Curvularia lunata WT strain CX-3, whose genome sequence is available, was used in this study (Gao et al., 2014). The Clpks18 deletion mutant (ΔClpks18) was offered by the Ministry of Agriculture Key Laboratory of Urban Agriculture (South) (MAKLUAS), Shanghai Jiao Tong University. Zea mays cultivars (ZHENGDAN-958) were used for virulence assays. The pCAMBIA1300 vector that contains the G418 resistance cassette is offered by MAKLUAS. This vector was used for complementation and overexpression of Clpks18 gene. Escherichia coli strain DH-5α was used to construct the vectors. Agrobacterium tumefaciens AGL-1 was used for transformation.

The pCAMBIA1300 vector that contains the G418 resistance cassette comprising the G418 resistance gene under the control of TrpC promoter and terminator from Aspergillus nidulans was used for gene overexpression. The full-length sequence of Clpks18 flanking with the promoter and terminator of TrpC was inserted into the HindIII-EcolI sites of pCAMBIA1300 to create plasmid pCAMBIA1300-C Clpks18 using an infusion kit (In-Fusion HD Cloning kits, Takara, Code: 639648). The plasmid carrying both the Clpks18 and the G418 resistance cassette was used to transform into the ΔClpks18 to generate the Clpks18-overexpression strain using the Agrobacterium tumefaciens-mediated transformation. The Clpks18 gene-complementation mutants were constructed as mentioned above, except changing the TrpC promoter to the promoter of C. lunata itself. The integration at the target sites and the complementation of the Clpks18 mutant were confirmed using the PCR. The sequences of primers for gene disruption, complementation, and PCR confirmation are shown in Supplementary Table 1. The overexpression of the Clpks18 mutants was obtained by the methods described above, except the plasmid Clpks18 flanking with TrpC promoter and the TrpC terminator.

Each strain was cultured in potato dextrose broth at 180 rpm at 28°C for 3–10 days. The growth rate of each strain was observed during 3–10 days. The conidia were harvested using sterile water to count on the 10th day. Each experiment was replicated three times.

The leaves of the maize ZHENGDAN-958 seedlings at the five-leaf stage were inoculated by smearing with conidial suspensions (1.0 × 10⁶ conidia/ml) with 0.05% Tween-20 (each leaf was smeared by 100 μl conidial suspensions). The maize seeds inoculated with C. lunata were grown in the greenhouse with 85–95% air humidity for 2 days. The inoculated leaves were collected on 1st, 3rd, 5th, and 7th day to observe the infection. The lesions area covered per unit area was calculated. Then, the maize leaves were decolored, and the hyphae were dyed for observation and calculation. At 24 hpi, the single conidium expansion area was calculated. At 120 hpi, the hyphae expansion area per unit area was calculated (240 μm × 182 μm, 43,680 μm²). All those tests were performed in triplicates.

M5HF2C toxin and (R)-(-)-mellein virulence assays were done as mentioned above. The five-leaf stage maize leaves inoculated by ΔClpks18 strain conidial suspension with M5HF2C toxin and (R)-(-)-mellein at different concentrations (M5HF2C toxin: 60–240 ng/100 μl; (R)-(-)-mellein: 80–320 ng/100 μl). At 24 hpi, the membrane injury of the maize leaves was measured by relative conductivity.

The infected leaves were collected at different time points. Then, they were sectioned, fixed, and stained for microscopic observation. The samples were observed using an optical microscope (Leica DM2500 Microscope and Leica DM6000 Microscope, Germany) and 120 kV Biology transmission electron microscope (Thermo Fisher Scientific, Tecnai G2 SpiritBiotwin, Czech).

For microscopic observation, inoculated leaves were immersed in 5 ml of alcoholic lactophenol [1 volume of phenol: glycerol: lactic acid: water (1:1:1:1) and 2 volumes of ethanol]. The leaves were evacuated for 15 min and then incubated at 65°C until the disappearance of chlorophyll pigment. The leaves were stained with 0.5% toluidine blue and decolorized using saturated chloral hydrate solution for an additional 2–24 h. Finally, the samples were stored in 50% glycerol for an optical microscope (Leica DM2500 Microscope). For transmission electron microscopy, the leaves were fixed with 2.5% glutaraldehyde in PBS (50 mM, pH 7.2) under vacuum infiltration until all the leaves sank to the bottom of the tube and stored at 4°C overnight. The tissues were then washed with 50 mM PBS and fixed in 1% osmic acid in PBS (50 mM, pH 7.2) at 4°C for 2 h. After three washes with 50 mM PBS, the tissues were dehydrated in an ethanol series (60–100%) and embedded in resin for 48 h. After finishing embedding, the samples were sectioned using Ultramicrotome (Leica EM UC7, Germany). Some samples were stained with toluidine blue for optical microscope (Leica DM6000 Microscope); the other samples were stained with uranyl acetate for 5 min and lead citrate for 5 min; and the grids were observed under the transmission electron microscope (Thermo Fisher Scientific, Tecnai G2 SpiritBiotwin) at 120 kV.

Each mutant and WT strain was cultured in PDB for 3 days and in Fries 3 media for 15 days. Total RNA of these strains was extracted using the TRIzol reagent (TRIzol™, Invitrogen, Code: 15596026, Unites States). The cDNA was synthesized using 1 mg RNA using the PrimeScript RT Reagent kit with gDNA Eraser (code: DRR047A; Takara, Japan). Real-time quantitative PCR (RT-qPCR) was carried out using a Hieff^® qPCR SYBR Green Master Mix (No Rox) kit (code:11201ES08; YEASEN, China). The PCR condition was described as mentioned in the kit above using a Light Cycler 96 Real-Time PCR System (Roch, United States). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene of C. lunata was used as the reference internal control for analyzing the relative expression levels using 2^–ΔΔCT method. The qPCR primers used in this study are shown in Supplementary Table 1. Each experiment was carried out in triplicates.

Wild-type and mutant strains were cultured in Fries 3 media (Supplementary Table 4) at 28°C, 180 rpm, for 15 days to extract metabolite. The extracellular metabolites were extracted using the fermented culture filtrates. The intracellular metabolites were extracted from cells homogenized with liquid nitrogen pestle. The method of metabolic extraction was shown in Supplementary Information (Smart et al., 2010). The extracted metabolites were analyzed using LC-MS (Sangster et al., 2006; Want et al., 2010). Chromatographic separation of the extracted metabolites was accomplished in an Acquity UPLC system equipped with an ACQUITY UPLC HSS T3 (150 × 2.1 mm, 1.8 μm, Waters) column maintained at 40°C. The temperature of the autosampler was 4°C. Gradient elution of analytes was carried out with 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) at a flow rate of 0.25 ml/min; 10 μl of each sample was injected after equilibration. An increasing linear gradient of solvent B (v/v) was used as follows: 0∼1 min, 2% B, 1–9.5 min, 2–50% B, 9.5–14 min, 50–98% B, 14–15 min, 98% B, 15–15.5 min, 98–2% B, 15.5–17 min, 2%. The ESI-MSn experiments were carried out using Thermo LTQ-Orbitrap XL mass spectrometer with a spray voltage of 4.8 and −4.5 kV in positive and negative modes, respectively. Sheath gas and auxiliary gas were set at 45 and 15 arbitrary units, respectively. The capillary temperature was maintained at 325°C. The voltages of capillary and tube were 35 and 50 V, and −15 and −50 V in the positive and negative modes, respectively. The Orbitrap analyzer scanned over a mass range of 89–1,000 m/z for the full scan at a mass resolution of 60,000. Data-dependent acquisition (DDA) MS/MS experiments were performed with a CID scan. The normalized collision energy was set at 30 eV. Dynamic exclusion was implemented with a repeat count of 2 and an exclusion duration of 15 s.

The aggregation level clustering was used in this experiment. Before the multivariate statistical analysis, the data were processed by mean-centering and scaled to unit variance conversion. Multivariate statistical analyses, such as principal component analysis (PCA), partial least squares-discriminant analysis (PLS-DA), and orthogonal partial least squares discriminant analysis (OPLS-DA), were used to analyze the results.

The mutants and the WT strains were cultured in Fries 3 medium for 15 days under the same conditions described in the metabolomic analysis to check the M5HF2C toxin production. The M5HF2C toxin and (R)-(-)-mellein in the maize leaves were extracted using methanol, H₂O, and formic acid (70:29:1) solution; dried using a vacuum rotary evaporator; and redissolved in methanol for UPLC-3Q-MS analysis. The extracts were analyzed using an ultra-high-performance liquid chromatography triple quadrupole mass spectrometer system (UHPLC 30A/Sciex Quadrupole 5500, AB Sciex) to determine the quantity of M5HF2C.

Clpks18 gene is essential for the melanization of hyphae and conidia in C. lunata. The ΔClpks18 strain resulted in the albino mycelia and conidia. The DHN-melanin production was recovered in the Clpks18 complementation strain (Figures 1A,B). The melanization was increased in the Clpks18 overexpression strain (Supplementary Figure 1). The expression of Clpks18 gene in the ΔClpks18-C strain was lower than the WT in the PDA medium (Figures 1C,D). Similarly, the melanization of mycelia and conidia was lower than the WT (Supplementary Figure 1). The melanization was positively correlated with the expression of the Clpks18 gene. The ΔClpks18 and ΔClpks18-C strains produced dense aerial mycelia compared with the WT and OE-Clpks18 strains. The aerial mycelia of WT and OE-Clpks18 strain were reduced during conidial production (Figure 1A and Supplementary Figure 1). The production of conidia in ΔClpks18 strain was reduced compared with the WT. The conidia of ΔClpks18-C strain were also reduced compared with the WT. The overexpression of Clpks18 increased the production of conidia (Table 1). However, the growth of ΔClpks18 and ΔClpks18-C was faster than that of WT and OE-Clpks18 (Figure 1E). It revealed that the lower expression of Clpks18 induces the C. lunata to grow faster and produce the aerial hyphae. On the contrary, the highest expression of Clpks18 induces the C. lunata to produce more conidia. In Fries 3 medium, the expression of Clpks18 in the WT strain was reduced compared with the OE-Clpks18 strain (Figure 1G). So, it revealed that the hyphal melanization of the WT strain was slower than that of the OE-Clpks18 strain (Figure 1F).

Deletion mutant of Clpks18 reduces virulence to maize compared with the WT and OE-Clpks18 strains. The lesions area of the maize leaves inoculated with WT and OE-Clpks18 strains at 7 days post-inoculation (dpi) was higher than the leaves inoculated with ΔClpks18 strains, which demonstrates that Clpks18 was important to the virulence (Figures 2A,B). The conidial germination on the surface of the maize leaves was observed at 24 h post-inoculation (hpi) to validate the role of Clpks18 gene on the infection. The results showed that the conidial germination was normal in all strains. The main differences observed between different strains were the accumulation of DHN-melanin in OE-Clpks18 and WT strains, which was higher than that in the ΔClpks18 strains. However, the hyphae could directly penetrate the cuticle on the leaf surface and grow in the intercellular spaces of the leaf cells (Figure 3A). The hyphae extended between the intercellular spaces of leaves on 5 dpi (120 hpi) (Figures 3A,B). The leaves inoculated with ΔClpks18 strain showed the presence of hyphae in the intercellular spaces of cells without much marked lesions. However, the quantity of hyphae is much lesser than the WT and OE-Clpks18 strains. Meanwhile, the hyphae of ΔClpks18 strain are slenderer than the WT and OE-Clpks18 strains (Figure 3B and Supplementary Figure 2). The results revealed that the Clpks18 gene was important for the pathogenicity of C. lunata.

To compare the secondary metabolism of the WT and mutants, they were cultivated in the Fries 3 medium. The relative expression of Clpks18 was significantly higher in the OE-Clpks18 strain than the WT strain in the Fries 3 medium compared with the PDA medium (Figures 1D,G). The relative expression of Clpks18 in OE-Clpks18 strain was 3.57 times higher than that of the WT strain, which coincides with the faster mycelial melanization of OE-Clpks18 strain compared with the WT strain (Figure 1F). At the same time, Clpks18 was also related to the production of M5HF2C toxin. The expression of the Clpks18 gene was positively correlated with the M5HF2C toxin production (Figure 4C and Supplementary Figure 3). To further explore the effect of Clpks18 on other secondary metabolic pathways, the metabolomic analysis of extracellular and intracellular metabolic contents of WT strains, ΔClpks18 strains, and OE-Clpks18 strains was performed; 98 metabolites significantly differed between the WT, ΔClpks18, and OE-Clpks18 strains (Figure 5A); 119 intracellular metabolites were identified in the WT, ΔClpks18, and OE-Clpks18 strains (Figure 5B). The enrichment analysis of intracellular metabolites showed that 16 pathways were obviously changed inside the cells (Supplementary Table 2); 7 pathways were found to be changed markedly in the extracellular secretion (Supplementary Table 3), which include the glycerolipid metabolism. We found that glycerin was extracellularly secreted into the medium of δClpks18 and OE-Clpks18 mutant than the WT (Figure 4B and Supplementary Figure 4). OE-Clpks18 mutants formed hyper-melanized mycelia and conidia; they might produce more glycerin in the hyphae than the WT strain, which generated higher turgor pressure in the hyphae than WT strains. Since the turgor pressure in the hyphae of OE-Clpks18 mutants is higher than the threshold value, they should release the glycerin into the extracellular space to reduce the excrescent turgor pressure. The lack of melanin in the cell wall of the ΔClpks18 mutants could not maintain the high turgor pressure of cells. In order to reduce the turgor pressure of ΔClpks18 mutants, they released more glycerin into the extracellular space to adapt themselves to turgor pressure which was lower than the WT strain. The extracellular and intracellular metabolite analysis showed that a compound identified as (R)-(-)-mellein (Supplementary Figure 5) was highly correlated with the relative expression of the Clpks18 gene (Tables 2, 3). The (R)-(-)-mellein was not detected in the extracellular secretion of ΔClpks18 mutants, while the (R)-(-)-mellein production was restored and secreted into the extracellular environment of the complemented Clpks18 (Figure 4A). The short-chain peptides such as His-Nap-OH, Tyr-Lys-OH, and N-Ac-Tyr-Val-Ala-Asp-CHO were positively correlated with the expression of Clpks18 gene extracellularly (Figure 4D and Supplementary Figure 4). In addition, the non-ribosomal peptide synthases (NRPS) were influenced by the Clpks18 gene. It was supposed that M5HF2C toxin, (R)-(-)-mellein, and short-chain peptides were the candidates for virulence factors.

To confirm the key virulence factor among the two, namely, M5HF2C toxin and (R)-(-)-mellein, they were detected in the maize leaves infected by the WT, ΔClpks18, and OE-Clpks18 strains at 3 dpi (72 hpi). It was found that the M5HF2C toxin (50.03 ng/g) and (R)-(-)-mellein (85.05 ng/g) were secreted into the host cells (Figure 6A and Supplementary Figures 6, 7). To verify which of (R)-(-)-mellein and M5HF2C is the more important pathogenic factor for C. lunata, these two compounds were used for the virulence assays. When ΔClpks18 conidial suspension is inoculated into maize leaves with (R)-(-)-mellein at a concentration of 800 ng/ml, the pathogenicity of ΔClpks18 mutant is largely restored to the level of WT. However, as ΔClpks18 conidial suspension is inoculated into maize leaves with M5HF2C toxin at a concentration of 600 ng/ml, the pathogenicity of ΔClpks18 mutant could not be restored to the level of WT (Figures 6B,C). At 24 hpi., the maize leaves infected by the WT, ΔClpks18, OE-Clpks18, and ΔClpks18 strains with different concentrations of M5HF2C toxin and (R)-(-)-mellein and fermentation of WT in liquor were collected for the relative conductivity test. The relative conductivity test indicated the membrane injury of the maize leaves. It showed that the ΔClpks18 in the early stages of infection could not cause obvious injury on the membrane of maize leaves cells. The leaves infected by WT, OE-Clpks18, and ΔClpks18 with different concentrations of (R)-(-)-mellein and fermentation of WT in liquor showed significant differences with blank control. At the same time, the single treatment with 160 ng/ml (R)-(-)-mellein can cause the relative conductivity to increase when compared with the blank control. The relative conductivity does not increase with the increase in the treatment concentration of (R)-(-)-mellein. The results revealed that (R)-(-)-mellein plays a key role in the early stages of C. lunata infection. Only as the ΔClpks18 conidia combine infection with the fermentation of WT in liquor, could the level of damage to the host cells’ membrane was restored to the level of WT infection. It implied that the fermentation of WT in liquor contains some substances that play a role in C. lunata infection except (R)-(-)-mellein (Figure 6D). Also, it suggests that the (R)-(-)-mellein secretion was correlated with the expression of Clpks18 gene. High hyphal content was observed in the phloem of maize leaves infected with WT and OE-Clpks18 strains at 5 days (at 120 hpi) using the transmission electron microscope. However, hyphal content in the phloem of maize leaves infected with the ΔClpks18 was very lower than the WT (Figure 6E).