The wild-type M. oryzae strain Guy11 and all generated strains were cultured on complete medium (CM) agar plates at 28°C (Talbot et al., 1993). Fungal mycelia were harvested from liquid CM and used for genomic DNA and RNA extractions. Protoplasts were prepared and transformed as described previously (Sweigard et al., 1992). Transformants were selected on TB3 medium with 250 mg/mL hygromycin B or 200 mg/mL bleomycin (Invitrogen, Carlsbad, CA, USA). For conidiation, mycelial blocks were maintained on straw decoction and corn agar medium (SDC) at 28°C for 7 days in the dark followed by 3 days of continuous illumination under fluorescent light (Zhang et al., 2009). For conidial production analysis, conidia were harvested and counted with a haemacytometer under a microscope. The final result of each sample = total number of conidia/colony area. Vegetative growth of Guy11, ΔMocpa2 mutant, and complemented strains was measured on CM, minimal medium (MM), and SDC medium for 7 days. Mycelial plugs were placed onto the freshly prepared MM agar plates with a range of concentrations of arginine or ornithine and cultured in the dark at 28°C for 7 days.

To generate a MoCPA2 gene replacement construct, 1.0-kb upstream and 1.0-kb downstream flanking sequences were amplified from M. oryzae genomic DNA by PCR using the primer pairs FL04503F1 (F)/FL04503F2 (R) and FL04503F3 (F)/FL04503F4 (R), respectively. The upstream flanking sequences were digested by XhoI and EcoRI, and cloned into a pCX62 vector containing a hygromycin B cassette to generate the plasmid pCX62-MoCPA2-UP. The plasmid pCX62-MoCPA2-UP and the downstream flanking sequence were digested by SpeI and SacI and ligated to generate the final construct pCX62-MoCPA2-U-D. Amino acid sequence alignments were performed using the CLUSTAL_W program. For complementation, the fragments containing the full-length coding region and a 1.5-kb native promoter region of the gene were amplified using primers FL04503F /FL04503R, and inserted into pYF11 containing bleomycin resistance and green fluorescent protein (GFP) genes to generate the pYF11-MoCPA2-GFP fusion construct, which was then transformed into protoplasts of the mutant. The complemented strains were screened by bleomycin resistance and GFP fluorescence observation. For construction of MoCPA2 domain deletion mutants, the fragments containing different domain deletion coding region and their 1.5-kb native promoter region were amplified using primers in Table S1, and inserted into pYF11 containing bleomycin resistance and GFP. The resulting constructs were then transformed into protoplasts of the mutant. The transformants were screened by bleomycin resistance and GFP fluorescence observation.

For a rice spraying assay, conidial suspensions (5 × 10⁴ spores/mL) were sprayed onto 2-week-old rice seedlings and placed in a moist chamber at 28°C for 24 h in darkness. They were then transferred to another moist chamber with a 12-h light/12-h dark photoperiod under fluorescent lights. Disease severity was assessed by taking photographs of 6-cm rice blades 7 days after inoculation. For a detached barley infection assay, conidial suspensions (10⁵, 10⁴, and 10³ spores/mL) with or without arginine were dropped onto the barley leaves, and photographed after 5 days. For a rice sheath assay, 100 μL of conidial suspensions (5 × 10⁴ spores/mL) were inoculated into the inner leaf sheath cuticle cells. After 48 h of incubation under humid conditions at 28°C, the leaf sheaths were observed under a microscope (Zhang et al., 2011).

To observe the plant defense response, infected leaf-sheath cells were prepared as described above (Qi et al., 2015) and examined under a microscope with ultraviolet radiation. Host derived reactive oxygen species (ROS) staining was performed as previously described (Chi et al., 2009). The experiments were performed in three independent biological experiments with three replicates.

The target gene probe and HPH probe were amplified with the primer pairs FL04503cF/FL04503cR (for MoCPA2) and FL1111/FL1112 (for HPH), respectively. Probe labeling, hybridization, and detection were performed with a DIG High Prime DNA Labeling and Detection Starter Kit (Roche Applied Science, Penzberg, Germany). Total RNA was isolated from frozen fungal mycelia using an RNA extraction kit (Invitrogen, USA). To measure the relative abundance of gene transcripts, RNAs were extracted from mycelia grown in CM liquid medium for 2 days at 28°C in a 150-rpm orbital shaker (Guo et al., 2011). To measure the relative abundance of MoCPA2 transcripts during diverse fungal development stages, the total RNA samples were extracted from mycelia grown in CM liquid medium, from conidia, and from plants inoculated with the Guy11 conidia (1 × 10⁸ spores/mL) for 8, 24, 48, and 72 h by the method described above (Guo et al., 2010; Dong et al., 2015). Total RNA was pretreated with DNase I and was then reverse transcribed (Invitrogen, Carlsbad, CA, USA). For the evaluation of transcription level of PR genes, RNA samples were extracted from the rice leaves inoculated with Guy11, the mutant (1 × 10⁸ spores/mL) and water (mock) for 48 h. qRT-PCR was performed on an ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. The stable-expression ACTIN gene (MGG_03982) amplified by primers FL04503RTF and FL04503RTR was used as internal control. All primers used in this study are listed in Table S1.

Each result was presented as the mean ± standard deviation (SD) of at least three replicated measurements. The significant differences between treatments were statistically evaluated by SD and one-way analysis of variance (ANOVA) using SPSS 2.0. The data between two specific different treatments were compared statistically by ANOVA, followed by F-test if the ANOVA result is significant at P < 0.05 and P < 0.01.

MoCPA2 (MGG_04503) was identified in the M. oryzae genome database as a homolog of yeast CPA2 which encodes a CPS. (http://www.broadinstitute.org/annotation/genome/magnaporthe_comparative/MultiHome.html). To investigate the role of MoCPA2 in M. oryzae, we first examined its transcript levels at various pathogen developmental stages using qRT-PCR analysis. Compared to the mycelial stage, the MoCPA2 transcript level was much higher in the conidial and early infectious stages, with a 2.7-fold increase in the conidial stage and a one- to three-fold increase in the early infectious stages at 8, 24, and 48 h post-infection (hpi). However, there were no expression differences in late infectious stages (72 hpi; Figure 1). These results suggested that MoCPA2 likely plays an important role in asexual development and early infection in the rice blast fungus. To further examine the biological function of MoCPA2, the MoCPA2 coding region was replaced with the hygromycin resistance cassette (HPH) via homologous recombination, and the successful gene deletion mutant was confirmed by Southern blot analysis (Figure S1).

To clarify the role of MoCpa2 in M. oryzae, we first observed the vegetative growth of the ΔMocpa2 mutant on CM, MM, and SDC media. After 7 days incubation, no significantly difference in colony diameter was observed between Guy11 and the ΔMocpa2 mutant on CM and SDC plates, but the mutant showed fewer aerial hyphae under the same conditions (Figure 2A). In contrast, the ΔMocpa2 mutant could not grow on MM plates (Figure 2A). In yeast, Cpa2 is known to catalyze the synthesis of an arginine precursor and is involved in arginine biosynthesis (Pierard and Schroter, 1978). Therefore, MM plates with various concentrations of arginine were prepared to test the growth of the mutant. The results showed that exogenous arginine was able to restore the growth rate but not the aerial hyphae of the mutant at low concentrations (Figure 2B). The ΔMocpa2 mutant displayed more aerial hyphae and became grayish white when the arginine concentration was increased, which was similar to that of the wild-type Guy11 strain at concentrations of 2.5 and 5.0 mM (Figure 2B). As the ornithine reacts with carbamoyl phosphate in the ornithine carbamoyl transferase step (Hilger et al., 1973; Pierard and Schroter, 1978), and is the precursor of arginine, we also added different concentrations of ornithine into MM medium to test the growth of the mutant. The results showed that exogenous ornithine was unable to restore the growth defects of the ΔMocpa2 mutant (Figure 2B). These results indicated that MoCpa2 plays a role in arginine utilization. Since nitric oxide synthase was known to be involved in the utilization of arginine (Durante et al., 2007). We further analyzed the activity of nitric oxide synthase in the mutant. The results revealed that the activity of nitric oxide synthase was decreased to 48.2 and 27.5% in the ΔMocpa2 mutant at 24 and 48 h post incubation, respectively (Figure S2), indicating that MoCpa2 is involved in the utilization of arginine medicated by the activity of nitric oxide synthase in M. oryzae.

Conidia are important structures for M. oryzae to infect the host plant (Kim et al., 2009; Liu et al., 2010). Therefore, we evaluated the ability of the ΔMocpa2 mutant to produce conidia by incubation the ΔMocpa2 mutant on SDC plates for 7 days. No conidia were collected from the ΔMocpa2 mutant compared with a large number of conidia produced by the wild-type Guy11 and complemented strains under the same conditions (Figures 3A,B), indicating that the ΔMocpa2 mutant was defective in asexual development. To determine whether this defect was related to arginine, exogenous arginine was added to SDC plates. In the presence of exogenous arginine, conidia production in the mutant was restored to 53 and 93% of that of the wild-type strains with arginine concentrations of 0.5 and 2.5 mM, respectively (Figures 3A,B). Meanwhile, exogenous ornithine was also added to the SDC plates, and no conidia were harvested even the concentration of ornithine increased to 2.5 mM (Figures 3A,B). These results indicate that MoCpa2 plays an essential role in conidiogenesis by controlling arginine biosynthesis in M. oryzae.

To gain insight into the functions of MoCpa2 in pathogenicity, ΔMocpa2 mutant and wild-type conidia were collected from SDC plates containing 2.5 mM arginine and sprayed onto 2-week-old rice seedlings. The ΔMocpa2 mutant caused tiny and restricted lesions in comparison to the typical blast lesions caused by the wild-type Guy11 and complemented strains (Figure 4A). A “lesion-type” scoring assay also confirmed this result. The mutant showed more type 1 and type 2 lesions, and caused fewer or no type 3–5 lesions (Figure 4B). These results suggested that even the conidia recovered by arginine cannot infect the host normally. Therefore, arginine as well as ornithine was added to conidial suspensions and inoculated onto detached barley leaves. The results showed that exogenous ornithine was unable to restore the virulence defects of the ΔMocpa2 mutant. In contrast, 0.5 mM arginine could partially restore the virulence of the mutant, while 2.5 mM arginine could completely restore this defect on barley leaves (Figure 4C). However, the defect was not suppressed on rice seedlings through the spraying assay (Figure S3A). These results suggest that MoCpa2 is important for pathogenicity by controlling arginine biosynthesis in the rice blast fungus. We also examined appressorium formation on an inductive surface, and found appressorium formation was not altered in the ΔMocpa2 mutant (Figure S3B).

To determine why the ΔMocpa2 mutant attenuates pathogenicity, we assayed infectious hyphae growth in rice sheath leaves. Four types of invasive hyphae were observed and analyzed. In the wild-type Guy11 and the complemented strains, over 80% of cells showed type 4 lesions, and fewer than 20% showed type 1 and type 2 lesions. In contrast, 15% of cells showed type 3 and type 4 lesions, and 85% showed type 1 and type 2 lesions in the ΔMocpa2 mutant (Figures 5A,B), indicating that the ΔMocpa2 mutant was defective in infectious hyphae growth. We further added exogenous arginine to conidial suspensions of the mutant, and found that the defect could be greatly restored by 2.5 mM arginine, but not by 0.5 mM arginine (Figures 5A,B). Infectious growth on rice sheathes was also observed by time course. The results showed that the ΔMocpa2 mutant rarely penetrated through the rice cuticle at 24 hpi, and that the invasive hyphae remained restricted to the primary infected cells at 48 and 72 hpi. In contrast, the invasive hyphae of the wild-type and complemented stains extended into neighboring cells under the same conditions (Figure 5C). These results suggest that MoCpa2 plays a critical role in penetration and infectious hyphal growth by affecting arginine biosynthesis, thus affecting pathogenicity.

The large subunit of CPS is responsible for binding to the inhibitor uridine monophosphate (UMP; Bueso et al., 1999). To further test the observed defects of the ΔMocpa2 mutant are due to arginine limitation or as a result total pathway disruption, the wild type Guy11 was treated with different concentrations of UMP. Growth rate of Guy11 was decreased under the treatment of UMP, and completely inhibited on MM plates with 5 mM UMP (Figure 6A). We also examined conidiation and found that on SDC medium with different concentrations of UMP, conidial production of Guy11 was significantly reduced, and was unable to produce conidia under 5 mM UMP (Figure 6B). Further infection assay on barley leaves showed that 0.5 mM UMP could cause defects on virulence of Guy11, when the concentration of UMP increased to 5 mM, the infection ability of Guy11 was completely abolished on barely leaves (Figure 6C). These results indicated that arginine biosynthetic pathway mediated by MoCpa2 is important for vegetative growth, asexual development, and pathogenicity of the rice blast fungus.

Plant defense responses are important during plant-microbe interaction. The recognition of microbe-associated molecules by plant immune systems results in the rapid generation of ROS in the infected cell, accumulation of pathogenesis-related (PR) proteins, and reinforcement of the cell wall (Bradley et al., 1992; Levine et al., 1994; Abramovitch et al., 2003; Tanaka et al., 2003; Torres and Dangl, 2005). Therefore, we speculated that the defects in infectious growth of the mutant partly resulted from the plant defense responses. ROS accumulation was examined by staining with 3,3′-diaminobenzidine (DAB) after inoculation for 48 h. Similar to the results in the wild-type Guy11 and complemented strains, no ROS accumulation was observed at the infection sites of the ΔMocpa2 mutant (Figure 7A). We further evaluated the expression of the PR genes, PR1a, AOS2, and PBZ1, which were previously shown by qRT-PCR analysis to be expressed at their highest levels during plant-pathogen interactions (Chi et al., 2009). The expression levels of the three PR genes did not increase in the ΔMocpa2 mutant compared to the wild-type Guy11 strain (Figure 7B), suggesting that deletion of MoCPA2 did not induce their expression. We also found that the expression of PBZ1 was much lower in rice inoculated with the ΔMocpa2 mutant compared to the wild type Guy11. As the expression of PBZ1 in the ΔMocpa2 mutant showed no significant difference from the mock, indicating that deletion of MoCPA2 did not induce the PBZ1 expression by 48 hpi. Therefore, we concluded that the ΔMocpa2 mutant cannot elicit plant defense responses at early stages of infection, and that the attenuated virulence of the mutant does not due to plant defense responses.

Structural analysis revealed that MoCpa2 contains an ATP grasp domain and two CPSase domains (Figure 8A). To investigate the roles of these three domains in M. oryzae, domain deletion constructs were fused with GFP to produce MoCpa2^ΔATP-GFP (deletion of the ATP domain), MoCpa2^ΔCPS1-GFP (deletion of the CPS1 domain), and MoCpa2^ΔCPS2-GFP (deletion of the CPS2 domain) constructs which were transformed into the ΔMocpa2 mutant. The resulting transformants were evaluated for GFP signals and analyzed. The tranformants expressing MoCpa2^ΔATP-GFP, MoCpa2^ΔCPS1-GFP, and MoCpa2^ΔCPS2-GFP showed growth, conidiation, and virulence phenotypes similar to those of the ΔMocpa2 mutant (Figures 8B–D). Further studies revealed that 2.5 mM exogenous arginine could restore the defects of each transformant (Figures 8B–D), suggesting that each domain is essential for the full function of MoCpa2 in M. oryzae. We also observed the localization of MoCpa2 and the domain deletion transformants. The results showed that MoCpa2 localized to mitochondria in M. orzyae (Figure S4A), and GFP signals of the domain deletion transformants were mainly observed in cytoplasm (Figure S4B), indicating ATP grasp domain and two CPSase domains are indispensable for proper localization of MoCpa2, thereby affecting the growth and development of the rice blast fungus.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.