Autophagy and cell wall integrity pathways coordinately regulate the development and pathogenicity through MoAtg4 phosphorylation in Magnaporthe oryzae

Autophagy and Cell wall integrity (CWI) signaling are critical stress-responsive processes during fungal infection of host plants. In the rice blast fungus Magnaporthe oryzae, autophagy-related (ATG) proteins phosphorylate CWI kinases to regulate virulence; however, how autophagy interplays with CWI signaling to coordinate such regulation remains unknown. Here, we have identified the phosphorylation of ATG protein MoAtg4 as an important process in the coordination between autophagy and CWI in M. oryzae. The ATG kinase MoAtg1 phosphorylates MoAtg4 to inhibit the deconjugation and recycling of the key ATG protein MoAtg8. At the same time, MoMkk1, a core kinase of CWI, also phosphorylates MoAtg4 to attenuate the C-terminal cleavage of MoAtg8. Significantly, these two phosphorylation events maintain proper autophagy levels to coordinate the development and pathogenicity of the rice blast fungus.


Introduction
Autophagy is a highly conserved catabolic process that mediates intracellular component degradation through the autophagosomes to maintain cellular homeostasis in response to cellular and environmental stresses [1,2], which can be categorized into nonselective and selective autophagy based on its selectivity for substrates [3].Nonselective autophagy is critical for pathogenicity, whereas selective autophagy (e.g., pexophagy and mitophagy) is dispensable for appressorium-mediated host infection in the rice blast fungus Magnaporthe oryzae [4][5][6].In the Baker's yeast Saccharomyces cerevisiae, the autophagosome formation correlates well with the amount of autophagy-related (ATG) protein 8 (Atg8)-phosphatidylethanolamine (PE), the lipidation form of Atg8 [7][8][9].At the same time, the cysteine protease Atg4 cleaves the C-terminus of the newly synthesized cytoplasmic Atg8 to produce an active variant that undergoes lipidation modification and further conjugates PE of the phagophore assembly site (PAS) [10,11].Once autophagosome forms at PAS, Atg4 deconjugates Atg8 from the Atg8-PE anchor on the autophagosome membrane, a critical process for Atg8 recycling and subsequent autophagosome fusion with vacuole [12,13].Studies in S. cerevisiae also found that Atg4-mediated deconjugation of Atg8-PE is inhibited by ULK/ULK protein kinase Atg1-dependent phosphorylation, ensuring Atg8 anchoring until autophagosome formation [1].
The cell wall integrity (CWI) signaling pathway is critical for maintaining strong but elastic cell wall during growth and development and also in responses to the external stress [14][15][16].In S. cerevisiae, CWI signals stimulate the small G protein Rho1 that binds to the protein kinase C (Pkc1) to activate the conserved mitogen-activated protein kinase (MAPK) cascade, composed of Bck1, Mkk1/2, and Slt2/Mpk1, which then targets the transcription factors Rlm1 and the SBF complex (Swi4 and Swi6) to govern the expression of cell wall biosynthesis and cell cycle genes, respectively [17][18][19].Many plant pathogenic fungi, including M. oryzae, Fusarium graminearum, and Claviceps purpurea, utilize this conserved signaling pathway to balance the cell wall homeostasis, which is critical for development and pathogenicity [20][21][22][23].In M. oryzae, deletions of the CWI MAPK cascade components MoMCK1, MoMKK1, or MoMPS1 lead to abnormal chitin distribution and virulence attenuation [22,24,25].Recent studies have also shown that deletions of these kinases result in aberrant autophagy by unknown mechanisms [16].
As the critical processes in response to environmental stresses, connections between autophagy and CWI signaling have been previously examined in S. cerevisiae and M. oryzae [16,[26][27][28].In S. cerevisiae, Slt2 promotes mitophagy by affecting mitochondrial recruitment to the PAS, and also regulates pexophagy without affecting nonselective autophagy [26][27][28].In M. oryzae, endoplasmic reticulum (ER) stress induced by abnormal protein synthesis activates MoAtg1-dependent MoMkk1 phosphorylation to enhance CWI signaling, thus promoting infection [16].At the same time, CWI signaling addresses cell wall stress by targeting protein production, implying a possible important relationship between the CWI signaling pathway and nonselective autophagy (hereafter "autophagy").However, the underlying mechanism of how autophagy coordinats with CWI signaling to regulate development and pathogenicity is unknown.Here, we revealed that two phosphorylation events of MoAtg4, MoMkk1-mediated phosphorylation in the cytoplasm and MoAtg1-mediated phosphorylation at PAS, collectively maintain proper autophagy and cellular homeostasis.

Cell wall stress-induced autophagy is dependent on CWI kinases in M. oryzae
To explore the coordination between autophagy and the CWI signaling pathway, we treated the wild-type strain Guy11 and CWI kinase mutant stains ΔMomck1, ΔMomkk1, and ΔMomps1 with cell wall stressor Calcofluor White (CFW) and performed autophagic body (AB) staining using monodansylcadaverine (MDC) [29][30][31].Vacuoles with AB were significantly increased after 5 h of treatment in Guy11 (~65%), but not in the mutants (Fig 1A and  1B).In addition, the expression of the autophagic marker protein RFP-MoAtg8 in these strains revealed that vacuole-localized RFP was significantly increased in Guy11, but not in the mutants (Fig 1C).This result was confirmed by Western blotting analysis that showed an elevated ratio of free RFP in Guy11 in comparison to the mutants (Fig 1D).We also used sodium dodecyl sulfate (SDS) as a second stressor and observed a similar result (S1A-S1C Fig) .In addition, the degradation of MoSec63-GFP, an ER-phagy marker of M. oryzae [32], was

PLOS PATHOGENS
detected under CFW treatment.These results showed that cell wall stress did not induce ERphagy (S1D Fig) .Together, these results indicate that cell wall stress-induced autophagy is dependent on CWI kinases in M. oryzae.

CWI kinase MoMkk1 interacts with and phosphorylates MoAtg4 in the cytoplasm
Previous studies identified 22 ATG proteins that are essential for autophagy in M. oryzae [6], among which several were found to interact with MoMkk1 via yeast-two-hybrid (Y2H) screening, including MoAtg3, 4, 5, and 16 that are directly involved in Atg8 lipidation [16].In addition, MoMkk1 mediates the crosstalk between the CWI signaling pathway and autophagy under ER stress [16].We, therefore, hypothesized that autophagy coordinates CWI signaling through direct phosphorylation of ATG proteins by MoMkk1.To test this hypothesis, in vitro phosphorylation assays were performed using phosphoprotein gel staining [16,33,34], and the results showed significantly increased phospho-fluorescence only between MoMkk1 and MoAtg4, but not MoAtg3, 5, or 16 (Fig 2A).This result suggested that MoMkk1 can phosphorylate MoAtg4.Consequently, we focused on the interaction between MoMkk1 and MoAtg4.
We next examined the localizations of MoMkk1 and MoAtg4 in M. oryzae.Both were localized to the cytoplasm with or without cell wall stress, though MoAtg4 was also accumulated at PAS that was marked with MoApe1-RFP (S2A and S2B Fig) .CFW treatment did not affect the numbers of cells with PAS-localized MoAtg4 (S2C Fig) .These results implied a direct interaction between MoMkk1 and MoAtg4.To test the physical interaction between the two, GST pull-down and co-immunoprecipitation (co-IP) assays were performed that showed positive interaction (Fig 2B and 2C).Furthermore, a bimolecular fluorescence complementation (BiFC) assay confirmed that the interaction occurs in the cytoplasm, and there was no obvious co-localization with the MDC dye (Figs 2D and S2D).This result indicated that MoMkk1 interacts with MoAtg4 in the cytoplasm.

MoMkk1-dependent MoAtg4 phosphorylation plays a role in the development and virulence of M. oryzae
To investigate the biological function of MoMkk1-dependent MoAtg4 phosphorylation, we replaced S97, S276, and T278 residues with aspartic acid (D) to mimic phosphorylation, and introduced nonphosphorylated MoATG4 3A -GFP and phosphorylation-mimic MoATG4 3D -GFP vectors into the ΔMoatg4 mutant, respectively.We found that ΔMoatg4/MoATG4 3A and ΔMoatg4/MoATG4 3D strains were defective in conidiation and appressorium formation.In addition, the phosphorylation-mimic strain showed more severe defects than nonphosphorylated ones (Figs 3A, S3A and S3B).The ΔMoatg4/MoATG4 3A strain showed a similar virulence as Guy11, in contrast to the attenuated one in the ΔMoatg4/MoATG4 3D strain (Fig 3B -3D).The invasive hyphae (IH) growth assay showed that about 30% of sites were type 4 in infection by Guy11, ΔMoatg4/MoATG4, and ΔMoatg4/MoATG4 3A , whereas only approximately 5% were type 4 in infection by the ΔMoatg4/MoATG4 3D strain, and none by ΔMoatg4 (Fig 3E).Collectively, these results indicated that MoMkk1-dependent MoAtg4 phosphorylation plays a role in the development and pathogenicity of M. oryzae.
To analyze the impact of cell wall stress on ΔMoatg4/MoATG4 3A and ΔMoatg4/MoATG4 3D strains, we investigated the growth, conidiation, and pathogenicity of the strains under CFW treatment.Although CFW inhibited 25% of the growth rate of ΔMoatg4/MoATG4 3A and 21% of ΔMoatg4/MoATG4 3D strains, it showed no difference compared to Guy11, and inhibitions of the strains' virulence by CFW treatment were also similar.In addition, CFW treatment did not impact conidiation (S4A-S4C Fig) .Since the CWI signaling pathway requires proper MAPK cascade phosphorylation and signal transduction [22], we examined the effect of MoMkk1-dependent MoAtg4 phosphorylation on CWI signaling through CFW staining.Chitin distribution of ΔMoatg4/MoATG4 3A and ΔMoatg4/MoATG4 3D strains is even and concentrated on the growing apices, and they also did not exhibit hyphal autolysis (S4D-S4F

MoMkk1-dependent MoAtg4 phosphorylation negatively controls autophagy by inhibiting MoAtg8 C-terminal cleavage
MoAtg4 was shown to regulate development and pathogenicity through autophagy function in M. oryzae [5,6].To test whether MoMkk1-dependent MoAtg4 phosphorylation fulfills MoAtg4-Flag were co-expressed in the Guy11 strain, and proteins were extracted for co-IP assay by Western blot analysis using anti-GFP and anti-Flag antibodies.Proteins from Guy11 co-expressing GFP and MoAtg4-Flag were used as the control.(D) BiFC observation in the strain co-expressing MoMkk1-nYFP and MoAtg4-cYFP during various developmental stages.The strains co-expressing empty nYFP with MoAtg4-cYFP and MoMkk1-nYFP with empty cYFP were used as controls.Scale bar: 5 μm.(E) Peptides of MoMkk1-dependent MoAtg4 phosphorylation were identified by LC-MS/MS analysis under CFW treatment.(F) GST-MoMkk1, His-MoAtg4, His-MoAtg4 S97A , His-MoAtg4 S276A , His-MoAtg4 T278A , and His-MoAtg4 3A fusion proteins were obtained for in vitro phosphorylation analysis.3A: replaced S97, S276, and T278 with A to mimic nonphosphorylated form of MoAtg4.Data from three independent experiments were used for statistical analysis by one-way ANOVA with Tukey's HSD.https://doi.org/10.1371/journal.ppat.1011988.g002biological function by modulating autophagy, we performed MDC staining.About 99% of vacuoles contained AB in Guy11, but only 10% in the ΔMoatg4 mutant.This autophagic defect was mostly restored by MoAtg4 3A , in which 75% of the vacuoles contained AB, and was slightly restored in the ΔMoatg4/MoATG4 3D strain (25%) (Fig 4A and 4B).In addition, transmission electron microscopy (TEM) showed that AB numbers per vacuole were significantly reduced in the ΔMoatg4/MoATG4 3D strain (Fig 4C and 4D).We also assessed the effect of MoMkk1-dependent MoAtg4 phosphorylation on cell wall stress-induced autophagy, and found that autophagy was significantly decreased in the ΔMoatg4/MoATG4 3D strain (S5A and S5B Fig) .These results demonstrated that MoMkk1-dependent MoAtg4 phosphorylation negatively regulates autophagy in M. oryzae.4E) is required for autophagy [5].We compared the MoAtg8 amino acid sequence with that of yeast Atg8 and found that MoAtg8 contains seven amino acid residues (7aa) following the active site glycine 116 (G116), instead of a single arginine residue in Atg8 (S5C Fig) .To test whether MoAtg4 phosphorylation regulates the C-terminal cleavage of MoAtg8, the recombinant His-MoAtg4, His-MoAtg4 3A , and His-MoAtg4 3D were incubated with GST-MoAtg8, respectively.The protease-dead MoAtg4 C206S , which causes a serious autophagic defect, was used as a negative control [1,5].The results showed that MoAtg4, but not MoAtg4 C206S , cleaves MoAtg8, and the nonphosphorylated MoAtg4 3A exhibited a similar cleavage to MoAtg4, while the cleavage in the phosphorylation-mimic MoAtg4 3D was reduced (Fig 4F and 4G).Moreover, an in vivo cleavage assay was performed in Guy11, ΔMoatg4, ΔMoatg4/MoATG4 3A , and ΔMoatg4/MoATG4 3D strains expressing MoAtg8-RFP, a MoAtg8 with a C-terminal RFP tag, under MM-N condition or CFW treatment.The bands of 7aa-RFP were detected in Guy11 and the ΔMoatg4/MoATG4 3A strains, while this cleavage was blocked in the ΔMoatg4 mutant and drastically impaired in the ΔMoatg4/MoATG4 3D strain (Figs 4H and S5D).
M. oryzae infection is dependent on appressoria that penetrate plant epidermis and appressorial development requires autophagy [2,31,35,36].We next observed the localization of MoAtg8-RFP in three-celled conidia and appressoria at the appressorium developmental stage.RFP fluorescence was mainly localized in vacuoles in conidium cells of Guy11 and ΔMoatg4/3A, while this localization pattern was not obvious in ΔMoatg4 and ΔMoatg4/3D in (C).(E) Schematic diagram of MoAtg8 processing mediated by MoAtg4.MoAtg4 cleaves the C-terminus (7aa) of the MoAtg8 precursor to produce an active variant, then deconjugates MoAtg8-PE after autophagosome formation.(F) His-MoAtg4, His-MoAtg4 3A , His-MoAtg4 3D , and His-MoAtg4 C206S (protease-dead) fusion proteins were respectively mixed with GST-MoAtg8.After incubation at 37˚C for 0, 1, or 2 h, in vitro, GST-MoAtg8 cleavage was analyzed by Western blot with anti-His and anti-GST antibodies.(G) Quantification of GST-MoAtg8 cleavage assays as shown in (E).(H) Extracted proteins from mycelia of Guy11, ΔMoatg4, ΔMoatg4/3A, and ΔMoatg4/3D strains expressing MoAtg8-RFP were treated with MM-N for 1 h and assayed for in vivo cleavage by Western blot using anti-RFP and anti-Actin antibodies.Conidia were inoculated to onion epidermis and proteins extracted at the appressorium developmental stage (8 hpi) and collected with anti-RFP agarose beads before analyzing MoAtg8 C-terminal cleavage.(I) pGBKT7 fused with MoATG4, MoATG4 3A , MoATG4 3D or MoATG4 C206S , and pGADT7 with MoATG8 were co-introduced into the yeast AH109 strain and grew on SD-Leu-Trp, and the transformants were transferred to SD-His-Leu-Trp, or SD-Ade-His-Leu-Trp.Three colony solutions diluted to different levels (10 0 to 10 −2 ) of each protein pair were inoculated on the same plate for binding analysis.(J) 10 mM GST-MoAtg8 was labeled with RED-NHS and mixed with His-MoAtg4, MoAtg4 3A , MoAtg4 3D , and MoAtg4 C206S at various ranges of concentrations.Recombinant proteins were collected in NT standard capillaries for MST analysis that determined binding affinities.The empty His protein was used as a control.(K) Proteins were extracted from ΔMoatg4/MoATG4, ΔMoatg4/3A, ΔMoatg4/3D, and ΔMoatg4/C206S strains expressing RFP-MoAtg8, which were treated with MM-N for 1 h, and co-IP assays were performed.Western blot analysis was performed using anti-RFP and anti-GFP antibodies.Proteins from ΔMoatg4 co-expressing GFP and RFP-MoAtg8 were used as a control.GST-G116: GST-MoAtg8 G116 .Different letters indicate statistically significant differences (Duncan's new multiple range test, p < 0.01). https://doi.org/10.1371/journal.ppat.1011988.g004

PLOS PATHOGENS
We next tested whether MoAtg4 phosphorylation affects its interaction with MoAtg8.The Y2H assay showed that, compared to MoAtg4, MoAtg4 3A and MoAtg4 C206S exhibit a similar binding, while MoAtg4 3D displays a significantly decreased interaction with MoAtg8 (Fig 4I).The microscale thermophoresis (MST) assay also showed that the interaction between MoAtg4 3D and MoAtg8 had a markedly higher dissociation constant (K d = 1.06 μM) than that of MoAtg4 with MoAtg8 (K d = 0.15 μM), MoAtg4 3A with MoAtg8 (K d = 0.10 μM), and MoAtg4 C206S with MoAtg8 (K d = 0.19 μM) (Fig 4J ), which verified that MoAtg4 3D weakens its interaction with MoAtg8.Co-IP analysis further revealed an attenuated interaction between MoAtg4 3D and MoAtg8 under MM-N condition or CFW treatment (Figs 4K and S5G).These results suggested that MoMkk1-dependent phosphorylation disturbs the interaction of MoAtg4 with MoAtg8.We also investigated whether the phosphorylation affects the interaction between MoMkk1 and MoAtg4 by Y2H, and found that the interaction between MoAtg4 3A or MoAtg4 3D  Although not covered in repeated LC-MS/MS data, S364, a conserved residue previously reported to be phosphorylated by Atg1 in S. cerevisiae, was mutated to evaluate the function in MoAtg1-mediated phosphorylation reaction (S6E Fig) .In vitro phosphorylation analysis showed that phospho-fluorescence was partially reduced in MoAtg4 S67A and MoAtg4 S364A , while MoAtg4 2A containing mutations of both S67A and S364A showed more significantly reduced phospho-fluorescence (S6C Fig) , indicating that S67 and S364 residues are important for MoAtg4 phosphorylation by MoAtg1.

MoAtg1-mediated MoAtg4 phosphorylation negatively regulates autophagy by inhibiting deconjugation of MoAtg8-PE
Since both MoAtg1 and MoAtg4 are essential for autophagy, we next investigated the effect of MoAtg1-mediated MoAtg4 phosphorylation on autophagy in M. oryzae.MDC staining showed that the number of vacuoles with AB was nearly 100% in Guy11, but reduced to 77% in ΔMoatg4/MoATG4 2A and 28% in ΔMoatg4/MoATG4 2D strains, respectively (S7A and S7B Fig) .Consistently, TEM showed defective autophagy in the aberrantly phosphorylated strains (S7C and S7D Fig) .These results indicated that MoAtg1-mediated MoAtg4 phosphorylation is critical for autophagy.
In yeast, Atg4 is responsible for Atg8-PE deconjugation and Atg8 recycling to promote the fusion of autophagosomes with vacuoles and flowing autophagosomes formation, which is inhibited by Atg1-mediated Atg4 phosphorylation [1,38].To assay whether the phosphorylation of MoAtg4 controls the deconjugation of MoAtg8-PE, we expressed MoAtg8 G116 with an N-terminal RFP tag (RFP-MoAtg8 G116 ), which could conjugate to PE without the initial step of MoAtg4-mediated cleavage [1,38], in the ΔMoatg4/MoATG4 2A and ΔMoatg4/MoATG4 2D strains.As indicated by vacuolar fluorescent dye CMAC [1,39], RFP signals were localized in the vacuoles of Guy11 and ΔMoatg4/MoATG4 2A strains but were restricted in the cytoplasm and the edge of vacuoles in ΔMoatg4 and ΔMoatg4/MoATG4 2D strains (S7E Fig) .Blocked deconjugation of MoAtg8-PE was also observed at the appressorium maturing stage in ΔMoatg4 and ΔMoatg4/MoATG4 2D strains.Western blotting analysis showed that protein levels of free RFP were lower in ΔMoatg4 and ΔMoatg4/MoATG4 2D (S7F Fig) .These findings suggest that the phosphorylation of MoAtg4 by MoAtg1 inhibits deconjugation of MoAtg8-PE.
To understand the underlying mechanism, we examined whether MoAtg4 phosphorylation by MoAtg1 affects its interaction with MoAtg8.Y2H results showed that the phosphorylationmimic MoAtg4 2D , but not the nonphosphorylated variant MoAtg4 2A , significantly reduced the affinity with MoAtg8 (S7G Fig) .Co-IP analysis was also performed with proteins extracted from strains expressing RFP-MoAtg8 G116 , and a weak interaction was detected between MoAtg4 2D and MoAtg8 (S7H

MoAtg4 phosphorylation mediated by MoMkk1 and MoAtg1 orchestrates autophagy and pathogenicity of M. oryzae
As MoMkk1 and MoAtg1 phosphorylate MoAtg4 at different amino acid residues to regulate the autophagy, development, and pathogenicity of M. oryzae, we examined the synergistic effect of the two phosphorylation events.All phosphorylation sites mediated by both kinases were mutated to A or D to construct the nonphosphorylated MoAtg4 5A or phosphorylationmimic MoAtg4 5D variant, respectively.Interestingly, MoAtg4 5A rescued defects of the ΔMoatg4 mutant in conidiation, appressorium formation, and pathogenicity, similar to MoAtg4 3A and MoAtg4 2A , while MoAtg4 5D did not show any restoring effect (S3A, S3B and S9A-S9E Figs).In short, MoMkk1 and MoAtg1 phosphorylate MoAtg4 to collectively regulate the development and pathogenicity of M. oryzae.
We next analyzed autophagic levels in ΔMoatg4/MoATG4 5A and ΔMoatg4/MoATG4 5D strains.Like ΔMoatg4, the ΔMoatg4/MoATG4 5D strain only had 10% vacuoles containing AB, with the AB number per vacuole significantly reduced compared to Guy11 (Fig 5A -5D).The in vitro and in vivo C-terminal cleavage assays of MoAtg8 showed that MoAtg4 5D could not cleave MoAtg8, similar to MoAtg4 C206S , and that the C-terminal cleavage did not occur at the mycelial or appressorial stage in the ΔMoatg4/MoATG4 5D strain (Fig 5E -5G).Interestingly, MoAtg4 5D could not interact with MoAtg8 (Fig 5H and 5I).These results indicated that MoAtg4 phosphorylation mediated by MoMkk1 and MoAtg1 collaborates to negatively regulate autophagy of M. oryzae.
In order to explore the relationship between MoMkk1-and MoAtg1-dependent MoAtg4 phosphorylation, we performed in vitro phosphorylation reactions of MoMkk1 with MoAtg4 2D and MoAtg1 with MoAtg4 3D .Interestingly, phosphorylation of MoMkk1 with MoAtg4 2D and MoAtg1 with MoAtg4 3D were significantly weaker than them with MoAtg4 (S9F Fig) , suggesting an antagonistic relationship between MoMkk1-and MoAtg1-dependent MoAtg4 phosphorylation.MoMkk1 S115D , the phosphorylated form by MoAtg1 [16], was also tested for phosphorylation with MoAtg4, and results showed that the phosphorylation level

Discussion
In eukaryotic cells, CWI signals are amplified by the MAPK pathway to maintain cell wall integrity, whereas autophagy is activated to degrade intracellular components for cellular homeostasis [15,40].Recently, ATG proteins in M. oryzae were found to enhance CWI signaling under ER stress generated by abnormal protein synthesis [16].In turn, the components of the CWI signaling pathway also affect autophagy.In the budding yeast, Slt2 regulates selective autophagy, but not nonselective autophagy, while CWI pathway components are important for nonselective autophagy in M. oryzae [16,[26][27][28].However, it is unknown whether CWI kinases regulate autophagy through ATG protein phosphorylation.Here, we found that MoMkk1 phosphorylates MoAtg4 to maintain proper autophagic levels under cell wall stress.It is worth noting that autophagy is a double-edged sword since either too low or too high levels are not conducive to normal survival and development [41,42].Although MoMkk1-dependent MoAtg4 phosphorylation negatively regulates autophagy, previous studies suggested that Pak1 (p21 [Rac1] activated kinase) phosphorylates Atg5 to promote human cell autophagy by strengthening its interaction with Atg16L [43]; therefore, the CWI signaling pathway may enhance autophagy through other factors under cell wall stress.Besides, our results indicate that cell wall stress does not induce ER-phagy, which is different from the CWI signaling pathway reported in yeast, highlighting a distinctive aspect of our investigation.
Our results further showed that the C-terminal cleavage of MoAtg8 is stronger in ΔMomck1 and ΔMomkk1, but remains unchanged in ΔMomps1 under cell wall stress.This is probably due to the absence of MoMkk1-dependent MoAtg4 phosphorylation, either through the blockage of cell wall stress transmission or deletion of kinase MoMkk1 in ΔMomck1 and ΔMomkk1 [22].In ΔMomps1, however, the normal level of MoAtg4 phosphorylation inhibits MoAtg8 cleavage and subsequent autophagy.Together, MoMkk1-dependent MoAtg4 phosphorylation mediates the coordination between autophagy and the CWI pathway in M. oryzae.
During M. oryzae infection of host plant, the fungus faces cell wall and nitrogen starvation stresses caused by cell wall remodeling and nutrient deficiency [6,14,31].How does the fungus overcome or respond to these conditions?Our studies indicated that MoMkk1 phosphorylates MoAtg4 under cell wall stress and that MoAtg1 phosphorylates MoAtg4 under nitrogen starvation stress, which are involved in autophagy regulation, thus affecting appressorium-mediated infection of host and pathogenicity.Therefore, our research demonstrated the importance of CWI kinase-dependent ATG protein phosphorylation in M. oryzae.
In yeast, the C-terminal cleavage of Atg8 and deconjugation of Atg8-PE mediated by Atg4 are important for autophagosome formation and autophagy [38,44].Atg1 phosphorylates Atg4 at PAS to enhance autophagy by protecting the Atg8-PE pool [1].We asked whether organisms employ Atg4 phosphorylation to accurately distinguish the two different processing of Atg8 and control autophagy.In this study, we showed that MoAtg1 phosphorylates MoAtg4 at S67 and S364 (the latter is same as yeast Atg4) to inhibit deconjugation of MoAtg8-PE similar to yeast, while MoMkk1-mediated MoAtg4 phosphorylation is responsible for the C-terminal cleavage of MoAtg8 in the cytoplasm.Our results suggested that the phosphorylation of MoAtg4 by different kinases regulates MoAtg8 processing in a delicate spatiotemporal manner.It has been reported that mammalian sterile20-like kinase Mst4, with the exception of Ulk1, phosphorylates Atg4B to promote autophagy [45].M. oryzae might be more similar to mammals than yeast in that there are additional MoAtg4 regulatory mechanisms.Furthermore, MoMkk1-and MoAtg1-mediated MoAtg4 phosphorylation inhibits the interaction of MoAtg4-MoAtg8.Interestingly, MoAtg4 C206S cannot cleave MoAtg8, but it interacts with MoAtg8 in a manner similar to MoAtg4.This suggests the phosphorylation of MoAtg4 may directly influence cleavage and modulate cleavage through altering MoAtg4-MoAtg8 interaction, ultimately regulating autophagy.
The deconjugation of Atg8-PE enhances autophagy by promoting the fusion of autophagosomes with vacuoles and also releasing and recycling Atg8 into the cytosol [12,46].Compared with de novo protein synthesis, recycling is a more efficient process.In M. oryzae, MoPtps protein phosphatases regulate MoOsm1, a Hog1 homolog, recycling into the cytoplasm through dephosphorylation to enable a rapid response to the external stress [47].Likewise, the late endosomes circulate the G-protein regulator MoRgs7 and the G-protein Gα subunit MoMagA to the plasma membrane to perceive the hydrophobic surface and promote cAMP signal transduction [48].Our studies revealed a similar recycling mechanism of MoAtg8 existed in M.
binding ability analysis of MoAtg4 5A and MoAtg4 5D with MoAtg8 by Y2H.(I) Co-IP assay to analyze the interaction between MoAtg4 5A or MoAtg4 5D and MoAtg8.5A/5D: All MoAtg4 phosphorylation sites (S97, S276, T278, S67, and S364) mediated by MoMkk1 and MoAtg1 were replaced by A/D to mimic nonphosphorylated/phosphorylated form of MoAtg4.GST-G116: GST-MoAtg8 G116 .Different letters indicate statistically significant differences (Duncan's new multiple range test, p < 0.05 or p < 0.01).https://doi.org/10.1371/journal.ppat.1011988.g005oryzae: MoAtg4 deconjugates and recycles MoAtg8 from the MoAtg8-PE anchor localized on the autophagosomal membrane to the cytoplasm.Of note, MoAtg1-mediated MoAtg4 phosphorylation may inhibit this process on immature autophagosomes.Although MoAtg8 recycling increases the MoAtg8 pool in the cytosol, which is available for subsequent autophagosome formation, earlier MoAtg8 recycling may prevent completion of current autophagosomes.Therefore, the inhibition of MoAtg8 deconjugation by MoAtg1-mediated phosphorylation may balance the MoAtg8 demand for autophagosome formation and recycling to modulate development and virulence.
One acceptor protein can be phosphorylated by multiple kinases is an efficient but complex regulation mechanism.MoMkk1 is a core detector that can be phosphorylated by MoMck1 and MoSep1, and as well as MoAtg1 under cell wall or ER stress [14,16].The mammalian Ulk1 is phosphorylated by the AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) to promote or inhibit its activity under glucose starvation or nutrient sufficiency conditions, respectively [49].How do cells sort out these phosphorylation relationships?Our study showed that MoMkk1-and MoAtg1-dependent MoAtg4 phosphorylation are contrary to each other, though it is difficult to quantify phosphorylation under cell wall and nutrition stress in vivo due to its complexity, as MoAtg4 phosphorylation is still detectable in the ΔMoatg1ΔMomkk1 double deletion mutant (S2E Fig) .MoAtg4 is localized in both the cytoplasm and at PAS, and MoMkk1 is mainly located in the cytoplasm.It has been reported that MoAtg1 is perceptible in the cytoplasm of M. oryzae [4], and yeast Atg1 gathers to PAS during autophagy [50,51], which probably causes more binding of MoAtg1 but not MoMkk1 and phosphorylates MoAtg4 at PAS.On the other hand, the antagonistic relationship of the two MoAtg4 phosphorylation events guarantees their collaborative but differential function.
In summary, MoAtg4 phosphorylation events mediated by MoAtg1 and MoMkk1 revealed novel crosstalk between autophagy and CWI signaling.

Strains and culture conditions
M. oryzae strain Guy11 was used as the wild-type strain in this study.Fusion constructs containing the MoAtg4 promoter, MoAtg4 (or sites-mutation variants) coding region, fragments of GFP, and bleomycin were transferred to ΔMoatg4 mutant to obtain complemented and phosphorylation site-mutations strains.All strains were cultured on CM at 28˚C in the dark.Mycelia blocks were cut from plates and cultivated in liquid CM for 2 d at 28˚C, and mycelia were harvested for DNA, RNA, and protein extraction.For the conidiation assay, strains were cultured on an SDC medium in the dark at 28˚C for 7 d, followed by 3 d illumination under fluorescent light before counting [52,53].Conidia were adjusted to a concentration of 1 × 10 5 / ml, and 30 μl of conidial solution was inoculated to hydrophobic slides.Conidial germination and appressorium formation were observed using a microscope following incubation in the dark for 2, 6, 10, and 24 h, respectively.

Virulence assay
For the virulence assay, conidia were harvested and resuspended to 5 × 10 4 /ml with 0.2% (w:v) gelatin.5 ml of solution was sprayed on two-week-old rice seedlings (Oryza sativa cultivar CO39).The inoculated plants were incubated in the dark at 28˚C and 90% humidity for 24 h, then transferred to 12 h light-dark environment for 6 d to assess leaf symptoms, which was quantified by measuring lesion area and quantitative PCR of M. oryzae genomic 28S rDNA relative to rice genomic RUBQ1 DNA [54].For IH growth analysis, 3-week-old rice sheaths were injected with conidial suspension (1 × 10 5 spores/ml) and cultured in the dark at 28˚C with high humidity for 36 h [55].

Yeast-two-hybrid (Y2H) assay
The coding sequences of corresponding genes were amplified and cloned into pGBKT7 or pGADT7 vectors, which were then introduced into the yeast AH109 strain after verification.Transformants grown on synthetic dextrose medium lacking tryptophan and leucine (SD-Leu-Trp) and were transferred to the medium lacking tryptophan, leucine, and histidine (SD-His-Leu-Trp) or tryptophan, leucine, histidine, and adenine (SD-Ade-His-Leu-Trp) [47].

GST pull-down assay
GST, GST-MoMkk1, and His-MoAtg4 were expressed in Escherichia coli BL21 cells.Cells were collected and washed 3 times with 1× PBS, then resuspended and lysed with a sonicator.After centrifugation with 13,000 g for 10 min, the GST or GST-MoMkk1 supernatant was mixed with 30 μl of glutathionesepharose beads (GE Healthcare, 10265165), which were washed 3 times using GST washing buffer (250 mM NaCl, 20 mM Tris, 2 mM EDTA, 2 mM EGTA, pH 7.5), then incubated at 4˚C for 2 h.The supernatant was discarded after centrifugation (13,000 g, 10 min) and the GST-or GST-MoMkk1-bound to GST beads were incubated with His-MoAtg4 protein solution for another 4 h.The beads were washed 5 times with GST washing buffer.Proteins were eluted with GST elution buffer (10 mM reduced glutathione in GST washing buffer, pH 8.0), then analyzed by Western blot analysis with anti-His (Abmart, M20001) and anti-GST (Abmart, M20007) antibodies, respectively.

Bimolecular fluorescence complementation (BiFC) assay
nYFP-MoMkk1 and MoAtg4-cYFP were expressed in Guy11.Transformants were isolated from plates containing 150 μg/ml hygromycin and 200 μg/ml zeocin and confirmed by fluorescent microscopy.For the BiFC assay, YFP signals were examined with confocal microscopy during different stages of the transformants.Strains expressing empty nYFP with MoAtg4--cYFP and nYFP-MoMkk1 with empty cYFP were used as controls.

In vivo phosphorylation analysis and phosphorylation site identification
To analyze the phosphorylation of MoAtg4 in M. oryzae, the MoATG4-GFP fusion construct was transferred into ΔMoatg4, ΔMomkk1, ΔMoatg1, and ΔMoatg1ΔMomkk1 mutants, respectively.Mycelia powder was added into 1 ml of lysis buffer (50 mM Tris, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100, pH 7.4) with or without 10 μl of phosphatase inhibitor cocktail 3 (Sigma-Aldrich, P0044).Extracted proteins were then purified with anti-GFP beads.Alkaline phosphatase (Sigma-Aldrich, P6774) at a final concentration of 2.5 U/ml was added to half of the sample without cocktail 3 and incubated with 1 mM MgCl 2 for 2 h at 37˚C.These protein samples were separated on 8% SDS-PAGE gel containing 20 μM phos-tag (NARD institute Limited company, 18D01) and 100 μM MnCl 2 .After electrophoresis at 80 V for 5-8 h, the gel was equilibrated firstly in transfer buffer containing 5 mM EDTA twice, furthermore equilibrated another twice in transfer buffer without EDTA, each time for 20 min.Proteins were transferred from the gel to PVDF membrane at a constant pressure of 80 V for 36-48 h on ice, and then analyzed for phosphorylation by Western blot analysis using an anti-GFP antibody.
To identify MoMkk1-or MoAtg1-dependent MoAtg4 phosphorylation sites, mycelia proteins were extracted from ΔMoatg4/MoATG4-GFP and ΔMomkk1/MoATG4-GFP strains treated with 1 mg/ml CFW for 5 h or without, and ΔMoatg4/MoATG4-GFP and ΔMoatg1/ MoATG4-GFP strains induced with MM-N for 5 h or without.After purification by anti-GFP beads, protein samples were separated on 12% SDS-PAGE and stained by Coomassie brilliant blue.MoAtg4-GFP bands were sliced, and phosphorylation sites were identified (Beijing Protein Innovation) by LC-MS/MS.Two independent experiments were used for phosphorylation site identification.

PLOS PATHOGENS
temperature for 1 h, then 10-fold of cold acetone was added to stop the reaction.Phosphorylation protein was stained by Pro-Q Diamond Phosphorylation Gel Stain (Thermo Fisher Scientific, P33301), the widely used phosphor-protein gel-staining fluorescence dye, in the dark at room temperature for 1 h.Proteins were pelleted with cold acetone and washed twice with 0.5 ml of cold acetone.The protein pellet then was dissolved in 200 μl of ddH 2 O after drying and measured phospho-fluorescence at 590 nm (excited at 530 nm) using a Cytation3 microplate reader (Biotek, Winooski, VT, USA).Groups without kinase or substrate proteins were set up as controls [14].

Microscopic observation
Strains were cultured in liquid CM and transferred to MM-N or CM containing 1 mg/ml CFW (both adding 2 mM PMSF) [2] for 5 h.50 μM MDC (Sigma-Aldrich, D4008) was used to stain AB at dark for 20-30 min and the observation was performed using a Zeiss LSM710 microscope.For the analysis of chitin distribution, hyphae cultured in CM were stained with CFW at a final concentration of 10 μg/ml in the dark for 5 min, being washed with distilled water before observation.

In vitro and in vivo MoAtg8 cleavage assays
GST-MoAtg8, His-MoAtg4, His-MoAtg4 3A , His-MoAtg4 3D , and His-MoAtg4 C206S proteins were expressed in E. coli BL21 cells. 1 μg His-MoAtg4, His-MoAtg4 3A , His-MoAtg4 3D or His-MoAtg4 C206S were mixed with 3 μg GST-MoAtg8 in a final volume of 1 ml.Following incubation at 37˚C for 0, 1, or 2 h, the loading buffer was used to stop the reaction.The in vitro cleavage of GST-MoAtg8 was analyzed by Western blot with anti-His and anti-GST antibodies.

Microscale thermophoresis (MST) analysis
Binding affinities of MoAtg4, MoAtg4 3A , MoAtg4 3D , and MoAtg4 C206S with MoAtg8 were measured by MST in a Monolith NT.Label Free instrument (Nano Temper Technologies, GMBH). 10 mM GST-MoAtg8 was labeled with RED-NHS and mixed with a range of concentrations of His-MoAtg4, His-MoAtg4 3A , His-MoAtg4 3D , and His-MoAtg4 C206S (1:1, v/v).An empty His protein was used as a control.Mixtures were loaded into NT.Label Free standard capillaries and measure with 40% MST power.Raw data was integrated and fitted to a binding model using MST analysis software (Version 1.5.41)[58].

Statistical analysis
Each experiment was performed with at least 3 replicate measurements and presented as mean ± standard deviation.One-way analysis of variance (ANOVA) with Tukey's HSD was used to determine significant differences for multi-sample experiments with one variable.Two-way ANOVA with Tukey's HSD was used to determine significant differences for multivariable analyses.

Fig 1 .
Fig 1. Cell wall stress-induced autophagy is dependent on CWI kinases.(A) Hyphae of Guy11, ΔMomck1, ΔMomkk1, and ΔMomps1 strains cultured in CM were treated without or with CFW for 5 h (both adding 2 mM PMSF), and autophagic bodies (AB) were observed by confocal microscopy after MDC staining.C: cytoplasm.V: vacuole.(B) Quantification of vacuoles with AB as shown in (A).(C) Quantification of cells with vacuole-localized RFP-MoAtg8 after CFW treatment.(D) Total mycelial proteins after CFW treatment were extracted and analyzed by Western blot analysis with anti-RFP and anti-Actin antibodies.The amount of free RFP was compared with the total amount of RFP-MoAtg8 and free RFP to quantify autophagic levels.CFW treatment: treated with 1 mg/ml CFW.Data (n = 100) from three independent experiments were used for statistical analysis by two-way ANOVA with Tukey's HSD.Scale bar: 5 μm.https://doi.org/10.1371/journal.ppat.1011988.g001

Fig 2 .
Fig 2. The CWI kinase MoMkk1 interacts with and phosphorylates MoAtg4 in the cytoplasm.(A) In vitro phosphorylation reaction of GST-MoMkk1 with His-MoAtg4, His-MoAtg3, GST-MoAtg5, or His-MoAtg16 fusion proteins in the presence of 50 μM ATP, and phosphorylation proteins were stained using Pro-Q Diamond Phosphorylation Gel Stain.Cytation3 microplate reader was then used to measure the Phospho-Fluorescence signal at 590 nm (excited at 530 nm).(B) GST-MoMkk1, empty GST, and His-MoAtg4 proteins were obtained for GST pull-down analysis.(C) MoMkk1-GFP and Fig).MoMps1 phosphorylation in ΔMoatg4/MoATG4 3A and ΔMoatg4/MoATG4 3D under CFW treatment was unaffected (S4G Fig), suggesting that MoMkk1-dependent MoAtg4 phosphorylation might not play a role in CWI signaling.

Fig 3 .Fig 4 .
Fig 3. MoMkk1-dependent MoAtg4 phosphorylation functions in the development and virulence of M. oryzae.(A) Strains cultured on SDC (straw decoction and corn) medium at 28˚C for 7 d in the dark, followed by 3 d of continuous illumination under fluorescent light for conidiation assay.(B) Conidial suspensions of Guy11, ΔMoatg4, ΔMoatg4/3A, ΔMoatg4/3D, and ΔMoatg4/MoATG4 strains were sprayed onto 2-week-old rice seedlings (cultivar CO39) for virulence analysis.Diseased rice leaves were photographed at 7 days post inoculation (dpi).(C) Quantification of diseased leaf area as shown in (B) by Image J. (D) Severity of rice blasts was evaluated by quantitative PCR of M. oryzae genomic 28S rDNA relative to rice genomic RUBQ1 DNA as shown in (B).(E) Close observation and statistical analysis of the invasive hyphae (IH) growth in rice leaf sheath at 36 hours post inoculation (hpi).Statistical analysis of IH at 100 appressorium penetration sites by rating from type I to type IV (type 1, no penetration; type 2, a single IH with no branch; type 3, IH extended but was limited in one plant cell; type 4, IH extended to neighboring cells).The arrow points to appressorium, and the asterisk indicates IH.Scale bar: 5 μm.3D: replaced S97, S276, and T278 with D to mimic phosphorylated form of MoAtg4.Different letters indicate statistically significant differences (Duncan's new multiple range test, p < 0.05 or p < 0.01).https://doi.org/10.1371/journal.ppat.1011988.g003 (S5E Fig).Western blot analysis further showed that MoAtg4 3D diminished ability to cleave MoAtg8 (Fig 4H).In contrast, MoAtg8 C-terminal cleavages were the same at the conidial stage in these strains (S5F Fig), suggesting that autophagy might not be induced.We also investigated MoAtg8 C-terminal cleavage in single-site phosphorylation mutations strains, ΔMoatg4/MoATG4 S97D , ΔMoatg4/MoATG4 S276D , and ΔMoatg4/MoATG4 T278D , and found that C-terminal cleavage was all weakened but still stronger than ΔMoatg4/MoATG4 3D (S5D Fig), suggesting all individual phosphorylation residue is involved in MoAtg8 C-terminal cleavage.These results indicated that MoMkk1-dependent MoAtg4 phosphorylation inhibits the C-terminal cleavage of MoAtg8.
and MoMkk1 was the same as MoAtg4 with MoMkk1 (S5H Fig).Taken together, MoMkk1-dependent MoAtg4 phosphorylation inhibited the C-terminal cleavage of MoAtg8 to negatively control autophagy.MoAtg1-mediated MoAtg4 phosphorylation is involved in the development and pathogenicity of M. oryzae Our previous analysis showed that MoAtg4 remains phosphorylated in the ΔMomkk1 mutant (S2E and S2F Fig), suggesting a possibility of phosphorylation by other protein kinases.Since the yeast and human Atg1/Ulk1 phosphorylate Atg4/Atg4B [1,37], we tested whether MoAtg1 also phosphorylates MoAtg4.The interaction between MoAtg1 and MoAtg4 was first observed at PAS (S6A and S6B Fig).The in vitro phosphorylation assay showed MoAtg1 phosphorylates MoAtg4 (S6C Fig).Further LC-MS/MS data revealed that the phosphorylation of MoAtg4 at S67 was detected in Guy11, but not in the ΔMoatg1 mutant, under nitrogen starvation, Guy11 and ΔMoatg1 without starvation (S6D Fig).
was similar, compared to MoMkk1 with MoAtg4 (S9G Fig), indicating that the action of MoAtg1 on MoMkk1 does not affect MoMkk1-dependent MoAtg4 phosphorylation.MoMkk1 phosphorylation in ΔMoatg4/MoATG4 3A and ΔMoatg4/MoATG4 3D treated with DTT showed no difference compared with ΔMomkk1/MoMKK1-RFP (S9H Fig), suggesting that MoMkk1-dependent MoAtg4 phosphorylation does not influence MoMkk1 phosphorylation of MoAtg1.In addition, MoAtg8 C-terminal cleavages in ΔMomkk1/MoMkk1 S115A and ΔMomkk1/MoMkk1 S115D were as the same as Guy11 with CFW treatment (S9I Fig), indicating that MoMkk1 phosphorylation by MoAtg1 does not affect the cleavage activity of MoAtg4.Interestingly, MoAtg8 C-terminal cleavage in ΔMomck1 and ΔMomkk1 under CFW treatment was stronger, while the cleavage in ΔMomps1 was the same as Guy11 (S9I Fig).We inferred that the cleavage in ΔMomps1 was inhibited by MoMkk1-dependent MoAtg4 phosphorylation, but the loss of MoAtg4 phosphorylation resulted in stronger MoAtg8 cleavage in ΔMomck1 and ΔMomkk1.Taken together, MoAtg4 phosphorylation mediated by MoMkk1 and MoAtg1 collectively govern the autophagy and pathogenicity in M. oryzae.

Fig 5 .
Fig 5. MoAtg4 phosphorylation mediated by MoMkk1 and MoAtg1 orchestrates autophagy.(A) AB observation after MDC staining.C: cytoplasm.V: vacuole.Scale bar: 5 μm.(B) Quantification of vacuoles with AB as shown in (A).(C) AB observation using TEM.The arrow points to AB. Scale bar: 0.5 μm.(D) Quantification of AB as shown in (C).(E) Cleavage assay of GST-MoAtg8 in vitro.(F) Quantification of GST-MoAtg8 cleavage assay as shown in (E).(G) In vivo cleavage assay of MoAtg8-RFP at the mycelium (nitrogen starvation for 1 h) and appressorium stages.(H) The MoAtg1 targets and phosphorylates MoAtg4 at PAS to inhibit deconjugation of MoAtg8-PE, whereas MoMkk1 phosphorylates MoAtg4 in the cytoplasm to reduce the C-terminal cleavage of MoAtg8.Together, they coordinate proper autophagic levels for regulating the development and pathogenicity of M. oryzae (Fig 6).

Fig 6 .
Fig 6.A model of autophagy and CWI pathways coordinating development and pathogenicity through MoAtg4 phosphorylation.MoAtg4 cleaves the MoAtg8 C-terminus (7aa) to activate the conjunction of MoAtg8 with phosphatidylethanolamine (PE) of PAS.Following autophagosome formation, MoAtg4 deconjugates MoAtg8 from PE anchors to promote the fusion of autophagosomes with vacuoles.The CWI kinase MoMkk1 phosphorylates MoAtg4 in the cytoplasm to attenuate the C-terminal cleavage of MoAtg8.On the other hand, The ATG protein kinase MoAtg1 phosphorylates MoAtg4 at PAS to inhibit the deconjugation of MoAtg8-PE.Two MoAtg4 phosphorylation events collectively maintain proper autophagic levels to govern the development and pathogenicity of M. oryzae.https://doi.org/10.1371/journal.ppat.1011988.g006