A VASt-domain protein regulates autophagy, membrane tension, and sterol homeostasis in rice blast fungus

ABSTRACT

Sterols are a class of lipids critical for fundamental biological processes and membrane dynamics. These molecules are synthesized in the endoplasmic reticulum (ER) and are transported bi-directionally between the ER and plasma membrane (PM). However, the trafficking mechanism of sterols and their relationship with macroautophagy/autophagy are still poorly understood in the rice blast fungus Magnaporthe oryzae. Here, we identified the VAD1 Analog of StAR-related lipid transfer (VASt) domain-containing protein MoVast1 via co-immunoprecipitation in M. oryzae. Loss of MoVAST1 resulted in conidial defects, impaired appressorium development, and reduced pathogenicity. The MoTor (target of rapamycin in M. oryzae) activity is inhibited because MoVast1 deletion leads to high levels of sterol accumulation in the PM. Site-directed mutagenesis showed that the 902 T site is essential for localization and function of MoVast1. Through filipin or Flipper-TR staining, autophagic flux detection, MoAtg8 lipidation, and drug sensitivity assays, we uncovered that MoVast1 acts as a novel autophagy inhibition factor that monitors tension in the PM by regulating the sterol content, which in turn modulates the activity of MoTor. Lipidomics and transcriptomics analyses further confirmed that MoVast1 is an important regulator of lipid metabolism and the autophagy pathway. Our results revealed and characterized a novel sterol transfer protein important for M. oryzae pathogenicity.

Abbreviations: AmB: amphotericin B; ATMT: Agrobacterium tumefaciens-mediated transformation; CM: complete medium; dpi: days post-inoculation; ER: endoplasmic reticulum; Flipper-TR: fluorescent lipid tension reporter; GO: Gene ontology; hpi: hours post-inoculation; IH: invasive hyphae; KEGG: kyoto encyclopedia of genes and genomes; MoTor: target of rapamycin in Magnaporthe oryzae; PalmC: palmitoylcarnitine; PM: plasma membrane; SD-N: synthetic defined medium without amino acids and ammonium sulfate; TOR: target of rapamycin; VASt: VAD1 Analog of StAR-related lipid transfer; YFP, yellow fluorescent protein.

Introduction

Autophagy is an evolutionarily conserved cellular pathway in eukaryotes. The process of autophagy involves the sequestering of cytosolic content in autophagosomes that are then delivered to the vacuole/lysosome for degradation or recycling [1,2]. Autophagy plays a vital role in a multitude of physiological and pathophysiological processes, such as stress, starvation, and microbial infection [3]. In recent decades, mounting evidence has shown that autophagy is crucial for the growth and infection-related development of plant pathogenic fungi [4]. In the rice blast fungus, Magnaporthe oryzae, the autophagy-related proteins MoAtg1-MoAtg10, MoAtg12, MoAtg14-MoAtg16 and MoAtg18 are essential for appressorium maturation, penetration peg formation, and glycogen accumulation [5,6]. Disruption of these ATG genes results in loss of pathogenicity in M. oryzae [7]. In Fusarium graminearum, the major causal agent of Fusarium head blight (FHB), autophagy is required for proper vegetative growth, asexual/sexual reproduction, full virulence, and toxin biosynthesis [8,9]. In the corn smut Ustilago maydis, Atg1 and Atg8 are important for normal budding of haploid sporidia, nutrient starvation survival, and pathogenic development [10]. Although the biological functions of autophagy have been progressively uncovered, the molecular and regulatory mechanisms of autophagy in plant pathogenic fungi are still largely unknown.

The highly conserved protein kinase Tor (target of rapamycin) is a central cell growth regulator involved in cellular metabolism, growth, and suppressing catabolic processes [11]. In animals, MTORC1 regulates autophagy by directly phosphorylating and inhibiting ULK1 (the autophagy-initiating kinase, a mammalian homolog of yeast Atg1) [12,13]. The MTORC1 protein is regulated by the upstream targets phosphoinositide 3-kinase (PI3K) and AKT, which are activated through nutrition signals and growth factors in mammalian cells [14]. In Saccharomyces cerevisiae, Tor1 and Tor2 form two complexes, TORC1 and TORC2. Like MTORC1, TORC1 is sensitive to rapamycin and mainly regulates cell growth and autophagy. TORC2 responds to osmotic stress by activating the stress response [15,16]. In M. oryzae, Tor pathways are also involved in modulating autophagy by coordinating with the carbon responsive regulator Abl1 (AMP activated protein kinase β subunit-like protein), glutaminolysis regulator Asd4 (ascus development 4), pre-mRNA processing protein Rbp35 (RNA-binding protein) and the glucose-metabolizing enzyme Tkl1 (transketolase), which are involved in host infection [17–20]. Recent evidence indicates that Imp1, MoSnt2, Abl1 and Asd4 also regulate autophagy via Tor signaling [2,18–21]. These studies have enriched the understanding of the upstream regulatory mechanism of autophagy in plant pathogenic fungi.

Sterols, essential components of eukaryotic membranes, are synthesized in the endoplasmic reticulum (ER) and delivered to the plasma membrane (PM) by lipid transfer proteins (LTPs) [22]. Cholesterol, a type of sterol, is found in animals and is required to build and maintain membranes, for example, by modulating membrane fluidity over a range of physiological temperatures [23]. Ergosterol is found in cell membranes of fungi and protozoa, serving similar functions as cholesterol in animal cells [24]. Regardless of whether they occur in animal cells or fungi, most sterols are distributed in the PM. In animals, they occupy 30–40% of the total lipid molecules of the PM, which account for 60–80% of total cellular cholesterol [25,26]. Although all sterols are synthesized in the ER, the ER contains only about 0.5–1% of cellular cholesterol [27]. Accordingly, mechanisms of sterol transport between the ER and the PM are important for cell survival in eukaryotic organisms.

Decades of studies have shown that membrane lipids are synthesized in the ER by multiple biosynthetic enzymes, and then transferred to other intracellular membrane systems via vesicular or non-vesicular trafficking pathways [28]. In the non-vesicular trafficking pathway, the oxysterol binding protein (OSBP) family transfers sterols and phosphatidylinositol 4-monophosphate (PtdIns4P) from the ER to the Golgi in animals [29]. In yeast, there are several classes of sterol-specific lipid transfer proteins, including non-vesicular transport Nvj2 and oxysterol binding protein (OSBP) homologs the Osh proteins, which transfer ceramide and sterol from the ER to the Golgi or to the PM, respectively [30,31]. In vesicular trafficking, polymerized cytosolic coat protein complex COPII exports a large amount of lipids from the ER membrane at very high rate. An intricate regulatory mechanism involving COPII-dependent vesicular transport was essential for cholesterol synthesis and homeostasis maintenance [32]. However, blocking vesicular traffic using genetic or pharmacological methods only partially inhibits transportation of glycerophospholipids and sterols to the PM, indicating that non-vesicular mechanisms function directly in lipid transfer between the ER and the Golgi [33].

Recently, a new family of ER membrane–anchored StART-like (also called VASt) domain-containing proteins were discovered and found to be conserved in yeast (Saccharomyces cerevisiae) and humans [22]. Members of this protein family are located at the ER–PM, ER–mitochondria, and ER–vacuole junctions, which transfer and regulate the distribution of sterols [22]. In animals, sterol transfer is likely controlled by GRAMD1A, GRAMD1B and GRAMD1C, which bind cholesterol [34]. In S. cerevisiae, there are six StART-like domain-containing proteins, Lam1, Ysp2/Lam2, Sip3/Lam3, Lam4, Lam5 and Lam6, that directly transfer sterols between the ER and the PM [22]. However, the functions of StART-like domain-containing proteins have not been explored in detail in plant pathogenic fungi.

In this study, a novel VASt domain protein was identified in M. oryzae and named MoVast1. We found that MoVast1 is anchored to the membrane system and is involved in sterol homeostasis in M. oryzae. In addition, we determined that the sterol content influences PM tension. By analyzing autophagy flux and lipidomics and transcriptomics profiles, we found that MoVast1 responds to external stress by regulating the sterol content in the PM. Also, MoVast1 regulates the activity of MoTor (and autophagy) by controlling the sterol distribution. Furthermore, phenotypic analyses confirmed that MoVast1 regulates conidiation and pathogenicity of the rice blast fungus. Our results reveal how sterols are distributed in the fungus M. oryzae and their importance to autophagy in this important rice pathogen.

Results

Identification of VASt domain-containing proteins involved in autophagy

Previous studies indicated that the autophagy process is dependent on core ATG proteins, which participate in the formation of autophagosomes [9,35]. To identify novel proteins that are potentially related to autophagy, we tagged autophagosomes using GFP-MoAtg8, and enriched proteins by co-immunoprecipitation (CoIP) followed by mass spectrometry. A total of 360 proteins were identified (Data S1). Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses showed that these proteins are involved in pathways such as protein translation, pyruvate metabolism, pyrimidine metabolism, glycolysis, the TCA cycle, and autophagy (Fig. S1A and S1B). To verify the functions of these proteins, 11 potential autophagy-regulated genes were deleted using a high-throughput gene knockout method (Table S1, Fig. S2A). We found that a VASt domain-containing protein MoVast1 (MGG_11211) is involved in the autophagy pathway.

To further confirm the interaction between MoVast1 and MoAtg8, we next performed a bimolecular fluorescence complementation assay (BiFC). The N-terminal fragment of MoVast1 was fused to the PKD2-YFPN vector and the N-terminal fragment of MoAtg8 was fused to the PKD5-YFPC vector. Then, the YFPN-MoVast1 and YFPC-MoAtg8 constructs were cotransformed into the wild-type strain Guy11. YFPN/YFPC-MoAtg8, and YFPN-MoVast1/YFPC were co-transformed into Guy11 as negative controls. YFP signals were only detected in the transformants co-expressing YFPN-MoVast1 and YFPC-MoAtg8, but not in the negative controls (Figure 1A), suggesting that MoVast1 interacts with MoAtg8 in vivo. The interaction of MoVast1 and MoAtg8 was also validated by the CoIP assay. GFP-MoAtg8 or GFP alone were separately transformed into the ΔMovast1 mutant complemented with MoVast1-Flag. The MoVast1 bands were detected in GFP-MoAtg8 immunoprecipitation lysis but were not found in GFP immunoprecipitation lysis (Figure 1B). These results show that MoVast1 interacts with MoAtg8 in M. oryzae.

Figure 1.

MoVast1 interacts with MoAtg8. (A) Visualization of the MoVast1-MoAtg8 interaction using the BiFC assay. YFP signals were observed in vegetative hyphae of the transformant harboring YFPN-MoVast1 and YFPC-MoAtg8. Bar: 10 μm. No detectable YFP signals were observed in the negative control transformant harboring YFPN and YFPC-MoAtg8, or YFPN-MoVast1 and YFPC. Bar: 10 μm. (B) Co-immunoprecipitation assay (CoIP) between MoVast1 and MoAtg8. MoVast1-Flag and GFP-MoAtg8 were co-expressed in the ΔMovast1 mutant, and proteins were detected using anti-Flag and anti-GFP antibodies. Lysed hyphal proteins were allowed to bind to GFP beads at 4°C and analyzed by immunoblot with anti-Flag and anti-GFP antibodies

MoVast1 is involved in conidiation, appressorium formation, and virulence

To understand the physiological functions of MoVast1 in M. oryzae, the MoVAST1 gene was deleted in the wild-type strain Guy11 (Fig. S2B and S2C). The three deletion mutants showed similar phenotypes (Fig. S3A and S3B). We selected one of the mutants, designated as ΔMovast1, and constructed the complementation strain (ΔMovast1 complemented with GFP-MoVast1) for further analysis. Compared with the wild-type strain Guy11, the ΔMovast1 mutant showed small colonies and fewer conidia when grown on complete medium (CM) for 9 d at 25ºC (Figure 2A, 2B and 2D). Next, we examined the conidia morphology and appressorium formation in Guy11, ΔMovast1, and the complementation strain. In the ΔMovast1 mutant, the conidia were short and abnormal compared to the wild type and the complementation strain (Figure 2C). The formation of appressoria was delayed on the artificial inductive surface, and the appressorial morphology was abnormal. The three types of appressoria (type 1 looks normal, type 2 is a vacuolation appressorium with multiple branches, type 3 is a mixture of type 1 and type 2) were observed in ΔMovast1, and a high proportion of type 2 and type 3 occurred in the ΔMovast1 mutant (Figure 2E and F). We concluded that MoVast1 is essential not only in mycelial growth and conidiation, but also in appressorium development.

Figure 2.

MoVast1 is involved in conidiation, appressorium formation, and virulence. (A) The colony morphology of Guy11, ΔMovast1, and the complementation strain in CM medium at 25°C for 9 d. (B) The conidia and conidiophores of Guy11, ΔMovast1, and the complementation strain. The conidia and conidiophores were induced for 24 h at 25°C after collection from CM medium at 25°C for 9 d. Bar: 50 μm. (C) The conidia morphology of Guy11, ΔMovast1, and the complementation strain. Bar: 10 μm. (D) The colony growth rate, conidiation, conidial germination and appressorium formation (at 6 hpi) in Guy11, ΔMovast1, and the complementation strain. Each strain was inoculated onto CM plates, and conidia were gathered using 5 ml water by filters. The data was analyzed by GraphPad Prism 7.0. Asterisks indicate statistically significant differences (*** p < 0.01). (E) The morphology of appressoria in Guy11, ΔMovast1, and the complementation strain on a hydrophobic plastic coverslip at 22°C for 24 h. Bar: 10 μm. (F) Different types of appressoria in Guy11, ΔMovast1, and the complementation strain. The data were analyzed by one-way analysis of variance (ANOVA) in Prism 7.0

To examine whether MoVast1 regulates the pathogenicity of M. oryzae, virulence assays were performed on two different susceptible plant hosts (rice and barley). Conidial suspensions (5 × 10⁴ conidia/ml) of ΔMovast1, wild type, or complementation strain were sprayed onto 2-week-old rice seedlings (CO-39). At 7 d post-inoculation (dpi), the ΔMovast1 mutant produced small, restricted lesions, in contrast to the typical fused lesions of the wild type and the complementation strain (Figure 3A). Host penetration assays were performed to gain further insight into the effects of the ΔMovast1 mutant on disease progression. Mycelial plugs of Guy11, ΔMovast1, and the complementation strain were inoculated on detached barley leaves. At 4 dpi, the wild type and the complementation strain caused severe lesions, while only small lesions could be detected in the ΔMovast1 mutant (Figure 3B).

Figure 3.

MoVast1 is involved in virulence. (A) Rice spraying assays to detect pathogenicity. The conidial suspension (5 × 10⁴ conidia/ml) of each strain were used for spraying and were cultured for 7 d at 25°C. (B) Disease symptoms on cut leaves of barley inoculated with mycelial plugs from Guy11, ΔMovast1, and the complementation strain at 25°C for 4 d. (C) Live-cell imaging of Guy11 and ΔMovast1 infecting susceptible CO-39 rice leaf sheaths at 6, 12 and 24 hpi. Bar: 10 μm. (D) Appressorium formation rate in Guy11 and ΔMovast1 at 6, 12, 24 hpi. Values are the mean of three independent replicates, *** p < 0.01. (E) Live-cell imaging of Guy11 and ΔMovast1 infecting susceptible CO-39 rice leaf sheaths at 48 hpi. Bar: 10 μm. (F) Three types of IH were quantified and statistically analyzed in Guy11 and ΔMovast1. Error bars represent the SD. (G) Appressorium collapse was analyzed in Guy11, ΔMovast1 and the complementation strain. The conidia were induced on a hydrophobic plastic slip for 24 h. Values are the mean of three independent replicates. Bar: 10 μm

To investigate how MoVast1 impacted infection-related development, detached rice leaf sheath assays were implemented in the ΔMovast1 mutant and the wild-type strain Guy11. At 6 h post-inoculation (hpi), more than 70% of appressoria were formed in Guy11, however, less than 10% of appressoria had formed in the ΔMovast1 mutant. At 12 hpi, there were no significant differences in appressorium formation between the wild-type Guy11 and the ΔMovast1 mutant. At 24 hpi, invasive hyphae (IH) were formed in Guy11, but no IH appeared in ΔMovast1 (Figure 3C and D). To gain further insights into the effects of MoVast1 on disease progression in rice, IH were examined on rice leaf sheaths at 48 hpi. In Guy11 and the complemented strain, nearly 90% of appressoria formed IH structures, and more than 50% of IH colonized into the adjacent cells. In contrast, only 20% of ΔMovast1 appressoria formed IH structures, and no IH could colonize into the adjacent cells at 48 hpi (Figure 3E and F). Considering that M. oryzae ruptured the cuticle of the host through the mechanical force imposed by high appressorium turgor, we evaluated the appressorium turgor pressure using the incipient cytorrhysis test as described previously [36]. The ratio of collapsed appressoria in ΔMovast1 was significantly higher than that of the wild type in 1.5 M glycerol in type 1 appressoria. The vacuolation appressoria (type 2 appressoria) could not collapse even in 3 M glycerol (Figure 3G). As a result, the morphology and turgor pressure were impaired in the ΔMovast1 mutant.

MoVast1 is localized in ER and PM sites and co-localizes with MoAtg8

To elucidate the subcellular localization of the MoVast1 protein in M. oryzae, GFP-MoVast1 fusion expression cassettes were transformed into the ΔMovast1 mutant. Under a fluorescence microscope, GFP-MoVast1 appeared in an ER-like structure in the dormant conidia (Figure 4A, 0 h). At 2 hpi in sterile water on a glass slide, GFP-MoVast1 appeared as puncta near the PM (PM labeled by FM4-64) (Figure 4B, 2 h). These results indicate that MoVast1 could localize to the PM.

Figure 4.

MoVast1 is localized to ER and PM sites. (A) The localization of MoVast1. The conidial suspension (5 × 10⁴ conidia/ml) of the ΔMovast1 mutant complemented with GFP-MoVast1 was placed on a glass slide then observed through a fluorescence microscope at 0 and 2 hpi after being dyed by FM4-64 for 5 min. Bar: 10 μm. Fluorescence densities of GFP-MoVast1 and FM4-64 were analyzed using ImageJ software. (B) The co-localization of MoVast1 and ER. GFP-MoVast1 and MoLhs1-mCherry were co-expressed in the ΔMovast1 mutant. Cultures were grown for 7 d in CM medium. The conidial suspension (5 × 10⁴ conidia/ml) was placed on a glass slide, then observed through a fluorescence microscope at 0 and 2 hpi. Bar: 10 μm. Fluorescence densities of GFP-MoVast1 and MoLhs1-mCherry were analyzed using ImageJ software. (C) The co-localization of MoVast1 and MoAtg8. GFP-MoVast1 and RFP-MoAtg8 were co-expressed in ΔMovast1. Cultures were incubated for 7 d in CM medium, and then transferred to CM liquid medium for 24 h. The hyphae were observed by a fluorescence microscope. Bar: 50 μm. Fluorescence densities of GFP-MoVast1 and RFP-MoAtg8 were analyzed using ImageJ software

To assess the nature of the punctate structures, an MoLhs1-mCherry (an ER marker) fusion protein was co-expressed with GFP-MoVast1 in Guy11. The localization of GFP-MoVast1 coincided completely with MoLhs1-mCherry in the dormant conidia (Figure 4B, 0 h). At 2 hpi in sterile water on a glass slide, GFP-MoVast1 appeared as puncta near the PM (Figure 4B, 2 h). Considering that MoVast1 interacts with MoAtg8, we carried out a colocalization assay between GFP-MoVast1 and RFP-MoAtg8. We found that most puncta of GFP-MoVast1 were colocalized with RFP-MoAtg8 in hyphae (Figure 4C). We conclude that MoVast1 is localized in ER and PM sites and co-localizes with MoAtg8.

MoVast1 negatively regulates autophagy

Autophagy is essential for triggering the developmental cell death program and degradation of the conidial components. Moderate autophagy contributes to storage energy and penetration peg development [7]. To test whether MoVast1 regulates autophagy, a GFP-MoAtg8 fusion protein was expressed in Guy11 and ΔMovast1. The autophagic flux was analyzed by fluorescence microscopy and western blotting. In nutrient-rich medium (CM), GFP-MoAtg8 fluorescence was mainly observed in the cytoplasm, and observable puncta surrounded the vacuole. However, few puncta were observed in ΔMovast1, and GFP-MoAtg8 fluorescence was predominant in vacuoles. After the hyphae of the wild-type strain Guy11 and the ΔMovast1 mutant were starved for 4 h in synthetic defined medium without amino acids and ammonium sulfate (SD-N), most of the GFP-MoAtg8 fluorescence was in vacuoles and a small number of puncta were observed in Guy11. However, few observable puncta could be found in ΔMovast1 (Figure 5A). We quantified the number of GFP-MoAtg8-labeled autophagosomes in CM and SD-N. Autophagosome number decreased significantly in ΔMovast1 in both CM and SD-N (Figure 5B).

Figure 5.

MoVast1 regulates the autophagy process. (A) The localization of GFP-MoAtg8 in Guy11 and the ΔMovast1 mutant. Strains were cultured in liquid CM medium for 2 d then transferred to SD-N medium for 4 hpi. Mycelia were stained with CMAC and photographs were obtained by fluorescence microscopy. Arrows indicate autophagosomes. Bar: 10 μm. (B) The number of autophagosomes in a hypha was counted in Guy11 and ΔMovast1. Statistics were calculated in ImageJ software. Asterisks indicate statistically significant differences (*** p < 0.01). (C) Immunoblot analysis of GFP-MoAtg8 proteolysis in Guy11 and ΔMovast1. The degradation rates were calculated using the formula: GFP: (GFP+GFP-MoAtg8). (D) Conidia autophagy in Guy11 and ΔMovast1. (E) The number of autophagosomes in Guy11 and ΔMovast1 from 0 hpi to 24 hpi. Data and statistics were obtained from ImageJ software. Bar: 10 μm

Next, the autophagic flux was analyzed by western blot through the amount of free GFP and total GFP-MoAtg8, which corresponds to vacuolar delivery and subsequent breakdown of GFP-MoAtg8. In the wild-type strain Guy11, the GFP-MoAtg8 band was strong and the free GFP band was weak when the hyphae were cultured in liquid CM medium for 2 d, while the free GFP band was comparatively strong in ΔMovast1. After transferring the hyphae to SD-N medium for 4 h, the signal from the free GFP band became strong and the fusion protein GFP-MoAtg8 band was light in the wild-type strain, however, the GFP-MoAtg8 band nearly disappeared in ΔMovast1 (Figure 5C).

To visualize infection-associated autophagy in M. oryzae, conidia were harvested from wild-type Guy11 and the ΔMovast1 mutant, and diluted to 5 × 10⁴ conidia/ml with water, then induced on hydrophobic plastic coverslips for 0, 4, 12 and 24 h at 22ºC. In Guy11, the GFP-MoAtg8 puncta appeared when induced for 0 h and 4 h, and were undetectable at 24 h, and most of the GFP fluorescence was in vacuoles at 12 h and 24 h. However, few puncta were observed in ΔMovast1 from 0 h to 24 h (Figure 5D). We quantified the number of GFP-MoAtg8-labeled autophagosomes in conidia during appressorium development. Autophagosome number decreased significantly in ΔMovast1 (Figure 5E). These results suggested that autophagy progresses faster in ΔMovast1 than in the wild-type Guy11.

MoVast1 interacts with MoAtg8, and GFP-MoVast1 colocalizes with RFP-MoAtg8. During M. oryzae development, the pattern of MoVast1 localization is the same as that of MoAtg8. In order to explore the localization pattern, RFP-MoVast1 and GFP-MoAtg8 were co-expressed in Guy11, and the co-localization of RFP-MoVast1 and GFP-MoAtg8 were observed. The GFP-MoAtg8 puncta appeared at 2 hpi, 6 hpi, and were undetectable at 12 hpi. Most of the free GFP fluorescence was observed in conidial vacuoles, and the GFP-MoAtg8 puncta were formed in appressoria at 6 hpi. The RFP-MoVast1 puncta also appeared at 2 hpi but disappeared until 18 hpi in conidia. However, the new RFP-MoVast1 puncta were re-formed in appressoria after 18 hpi and MoVast1 were rarely colocalized with MoAtg8 in both conidium and appressorium (Fig. S4A). We infer that the interaction between MoVast1 and MoAtg8 is quick and instantaneous because conidium germination and appressorium formation are highly active processes and MoVast1 would be degraded after autophagy is initiated. So, we detected the content of MoVast1 at 6 hpi, 12 hpi and 24 hpi in SD-N by western blot. GFP-MoAtg8 was degraded by 6 hpi and the MoVast1 was also degraded by 6 hpi under starvation (Fig. S4B and S4C). These results indicate that MoVast1 could be degraded by the autophagy pathway.

MoTor activity is compromised in the ΔMovast1 mutant

Previous studies showed that the activity of Tor is a critical modulator of the autophagy process [18–20,37]. To test whether MoVast1 regulates the activity of MoTor, the relationship between MoVast1 and MoTor kinase was verified by treatment with rapamycin, a specific inhibitor of the Tor kinase. Compared with the wild-type Guy11, the ΔMovast1 mutant was more sensitive to rapamycin when cultured on CM medium for 5 d (Figure 6A and B). To verify the regulation of MoTor activity by MoVast1, we detected the phosphorylation status of Rps6, a marker of the activity of TORC1 [38]. Western blot analysis showed that the phosphorylation level of Rps6 was significantly decreased in the ΔMovast1 mutant compared with Guy11 when treated with rapamycin (Figure 6C and D). In addition, inhibition of MoTor activity increased autophagy flux by MoAtg8/MoAtg8-PE turnover and GFP-MoAtg8 degradation (Figure 6E and F). These data indicate that MoTor is a negative regulator of the autophagy pathway, and MoVast1 regulates autophagy likely by promotion of MoTor activity.

Figure 6.

MoTor activity was inhibited in ΔMovast1. (A) Growth of Guy11, ΔMovast1, and the complementation strain on CM agar medium containing 100 ng/ml rapamycin for 5 d at 25°C. (B) The relative growth rate of Guy11, ΔMovast1 and the complementation strain. Asterisks indicate statistically significant differences (*** p < 0.01). (C) Phosphorylation analysis of MoRps6 in Guy11 and ΔMovast1. Strains were cultured in liquid CM medium for 2 d and then half volume was transferred to a new bottle treated with 30 ng/ml rapamycin for 4 h. Proteins were extracted using the TCA-Acetone-SDS method and a western blot by 12.5% SDS-PAGE. (D) The phosphorylation level of MoRps6 was analyzed using ImageJ software. Asterisks indicate statistically significant differences (*** p < 0.01). (E) Immunoblot analysis of MoAtg8/MoAtg8-PE turnover in Guy11 and the ΔMovast1 mutant when treated with SD-N medium and 30 ng/ml rapamycin for 4 h at 25°C. (F) Immunoblot analysis of GFP-MoAtg8 proteolysis in Guy11 and the ΔMovast1 mutant when treated with 30 ng/ml rapamycin for 4 h at 25°C

The VASt domain and GRAM domain are essential for MoVast1 function(s)

Protein domains have specialized tasks that form structural, evolutionary, and functional units of proteins [39]. The homology analysis indicated that the VASt domain is highly conserved from fungi to humans. In M. oryzae, only two VASt domain-containing proteins were identified, and we named them MoVast1 and MoVast2. Pfam analysis showed that MoVast1 contains a GRAM domain and a VASt domain, and MoVast2 contains a BAR domain, a PH-like domain and a VASt domain. In humans, GRAMD1A, GRAMD1B and GRAMD1C all contain both a VASt domain and a GRAM domain. In budding yeast, Lam1/Ysp1, Ysp2, Sip3, Lam4, Lam5 and Lam6 were identified as VASt domain-containing proteins. Additionally, Ysp1 and Sip3 were divergent compared to Ysp2, Lam4 to Lam6, and GRAMD1A to GRAMD1C. MoVast1 is similar to Ysp2 (identity 37.84%) (Figure 7A).

Figure 7.

The predicted domains of MoVast1. (A) There were only two VASt domains found in M. oryzae, while three VASt domains were found in humans and six in yeast. Visualization of the VASt domain, all containing the GRAM domains or Bin/amphiphysin/RVS (BAR) domain. (B) Alignment of rice blast, yeast, and human VASt domains using Clustal X. Secondary structure is indicated above the alignment (sheets–green arrows, helices–gray). (C) The predicted structure of VASt domain of MoVast1 in M. oryzae using SWISS-MODEL (https://swissmodel.expasy.org/interactive). (D) The VASt domain harbors a large hydrophobic cavity. Surface of the VASt model colored according to residue hydrophobicity

Although sequence-based phylogeny could identify protein domains of VASt, structure-based phylogenetic network inference could improve the resolution of deep evolutionary relationships and assist in the inference of protein function. To predict the 3D structure of the VASt domain in M. oryzae, we submitted the VASt-related 149 amino acid sequence to the homology and threading structure prediction server SWISS-MODEL (https://swissmodel.expasy.org/interactive). The best model matched the membrane-anchored lipid-transfer protein Ysp2. The predicted structure of the VASt domain in M. oryzae contains three α helices (α1 to 3), six β-sheets (β1 to 6), and two loops (Ω1 and 2), and encompasses a large hydrophobic cavity (Figure 7B-7D). This model had a high score in the top 10 templates and the coverage was 97%. These results indicate that more than 90% confidence could be estimated in quality predictions.

To explore the domain functions in pathogenicity in more detail, we focused on the GRAM and VASt domains of MoVast1. The full-length CDS (ΔMovast1::MoVast1), full-length CDS without the GRAM domain (MoVast1ΔGRAM), full-length CDS without the VASt domain (MoVast1ΔVASt), GRAM domain (GFP-GRAM), and VASt domain (GFP-VASt) fused to GFP were separately expressed in the ΔMovast1 mutant. As expected, only ΔMovast1::MoVast1 completely restored the defects of ΔMovast1. MoVast1ΔGRAM, MoVast1ΔVASt, and GFP-GRAM have similar phenotypes as ΔMovast1. However, GFP-VASt showed smaller colonies compared to ΔMovast1, but the pathogenicity was more severe than ΔMovast1 and the stress tolerance defects were restored (Fig. S5A, S5B and S6). Next, we observed the localization of MoVast1, GFP-GRAM, GFP-VASt, MoVast1ΔGRAM and MoVast1ΔVASt in M. oryzae. MoVast1ΔGRAM and MoVast1ΔVASt strains had some puncta in hyphae, but significantly fewer compared with the MoVast1 strain. However, no puncta were found in GFP-GRAM and GFP-VASt strains (Fig. S5C). From these results, we infer that the VASt and GRAM domains function directly in conidiation and pathogenicity, and localization of MoVast1.

The T902 site affects the functions of MoVast1

Previous studies showed that the VASt domain (also called StART-like domain) could bind sterols and aid in the transfer of sterols from the ER to the PM via the T921 binding site in yeast Ysp2 protein [40]. To verify the functions of the active site in MoVast1 of M. oryzae, GFP-MoVast1 and GFP-MoVast1^T902D were separately expressed in the ΔMovast1 mutant. The GFP-MoVast1^T902D cassette could not restore the growth, pathogenicity and stress tolerance defects of ΔMovast1 (Figure 8A, 8B, S7A and S7B), indicating that the MoVast1 (T902) site is essential for the functions of MoVast1.

Figure 8.

The T902 site affects the functions of MoVast1. (A) The colony morphology of Guy11, ΔMovast1, MoVast1 (T902) site mutagenesis strain, and the complementation strain on CM medium at 25°C for 7 d. (B) Disease symptoms on cut leaves of barley inoculated with mycelial plugs of Guy11, ΔMovast1, MoVast1 (T902) site mutagenesis strain and the complementation strain cultured on CM at 25°C for 4 d. (C) The localization of GFP-MoVast1 in conidia. The conidial suspensions (5 × 10⁴ conidia/ml) of the ΔMovast1 mutant complemented with GFP-MoVast1 were observed by fluorescence microscopy at 0 and 2 hpi. Bar: 10 μm. (D) The localization of GFP-MoVast1^T902D in conidia. The conidial suspension (5 × 10⁴ conidia/ml) of the ΔMovast1 mutant complemented with MoVast1^T902D were observed by fluorescence microscopy at 0 and 2 hpi. Bar: 10 μm

In order to explore the functions of the T902 site, the subcellular localization of MoVast1^T902D was detected in M. oryzae. In the GFP-MoVast1 strain, the fluorescence was mainly localized in the ER in dormant conidia collected from CM media. The same fluorescence pattern was seen in the GFP-MoVast1^T902D strain. The punctate structures appeared near the PM when the conidia were incubated on glass slides for 2 h. However, few puncta appeared in the GFP-MoVast1^T902D strain (Figure 8C and D). These results indicate that the 902 site of MoVast1 is essential for the biological function and localization of MoVast1.

VASt domain and GRAM domain can respectively bind sterols and PtdIns3P

To further clarify the functions of the VASt and GRAM domains in M. oryzae, we performed in vitro lipid-binding assays. GFP-VASt and GFP (negative control) were purified using GFP beads and incubated with 10 μM ergosterols for 40 min, then stained with 50 μg/ml filipin for 2 h and observed by fluorescence microscopy with a UV filter. The fluorescence could be observed in GFP-VASt beads (Figure 9A), indicating that VASt can bind ergosterols. In addition, we tested the membrane-binding properties of VASt using 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) liposomes containing ergosterols. GFP-VASt, GFP-VASt^T902D and GFP were purified using GFP beads and incubated with DOPC and ergosterol (ERG)-containing liposomes. GFP-VASt and GFP-VASt^T902D could bind ergosterols; however, GFP-VASt^T902D showed a significantly weaker ability to bind to the DOPC-ERG liposomes (Figure 9B). Furthermore, we detected whether the GRAM domain could bind lipids using pip strips. As shown in Figure 9C, GRAM could interact with PtdIns3P. These data indicate that the GRAM domain and VASt domain may have separate functions in MoVast1.

Figure 9.

VASt domain and GRAM domain can bind sterols and PtdIns3P, respectively. (A) The VASt domain can bind sterols based on the sterols staining assay. GFP-VASt and GFP were purified using GFP beads and then observed by UV filter after they were stained with 50 µg/ml filipin for 2 h. Bar: 50 μm. (B) Liposome pull-down assay. The DOPC (100%) and DOPC-ERG (molar ratio 80:20) liposomes were mixed with the purified GFP-VASt, GFP-VASt^T902D and GFP protein, and incubated at 4°C for 40 min. The residual supernatant protein and the liposome-bound protein were separated by centrifugation at 17,000 × g for 30 min and then washed three times with TBS buffer. The proteins in the liposome pellet were analyzed using SDS-PAGE. (C) GRAM domain can bind PtdIns3P in protein–lipid overlay assay. PIP Strips were incubated with 0.5 µg/ml purified Flag-GRAM for 2 h, and bound proteins were detected by immunostaining with anti-Flag antibody. LPA, lysophosphatic acid; LPC, lysophosphocholine; PtdIns, phosphatidylinositol; PE, phosphatidylethanolamine; PC, phosphatidylcholine; S1P, sphingosine-1-phosphate; PA, phosphatidic acid; PS, phosphatidylserine

Sterol and sphingolipid synthesis is altered in the ΔMovast1 mutant

To test the function of MoVast1 in sterol transfer, we examined the growth of the ΔMovast1 mutant in the presence of amphotericin B (AmB), an antifungal agent that binds to sterols on PM of sensitive fungi and damages the permeability of the membrane. The ΔMovast1 mutant showed increased sensitivity to AmB compared to the wild-type Guy11 and the complementation strain (Figure 10A and B), suggesting that the content of sterols was higher in the ΔMovast1 mutant. Furthermore, to determine the intracellular distribution of sterols, we stained the hyphae with filipin, a sterol-sensitive fluorescent dye that interacts with sterol [41]. The filipin fluorescence in the PM and cytosol were quantified by fluorescent intensity values of the wild type and the ΔMovast1 mutant [42]. The filipin staining intensity was significantly higher in the ΔMovast1 mutant compared to the wild-type strain regardless of whether it was in the PM or cytosol (Figure 10C-10E), indicating higher sterol content in the ΔMovast1 mutant. In addition, we also checked the content of ergosterol using an HPLC assay, and the result also showed higher content of ergosterol in the ΔMovast1 mutant compared to the wild-type strain (Figure 10F). However, the content of ergosterol was significantly lower in the GFP-VASt mutant and was more sensitive to fenpropimorph (a sterol synthesis inhibitor) (Fig. S8A-S8C). These results suggest that MoVast1 plays a key role in ergosterol homeostasis in the PM. Lack of MoVast1 resulted in sterol accumulation and expression of VASt resulted in sterol reduction in M. oryzae.

Figure 10.

Sterols and sphingolipid synthesis were disturbed in the ΔMovast1 mutant. (A) The growth of Guy11, ΔMovast1, and the complementation strain in CM medium added with 2 μM amphotericin B for 5 d. (B) The relative growth rate of Guy11, ΔMovast1, and the complementation strain treated with 2 μM amphotericin B for 5 d in CM medium. Asterisks indicate statistically significant differences (*** p < 0.01). (C) Filipin fluorescence staining in Guy11, ΔMovast1, and the complementation strain. The conidial suspension (5 × 10⁴ conidia/ml) of each strain were incubated for 6 h at 25°C, then treated with 50 μg/ml filipin solution for 2 h and observed with fluorescence microscopy using a UV filter. Bar: 10 μm. (D) The relative plasma membrane sterols content detection by fluorescence intensity in Guy11, ΔMovast1, and the complementation strain. Asterisks indicate statistically significant differences (*** p < 0.01). (E) The relative cytoplasmic sterols content detection by fluorescence intensity in Guy11, ΔMovast1, and the complementation strain. Asterisks indicate statistically significant differences (*** p < 0.01). (F) Relative abundance of ergosterol detection in each strain by HPLC. Data presented are the mean ± s.d. (n = 3). Asterisks indicate statistically significant differences (*** p < 0.01). (G) The colony morphology of Guy11, ΔMovast1, and the complementation strain on CM agar medium containing 2 μM myriocin for 5 d at 25°C. (H) The relative growth rates of Guy11, ΔMovast1 mutant, and the complementation strain. Asterisks indicate statistically significant differences (*** p < 0.01). (I) (i) Appressoria formation in Guy11, ΔMovast1, and the complementation strain. Conidia with or without 2 μM myriocin were inoculated on hydrophobic plastic coverslips at 25°C for 12 h. Bar: 10 μm. (ii) The pathogenicity of Guy11, ΔMovast1, and the complementation strain on barley. The conidial suspension (5 × 10⁴ conidia/ml) with or without 2 μM myriocin of each strain were dropped on cut leaves of barley for 4 d. (J) Different types of appressoria formation in Guy11, ΔMovast1, and the complementation strain. Type1, normal appressorium; Type2, no appressorium; Type3 and Type4, abnormal appressorium. The data were analyzed using Prism 7.0 software

Sphingolipids and sterols are the main components of cell membranes [43]. To test whether ergosterol homeostasis contributes to the sphingolipid synthesis system, myriocin (a sphingolipids synthesis inhibitor) was used to check the sensitivity in wild-type Guy11, the ΔMovast1 mutant, and the complementation strain. The growth of the ΔMovast1 mutant was significantly inhibited compared to Guy11 and the complementation strain (Figure 10G and H), suggesting that sphingolipid synthesis was reduced in ΔMovast1. Next, we examined the virulence of ΔMovast1 on barley using a conidial suspension supplemented with or without 2 μM myriocin. Under the myriocin-absent condition, the wild-type Guy11 and the complementation strain formed appressoria and produced expanded lesions. However, the ΔMovast1 mutant could only form restricted lesions even though the mutant could form normal appressoria. When 2 μM myriocin was added to the conidial suspension, the expanded lesions were generated normally in the wild type and the complementation strains. However, almost no lesions were found on leaves inoculated with the ΔMovast1 mutant (Figure 10I). In addition, the appressorium morphology of ΔMovast1 was impaired when treated with myriocin.

In order to characterize the appressorium morphology, it was classified into the following four types: Type 1, normal appressoria; Type 2, no appressoria; Type 3 (long oval shape) and Type 4 (heart shape), abnormal appressoria. The four types were classified and analyzed statistically in the various strains. In Guy11 and the complementation strain, nearly 70% of appressoria were Type 1; this is in contrast to the less than 30% that were Type 1 in ΔMovast1. In addition, 50% of the appressoria in ΔMovast1 were Type 3 and Type 4 (Figure 10J). These results indicate that the changes of the ergosterol homeostasis at the PM are important for sphingolipid content. Thus, sphingolipid synthesis plays an important role in the formation of appressoria and pathogenicity of M. oryzae.

MoVast1 responds to hyper-osmotic stress and monitors the PM tension

Protein homeostasis can respond to changes in the external environment in M. oryzae [44]. To verify the function of MoVast1 in response to such stress, we tested the sensitivity of ΔMovast1 to hyper-osmotic shock. When treated with 0.5 M NaCl and 1 M sorbitol, the ΔMovast1 mutant showed more sensitivity than the wild-type Guy11 or the complementation strain (Figure 11A and B). Furthermore, the phosphorylation levels of Hog1, which responds to hyper-osmotic stress, were low when the strains were cultured in CM medium for 2 d. However, when treated with 0.5 M NaCl for 20 min, the phosphorylation level of Hog1 significantly increased both in Guy11 and ΔMovast1 and decreased until treatment for 60 min (Figure 11C). These results indicate that the high-level phosphorylation of Hog1 could respond to hyper-osmotic stress in both the wild type and ΔMovast1 mutant. Next, we observed the autophagy level when treated with hyper-osmotic stress. Autophagy was induced by hyper-osmotic stress in the ΔMovast1 mutant (Figure 11D). We inferred that autophagy can be induced by hyper-osmotic stress and that MoVast1 is important for such a response to hyper-osmotic stress but is independent of the Hog1 pathway in M. oryzae.

Figure 11.

MoVast1 changes PM tension and response to cell and hyper-osmotic stress. (A) Growth of M. oryzae on CM agar medium containing 0.5 M NaCl or 1 M sorbitol for 7 d at 25°C. (B) Statistical analysis of the mycelial growth rate under hyper-osmotic stress. Asterisks represent significant differences (*** p < 0.01). (C) Phosphorylation analysis of MoHog1 in Guy11 and ΔMovast1 mutant when treated with 0.5 M NaCl. (D) Immunoblot analysis of MoAtg8/MoAtg8-PE turnover in Guy11 and ΔMovast1 when treated with 0.5 M NaCl for 5 min, 20 min, 40 min and 60 min. (E) MoVast1 impact on PM tension. Conidial suspensions (5 × 10⁴ conidia/ml) of each strain were incubated for 6 hpi. The lifetime of the Flipper-TR probe was determined by fluorescence lifetime imaging microscopy (FLIM). Bar: 10 μm. (F) The lifetime in Guy11 and ΔMovast1 mutant. Error bars represent propagated error of mean values for three independent experiments and asterisks represent significant differences (*** p < 0.01). (G) Phosphorylation analysis of MoYpk1 and MoRps6 in Guy11 and the ΔMovast1 mutant. Strains were cultured in liquid CM medium for 2 d and treated with 0.5 M NaCl for 5 min, 20 min, 40 min and 60 min. Proteins were extracted using the TCA-Acetone-SDS method followed by a western blot with 12.5% SDS-PAGE. (H) The phosphorylation level of MoYpk1 was analyzed using ImageJ software. Data presented are the mean ± s.d. (n = 3). (I) The phosphorylation level of MoRps6 was analyzed using ImageJ software. Data presented are the mean ± s.d. (n = 3)

Recent studies showed that hyper-osmotic shock can inhibit the activity of TORC2 by regulating the PM tension in yeast [45]. Flipper-TR, a new fluorescent membrane tension probe, was used to verify the response of MoVast1 to hyper-osmotic shock by regulation of the PM tensions in M. oryzae. Flipper-TR contains twisted push–pull fluorophore that stably integrates into the PM, and thus its fluorescence characteristics are sensitive to mechanical forces acting on the membrane [45]. Definitely, the fluorescence lifetime changes linearly with the PM tension in yeast and mammalian membranes, and can be determined using fluorescence lifetime imaging microscopy (FLIM) [45]. It is a valuable tool to monitor the PM tension changes in situ. We tested the membrane tension in ΔMovast1 and the wild-type Guy11 with the Flipper-TR probe using FLIM. We found that the PM tension was significantly increased in the ΔMovast1 mutant compared to the wild type, while in the GFP-VASt mutant, the PM tension was significantly decreased (Figure 11E and F). This indicates that loss of MoVast1 could increase the PM tension.

Next, we checked the relationship between the PM tension and the activity of MoTor. In yeast, the activity of TORC2 was detected by the phosphorylation of Ypk1 in the T662 site. In our research, we used the orthologous MGG_06599 (named MoYpk1) to test the activity of MoTor in M. oryzae. At first, to detect whether the phosphorylation of MoYpk1 (S619) (homologous with Ypk1 [T662]) can represent the activity of TORC2, we used the TORC2 inhibitor palmitoylcarnitine (PalmC), which reduces the activity of TORC2 by decreasing the PM tension [45]. Western blot analysis showed that the phosphorylation level of MoYpk1 was significantly decreased when treated with PalmC (Fig. S9A), indicating that MoYpk1 (S619) can monitor TORC2 activity. We also detected the formation of appressoria when treated with PalmC and found that appressorium development was impaired when the activity of TORC2 decreased (Fig. S9B and S9C). Next, we detected the phosphorylation of MoYpk1 (S619) when treated with hyper-osmotic shock using 0.5 M NaCl. As expected, the phosphorylation of MoYpk1 (S619) decreased rapidly after 5 min and was restored in 60 min in the wild-type Guy11. Although the phosphorylation of MoYpk1 (S619) also decreased at 5 min, it was not restored within 60 min in the ΔMovast1 mutant (Figure 11G and H). We also used MoRps6 to monitor TORC1 activity, but there were no significant differences between wild type and the ΔMovast1 mutant when treated with 0.5 M NaCl (Figure 11G and I). These results indicate that MoVast1 responds to hyper-osmotic shock by regulating the activity of TORC2.

Lipid metabolism and the autophagy pathway were disturbed in ΔMovast1.

In order to further explore the functions of MoVast1 in lipid metabolism and the autophagy pathway, the lipidomics and transcriptomics profiles were analyzed between ΔMovast1 and the wild-type Guy11 and Guy11 treated with rapamycin. Gene Ontology (GO) enrichment of differentially expressed genes (DEGs) was performed to explore multiple aspects including the carbohydrate metabolic process, melanin metabolic process, phosphate ion transport, and enzyme activities (Figure 12A). In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed that steroid biosynthesis and glycolipid metabolism were the most highly enriched pathways, which indicated that lipid metabolism was disturbed in the ΔMovast1 mutant (Figure 12B). Previous data showed that loss of MoVast1 decreased the activity of Tor. Next, we detected whether Tor signaling mediated lipid metabolism using RNA-Seq analysis. The results showed that the lipid metabolism-related genes were also enriched (Fig. S10A).

Figure 12.

Lipidomics analysis verifies that MoVast1 is involved in lipid metabolism and the autophagy pathway. (A) Gene Ontology (GO) analysis of differentially expressed genes (DEGs) in the wild-type Guy11 and the ΔMovast1 mutant. DEGs were classified under three categories as indicated, BP: biological process, CC: cellular component, MF, molecular function. The top 10 GO terms are shown for each category. (B) KEGG analysis comparing the wild-type Guy11 and the ΔMovast1 mutant samples, revealing significant enrichment of many pathways. (C) Heatmap showing the lipidomic analysis of the ΔMovast1 mutant vs the wild-type Guy11. Each rectangle represents a lipid colored by its normalized intensity scale from black (increased level) to white (decreased level). The hierarchical clustering analysis was based on identified lipid metabolites with significant changes in quantity. LPC, lysophosphatidyl choline, PC phosphatidylcholines, PE phosphatidylethanolamine, SM sphingomyelin, Cer ceramides, TG Triglyceride, DG, diacylglycerol. (D) KEGG analysis showing significantly different processes involved in lipid metabolism

To further explore how MoVast1 perturbed lipid metabolism, a widely targeted lipidomic analysis was performed between ΔMovast1 and the wild-type Guy11. Orthogonal projection to latent structures discriminant analysis (OPLS-DA) between Guy11 and ΔMovast1 samples was carried out for both the GC-MS and LC-MS data (Fig. S10B, Data S2), and a combination of variable importance in projection (VIP) value analysis and Student’s t-tests revealed 116 differentially accumulated metabolites in the GC-MS data set (Data S2). To better demonstrate lipid profile changes between ΔMovast1 and Guy11, a heatmap was constructed according to the identified lipid list (Data S3). As shown in Fig. 12 C, lipid features were changed, the content of phosphatidylethanolamine (PE), lysophosphatidyl choline (LPC) and phosphatidylcholines (PC) were increased, however, the contents of triglyceride (TG) were decreased significantly. These results indicated that MoVast1 strongly perturbed lipid homeostasis and biological experiments further verified that lipid degradation was delayed in the ∆Movast1 mutant (Figure 12C and Fig. S11). To map the landscape of metabolic-transcriptional alterations in the context of MoVast1, we performed an integrated transcriptomic and lipidomic analysis to simultaneously map genes and lipid metabolites in different pathways. Integrative network modeling specifically revealed that the autophagy pathway, glycerolipid metabolism, and the sphingolipid and sterol synthesis pathway were most dramatically modulated by MoVast1 (Figure 12D). Collectively, we demonstrated that the loss of MoVast1 triggered significant lipid metabolic changes in M. oryzae.

Additionally, transcriptome data analysis showed that MoVast1 also regulates many metabolic pathways by regulating the expression of other genes involved in ergosterol biosynthesis, melanin biosynthesis, amino acid metabolism, carbohydrate metabolism, energy homeostasis, and lipid catabolism. Furthermore, we verified the expression of 28 DEGs by qPCR. In ΔMovast1, ergosterol biosynthesis genes (MGG_04432, MGG_03765, MGG_04346 (ERG6), MGG_06139), glycerolipid metabolism genes (MGG_00220, MGG_03900, MGG_09805, MGG_07890, MGG_10005), carbohydrate metabolism genes (MGG_02530, MGG_04099, MGG_06777), transporter genes (MGG_15183, MGG_05833, MGG_10410), and cytochrome P450 genes (MGG_05215, MGG_01391, MGG_09198, MGG_01924) were significantly downregulated compared with Guy11. However, melanin metabolic genes (MGG_07216, MGG_5059, MGG_02252, MGG_7219, MGG_07215), heat shock response genes (MGG_04437 (HSP78), MGG_05719 (HSP30), MGG_03329), and PKC genes (MGG_01393, MGG_08709, MGG_14499, MGG_16375, MGG_02944) were significantly upregulated compared with those of Guy11 (Fig. S12). These results indicate that MoVast1 could regulate several metabolic pathways by down- or upregulating genes in M. oryzae.

DISCUSSION

Accumulating evidence indicates that autophagy is crucial for virulence of many plant pathogenic fungi, such as M. oryzae, Colletotrichum lindemuthianum, F. graminearum, and U. maydis [44], however, the regulatory mechanisms of autophagy in pathogenic fungi are still unclear. In this study, we report the roles of MoVast1 in the regulation of pathogenicity and autophagy in M. oryzae. We show that MoVast1 contributes significantly to fungal development and pathogenicity. Through controlling the homeostasis of sterols and sphingolipid, MoVast1 modulates the PM tension and thus the activity of MoTor kinase, and thus adjusts the intensity of the rice blast fungal autophagy.

The PM forms a selectively permeable barrier and dynamic interface. Meanwhile, many fate-determining signaling decisions occur on the PM [46]. PM tension is defined as the in-plane force counteracting surface expansion and plays a vital role in transferring and integrating information in cells and tissues [45,47]. The mechanisms by which the PM tension is sensed and regulated is still very limited. Current research shows that mechanical stretching of the PM, inhibition of sphingolipid biosynthesis and hypo-osmotic shock increase the PM tension [45], and TORC2 is a regulator of PM tension homeostasis in yeast [45,48]. In this study, we found that the sterol-associated protein MoVast1 binds to sterols and PtdIns3P by its VASt and GRAM domains, maintains PM homeostasis, and modulates PM tension. When MoVast1 is deleted, there is an excessive accumulation of sterols in the PM, resulting in higher PM tension, which inhibits MoTor activity and sphingolipid synthesis ((Figure 13).

Figure 13.

Model of MoVast1 regulating lipid homeostasis and autophagy. MoVast1, a sterol binding protein, can maintain the homeostasis of the plasma membrane and modulate PM tension. Furthermore, the activity of Tor can be regulated by changes in PM tension. In ΔMovast1 and MoVast1^T902D mutant, the sterol content accumulated in the PM and the excessive accumulation of sterol caused high PM tension. In order to restore the homeostasis of the plasma membrane, the activity of MoTor was downregulated. Afterward, sphingolipid synthesis was inhibited. In addition, MoTor activity was inhibited in the ΔMovast1 mutant and the autophagy process was able to resume

Sterols are the main nonpolar lipids of eukaryotic cell membranes and function directly in biological activities, such as endocytosis, protein sorting, establishment of membrane curvature, stabilization of membrane proteins, and regulation of receptor proteins [43,49,50]. Biosynthesis and trafficking of sterols are the most conserved biosynthetic pathways in eukaryotic cells. The pathways of sterol biosynthesis are similar in different organisms and differ only in the final stages [51]. In yeast, ergosterol biosynthesis requires 25 Erg enzymes, which catalyze the transition from the precursor squalene to the main product ergosterol. All of the downstream intermediates are insoluble in water, so the corresponding reaction steps take place at the membrane of ER [24]. However, a sterol gradient was maintained within the cells and up to 90% of cellular sterols were transferred to the PM [27]. The two main families of non-vesicular transport proteins, Osh and Lam, contribute to sterol trafficking in yeast. Osh family proteins (Osh1-Osh7) transfer sterols/PI4P from the ER to the PM, while the Lam proteins (containing sterol-binding StART [steroidogenic acute regulatory transfer]-like domain/VASt domain) transfer sterols from the PM to the ER [22,30]. In M. oryzae, two Lam family proteins (MoVast1 and MoVast2) were found. MoVast1 contains the GRAM domain and the VASt domain, both of which are essential for conidiation and pathogenicity (Fig. S5). Deletion of MoVast1 led to decreased stress resistance in M. oryzae under conditions such as hyper-osmotic pressure, rapamycin treatment, myriocin treatment and low temperature (Fig. 6A, 10A, 11A and S6). However, overexpression of the VASt domain can partially restore resilience to these stresses, and the conidiation and pathogenicity defects (Fig. S5 and S6). MoVast2, the other Lam family protein in M. oryzae, does not affect the conidiation and pathogenicity process of M. oryzae, but is involved in the ability of cells to resist stress and interact with MoVast1 and MoAtg8 (data not shown).

Apart from sterols, eukaryotic cells contain high levels of sphingolipids in the PM [27]. According to the “umbrella” model, a relatively large hydrophilic group of the sphingolipid molecule is necessary to compensate for excess sterols in a particular region of the membrane [52]. Therefore, sterols affect lipid membranes directly by modulating their physicochemical properties and indirectly by regulating activity and sorting of membrane-associated proteins [43]. Here, we found that MoVast1 can regulate the activity of Tor by modulating levels of sterol and sphingolipids through its VASt domain in M. oryzae.

As in mammalian cells, a crucial threshold of autophagy is critical for progression of programmed cell death in M. oryzae [4]. Launch of autophagy (resulting from MoTor inhibition, rapamycin treatment, or deletion of MoVast1) activates the pro-death signals that contribute to conidial collapse, and likely promotes trafficking of cellular contents to incipient appressoria. Upon penetration into host cells, a metabolic switch to glucose metabolism within IH enhances ATP production that promotes Tps1 activities, which in turn activates Tor signaling pathway, leading to mitosis that facilitates early biotrophic growth [4,53]. This suggests that the precise regulation of autophagy activities in conidia is essential for plant infection by M. oryzae. Overstimulation of autophagy by constant suppression of MoTor activity would hinder host infection by M. oryzae [2]. Therefore, the mechanisms that regulate Tor activity might be a switch for initiation and termination of autophagy. Our results uncovered new aspects of MoVast1 that regulate sterol homeostasis in PM and contribute to such a regulatory function for autophagy in response to stress conditions in M. oryzae.

The Tor signaling pathway is an exceedingly complex pathway that is regulated by many factors, such as nutrient signaling, oxidative metabolism, energy and environmental stress [54,55]. Previous studies have shown that the sterol and PtdIns-4-P-binding protein Kes1 mediates sphingolipid signaling and regulates activation of Tor by control of nitrogen signaling [56]. Recently, researchers found that the yeast Lam6/Ltc1 and Lam4/Ltc3 and Ysp2/Ltc4 functioned separately at the vacuole and PM. Membrane domains were created to partition upstream regulators of the TORC1 and TORC2 signaling pathways and adjust cellular stress responses with sterol homeostasis [57]. In addition, current research shows that mechanical stretching of the PM and inhibition of sphingolipid biosynthesis increase the PM tension, and that TORC2 is a regulator of the PM tension homeostasis in yeast [45]. These results suggested that PM homeostasis also participates in Tor signaling. In this study, we identified a new sterol-associated protein MoVast1 in M. oryzae that maintains the homeostasis of sterols and sphingolipid and modulates the PM tension by binding to sterols and PtdIns3P via its VASt and GRAM domains. Loss of MoVast1 resulted in excessive accumulation of sterols in the PM and increased PM tension. Then, the MoTor activity and sphingolipid synthesis were inhibited, and the autophagy activity was increased.

In conclusion, this study uncovers a critical link between homeostasis of sterols, MoTor signaling and autophagy induction in M. oryzae and reveals a novel molecular switch for regulation of autophagy by the sterol-transfer protein MoVast1. MoVast1 maintains sterol concentration in the PM to adjust its tension in order to positively regulate MoTor activities and thus autophagy processes. Future research should determine the precise molecular functions of VASt-domain proteins and deduced crystal structure(s), which will help for molecular docking of small molecule inhibitors of MoVast1, with important implications for developing new therapeutic strategies against the rice blast fungus.

Materials and Methods

Mass spectrometry analysis

To isolate autophagy-related proteins, GFP-MoAtg8 was transferred into the ΔMoatg8 mutant using the Agrobacterium tumefaciens-mediated transformation (ATMT) method [58]. Hyphae were cultured for 2 d in liquid CM medium (10 g D-glucose [Sango Biotech, A100188], 2 g peptone [Sango Biotech, A100849], 1 g yeast extract [Sango Biotech, A100850], 1 g casamino acid [Sango Biotech, A100851], 1.5 g KH₂PO₄ [Sango Biotech, A600445], 6 g NaNO₃, 0.5 g KCl [Sango Biotech, A610440], 0.5 g MgSO₄ · 7H₂O [Sango Biotech, A610329], 1 ml vitamin solution [0.01 g biotin {Sango Biotech, A100340}, 0.01 g pyridoxine{Sigma-Aldrich, P9755}, 0.01 g thiamine {Sigma-Aldrich, T1270}, 0.01 g riboflavin {Sango Biotech, A600470}, 0.01 g p-aminobenzoic acid {Sango Biotech, A50030}, 0.01 g nicotinic acid {Sango Biotech, A610660-0250} in 100 ml of distilled water], 1 ml trace elements [0.15 g Na₂MoO₄ · 5H₂O {Sigma-Aldrich, 243,655}, 0.16 g CuSO₄ · 5H₂O {Sango Biotech, A600063}, 0.17 g CoCl₂ · 6H₂O{Sango Biotech, A600316}, 0.5 g MnCl₂•4H₂O {Sango Biotech, A500331}, 0.5 g FeSO₄ · 7H₂O {Sango Biotech, A600461}, 1.1 g H₃BO₃ {Sigma-Aldrich, B6768}, 2.2 g ZnSO₄ · 7H₂O{Sango Biotech, A602906}, Na₄EDTA {Sango Biotech, A100105-0500} in 100 ml of distilled water] in 1 L of distilled water, PH 6.5), then transferred to SD-N medium (1.7 g Yeast Nitrogen Base without Amino Acids [Sango Biotech, A6100507], 20 g D-glucose [Sango Biotech, A100188], 2 g Asparagine [Sango Biotech, A694341], 1 g NH₄NO₃) for 4 hpi with 2 mM phenylmethylsulfonyl fluoride (Sango Biotech, A100754). The mycelial mat was ground in liquid nitrogen and then transferred to extract buffer (50 mM Tris, 150 mM NaCl, 1 mM EDTA, pH 7.4). The powder was centrifuged for 20 min at 7000 × g and 4ºC and the supernatant was transferred to a new centrifuge tube then incubated with anti-GFP beads (ChromoTek, gt-250) for 6 h at 4ºC and washed three times with TBST (Sango Biotech, C520009). Deposits were collected and 200 μl 0.3 M glycine buffer (pH 3.0) was added to the tube, which was then incubated for 5 min on ice. The supernatant (centrifuged for 2 min at 2000 × g) was transferred to a new centrifuge tube and sonicated for 5 min on ice. The supernatant was removed after centrifugation at 100,000 × g for 30 min.

To identify the autophagy-related proteins, LC-MS/MS analysis was performed on a Q Exactive mass spectrometer (Thermo Fisher Scientific) as described by Zhang et al [59]. The MS data were analyzed using MaxQuant software.

Gene knockout and complement system

In these assays, genes were deleted using the high-throughput gene knockout system designed by Lu et al [58] with slight modifications. Briefly, the 1.2- to 1.4-kb upstream fragment (UF) and 1.2- to 1.4-kb downstream fragment (DF) flanking sequences of each target gene were cloned from M. oryzae genomic DNA with gene-specific primers (Table S2). The target gene was replaced with a hygromycin-resistance gene (HPH). These three fragments (UF, DF and HPH) were fused with the linearized (HindIII/XbaI-linearized) vector pKO1B (gifted by Prof. Lu) [58] with a fusion enzyme (Vazyme Biotech Co., Ltd, C113-02). The strategy of targeted gene-deletion used ATMT [58]. For complementation, the situ complementation method was used. The UF, target gene, glufosinate ammonium resistance selection marker gene BAR and DF fragments were fused with the linearized (HindIII/XbaI -linearized) vector pKO1B. The resulting constructs were transformed into the gene-deletion mutants using the ATMT method. To identify gene-deletion mutants, the alternative transformants were screened using PCR as described by Zhu et al [36].

Strains and phenotypic analyses

All strains in this study were cultured on complete medium (CM) [2]. For the growth assay, strains were cultured in 9 cm CM medium for 9 d at 25°C. The conidiophore formation, conidial germination and appressorium formation assays were tested as reported by Liu et al [60]. For the virulence assay, the conidial suspension (5 × 10⁴ conidia/ml) in 0.2% gelatin (Sango Biotech, A600908) was sprayed on 21-d-old rice seedlings (CO-39; International Rice Research Institute, IRRI). Disease symptoms were observed and photographed after 7 d. To examine the pathogenicity of the strains, the detached barley leaves were inoculated with mycelial plugs of the Guy11 strain (International Rice Research Institute, IRRI) and mutant strains for 4 d.

Growth inhibition assay

To detect the response of mutants to different stresses and quantify the inhibitory effects of chemicals, rapamycin (Sigma-Aldrich, 53,123–88-9), myriocin (APExBIO, B6064) and amphotericin B (APExBIO, B1885) were separately added in solid CM media. The relative growth rates were calculated using the following formula: growth rate = (the diameter of strain treated with chemicals):(the diameter of untreated strain). All assays were repeated three times.

Autophagy assays

In order to detect the autophagy process, the GFP-MoAtg8 fusion with native promoter vector was transformed into Guy11 and the mutants using the ATMT method as previously described [60]. The GFP-MoAtg8 transformed strains were grown in liquid CM medium for about 2 d at 25°C, and then shifted to fresh liquid CM medium containing 30 ng/ml rapamycin or shifted to nitrogen starvation (SD-N) medium for 4 and 8 h to induce autophagy. The hyphae induced for 4 h were analyzed using a microscope (eclipse 80i 100 × oil). Immunoblots of GFP-Atg8 fusion proteins were probed with anti-GFP (GFP 1:10,000; Abcam; ab32146). The amounts of free GFP and GFP-MoAtg8 were calculated using densitometric analysis (ImageJ software) after western blotting. For endogenous MoAtg8/MoAtg8-PE assays, Guy11 and the ΔMovast1 mutant were cultivated in CM liquid for 2 d, then transferred to fresh liquid CM medium contains 30 ng/ml rapamycin, a solution containing 0.5 NaCl, or SD-N medium for 4 and 8 h. Proteins were isolated using the TCA method and analyzed by western blotting (13.5% SDS-PAGE in the presence of 6 M urea [Sango biotech, A510907] with anti-Atg8 [1:2000, BML Beijing Biotech; PM090]). Degradation trends were analyzed using ImageJ software (NIH).

Determination of membrane tension

To determine the PM tension, we used a novel probe Flipper-TR (for “fluorescent lipid tension reporter”) (Spirochrome, SC020). The method followed a previously described protocol [45]. The final concentration of Flipper-TR was 2 μM. Conidia were gathered and diluted to 5 × 10⁴ conidia/ml with double distilled water, then placed on a glass slide and incubated for 6 hpi at 25°C. At 6 hpi, the water was replaced with the staining solution so that all cells were covered with solution and incubated at 25°C for 15 min before imaging, as described by Colom Adai and Margot Riggi [47,59]. For lifetime measurements, FLIM imaging was performed using a Nikon Eclipse TiE-A1 plus microscope equipped with a time-correlated single-photon counting module from PicoQuant. The data was analyzed using ImageJ software.

Filipin fluorescence staining assay

To test the sterol content in the PM of the ΔMovast1 mutant and the wild-type Guy11, conidia were gathered and diluted to 5 × 10⁴ conidia/ml with double distilled water. Then, conidia were placed onto a glass slide and cultured for 6 hpi at 25°C. At 6 hpi, we replaced water with 50 μg/ml filipin (APExBIO, B6034). All cells were covered with filipin solution and incubated at 25°C for another 2 hpi. Before imaging, the cells were washed three times with PBS (Sango Biotech, E607008) then viewed under a UV filter set (Nikon 63 × oil). The image was analyzed using ImageJ software.

Protein-lipid overlay assays

For the sterol binding assay, we used a liposome pull-down assay as described previously [61]. For prepared liposomes, ergosterol (ERG) and DOPC were obtained from Sigma-Aldrich (45,480, P6354). The DOPC-ERG were dissolved in chloroform and mixed to the desired molar ratio and dried in a vacuum with a rotary evaporator. The dried lipid thin-films were swollen by adding 1 ml 0.3 M sucrose (Sango Biotech, A502792) and sonicated for 30 min. Subsequently, the lipid solution was extruded ten times through a poly-carbonate filter with a pore size of 0.1 μm and then 0.2 ml DOPC solution and 1 mL TBS (20 mM Tris·HCl, pH 8.0, 0.1 M NaCl) were added and the mixture was centrifuged at 17,000 × g for 30 min. The supernatant was discarded, and the pellet was resuspended with 700 μl TBS. For the liposome pull-down assay, the DOPC and DOPC/ERG liposomes of lipids were mixed with GFP-VASt, GFP-VASt^T902D and GFP and incubated at 4°C for 40 min. The protein in the supernatant and the liposome-bound protein were separated by centrifugation at 17,000 × g for 30 min. The pellet was washed with 1 ml of the TBS buffer, and the proteins in the supernatant and in the liposome pellet were analyzed using SDS-PAGE.

For the PtdIns3P binding assay, we used PIP strips (Echelon Biosciences, P-6001). Briefly, the membrane was blocked using TBST with 3% fatty acid–free BSA (Sango Biotech, A500023), and gently agitated for 1 h at room temperature. Then, it was incubated using 0.5 µg/mL of the Flag-GRAM protein in TBS-T with 3% fatty acid-free BSA for 2 h at room temperature. After incubation, the membranes were washed three times with TBST and then detected using Flag antibody (HUABIO, M1403-2).

Transcription analyses by RNA-seq

Total RNA was extracted from wild-type strain Guy11, the ΔMovast1 mutant, and the ΔMovast1 mutant complemented with VASt domain grown on CM agar medium for 7 d, then transferred to liquid CM medium for 24 h. Every strain had three biological replicates. Total RNA was extracted using RNA 6000 nano kit (Agilent, 5067–1511) and detected by Illumina HiSeq 2500 platform (Illumina Inc., USA). The differential expression genes analysis was conducted using the DESeq R package (https://www.rdocumentation.org/packages/DESeq2/versions/1.12.3) and the differential genes were verified by qRT-PCR.

Widely targeted metabolomics profiling

For widely-targeted metabolomics analysis, the samples were extracted and analyzed using an LC-ESI-MS/MS system (UPLC, Shim-pack UFLC SHIMADZU CBM A system). The analytical conditions were described as Song et al [62]. The data analysis implemented in SIMCA-P version 14.1 (Umetrics AB, Sweden) using Principal-component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA). Differentially metabolites were analyzed using the KEGG database (http://www.kegg.jp/).

Phosphorylation level analysis

To detect the phosphorylation level, hyphae were inoculated in CM liquid medium for 2 d and proteins were extracted by TCA-SDS methods [35]. To detect phosphorylated MoRps6 level, we used an anti-phospho-Rps6 (S235/S236) antibody (Cell Signaling Technology, 2211) and the anti-Rps6 antibody (Abcam, ab40820) was used as a control. To detect phosphorylated MoYpk1 level, the anti-phospho-MoYpk1 (S619) antibody and anti-MoYpk1 (Prepared by ABclonal Biotechnology Co., Ltd) were used. To detect phosphorylated MoHog1 level, we used an anti-MAPK14 (T180/Y182) antibody (Cell Signaling Technology, 9211S) and the anti-MAPK14 antibody (Cell Signaling Technology, 9212S) was used as a control. The overall control was anti-ACTB/beta-actin (ABclonal, AC004).

Statistical methods

In fluorescence and immunoblot analysis, the intensity of fluorescence was determined by ImageJ software. For analyzing statistical significance, each result is presented as the mean ± SD of at least three replicated measurements. The probability value (p-value) was statistic by the two-sample Student’s t-test using GraphPad Prism version 7.0.

Accession number

All the genes from this article can be found in the GenBank database libraries under thefollowing accession numbers: MoVAST1 (MGG_11211), MoVAST2 (MGG_08887), MoYPK1 (MGG_06599), MoTOR (MGG_15156), MoRPS6 (MGG_03236), MoATG1 (MGG_06393), MoATG8 (MGG_01062).

Funding Statement

This work was supported by the National Natural Science Foundation of China [31972216]; National Natural Science Foundation of China [31970140]; Ministry of Agriculture of China [2016ZX08009003-001-005]; Merit Review Grant [I01BX002924]; National Science and Technology Major Project [2018ZX08001-03B]; NIH [AI125770].

Disclosure statement

Dr. Maurizio Del Poeta, M.D. is a Co-Founder and Chief Scientific Officer (CSO) of MicroRid Technologies Inc. All other authors declare no conflict of interest.

Supplemental material

Supplemental data for this article can be accessed here.

Supplementary material

Supplemental data for this article can be accessed here.