The wild-type V. mali strain LXS080901 isolated and preserved by our laboratory was grown on potato dextrose agar (PDA) at 25°C in the dark. The gene deletion mutants and complementation strains were cultured on PDA supplemented with 100 mg/ml hygromycin B or geneticin G418 (Sigma, St. Louis, MO, USA). Escherichia coli strains were grown in Luria-Bertani (LB) with appropriate antibiotics at 37°C. Agrobacterium tumefaciens strains were grown in LB with appropriate antibiotics at 28°C.

Total RNA was isolated from fresh mycelia using RNAiso Plus Kit (TaKaRa, Dalian, China). Subsequently, 5 μg of total RNA from each sample was reverse transcribed to cDNA using a HiScript II 1st-Strand cDNA Synthesis Kit (+gDNA wiper) (Vazyme, Nanjing, China). The gene VmXyl2, which is predicted to have xylanase activity and shows high transcript levels during V. mali infection, was cloned and sequenced.

The amino acid sequences of xylanases from other strains in this study were obtained from the NCBI GenBank. All the homology searches were carried out on the NCBI BLAST server. The phylogenetic tree was inferred using the Maximum Likelihood (ML) method implemented in MEGA 7.0, with 1000 bootstrapping replicates. The DNAMAN 6.0 was used to perform multiple sequence alignments of VmXyl2 and other well-characterized xylanases from the GH11 family.

Mycelia grown on PDA for 3 days were used to inoculate apple twigs (Meng et al., 2021). The bark tissues were sampled at 0, 6, 12, 24, 48, 72, 96 and 120 hours post inocualtion (hpi). The RNA was extract from bark tissues, and then the cDNA was synthesized. All RT-qPCR experiments were conducted with SYBR Master Mix (TaKaRa, Dalian, China), following the manufacturer’s protocol. The EF1-a gene was used as an endogenous reference. The whole experiment was repeated twice, and three replicates were included in each experiment. All primers used in these assays are listed in Supplementary Table S1 .

To obtain VmXyl2 gene deletion mutants, PEG-mediated protoplast transformation was carried out to get homologous recombination as described previously (Meng et al., 2021). The gene deletion cassette with three components used the hygromycin B phosphotransferase gene (HPH) as a selective marker for gene deletion ( Supplementary Figure S1 ). Upstream and downstream fragments of VmXyl2 genes were amplified from genomic DNA of the wild-type strain LXS080901 using the gene-specific primers ( Supplementary Table S1 ). The HPH gene was amplified from the vector pBS. The gene deletion cassette was generated by double-joint PCR, and the result was confirmed by sequencing. The cassettes were later transformed into the protoplasts of V. mali LXS080901, and the transformants were screened by culturing on medium with 100 μg/ml hygromycin B. The putative gene deletion mutants were validated by PCR using four primer pairs ( Supplementary Figure S1 and Supplementary Table S1 ). For generating the VmXyl2 complementation strains, the fragment containing the full-length coding of VmXyl2 and its native promoter region are constructed into pYF11 vector by homologous recombination. Then transformed into the VmXyl2 deletion mutant through PEG-mediated transformation. The transformants were confirmed by PCR.

Mycelial plugs (diameter of 5 mm) cut from actively growing colony edges of the wild-type strain, gene deletion mutants, and complemented mutants were transferred to PDA plates. The plates were then incubated at 25°C before the colony shape, color, and diameters were assessed. For the pycnidia formation experiment, the V. mali strains were cultured on PDA plates for 3 days at 25°C and then induced for 30 days under UV light (365 nm), and the number of pycnidia was counted.

Pathogenicity assays were performed using apple leaves and 1-year-old twigs (M. domestica ‘Fuji’) were collected from the greenhouse at Qingdao Agricultural University, Qingdao, China. The detached leaves and twigs were sterilized with 75% ethanol, and wounds were made as described by Yu et al. (2018). Mycelial plugs were used to inoculate the wounds. The inoculated leaves and twigs were placed in trays, which were maintained under conditions of high humidity, 25 °C and darkness. The lesion length was measured, and the development of the lesions was photographed at several time points. The assays were repeated three times, and at least 15 leaves and twigs were included in each treatment.

The evaluation of xylanase activity was conducted following the 3,5-dinitrosalicylic acid (DNS) method described by Yu et al. (2018). This process involved a reaction mixture comprising the purified recombinant protein or culture filtrate and 0.5% beechwood xylan dissolved in 50 mM sodium citrate buffer at pH 5.0. The mixture was incubated at 50°C for 30 minutes. Subsequently, DNS solution was added to the mixture, which was then boiled for 5 minutes. Absorbance was measured at 540 nm. Xylanase activity was defined as the enzyme quantity necessary to catalyze the release of 1.0 mmol of xylose per minute under the conditions of pH 5.0 and 50°C. The activity was quantified in units per minute per milligram of protein (U/mg).

Full length VmXyl2 cDNA was amplified and cloned into was amplified and cloned and inserted into the EcorI and BssHII sites of the pET-32a vector. VmXyl2 recombinant protein was expressed in E. coli strain DE3 cells. Expression was induced by incubation with 0.3 mM isopropyl-β-D-thiogalactopyranoside (IPTG) for 24 h at 16°C. Cells were collected by centrifugation at 5000 × g for 10 min. For protein extraction, cells were resuspended in lysis buffer (20 mM sodium hydrogen phosphate, 300 mM NaCl, pH 7.4) supplemented with 1 mg/ml lysozyme, 1 mM phenylmethanesulfonyl fluoride (PMSF), and 1.98 mM β-mercaptoethanol and then subjected to sonication and centrifugation at 10,000 × g for 10 min. VmXyl2 was purified by affinity chromatography using Ni-NTA resin (Thermo Scientific, Waltham, MA, USA) following the manufacturer’s instructions.

Tobacco plants were grown in a climate chamber (16-h photoperiod, 22°C, 65% relative humidity). A. tumefaciens strain GV3101 carrying pGR106 or TRV vectors cultured as above. The bacterial cells were pelleted and resuspended in MES buffer (10 mM MgCl₂, 1.0 mM MES, and 200 μ acetosyringone, pH 5.7) in the dark for 3 h at room temperature before infiltration. For agro-infiltration assays, A. tumefaciens cells in suspension were combined with a silencing suppressor at a suitable ratio to achieve a final optical density (OD₆₀₀) of 0.6. Subsequently, this mixture was infiltrated into the leaves of 6-week-old tobacco plants using a syringe without its needle attached.

For Tobacco Rattle Virus (TRV)-mediated gene silencing, A. tumefaciens cultures expressing TRV2 constructs and those expressing TRV1 were mixed at a 1:1 ratio to a final OD₆₀₀ of 0.8 and then infiltration into primary leaves of four-leaf-stage tobacco seedlings. BCL2-associated X (BAX) and empty vector were used as controls. Three weeks after treatment with TRV2 constructs, plants were used for corresponding assays.

Y2H assays for examining the interactions among pGAD-MP1/MP2 and pGBK-VmXyl2. pGADT7 and pGBKT7 fused with specific genes were cointroduced into the yeast competent cell AH109 strains (Weidi, Shanghai, China). Yeast cotransformants expressing the bait and prey constructs were isolated on SD-Leu-Trp plates for 2 days and screened by culturing on SD-Trp-Leu-His-Ade plates with 1 mM X-α-gal for 3 days. Yeast transformants co-transformed with pGADT7‐T and pGBKT7‐53 were used as positive controls, and transformants co-transformed with pGADT7‐T and pGBKT7‐Lam served as negative controls.

The molecular weight of VmXyl2 protein were analyzed and predicted using https://prosite.expasy.org/prosite.html. Signal peptides were analyzed with DetaiBio (http://www.detaibio.com/tools/index.php?r=signal-peptide/index). The hydrophobic amino acid sequence was analyzed with https://web.expasy.org/protparam/. The relationships among the xylanase protein family were analyzed through a Blast analysis of the NCBI website (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

All treatments were performed in three independent biological experiments with three replicates. All statistical analysis was conducted using SPSS software (Version 19.0, SPSS Inc., Shanghai, China). All the data collected were subjected to analysis of variance (ANOVA) followed by Duncan’s multiple range tests. The asterisks indicate a statistically significant difference with the wild-type strain (*p < 0.05; **p < 0.01).

The gene VmXyl2 was amplified by PCR using cDNA of V. mali as a template and were confirmed by sequencing. The cDNA of VmXyl2 contains an open reading frame of 681 bp that encodes a protein with 226 amino acid and a calculated molecular mass of 23.84 kDa. Using the Signal P5.0 server, VmXyl2 was predicted that the first 19 amino acids are its signal peptides.

Phylogenetic tree was constructed with the characterized xylanase proteins from various strains. The genomic sequence of V. mali, as published, revealed the presence of five candidate genes that are responsible for encoding xylanases (Yin et al., 2015). Phylogenetic analysis indicated that five genes encoding xylanase in V. mali belonged to two families ( Figure 1A ). VmXyl1, VmXyl4, and VmXyl5 are members of the GH10 family, while VmXyl2 and VmXyl3 are members of GH11 family. The sequence alignment of VmXyl2 and other xylanase from GH11 family showed that VmXyl2 possesses the catalytic residues associated with their activity. Specifically, Glu122 and Glu213 serve as the catalytic sites for enzyme activity ( Figure 1B ). Both VmXyl2 and the reference sequences exhibit a conserved motif consisting of 25 amino acids.

The transcript levels of VmXyl2 at various time points (0, 6, 12, 24, 48, 72, 96, and 120 hpi) were determined during infection of apple twigs by V. mali ( Figure 1C ). The transcript level of VmXyl2 was significantly enhanced from 12 hpi, gradually increased and reached the highest with a 15.02-fold change at 72 hpi. Overall, the significant upregulation of VmXyl2 during infection indicates its potential involvement in the pathogenicity of V. mali.

To conduct the functional analysis of VmXyl2 in V. mali, we generated a mutant strain with a targeted gene deletion ( Supplementary Figure S1A ). These transformants were further confirmed by PCR using the primer pairs presented in SupplementaryTable S1 ( Supplementary Figure S1B ). Corresponding results obtained from these assays showed that VmXyl2 was successfully replaced with HPH. In addition, the complementation strain ΔVmXyl2-C was created by introducing the native promoter of the VmXyl2 coding region into the ΔVmXyl2 mutant. The complementation strain ΔVmXyl2-C was verified by PCR ( Supplementary Figure S1C ).

To investigate the influence of VmXyl2 on the growth and development of V. mali, we evaluated the colony morphology, growth rate, and pycnidia formation of the gene deletion mutant ΔVmXyl2 and the wild-type strain. However, we did not observe any significant differences in colony morphology or growth rate between ΔVmXyl2 and the wild-type strain ( Figures 2A, B ). All strains were able to form pycnidia on PDA under UV-light (365 nm). The ΔVmXyl2 strain produced fewer than 41 pycnidia per plate, whereas the wild-type strain produced over 96 pycnidia per plate. Additionally, the complementation strain ΔVmXyl2-C by reintroducing VmXyl2 restored pycnidia formation to that of the wild-type strain ( Figures 2C, D ). The results suggested that VmXyl2 has no impact on the vegetative growth, but it does affect the pycnidia formation of V. mali.

Pathogenicity assays were conducted on the detached apple leaves and twigs to investigate the role of VmXyl2 in disease development. The results indicated that the ΔVmXyl2 strain exhibited significantly reduced virulence towards apple leaves and twigs, in contrast to the wild-type strain ( Figure 3 ). The wild-type strain typically exhibited symptoms of necrosis and canker. The ΔVmXyl2 strain exhibited a reduction of over 63.81% and 60.98% in the average lesion size on both apple leaves and twigs. Moreover, the complementation strain ΔVmXyl2-C restored the highly virulent phenotype, exhibiting the same symptoms on both apple leaves and twigs. The results indicated that VmXyl2 plays a crucial role in the pathogenicity of V. mali.

To assess the effect of VmXyl2 deletion on the ability of V. mali to utilize xylan, we compared the growth rate of the wild-type and mutant strains on minimal medium supplemented with xylan as a sole carbon source. The growth rate of ΔVmXyl2 mutant was significantly affected by the deletion of VmXyl2, the deletion mutant grew 34.31% slower than the wild-type strain ( Figures 4A, B ). We further examined the effect of deleting VmXyl2 on xylanase activity in the culture filtrates of the wild-type and mutant strains. Xylan was found to induce the production and secretion of xylanase. The xylanase activity of the ΔVmXyl2 mutant was reduced 41.22% compared to the wild-type stain ( Figure 4C ). The reintroduction of the native gene VmXyl2 resulted in the restoration of growth rate and xylanase activity to the levels comparable to the wild-type strain ( Figure 4 ). The results indicated that VmXyl2 is involved in the absorption of xylan and carry xylanase activity in V. mali.

The recombinant protein VmXyl2 was successfully purified to study its ability to induce cell death in both host and non-host plants ( Supplementary Figure S2A ). VmXyl2 was found to cause significant cell necrosis in the leaves of Nicotiana tabacum cv. Samsun, M. domestica, Arabidopsis thaliana and Fragaria ananassa, while no necrotic effects were observed in N. benthamiana leaves ( Figures 5A–E ). Cell necrosis was observed at a minimum concentration of 50 nM VmXyl2. The size of necrosis increased with higher concentrations of purified protein, ranging from 50 nM to 200 nM ( Supplementary Figure S2B ). Moreover, VmXyl2 was discovered to stimulate the accumulation of hydrogen peroxide ( Supplementary Figure S2C ) and the upregulation of defense-related genes in tobacco and apple leaves ( Supplementary Figure S3 ).

To investigate the impact of VmXyl2-inducing cell necrosis on V. mali infection, the pathogen was inoculated to apple leaves following treatment with purified protein. Bovine serum albumin (BSA) was used as a control at the same concentration to VmXyl2. Compared to BSA treatment, the lesion size exhibited a significant increase after VmXyl2 treatment ( Figure 5F ), which indicates that VmXyl2-inducing cell necrosis promotes the infection of V. mali on apple leaves.

VmXyl2 contains a signal peptide encoding 19 amino acids at N-terminus, implying VmXyl2 might be a secreted protein ( Figure 6A ). In order to confirm the previous hypothesis that VmXyl2 is a secreted protein to induce cell death, we transiently expressed the full length VmXyl2 and VmXyl2^-SP (without signal peptide) in tobacco and tomato plants through agro-infiltration. The result showed that VmXyl2, which contains signal peptide, was able to induce cell necrosis in N. tabacum cv. Samsun and tomato leaves. However, the absence of the signal peptide in VmXyl2^-sp eliminated its abiltiy to induce cell death ( Figure 6B ). The results suggest that the signal peptide is essential for VmXyl2-mediated cell death.

Plant receptors frequently recognize specific small protein epitopes to induce plant cell death. The amino acid sequence analysis revealed that the conserved sequence of VmXyl2, which consists of a 25-residue pepide (C^111-135), shares 81.71% identity with five other xylanases ( Figure 1B ). To elucidate the function of the 25-residue peptide, we synthetically produced this small peptide segment. It was found that the conserved region C^111-135 induced the same cell death as full-length VmXyl2 in N. tabacum cv. Samsun leaves ( Figure 6C ).

The sequence alignment revealed that VmXyl2 contains two potentially highly conserved catalytic residues (E122 and E213), which are crucial for its xylanase activity ( Figure 4A ). To investigate whether the ability of VmXyl2 to induce plant cell necrosis is dependent on its hydrolase activity, we obtained the site-directed mutant protein that two glutamic acid residues were substituted by aspartate ( Supplementary Figure S4B ). Enzymatic assays with site-directed mutant protein showed that xylanase activity was nearly eliminated ( Supplementary Table S2 ). Surprisingly, the mutant protein lacked xylanase activity but still retained the cell death-inducing activity, similar to the wild-type protein (VmXyl2). The mutant protein and VmXyl2 caused similar visible cell death symptoms in N. tabacum cv. Samsun and tomato leaves ( Supplementary Figures S4C, D ). The results suggests that the cell death-inducing activity of VmXyl2 is independent of its hydrolase activity.

BRI1-associated kinase-1 (BAK1) and Suppressor of BIR1-1 (SOBIR1) are the Leucine-Rich Repeats Receptor-Like Kinase (LRR-RLK), and it was found that BcXyl1-inducing cell death was mediated by BAK1 and SORBIR1 (Yang et al., 2018). To explore the mechanism of VmXyl2-inducing cell necrosis, we silenced the expression levels of BAK1 and SOBIR1 in tabacco plants. The qPCR analysis validated a significant reduction in the expression levels of BAK1 or SOBIR1 upon inoculation with TRV::BAK1 or TRV::SOBIR1, showing an expression level less than 20% compared to inoculation with TRV::GFP ( Figures 7A, B ). After three weeks of viral inoculation to silence BAK1 in N. tabacum cv. Samsun, the transient expression of VmXyl2 through agro-infiltration with VmXyl2 expression constructs did not induce cell necrosis. However, the results obtained from plant with silenced SOBIR1 demonstrated that VmXyl2 induced cell death, which was inconsistent with the findings in plants with silenced BAK1 ( Figure 7C ). In all assays, the positive control BAX induced cell death in plants. Based on these results, we concluded that BAK1, but not SOBIR1, was necessary for VmXyl2-inducing cell death in N. tabacum cv. Samsun.

The tomato LRR-RLK, LeEix2, can recognize xylanase TvEIX and induce cell death in plants (Ron and Avni, 2004; Pizarro et al., 2018b). By analyzing the whole-genome sequence of M. domestica, we obtained two candidate LRR-RLKs, named Mp1 and Mp2, which exhibited the highest homology to LeEix2. The yeast two-hybrid method was used to verify the interaction between two LRR-RLKs and VmXyl2. The results showed that only Mp2 exhibited interaction with VmXyl2 in vitro ( Figure 8A ). To further confirm the interaction in vivo, we performed a tobacco infiltration experiment. Previous experiments showed that VmXyl2 alone did not induce cell necrosis in N. benthamiana leaves ( Figure 5A ). We subsequently co-infiltrated VmXyl2 with Mp1 or Mp2 into the leaves of N. benthamiana. The results showed that cell necrosis was observed only in leaves co-infiltrated with VmXyl2 and Mp2 ( Figure 8B ). This suggests an in vivo interaction between VmXyl2 and Mp2, and highlights the requirement of Mp2 for VmXyl2-induced cell necrosis in N. benthamiana leaves.