Phenazine biosynthesis protein MoPhzF regulates appressorium formation and host infection through canonical metabolic and noncanonical signaling function in Magnaporthe oryzae

Summary

Microbe-produced secondary metabolite phenazine-1-carboxylic acid (PCA) facilitates pathogen virulence and defense mechanisms against competitors. Magnaporthe oryzae, a causal agent of the devastating rice blast disease, needs to compete with other phyllosphere microbes and overcome host immunity for successful colonization and infection.

However, whether M. oryzae produces PCA or it has any other functions remains unknown. Here, we found that the MoPHZF gene encodes the phenazine biosynthesis protein MoPhzF, synthesizes PCA in M. oryzae and regulates appressorium formation and host virulence.

MoPhzF, likely acquired through an ancient horizontal gene transfer event and has a canonical function in PCA synthesis. In addition, we found that PCA has a role in suppressing the accumulation of host-derived reactive oxygen species (ROS) during infection. Further examination indicated that MoPhzF recruits both the endoplasmic reticulum membrane protein MoEmc2 and the regulator of G-protein signaling MoRgs1 to the plasma membrane (PM) for MoRgs1 phosphorylation, which is a critical regulatory mechanism in appressorium formation and pathogenicity.

Collectively, our studies unveiled a canonical function of MoPhzF in PCA synthesis and a noncanonical signaling function in promoting appressorium formation and host infection.

Introduction

The rice plant serves as the host for a diverse array of microorganisms, providing an open and inviting habitat for microbial colonization and growth (Doni et al., 2022; Wang et al., 2022). Pseudomonas species stands out as a prominent rice phyllosphere bacterium that produces the secondary metabolite phenazine-1-carboxylic acid (PCA) to facilitate virulence and defense against competitors (Upadhyay & Srivastava, 2011; Biessy et al., 2019; Roman-Reyna et al., 2020; Rani et al., 2021). PCA production is required for full pathogenicity of the plant pathogenic bacterium P. syringae pv. tomato (Pst) DC3000 in tomato (Wen et al., 2016). In addition, PCA promotes biofilm development to render P. aeruginosa more resistant to antimicrobial agents by facilitating ferrous iron acquisition and functions as a virulence factor contributing to fast kill of Caenorhabditis elegans (Wang et al., 2011; Cezairliyan et al., 2013; Sakhtah et al., 2016). Furthermore, PCA secreted by P. fluorescens 2–79 and P. aureofaciens 30–84 could inhibit the wheat take-all disease caused by Gaeumannomyces graminis var. tritici and ensure host survival in the plant rhizosphere and soil (Thomashow & Weller, 1988; Mazzola et al., 1992). In the naturally suppressive soils of Châteaurenard (France), PCA-producing Pseudomonas spp., in combination with non-pathogenic Fusarium oxysporum, exerts suppressive effects on Fusarium wilt through direct antibiosis and/or indirect competition for iron resources (Mazurier et al., 2009; Biessy & Filion, 2018).

In bacteria, PCA synthesis is a complex process mediated by a highly conserved phenazine biosynthetic operon consisting of seven unique phenazine biosynthesis (phz) genes, typically named phzA-G in P. chlororaphis, P. fluorescens, and P. aeruginosa (Mazurier et al., 2009; Biessy & Filion, 2018). The proteins encoded by these genes within the phenazine operon orchestrate the conversion of the phenazine biosynthesis precursor, chorismic acid, into PCA (Blankenfeldt et al., 2004; Mavrodi et al., 2004; Parsons et al., 2004). PhzF exhibits significant structural similarity to members of the diaminopimelate epimerase (DapF) fold family of proteins and catalyzes an allylic rearrangement to yield a 3-oxo derivative of trans-2,3-dihydro-3-hydroxyanthranilic acid (DHHA) and dimerizes DHHA to PCA (Blankenfeldt et al., 2004; Parsons et al., 2004). Previous studies have shed light on the complex evolutionary history of phenazine genes and suggested that the evolution and dispersal of phenazine genes are driven by mechanisms ranging from conservation in Pseudomonas spp. to horizontal gene transfer (HGT) in Burkholderia spp. and Pectobacterium spp. (Fitzpatrick, 2009; Mavrodi et al., 2010). Indeed, it was reported that fungi, including Candida parapsilosis, Schizosaccharomyces pombe, and Nesidiocoris tenuis, acquired PhzF homologs from bacterial sources through HGT (Fitzpatrick et al., 2008; Ferguson et al., 2021).

Several studies have identified that the endophytic Nigrospora oryzae and sea anemone-derived Emericella sp. SMA01 produce PCA (Thanabalasingam et al., 2015; Yue et al., 2021). The plant-pathogenic fungus Truncatella angustata also produces PCA as a competitive mechanism against other fungi pathogenic to grapevine (Cimmino et al., 2021). Recent studies showed that the rice blast fungus Magnaporthe oryzae is more resistant to PCA than other fungal pathogens, including Rhizoctonia solani, Alternaria solani, and Phytophthora capsici (Yu et al., 2018; Xiong et al., 2019; Zhu et al., 2019; Li et al., 2021). M. oryzae infects the rice host, where it must face the daunting task of overcoming host defenses and contending with other microorganisms (Zhang et al., 2016; Liu et al., 2020). However, whether M. oryzae produces PCA or its infection involves other functions of PCA remains unknown.

Here, we successfully detected PCA in M. oryzae using liquid chromatography-tandem mass spectrometry (LC-MS-MS) and characterized the function of the MoPHZF gene involved in PCA biosynthesis. Through genetic, biochemical, and bioinformatic analyses, we found that MoPhzF is responsible for PCA production in M. oryzae. PCA suppresses host reactive oxygen species (ROS) accumulation to promote infection. Evidence also suggested that M. oryzae might acquire MoPHZF through HGT. Interestingly, MoPhzF has evolved a critical regulatory function in the growth, conidiation, and appressorium formation of M. oryzae. We have provided evidence indicating that MoPhzF functions as the scaffold protein to recruit MoEmc2 and MoRgs1 to the plasma membrane where MoRgs1 phosphorylation occurs.

Materials and Methods

Chemicals

A highly purified phenazine-1-carboxylic acid (purity > 99.88%) was purchased from Shanghai Haohong Biomedical Technology Co., Ltd. It was dissolved in DMSO, prepared as stock solutions at 20 mg/mL, and stored at 4°C before further use.

Strains and cultural conditions

Magnaporthe oryzae Guy11 was used as the wild-type strain in this study. All strains were cultured on the complete medium (CM) for 3–15 days in the dark at 28°C (Talbot et al., 1993; Liu et al., 2021; Guo et al., 2023). Mycelia were harvested from the liquid CM for 2 days in darkness for extraction of DNA, RNA, and protein (Yin et al., 2020; Feng et al., 2021).

Extraction and identification of phenazine-1-carboxylic acid

Strains were harvested in liquid CM for 48 h. Mycelia were ground into a fine powder in liquid nitrogen and lyophilized for 24 h. Phenazine-1-carboxylic acid was extracted following previously established procedures (Song et al., 2020). Samples were precisely weighed and mixed with ultrapure water. The mixtures were initially acidified to pH 2.0 with 12 M HCl followed by extraction with three volumes of chloroform under vigorous shaking. The supernatant was collected by centrifugation and completely evaporated in a rotary evaporator. The resulting residues were dissolved in acetonitrile and used for analysis.

The extracts of all strains were filtered through a syringe filter (0.22 μm) and subjected to liquid chromatography-tandem mass spectrometry (LC-MS-MS) for the characterization of secondary metabolites. The highly purified PCA (purity > 99.88%) was detected by LC-MS-MS in positive ion mode as standard. Separations were achieved on a ACQUITY UPLC^® BEH C18 column (2.1×50 mm, 1.7 μm particle size). A gradient used to separate the metabolites consisted of 0.1% formic acid in water (Solvent A) and acetonitrile with 0.1% formic acid (Solvent B) (Shahid et al., 2017). The total LC-MS-MS run was 4 min with a flow rate of 1 mL/min, and the injection volume was 0.5 μL.

Gene identification and phylogenetic analysis

To identify the homologous proteins related to the phenazine biosynthetic pathway, the associated proteins from model strain Pseudomonas aeruginosa PAO1 were used as a query in BLASTp search against the NCBI non-redundant protein database. To investigate the distribution of MoPhzF homologs in other organisms, the MoPhzF protein sequence was used as a query in BLASTp search against the NCBI non-redundant protein database. The top 200 database hits (E-value < E-70, bit-score > 230) were downloaded by BLASTp search. The phylogeny construction was conducted using the neighbor-joining method and 1,000 bootstrap replicates (with bootstrap values in black showing on respective branches) (Li et al., 2022). The resulting tree was midpoint rooted, and the tree layout was further edited using MEGA5.2 (Xia et al., 2021).

Targeted gene deletion and complementation

The MoPHZF gene deletion mutants were generated by the standard one-step gene replacement strategy. First, two approximately 1.0 kb of sequences flanking the MoPHZF were amplified with primer pairs in Supplementary Table 1. Afterwards, the resulting PCR products were digested with restriction endonucleases and ligated with the hygromycin-resistant gene cassette released from pCX62. Finally, the recombinant insert was sequenced. The 3.4 kb fragment included the flanking sequences, and the HPH cassette was amplified and transformed into Guy11 protoplasts. Putative mutants were screened by PCR and confirmed by Southern blotting analysis.

For complementation and amino acid substitution experiments, complement fragments were amplified with primer pairs (Table S1) and inserted into the pYF11 vector (bleomycin resistance). After sequencing, the fused-pYF11 plasmids were transformed into the mutant strains through PEG-mediated transformation (Guo et al., 2023).

Assays for vegetative growth and conidiation

For vegetative growth, small agar blocks were cut from the edge of 5-day-old cultures and placed onto CM, minimal medium (MM), as well as straw decoction and corn (SDC: 100 g of straw, 40 g of corn powder, 15 g of agar in 1 L of distilled water) medium followed by incubation in the dark at 28°C (Yin et al., 2019; Zhang et al., 2019). The radial growth was measured following incubation for 7 days (Yin et al., 2019).

For conidia production, strains were grown on SDC medium at 28°C for 7 days in the dark, followed by constant illumination for 3 days (Yu et al., 2021). Conidia harvested from 10-day-old SDC cultures were filtered through two layers of Miracloth (EMD Millipore Corporation, 475855–1R), resuspended in sterile water, and counted using a hemocytometer (Liu et al., 2020).

Appressorium formation assay and intracellular cAMP level measurement by high-performance liquid chromatography

For appressorium formation, conidia harvested from 10-day-old SDC cultures were filtered through two layers of Miracloth and washed with sterilized water three times. Droplets (20 μL) of conidial suspension (5×10⁴ spores/mL) were placed on microscope cover glass (Fisher Scientific, 16938) under humid conditions at 28°C in darkness, and the samples were microscopically observed (Yin et al., 2019). The percentages of appressorium formation were determined by microscopic examination at intervals.

The intracellular cyclic adenosine monophosphate (cAMP) content was determined according to the method described previously (Liu et al., 2016; Li et al., 2019). All strains cultured in liquid CM for 48 h were harvested, frozen in liquid nitrogen, and lyophilized for 24 h. 1 mg of mycelia was mixed with 20 μL of 6% (w/v) trichloroacetic acid (TCA) solution and kept on ice for 10 min. The mixtures were centrifuged at 1377 g for 15 min at 4°C, and the supernatant was collected and washed twice with five times the volume of anhydrous ether. The extract solution was filtered with 0.22 μm filter membrane and collected for HPLC. HPLC analysis was performed as reported (Liu et al., 2016; Li et al., 2019).

Protein extraction and Western blot analysis

For total protein extraction, strains were cultured in liquid CM with shaking for 48 h. Germinating conidia on an inductive surface for 3 h or mycelia were collected and ground into fine powder in liquid nitrogen, then resuspended in 1 mL lysis buffer (10 mM Tris-HCl, pH 7.5; 150 mM NaCl; 0.5 mM EDTA; 1% Triton X-100) with 2 mM PMSF and proteinase inhibitor cocktail. The lysates were placed onto ice for 30 min and shaken once every 10 min. Cell debris was removed by centrifugation at 13,000 g for 10 min at 4°C, and the supernatant lysates were collected as total proteins (Li et al., 2019; Liu et al., 2020).

Total proteins were separated by 12% SDS-polyacrylamide gels with Western blotting analysis. For Flag-tagged protein detection, samples were analyzed with the anti-Flag antibody (mouse, 1:5000; Abmart, M20018) and the anti-mouse secondary antibody (1:10,000, LI-COR; IRDye, C70301–02). For GFP-tagged protein detection, samples were analyzed with the anti-GFP antibody (mouse, 1:5000; Abmart, 293967) and the anti-mouse secondary antibody (1:10,000, LI-COR; IRDye, C70301–02). For RFP-tagged protein detection, samples were analyzed with the anti-RFP antibody (mouse, 1:5000; Chromotek, 6g6–100) and the anti-mouse secondary antibody (1:10,000, LI-COR; IRDye, C70301–02). Blot signals were detected and analyzed using the ODYSSEY infrared imaging system (Version 2.1).

Phosphorylation analysis through Phos-tag gel electrophoresis

The MoRgs1-GFP fusion construct was introduced into the wild-type (Guy11) and ΔMophzf strains. The total protein extracted from transformants during appressoria stages was resolved on 10% SDS-PAGE prepared with 50 μM acrylamide-dependent Phos-tag ligand and 100 μM MnCl₂ as described (Liu et al., 2020; Yin et al., 2020; Yu et al., 2021). Gel electrophoresis was performed with a constant voltage of 80 V for 12 h. Before transferring, gels were equilibrated in transfer buffer with 5 mM EDTA for 20 min three times and followed by transfer buffer without EDTA for another 20 min. Protein transfer from the Mn²⁺-phos-tag acrylamide gel to the PVDF membrane was performed for 48 h at 80 V at 4°C, and then the membrane was analyzed by Western blotting using the anti-GFP antibody (Liu et al., 2020; Yu et al., 2021).

Yeast two-hybrid (Y2H) and co-immunoprecipitation (co-IP) assays

The bait construct was generated by cloning full-length MoPhzF cDNA into pGBKT7. The prey constructs of AD:MoRgs1, AD:MoEmc2, and AD:MoCk2 were obtained previously (Yu et al., 2021). Prey and bait constructs were confirmed by DNA sequencing and co-transformed into the yeast strain AH109 following the recommended protocol (BD Biosciences Clontech). Transformants grown on the synthetic medium lacking leucine and tryptophan (SD-Leu-Trp) were transferred to the synthetic medium lacking leucine, tryptophan, adenine, and histidine (SD-Leu-Trp-Ade-His). Transformants screened by synthetic dextrose medium minus leucine, tryptophan, adenine, and histidine (SD-Leu-Trp-Ade-His) were selected. Yeast strains for positive and negative controls were provided by the BD library construction and screening (Hu et al., 2022; Guo et al., 2023).

To confirm in vivo interactions among MoPhzF-MoRgs1, MoPhzF-MoEmc2, and MoRgs1-MoEmc2, full-length gDNA with a 1.5 kb native promoter region of MoPhzF was cloned into pHZ126 labeled by the FLAG tag (Liu et al., 2020). For MoPhzF, since we could not detect any GFP using its native promoter, the strong constitutively activated ribosomal protein 27 (RP27) promoter was used (Zhang et al., 2019; Liu et al., 2021). Constructs of MoRgs1-GFP and MoEmc2-RFP were obtained previously (Yu et al., 2021). Different pairs of specific constructs were co-transformed into the protoplasts of Guy11 or ΔMophzf. Total proteins were isolated from different positive transformants during appressoria stages and incubated with anti-GFP affinity beads (Smart lifesciences, SA070001) at 4°C for 4 to 12 h with gentle shaking. Proteins bound to the beads were eluted after a series of washing steps by 1×PBS. Elution buffer (200 mM glycine, pH 2.5) and neutralization buffer (1 M Tris, pH 10.4) were used for the elution process. Total, suspension, and eluted proteins were analyzed by Western blot using GFP (mouse, 1:5000; Abmart, 293967) or FLAG (mouse, 1:5000; Abmart, M20018) specific antibodies or RFP (mouse, 1:5000; Chromotek, 6g6–100).

Bimolecular fluorescence complementation (BiFC) assay

For the BiFC assay, MoPhzF fusion protein constructs were generated by cloning full-length gDNA with a 1.5 kb native promoter region into pHZ65 and pHZ68 (BiFC vectors from Dr. J. R. Xu of Purdue University). MoEmc2-YFP^N and MoRgs1-YFP^C fusion constructs were obtained previously (Yu et al., 2021). MoPhzF-YFP^N and MoRgs1-YFP^C, and MoPhzF-YFP^C and MoEmc2-YFP^N were introduced into the protoplasts of Guy11, respectively. Transformants resistant to both hygromycin and zeocin were isolated and confirmed by PCR analyses. YFP signals were examined with a confocal fluorescence microscope (Zeiss LSM710, 63×oil) (Yin et al., 2020).

Localization observation of MoRgs1-GFP, MoEmc2-GFP, and MoPhzF-GFP

To investigate the cellular localization of MoRgs1, MoEmc2, and MoPhzF, MoPhzF-GFP, MoRgs1-GFP, and MoEmc2-GFP fusion constructs were transformed into Guy11, ΔMophzf, ΔMorgs1, or ΔMoemc2 strains. Conidia harvested from 10-day-old SDC medium plates were germinated on an inductive surface for 3 h, and microscopy observation was carried out (Zeiss LSM710, 63×oil). The filtered channels set: GFP (excitation spectra: 488 nm, emission spectra: 510 nm). Insets highlight areas analyzed by line-scan (Li et al., 2019).

Virulence and rice sheath penetration assays

Conidia were harvested from 10-day-old SDC agar cultures, filtered through two layers of Miracloth, and resuspended to a concentration of 5×10⁴ spores/mL in a 0.2% gelatin solution. For spray inoculation, 5 mL of a conidial suspension of each treatment was sprayed onto two-week-old rice seedlings (Oryza sativa cv. CO39). Inoculated seedlings were kept in a growth chamber at 28°C with 90% humidity for the first 24 h and then subjected to a light-dark cycle for 5 to 7 days (Yu et al., 2021; Guo et al., 2023). The degree of disease lesions was analyzed by ImageJ.

For the penetration assay, conidial suspensions (2×10⁵ spores/mL) were injected into the 3-week-old rice sheath using 1 mL syringes, and the inner epidermises of infected sheaths were harvested at 36 h for microscopic observation (Liu et al., 2021).

Observation of septin ring formation

To examine the septin ring formation in M. oryzae, MoSep3-GFP fusion constructs were transformed into Guy11 and ΔMophzf. Conidial suspensions (5×10⁴ spores/mL) of Guy11 and ΔMophzf were inoculated on an inductive surface for 24 h in the presence or absence of PCA (5 μg/L), and observed under a confocal fluorescence microscope (Zeiss LSM710, 63×oil).

ROS observation

For 3,3’-diaminobenzidine (DAB) staining, the infected rice sheaths were incubated with 1 mg/mL DAB solution (pH3.8) at room temperature for 8 h in the dark and destained with clearing solution (ethanol:acetic acid, 94:4, v/v) for 4 h (Zhang et al., 2019; Liu et al., 2021). The inner epidermises of infected sheaths were harvested and observed under a microscope (Zeiss, Axio Observer A1).

For 5-(and −6)-chloromethyl-2’, 7’-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H₂DCFDA) staining, the epidermal layers of infected rice leaf sheaths were incubated in 1 mL water for 5 min to remove wound-induced ROS and followed by staining with 2 μM CM-H₂DCFDA (Molecular Probes Life Technologies, Eugene, OH, United States) in the dark for 30 min at room temperature. The incubated sheath samples were washed three times with 1×PBS buffer for 5 min in the dark and the cells were observed immediately using a confocal fluorescence microscope (Chen et al., 2020; Dangol et al., 2021; Hu et al., 2022).

Nitroblue tetrazolium (NBT) staining

Conidia harvested from 10-day-old SDC cultures were inoculated on microscope cover glass (Fisher Scientific, 16938) for the indicated time under humid conditions at 28°C in darkness, and followed by staining with 0.1% NBT solution for 20 min (Kou et al., 2017; Liu et al., 2019). The samples were observed under a microscope (Zeiss, Axio Observer A1).

RNA isolation and quantitative RT-PCR assay

To detect the transcription of respiratory burst oxidase homologous genes during the invasive hyphae growth stage in the host, total RNA was extracted from rice leaves sprayed with conidial suspensions of Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF strains following PCA treatments at 48 hpi.

RT-qPCR was used to evaluate the gene expression. Total RNA was isolated using the Total RNA Extraction Kit (PuDi, China) and reverse transcribed into first-strand cDNA with a reverse transcription kit (Vazyme, Nanjing). For gene expression analysis, RT-qPCR was performed on the Applied Biosystems 7500 Real-Time PCR System with ChamQ SYBR^® qPCR Master Mix (Vazyme, Nanjing). Primer Premier 5.0 (PREMIER Biosoft, USA) was used to design the primers, as listed in Table S1. The OsACTIN gene (XM_015774830.2) was used as the reference gene for normalization, and relative expression levels were calculated using the 2^−ΔΔCt method, as previously described (Hu et al., 2022).

Identification of rice proteins interacting with phenazine-1-carboxylic acid

To identify the rice proteins that interact with phenazine-1-carboxylic acid, chemical proteomics by the LiP-SMap approach was performed according to previous studies (Piazza et al., 2018; Zhang et al., 2022). Two-week-old seedlings of rice cv. CO39 were collected and lyophilized for 24 h, then ground into fine powder in liquid nitrogen and resuspended in precooled native lysis buffer (20 mM HEPES, 150 mM KCl, 10 mM MgCl₂ pH 7.5). After homogenized for 4 min in 35 Hz, the mixture was placed onto the ice for 30 min and shaken once every 5 min followed by centrifugation at 12,000 g for 15 min at 4°C. The supernatant was collected and aliquoted in equivalent volumes containing 100 μg of proteins each.

33 nmol phenazine-1-carboxylic acid (dissolved in DMSO) was added to each aliquot and incubated at 25°C for 10 min. Limited proteolysis was conducted by adding proteinase K (Sangon Biotech, China) to all the proteome-metabolite samples at a 1:100 enzyme-substrate ratio, and the generated protein fragments were then completely digested by trypsin with an enzyme-substrate ratio of 1:50 to generate peptides for mass spectrometry analysis. Peptide samples were separated and analyzed with a nano-UPLC (EASY-nLC1200) coupled to a Q Exactive HFX Orbitrap instrument (Thermo Fisher Scientific) with a nano-electrospray ion source. Proteins and peptides were queried against the Rice UniProt FASTA database.

Statistical analysis

The experiments were performed in triplicates, and statistical analyses were conducted using SPSS software (SPSS Inc., Chicago, IL, USA). The results were represented as means ± standard deviations (SD). The significant differences between samples were statistically determined by one-way analysis of variance (ANOVA) comparison and followed by Duncan’s new multiple range tests if the ANOVA analysis is significant at p < 0.05.

Results

Identification and characterization of MoPhzF from M. oryzae

To identify potential phenazine biosynthesis genes in M. oryzae, the phenazine biosynthesis protein PhzF1 of Pseudomonas aeruginosa PAO1 was used to search against the M. oryzae genome. MGG_12726 was identified as the locus encoding a phenazine biosynthesis protein homolog named MoPhzF. MoPhzF consists of 325 amino acids, including a PhzC-PhzF domain (Fig. 1a), and it shares an overall 22.22% amino acid sequence similarity with PhzF1 of P. aeruginosa (Fig. S1). The conserved PhzC-PhzF domain is responsible for the dimerization of two trans-2,3-dihydro-3-hydroxyanthranilic acid (DHHA) molecules to generate PCA in P. fluorescens (Blankenfeldt et al., 2004; Parsons et al., 2004). MoPhzF is predicted to harbor a catalytic active site, Glu77 (https://www.uniprot.org/uniprot/G4N988, Fig. 1a). To test whether this residue is critical for MoPhzF function, we replaced Glu77 with aspartic acid. Previous studies indicated that (2S,3S)-trans-2,3-Dihydro-3-hydroxyanthranilic acid (DHHA) is the substrate of bacterial PhzF (Blankenfeldt et al., 2004; Parsons et al., 2004). MoPhzF and MoPhzF^E77D structures were predicted by AlphaFold2 (Fig. 1b), and a docking analysis of MoPhzF and MoPhzF^E77D with DHHA was performed. As illustrated in Fig. 1c, MoPhzF^E77D lost the binding ability to DHHA, compared to MoPhzF, indicating a critical role of Glu77 in MoPhzF function.

Fig. 1. The MoPhzF domain structure and critical active site for its binding ability to trans-2,3-dihydro-3-hydroxyanthranilic acid (DHHA).

(a) A schematic representation of MoPhzF. PhzC-PhzF domain (blue rectangle) and the catalytic active site (Glu77, 77E, orange line) were predicted by UniProt (https://www.uniprot.org/uniprot/G4N988). aa, amino acid. (b) Structures of MoPhzF and MoPhzF^E77D predicted with AlphaFold2. (c) Docking analysis of MoPhzF and MoPhzF^E77D with DHHA. Binding affinity is obtained by AutoDock Vina. Glu77 and Asp77 are indicated by orange circles. DHHA is indicated by a blue circle.

To examine the distribution of MoPhzF homologs in other microorganisms, BLASTp search against the NCBI non-redundant database was performed, which showed that the top 200 hits were from ascomycete fungi and bacteria (E-value < E-70, bit-score > 230). MoPhzF is most homologous to PhzF proteins from Pyricularia pennisetigena, Plectosphaerella cucumerina, and Cladophialophora bantiana (Fig. S2). In addition, MoPhzF is homologous to PhzF proteins from bacteria, including P. viridiflava, Tardiphaga robiniae, Paraburkholderia fungorum, and Burkholderia sp. S171 (Fig. S2). To infer evolutionary events that may explain the wide distribution of PhzF, we constructed a phylogenetic tree based on the alignment of MoPhzF and its homologs. The phylogenetic analysis showed that MoPhzF lies in a strongly supported clade with ascomycete fungi, and this clade is surrounded by bacterial clades (Fig. S2). The results suggested that the ancestor of PHZF in ascomycetes might have been horizontally transferred from bacteria.

MoPhzF is required for phenazine-1-carboxylic acid (PCA) production

To characterize the function of MoPhzF, we obtained two ΔMophzf mutant strains by replacing the coding region with the hygromycin-resistance cassette, which was then verified by Southern blotting analysis (Fig. S3). Two independent mutants showed similar defective phenotypes and thus only one mutant strain was used for further studies. Since MoPhzF is predicted to be a phenazine biosynthesis PhzF family protein, the phenazine-1-carboxylic acid (PCA) level was examined by LC-MS-MS. We found that the molecular weight of PCA is 224.21, with a retention time of 2.27 min, and two typical peaks with a mass of 225.11 and 207.10 Da, respectively (Fig. 2a,b). A similar peak was found in both the wild-type Guy11 and ΔMophzf/MoPHZF complemented strains, with PCA concentrations of about 4.35 ng/g and 5.11 ng/g, respectively (Fig. 2c,d). No such a peak was found in the ΔMophzf mutant strain (Fig. 2c). These findings indicated that MoPhzF is required for PCA production in M. oryzae.

Fig. 2. MoPhzF is essential for phenazine-1-carboxylic acid (PCA) production.

(a) Mass spectrum of highly purified PCA in positive ion mode. The ion peaks pointed by the arrows indicate the presence of PCA with m/z 225.11 and the fragment m/z of 207.10. (b) The presence of a single peak pointed by the arrows at 2.27 min indicates that high purity PCA was used as the standard. (c) Liquid chromatography-tandem mass spectrometry (LC-MS-MS) analysis of phenazine-1-carboxylic acid in Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF. (d) A standard curve of PCA and the content of PCA in different strains. Data were represented by means ± standard deviations. Data sets marked with asterisks are significantly different from the Guy11 strain under the same conditions (Student’s t-test, ** p < 0.01). PCA, phenazine-1-carboxylic acid. All experiments were conducted with three biological repetitions and three replicates.

To test whether the catalytic active site, Glu77, is critical for MoPhzF function in PCA production, we replaced Glu77 with aspartic acid using site-directed mutagenesis, and the resulting MoPhzF^E77D was expressed in the mutant strain (ΔMophzf/MoPHZF^E77D) and PCA quantified by LC-MS-MS. Unlike Guy11 or ΔMophzf/MoPHZF, ΔMophzf/MoPHZF^E77D barely produced any PCA (Fig. 2c,d), indicating a critical role of Glu77 in MoPhzF-mediated PCA production for MoPhzF function in PCA production, which was consistent with the docking analysis of MoPhzF and MoPhzF^E77D with DHHA.

MoPhzF contributes to M. oryzae development independent of PCA production

To determine whether MoPhzF has a role in the growth and development of M. oryzae, we compared the vegetative growth and conidiation of the ΔMophzf mutant with the wild-type strain Guy11, and the complemented mutant strain ΔMophzf/MoPHZF. The ΔMophzf mutant exhibited significant defects in vegetative growth and conidiation (Fig. S4a–d). Considering that the appressorium plays a key role in host invasion (Fernandez & Orth, 2018), we measured appressorium formation rates on the artificial inductive surfaces at continuous time points. The results showed that ΔMophzf exhibited a delay in appressorium formation (Fig. S4e). Since previous studies linked appressorium formation directly to cAMP signaling regulated by the adenylate cyclase MoMac1 that positively regulates cAMP production (Choi & Dean, 1997; Liu et al., 2016). The intracellular cAMP levels were estimated by high-performance liquid chromatography (HPLC). The ΔMophzf mutant has a relatively higher cAMP level than that of Guy11 and ΔMophzf/MoPHZF (Fig. S4f). Consistent with this result, the expression of MoMAC1 was significantly up-regulated in the ΔMophzf mutant, compared to Guy11 (Fig. S4g).

To determine whether the development deficiency in ΔMophzf was due to the lack of PCA, we examined the phenotype of the ΔMophzf/MoPHZF^E77D strain. Intriguingly, vegetative growth, conidiation, appressorium formation, intracellular cAMP levels, and the expression of MoMAC1 were all similar to Guy11 and the complemented strain (Fig. S4). In addition, the exogenous addition of PCA had no effects on the vegetative growth, conidiation, and appressorium formation in ΔMophzf (Table S2), indicating that the function of MoPhzF in the growth and development of the fungus is independent of PCA production.

MoPhzF regulates appressorium formation by mediating MoRgs1 membrane recruitment and protein phosphorylation

Given that an appressorium provides the gateway for establishing the blast disease (Eseola et al., 2021) and the ΔMophzf mutant exhibited a defect in appressorium formation, we sought to identify the functional relevance of MoPhzF interactors known to be involved in appressorium formation. In previous studies of MoRgs1, a key regulator of G-protein signaling (RGS) known to be crucial for appressorium formation and pathogenicity (Liu et al., 2007; Zhang et al., 2011; Yu et al., 2021), MoPhzF was identified as a candidate target. We thereby verified the interaction between MoRgs1 and MoPhzF by Y2H, BiFC, and co-IP assays (Fig. 3a–c). Since MoRgs1 undergoes protein phosphorylation (Yu et al., 2021), we hypothesized that MoPhzF might have a role in regulating MoRgs1 phosphorylation. To test this hypothesis, we performed in vivo phosphorylation assays using Mn²⁺-phos-tag gel electrophoresis analysis following the introduction of the MoRgs1-GFP fusion construct into Guy11 and the ΔMophzf mutant. Total proteins were extracted and treated either with or without a phosphatase or a phosphatase inhibitor (PI), and mobility shifts were examined by immunoblotting with the anti-GFP antibody as previously described (Liu et al., 2020; Yu et al., 2021). A specific mobility shift of MoRgs1 corresponding to the unphosphorylated MoRgs1 was observed in the phosphatase-treated wild-type cells but not in the untreated or PI-treated cells (Fig. 3d). No similar band shifts were observed in the ΔMophzf mutant (Fig. 3d), suggesting that MoPhzF is indeed involved in MoRgs1 phosphorylation.

Fig. 3. MoPhzF regulates appressorium formation and pathogenicity by mediating MoRgs1 protein phosphorylation.

(a) Yeast two-hybrid analysis of the interaction between MoPhzF and MoRgs1. MoPhzF and MoRgs1 cDNA was inserted into pGBKT7 and pGADT7, respectively. Yeast transformants co-expressing the bait and prey constructs were incubated on SD-Leu-Trp plates and replicated onto SD-Leu-Trp-His-Ade plates. Transformants co-expressing pGADT7-RECT and pGBKT7–53, and pGADT7-RECT and pGBKT7-Lam were used as the positive control and negative control, respectively. (b) Bimolecular fluorescence complementation assay showing the interaction between MoPhzF and MoRgs1. The transformant co-expressing the MoPhzF-YFP^N and MoRgs1-YFP^C was observed for fluorescence using confocal fluorescence microscopy (Zeiss LSM710, 63×oil) during the germ tube hooking stage (3 h). Strains co-expressing MoPhzF-YFP^N and empty YFP^C, MoRgs1-YFP^C and empty YFP^N, and empty YFP^C and empty YFP^N were used as negative controls. YFP, yellow fluorescent protein. Bar represents 10 μm. (c) Co-immunoprecipitation assay for the interaction between MoPhzF and MoRgs1. Proteins were extracted from transformants co-expressing MoPhzF-Flag and MoRgs1-GFP during the germ tube hooking stage (3 h) and incubated with anti-GFP agarose beads and eluted. Total and eluted proteins were analyzed by Western blotting, and the presence of MoRgs1 and MoPhzF were detected with anti-GFP and anti-Flag antibodies, respectively. GFP was used as the negative control. T: Total proteins. E: Eluted proteins. Coomassie brilliant blue (CBB) staining indicates loading controls. (d) In vivo phosphorylation analysis of MoRgs1 in Guy11 and the ΔMophzf mutant. MoRgs1-GFP proteins were extracted from transformants during the germ tube hooking stage (3 h) and treated with phosphatase or phosphatase inhibitors, and analyzed by Mn²⁺-Phos-tag SDS-PAGE and normal SDS-PAGE with the anti-GFP antibody, respectively. PI, Phosphatase inhibitor. Phosphorylated-MoRgs1-GFP, P-MoRgs1:GFP. (e) Appressorium formation assay. Conidia of Guy11, ΔMophzf, ΔMophzf/MoRGS1^5A, ΔMophzf/MoRGS1^5D, and ΔMophzf/MoPHZF were incubated on artificial inductive surfaces. Appressorium formation rates at different time points were calculated and analyzed. The percentage at a given time was recorded by observing at least 100 conidia for each strain, and calculated with three replicates. Data were represented by means ± standard deviations, and different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). (f) Quantification of intracellular cAMP levels. Intracellular cAMP from mycelia cultured in the liquid CM for 48 h were exacted, and levels were quantified by high-performance liquid chromatography (HPLC). Data were represented by means ± standard deviations, and different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). (g) Pathogenicity assay of Guy11, ΔMophzf, ΔMophzf/MoRGS1^5A, ΔMophzf/MoRGS1^5D, and ΔMophzf/MoPHZF. Rice (Oryza sativa cv. CO39) seedlings were sprayed with conidial suspensions (5×10⁴ spores/mL), and lesions were photographed at 7 days post-inoculation. (h) The disease lesion areas were assessed by Image J. Data were represented by means ± standard deviations, and different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). (i) Leaf sheaths of 3-week-old rice seedlings were inoculated with conidial suspensions (2×10⁵ spores/mL) of Guy11, ΔMophzf, ΔMophzf/MoRGS1^5A, ΔMophzf/MoRGS1^5D, and ΔMophzf/MoPHZF. Detailed observation and statistical analysis for invasive hyphae in rice sheath cells at 36 h post-inoculation. Appressorium penetration sites (n = 100) were observed, and invasive hyphae were rated from type 1 to 4 (type1, no penetration with only appressoria; type 2, only with a penetration peg or primary invasion hyphal; type 3, secondary invasive hypha extended but was limited in one plant cell; type 4, invasive hyphae extended to surrounding cells). Data were represented by means ± standard deviations. Bar represents 20 μm. All experiments were conducted with three biological repetitions.

Considering that MoPhzF is not a kinase, we hypothesized that phosphorylation might occur through an interaction between MoPhzF and a protein kinase. As a previous study identified five MoCk2 casein kinase-dependent serine phosphorylation sites (S396, S399, S585, S696, and S700) in MoRgs1 and constitutively activated MoRgs1 phosphorylation (MoRgs1^5D) could partially inhibit the defect in appressorium formation and pathogenicity (Yu et al., 2021), we tested the effects of phosphomimic MoRgs1^5D and constitutively unphosphorylated MoRgs1^5A on the ΔMophzf mutant. As we hypothesized, appressorium formation was elevated in the ΔMophzf/MoRGS1^5D strain when compared with ΔMophzf and ΔMophzf/MoRGS1^5A (Fig. 3e). In addition, since MoRgs1 phosphorylation is required for its GAP function, which is critical for maintaining the intracellular cAMP level in M. oryzae (Yu et al., 2021), we measured cAMP levels in all of the strains by HPLC. The ΔMophzf/MoRGS1^5D strain displayed a lower cAMP level than that of the ΔMophzf and ΔMophzf/MoRGS1^5A strains (Fig. 3f). Furthermore, rice seedlings virulence assays showed no difference between ΔMophzf/MoRGS1^5A and ΔMophzf, while ΔMophzf/MoRGS1^5D exhibited larger lesion areas (Fig. 3g,h). Consistent with this result, MoRgs1^5D partially rescued the virulence defect of the ΔMophzf mutant (Fig. 3i). Finally, MoRgs1^5D partially suppressed the defect of mycelial growth and conidiation in the ΔMophzf mutant (Fig. S5). These findings collectively indicated that the disturbance of MoRgs1 phosphorylation is, to some extent, responsible for the defect in the development and virulence of ΔMophzf.

MoPhzF functions as a scaffold protein to recruit MoRgs1 and MoEmc2 to the plasma membrane

Previous studies suggested that MoEmc2, an endoplasmic reticulum membrane protein complex subunit, modulates the subcellular localization of the casein kinase 2 MoCk2 for MoCk2-dependent MoRgs1 phosphorylation (Yu et al., 2021). To test whether MoPhzF is involved in this process, we examined the interaction between MoPhzF and MoEmc2, as well as MoPhzF and MoCk2, by Y2H, BiFC, and co-IP assays. We found that MoPhzF interacts with MoEmc2, but not with MoCk2, in vivo and in vitro (Fig. 4a–c, S6).

Fig. 4. MoPhzF is required for MoEmc2-mediated MoRgs1 membrane recruitment and phosphorylation.

(a) Yeast two-hybrid analysis of MoPhzF-MoEmc2 interaction. MoPhzF cDNA was inserted into pGBKT7, and MoEmc2 cDNA was inserted into pGADT7. Yeast transformants co-expressing the bait and prey constructs were incubated on SD-Leu-Trp plates and screened on SD-Leu-Trp-His-Ade plates. pGADT7-RECT and pGBKT7–53 were used as the positive control, and pGADT7-RECT and pGBKT7-Lam were used as the negative control. (b) Bimolecular fluorescence complementation assay for the interaction between MoPhzF and MoEmc2. The transformant co-expressing the MoPhzF-YFP^C and MoEmc2-YFP^N was observed with confocal fluorescence microscopy (Zeiss LSM710, 63×oil) during the germ tube hooking stage (3 h). Strains co-expressing MoPhzF-YFP^C and empty YFP^N, MoEmc2-YFP^N and empty YFP^C, and empty YFP^C and empty YFP^N were used as negative controls. YFP, yellow fluorescent protein. Bar represents 10 μm. (c) Co-immunoprecipitation assay for the interaction between MoPhzF and MoEmc2. Total proteins were extracted from transformants co-expressing MoPhzF-GFP and MoEmc2-RFP, incubated with anti-GFP agarose beads, and eluted. Total and eluted proteins were analyzed by Western blotting, and the presence of MoPhzF and MoEmc2 was detected by Western blotting with anti-GFP and anti-RFP antibodies, respectively. GFP was used as the negative control. T: Total proteins. E: Eluted proteins. Coomassie brilliant blue (CBB) staining indicates loading controls. (d) Co-immunoprecipitation assays for the interactions between MoRgs1 and MoEmc2 in Guy11 and ΔMophzf, respectively. Proteins were extracted and eluted as above. The presence of MoRgs1 and MoEmc2 was analyzed by Western blotting with the corresponding antibodies. T: Total proteins. E: Eluted proteins. Coomassie brilliant blue (CBB) staining indicates loading controls. (e,f) Subcellular localization of MoRgs1-GFP and MoEmc2-GFP was visualized in Guy11 and ΔMophzf during the germ tube hooking stage (3 h), respectively. White arrows indicated the regions where the fluorescence intensity was measured by line-scan analysis. The percentage of a pattern shown in the image was calculated by observation for 50 germinated conidia that were randomly chosen. PM, plasma membrane. Bar represents 10 μm. All experiments were conducted with three biological repetitions.

Interestingly, MoPhzF interacted with both MoRgs1 and MoEmc2 on the PM (Fig. 3b, 4b). Thus, we hypothesized that MoPhzF might function as a scaffold protein to recruit MoRgs1 and MoEmc2 to PM for MoRgs1 phosphorylation. To test this hypothesis, we examined the interaction between MoRgs1 and MoEmc2 by co-IP, and found that the MoRgs1-MoEmc2 interaction was abolished in the ΔMophzf mutant (Fig. 4d). In addition, we examined the subcellular localization of MoRgs1 and MoEmc2 in Guy11 and the ΔMophzf mutant during the germ tube hooking stage, and the results showed that MoRgs1 and MoEmc2 PM localization is disrupted in ΔMophzf, but not in Guy11 (Fig. 4e,f). Collectively, these findings demonstrated that MoPhzF is involved in MoRgs1 phosphorylation through the recruitment of MoRgs1 and MoEmc2 to PM.

Finally, we have divided MoPhzF into two regions based on its sequence, an N-terminal domain (NTD) and a C-terminal domain (CTD), and tested their interactions with MoRgs1 and MoEmc2 by Y2H, which showed that CTD, but not NTD, interacts with both MoEmc2 and MoRgs1 (Fig. S7).

MoPhzF-mediated PCA production is required for invasive hyphae growth and full virulence of M. oryzae

To examine the specific role of MoPhzF in virulence, conidial suspensions (5×10⁴ spores/mL) of Guy11, the ΔMophzf mutant, and the complemented strain (ΔMophzf/MoPHZF) were sprayed onto 2-week-old CO39 rice seedlings. The ΔMophzf mutant caused fewer and smaller lesions compared with Guy11 and ΔMophzf/MoPHZF 7 days after inoculation, with 17.66%, 1.33%, and 17.86% in Guy11, ΔMophzf, and ΔMophzf/MoPHZF, respectively (Fig. 5a,b). We further examined the penetration and invasive hyphal extension in rice sheath cells. After incubation with conidia suspension for 36 h, Guy11 and ΔMophzf/MoPHZF showed about 80% of type 3 (invasive hyphal extended but limited in one cell) and type 4 (invasive hyphae extended to surrounding cells) infectious hyphae (Fig. 5c). By contrast, more than 60% of penetration sites showed type 1 (no penetration of appressoria) and type 2 (only with a penetration peg or a primary invasion hypha) in the ΔMophzf mutant (Fig. 5c).

Fig. 5. MoPhzF-mediated PCA production is required for the invasive hyphae growth and full virulence of M. oryzae.

(a) Two-week-old rice (Oryza sativa cv. CO39) seedlings were sprayed with conidial suspensions (5×10⁴ spores/mL) of Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF following 5 μg/L PCA treatment and photographed 7 days post-inoculation. DMSO treatment was used as a negative control. (b) Diseased leaf areas were assessed using ImageJ. Data were represented by means ± standard deviations, and columns marked with different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). (c) Rice leaf sheaths of 3-week-old rice seedlings were inoculated with conidial suspensions (2×10⁵ spores/ml) of Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF following 5 μg/L PCA treatment. Detailed observation and statistical analysis for infectious growth in rice sheath cells were made at 36 h post-inoculation. Appressorium penetration sites (n = 100) were observed, and invasive hyphae were rated from type 1 to 4 (type1, no penetration with only appressoria; type 2, only with a penetration peg or primary invasion hyphal; type 3, secondary invasive hypha extended but was limited in one plant cell; type 4, invasive hyphae extended to surrounding cells). Data were represented by means ± standard deviations. Bar represents 20 μm. PCA, phenazine-1-carboxylic acid. All experiments were conducted with three biological repetitions.

To test whether MoPhzF-mediated PCA production is associated with virulence, rice spray and invasive growth assays were performed. We found that the ΔMophzf/MoPHZF^E77D strain showed pin-sized specks and defects in the invasive hyphae growth, similar to those produced by the ΔMophzf mutant (Fig. 5). Moreover, PCA was added to conidial suspensions at a final concentration of 5 μg/L and sprayed onto rice seedlings to evaluate the effect of PCA on virulence further. PCA treatment partially rescued the virulence of ΔMophzf and ΔMophzf/MoPHZF^E77D (Fig. 5a). Consistent with this result, the percentage of diseased lesion area was dramatically increased to 9.64% and 11.14% in ΔMophzf and ΔMophzf/MoPHZF^E77D after treatment compared with 1.33% and 3.68% untreated (Fig. 5b). In contrast, there was no significant differences in lesion areas by Guy11 and ΔMophzf/MoPHZF between PCA-treated and untreated (Fig. 5b). Meanwhile, PCA treatment also promoted the percentages of type 3 and type 4 invasive hyphae from 40% to 60% in ΔMophzf and ΔMophzf/MoPHZF^E77D (Fig. 5c). Taken together, these results suggested that MoPhzF-mediated PCA production contributes to the invasive hyphae growth and full virulence of M. oryzae.

Previous studies indicated that appressorium-mediated plant infection requires the timely assembly of a higher-order septin ring structure at the base of the appressorium in M. oryzae (Dagdas et al., 2012). Considering that ΔMophzf exhibits defects in appressorium formation and virulence, we hypothesized that MoPhzF might be involved in septin-dependent host infection. Given that four septin GTPases (Sep3, Sep4, Sep5, and Sep6) are known to form a heteromeric septin ring at the appressorium pore (Dagdas et al., 2012), septin ring organization was visualized through expressing MoSep3-GFP in Guy11 and ΔMophzf, respectively. As shown in Fig. S8, the appressorium normally formed a large septin ring in Guy11. However, there was about 40% of appressoria formed disorganized masses in the ΔMophzf mutant (Fig. S8). To determine whether MoPhzF-mediated PCA production has a role in septin-dependent plant infection, PCA was applied to conidial suspensions at a final concentration of 5 μg/L. The result showed that there were no differences between PCA-treated and untreated ΔMophzf (Fig. S8). Collectively, these findings indicated that MoPhzF is also involved in septin ring formation, which is independent of PCA production.

MoPhzF-mediated PCA production suppresses host ROS accumulation

During its interaction with host cells, M. oryzae needs to suppress host immunity, including ROS generation, to promote its infection (Guo et al., 2011; Kou et al., 2019; Liu et al., 2021). As the ΔMophzf mutant showed reduced virulence in rice, we tested whether MoPhzF is involved in host ROS suppression by quantifying ROS production using DAB and CM-H₂DCFDA staining in infected rice sheaths at 36 h. We found ΔMophzf infection resulted in significantly more staining than Guy11 and ΔMophzf/MoPHZF (Fig. 6a–c). Consistently, ΔMophzf infection drastically increased the expression of OsRBOHD, marker genes for ROS production (Fig. 6d). These data demonstrated that MoPhzF suppresses host ROS accumulation.

Fig. 6. MoPhzF-mediated PCA production is involved in suppressing ROS accumulation in the host.

(a) DAB was used to stain leaf sheaths injected with conidia suspensions of Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF following 5 μg/L PCA treatment at 36 h post-inoculation. The DMSO solvent treatment was used as a control. Bar represents 20 μm. (b) Statistical analysis of infected rice cells stained by DAB. Error bars represent standard deviations, and different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). (c) CM-H₂DCFDA staining on infected leaf sheaths by Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF following 5 μg/L PCA treatment at 36 h post-inoculation. Percentages of leaf sheath cells stained by CM-H₂DCFDA were shown in the image. Data were represented by means ± standard deviations, and different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). Bar represents 20 μm. (d) The transcriptional level of OsRBOHD in the infected rice was assayed using RT-qPCR. RNA samples were collected from rice leaves sprayed with conidial suspensions (2×10⁵ spores/mL) of Guy11, ΔMophzf, ΔMophzf/MoPHZF^E77D, and ΔMophzf/MoPHZF following 5 μg/L PCA treatment at 48 h post-inoculation. RT-qPCR was used to evaluate gene expression, with OsACTIN as the internal reference gene. Data were represented by means ± standard deviations, and columns marked with different letters indicate significant differences (Duncan’s new multiple range test, p < 0.05). PCA, phenazine-1-carboxylic acid. All experiments were conducted with three biological repetitions.

To test whether MoPhzF-mediated PCA production plays a role in host ROS suppression, ROS accumulation and OsRBOHD expression levels in rice cells infected by ΔMophzf/MoPHZF^E77D with or without PCA treatment were examined. We found that 54% of rice cells were stained by DAB, and 79% of rice cells were stained by CM-H₂DCFDA after ΔMophzf/MoPHZF^E77D infection (Fig. 6a–c). A similar trend was also observed after ΔMophzf infection (Fig. 6a–c). Meanwhile, the transcript level of OsRBOHD in ΔMophzf/MoPHZF^E77D infected cells was higher than that by Guy11 or ΔMophzf/MoPHZF. Still there was no difference between ΔMophzf/MoPHZF^E77D and ΔMophzf (Fig. 6d). PCA treatment dramatically reduced the percentage of infected cells stained by DAB and CM-H₂DCFDA after ΔMophzf and ΔMophzf/MoPHZF^E77D infection (Fig. 6a–c). The high expression of OsRBOHD in ΔMophzf and ΔMophzf/MoPHZF^E77D was significantly down-regulated following PCA treatment (Fig. 6d). However, the expression levels of OsBAK1, as well as genes associated with antioxidant production (OsVTC1, OsGS, OsGPP, OsGME1), in ΔMophzf/MoPHZF^E77D-infected cells were similar to those observed in Guy11 infection (Fig. S9). Our findings demonstrated that MoPhzF-mediated PCA production suppresses the host ROS accumulation.

Given that ROS is also important for signaling and septin-mediated cytoskeletal reorganization (Liu et al., 2020), ROS accumulation in Guy11 and ΔMophzf during appressorium formation in the presence or absence of PCA (5 μg/L) were analyzed using NBT staining. Dark blue precipitates were observed at the tips of germ tubes and early appressoria in both Guy11 and ΔMophzf (Fig. S10a,b). Moreover, there was no difference in the distribution of the NBT precipitate in Guy11 between PCA-treated and untreated (Fig. S10a,b). MoNOX2 expression was also examined using RT-qPCR analysis, and the result showed that there was no significant difference in Guy11 between PCA-treated and untreated (Fig. S10c). These data indicated that MoPhzF is not involved in the ROS accumulation of M. oryzae.

PCA targets proteins involved in host plant photosynthesis

To further explore the mechanism of how PCA suppresses host ROS accumulation, we used LiP-small molecule mapping (LiP-SMap) of the chemical proteomics approach to identify rice host proteins interacting with PCA. Significant changes in the abundance of half-tryptic peptides (P-value < 0.05 and fold change ≥ 1.5 or ≤ 0.67) were found, and a total of 259 differentially expressed proteins were identified (Table S3). KEGG pathway analysis showed that putative interaction proteins were highly enriched in pathways related to photosynthesis (Fig. S11), suggesting that PCA might target these groups of important host proteins.

Discussion

Previous studies demonstrated that PCA produced by rice phyllosphere bacteria and fungi plays an important role in virulence and defense against their competitors (Thanabalasingam et al., 2015; Costa et al., 2018; Zhu et al., 2018; Cimmino et al., 2021; Yue et al., 2021; Serafim et al., 2023). We extended antifungal tests to assess sensitivities of various pathogens to PCA and found that they were sensitive or resistant to PCA (Table S4). Recent studies showed that the rice blast fungus Magnaporthe oryzae is more resistant to PCA than several fungal pathogens (Yu et al., 2018; Xiong et al., 2019; Zhu et al., 2019; Li et al., 2021). However, whether M. oryzae produces PCA and the biological function of PCA remain unknown. In the present study, we not only detected PCA from extracts of M. oryzae by LC-MS-MS, but also characterized the function of the PCA biosynthesis gene MoPHZF. In addition to providing evidence suggesting that MoPhzF was acquired through HGT, we found that MoPhzF is essential for PCA synthesis and PCA suppresses host ROS accumulation to promote M. oryzae infection. In addition to this canonical function, MoPhzF has evolved a noncanonical signaling function to mediate MoEmc2 and MoRgs1 membrane recruitment and MoRgs1 phosphorylation (Fig. 7).

Fig. 7. A proposed model depicting an ancient event of PhzF acquisition and functions of MoPhzF in regulating the development, appressorium formation, and pathogenicity of Magnaporthe oryzae.

M. oryzae acquires the phenazine biosynthesis gene PHZF likely through an ancient horizontal gene transfer (HGT) event from bacteria, and MoPhzF plays dual roles in M. oryzae. M. oryzae-secreted phenazine-1-carboxylic acid (PCA) overcomes host immunity during infection by suppressing host ROS accumulation. At the same time, MoPhzF mediates MoRgs1 and MoEmc2 recruitment to the plasma membrane (PM), where MoRgs1 undergoes phosphorylation to regulate G-protein/cAMP signal transduction. GPCR: G-protein coupled receptor. Gα, Gβ, and Gγ: heterotrimeric G-protein α, β, and γ subunit, respectively.

Previous studies have suggested that Candida parapsilosis, Schizosaccharomyces pombe, and Nesidiocoris tenuis all acquired PhzF homologs from bacteria via HGT (Fitzpatrick et al., 2008; Ferguson et al., 2021). HGT is a natural biological process that facilitates the exchange of genetic material between distantly related lineages, which allows organisms to acquire novel traits and exploit unexplored resources (Bublitz et al., 2019; Li et al., 2022). For instance, Thinopyrum elongatum gained Fusarium head blight resistance gene Fhb7 through HGT from an endophytic Epichloë species, which confers broad resistance to Fusarium species by detoxifying trichothecenes (Wang et al., 2020). The whitefly Bemisia tabaci has acquired a phenolic glucoside malonyltransferase gene BtPMaT1 through a plant-to-insect HGT event, which allows whiteflies to detoxify plant defense compounds phenolic glucosides (Xia et al., 2021). In light of these findings, our investigation suggesting the possibility of PHZF HGT among ascomycete fungi is likely valid. Significantly, our studies revealed that MoPhzF remained the canonical function in PCA synthesis. In bacteria, PCA synthesis is mediated by a group of highly conserved and clustered PHZ genes (Mavrodi et al., 2010; Yu et al., 2018). Except for PhzA and PhzB, all proteins encoded by the seven-gene phenazine operon are absolutely essential for converting chorismic acid to PCA (McDonald et al., 2001; Ahuja et al., 2004; Mavrodi et al., 2004). Previous work showed that in Pst DC3000 phz homologs consist of phzC/D/E/F/G, and deletion of phzF blocked PCA production (Wen et al., 2016). However, Xanthomonas oryzae pv. oryzae (Xoo) PXO99^A could not produce PCA, although it contains phzC/E1/E2/F/G (Wen et al., 2016). In the pathogenic fungus M. oryzae, PHZ genes are composed of PHZC/D/E/F/G (Table S5) that are scattered in the genome characteristics of eukaryotic organisms, similar to PHZ genes in Pst DC3000. Phylogenetic trees for the other MoPHZ genes (MoPHZC/D/E/G) were also constructed and the results showed that none of them was likely acquired through an ancient horizontal gene transfer event (Fig. S12–S15). In addition, we also generated the deletion mutant of MoPhzG (MGG_01535), and the result showed that the deletion of MoPHZG caused defects in the PCA generation and pathogenicity of M. oryzae (Fig. S16). Therefore, we considered that all of the PhzC/D/E/F/G may be required for PCA production in M. oryzae, which is similar to bacteria.

We found that MoPhzF-mediated PCA production is independent of MoPhzF function in vegetative growth, conidiation, appressorium formation, intracellular cAMP level regulation, and septin ring formation. Previous studies indicated that the regulator of G-protein signaling protein MoRgs1 functions as a GTPase-activating protein for Gα subunit MoMagA upon phosphorylation by the casein kinase 2 MoCk2 and MoEmc2, thereby governing the G-protein/cAMP signaling required for appressorium formation (Zhang et al., 2011; Yu et al., 2021). Intriguingly, we found that MoPhzF interacts with MoRgs1 and has a positive role in regulating MoRgs1 phosphorylation, consistent with the elevated intracellular cAMP levels in the ΔMophzf mutant (Fig. S4f). Moreover, MoPhzF is required for not only the interaction between MoEmc2 and MoRgs1 but also their subcellular localization in PM during the germ tube hooking stage, whereas MoRgs1 and MoEmc2 had no effects on the localization of MoPhzF (Fig. 4, S17). We have previously also reported that the regulator of G-protein signaling protein MoRgs7 serves as a GPCR-like receptor to detect environmental hydrophobic cues (Zhang et al., 2011; Li et al., 2019). Upon sensing a hydrophobic surface, MoRgs7 undergoes endocytosis to internalize environmental cues and activate cAMP signaling, leading to appressorium formation (Li et al., 2019; Xu et al., 2023). We here found that MoPhzF has a role in recruiting MoRgs1 and MoEmc2 to PM for MoRgs1 phosphorylation, suggesting noncanonical functions of other proteins might also serve as protein chaperones to mediate signal transduction pathways important in the appressorium formation and host infection of M. oryzae. In addition, we also noticed that there was no PX domain or trans-membrane domain in MoPhzF. However, Gene Ontology analysis demonstrated that several MoPhzF-interactors are involved in the membrane organization process (Fig. S18). Thus, we hypothesized that MoPhzF’s presence at the plasma membrane is facilitated through its interaction with those plasma membrane proteins, akin to the mechanism observed in SiPTl1s, as demonstrated in the study (Huangfu et al., 2021). The ability of MoPhzF to associate with plasma membrane proteins suggests a potential role as a scaffold protein. This putative scaffold function could involve facilitating the recruitment of interacting partners, such as MoEmc2 and MoRgs1, to the plasma membrane despite the absence of traditional membrane-targeting motifs in MoPhzF.

Previous studies in P. syringae pv. tomato DC3000 showed that the deletion of phzF not only impedes PCA production but also significantly reduces pathogenicity in tomato plants (Wen et al., 2016). We also found that the ΔMophzf mutant exhibited defects in PCA production and virulence (Fig. 2, 5). To test whether the function of PCA is concentration-dependent, an overexpression line of MoPHZF which exhibited strong fluorescence and elevated mRNA levels (Fig. S19a,b) was generated by introducing MoPhzF-GFP with the strong constitutively activated ribosomal protein 27 (RP27) promoter into the protoplasts of Guy11. However, there were no significant differences in PCA production between Guy11 and the overexpression strain (Fig. S19c). In P. aeruginosa PA1201, PCA production was improved by blocking PCA conversion, increasing supplies of chorismate, and increasing the expression of phzA2-G2 and translation of phzA1-G1 (Jin et al., 2015). We also found that MoPHZC and MoPHZG expression in the overexpression strain were similar to those observed in Guy11 (Fig. S19b) Thus, the overexpression of MoPHZF might have no effect in improving PCA production. In addition, to address the potential impact of a fitness assay spanning multiple generations in rice, involving the ΔMophzf mutant with or without PCA-producing bacteria that infect rice, we have conducted an antifungal test in vitro to assess the sensitivity of ΔMophzf to PCA. The results revealed a significant suppression of mycelial growth by PCA, and the suppression was dose-dependent (Fig. S20).

As a fungal pathogen, M. oryzae subverts host immunity, including suppressing ROS accumulation, to adapt to the host environment and establish rapid colonization within the host tissue (Guo et al., 2011; Liu et al., 2020; Hu et al., 2022; Liu & Zhang, 2022). The ΔMophzf mutant was unable to suppress host ROS accumulation (Fig. 6). Thus, MoPhzF appears to play a role in infection and host colonization by suppressing host ROS accumulation, while MoPhzF is not involved in ROS accumulation (Fig. S10). Notably, it is different from previous studies that PCA results in intracellular ROS accumulation in Botrytis cinerea, Xanthomonas oryzae pv. oryzae, and Pseudopestalotiopsis camelliae-sinensis (Xu et al., 2015; Simionato et al., 2017; Yin et al., 2021), we found that the lower concentration of PCA suppressed host ROS accumulation. In revealing this new function of PCA in host-pathogen interactions, our chemical proteomic analysis revealed that host target proteins of PCA are enriched in the photosynthesis pathway (Fig. S11). Photosynthesis is a high-rate redox metabolic process, and rapid transients of photon capture, electron fluxes, and redox potentials during photosynthesis cause ROS to be released (Dietz et al., 2016; Foyer & Hanke, 2022). Given the established connections between photosynthesis and environmental stresses, such as salt stress, and pathogenic infections by X. oryzae pv. oryzae and Sinorhizobium meliloti 1021 (Vo et al., 2021; Xiong et al., 2021; Wang et al., 2022), our findings suggested that M. oryzae PCA may play a pivotal role in undermining host immunity by disrupting the normal functioning of these processes, particularly in ROS production. In addition, we suggested that MoPhzF likely suppresses host ROS accumulation by regulating OsRBOHD expression (Fig. 6). However, OsRbohD was not identified in the list of targets by LiP-small molecule mapping. We propose that PCA might act on these targets through a mechanism that could involve specific motifs or docking sites for PCA binding. Further investigations into the precise molecular interactions and motifs involved in PCA binding to these targets will be valuable for a more comprehensive understanding of PCA’s mode of action on differential gene expression.

Here, we unraveled the dual functional roles of MoPhzF in metabolic and signal transduction critical for appressorium formation and pathogenicity in M. oryzae. A previous study showed that MoAa91, an auxiliary activity family 9 protein (Aa9) homolog, also exhibits diverse dual functions. Specifically, MoAa91 not only acts as a novel signaling molecule to govern appressorium development on the artificial inductive surface but also functions as an effector to compete with a chitin elicitor-binding protein precursor (CEBiP) for chitin binding, thereby suppressing chitin-induced host immunity (Li et al., 2020) This example reinforces the notion that proteins, including MoPhzF, can exhibit diverse roles beyond what may be traditionally associated with their primary functions.

Supplementary Material

Supinfo

Fig. S1. Amino acid sequence alignments of phenazine biosynthesis proteins from Magnaporthe oryzae and Pseudomonas aeruginosa.

Fig. S2. Phylogenetic alignments of phenazine biosynthesis PhzF proteins of fungal and bacterial origins.

Fig. S3. Southern blot analysis of the MoPHZF gene knockout mutant.

Fig. S4. Contribution of MoPhzF to the development of M. oryzae is independent of PCA production.

Fig. S5. Activated MoRgs1 phosphorylation can partly restore vegetative growth and conidia production deficiency in the ΔMophzf mutant.

Fig. S6. MoPhzF does not interact with MoCk2.

Fig. S7. The C-terminus of MoPhzF interacts with MoRgs1 and MoEmc2.

Fig. S8. The function of MoPhzF in regulating septin ring is independent of PCA production.

Fig. S9. The transcriptional levels of OsBAK1 and genes associated with antioxidant production (OsGS, OsVTC1, OsGPP, OsGME1) in infected rice.

Fig. S10. MoPhzF is not involved in ROS accumulation of M. oryzae.

Fig. S11. The KEGG analysis for differentially expressed proteins of phenazine-1-carboxylic acid treatment group vs the control group.

Fig. S12 Phylogenetic alignments of phenazine biosynthesis protein PhzC of fungal and bacterial origins.

Fig. S13. Phylogenetic alignments of phenazine biosynthesis protein PhzD of fungal and bacterial origins.

Fig. S14. Phylogenetic alignments of phenazine biosynthesis protein PhzE of fungal and bacterial origins.

Fig. S15. Phylogenetic alignments of phenazine biosynthesis protein PhzG of fungal and bacterial origins.

Fig. S16. MoPhzG is required for PCA production and full virulence of M. oryzae.

Fig. S17. Subcellular localization of MoPhzF during the germ tube hooking stage.

Fig. S18. Gene Ontology analysis of MoPhzF-interactors based on cellular component and biological process.

Fig. S19. The fluorescence of MoPhzF-GFP, transcriptional levels of MoPHZ genes, and the content of PCA in an overexpression line of MoPHZF.

Fig. S20. PCA inhibits mycelial growth of the ΔMophzf mutant.

Table S1. Primers used in this study.

Table S2. Phenotype analysis of Guy11 and ΔMophzf following 5 μg/L PCA treatment.

Table S3. Differentially expressed proteins of phenazine-1-carboxylic acid treatment group vs the control group.

Table S4. Inhibitory efficacy of phenazine-1-carboxylic acid against various pathogens.

Table S5. Phz proteins identified from the M. oryzae genome.

Data availability

The data that supports the findings of this study are available in the Supporting Information of this article.