (Mirocha et al., 1998). The DON extracts were purified with PuriToxSR DON column TC-T200 (Trilogy analytical laboratory), and amounts of DON and ergosterol in each sample were determined using a Waters 1525 HPLC system (Liu et al., 2013). The experiment was repeated three times.

Yeast strains and complementation assays

The full-length cDNA of each gene tested was amplified using primer pairs from 89 to 100 as listed in Table S2. The PCR product was digested with appropriate enzymes and cloned into the pYES2 vector (Invitrogen), and then transformed into the corresponding yeast mutant. Yeast transformants were selected on synthetic medium lacking uracil (Clontech, Palo Alta, CA, USA). Additionally, the WT strain BY4741 and the mutant transformed with the empty pYES2 vector were used as controls. For complementation assays, the yeast transformants were grown at 30 or 37°C on YPG medium (1% yeast extract, 2% bactopeptone, 2% galactose) supplied with various stress agents as indicated in the legend of Fig. 4. The experiments were repeated three times.

Yeast two-hybrid assays

In order to construct plasmids for yeast two-hybrid analyses, the coding sequence of each tested gene was amplified from the cDNA of PH-1 with the primer pairs from 101 to 118 as indicated in Table S2. The cDNA fragment was inserted into the yeast GAL4 binding domain vector pGBKT7 and GAL4 activation domain vector pGADT7 (Clontech, Mountain View, CA, USA), respectively. The pairs of yeast two-hybrid plasmids were co-transformed into S. cerevisiae strain AH109 following the LiAc/SS-DNA/PEG transformation protocol (Schiestl & Gietz, 1989). In addition, a pair of plasmids, pGBKT7-53 and pGADT7, served as a positive control. The following pairs of plasmids were used as negative controls: pGBKT7-Lam and pGADT7; pGBKT7 and pGADT7-FgPP2A; pGBKT7 and pGADT7-FgSIT4; pGBKT7 and pGADT7-FgPPG1; pGADT7 and pGBKT7-FgTIP41; pGADT7 and pGBKT7-FgTAP42; and pGADT7 and pGBKT7-FgMSG5. Transformants were grown at 30°C for 3 d on synthetic medium lacking Leu and Trp, and then transferred to the medium stripped of His, Leu and Trp and containing 5 mM 3-aminotriazole (3-AT) to assess binding activity (Jiang et al., 2011b). The transformants were also assayed for β-galactosidase activity following the previously published protocol (Zhou et al., 2011b). Three independent experiments were performed.

Western blotting assay

Six mycelial plugs of each tested mutant were inoculated into 150 ml potato dextrose broth (PDB) and incubated at 25°C with agitation (200 rpm) for 36 h. Mycelia were harvested, washed with deionized water, and then ground in liquid nitrogen. Approximately 200 mg finely ground mycelia were resuspended in 1 ml of extraction buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM EDTA, 1% Triton X-100, 2 mM PMSF) and 10 μl of protease inhibitor cocktail (Sangon, Shanghai, China). After homogenization with a vortex shaker, the lysate was centrifuged at 10 000 g for 20 min at 4°C. Then, 100 μl of supernatant was mixed with an equal volume of 2× loading buffer and boiled for 5 min. Subsequently, 15 μl of each sample was loaded onto SDS-PAGE gels. The proteins separated on SDS-PAGE gels were transferred onto a polyvinylidene fluoride membrane with a Bio-Rad electroblotting apparatus. The FgMgv1 kinase and its phosphorylated protein were detected with the PhophoPlus p44/42 MAP kinase antibody kit (Cell Signaling Technology, Beverly, MA, USA). The horseradish peroxidase-conjugated secondary antibody and chemiluminescent substrate (Santa Cruze Biotechnology, Santa Cruze, CA, USA) were used for antigen antibody detections. The monoclonal anti-FLAG (Abmart, Shanghai, China) was used at a 1 : 1000–1 : 2000 dilution for immunoblot analysis. The experiment was repeated three times.

Affinity purification and mass spectrometry analysis

Protein extraction was performed as already described. Approximately 50 μl of anti-FLAG agarose (Abmart, Shanghai, China) was added to capture FgTap42, FgSit4 or FgPpg1-interacting proteins, following the manufacturer's instructions. After incubation at 4°C overnight, the agarose was washed three times with 500 μl of TBS (20 mM Tris-HCl, 500 mM NaCl, pH 7.5). Proteins binding to the beads were immediately eluted with 60 μl of elution buffer (0.2 M glycine, pH2.5). Eluant was instantly neutralized with 3 μl of neutralization buffer (1.5 M Tris, pH9.0) and digested with trypsin using a previous described protocol (Tao et al., 2005; Zhou et al., 2007). Tryptic peptides were analyzed as described previously (Ding et al., 2010).

Co-immunoprecipitation (Co-IP) assay

FgSIT4 and FgTIP41 were separately amplified and cloned into pHZ126 (Zhou et al., 2011a) by the yeast gap repair approach (Bruno et al., 2004) to generate the 3× FLAG fusion constructs. A similar approach was employed to generate the GFP fusion construct for the pFL1 vector containing FgTAP42 or FgPPG1. The resulting fusion constructs were verified by DNA sequencing and transformed into the WT PH-1. Transformants expressing the fusion constructs were confirmed by Western blot analysis. For Co-IP assays, total proteins were isolated and incubated with the anti-FLAG agarose as described above. Proteins eluted from agarose were analyzed by Western blot detection with monoclonal anti-FLAG and monoclonal anti-GFP antibodies (Abmart, Shanghai, China).

Conventional nucleic acid manipulations

The probes for Southern hybridization analyses of the mutants were labeled with digoxigenin (DIG) using the high prime DNA labeling and detection starter kit II according to the manufacturer's protocol (Roche Diagnostics, Mannheim, Germany).


Rapamycin exhibits a strong inhibition against F. graminearum

Before we investigated the TOR signaling pathway in F. graminearum, we assayed the sensitivity of the WT strain PH-1 to rapamycin. Carbendazim and tebuconazole – two fungicides widely applied for FHB management – were used as controls. Radial growth of PH-1 was severely inhibited on PDA amended with 0.25 μg ml−1 rapamycin (Fig. 1a). Microscopic examination showed that hyphae treated with rapamycin had more branches than the untreated hyphae (Fig. 1b). Interestingly, rapamycin treatment led to increased hyphal septation frequency in F. graminearum (Fig. 1b), indicating that the TOR pathway is involved in the regulation of septum formation or cell wall integrity. In addition, rapamycin-treated hyphae contained many lipid droplets, as visualized by Nile Red staining (Fig. 2a), suggesting that the TOR pathway is involved in the regulation of lipid metabolism in F. graminearum. Rapamycin is also known to induce autophagy in P. anserina and Aspergillus oryzae (Dementhon et al., 2003; Pinan-Lucarre et al., 2005; Kikuma et al., 2006). Because autophagy is involved in nutrient recycling during starvation and plays an important role in growth and development in filamentous fungi (Pollack et al., 2009), we tested whether rapamycin triggers autophagy in F. graminearum. As shown in Fig. 2(b), similar to the autophagy induced by a minimal medium lacking an N source, the autophagy was readily observed in hyphae of F. graminearum treated with 0.25 μg ml−1 rapamycin for 4 h. In addition, rapamycin also exhibited a strong inhibitory effect on asexual reproduction: F. graminearum was unable to produce conidia in MBB amended with 250 ng ml−1 rapamycin (Fig. 2c); with 2.5 ng ml−1 rapamycin, c. 90% inhibition of conidiation was observed. Collectively, these results suggest the likelihood that the TOR pathway plays important roles in vegetative growth and differentiation in F. graminearum.

Figure 1.

Effects of rapamycin on mycelial growth, hyphal branching and septum formation in Fusarium graminearum. (a) Inhibitory effect of carbendazim, tebuconazole and rapamycin against F. graminearum. The wild-type (WT) strain PH-1 was incubated for 3 d on potato dextrose agar (PDA) amended with 0.25 μg ml−1 of carbendazim, tebuconazole or rapamycin. (b) Hyphal morphology of the WT strain PH-1 grown on PDA with 0.25 μg ml−1 rapamycin showed more branches and increased septation than the untreated controls. In the lower panel, the lower and upper images are the hyphae stained or unstained with calcofluor white, respectively.

Figure 2.

Impacts of rapamycin on lipid-body accumulation, autophagesome formation, and conidiation of Fusarium graminearum. (a) Rapamycin treatment resulted in the accumulation of lipid droplets in hyphae of the wild-type PH-1, but to a much lesser extent in hyphae of ∆FgSIT4 and ∆FgPPG1. (b) Rapamycin-induced autophagesome formation in F. graminearum. Typical autophagesomes were visible within hyphae of PH-1 treated with 0.25 μg ml−1 rapamycin. The cultures growing in minimal medium (MM) and MM without nitrogen (N) source (MM-N) were used as negative and positive controls, respectively. (c) Conidiation was assayed after incubation at 25°C for 7 d in mung bean broth (MBB) amended with rapamycin at different concentrations as indicated in the figure. Error bars denote standard errors of three repeated experiments. Bars sharing the same letter denote values that are not significantly different at = 0.05.

Deletion of FgFKBP12 leads to resistance to rapamycin in F. graminearum

In S. cerevisiae, rapamycin does not directly target Tor kinase but, rather, binds to Fkpb12. Rapamycin and Fkbp12 form a gain-of-function complex, and this negatively regulates the Tor kinase activity (Heitman et al., 1991). F. graminearum harbors one FKBP12 ortholog (FGSG_09690, named FgFKBP12) encoding a protein with 57% similarity to S. cerevisiae Fkbp12. Using a targeted gene deletion strategy, we generated a FgFKBP12 deletion mutant (ΔFgFKBP12). Morphological analyses showed that mycelial growth of the mutant was comparable to that of the WT progenitor on PDA (Fig. S2a), suggesting that FgFkbp12 is dispensable for hyphal growth. Fungicide sensitivity tests showed that ΔFgFKBP12 confers resistance to rapamycin and FK506 in F. graminearum (Fig. S3), but not to other fungicides including carbendazim and tebuconazole (data not shown). The sensitivity to rapamycin and FK506 was restored in the complemented strain (ΔFgFKBP12-C) (Fig. S3).

We found two additional genes, FGSG_01408 (named FgFKBP20) and FGSG_01059 (FgFKBP54), encoding proteins with 45% and 49% similarity to Fkbp12. These gene deletion mutants did not show recognizable changes in growth on PDA and in sensitivity to rapamycin (Fig. S2a,b). These results indicate that FgFkbp12, rather than FgFkbp20 and FgFkbp54, is associated with rapamycin toxicity in F. graminearum.

The S1866L mutation in the FgTor kinase confers rapamycin resistance

Saccharomyces cerevisiae has two TOR kinase genes; TOR1 is dispensable, whereas TOR2 is essential (Heitman et al., 1991). However, in silico analysis revealed that FGSG_08133 (FgTOR) is the only TOR kinase gene predicted in F. graminearum. To determine its function, we targeted FgTOR for gene deletion. We recovered > 80 hygromycin-resistant transformants; however, all of them were ectopic mutants and we failed to retrieve a null mutant. These results suggested that the deletion of FgTOR in F. graminearum may be lethal, which is consistent with an earlier report on systematic characterization of the F. graminearum kinome (Wang et al., 2011).

Concurrently, we obtained a RR mutant by UV mutagenesis that grew remarkably better than PH-1 on PDA amended with 10 μg ml−1 rapamycin (Fig. S4a). We amplified and sequenced FgTOR and FgFKB12 genes from this mutant. While no mutation was found in FgFKBP12, we discovered a single base-pair mutation (C to T at nucleotide 5597) in FgTOR resulting in a substitution of serine to leucine at codon 1866 (S1866L) (Fig. 3a). Intriguingly, this mutation is predicted to be located in the second alpha-helix of the Fkbp12-rapamycin binding (FRB) domain (Fig. S4b). Based on these observations, we hypothesized that FgTorS1866L is no longer recognized by or binding to the FgFkbp12-rapamycin complex. To test this, the physical interaction of FgFkbp12 with the FRB domain of FgTor (named FgFRB) was examined by yeast two-hybrid (Y2H) assays. Due to the fact that the WT yeast strain is sensitive to rapamycin, we generated a rapamycin-resistant yeast strain AH109R by UV radiation before this Y2H experiment. The pair of yeast two-hybrid plasmids was co-transformed into the rapamycin-resistant strain AH109R and, as shown in Fig. 3(b), FgFkbp12 was unable to interact with FgFRB without rapamycin treatment. By contrast, when the medium was supplemented with 1 μg ml−1 rapamycin, FgFkbp12 interacted strongly with FgFRB. As expected, FgFkpb12 was unable to interact with the mutated FgFRBS1866L, regardless of rapamycin presence or absence. These results indicate that the point mutation (S1866L) prevents Fkbp12-rapamycin complex from binding to the FgFRB domain of FgTor.

Figure 3.

Yeast two-hybrid assays of interactions of FgFkbp12 with the Fkbp12-rapamycin binding domain of FgTor kinase (FgFRB). (a) Schematic representation of Fusarium graminearum full-length FgTor, showing the HEAT repeats, FAT domain, FRB domain, PI3/PI4 kinase catalytic domain and FATC domain. The numbers in brackets indicate the locations of amino-acid (aa) residues in each domain. Red arrow marks the point mutation in rapamycin-resistant mutant of F. graminearum. (b) Serial dilutions of yeast cells (cells ml−1) transferred with the bait and prey constructs indicated in the figure were assayed for growth on yeast minimal synthetic defined base (SD) depleted of leucine, tryptophan and histidine, but amended with (+) or without (−) 1 μg ml−1 rapamycin. The pair of plasmids pGBKT7-53 and pGADT7 was used as a positive control. The pair of plasmids pGBKT7-Lam and pGADT7 was used as a negative control. The same set of yeast transformants were also assayed for β-galactosidase activity.

FgTap42 interacts with FgSit4, FgPp2A and FgPpg1 in F. graminearum

In S. cerevisiae, Tor kinases execute their functions at least in part by regulating Tap42, a phosphatase 2A-associating protein (Di Como & Arndt, 1996). The F. graminearum genome has a single TAP42 homolog (FGSG_09800, named FgTAP42). The predicted amino acid sequence of FgTap42 shares 29% similarity to S. cerevisiae Tap42. In order to investigate the function of FgTap42, we targeted FgTAP42 for gene deletion. Altogether, we obtained 65 transformants from four independent transformation experiments. However, none proved to be the FgTAP42 targeted gene deletion mutant, which led us to conclude that FgTAP42 may be an essential gene in F. graminearum. This conclusion is in agreement with a previous finding that Tap42 is an essential protein in S. cerevisiae (Di Como & Arndt, 1996).

In order to further determine the function of FgTap42, we assayed whether FgTAP42 could complement a temperature-sensitive S. cerevisiae tap42-11 mutant that can grow at 30°C but not at 37°C. The full-length FgTAP42 cDNA was cloned into pYES2 and transformed into the tap42-11 mutant. The growth defect of tap42-11 at 37°C was partially restored in the yeast transformant expressing FgTAP42 (Fig. 4a), indicating that the functions of Tap42 orthologs may be conserved in the budding yeast and filamentous fungi.

Figure 4.

Complementation of the yeast mutants with Fusarium graminearum gene counterparts. (a) F. graminearum FgTAP42 partially complemented a yeast temperature-sensitive (ts) mutant tap42-11. The BY4741-derived ts mutant tap42-11 was transformed with pYES2 or pYES2-FgTAP42. Yeast cells containing BY4741, tap42-11+ pYES2 or tap42-11+ pYES2-FgTAP42 were spotted on yeast extract/peptone medium containing 2% galactose medium (YPG) and incubated at 30 or 37°C for 4 d. (b) FgPP2A, FgSIT4 and FgPPG1 partially complemented the growth defect of yeast SIT4 deletion mutant under 37°C. The yeast SIT4 mutant was complemented with FgPP2A, FgSIT4 or FgPPG1 cDNA to generate the strain BY4741ΔSIT4+ pYES2-FgPP2A, BY4741ΔSIT4+ pYES2-FgSIT4, or BY4741ΔSIT4+ pYES2-FgPPG1, respectively. The wild-type (WT) strain BY4741 and SIT4 mutant BY4741ΔSIT4 transformed with empty pYES2 vector were used as controls. (c) FgTIP41 partially restored rapamycin sensitivity of the yeast TIP41 mutant. The yeast TIP41 mutant was complemented with FgTIP41 to generate the strain BY4741ΔTIP41+ pYES2-FgTIP41. The WT strain BY4741 and TIP41 mutant BY4741ΔTIP41 transformed with empty pYES2 vector were used as controls. (d) FgAREA partially complements the Saccharomyces cerevisiae GLN3 mutant. The yeast GLN3 mutant was complemented with FgAREA cDNA to generate the strain BY4741ΔGLN3+ pYES2-FgAREA. The WT strain BY4741 and GLN3 mutant BY4741ΔGLN3 transformed with empty pYES2 vector were used as controls. Serial dilutions of cell suspension of each strain were spotted on yeast peptone galactose medium (YPG) under different stresses as indicated in the figure.

In S. cerevisiae, Tap42-phosphatase complexes associate with TORC1 and exist mainly on membrane structures (Yan et al., 2006). Comparisons in the subcellular localization of FgTap42 in the budding yeast and that in F. graminearum may provide some clues for understanding its biological functions. As shown in Fig. 5(a), FgTap42 was distributed mainly in cytoplasm in the S. cerevisiae tap42-11 strain. However, in F. graminearum, FgTap42 was localized in the cytoplasm in mycelia and near the nucleus in conidia (Fig. 5b,c), indicating that FgTap42 may perform distinct functions in mycelia and in conidia. These results are consistent with the observation that rapamycin is highly effective against hyphal growth (Fig. 1), but not against conidial germination of F. graminearum (data not shown).

Figure 5.

Subcellular localization of FgTap42 in Saccharomyces cerevisiae and Fusarium graminearum. (a) Temperature-sensitive S. cerevisiae tap42-11 was complemented with GFP-tagged FgTap42. FgTap42-GFP was mainly observed in cytoplasm in S. cerevisiae. In F. graminearum, FgTap42-GFP was mainly localized to cytoplasm in hyphae (b) and to the nucleus in conidia (c). Nuclei in F. graminearum conidia were stained with 4′6-diamidino-2-phenylindole (DAPI). DIC, differential interference contrast; bars, 10 μm.

Tap42 interacts with Tip41 (a Tap42-interacting protein) in S. cerevisiae (Jacinto et al., 2001). However, interaction between FgTap42 and FgTip41 (FGSG_06963) in F. graminearum was not observed in the Y2H assays (Fig. 6a). In S. cerevisiae, Tap42 also interacts with the catalytic subunit of type 2A protein phosphatases (PP2As), including Pph3, Pph21, Pph22 and the type 2A-like phosphatases, Sit4 and Ppg1 (Di Como & Arndt, 1996; Wang et al., 2003). The F. graminearum genome contains three genes, FGSG_09815, FGSG_01464 and FGSG_05281, named FgPP2A, FgSIT4 and FgPPG1, respectively, that encode putative orthologs of Pph21/Pph22/Pph3, Sit4 and Ppg1 (Tables S1, S3). Y2H assays showed that FgTap42 interacts with FgPp2A, FgSit4 and FgPpg1 (Fig. 6a). The interaction of FgTap42 with FgSit4 and FgPp2A was further confirmed by affinity capture and co-immunoprecipitation assays (Table 1, Fig. 6b). In addition, FgPpg1 also interacted with FgTip41 in Y2H and Co-IP assays (Fig. 6a,b). These results suggest that the interaction framework of FgTap42, FgTip41 and FgPpg1 in F. graminearum is different from that observed in S. cerevisiae.

Table 1. Putative FgTap42, FgSit4 and FgPpg1 interacting proteins identified by affinity capture assays in Fusarium graminearum
ProteinsPutative functions
FgTap42 interacting proteins
FGSG_01464Serine/threonine-protein phosphatase FgSit4
FGSG_09815Serine/threonine-protein phosphatase FgPp2A
FGSG_01369Protein phosphatase PP2A regulatory subunit B
FGSG_02646Hypothetical protein
FGSG_05430Protein phosphatase PP2A regulatory subunit A
FGSG_06306T-complex protein 1 subunit eta
FGSG_06313T-complex protein 1 subunit zeta
FGSG_08704Hypothetical protein (FgKog1)
FGSG_10856Hypothetical protein
FgSit4 interacting proteins
FGSG_10524Sit4-associated protein
FGSG_05012Putative transcription factor
FGSG_06166Rab GDP-dissociation inhibitor
FGSG_08737Transcription factor (virulence-specific)
FGSG_09747Hypothetical protein
FgPpg1 interacting proteins
FGSG_06977Dual-specificity phosphatase similar to yeast Msg5 (FgMsg5)
FGSG_05430Protein phosphatase PP2A regulatory subunit A
FGSG_09019Putative transcription factor
FGSG_09709Putative transcription factor
FGSG_10980Hypothetical protein
Figure 6.

Yeast two-hybrid and co-immunoprecipitation (Co-IP) analyses of interactions among FgTap42, FgTip41 and the protein phosphatases FgPp2A, FgSit4 and FgPpg1 from Fusarium graminearum. (a) Serial dilutions of yeast cells (cells ml−1) transferred with the bait and prey constructs indicated in the figure were assayed for growth on yeast minimal synthetic defined base (SD) depleted of leucine, tryptophan, and histidine. The pair of plasmids pGBKT7-53 and pGADT7 was used as a positive control. The pair of plasmids pGBKT7-Lam and pGADT7 was used as negative control. The same set of yeast transformants were also assayed for β-galactosidase activity. (b) Co-IP assays. Immunoblots of total proteins extracted from F. graminearum transformants co-expressing the GFP and FLAG fusion constructs as indicated and proteins eluted from anti-FLAG agarose were detected with monoclonal anti-FLAG and monoclonal anti-GFP antibodies, respectively. Flow-through solution and total proteins isolated from the wild-type progenitor PH-1 served as controls.

FgPp2A, FgSit4 and FgPpg1 could partially complement the yeast sit4 mutant

In order to characterize the functions of FgPP2A, FgSIT4 and FgPPG1, we tested whether these genes can complement the yeast sit4 mutant (Hayashi et al., 2005). An expression vector pYES2 containing the full-length FgPP2A, FgSIT4 or FgPPG1 cDNA was transformed into the sit4 mutant, BY4741ΔSIT4. As a negative control, the mutant was also transformed with an empty pYES2 vector. As shown in Fig. 4(b), the growth defect of the yeast BY4741ΔSIT4 mutant at 37°C was partially restored by F. graminearum FgPP2A, FgSIT4 and FgPPG1. These results indicated that FgPP2A, FgSIT4 and FgPPG1 could function as the type 2A-like phosphatases in S. cerevisiae.

FgSit4 and FgPpg1 are important for hyphal development, virulence and sexual reproduction in F. graminearum

In order to determine the biological functions of FgPP2A, FgSIT4 and FgPPG1, we attempted to delete each gene in F. graminearum. For FgPP2A, we failed to obtain a null mutant after screening over 100 hygromycin-resistant transformants, indicating that it is likely an essential gene in F. graminearum. By contrast, the FgSIT4 deletion mutant (ΔFgSIT4) and FgPPG1 deletion mutant (ΔFgPPG1) grew remarkably more slowly than the WT strain on PDA (Fig. 7a). Nile Red staining showed that only few lipid droplets were observed in the hyphae of these mutants treated with 0.25 μg ml−1 rapamycin (Fig. 2a), indicating that accumulation of lipid droplets in F. graminearum triggered by rapamycin is dependent on FgSit4 and FgPpg1.

Figure 7.

Effects of FgSIT4, FgPPG1 or FgTIP41 deletion on mycelial growth and sexual development of Fusarium graminearum. (a) The wild-type PH-1, ΔFgSIT4,ΔFgPPG1, ΔFgTIP41 and the complemented strains ΔFgPPSIT4-C, ΔFgPPG1-C and ΔFgTIP4-C were incubated on potato dextrose agar (PDA) for 3 d. (b) Deletion of FgSIT4 or FgPPG1, but not FgTIP41, impaired the ability of F. graminearum to produce perithecia on carrot agar.

Sexual reproduction is a critical aspect of the F. graminearum lifecycle (Trail, 2009; Min et al., 2012). Deletion of either FgSIT4 or FgPPG1 abolished the production of perithecia on carrot agar (Fig. 7b). In infection assays with flowering wheat heads, ΔFgSIT4 and ΔFgPPG1 caused scab symptoms only in the inoculated spikelets (Fig. 8a). Under the same conditions, however, scab symptoms developed in > 90% spikelets when wheat heads were point-inoculated with the WT PH-1 or the complemented strains ΔFgSIT4-C and ΔFgPPG1-C (Fig. 8a).

Figure 8.

Effects of FgSIT4, FgPPG1 or FgTIP41 deletion on virulence and deoxynivalenol (DON) biosynthesis in Fusarium graminearum. (a) Wheat heads of susceptible cultivar Jimai22 were point-inoculated with conidial suspension of the wild-type PH-1, ΔFgSIT4, ΔFgPPG1, ΔFgTIP41 and the complemented strains ΔFgSIT4-C, ΔFgPPG1-C and ΔFgTIP4-C. Infected wheat heads were photographed 15 d post inoculation. The wheat head inoculated with water was used as a negative control (NK). (b) The levels of DON (per mg fungal ergosterol) produced by each strain in infected wheat kernels were detected after 20 d of inoculation. Error bars denote standard errors of three replicated experiments. Bars sharing the same letter denote values that are not significantly different at = 0.05.

FgSit4 and FgPpg1 show different roles in regulating conidiation and DON biosynthesis

Because DON is one of the important mycotoxins produced by F. graminearum, we assayed DON biosynthesis in the ΔFgSIT4 and ΔFgPPG1 mutants. After 20 d of incubation on sterilized wheat kernels, ΔFgPPG1 produced no detectable DON while the level of DON produced by ΔFgSIT4 was not significantly different from that of the WT strain (Fig. 8b). These results indicated that ΔFgSIT4 and ΔFgPPG1 play different roles in asexual development and secondary metabolism in F. graminearum.

When grown in MBB, ΔFgPPG1 produced relatively few conidia, whereas the ΔFgSIT4 produced similar numbers of conidia as the WT strain (Fig. 9a). Interestingly, conidia of ΔFgSIT4 and ΔFgPPG1 had fewer septa than those of the WT strain (Fig. 9b), indicating that FgSit4 and FgPpg1 may be involved in regulating septum formation.

Figure 9.

Impacts of FgSIT4 and FgPPG1 deletion on conidiogenesis of Fusarium graminearum. (a) Numbers of conidia were examined after the wild-type PH-1, ΔFgSIT4, ΔFgSIT4-C, ΔFgPPG1 and ΔFgPPG1-C strains were incubated in mung bean broth (MBB) for 7 d. Error bars in each column denote standard errors of three replicated experiments. Bars sharing the same letter denote values that are not significantly different at = 0.05. (b) Percentage of conidia with different septum numbers in PH-1, ΔFgSIT4, ΔFgSIT4-C, ΔFgPPG1 and ΔFgPPG1-C. A total of 200 conidia were examined for each strain.

Involvement of FgSit4 and FgPpg1 in cell wall integrity

In order to further explore the specific function of FgSit4 and FgPpg1 in F. graminearum, we prepared serial dilutions of spores of each mutant and placed them under a variety of stress conditions. ΔFgSIT4 and ΔFgPPG1 exhibited dramatically increased sensitivity to the cell wall damaging agents Congo red and calcofluor white (Fig. 10), but not to osmotic stress mediated by NaCl, KCl and sorbitol, and oxidative stress mediated by H2O2 and paraquat (data not shown). Western blot analysis showed that the phosphorylation level of FgMgv1 – a core MAP kinase in the cell wall integrity (CWI) pathway – was decreased considerably in ΔFgSIT4 and ΔFgPPG1 (Fig. 11b).

Figure 10.

Sensitivity of FgSIT4, FgPPG1, and FgMSG5 deletion mutants of Fusarium graminearum to cell wall damaging agents. Serial dilutions of conidial suspensions of the wild-type (WT), ΔFgSIT4, ΔFgSIT4-C, ΔFgPPG1, and ΔFgPPG1-C were spotted onto minimal medium (MM) containing 2 mg ml−1 congo red or 100 μg ml−1 calcofluor white. Colony growth was inspected after 2 d of incubation at 25°C. For ΔFgMSG5, serial dilutions of conidia suspensions of the WT and ΔFgMSG5 were spotted onto MM containing 3 mg ml−1 congo red or 200 μg ml−1 calcofluor white. Colony growth was inspected after 3 d of incubation at 25°C.

Figure 11.

Analysis of the relationships among FgSit4, FgPpg1, FgMgv1 and FgMsg5 of Fusarium graminearum. (a) Yeast two-hybrid assay of interactions among these proteins. Serial dilutions of yeast cells (cells ml−1) transferred with the bait and prey constructs indicated in the figure were assayed for growth on yeast minimal synthetic defined base (SD) depleted of leucine, tryptophan, and histidine. The pair of plasmids pGBKT7-53 and pGADT7 was used as a positive control. The pair of plasmids pGBKT7-Lam and pGADT7 was used as negative control. The same set of yeast transformants were also assayed for β-galactosidase activity. (b) Phosphorylation of FgMgv1 in the FgSIT4, FgPPG1, and FgMSG5 deletion mutants. Total proteins were isolated from hyphae of the wild-type PH-1, ΔFgSIT4, ΔFgPPG1 and ΔFgMSG5. FgMgv1 and phosphorylated FgMgv1 proteins were detected using the yeast Mpk1 (yN-19) antibody and phospho-p44/42 MAP kinase antibody, respectively.

Both FgSit4 and FgPpg1 function as phosphatases that may dephosphorylate target proteins. How could the disruption of FgSIT4 or FgPPG1 lead to decreased phosphorylation of FgMgv1? To resolve this paradox, affinity capture assays were performed for FgPpg1. Briefly, FgPpg1 was tagged with 3× FLAG, and proteins co-purified with FgPpg1-3× FLAG were analyzed by mass spectrometry. Interestingly, FgMsg5 (encoded by FGSG_06977) was co-purified with FgPpg1-3× FLAG (Table 1). BLAST analysis revealed that FgMsg5 is homologous to S. cerevisiae ScMsg5, which functions as a dual-specificity phosphatase that dephosphorylates phosphorylated Slt2 (Flandez et al., 2004). Therefore, we hypothesized that FgMsg5 is a negative regulator of the CWI pathway in F. graminearum. Western blot revealed that the phosphorylation level of FgMgv1 in ΔFgMSG5 was elevated compared with the WT progenitor (Fig. 11b). In addition, ΔFgMSG5 revealed increased resistance to the cell damaging agent calcofluor white (Fig. 10). Y2H assays further showed that FgMsg5 interacts physically not only with FgSit4 and FgPpg1, but also with FgMgv1 (Fig. 11a), an ortholog of S. cerevisiae Slt2. These results indicate that FgMsg5 is a negative regulator of FgMgv1, and that FgSit4 and FgPpg1 positively regulate the phosphorylation level of FgMgv1 via FgMsg5.

FgTip41 is involved in the regulation of hyphal growth, DON production and virulence

In S. cerevisiae, deletion of the Tap42-interacting gene, TIP41, confers rapamycin resistance but has no detectable effect on cell growth under normal culture conditions (Jacinto et al., 2001). F. graminearum has one TIP41 ortholog, FgTIP41 (FGSG_06963). When it is expressed in yeast, FgTIP41 partially restored rapamycin sensitivity in the tip41 mutant (Fig. 4c). However, the FgTIP41 deletion mutant (ΔFgTIP41) had no detectable changes in sensitivity to rapamycin (data not shown). Furthermore, we failed to detect any physical interaction between FgTap42 and FgTip41 by Y2H assays (Fig. 6a). Unlike the tip41 mutant of S. cerevisiae, deletion of FgTIP41 in F. graminearum led to a considerable defect in hyphal growth on PDA, and the defect was fully restored in the complemented strain ΔFgTIP41-C (Fig. 7a). These results indicate that the functions of FgTip41 in F. graminearum differ from its ortholog in the budding yeast.

In infection assays with flowering wheat heads, the deletion of FgTIP41 resulted in considerable reduction in virulence (Fig. 8a). Because ΔFgTIP41 was impaired in virulence and DON is considered as an important virulence factor in F. graminearum (Proctor et al., 1995; Seong et al., 2009), we analyzed the effects of FgTIP41 deletion on DON biosynthesis. After 20 d, the amount of DON produced by ΔFgTIP41 on wheat kernels was decreased > 60% in comparison with the WT strain (Fig. 8b). These results indicated that FgTip41 is involved in regulation of DON biosynthesis, subsequently affecting virulence in F. graminearum.

FgAreA is associated with conidiation, mycotoxin production and pathogenicity

In S. cerevisiae, the transcriptional activators Gln3 and Gat1 (Nil1) are phosphorylated in a TOR-dependent manner under N-sufficient conditions (Rohde & Cardenas, 2004). In the filamentous fungus Aspergillus nidulans, the major regulator of N metabolism, the GATA factor AreA (Kudla et al., 1990), is an ortholog of Gln3 and Gat1. The F. graminearum genome contains one AREA ortholog (FGSG_08634) named FgAREA. When expressed in yeast, FgAREA could partially restore tolerance of the Gln3 mutant to rapamycin and paraquat (Fig. 4d). Consistent with an earlier report on systematic characterization of AreA in F. graminearum (Min et al., 2012), ΔFgAREA was impaired in growth on a medium containing nitrate as the sole N source, and showed severely reduced virulence and DON production on wheat heads (Fig. S5).


The benzimidazole and triazole fungicides, which have been used widely for the control of FHB, normally provide only c. 50% reduction of FHB index and 40% reduction in DON (Blandino et al., 2006). Therefore, the initial objective of this study was to test the efficacy of the macrolide antibiotic, rapamycin, against F. graminearum. We found that rapamycin strongly inhibits hyphal growth of F. graminearum. In comparison with tebuconazole and carbendazim, rapamycin showed a stronger inhibitory effect. The cytological examination showed that rapamycin treatment led to the accumulation of lipid droplets in F. graminearum hyphae (Fig. 2a), which has not been reported previously in S. cerevisiae and other fungi. Similar to F. graminearum, Drosophila contains a single Tor kinase. Tor kinase in Drosophila is inactive under rapamycin treatment, and cells lacking Tor activity result in lipid vesicle aggregation (Zhang et al., 2000). Based on these results, we presumed that rapamycin may have similar effects on lipid metabolism in filamentous fungi and in insects.

In S. cerevisiae, rapamycin exerts antifungal activity via its interaction with the prolyl isomerase Fkbp12 to form a binary complex, which binds to the conserved FRB domain of Tor kinases. Fkbp12 catalyzes cis-trans peptidyl-prolyl isomerization, a rate-limiting step in protein folding (Hur & Bruice, 2002). In our current study, we found that the FgFKBP12 deletion mutant of F. graminearum was resistant to rapamycin and FK506 (Fig. S3). Resistance of FKBP12 deletion mutants to rapamycin has been documented in several other fungi, including A. nidulans, B. cinerea, Candida albicans, Cryptococcus neoformans, Fusarium fujikuroi and Mucor circinelloides (Cruz et al., 1999, 2001; Fitzgibbon et al., 2005; Teichert et al., 2006; Melendez et al., 2009; Bastidas et al., 2012). In addition, rapamycin affects sexual development rather than vegetative growth, and this development is mediated by Fkbp12 via a mechanism independent of the TOR pathway, as shown in fission yeast (Weisman et al., 2001). In addition, the deletion of FKBP12 (named BcPIC5) in B. cinerea strain T4 led to reduced virulence on Arabidopsis thaliana. By contrast, deletion of BcFPKP12 did not impair virulence in B. cinerea strain B05.10 on tomato and grape fruit, and even led to faster colonization of apple sections and cucumber cotyledons (Melendez et al., 2009). These results strongly indicate that roles of Fkbp12 orthologs vary considerably in different fungi, likely in a species-specific or even strain-specific manner.

In S. cerevisiae, the Tor downstream effector Tap42 is an essential protein (Di Como & Arndt, 1996). Similarly, we were unable to obtain a F. graminearum FgTAP42 null mutant. Complementation of a temperature-sensitive S. cerevisiae tap42-11 mutant with FgTAP42 further indicates that Tap42 protein has conserved functions in yeast and filamentous fungi. In S. cerevisiae, the interaction of Tap42 with PP2A or PP2A-like proteins (e.g. Pph3, Pph21, Pph22, Sit4 and Ppg1) is regulated by the TOR signaling pathway in response to rapamycin treatment or nutrient deprivation. Inactivation of Tor induces a rapid activation of PP2As, which is accompanied by their dissociation from Tap42, suggesting that Tap42 acts as an inhibitor of PP2As (Jacinto et al., 2001). Other studies have indicated that Tap42 may play a positive role in stimulating phosphatase activity, at least for certain substrates of the phosphatase (Cherkasova & Hinnebusch, 2003; Duvel et al., 2003). Among the five protein phosphatases interacting with Tap42, Sit4 is the only one that plays a critical role in cell growth and proliferation in S. cerevisiae (Di Como & Jiang, 2006). Deletion of PPH3, PPH21, PPH22 or PPG1 had little effect on cell growth (Jiang, 2006). Deletion of SIT4, however, led to a variety of phenotypic changes, such as decreased vegetative growth (Yoshikawa et al., 2011), de-repressed growth in a high glucose environment (Yoshikawa et al., 2011), and increased sensitivity to hygromycin and to high temperature (Auesukaree et al., 2009). In addition, Sit4 is required for proper telomere function (Hayashi et al., 2005), monovalent ion and pH homeostasis (Masuda et al., 2000), and initiation of translation (Montero-Lomeli et al., 2002). In F. graminearum, we identified three FgTap42-interacting PP2A and PP2A-like proteins (e.g. FgPp2A, FgSit4 and FgPpg1), and all of these genes partially complemented the yeast sit4 mutant (Fig. 4b). Unlike S. cerevisiae, where only Sit4 plays a critical role in cell growth, all three FgSit4-interacting phosphatases are important for fungal growth in F. graminearum; the deletion of FgPP2A is predicted to be lethal, and the deletion of either FgSIT4 or FgPPG1 led to severely reduced mycelial growth.

Comparative analysis showed that FgSit4 and FgPpg1 are homologous to S. cerevisiae Sit4 and Ppg1, respectively, whereas FgPp2A is homologous to Pph3, Pph21 and Pph22 (Table S3). Similar to S. cerevisiae Sit4, FgSit4 also plays important roles in regulation of various cellular processes, including mycelial growth, virulence and sexual development in F. graminearum. In S. cerevisiae, while deletion of PPH21 or PPH22 did not result in notable phenotypic deficiencies, double mutation eliminated 80–90% of total PP2A activity in the cell and led to severely crippled growth (Sneddon et al., 1990). The residual PP2A activity in the absence of Pph21 and Pph22 is believed to be contributed by Pph3, because the deletion of PPH3 is lethal in the Pph21/Pph22 double mutant (Sneddon et al., 1990). Because F. graminearum contains a single FgPP2A, which is homologous to PPH3, PPH21, and PPH22 (Table S3), it was not surprising to discover that the deletion of FgPP2A in F. graminearum is lethal.

In S. cerevisiae, deletion of PPG1 has no obvious effects on cell growth (Jiang, 2006). However, deletion of its ortholog FgPPG1 in F. graminearum resulted in various defects, including reduced mycelial growth and impaired asexual and sexual development. Subsequently, we hypothesized that the important functions of FgPpg1 in F. graminearum are mediated by its interaction with FgTip41. Tip41 was identified as a binding partner of Tap42 in the budding yeast, and has been suggested as a regulator of the rapamycin-sensitive signaling pathway by competing for Tap42 against Sit4 (Jacinto et al., 2001). Interestingly, we found that FgTip41 interacts with with FgPpg1 rather than FgTap42, which may explain why FgPpg1 was not co-purified with FgTap42 in our affinity capture assays (Table 1). In addition, we also observed the interaction between FgPpg1 and FgTap42 in the Y2H assay, indicating that FgTap42, FgPpg1 and FgTip41 may form a heterotrimer in F. graminearum. To our knowledge, this is the first report on the existence of this heterotrimer (FgTip41:FgPpg1:FgTap42) in a fungal species. Our finding is in contrast with the model proposed in S. cerevisiae, but in agreement with recently published study on human proteins. The mammalian ortholog of Tip41 (TIPRL) does not interact with the mammalian Tap42 ortholog (α4), but TIPRL and α4 can bind PP2A simultaneously, forming a stable ternary complex, which regulates the activity of PP2A (Murata et al., 1997).

FgSit4 and FgPpg1 appear to be involved in several functions including vegetative development and virulence in F. graminearum. In this study, we found for the first time that FgSit4 and FgPpg1 positively regulate phosphorylation of FgMgv1 via interacting with a negative regulator FgMsg5, which is a negative regulator of the CWI pathway (Flandez et al., 2004). Because the CWI pathway is known to carry out various cellular functions including regulating virulence, mycelial growth, sexual and asexual development, and secondary metabolism in F. graminearum (Hou et al., 2002), FgSit4 and FgPpg1 may be involved in regulating various cellular processes partially via the CWI pathway in F. graminearum. In addition, SAGE analyses of gene expression profiling revealed that several transcription factor genes (FGSG_05304, FGSG_06542 and FGSG_01877) were highly upregulated in the FgPPG1 deletion mutant (data not shown). Similar to the FgPPG1 mutant, deletion mutants of these transcription factor genes exhibit reduced mycelial growth rate, decreased DON production, less conidiation or impaired virulence on wheat head (Son et al., 2011), indicating that FgPpg1 may regulate a variety of regulatory pathways via different transcription factors (Fig. 12).

Figure 12.

A proposed target of the rapamycin (TOR) pathway in Fusarium graminearum. Rapamycin forms a complex with FgFkbp12, and this complex can bind to and inhibit FgTor, the key component of TOR complex. Meanwhile, FgTap42, a downstream component of FgTor complex, exhibits its function by interacting with three phosphatases FgPp2A, FgSit4 and FgPpg1. FgSit4 and FgPpg1 are involved in the regulation of cell wall integrity via a negative regulator FgMsg5. In addition, FgPpg1 also regulates deoxynivalenol (DON) biosynthesis and virulence via its downstream transcription factors FgAreA, FGSG_09019 and FGSG_09709. The function of FgPp2A remains inconclusive. Solid arrows and solid with perpendicular lines indicate the positive and negative regulation patterns, respectively. The dashed arrows indicate putative connections.

Inhibition of TOR by rapamycin causes a nutrient stress response in S. cerevisiae. Both N starvation and rapamycin treatment resulted in rapid dephosphorylation of Gln3 by Sit4 (Beck & Hall, 1999). Dephosphorylated Gln3 is imported into the nucleus, and activates genes involved in the assimilation of alternative N sources (Beck & Hall, 1999; Cooper, 2002). The functions of Tor in N metabolism remain controversial in the filamentous fungus A. nidulans. Tor was proposed to have only a minor role in N metabolism in A. nidulans because mutations in the TOR pathway genes did not have notable effects on the N utilization phenotype (Fitzgibbon et al., 2005). However, in Fusarium oxysporum, rapamycin treatment increases transcript levels of genes involved in N catabolism, such as NIT1, supporting a role of TOR in N catabolite repression (Lopez-Berges et al., 2010). This finding is in agreement with a previous observation that rapamycin activates expression of N metabolism-related genes encoding ammonium permease and uricase in Fusarium fujikuroi (Teichert et al., 2006). Similar to these findings, we found that rapamycin treatment led to increased expressions of several genes involved in N metabolism in F. graminearum, such as genes encoding 2-nitropropane dioxygenase and NAD-specific glutamate dehydrogenase (data not shown). Furthermore, the F. graminearum ortholog of Gln3, FgAreA, could fully complement the budding yeast GLN3 deletion mutant (Fig. 4d). In addition, ΔFgAREA and ∆FgPPG1 showed similar defects in conidiation, DON production and pathogenicity. Collectively, we propose that FgAreA is one of the FgPpg1 downstream components in the TOR signaling pathway in F. graminearum (Fig. 12).

In conclusion, we found that the TOR pathway plays critical roles in regulating vegetative differentiation and virulence in F. graminearum, which advances our understanding of the pathogenesis of plant pathogenic fungi.


This research was supported by National Key Basic Research and Development Program (2013CB127802), National Science Foundation (31170135), Special Fund for Agro-scientific Research in the Public Interest (no. 201303023), and China Agriculture Research System (CARS-3-1-15).