Botrytis cinerea wild type (WT) strain B05.10 was used in this study. B. cinerea strains were routinely grown on potato dextrose agar (PDA, Difco) and maintained at 22°C under continuous fluorescent light. Conidia were obtained from 7-days-old cultures. The Agrobacterium tumefaciens strain GV3101 was used for Agrobacterium-mediated transient expression in plant leaves. Escherichia coli strain DH5α was used to propagate plasmids. Bean (Phaseolus vulgaris cv. French bean, genotype N9059) and tobacco (Nicotiana benthamiana) plants were grown in greenhouse (16: 8 h, 25: 22°C, day: night).

Pathogenicity assays of B. cinerea on beans were performed as described (Schumacher et al., 2012) that the conidia harvested from 7-days-old agar cultures were re-suspended in Gamborg’s B5 medium supplemented with 2% glucose and 10 mM KH₂PO₄/K₂HPO₄, pH 6.4. The primary leaves of 9-days-old beans were inoculated with 7.5 μl conidial suspensions (2 × 10⁵ conidia/ml) and the infected bean plants were incubated in the humid chambers at 22°C for 72 h, and the lesion diameter was measured.

The genomic sequence database of B. cinerea at JGI¹ was used to characterize B. cinerea genes. The GPI Modification Site Prediction² was used to analyze potential GPI modification site. The SMART MODE³ was used to analyze signal peptide sequence and protein domain. Database NCBI was used for Blastp analysis. The Clustal W and Jalview program were used for mature proteins alignments. MEGA 5 program was used to generate phylogenetic tree with unrooted neighbor-joining method.

To explore the expression pattern of BcCFEM1 during early infection stage, the B. cinerea conidial suspensions (5 × 10⁵ conidia/ml) were sprayed onto the beans leaves (9-days-old). The inoculated leaves and the hyphae growing on plates (solid Gamborg’s B5 medium containing 2% glucose) as control were harvested at 12, 24, 36, 48, 60, and 72 hours post inoculation (hpi), respectively, and then frozen at -80°C for total RNA extraction. To compare expression level of BcCFEM1 between WT strain, BcCFEM1 deletion mutant and BcCFEM1 over-expression strain, the mycelial agar disks of these strains were inoculated to the cellophane of PDA at 22°C for 3 days, and mycelia were then collected and stored at -80°C for total RNA extraction.

The genomic DNA of fungi were isolated with Axygen DNA kit (Axygen, United States) according to the manufacturer’s protocols. The total RNA samples of fungal strains and plants were isolated using the Axygen RNA kit (Axygen, United States) according to the manufacturer’s instructions and stored at -80°C for further study. The DNase I (TAKARA Biotechnology, Dalian, China) treatment was performed to eliminate residual genomic DNA and the RevertAid^TM First Strand cDNA Synthesis Kit (Thermo Scientific, Lithuania) was used to generate the first strand cDNA.

Quantitative RT-PCR (qRT-PCR) was performed by the using of CFX96 Touch^TM Real-Time PCR Detection System (Bio-Rad, Hercules, CA, United States) and SYBR^®Premix Ex Taq^TM II (TAKARA Biotechnology, Dalian, China), according to the manufacturer’s instructions. Primers were designed across or flanking an intron (Supplementary Table S1). The relative expression of B. cinerea Bcgpdh (BC1G_05277) gene (Ma et al., 2017) and P. vulgaris Actin-11 (GenBank: EH040443.1) gene (Oliveira et al., 2015) were respectively used as reference genes for normalizing the RNA sample. For each gene, qRT-PCR assays were respectively repeated at least twice, with each repetition having three independent replicates.

Southern blot analysis was performed following the method described (Wei et al., 2016) that 15–20 μg of genomic DNA of each strain was digested with XbaI, size-fractionated through a 0.8% agarose gel and mounted on positively charged nylon membrane. The nylon membrane was then hybridized with a probe amplified from hygromycin hph gene by primer pairs (Supplementary Table S1) and labeled with digoxigenin (DIG)-dUTP using the PCR DIG Probe Synthesis Kit (Roche, Mannheim, Germany) following the manufacturer’s protocols.

The BcCFEM1 replacement constructs was generated as described (Ma et al., 2017) that the 5′(547 bp)- and 3′(536 bp)- flanks of the BcCFEM1 ORF were amplified from genomic DNA of the WT strain B05.10 with primer pairs (Supplementary Table S1). The fragments were then respectively cloned into the upstream and downstream of hph cassette using Gibson Assembly Master Mix kit (New England Biolabs, Ipswich, MA, United States) according to the manufacturer’s protocols.

For complementation assays, the PCR product containing a 2-kb upstream sequence, a full-length BcCFEM1 gene coding region and a 1-kb downstream sequence was amplified from the genomic DNA of WT strain using primer pairs (Supplementary Table S1), and cloned into the plasmid p3300neoIII to generate the complementary vector p3300neoIIIBcCFEM1 as described (Zhang et al., 2016).

To construct the over-expression vector of BcCFEM1, the full-length BcCFEM1 ORF fused with a HA tag at C terminus was cloned into hph cassette-containing vector p3300hyg under the manipulation of PtrpC promoter and TtrpC terminator as described (Wei et al., 2016). To construct the expression vector of SP-EGFP-HA, EGFP ORF fused with BcCFEM1 signal peptide and HA tag at N and C terminus respectively was cloned into hph cassette-containing vector p3300hyg under the manipulation of PtrpC promoter and TtrpC terminator.

For transient expression in N. benthamiana, the sequence encoding Arabidopsis thaliana PR3 (TAIR ID: AT3G12500) signal peptide-BcCFEM1-HA tag fusion protein was cloned into pCAMBIA3300 under the manipulation of CaMV 35S promoter and NOS terminator. The transient expressions of other genes were also performed using this method.

Protoplasts preparation and genetic transformation of B. cinerea were performed as previously described (Ma et al., 2017).

To assay growth rate, the WT strain, BcCFEM1 deletion mutants, and BcCFEM1-HA over-expression transformants were cultured on PDA at 22°C for 2 days. Then, the mycelial agar disks were taken from the active colony edge and inoculated on the center of the PDA petri dish at 22°C to examine the hyphal growth rate. The colony morphology of these strains were examined after being grown on PDA plate for 7 days at 22°C under 24 h fluorescent light photoperiod and/or 14 days at 22°C under dark condition. For stress tolerance assay, indicated strains were respectively inoculated onto PDA plates containing 1 M NaCl, 1 M Sorbitol, 0.4 mg/ml Calcofluor White (CFW), 20 mM H₂O₂, 0.8 mg/ml Congo Red (CR) and 0.02% SDS as previously described (Ma et al., 2017). To analyze the conidial germination, 25 μl conidial suspensions (1 × 10⁵ conidia/ml) of each strain was spotted on hydrophobic coverslips and incubated at 22°C for 10 h. Then, the conidial germination was examined by Nikon Eclipse 80i microscope (Nikon, Tokyo, Japan), under bright-field model using 40× fold magnification.

Agrobacterium-mediated transient expression was performed using leaf infiltration as described (Kettles et al., 2017). Protein extractions from B. cinerea and the immunoblot analysis were performed as described (Lyu et al., 2016).

To confirm whether the BcCFEM1-HA fusion protein was secreted outside the cell to extracellular matrix, 0.2 g fresh mycelia was cultured in 5 ml potato dextrose broth (PDB, Difco) in 6-well-cell culture plate at 22°C with agitation at 180 rpm for 3 days. The liquid mediums were then filtered with 0.45 μm Minisart^®non-pyrogenic filter (Sartorius-Stedim Biotech, Germany) to eliminate residual mycelia. The Amicon^®Ultra-4 Centrifugal Filter Devices (10 ml, 10 kDa, Merck Millipore, Billerica, MA, United States) was used to clean and concentrate the target secretory protein solution. The clean samples were dissolved in 250 μl PBS solution and frozen at -80°C for further immunoblot analysis.

Statistical tests were performed using Origin 7.5 (OriginLab Corporation, Northampton, MA, United States). Data significance was analyzed by the using of ANOVA (one-way, P ≤ 0.01). In all graphs, results represent the mean value of three independent experiments ± SD. Different letters or asterisks in the graphs indicate statistical differences, P ≤ 0.01.

The B. cinerea BC1G_15201 gene is a single copy gene, consisting of two exons, encoding 215 amino acid. The SMART MODE analysis result indicated that the initial N terminus 17 amino acids encode signal peptide and 18–86 amino acids encode CFEM (common in several fungal extracellular membrane proteins) domain (Pfam ID: PF05730) (Figure 1A). No transmembrane helices of this protein were predicted, indicating that BC1G_15201 is a secretory protein possibly. BLAST searches for homologous sequences of BC1G_15201 only resulted in similarity with sequences from Sclerotinia sclerotiorum (SS1G_07295, E-value: 2E-52, 58% identity), Metarhizium album (MAM_04868, E-value: 2E-15, 36% identity), Phialocephala subalpina (PAC_04512, E-value: 2E-13, 37% identity), Colletotrichum gloeosporioides (CGLO_03584, E-value: 2E-09, 45% identity), Trichoderma gamsii (TGAM01_02421, E-value: 6E-09, 58% identity) that match to CFEM domain. Phylogenetic and multiple sequence alignment analysis revealed sequences conservation (Figures 1B,C), especially the eight cysteine residues (C²⁵, C²⁹, C⁴⁰, C⁴⁷, C⁴⁹, C⁶⁴, C⁶⁹, and C⁸⁵) in BC1G_15201 are well-conserved in all of these homologs (Figure 1C), which may be involved in the formation of disulfide bonds and play significant roles in the structure and function of BC1G_15201.

Therefore, we named BC1G_15201 protein “BcCFEM1,” as this is the first report of B. cinerea CFEM protein.

GPI Modification Site Prediction⁴ was used to predict potential GPI modification site. The result demonstrated that the amino acids G¹⁹² and A¹⁹³ are predicted to be the best and second best of potential GPI-modification site, respectively (Supplementary Figure S1). From this result, we inferred that BcCFEM1 contains a putative GPI-anchored site, which possibly anchored to the outer layer of the plasma membrane through a C-terminal GPI anchor or transferred to the cell wall as other fungal GPI-anchored proteins (Krüger et al., 2016).

As mentioned above that BcCFEM1 contains a putative signal peptide and maybe a secretory protein, to verify this hypothesis further, SP-EGFP-HA expression strain, which EGFP fused with BcCFEM1 signal peptide and HA tag at N and C terminus respectively, was constructed and cultured in PDB medium for shake culture. Western blot result demonstrated that SP-EGFP-HA fusion protein could be detected in both culture medium and mycelium (Figure 2). The result indicated that the signal peptide from BcCFEM1 can mediate secretion of protein effectively.

In addition, we also constructed over expression strain OEBcCFEM1-HA and cultured in PDB medium. Unexpectedly, Western blot result demonstrated that BcCFEM1-HA fusion protein could not be detected in both mycelium and culture medium (Figure 2), indicating that BcCFEM1-HA fusion protein may be cleaved or modified.

To explore the possible contribution of BcCFEM1 to B. cinerea, we determined the expression patterns of BcCFEM1 during infection stages following inoculation on beans with conidia. Results indicated that when B. cinerea was inoculated on bean leaves, the transcript levels of BcCFEM1 increased from 12 to 36 hpi, and then decreased, but the expression level was still higher than at 0 hpi (Figure 3). However, when inoculated on solid Gamborg’s B5 medium, the expression of BcCFEM1 did not vary to any great extent, with the highest expression being observed at late stage (72 hpi) (Figure 3). Therefore, the expression levels of BcCFEM1 must be induced by its interaction with host plants, suggesting that this gene may play significant roles at the early stages of infection.

In addition, transient expression of BcCFEM1-HA using Agrobacterium-infiltration method caused obvious chlorosis in N. benthamiana leaves (Figure 4), indicating that BcCFEM1 may function as a virulence factor.

We firstly constructed BcCFEM1 deletion, complementary and over-expression strains, and these strains were confirmed by PCR, RT-PCR, qRT-PCR, and Southern blot (Supplementary Figure S2).

To establish whether BcCFEM1 is required for B. cinerea physiological processes, the phenotype of WT, BcXYG1 deletion mutant ΔBcCFEM1, BcCFEM1 complementary strain BcCFEM1-Com and BcCFEM1 over-expression strain OEBcCFEM1 were assayed. The result showed that there was not any obvious difference in growth rate between these strains when cultured on PDA medium (Figure 5A). In addition, the conidial morphology and germination were also assayed. However, we did not detect any differences in spore germination between the WT, ΔBcCFEM1, BcCFEM1-Com, and OEBcCFEM1 strains (Figure 5B). Further more, the colony morphology of each strain under continuous fluorescent light or dark condition was also analyzed. Likewise, over expression or deletion of BcCFEM1 gene had no visible effect on colony morphology and sclerotial production on PDA plate (Figure 5C).

As described above that the BcCFEM1 was induced to express highly at early infection stage and could cause chlorosis in tobacco leaves, inferring that BcCFEM1 may be involved in fungal virulence. To study the possible effect of BcCFEM1 on B. cinerea virulence, infection was carried out on bean leaves. Results demonstrated that the pathogenicity of ΔBcCFEM1 was decreased compared to WT, BcCFEM1-Com and OEBcCFEM1 strains, but over expression of BcCFEM1 had no visible effect on virulence (Figure 6A).

Although WT, ΔBcCFEM1, BcCFEM1-Com, and OEBcCFEM1 cultures developed smooth and uniform mycelium that differentiated to form abundant spores within 7 days, conidial production of ΔBcCFEM1 was reduced compared with WT, BcCFEM1-Com and OEBcCFEM1 strains (Figure 6B).

In addition, to investigate whether BcCFEM1 is involved in external stress tolerance, growth assays of indicated strains on PDA media supplemented with salt stress (1 M NaCl), osmotic stress (1 M Sorbitol), H₂O₂ (20 mM) and cell wall stress (0.3 mg/ml CFW, 0.02% SDS, and 0.5 mg/ml CR) were respectively analyzed as previously described (Pérez-Hernández et al., 2017). The results showed that the ΔBcCFEM1 strain was more hypersensitive to all of these compounds compared with WT, BcCFEM1-Com and OEBcCFEM1 strains (Figure 6C). However, no obvious change in stress tolerance between WT, BcCFEM1-Com and OEBcCFEM1 strain was detected (Figure 6C). These results indicated that BcCFEM1 may positively regulate the sensibility of B. cinerea to cell wall stressors, oxidative stress and osmotic stabilizers.