Transcriptome analysis and functional validation reveal a novel gene, BcCGF1, that enhances fungal virulence by promoting infection‐related development and host penetration

Abstract Simultaneous transcriptome analyses of both host plants and pathogens, and functional validation of the identified differentially expressed genes (DEGs) allow us to better understand the mechanisms underlying their interactions. Here, we analyse the mixed transcriptome derived from Botrytis cinerea (the causal agent of grey mould) infected tomato leaves at 24 hr after inoculation, a critical time point at which the pathogen has penetrated and developed in the leaf epidermis, whereas necrotic symptoms have not yet appeared. Our analyses identified a complex network of genes involved in the tomato–B. cinerea interaction. The expression of fungal transcripts encoding candidate effectors, enzymes for secondary metabolite biosynthesis, hormone and reactive oxygen species (ROS) production, and autophagy‐related proteins was up‐regulated, suggesting that these genes may be involved in the initial infection processes. Specifically, tomato genes involved in phytoalexin production, stress responses, ATP‐binding cassette transporters, pathogenesis‐related proteins, and WRKY DNA‐binding transcription factors were up‐regulated. We functionally investigated several B. cinerea DEGs via gene replacement and pathogenicity assays, and demonstrated that BcCGF1 was a novel virulence‐associated factor that mediates fungal development and virulence via regulation of conidial germination, conidiation, infection structure formation, host penetration, and stress adaptation. The fungal infection‐related development was controlled by BcCGF‐mediated ROS production and exogenous cAMP restored the mutant infection‐related development. Our findings provide new insights into the elucidation of the simultaneous tactics of pathogen attack and host defence. Our systematic elucidation of BcCGF1 in mediating fungal pathogenesis may open up new targets for fungal disease control.


| INTRODUC TI ON
The necrotrophic fungal pathogen Botrytis cinerea causes grey mould on over 1,000 plant species (Fillinger and Elad, 2016). The pathogen infects almost all vegetable and fruit crops, including numerous economically important crops such as grapevine, strawberry, and raspberry, and annually causes enormous economic losses worldwide.
The pathogen attacks many plant organs, including stems, fruits, leaves, and flowers, at both pre-and post-harvest stages (Dean et al., 2012). In the field, the main infection source of this pathogen is conidia that form appressoria or appressorium-like structures on host surfaces to facilitate infection (Gourgues et al., 2004;Choquer et al., 2007). In addition, B. cinerea is the most extensively studied necrotrophic fungal pathogen due to the availability of its genome sequence, its ease of study (e.g., easy to obtain gene knockout mutants or achieve gene silencing), as well as its economic relevance (Dean et al., 2012).
B. cinerea employs a range of toxic molecules to kill and decompose plant tissue prior to converting it into fungal biomass (Williamson et al., 2007). To initiate host infection, B. cinerea conidia on plant surfaces germinate and form appressoria for host penetration. Conidia of the pathogen hardly germinate in the presence of only water. Therefore, gluconeogenesis and glucose play a crucial role in the initiation of conidial germination; gluconeogenesis allows the pathogen to cope with the limitation of glucose and/or other carbon sources in the infection niches (Liu et al., 2018). Shortly after conidial germination, the germ tubes cease their polarized cell growth and start to form swollen melanized structures called appressoria. Besides appressoria, B. cinerea also forms highly melanized specialized hyphal networks or clumps of hyphal structures called infection cushions (Marschall and Tudzynski, 2016b). Infection structures (IFSs), including appressoria and infection cushions, are required for the pathogen to penetrate host cells. The fungal septin protein Sep4 is essential for pathogens to initiate IFS formation and host penetration (Feng et al., 2017). Plant cell walls are barriers that impair pathogen host colonization; however, they are also important reservoirs of energy-rich sugars. To invade plant cells and to degrade or exploit cell wall components, B. cinerea secretes enzymes that disassemble cell wall polysaccharides during infection (Blanco-Ulate et al., 2014). Interestingly, to disarm host resistance, the fungus also employs small RNAs as effectors that are delivered to the host plant cells where they can be loaded by the RNA silencing machinery and target plant defence genes (Weiberg et al., 2013). Although knowledge about B. cinerea pathogenesis has greatly expanded in the last two decades, mechanisms of pathogen factors that mediate the fungal pathogenesis and/or disarm host resistance remain obscure.
Once a pathogen is detected, the host plant activates signalling networks through the generation of small signalling molecules and coordination of hormonal signalling pathways to initiate defence mechanisms to the pathogen attack (AbuQamar et al., 2017). Plant responses to B. cinerea infection include host cell death, production of various secondary metabolites, antimicrobial peptides, and hormones, for example ethylene (ET), salicylic acid (SA), abscisic acid (ABA), and jasmonate (JA), as well as accumulation of reactive oxygen species (ROS), callose, and a variety of other cell wall modifications (Mengiste, 2012). ABA, ET, and auxin, are known to be actively involved in plant defence against B. cinerea (Navarro et al., 2006;El Oirdi et al., 2011;Windram et al., 2012;Sivakumaran et al., 2016). Other phytohormones, for example brassinosteroids, also regulate plant immunity by mainly interacting with transcription factors, or through camalexin biosynthesis and callose deposition.
In addition, studies have notably provided evidence for cross-talk among SA, JA, and ET or with other hormones in regulating plant defence responses to B. cinerea (for review, see AbuQamar et al., 2017).
Long non-coding RNAs (lncRNAs) play important roles in gene expression and silencing pathways for several biological processes in eukaryotes. Host lncRNAs also play roles in plant defence. In response to pathogen infection, lncRNA-dependent immune systems protect host plants via lncRNAs' regulation of pathogen-associated molecular patterns (PAMPs) and other effectors (Zaynab et al., 2018). Arabidopsis cells secrete exosome-like extracellular vesicles to deliver small RNAs (sRNAs) into B. cinerea to silence fungal genes critical for pathogenicity, indicating that Arabidopsis has employed exosome-mediated cross-kingdom RNA interference as part of its immune responses during the evolutionary arms race with the pathogen (Cai et al., 2018). Despite some understanding of the antifungal mechanisms, more extensive studies on plant resistance to the pathogen are necessary, especially on host plants that are internally colonized by aggressive B. cinerea strains without displaying any signs of disease or stress (Shaw et al., 2016;Veloso and van Kan, 2018).
In B. cinerea pathosystems, host plant and/or pathogen transcriptomes have been investigated in Arabidopsis (Mulema and Denby, 2012), tomato (Blanco-Ulate et al., 2013;Smith et al., 2014;Vega et al., 2015;Rezzonico et al., 2017), lettuce (De Cremer et al., 2013), and cucumber (Kong et al., 2015). These reports greatly increase our understanding of the mechanisms underlying B. cinerea interaction with its diverse hosts. However, plant or pathogen factors identified in the transcriptome analysis during their interactions remain to be functionally validated.
In this study, we investigated the simultaneous gene expression profiles of B. cinerea and tomato and identified a complex network of differentially expressed genes (DEGs) involved in the early stage of the B. cinerea-tomato interaction. We also functionally investigated several up-regulated fungal genes and demonstrated that BcCGF1 is a novel B. cinerea factor that mediates the process of fungal development and infection via facilitating ROS production. Our work provides new insights into the molecular mechanisms underlying tomato-B. cinerea interaction at the early stage, which may be crucial to protect plants from damage caused by necrotrophic fungal pathogens.

| DEGs in tomato and B. cinerea at the early interacting stage (24 hr post-inoculation)
Transcriptomic analyses of B. cinerea-infected tomato leaves as well as plant and pathogen controls (uninfected tomato plants and B. cinerea B05.10 cultured in ½ × potato dextrose broth, PDB) generated a total of 83,128,534 raw reads. After removing the adapter sequences and low-quality reads, 70,186,664 clean reads remained (Table S1). From these, 96.2% and 84.4% of the clean reads in control and infected tomato, respectively, were mapped to the reference tomato genome. Likewise, 95.2% and 6.1% of the clean reads in B. cinerea at the noninfection and infection stages, respectively, were mapped to the B. cinerea genome (Tables S1 and S2 Tables S7 and S8) and KEGG pathways ( Figure S2, and Tables S9 and S10) of the DEGs were similar to those previously reported (De Cremer et al., 2013;Kong et al., 2015;Rezzonico et al., 2017).

| Characteristics of up-regulated genes in B. cinerea-infected tomato leaves
To understand early plant responses to B. cinerea infection, we selected some up-regulated tomato genes for further functional annotation (Tables 1 and S3). Presumably, these transcripts play important roles in tomato plants against B. cinerea infection.

| Genes related to ATP-binding cassette transporters
ATP-binding cassette (ABC) transporters are involved in improving plant immunity (Mulema and Denby, 2012

| Genes associated with autophagy machinery and WRKY DNA-binding transcription factors
Autophagy is a mechanism of the cell that disassembles and recycles unnecessary or dysfunctional cellular components. The up-regulation of an autophagy-related (ATG) gene SlATG8 (Solyc08g078820.2) (

| Genes related to PAMP receptors, PR proteins, and other factors involved in host defence
Fungal cell wall components such as oligosaccharides and chitin are fungal PAMPs that are recognized by host plant receptors and activate defence responses. Fungal polygalacturonases (PGs) are important components essential for virulence and detected by host plants (Windram et al., 2012). The tomato chitin elicitor receptor kinase gene Solyc04g072000.2 (LOC101256086) that encodes chitinase (Liu et al., 2012), and genes Solyc02g070910.1 (LOC101259509) and Solyc07g006770.2 (LOC101243723) encoding flagellin-sensitive

| Characteristics of up-regulated genes in B. cinerea at early stage of infection
To understand the possible roles of pathogen genes in host infection, we characterized some up-regulated fungal DEGs, including BCIN_12g01020 encoding oxaloacetate acetylhydrolase (OAH) and BCIN_03g01540 encoding choline dehydrogenase (Tables 2 and   S5). The pathogen may orchestrate these DEGs to facilitate host infection.

ROS production
Cytochrome P450 monooxygenases are essential for ABA biosynthesis in B. cinerea (Siewers et al., 2004). ABA decreases tomato resistance to B. cinerea via reduction of NO production, which also suppresses both ROS and ET production (Sivakumaran et al., 2016). Two P450 monooxygenase genes (BCIN_07g05430 and BCIN_03g06490) involved in ABA biosynthesis in B. cinerea were upregulated (Table S5). NADPH oxidases are required for ROS production and are involved in conidial germination, differentiation, and vegetative and pathogenicity development as well as virulence in B. cinerea (Marschall and Tudzynski, 2016a;Cao et al., 2018;Hou et al., 2020). NADPH oxidase genes BcNOXA (BCIN_05g00350) and BcNOXB (BCIN_02g04930) were also up-regulated (Table S5).

F I G U R E 1
The protein-protein interaction (PPI) networks of tomato and Botrytis cinerea in the early stage of their interaction. The PPI networks of tomato ATP-binding cassette (ABC) transporters (a) and B. cinerea autophagy-related proteins (b) in the early stage of tomato-B. cinerea interaction. The proteins with green edges are encoded by the corresponding up-regulated tomato ABC transporter genes (a) or B. cinerea autophagy-related genes (b). The other interaction proteins in the networks are from STRING (https://strin g-db.org/), Uniprot (https://www.unipr ot.org/), and NCBI (https://www.ncbi.nlm.nih.gov/). For the size of the circles, the bigger the circles, the more connecting strings/lines or interacting proteins the circles have. For the filled colours in the circles, the more connecting lines, the closer the colour to orange (ranging from yellow to orange). For the colours of the connecting lines, red indicates that proteins directly interact with the interested proteins (nodes with green edges) and black indicates other relationships of interactions. The bold connecting lines indicate that the combined score (0-1, the higher the score, the more reliable the interaction relationship) is greater than 0.95

| Genes encoding candidate effectors
B. cinerea effectors play important roles in establishment of infection via suppression of plant innate immunity (Weiberg et al., 2013;Heard et al., 2015). Through prediction of secretory signal peptide and the cellular localization of protein encoded by each up-regulated fungal transcript, we identified a larger number of transcripts (88/621) encoding putative secreted proteins that may be B. cinerea candidate effectors (Table S13). Among them, several candidate effector genes were highly up-regulated, for example the expression levels of BCIN_16g01820 and BCIN_10g05620, encoding a hypothetic protein and a pectate lyase, respectively, were over 150-and 60-fold of that of the control (Table 2).

| Functional analysis of up-regulated B. cinerea genes
Each DEG with unknown function may constitute a potential entry point for investigating the novel mechanism by which the pathogen secures a successful infection. To investigate gene function at early stage of infection, we representatively selected several up-regulated genes (with known and unknown function) and performed quantitative reverse transcription PCR (RT-qPCR) to confirm their expression levels (Figure 2a). We then generated gene knockout (KO) mutants as previously described (Feng et al., 2017) and conducted pathogenicity assays. Consistent with previous findings (Schumacher et al., 2015;Ren et al., 2017), the mutants lacking BcATG1, BcLAE1, or BcVEL1 lost their virulence ( Figure S5). However, loss of individ-  (Figure 2c,d), suggesting that the factor is required for a successful infection.

| BcCGF1 is a novel virulence-associated gene required for virulence of B. cinerea
Among the fungal DEGs, we were particularly intrigued with these novel DEGs; BCIN_16g01820, encoding a hypothetical protein (Table 2), represented one such DEG because it had not been previously implicated as a factor that mediates fungal pathogenesis. We thus focused on the functional analysis of this DEG. Bioinformatics analysis indicated that the hypothetical protein contained 255 amino acid residues and shared 66% identity with the spore germination protein (PQE13889.1) of Rutstroemia sp. NJR-2017a BBW ( Figure S3); however, the protein has not been previously characterized. Our functional analysis demonstrated that BCIN_16g01820 was involved in B. cinerea conidial germination (see below), therefore BCIN_16g01820 was designated as B. cinerea conidial germination-associated factor 1 (BcCGF1). BcCgf1 is conserved among some known fungi and putatively conserved domains have not been detected in the protein ( Figure S3).
To A similar result was observed when mycelial plugs of the tested strains were used to inoculate green bean leaves ( Figure S6).

Complementation of the mutants with the wild-type (WT) CGF1
allele completely rescued the defects of the mutants (Figures 3   and 4). The findings demonstrate that BcCGF1 is a novel virulence-associated factor required for conidial development, virulence, and in planta development of the pathogen.

| BcCGF1 is required for B. cinerea IFS development and host penetration
To investigate the reduction in virulence of the ΔBccgf1 mutants, we evaluated the capability of the WT, ΔBccgf1, and ΔBcgf1-C strains to form IFSs that play crucial roles in pathogen host penetration. We inoculated the tested strains on inductive surfaces (Feng et al., 2017)

| BcCGF1 mediates B. cinerea conidiation, conidial morphogenesis, and germination, but is dispensable for sclerotial formation and germination
To determine the roles of BcCGF1 in fungal development, we inoculated the WT, ∆Bccgf1, and ∆Bccgf1-C strains on glass slides or CM plates and performed fungal development assays. Disruption of BcCGF1 reduced conidiation of the ∆Bccgf1 mutants (Figure 6a,b). The mutant conidia displayed an abnormal morphology (Figure 6a, lower panel).
Quantitative analysis demonstrated that conidia produced by the mutants were smaller in both length and width than those produced by the WT and complemented strains (Figure 6c). The subsequent conidial germination assay indicated that disruption of BcCGF1 reduced the conidial germination rate and germ tube development (Figure 6d-f).
Complementation of the mutants with the WT BcCGF1 rescued the defects ( Figure 6). However, loss of BcCGF1 did not impair the production, morphology, or germination of the mutant sclerotia (Figure 6g-k).
These findings indicate that BcCGF1 is required for conidiation, conidial morphogenesis and germination, and germ tube development, but dispensable for sclerotium formation and germination, of the pathogen.

| BcCGF1 mediates B. cinerea osmotic and oxidative stress adaptation as well as cell wall integrity
To test whether BcCGF1 plays a role in B. cinerea adaptation to infectionrelated stresses, we inoculated the WT, ΔBccgf1, and complemented strains on CM supplemented with the assorted stress-mimicking agents, including osmotic stress agents NaCl and KCl, the oxidative stress agent H 2 O 2 , and the cell wall-disturbing agents sodium dodecyl sulphate (SDS) and Congo Red (CR) (Feng et al., 2017), and compared the radial growth rates of these strains. The mutant strains displayed reductions in conidial germination and subsequent hyphal growth (Figure 7a,c) and reductions in radial growth of mycelia on CM containing the assorted stress agents (Figure 7b,d). Further analysis indicated that the relative mycelial radial growth inhibition (Cao et al., 2018;Hou et al., 2020) of the ΔBccgf1 strains significantly increased when cultured on CM containing the stress-mimicking agents compared to that of the control and complemented strains (Figure 7e,f). These data suggest that BcCGF1 is required for pathogen osmotic and oxidative stress adaptation as well as cell wall integrity.

| Loss of BcCGF1 impairs endogenous ROS production and cAMP rescues the infection-related development in the BcCGF1 deletion strains
Endogenous ROS and the cAMP signalling pathway play crucial roles Data represent means ± SD from at least four independent experiments. **, ***Significance at p < .01 and p < .001, respectively nitroblue tetrazolium (NBT) staining approaches (Cao et al., 2018;Hou et al., 2020). DAB is converted to dark-brown polymers in the presence of H 2 O 2 and NBT forms a dark-blue water-insoluble formazan precipitate on reduction by superoxide radicals. We also determined the influence of cAMP signalling on conidial germination and/or IFS formation of the tested strains as previously described (Feng et al., 2017;Cao et al., 2018;Hou et al., 2020). Our data demonstrate that the ∆Bccgf1 mutants are difficult to be stained in the germ-tubes and growing mycelia during conidial germination, appressorium, and infection structure formation (Figures 8a and S7); quantification analysis indicated that the relative ROS production by the mutants was significantly reduced (Figure 8b). These findings demonstrate that loss of BcCGF1 in B. cinerea impairs endogenous ROS production.
The analysis of cAMP influence on the fungal infected-related development indicated that exogenous cAMP (50 µM) significantly facilitates conidial germination, appressorium, and IFS formation of F I G U R E 4 BcCGF1 is required for virulence of Botrytis cinerea. Host leaves and fruits were inoculated, incubated at room temperature in the dark, and at the indicated times after inoculation the inoculated materials were photographically documented and quantitatively analysed. Intact leaves were inoculated with the indicated strains via the conidial inoculation approach (10 5 conidia/ml, 8 μl) and the diseased tomato (a) or green bean (b) leaves at the indicated times after inoculation were photographically documented. Quantification of the relative lesion sizes on inoculated tomato (c) and green bean (d) leaves. Diseased apple fruits induced by conidia (10 5 conidia/ml, 5 μl) (e) or mycelial plugs (f) of the indicated strains at the indicated times after inoculation and quantification of the relative lesion sizes induced by conidia (g) or mycelial plugs (h) of the strains as shown in (e) and (f). Data represent means ± SD from at least four independent experiments. *, ***Significance at p < .05 and p < .001, respectively the mutants compared to that of the strains in the absence of cAMP . Consistent with the previous reports (Feng et al., 2017;Hou et al., 2020), exogenous cAMP in the WT and ΔBccgf1-C strains resulted in a promotion of conidial germination, and an earlier emergence and facilitation of IFS development (Figure 8c-g). These data suggest that cAMP signalling regulates the pathogen infectionrelated development downstream of BcCgf1.

| D ISCUSS I ON
In this study, we performed simultaneous transcriptome analyses of tomato and B. cinerea during their early stage of interaction and identified 720 and 621 DEGs in tomato and B. cinerea, respectively.
We used B. cinerea WT strain cultured in liquid medium, growing from spore suspension as a control. This culture condition can mimic the fungal infection environment inside host cells/tissues that may be hypoxic, and thus can minimize the interference of fungal genes that respond only to low-oxygen conditions (Kawahara et al., 2012). After DEG enrichment analysis, we identified the tomato and B. cinerea genes and pathways that may play roles in their early stage of interaction and functionally investigated the gene functions of some fungal DEGs, including BcCGF1.
Similar to the previous findings (Blanco-Ulate et al., 2013;Smith et al., 2014;Vega et al., 2015;Rezzonico et al., 2017), we identified a large number of up-regulated tomato genes that may increase plant defence against B. cinerea. These genes are mainly associated with ROS and phytohormone production, stress responses, and PR proteins, and so on (Tables 1 and S3 . Deletion of BcCGF1 in B. cinerea did not affect sclerotial production. The wild type (B05.10), ΔBccgf1, and ΔBccgf1-C strains were inoculated on complete medium (CM) plates at 20 °C in darkness after 14 days. Sclerotial production by each strain was photographically documented (g) and quantitatively determined (h). (i) Quantification of the sizes of sclerotia produced by the indicated strains at 14 days. (j) Sclerotial germination of the indicated strain on CM plates at the indicated hpi. (k) Relative sclerotial growth of the indicated strains (nine sclerotia/ plate). Data represent means ± SD from at least three independent experiments in which triplicate plates were analysed for each strain in each experiment. *, **, *** significance at p < .05, p < .01, and p < .001, respectively B. cinerea (Figure 1a and Table S11). The plant JA signalling pathway is activated mainly against necrotrophic pathogens (El Oirdi et al., 2011). The up-regulation of Solyc12g009220.1 (JAZ2) (Table S3) (Table S3), encoding CAMTAs or calcium-related factors, were up-regulated. The up-regulation of these new tomato genes suggests that they may be required for tomato defence against pathogen attack or maintenance of survival of the infected cells.

F I G U R E 7
BcCGF1 mediates osmotic and oxidative stress adaptation as well as cell wall integrity of Botrytis cinerea. Conidial (5 × 10 5 conidia/ml, 5 μl) (a) and mycelial radial growth (mycelial plugs 5 mm in diameter) (b) of the wild type (B05.10), ΔBccgf1, and ΔBccgf1-C strains of B. cinerea on complete medium (CM) supplemented with the osmotic stress agents NaCl (1 M) and KCl (1 M), the oxidative stress agent H 2 O 2 (6 mM), the cell wall disturbing agents sodium dodecyl sulphate (SDS, 0.01%) or Congo Red (CR, 300 μg/ml). Quantification of the relative mycelial growth of the indicated strains growing from inoculated conidia (c) or mycelial plugs (d) on CM supplemented with the indicated stress-mimetic agents as presented in (a) and (b). The relative mycelial growth inhibition of the indicated strains growing from inoculated conidia (e) or mycelial plugs (f) on CM supplemented with the indicated stress-mimetic agents as presented in (a) and (b). Representative images are from one experiment (4 days post-inoculation). Data represent means ± SD from three independent experiments in which triplicate plates were examined for each strain in each experiment. *, **, *** significance at p < .05, p < .01, and p < .001, respectively Autophagy facilitates programmed cell developmental changes that occur during cellular remodelling, and thus serves as an adaptive mechanism for cellular nutrient starvation. Therefore, it is not surprising that some ATGs, including BcATG1, BcATG2, and BcATG13, were up-regulated in the pathogen during initial infection. At early infection stage, B. cinerea may be in a condition of nutrient limitation and autophagy allows the pathogen to cope with nutrient starvation in the infection niches. The interactions of the autophagy proteins F I G U R E 8 BcCGF1 regulates endogenous reactive oxygen species (ROS) production in Botrytis cinerea and cAMP rescues the infectionrelated development in the BcCGF1 deletion strains. Conidia of the wild type (B05.10), ∆Bccgf1, and ∆Bccgf1-C strains were cultivated on complete medium (CM) or CM supplemented with diphenylene iodonium (DPI) or dithiothreitol (DTT) at 20 °C for the indicated lengths of time, and then the germ tubes or hyphae/mycelia of the tested strains were stained with 3,3′-diaminobenzidine (DAB) or nitroblue tetrazolium (NBT) solution as previously described (Cao et al., 2018;Hou et al., 2020).  (Ren et al., 2017(Ren et al., , 2018a(Ren et al., , 2018bLiu et al., 2019a).
Glucose metabolism is one of the basic characteristics of B. cinerea host infection. Glucose mediates pathogen development and pathogenesis via initiating conidial germination, mediating infection structure development, and host penetration (Liu et al., 2018).  (Table S5), suggesting that these enzymes may function in the pathogen CWD and break the plant physical barrier.
Unsurprisingly, a large number of genes encoding glucosidases were up-regulated in B. cinerea during infection (Table S5). We generated the KO mutants of the six β-glucosidase genes (BcBGL1 to BcBGL6) with up-regulation levels ranging from 3-to 26-fold (Table S5) and performed pathogenicity assays for these mutants; however, disruption of these genes did not impair virulence of the mutants (Figure 2b), suggesting that these individual β-glucosidases may be dispensable for virulence of the pathogen. This may result from the functional redundancy of these enzymes in B. cinerea, thus loss of one or two of them did not impair virulence of the pathogen.
Our functional analysis suggested that fungal GMC oxidoreductase BCIN_03g01540 is involved in B. cinerea host infection.
The GMC superfamily is a large and functionally diverse family of oxidoreductases that share a common structural fold. Fungal GMC enzymes of this superfamily that exhibit lignocellulose degradation include aryl-alcohol oxidoreductase, alcohol oxidase, cellobiose dehydrogenase, glucose oxidase, glucose dehydrogenase, pyranose dehydrogenase, and pyranose oxidase (Sutzl et al., 2019). In polyporales, the members of the GMC oxidoreductase superfamily also play a central role in degradation of plant polymers because they generate extracellular H 2 O 2 , acting as the ultimate oxidizer in both white-rot and brown-rot decay (Ferreira et al., 2015).

| Fungal and plant materials
B. cinerea WT B05.10 and its derived mutant and complemented strains were used in this study (Table S14). All strains were maintained on potato dextrose agar (PDA) or CM as previously described (Liu et al., 2018). The plant materials used in this work included tomato, green bean, strawberry, and apple leaves or fruits.   (Trapnell et al., 2009;.

| Transcriptome analysis
The DEGs were identified from RNA-Seq data with the cut-off of corrected p value < .01. Analyses of the biological information of DEGs were performed as previously described (Hou et al., 2020).
The database used in the analyses included EggNOG (evolution-

| RT-qPCR
Total RNA for RT-qPCR assay was extracted using RNAiso plus kid (TaKaRa), and cDNA was prepared using PrimerScript RT Reagent Kit with gDNA Eraser (TaKaRa) according to the manufacturer's instructions. RT-qPCRs were carried out as previously described (Feng et al., 2017).

| Generation of gene deletion mutants and complemented strains
Generation of gene KO mutants (including ΔBccgf1, ΔBclae1, ΔBcvel1, ΔBcatg1, ΔBCIN_03g01540, and ΔBcbgl1 to ΔBcbgl6) and the genetic complemented strain ΔBccgf1-C was performed as previously described (Feng et al., 2017). Briefly, vector pXEH containing the HPH cassette was used for replacement of the targeted genes. The 5′ and 3′ homologous flanks of the targeted gene were amplified and cloned into pXEH upstream and downstream of HPH, respectively. The gene KO vector was transformed into Agrobacterium tumefaciens AGL-1. The resultant deletion transformants were screened on PDA with 100 μg/ml hygromycin. Vector pSUL conferring resistance to chlorimuron-ethyl was used for complementation of the ∆Bccgf1. A fragment containing BcCGF1 (1219 bp upstream and 686 bp downstream of the coding region of BcCGF1) was amplified by PCR and cloned into pSUL vector to generate the complementation vector. The complementary vector was transformed into A. tumefaciens AGL-1 and the resultant transformants were screened on defined complex medium containing 100 μg/ml chlorimuron-ethyl (Rolland et al., 2003). Diagnostic PCR was performed to verify the integration events of the selected transformants. The gene KO and complemented strains were further confirmed by RT-qPCR (Weiberg et al., 2013). The primers used in the experiments are listed in Table S15.

| Fungal growth and pathogenicity assays
For growth assays, conidia of B. cinerea WT, gene KO, and complemented strains were harvested with PDB, and conidial suspension was adjusted to approximately 10 5 conidia/ml. Conidial suspension (10 μl) was dropped onto glass slides to observe conidial morphology and germination, or onto CM plates to observe mycelial growth.
Stress adaptation assays and the influence of cAMP on the development of the tested strains were performed as previously described (Feng et al., 2017;Hou et al., 2020). For pathogenicity assay, droplets of conidial suspension (5 × 10 5 conidia/ml in ½ × PDB, 5 μl) of the tested strains were dropped on host leaves/surfaces. When mycelial plugs were used in these assays, mycelial plugs (5 mm in diameter) taken from a 3-day-old culture of the tested strains were used, unless otherwise indicated. The inoculated slides or plant materials were incubated in the dark in a moistened box at 21 °C. Conidial germination, appressorium, and infection cushion formation were observed and counted under a microscope. At least 100 conidia were counted per replicate in each experiment. At least three independent experiments (triplicate samples examined for each treatment in each experiment) were performed. The assays were performed as previously described (Feng et al., 2017;Cao et al., 2018;Liu et al., 2018).

| Cytological assay
Preparation of conidia and onion epidermal cells as well as sample lactophenol blue staining were performed as previously described (Liu et al., 2018(Liu et al., , 2019bHou et al., 2020). The infected samples were microscopically observed, photographically documented, and analysed at 20 hpi.

| Detection and quantification of ROS production
Detection and quantification of ROS production during conidial germination and infection structure formation were performed as previously described (Feng et al., 2017;Cao et al., 2018;Hou et al., 2020). The software ImageJ (http://rsbweb.nih.gov/ij/) was used to quantify ROS production.

| Statistical analysis
The quantitative data in this study were derived from at least three independent experiments with triplicate treatments examined unless otherwise indicated. The data of controls, including conidial germination, mycelial growth, and lesion size, in each independent experiment were normalized as 1 or 100%. The significance between the data were assessed using the Student's t test and p < .05 was considered as a significant difference.