B. cinerea wild-type strain B05.10 was used in this study. The fungus was grown on potato dextrose agar (200 g potato, 20 g glucose per liter, 2% agar), and incubated for 7–30 days at 20°C Knockout transformants were cultured on PDA amended with 75 μg/mL hygromycin B (Calbiochem, San Diego, CA, USA). Complementary transformants were cultured on PDA amended with 75 μg/mL hygromycin B and 50 μg/mL G418. Escherichia coli strain DH5α was used to propagate all of the plasmids, and Agrobacterium tumefaciens strain EHA105 was used for the transformation of fungi. Seedlings of Arabidopsis thaliana (ecotype Columbia-0) were grown in the greenhouse at 20 ± 2°C for 1 month under a 12 h light/dark cycle with 70% relative humidity.

BcDIM5, BcDIM2, and BcHP1 were knocked out in the WT strain B05.10 using the hygromycin B-resistance gene (Hyg) to replace the partial sequences of BcDIM5, BcDIM2, and BcHP1 (Figure S1). To construct the ΔBcDIM5 disruption vector, a 753-bp DNA fragment named 3U with the EcoRI restriction sites at 5′- and 3′-terminus and a 740-bp DNA fragment named 3D with the EcoRI restriction sites at 5′- and 3′-terminus were PCR-amplified from genomic DNA of strain B05.10. Using pKS1004 vector as the template, the 740-bp HY fragments were amplified by primers HYG-F and HY-R, and the 1133-bp YG fragments were amplified by primers YG-F and HYG-R. The 3U and HY fragments were fused with PCR to obtain a 1465-bp fragment named 3UHY, and 3D and YG fragments were fused with PCR to obtain a 1936-bp fragment named 3DYG. After digestion with enzymes EcoRI, the DNA fragments 3UHY and 3DYG were separately collected and inserted into the linearized plasmid pBluescriptII KS1004 for sequencing. 3UHY and 3DYG were individually linearized and transformed separately into the protoplasts of B. cinerea wild type strain B05.10 using the PEG-mediated protoplast transformation technique (Wei et al., 2013). For complementation assays, the 5.75-kb PCR product containing a 3-kb upstream sequence (which contained the endogenous promoter), a full-length BcDIM5 gene coding region, and a 1-kb downstream sequence was amplified from the wild type strain B05.10 genomic DNA using primers DIM-5-CF-BamHI and DIM-5-CR-BamHI and cloned into the same sites of p3300neoIII to generate the BcDIM5 complementary vector p3300neoIII-BcDIM5, which was then transformed into BcDIM5 knockout transformants using the A. tumefaciens-mediated transformation technique.

The wild-type strain and knockout transformants were cultured on PDA plates covered with cellophane membranes, and mycelia were harvested at 3 days post inoculation (dpi). Genomic DNA was extracted using the CTAB method. Southern blot analysis was conducted following the method described by Wei et al. (2016). The genomic DNA of 15 μg was completely digested with HindIII, separated by electrophoresis on 0.8% agarose gel and transferred onto a Hybond N+ membrane (Amersham Pharmacia Biotech). The probe was amplified the specific fragment of hygromycin hph gene and be labeled by DIG (GE healthcare). The nylon membranes were autoradiographed and analyzed using Bio-imaging analyser BAS-1800II (FUJIFILM, Tokyo, Japan).

Pathogenicity tests of transformed and wild-type B. cinerea strains were performed on Arabidopsis, tomato, soybean by the inoculation of detached leaves with young non-sporulating mycelium or conidial suspensions. Leaves were harvested from 4-week-old plants and placed in a transparent plastic box lined with tissue moistened with sterile water. Leaves were inoculated with 2-mm-diameter plugs of 3-day-old mycelium. Alternatively, conidia were collected from 10-day-old plates and suspended in water to a final concentration of 5 × 10⁵ conidia/mL. Droplets of 5 μL were applied to the leaves. Storage boxes containing inoculated leaves were incubated in a growth cabinet at 20°C with 16 h of daylight. Disease development on leaves was recorded daily as the radial spread from the inoculation point to the lesion margin. Pathogenicity assays on leaves were repeated three times using at least three leaves per assay.

The mRNA transcripts were measured using a SYBR Green I real-time PCR assay in a CFX96 real-time PCR detection systems (Applied Biosystems). The thermal cycling conditions were 95°C for 2 min for predegeneration, 40 cycles of 95°C for 20 s for denaturation, 60°C for 20 s for annealing and 72°C for 20 s for extension. All of the reactions were run in triplicate by monitoring the dissociation curve to control the dimers. The B. cinerea β -tubulin gene was used as reference gene. qRT-PCR was used to examine the expression of B. cinerea Bcsod1 (BC1G_00558), BcSpl1 (BC1G_02163), BcVeA (BC1G_02976), BcBOT2 (BC1G_06357), Bcmp3 (BC1G_07144), BcVELB (BC1G_11858), BcPKS (BC1G_15837), Bcbot1 (BC1G_16381), BcDIM2 (BC1G_12419), BcHP1 (BC1G_06432), and BcDIM5 (BC1G_11188). The primers for qRT-PCR are listed in Table S1.

Mycelia were harvested and ground in liquid nitrogen. Two hundred milligrams powders were resuspended in 400 μL cell lysis buffer on ice (Beyotime, China). Total proteins of mycelium extract were centrifuged at 12,000 rpm at 4°C for 10 min, and the supernatant was used for western-blot analysis. Proteins were separated by 12% PAGE and electroblotted to a nitrocellulose membrane at 25 V for 40 min. The membrane was blocked with Tris-buffered saline plus Tween 20 containing 5% skim milk powder for 2 h at room temperature. After incubation with primary antibody and then with secondary antibody, the membrane was transferred for protein detection using a Thermo SuperSignal West Pico kit (Thermo Scientific). The sources and dilutions of antibodies were as follows: Histone H3 monoclonal antibody (Abmart, at 1: 2000), Histone H3 tri methyl K9 monoclonal antibody (Abmart, at 1: 2000), and goat anti-mouse horseradish peroxidase-conjugated antibody (Abmart, at 1: 5000).

The B. cinerea BC1G_11188 gene (GenBank accession: XM_001550366.1) is a single copy gene consisting of 4 exons and 3 introns and encoding a peptide of 357 amino acid residues. The protein contains the typical domain structures of H3K9 methyltransferase (Figure 1A), including a PreSET domain, a SET domain, and a PostSET domain (Rea et al., 2000; Tamaru and Selker, 2001; Adhvaryu et al., 2005). SET domain contains two conserved sequences of NHXCXPN and DY, which can form an AdoMet binding site and the catalytic activity site of methyl transferase, and participates in the formation of DIM5 hydrophobic structure (Zhang et al., 2002). PreSET domain contains nine conserved cysteine residues that can be combined with three zinc ions to form a zinc cluster. PostSET domain contains three conserved cysteine residues involved in the binding of DIM5 to AdoMet (Figure 1B). Similar to the histone lysine methyltransferases of NcDIM5, SpClr4, HsSUVH1, HsSUVH2, and AtSUVH4, BC1G_11188 includes cysteine-rich sequences that flank a SET domain (Figure 1B). The results of Bioinformatics Toolkit HHpred analysis indicate that the structure of BC1G_11188 is similar to the three-dimensional structure of histone lysine n-methyltransferase (Template: c4qeoA, Confidence: 100%; Figure 1C). In summary, the protein coded by BC1G_11188 resembles histone H3 lysine 9 methyltransferase. Thus, we designated this gene which was derived from the homologs of DIM5 in N. crassa as “BcDIM5.”

To examine the function of BcDIM5, the BcDIM5 knockout transformants were generated through homologous recombination of the BcDIM5 open reading frame with a gene conferring hygromycin resistance (HYG) (Figure S1A). Positive knockout transformants were amplified with PCR (Figure S1B), and three BcDIM5 knockout transformants (21, 24, and 25) were identified. Reverse transcription PCR was employed to detect the BcDIM5 gene expression. The results showed that there were no transcripts in the three knockout transformants (Figure S1C). Two transformants (ΔBcDIM5-21 and ΔBcDIM5-24) were further confirmed by Southern blot analysis (Figure S1D). The mutant ΔBcDIM5-21 was used for phenotypic analysis and complemented with the full-length gene of wild typeBcDIM5. After being verified by PCR, the complemented strain BcDIM5-C1 was chosen for further study (Figure S1E).

To determine the role of BcDIM5 in the growth and development of B. cinerea, we compared the hyphal growth rate, the number of conidiophores and dry weight of sclerotia of the wild-type strain, ΔBcDIM5-21 strain and the BcDIM5-C1 complemented strain. Although BcDIM5 did not affect the morphology of tip hyphae (Figure S2), the hyphal growth rate was significantly reduced in ΔBcDIM5-21 strains, while this phenotype was restored after complementation (Figure 2D). BcDIM5 affected the development of conidiophores and sclerotia, and the wild-type strain could averagely produce 2 × 10⁷ conidiophores per plate, while ΔBcDIM5-21 strains only produced 3 × 10⁶ conidiophores per plate, and the conidiophore production of BcDIM5-C1 complemented strain was restored to 1.1 × 10⁷ conidiophores per plate (Figures 2A,E). Although ΔBcDIM5-21 strain was able to produce conidiophores, the time required for conidiophore initiation was about 3–4 days longer than that for the normal strain. The dry weight of sclerotia produced by wild-type B. cinerea was 217 mg per plate, while ΔBcDIM5-21 strains only produced 57 mg sclerotia per plate, and the sclerotia was small and round. The sclerotia production of BcDIM5-C1 complemented strain was restored to 200 mg per plate, which was similar to that of the wild type (Figures 2B,C,F). These data clearly suggest that BcDIM5 knockout has a deleterious effect on hyphal growth and the development of conidiophores and sclerotia, meanwhile, the reduced hyphal growth rate of ΔBcDIM5-21 strains may delay conidiophore development and the sclerotia formation.

To determine the role of BcDIM5 in pathogenicity, detached Arabidopsis leaves were inoculated with conidia of the wild-type strain, ΔBcDIM5-21 strain and BcDIM5-C1 complemented strain. At 2 dpi, the inoculation of the wild-type strain B05.10 caused the formation of necrotic lesions with 5–6 mm diameter, and the leaves completely rotted away at 5 dpi (Figure 3). However, the leaves inoculated with the ΔBcDIM5-21 strains exhibited no necrotic lesions at 2 dpi, and only slight necrotic lesions in the infection site at 5 dpi. The pathogenicity of the BcDIM5-C1 complemented strain was only partially restored, as the diameter of the necrotic lesions was half of that induced by wild-type strain (Figure 3). We further used mycelium pieces to inoculate soybean and tomato leaves. Similar results were observed in ΔBcDIM5-21 and ΔBcDIM5-24 mutants. The virulence of the BcDIM5 knockout transformants was significantly reduced, and only small lesions were observed on the detached leaves of soybean and tomato (Figure S3), indicating that the virulence reduction is not host species-specific. These results indicate that BcDIM5 plays crucial roles in the virulence.

To determine whether BcDIM5 regulates the expression of genes associated with pathogenesis, qRT-PCR was performed to analyze the expression of eight genes involved in the pathogenic process of B. cinerea based on previous studies. These eight genes included a Cu-Zn-superoxide dismutase gene Bcsod1, which is slightly increased by 0.5 mM H₂O₂ and significantly increased by 2.0 mM H₂O₂ (Rolke et al., 2004); a cerato-platanin family protein gene BcSpl1, which was induced in tobacco systemic resistance to two plant pathogens (Frías et al., 2013); two velvet-like genes BcVEA and BcVELB, which act as negative regulators for conidiation and melanin biosynthesis in B. cinerea (Yang et al., 2013); a presilphiperfolan-8 beta-alpha synthase gene BcBOT2, which encode the sesquiterpene synthase that is responsible for the committed step in the biosynthesis of botrydial (Wang et al., 2009); a mitogen-activated protein kinase gene Bcmp3, which is required for normal saprotrophic growth, conidiation, plant surface sensing and host tissue colonization (Rui and Hahn, 2007), a polyketide synthase gene BcPKS (Dalmais et al., 2011) and a cytochrome P450 monooxygenase gene Bcbot1 (Siewers et al., 2005). Expression analyses revealed that all of these eight genes were significantly down-regulated in the BcDIM5 knockout transformants compared with in the wild type (Figure 4). These data indicate that many pathogenic genes may be regulated by BcDIM5.

To investigate the possible effects of DIM5 on the trimethylation of H3K9, we carried out western-blot to analyze the histone methylation level of the wild-type strain, ΔBcDIM5-21 and BcDIM5-C1 (Figure 5). All the strains presented equally robust signals when detected using the antibody histone H3. Notably, trimethylated H3K9 was detectable in bulk from the wild-type strain and BcDIM5-C1, while knockout transformant ΔBcDIM5-21 extinguished the trimethylated H3K9 signal, suggesting that BcDIM5 is dominantly responsible for the trimethylation of H3K9, and DIM5 may generate trimethylated H3K9 from unmodified histone H3 in vivo.

In N. crassa, DIM5 trimethylates H3K9, and H3K9me3 directs DNA methylation through a complex containing heterochromatin protein 1 (HP1) and the DNA methyltransferase DIM2. We also investigated whether DIM2 and HP1 play important roles in B. cinerea development and virulence. Based on the homology to N. crassa DIM2 and HP1, we cloned BcDIM2 (BC1G_12419) and BcHP1 (BC1G_06432) from B. cinerea. BcDIM2 contains typical domain structures of DNA methyltransferase and BcHP1 contains typical domain structures of CD (chromo domain) and CSD (chromo shadow domain; Paro and Hogness, 1991; Aasland and Stewart, 1995). RT–PCR analysis showed that BcDIM5, BcDIM2, and BcHP1 were all expressed in mycelium growth, sclerotial development and infection stages (Figure S4). BcDIM5 was constitutively expressed in these three stages, and BcDIM2 was significantly up-regulated during the infection stage, while BcHP1 was significantly down-regulated during the sclerotial development. Through homologous recombination methods, we got two knockout transformants ΔBcDIM2-16 and ΔBcHP1-7. There were no significant changes in the phenotype and number of produced conidiophores as well as in hyphal growth rate, dry weight of sclerotia and virulence among the wild-type, ΔBcDIM2-16 and ΔBcHP1-7 strains (Figure 6). Therefore, it can be concluded that BcDIM5 is required for the development and virulence whereas BcDIM2 and BcHP1 are not required for virulence in B. cinerea.

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.