MoGT2 Is Essential for Morphogenesis and Pathogenicity of Magnaporthe oryzae

The ascomycete fungus Magnapothe oryzae is the causal agent of rice blast disease, leading to severe loss in cultivated rice production worldwide. In this study, we identified a conserved type 2 glycosyltransferase named MoGt2 in M. oryzae. The mogt2Δ targeted gene deletion mutants exhibited pleiotropic defects in vegetative growth, conidiation, stress response, hyphal appressorium-mediated penetration, and pathogenicity. Furthermore, conserved glycosyltransferase domains are critical for MoGt2 function. The comparative transcriptome analysis revealed potential target genes under MoGt2 regulation in M. oryzae conidiation. Identification of potential glycoproteins modified by MoGt2 provided information on its regulatory mechanism of gene expression and biological functions. Overall, our study represents the first report of type 2 glycosyltransferase function in M. oryzae infection-related morphogenesis and pathogenesis.

firmed by real-time PCR (RT-PCR) (Fig. S1C) and selected for phenotypic analysis. To confirm the phenotypic defects of mogt2Δ mutants resulted from the deletion of MoGT2, we complemented the mogt2Δ-39 strain with a MoGT2-GFP fusion gene (Cterminal green fluorescent protein [GFP] tagging vector), and one complemented (MoGT2-com) strain was confirmed by RT-PCR (Fig. S1C).
Although in the MoGT2-com strain, a green fluorescent protein (GFP) was tagged at the C terminus of MoGt2 protein, we were unable to observe the subcellular localization of the MoGt2-GFP fusion protein (data not shown). We performed immunoblotting with this MoGT2-GFP complemented mogt2Δ strain, using anti-GFP antibody. An 83-kDa band of the expected size of MoGt2-GFP fusion protein was detected (Fig. S1D). This confirmed that a MoGt2-GFP fusion protein was successfully expressed in the complemented strains. Meanwhile, abundant GFP peptide (of 27 kDa) was also detected (Fig. S1D), indicating that a cleavage occurred between MoGt2 and GFP. As such, the genetic complementation strain could not be used to visualize subcellular localization of MoGt2, but could still be used for assessing MoGt2 function in M. oryzae growth, asexual development, and pathogenicity as described in the following sections.
MoGT2 is necessary for vegetative growth. To investigate the role of MoGT2 in mycelial growth, we tested the growth rate of mycelia from each strain and found that mogt2Δ mutants grew slower than the wild-type or complemented strains, when cultured on CM (complete medium), MM (minimal medium) or PDA (potato dextrose agar) ( Fig. 2A and B). When grown in liquid CM for 2 days, the mogt2Δ mutants formed , Aspergillus nidulans hypothetical protein (XP_682338), and C. neoformans Cps1 (AAQ92917). The evolutionary history was inferred using the neighbor-joining method (46). The optimal tree with the sum of branch length ϭ 1.76055829 is shown. The percentages of replicate trees in which the associated taxa clustered together in the bootstrap test (1,000 replicates) are shown next to the branches (47). The evolutionary distances were computed using the p-distance method (48) and are in units of the number of amino acid differences per site (labeled at the nodes). The rate variation among sites was modeled with a gamma distribution (shape parameter ϭ 1.2). All positions with less than 50% site coverage were eliminated. The position of MoGt2 in the phylogenetic tree is indicated by gray highlighting. Asterisks denote the fungal Gt2 or Cps1 proteins characterized in pathogenic fungi (11)(12)(13). Domain annotation was performed using the SMART website (http://smart.embl-heidelberg.de/). The amino acid residue number of the annotated domains is indicated. TM, transmembrane region; catalytic, glycosyl transferase domain. small compact mycelial masses, in contrast to the bigger sparse mycelium formed by the wild-type (WT) or complemented (MoGT2-com) strain ( Fig. 2A). By calcofluor white (CFW) staining, we found that the mogt2Δ vegetative hyphae contained more septa and the distance between two septa appeared shorter than those in the WT or MoGT2-com strain (Fig. 2C). These results indicated that MoGT2 is required for proper vegetative growth in M. oryzae.
MoGT2 is essential for asexual sporulation. Asexual spores play an essential role in the disease cycle of M. oryzae (14). To assess the role of MoGT2 in asexual sporulation, we observed conidiophore differentiation and conidium production. No conidia or typical conidiophores were observed in the mogt2Δ mutants, while the wild-type strain formed normal conidia and conidiophores (Fig. 3A). The ability to form asexual spores was further evaluated by carefully washing the surface of different strains cultured for 10 days (16-h light/8-h dark cycle). No conidia were harvested from the mogt2Δ mutants, whereas the wild-type strain produced (29.6 Ϯ 1.96) ϫ 10 6 spores per plate, and the complementation strain produced (24.73 Ϯ 1.57) ϫ 10 6 spores per plate (P Ͼ 0.05, WT versus MoGT2-com strain). Furthermore, we tried different media for inducing conidiation, including PA (prune agar) and CM, as well as starvation conditions. Neither of them could stimulate conidiation in the mogt2Δ mutant. We also tried to scrape off aerial hyphae and incubate further under humid conditions, which also failed to induce conidiation in the mogt2Δ mutant.
Next we perform quantitative real-time PCR (qRT-PCR) analysis to check the expression levels of conidiation-related genes CON7 and HTF1 and found significantly reduction of these two genes in the mogt2Δ mutant (Fig. 3B), indicating that MoGT2 may regulate M. oryzae asexual sporulation through (maybe indirectly) regulation of these conidiation-related genes' expression. Overall, we conclude that MoGt2 is essential for M. oryzae conidiation.
MoGT2 is essential for pathogenicity and appressorium-like structure formation from mycelia. To determine the role of MoGT2 in plant infection, we performed infection assays with leaf explants. Since the mogt2Δ mutants were unable to produce conidia, we used mycelial plugs of these strains for inoculation on the surface of 7-day-old barley or 2-week-old rice leaves. After 5 days, the wild-type strain caused typical rice blast lesions on both intact and abraded leaves, while the mogt2Δ mutants were nonpathogenic (Fig. 3C). When inoculated on abraded barley or rice leaves, the mogt2Δ mutants were still unable to cause disease symptoms, suggesting that in planta growth was also impaired (Fig. 3C). The loss of pathogenicity was fully restored in the complementation strain (Fig. 3C, Com). The mogt2Δ mycelia were unable to cause disease lesion as the WT or complementation mycelia did, suggesting that MoGt2 may play a role in host infection mediated by mycelia.
M. oryzae can form appressorium-like structures (ALSs) at hyphal tips to penetrate plant cuticles and develop invasive hyphae (23). We also harvested the mycelium of the wild type or mogt2Δ mutants and induced ALSs on hydrophobic GelBond film surfaces. We found that mogt2Δ mutants were unable to form ALS (Fig. 3D): thus, we conclude that MoGt2 is essential for ALS formation.
MoGT2 is involved in stress response. The fungal cell wall plays an important role in hyphal development and full virulence (24)(25)(26). We evaluated the effect of MoGT2 disruption on stress tolerance, by assessing the growth of wild-type or mutant mycelia on CM supplemented with salt stress (0.7 M NaCl or 1.0 M KCl), the osmotic stress (1.0 M sorbitol), or cell-wall-perturbing reagents (0.01% sodium dodecyl sulfate [SDS] or 200 g/ml Congo red [CR]). The mogt2Δ mutants showed significantly elevated sensitivity to various stressful conditions, as the size of mutant colonies was obviously reduced compared to those of the mutant under the untreated condition or WT colonies under the same treatment (Fig. 4A). Quantification of the growth inhibition rate based on colony diameter confirmed that the mogt2Δ mutants were more sensitive to these stressful conditions than the WT (Fig. 4B), suggesting that MoGT2 plays an important role in stress tolerance in M. oryzae.
We noticed that after growth in liquid CM for 60 h, the hyphae of mutants became noticeably darker than those of the wild-type or complemented strains (see Fig. S2A in the supplemental material), suggesting excess melanin accumulation in the mogt2Δ mutants. Consistent with this, transcription levels of melanin biosynthesis genes ALB1 and BUF1 (27) were significantly upregulated in the mogt2Δ mutants compared to that of the wild-type strain (Fig. S2B), indicating that MoGT2 is involved in regulation of melanin biosynthesis. We then used transmission electron microscopy (TEM) to examine the cell wall structure. However, no obvious differences in cell wall were observed between the wild type and the mogt2Δ mutants (Fig. S2C). Taken together, our results suggest that MoGT2 plays an important role in response to various stresses.
MoGT2 regulates hyphal hydrophobicity. Surface hydrophobicity is important for pathogenicity in plant-pathogenic fungi, including the rice blast fungus (28)(29)(30). MPG1 mutants showing an "easily wettable" phenotype, due to loss of hydrophobin production, and displayed defects in appressorium formation and disease symptom development (28). We observed that the colonies of the mogt2Δ mutants were morphologically distinct from the wild-type strain and failed to form appressorium; therefore, we intended to check the hydrophobicity of the mogt2Δ mutants. The 10-l drops of water or detergent solutions (0.2% SDS and 50 mM EDTA) were, respectively, placed on the surface of the wild-type or mogt2Δ strain. We found that drops of water remained intact on the surface of the wild-type colonies after 24 h of incubation. However, the hyphae on the surface of mogt2Δ mutants were gradually infiltrated (Fig. 4C). When treated with detergent solutions, drops of solutions immediately soaked into the surface of mogt2Δ mutants and rapidly expanded to the surrounding aerial hyphae compared to the wild-type strain (Fig. 4C). These results showed that MoGT2 regulates the hydrophobicity of aerial hyphae in M. oryzae.
We reasoned the wettable phenotype of mogt2Δ mutants may be attributed to downregulation of hydrophobin gene MPG1. To test this idea, we carried out quantitative RT-PCR (qRT-PCR) analysis and found that the expression levels of MPG1 were significantly reduced in the mogt2Δ mutants (Fig. 4D), indicating that MoGT2 is required for expression of MPG1.
Conserved DxD and QxxRW motifs are required for MoGT2 function. It has been reported that a number of type 2 glycosyltransferases contain conserved DxD and QxxRW motifs, which are located in nucleotide-binding and acceptor-binding domains, respectively (31)(32)(33). The DxD motif is involved in Rib and Mn phosphate coordination. Sequence analysis revealed that these two motifs are also present in MoGt2 (Fig. 5A).
To study the function of these motifs, we generated point mutation constructs pGt2 D156R , pGt2 D158R , and pGt2 Q301R and introduced them into the mogt2Δ-39 mutant, respectively. The resulting GT2 D156R and GT2 D158R mutants showed similar phenotypes to the mogt2Δ mutant, including defective colony growth and loss of conidiation and pathogenicity (Fig. 5B). On the other hand, we observed that introduction of GT2 Q301R fragment could partially restore the vegetative growth, conidiation, and pathogenicity ( Fig. 5B and C). Quantification of two GT2 Q301R strains' conidium production was (0.20 Ϯ 0.05) ϫ 10 6 and (0.30 Ϯ 0.09) ϫ 10 6 conidia per plate, respectively, levels of both of which were significantly reduced compared to the wild-type strain [(29.6 Ϯ 1.96) ϫ 10 6 spores per plate; P Ͻ 0.01]. Infection with mycelial plugs of the GT2 Q301R strain caused reduced disease lesion on the barley leaf explants (Fig. 5B). We further tested the pathogenicity of the GT2 Q301R mutant by inoculating its conidia onto barley or rice leaf explants, with the WT conidiation as a control. The results showed that the GT2 Q301R conidia were also weak in pathogenicity, compared to the wild-type conidia (Fig. 5C). Overall, these results suggested that the conserved DxD and QxxRW domains are necessary for the full function of MoGt2.
Altered glycoproteins in M. oryzae conidiation due to loss of MoGT2. To screen for the potential protein substrate(s) of the glycosyltransferase MoGt2 in M. oryzae during conidiation, we performed an analysis of protein glycosylation profiles with the wild-type and mutant strains. Total protein extracts from the wild-type or mogt2Δ mutant strain, cultured on solid medium and exposed to light for 12 to 16 h to induce conidiation, were subjected to SDS-PAGE and stained for glycoproteins. As shown in Fig. 6A, two bands, of approximately 100 to 140 kDa and 75 kDa, respectively, were present in the wild-type samples while absent in the mogt2Δ mutant. We cut down these two gel bands and sent them for mass spectrometry (MS) identification. In Table 1, we summarize the possible protein or proteins identified as band 1 or band 2, respectively; detailed information for peptide and protein identification is included in Data Set S1 in the supplemental material. We noticed that band 1 was most likely  a coiled-coil protein-containing protein, aminopeptidase 2, or a nuclease domaincontaining protein 1 (Table 1). Band 2 could also be a coiled-coil protein-containing protein (different from band 1), Hsp70, Hsp80/Hsp90, or Hsp70-like protein. A typical Hsp70 (MoSsb1) was reported critical for M. oryzae growth and pathogenicity and regulates the cell wall integrity (CWI) pathway governed by the mitogen-activated protein kinase (MAPK) signaling pathway (34). This Hsp70 (MoSsb1) protein was among the predicted band 2 proteins encoded by MGG_02503, as listed in Table 1.
Differentially expressed genes in the mogt2⌬ mutant during conidiation. We also performed a transcriptome analysis between the wild-type and mutant strains under the conidiation condition. We identified 3,808 differentially expressed genes (DEGs; |log 2 | Ն 1 and P Յ 0.05), of which 1,786 overlapped in the three biological replicates (see Data Set S2 in the supplemental material). These DEGs were enriched in metabolism, genetic information processing, environmental information processing, cellular processes, and organismal systems (Fig. 6B). Particularly, a conidiation-related gene, COS1 (35), was found significantly reduced in the mogt2Δ mutant compared to the WT (Data Set S2), which may support its function in M. oryzae conidiation. N-glycan biosynthesis was shown to be differentially regulated in the mogt2Δ mutant (Data Set S2), thus providing an explanation for the mutant's altered cell wall integrity, although no morphological difference was observed by TEM (Fig. S2C). Genes involved in DNA repair or replication, protein translation, and posttranslational modification were also among the DEGs (Fig. 6B; Data Set S2). Interestingly, we noticed the autophagy pathway was enriched (Data Set S2). The MAPK pathway responsible for osmotic response was differentially regulated, consistent with the elevated sensitivity of the mogt2Δ mutant under osmotic or cell wall stresses (Fig. 4). We infer that MoGt2 may regulate these important metabolic and environmental response processes to fulfill its function in M. oryzae conidiation. On the other hand, MoGt2 may also regulate the CWI pathway and oxidative response during host infection, which is important for fungal pathogenicity (26,36,37).

DISCUSSION
In this study, we identified and functionally characterized a predicted type 2 glycosyltransferase, MoGt2, in M. oryzae. MoGt2 is highly conserved among several filamentous fungi (Fig. 1) and might have conserved functions in fungal development and/or pathogenicity. Targeted deletion of MoGT2 resulted in impairment of vegetative growth, conidiation, stress response, hyphal appressorium-mediated penetration, and pathogenicity, suggesting an important role of MoGT2 in infection-related morphogenesis and pathogenesis in M. oryzae. As a member of the group 2 glycosyltransferase protein family, Gt2 contains the conserved DxD and QxxRW motifs. Our site-directed mutagenesis analysis confirmed that DxD and QxxRW motifs are critical for MoGt2 function.
Previous studies revealed that type 2 glycosyltransferase, GT2, is essential for hyphal growth in Z. tritici and F. graminearum (11). In M. oryzae, mogt2Δ mutants showed a similar phenotype to Z. tritici and F. graminearum gt2 mutants, further confirming a conserved function of Gt2 in fungal hyphal development. Such reduced mycelial growth in the mogt2Δ mutants may be due to shortening of interseptal distances as visualized by CFW staining.
The mogt2Δ mutant failed to produce conidia under several tested culture conditions, including CM, PA (prune agar), MM, or N-deplete medium. To assess MoGt2 function in M. oryzae pathogenicity, we performed an infection assay using the mycelial plugs as the mogt2Δ mutant did not produce conidia. We found that the mogt2Δ mutant was unable to invade host tissue or form an appressorium-like structure from mycelia to penetrate the host cuticle. Therefore, we reasoned that the loss of virulence in the mogt2Δ mutant may be caused by deficiency in appressorium-like structure formation. To get a better understanding of the role of MoGT2 in infection awaits silencing of this gene only during infection and observation of pathogenicity under such a condition. In addition, we found the mogt2Δ mutants showed an "easily wettable" phenotype and reduction in hydrophobin gene MPG1 transcription (Fig. 4D), which may account for impairment of conidium production and hyphal growth.
A MoGT2-GFP fragment was reintroduced into the mogt2Δ mutant and able to restore all phenotypes, indicating that the ectopically expressed MoGt2-GFP fusion protein is functional. However, we failed to detect visible GFP signal in the complementation strain under mycelial growth or conidiation or during the infection stage. The predicted topology of MoGt2 is that its C terminus is outside the plasma membrane, so we infer that the C-terminal GFP was cleaved and released to the extracellular space and therefore could not be used to assess subcellular localization of MoGt2.
Cell wall is an important structure that is responsible for maintaining cell shape and is also critical for cell expansion during growth and morphogenesis (49). In M. oryzae, cell wall integrity was essential for fungal pathogenesis (26,36,37). In this study, deletion of MoGT2 led to increased sensitivity to distinct stresses, including the osmotic stress and cell-wall-perturbing reagents. In N. crassa, it has also been reported that cps-1 deletion mutants are sensitive to cell wall perturbation reagents and play a critical role in cell wall biogenesis (12). However, TEM observation showed no obvious differences in cell wall ultrastructure between the wild type and mogt2Δ mutants (Fig. S2C).
By liquid chromatography-tandem MS (LC-MS/MS), we tried to identify two bands cut from SDS-PAGE for the glycoproteins present in the wild-type strain but absent in the mutant during conidiation. We identified two coiled-coil domain-containing proteins, several heat shock proteins, aminopeptidase 2, and nuclease domain-containing protein 1. Particularly, a typical Hsp70 protein, MoSsb1, was recently reported to be important for M. oryzae growth, conidiation, and pathogenicity and regulates the CWI pathway through interaction with MAPK MoMkk1. The mossb1Δ mutant displayed similar phenotypes to the mogt2Δ mutant, except that the mossb1Δ mutant produced conidia but of abnormal morphology (34). We infer that MoGt2 may regulate M. oryzae growth, pathogenicity, and CWI through glycosylation of MoSsb1, and other (unidentified) substrates may contribute to conidiation. However, we failed to predict any glycosylation residue on MoSsb1 by using the NetNGlyc 1.0 Server (Table 1). Whether MoSsb1 is actually glycosylated by MoGt2, as well as the biological relevance of such posttranslational modification, awaits further investigation. Verification of other potential substrates listed in Table 1 would also be of interest.
We also performed comparative transcriptome analysis to investigate the mechanism of MoGt2 function. A conidiation-related gene, COS1 (35), was found significantly reduced in the mogt2Δ mutant compared to the WT (Data Set S2). Two conidiationrelated genes, CON7 and HTF1, were not among the filtered DEGs (|log 2 | Ն 1 and P Յ 0.05) but were shown by qRT-PCR analysis to be downregulated in the mogt2Δ mutant (Fig. 3B). The melanin biosynthesis genes ALB1 and BUF1 were found significantly upregulated, and the hydrophobin-encoding gene MPG1 was downregulated, in the mogt2Δ mutant compared to the WT, which was consistently supported by the comparative transcriptome analysis (Data Set S2) and qRT-PCR ( Fig. 4D; Fig. S2B). This confirms that the results from comparative transcriptome analysis were reliable. Functional investigation of the candidate DEGs may help further elucidate the MoGt2 functional mechanism. Nuclease domain-containing protein 1 was identified as a potential glycoprotein in the wild-type strain during conidiation, which also contains a Tudor domain that was reported present in RNA-binding proteins (38). This potential RNA-binding nuclease may account for the DEGs between the wild type and the mogt2Δ mutant and is likely subject to regulation through glycosylation. Three glycosylation residues (Asn in the Asn-Xaa-Ser/Thr sequon) could be predicted in this protein (Table 1). Such a hypothesis needs further verification in the future.
Overall, our study identified a type 2 glycosyltransferase, MoGt2, in M. oryzae, responsible for vegetative hyphal growth, conidiation, pathogenicity, and CWI. In the present study, we did not perform an in vitro experiment to test the glycosyltransferase activity of MoGt2, nor did we confirm the biological function of candidate glycoproteins in M. oryzae conidiation and their relationship with MoGt2. Future investigation of these aspects will further reveal the cellular and molecular mechanisms of MoGt2 function in fungal development and pathogenicity.

MATERIALS AND METHODS
Strains and culture conditions. Magnaporthe oryzae strain Guy11 was used as the wild-type strain. The wild-type strains and corresponding transformants generated in this study were grown on CM at 25°C for 10 days. For the stress sensitivity test, the mycelial plugs of each strain were cultured on CM plates, to which were added 0.7 M NaCl, 1 M KCl, 1 M sorbitol, 0.01% SDS, and 0.2 mg/ml Congo red (CR), respectively. The diameters of the colonies were recorded 10 days after inoculation. Conidial development was assessed by harvesting conidia from the surface of 10-day-old plate cultures and determining the concentration of the resulting conidial suspension using a hemocytometer (Corning). Means and standard deviations were calculated based on three independent experiments.
Nucleic acid manipulation, qRT-PCR, and Southern blotting. General procedures for nucleic acid analysis followed standard protocols (39). Total RNA was extracted using PureLink RNA minikit (Invitrogen, USA) and used to synthesize first-strand cDNA using PrimeScript RT (TaKaRa). The qRT-PCR was performed on the QuantStudio 6 Flex (Applied Biosystems, USA) by using SYBR green PCR master mix (Vazyme, Nanjing, China) as per the manufacturer's instruction. Genomic DNAs were extracted from vegetative hyphae with the cetyltrimethylammonium bromide (CTAB) protocol (40). Southern blot analysis was performed with the digoxigenin (DIG) High Prime DNA labeling and detection starter kit II (Roche, Mannheim, Germany). The primers used in this study are listed in Table S1 in the supplemental material.
Plasmid constructs and fungal transformants. To construct the GT2 gene replacement vector pKO-GT2, the 1.5-kb upstream and 1.5-kb downstream sequences of MoGT2 were amplified with primer pairs LB F/LB R and RB F/RB R, respectively. The two flanking sequences were cloned into the pFGL821 vector to generate pKO1191.Then pKO1191 was transformed into Guy11 protoplasts to generate homologous recombinants, as previously described (28).
For GT2 complementation, a 3.7-kb fragment, including a 1.8-kb native promoter region and 1.9-kb full length of the MoGT2 gene, and the enhanced green fluorescent protein (eGFP) gene were amplified and then cloned into pCB1532 to create pGT2:eGFP (a C-terminal GFP tagging vector), according to the manufacturer's instructions of the One Step cloning kit (Vazyme, Nanjing, China).
For generating the GT2 D156R mutant, site-directed mutagenesis introduced D156R R/D156R F to replace D with R (GAT to CGA) in the Gt2 domain. Fragments of 2.1 and 1.8 kb were amplified with primer pairs Gt2 up F/D156R R and D156R F/Gt2 down R, respectively. Two PCR products were cloned into pCB1532 to create pD151R with the One Step cloning kit, as described above. The resulting plasmid was transformed into the mogt2Δ-39 mutant to generate the GT2 D156R mutant. Similar strategies were used to generate other site-directed mutants.
Pathogenicity assay. Two-week-old seedlings of the rice cultivar CO39 and 7-day-old seedlings of the barley cultivar Golden Promise were used for infection assays. The mycelium plugs from 10-day-old CM cultures were placed onto the leaf surface. Wounded leaves were prepared by removing the surface cuticle by abrasion with an emery board (41). The inoculated plants were incubated in a plastic plate with full humidity at 25°C. The disease lesions were examined and photographed at 5 days postinoculation.
Transmission electron microscope. Mycelium cultured in liquid CM for 2 days was processed for transmission electron microscopy (TEM). TEM sample treatment was performed as described previously (42), and treated samples were observed under a transmission electron microscope (Hitachi H-7650).
Staining assays. For calcofluor white staining, mycelia were stained with 10 g/ml CFW (18909; Sigma-Aldrich) for 10 min in the dark and then washed twice with phosphate-buffered saline (PBS) buffer.
Glycoprotein staining. The wild-type and mutant strains were grown in liquid CM for 2 days before harvesting the mycelia. Total protein extract from the mycelia was subjected to SDS-PAGE and stained using a Pierce glycoprotein staining kit (24562; Thermo Scientific).
Mass spectrometry. The protein gel band cut from the SDS-PAGE gel was digested with trypsin for 20 h at 37°C. The digested peptides were loaded onto a reverse-phase trap column (Thermo Scientific Acclaim PepMap100, 100 m by 2 cm; nanoViper C 18 ) connected to the C 18 reverse-phase analytical column (Thermo Scientific Easy column, 10 cm long, 75-m inner diameter, 3-m resin) in buffer A (0.1% formic acid) and separated with a linear gradient of buffer B (84% acetonitrile and 0.1% formic acid) at a flow rate of 300 nl/min (0 to 35% buffer B for 50 min, 35 to 100% buffer B for 5 min, hold in 100% buffer B for 5 min). Bioinformatic analysis of raw LC-MS/MS data followed the established protocol (43) and was performed by Applied Protein Technology Co., Ltd. (Shanghai, China).
Western blot analysis. Total protein was extracted from mycelium cultured in liquid CM for 2 days and separated on a 10% SDS-PAGE gel, before being transferred to polyvinylidene difluoride (PVDF) membrane. Detection of GFP fusion proteins was carried out using anti-GFP antibody (A6455, rabbit, 1:5,000; Invitrogen Molecular Probes). The SuperSignal West Pico chemiluminescent kit (34580; Thermo Scientific) was used for signal detection.
RNA-seq and transcriptome analysis. High-throughput RNA sequencing (RNA-seq) and transcriptome analysis were performed by Gene Denovo Co. (Guangzhou, China) using the reported protocols (43). Short reads were mapped to the complete genome of M. oryzae (https://www.ncbi.nlm.nih.gov/ genome/?termϭmagnaportheϩoryzae) using Tophat (44). Genes with a fold change of 2 and a falsediscovery rate (FDR) of 0.05 in a comparison of significant differentially expressed genes (DEGs) were subjected to enrichment analysis of Gene Ontology (GO) functions and KEGG pathways following our established protocol (43).

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.