In this study, M. oryzae strain P131 was regarded as a wild type, and all the fungal strains were cultured on oatmeal tomato agar plates at 28°C to assess growth and conidiation capacity. To test various stress sensitivity, all strains were grown on complement medium (CM) agar added with 0.1 mg/ml CFW (Sigma-Aldrich, USA), 0.2 mg/ml Congo Red (Sigma-Aldrich), 0.005% sodium dodecyl sulfate (SDS), 0.5 M NaCl, or 10 mM H₂O₂. Then the colony diameters were measured after 5 days of growth (Liu et al., 2018). For calcium sensitivity assessment, the strains were cultured on minimum medium (MM) supplemented with 200 mM CaCl₂ or/and 0.2 ng ml⁻¹ FK506 for 5 days. Genome DNA and total RNA were extracted from mycelium after 36 h of culture in CM (180 rpm).

The protein sequence of MoRCN1 was regarded as a query to blast homolog proteins from other different species on the Enzembl Fungi website (http://fungi.ensembl.org). Clustal_W was used to align the amino acid sequences of homologous proteins in the species (Larkin et al., 2007). The phylogenetic tree was accomplished using MEGA7, and the percentage of replicate in which associated taxa clustered together in the bootstrap test (1,000 replicates) was shown next to the branches (Kumar et al., 2016).

For gene deletion vector construction, 1.5-kb upstream and downstream fragments of MoRCN1 were separately amplified from the genome DNA of P131 and then cloned into the deletion vector pKNH (Supplementary Table 6) (Chen et al., 2014). The constructed deletion vector pKNH-RCN1 was transformed into protoplasts of P131 to gain the MoRCN1 deletion transformants. For gene complementation assay, a fragment containing a 1.5-kb native promoter region, coding region, and 0.5-kb terminator region of MoRCN1 was cloned to the pKN vector (Supplementary Table 6) (Chen et al., 2014). The constructed vector pKN-RCN1 was transformed into protoplasts of the ΔMorcn1 mutant, and subsequent transformants were selected by 400 μg/ml neomycin (Amresco, Solon, OH, USA) and confirmed by the PCR-mediated method. All the primers used in this study are shown in Supplementary Table 7 (see Supporting Information).

For virulence test on rice, conidia suspensions with a concentration of 5 × 10⁴ conidia/ml in 0.025% Tween 20 were sprayed onto 21-day-old rice leaves (Oryza sativa cv. LTH). The rice samples were kept in a growth chamber at 28°C with 90% humidity in dark for 36 h, then exposed to light. Lesion formation could be observed at 5 days after incubation. To observe the infection process, 7-day-old barley (Hordeum vulgare cv. E9) were inoculated with conidia suspension (3 × 10⁴ conidia/ml) of different strains, and the infection process was observed at different time points using a Nikon Ni90 microscope (Nikon, Tokyo, Japan).

Total RNA samples were extracted from CM media cultivating for 36 h (28°C, 160 rpm) using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and were treated with DNaseI to digest genome DNA. A total of 1 mg RNA was used to synthesis cDNA, which was used for the template of qRT-PCR to evaluate the gene expression level. The qRT-PCR experiment was performed by using an SYBR Green PCR Master Mix kit (Takara, Dalian, China) on an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA).

To confirm the interaction between MoRCN1 and MoCNA, the coding region of MoRCN1 was cloned into vector pKN-Flag to gain pKNFlag-RCN1, and MoCNA was cloned into pRGTN (containing GFP) to gain pRGTN-CNA. The pKNFlag-RCN1 and pRGTN-CNA were co-transformed into the WT strain, and the resulting transformants were selected by GFP signal detection and western blot with an anti-FLAG antibody. To perform the Co-IP experiment, the total number of proteins of the selected transformants were extracted and incubated with 20 μl of anti-FLAG (DYKDDDDK) magnetic beads magarose beads (B26101, Bimake, Beijing, China) for 4 h at 4°C. The magnetic beads were washed three times using TBS buffer (50 mM Tris, 0.15 M NaCl, and pH7.4), and proteins were eluted from the anti-FLAG beads and analyzed by western blot with the anti-GFP antibody. The pKNFlag-RCN1 and pRGTN-CNA vectors were also, respectively, transformed into the WT strain and used as controls and analyzed by western blot.

The full-length cDNA of MoRCN1 was cloned into pGBKT7 as the bait construct pGBKT7-RCN1, and the cDNA of MoCNA was cloned into pGADT7 as the prey construct pGADT7-CNA. The pGBKT7-RCN1 and pGADT7-CNA were co-transformed into the yeast strain AH109 according to the manufacturer's instructions (Clontech, San Francisco, CA, USA). All transformants were assayed with 1 × 10⁶ cells/μl droplet on SD-Leu-Trp-His and SD-Leu-Trp-His with 20 mg/ml X-α-gal plates. Self-activation showed that both pGBKT7-RCN1 and pGADT7-CNA had no self-activation. The interaction of pGBKT7-53 and pGADT7-T was used as a positive control, and the interaction of pGBKT7-Lam and pGADT7-T was used as a negative control.

The DAB (3,3-diaminobenzidine, Sigma-Aldrich, USA) solution with a concentration of 0.1 mg/ml was used to detect ROS in host cells infected by different strains. For CFW staining, mycelia and conidia were harvested from CM medium and stained with 10 μg/ml CFW (Sigma-Aldrich, United States) for 10 min in the dark, then observed under the Nikon microscope Ni90 (Nikon, Tokyo, Japan). Glycogen was detected by KI/I₂ staining solution (60 mg/ml KI and 10 mg/ml I₂) (Thines et al., 2000). Lipid droplets were stained using Nile Red solution, consisting of 50 mM Tris/maleate buffer (pH 7.5), 20 mg/ml polyvinylpyrrolidone, and 2.5 mg/ml Nile Red (Sigma-Aldrich, St. Louis, MO, USA). For different kinds of staining assays, the spore suspension (5 × 10⁵ spores/ml) was dropped on plastic coverslips and kept in growth chamber and stained for 15 min at different time points, then detected using an epifluorescence microscope (Ni90, Nikon, Tokyo, Japan).

The MoRCN1:GFP fusion vector was generated by cloning the MoRCN1 promoter and coding region into the N-terminus of GFP in the vector pGTN (Supplementary Table 6). The resulting pGTN-RCN1 vector was transformed into the ΔMorcn1 mutant, and the positive transformants were used to observe subcellular localization at different developmental stages and infection processes using a confocal microscope Leica TCS SP8 (Leica Microsystems, Germany).

The wild-type strain P131 and the ΔMorcn1 mutants were incubated for 48 h in CM medium with or without 200 mM CaCl₂ (WT, WT[Ca²⁺], ΔMorcn1, ΔMorcn1[Ca²⁺]). Three biological replicates were used for each sample. Total RNA from each mycelium sample was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA). The mRNA was enriched and purified using poly-T oligo-attached magnetic beads and fragmented into small pieces using divalent cations. The RNA fragment was then used to synthesize cDNA, “A” base was added, and adapter was connected. After purification and enrichment by PCR amplification, the samples were used to obtain the final cDNA library for sequencing using combined probe anchor synthesis sequencing methods (Beijing Genomics Institution, Beijing, China).

Differentially expressed genes were analyzed by the DESeq function in the DESeq2 package, and expression with log2|FC| > 1 with <0.05 p-adjust values was defined as DEGs. Gene Ontology enrichment analysis was performed using AgriGO toolkits (Du et al., 2010) using p-adjust value cutoffs (0.05) for significance. KEGG enrichment of metabolic pathways was analyzed using the R package clusterProfilter version 3.16 (Yu et al., 2012). Pathview (Version 3.11) was used to produce detailed maps for the selected pathways (Luo et al., 2017).

Using the S. cerevisiae RCN1 protein sequence as a query to perform a BLASTP search in the Magnaporthe genome database (Ensembl Fungi), the M. oryzae MoRCN1 (MGG_03218) was identified. Multiple protein sequence alignments indicated that M. oryzae MoRCN1 protein sequence shares high homology with numerous proteins of different eukaryotes, including S. cerevisiae, Fusarium oxysporium, C. neoformans, Danio rerio, and Homo sapiens (Supplementary Figure 1A). The M. oryzae MoRCN1 protein is predicted to be composed of 266 amino acid residues that contain a calcipressin domain (Supplementary Figure 1B). Phylogenetic tree analysis was analyzed using MEGA7 software, which showed that the MoRCN1 protein shares the highest homology with ascomycete fungi, including Neorospora crassa, Fusarium oxysporum, F. graminearum, while is not closely related to basidiomycete fungi Ustilago maydis and C. neoformans (Supplementary Figure 1C).

To gain insight into the functions of MoRCN1 of M. oryzae, we first examined its expression pattern at developmental stages, including in mycelium, conidium, germ tube, appressoria, and invasive hypha at 18, 24, and 48 h post inoculation (hpi). According to the qRT-PCR result, in comparison to the hyphal stage, the expression of MoRCN1 was highly expressed in conidium, appressorium, primary infection hypha, and late infection hypha (Supplementary Figure 2), suggesting it plays important roles in these stages.

To further determine the functional roles of MoRCN1 in M. oryzae, we subsequently disrupted this gene in the wild-type strain P131 by using a homologous recombination strategy (Supplementary Figure 3A). After a PCR-based screening (Supplementary Figure 3B), two independent MoRCN1 deletion mutants (KO-1 and KO-2) were obtained. We also performed a complementation experiment by introducing the native promoter-driven MoRCN1 coding region into KO-1. Most of the complementation transformants were recovered to the wild type in growth, conidiation, and virulence, and one of which, cRCN1, was chosen for further analysis.

To investigate the roles of MoRCN1 in vegetative growth, we observed the colony growth and mycelial morphology of the ΔMorcn1 mutants. Compared with that of the wild type, we found a severe growth reduction in the ΔMorcn1 mutants, since the colony diameter of the ΔMorcn1 mutants was around 1.6 cm shorter than that the wild type after cultivating on OTA plates for 120 h at 28°C (Figures 1A,B). Staining of mycelial tip cells with Calcofluor White showed that ΔMorcn1 appeared to form shorter hyphal compartments than the wild type (Figures 1C,D).

We also tested whether the deletion of MoRCN1 affected the conidia development. We found that the ΔMorcn1 mutants formed sparse conidia on the conidiophores when observed under the microscope. In contrast, the wild-type and complement strains formed dense conidia on the conidiophores (Figure 1E). Coordinately, the ΔMorcn1 mutants formed only 42% of the wild-type conidia on OTA plates (Figure 1F). These results showed that MoRCN1 is required for asexual growth and conidia formation.

To test whether disruption of MoRCN1 affected virulence of M. oryzae, the barely leaves and rice seedlings were sprayed with conidia from different strains. A severe reduction in virulence was observed in the ΔMorcn1 mutant-infected plants compared with those infected by the wild type (Figures 2A,B). ΔMorcn1 mutants are also cannot spread well on the wounded rice leaves, suggesting a defect in host cell invasion (Figure 2C). We then observed whether deletion of MoRCN1 affected appressorium formation. On hydrophobic surface, more than 90% of conidia in the wild type formed appressoria at 12 hpi, but only 43% of conidia in the Δrcn1 mutants formed appressoria, indicating that MoRCN1 plays a role in appressorium formation (Figures 2D,E). To further determine why the ΔMorcn1 mutants were attenuated in virulence, we observed the cellular infection processes in barley epidermis cells. At 25 hpi, around 60% of the ΔMorcn1 mutant conidia still remained at the appressorium stage and 20% formed primary invasive hyphae, while 75% of the wild-type conidia formed invasive hyphae and 45% formed more than two branches. At 30 hpi, nearly 70% of the wild-type IH formed more than two branches, whereas it was no more than 20% in the ΔMorcn1 mutants (Figures 2F–H). Taken together, MoRCN1 is indispensable for appressorium formation, penetration, and invasive growth in the host cells.

Since the formation of a functional appressorium is accompanied by the degradation of fatty acids and the utilization of glycogen, we next detected the cellular distribution of glycogen and lipids during the appressorium development. Glycogen can be stained with I₂/KI solution, while the lipid droplets can be detected by Nile Red staining. In wild type, the glycogen was translocated from conidia to nascent appressoria, and rapidly consumed after 12 hpi, and disappeared at 24 hpi. While in the ΔMorcn1 mutants, the glycogen of the ΔMorcn1 mutants is utilized much slower and were still noticeable at 24 hpi (Figure 3A). Similarly, in wild type, lipid droplets were transferred from conidia to nascent appressoria and were consumed after 8 hpi, while in the ΔMorcn1 mutants, degradation of the lipid droplets was much slower, and a large amount of lipid droplets could be still detected at 24 h (Figure 3B). These results indicated that deletion of MoRCN1 affects glycogen and lipid metabolism, which is important for conidia storage utilization and functional appressorium formation for host penetration.

Retardation in the invasive growth of M. oryzae is usually caused by failure to suppress host ROS accumulation. To test whether ΔMorcn1 mutants could induce host ROS accumulation, the M. oryzae-infected barley epidermis cells at 30 hpi were stained with 3, 3-diaminobenzidine (DAB). Approximately 70% of the ΔMorcn1 mutant-infected host cells had ROS accumulation with reddish-brown color, while no more than 30% of the wild-type strain infected host cells were detected, confirming that the ΔMorcn1 mutants were impaired in the detoxification of ROS produced by the host cells (Figures 4A,B). This result indicated that MoRCN1 is important for suppressing host defense response, especially the oxidative stresses.

To further determine the role of MoRCN1 in stress response, we tested the sensitivities of the ΔMorcn1 mutants to different cellular stress agents. The tested strains were cultured on CM plates amended with different cellular stress agents, including those involved in cell wall disruption (0.2 mg/ml CR, 0.1 mg/ml Calcofluor White [CFW], and 0.005% SDS), osmotic stress (0.5 M NaCl), and oxidative stress (10 mM H₂O₂). The result showed that ΔMorcn1 mutants were more sensitive to all the above mentioned stresses (Figures 5A,B), suggesting that MoRCN1 plays a central role in responding to different kinds of stresses, including cell wall integrity and osmotic and oxidative stress responses.

To study the subcellular localization of MoRCN1, the GFP was fused with the C-terminus of MoRCN1 protein and transformed into ΔMorcn1 mutants. The transformants were all recovered to the wild type, and one of which was used for subcellular observation. Through the fluorescence microscopy, we observed that MoRCN1-GFP was distributed in the cytoplasm at the stages of conidia, vegetative hyphae, appressorium, and invasive hyphae, with a stronger fluorescence accumulation around the nuclei (Figure 6A). Since high concentration of calcium can activate the calcium signaling pathway and induce calcineurin to dephosphorylate target proteins in eukaryotes (Park et al., 2019), we wonder whether calcium ions could also induce nuclear localization of MoRCN1. As expected, we observed MoRCN1-GFP was evidently translocated into the nuclei under 200 mM CaCl₂ treatment in vegetative hyphae (Figure 6B). This result showed that MoRCN1 was located in the cytoplasm under normal conditions, and a calcium ion could promote its transfer into the nucleus, suggesting a role of MoRCN1 in calcium signaling response.

MoRCN1 has been reported to be a calcipressin that regulates calcineurin by directly binding in Botrytis cinerea, Aspergillus fumigatus, and C. neoformans (Görlach et al., 2000; Pinchai et al., 2009; Harren et al., 2012). To experimentally prove whether MoRCN1 interacts with the calcineurin subunit A (MoCNA) in M. oryzae, we performed a yeast two hybrid (Y2H) assay. As expected, MoCNA clearly interacted with its putative regulator MoRCN1 in vitro (Figure 7A).

To further confirm the interaction of MoRCN1 and MoCNA in vivo, we performed a co-immunoprecipitation (CoIP) assay. We generated the MoRCN1-3× FLAG and MoCNA-GFP constructs, which were co-transformed into the wild-type strain. As a control, the two vectors were also, respectively, transformed into the wild-type strain. The MoRCN1-interacted proteins were enriched from the total protein with anti-FLAG magnetic beads, and the eluted protein from the beads was detected with GFP antibody. As shown in Figure 7B, the MoRCN1-FLAG fusion protein co-immunoprecipitated with the MoCNA. Thus, MoRCN1 also interacts with MoCNA in vivo.

In the wild-type strain, MoCNA-GFP was clearly located in the cytoplasm and the septum but not in the nucleus. In the ΔMorcn1 mutants, the localization of MoCNA-GFP was not changed, but its fluorescence intensity was dramatically decreased compared with that in the wild type (Figure 7C). Under the treatment of 200 mM CaCl₂, the localization of MoCNA-GFP was not changed either (Supplementary Figure 4). Subsequent western blot analysis also confirmed a significant reduction in the protein level of MoCNA-GFP in the ΔMorcn1 mutants (Figure 7D). The qRT-PCR experiment demonstrated only a slight reduction in expression of MoCNA in ΔMorcn1 mutants compared with the wild type, suggesting that the reduction in the protein level should be due to a decrease in its transcription level (Figure 7E). We also detected an increased expression level of MoCNB, the regulatory subunit of calcineurin, in the ΔMorcn1 mutants (Figure 7E), suggesting a compensation effect resulted by protein reduction in the calcineurin. Together, these results indicated that MoRCN1-regulated protein stability of the calcineurin regulatory subunit MoCNA in M. oryzae.

To further investigate whether MoRCN1 plays a role in the calcineurin pathway, we compared the growth of wild-type and ΔMorcn1 strains on minimal medium under high-calcium levels (200 mM CaCl₂) with or without the calcineurin inhibitor FK506. Interestingly, the ΔMorcn1 strain exhibited a greater hyphal growth than the wild-type strain in the medium containing CaCl₂ (Figures 7F,G). However, in the presence of FK506, the increased calcium tolerance of the ΔMorcn1 strain was completely reversed, indicating that MoRCN1 contributed to calcium tolerance in a calcineurin-dependent manner (Figures 7F,G), suggesting that MoRCN1 may play a negative feedback regulatory role in response to calcium tolerance.

To understand the impact of high Ca²⁺ concentration on gene expression at the transcriptional level in the wild-type and ΔMorcn1 strains, we performed a comparative transcriptomic analysis in vegetative hyphae treated with 200 mM CaCl₂ for 1 h. In WT, compared with that under the normal condition, the expression of 1,666 genes was increased under the calcium ion-treatment condition, while that of 804 genes was decreased (Figure 8A). In ΔMorcn1 mutants, under the calcium ion-treatment condition, we identified 337 upregulated genes and 491 downregulated genes compared with that of the wild-type strains (Figures 8A,B; Supplementary Figure 5). This result indicates that MoRCN1 plays both activator and repressor roles in gene expression, and relatively more genes are activated by MoRCN1. Interestingly, around one half of the downregulated genes in ΔMorcn1 mutants were also upregulated genes in the wild-type strain induced by calcium ions (Figure 8C), indicating that MoRCN1-mediated calcium response accounts for a large part of the calcium response of M. oryzae.

We also compared MoRCN1-regulated genes with the MoCRZ1-binding genes, which was reported by Kim et al. (2010). MoCRZ1 is an important calcium-responsive transcription factor in M. oryzae (Kim et al., 2010). As shown in Figure 8D, 31 downregulated genes of ΔMorcn1 were also listed in the MoCRZ1-binding genes, but only one gene was listed in the upregulated genes of ΔMorcn1, suggesting that MoRCN1 plays an important role in activating MoCRZ1, the C2H2 transcription factor activated by calcineurin dephosphorylation (Kim et al., 2010).

Hierarchical clustering of MoRCN1-regulated genes gives an overview of calcium induction and highlights major clusters (Figure 8E). We also searched the previously published genes in the list of MoRCN1-regulated genes and found some upregulated genes in ΔMorcn1, including Rbf1, Gel5, RAS2, XYR1, CONx4, MoHEX13, MoVelC, BAS3, and PTH11. Among them, RAS2 and PTH11 are known as appressorium formation regulators (DeZwaan et al., 1999; Park et al., 2006). MoVelC and CONx4 are involved in development of conidial morphology (Kim et al., 2014; Cao et al., 2016). MoHEX13 and BAS3 are effector genes that function in invasive growth (Mosquera et al., 2009). RBF1 is a virulence gene that is essential for the focal BIC formation during invasive growth (Nishimura et al., 2016) (Figure 8F). These data suggest that MoRCN1 could negatively regulate different genes involved in development and infection upon high calcium induction.

Gene Ontology (GO) enrichment analysis of differentially expressed genes showed that biological processes of interaction with the host via secretary system, secondary metabolic process, secondary metabolite biosynthesis, ascospore wall assembly, lipid homeostasis, membrane raft assembly, as well as transmembrane transport were significantly enriched, most of which were related to stress response. For cellular components, the membrane raft was enriched. For molecular function, heme binding, monooxygenase activity, oxidoreductase activity, and electron transfer activity were enriched (Figure 9A), most of which were related to energy metabolism. In sum, results of the GO enrichment analysis were consistent with the fact that the MoRCN1 is involved in response to different stresses.

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis shows that the MoRCN1-regulated genes induced by calcium were mainly involved in pathways related to amino acid metabolisms (tryptophan metabolism, tyrosine metabolism, phenylalanine metabolism, and glycine, serine, and threonine metabolism), lipid metabolism (fatty acid degradation, ether lipid metabolism, and glycerolipid metabolism), secondary metabolite metabolism (thiamine metabolism, indole diterpene alkaloid biosynthesis, diterpenoid biosynthesis, arachidonic acid metabolism, riboflavin metabolism, nicotinate and nicotinamide metabolism, staurosporine biosynthesis, and aflatoxin biosynthesis), and ABC transporters (Figure 9B). These results suggest that MoRCN1 is important for metabolism pathways for invasive growth.