The V. mali wide type strain 03–8 used in this study was deposited in the Laboratory of Pathogen Biology and Integrated Control of Fruit Tree Diseases, College of Plant Protection, Northwest A&F University, Yangling, Shaanxi, China. The wild type and mutant strains generated in this study were maintained on a potato dextrose agar (PDA, 20% potato, 2% glucose, and 1.5% agar) medium at 25°C in the dark. Liquid potato dextrose broth (PDB) was used as the growth medium in the mycelium collection.

The VmDCL sequences were isolated from the available V. mali genome database using the DCL sequences of M. oryzae and N. crassa as the probe sequences for the in silico cloning. The domain architecture was analyzed using InterProScan¹. A phylogenetic comparison of the deduced protein sequences and the corresponding proteins from other species was performed with DNAMAN version 6.0.

The fungal genomic DNA was extracted using the CTAB method (Möller et al., 1992). For the RNA isolation, apple twigs infected with V. mali strain 03–8 were sampled at 48 hpi (hours post inoculation). RNA samples from 03–8 mycelium and apple twigs infected by 03–8 were isolated using the TRIzol^TM Reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer’s instructions. Contaminated genomic DNA was removed using DNase I (Invitrogen, Carlsbad, CA, United States). DNA/RNA degradation was monitored on 1% agarose gels. A NanoDrop^TM 1000 spectrophotometer (Thermo Fisher Scientific, United States) was used to assess the DNA/RNA purity, concentration and integrity.

In total, three micrograms of total RNA were used to synthesize the first-strand cDNA using an RT-PCR system (Thermo Fisher Scientific, United States) according to the manufacturer’s instructions. The quantitative PCR amplifications were performed on a CFX96 Real-Time System (Bio-Rad) with SYBR Green (Invitrogen). The Glyceraldehyde-6-phosphate dehydrogenase gene (G6PDH) of V. mali was used as the internal control gene (Yin et al., 2013). The relative expression levels of each gene were calculated by the 2^-ΔΔCT method (Livak and Schmittgen, 2001). Three biological replicates were performed for each experiment. All primers are listed in Supplementary Table S1.

The vector pBIG2RHPH2-GFP-GUS, which contains a hygromycin resistance gene, was used to construct the gene deletion vector. The double-joint PCR approach was used for the vector construction and generating the gene knock-out mutants (Yu et al., 2004). For the single gene knock-out mutants, the strategy consisted of replacing the VmDCLs with the hygromycin-phosphotransferase (hph) cassette, which was amplified from PHIG2RHPH2-GFP-GUS. The upstream and downstream flanking sequences were amplified with the 1F/2R and 3F/4R primers, respectively. After ligation with the hph cassette, the ligated PCR product was transformed into 03–8 according to the method described by Gao et al. (2011). Positive transformants were selected using hygromycin B at a final concentration of 100 μg/mL. The putative knockout mutants identified by screening with the primers 5F and 6R were further analyzed by PCR with the primers 7F and H855R and H856F and 8R to confirm the gene replacement events (Supplementary Figure S1). For the double gene knock-out mutants, the single gene knock-out mutant (ΔVmDCL2) was used to generate the protoplast, and another vector containing a geneticin resistance gene NEO was used to replace the ORF of the second gene. The detailed method used for the mutant generation is similar to that described above. The putative mutants were further confirmed by a Southern blot analysis with probes for the HYG/NEO fragments according to the manufacturer’s instructions of the DIG DNA Labeling and Detection Kit II (Roche, Mannheim, Germany). For the complementation assays, the VmDCLs were amplified using the primer pairs Com-F and Com-R. The resulting PCR products were cloned into the vector pFL2 using the yeast gap repair approach (Bruno et al., 2004; Zhou et al., 2011). The resulting fusion constructs rescued from the Trp⁺ yeast transformants were confirmed by sequencing and transformed into their respective V. mali deletion mutant. The geneticin-resistant transformants expressing the complementing constructs were identified by PCR using the primer pairs Com-F and Com-R. The primers used for the gene deletion and complementation are listed in Supplementary Table S1.

To detect the growth rate of the colonies, VmDCL mutant strains (ΔVmDCLs) and 03–8 mycelium were placed onto the center of a PDA medium plate and were cultured in the dark at 25°C. The size of the colonies was examined after 24, 48, and 72 h. To observe the morphology of the colonies and hyphae, the colonies were photographed, and the hyphae morphology was observed using samples from the colony edge of each strain cultured in the dark at 25°C for 2 days under a microscope. To assay the stress response defects in the ΔVmDCLs, mycelium of the ΔVmDCLs and 03–8 were placed onto a PDA medium plate containing 0.05% H₂O₂ and 0.5 M KCl, and then, the plates were cultured in the dark at 25°C. The size and morphology of the colonies were examined after 3 days. The significance was analyzed using SPSS18 software. All experiments were performed in triplicate with three Petri dishes in each replication.

‘Fuji’ apple (Malus domestica Borkh. cv ‘Fuji’) twigs were used to detect the pathogenicity of the ΔVmDCLs and 03–8. The test was conducted using the detailed methods described by Wei et al. (2010). The statistical significance of the differences was analyzed using SPSS18 software. The infection assays were repeated three times with six twigs per replicate.

Small RNA libraries were prepared following the detailed protocol provided by the genome sequencing company (Novogene, China). Three individual biological replicates of each sample (ΔVmDCL1, ΔVmDCL2, ΔVmDCL1DCL2 and 03–8) were prepared, and the three sample replicates were mixed for the RNA isolation. For the small RNA library construction, three micrograms of total RNA from each sample were used as the input material. The sequencing libraries were generated using the NEBNext^® Multiplex Small RNA Library Prep Set for Illumina^® (NEB, United States) following the manufacturer’s recommendations, and index codes were added to attribute sequences to each sample. The library quality was assessed on an Agilent Bioanalyzer 2100 system using DNA High Sensitivity Chips. The clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq SR Cluster Kit v3-cBot-HS (Illumia) according to the manufacturer’s instructions. After the cluster generation, the library preparations were sequenced on an Illumina HiSeq 2500/2000 platform, and 50 bp single-end reads were generated.

Raw data (raw reads) in the fastq format were first processed through custom Perl and python scripts. During this step, clean data (clean reads) were obtained by removing reads containing ploy-N, reads with 5′ adapter contaminants, reads without 3′ adapters or insert tags, reads containing ploy A, T, G or C, and low-quality reads from the raw data. The clean reads were mapped onto the reference sequence using Bowtie (Langmead et al., 2009) without a mismatch to confirm the sequence accuracy. The number of total unique sRNAs from the different samples was calculated to compare the difference among the ΔVmDCLs and 03–8. The common and specific sequences among the ΔVmDCLs and 03–8 were also compared. To preliminarily estimate the sRNAs varieties, the length distribution of each sample was also analyzed.

In total, there are two DCL homologous sequences in V. mali. The genes were designated VmDCL1 and VmDCL2, and the sequences were deposited in GenBank with accession numbers KUI68257.1 and KUI73784.1, respectively. The InterProScan analyses showed that the deduced VmDCL1 protein contains five domains, which are known as RNA helicase domains (DEAD and hel C), dsRBD and Ribonuclease domains (RNase III a and RNase III b) and are characteristic of DCL proteins. Compared with VmDCL1, VmDCL2 also contains an additional special domain, i.e., DSRM (Figure 1A). In addition, the phylogenetic analysis suggests that the VmDCLs are clustered with the DCLs from fungi as expected, and they are close to the DCLs from Cryphonectria parasitica in evolution (Figure 1B).

As shown in Supplementary Figure S2, the VmDCL1 single-gene deletion mutants (ΔVmDCL1), VmDCL2 single-gene deletion mutants (ΔVmDCL2) and the VmDCL1DCL2 double-gene deletion mutants (ΔVmDCL1DCL2) were identified via screening with different PCR primers. The results indicated that we obtained putative knockout mutants of VmDCL1, VmDCL2 and VmDCL1DCL2. Subsequently, all putative knockout mutants were further verified by Southern blot hybridization. Furthermore, no hybridization signal was detected in the wild type strain 03–8 using the HYG probe amplified with the primers HYG/F and HYG/R, but there was a 4.3 and 6.0 Kb band in the VmDCL1 and VmDCL2 mutants, respectively; meanwhile, a NEO probe amplified with the primers NEO/F and NEO/R was used to hybridize with the genomic DNA of ΔVmDCL1DCL2, and a single corresponding 3.5 Kb band was detected. Thus, the single copy mutants of VmDCL1, VmDCL2 and VmDCL1DCL2 were generated successfully (Figure 2).

To investigate whether VmDCLs play a critical function in the growth and development of V. mali, the growth rate and morphology of the colony and hypha of the mutants (ΔVmDCL1, ΔVmDCL2 and ΔVmDCL1DCL2) were compared with those of 03–8. No obvious difference in the colonial morphology was observed between the mutants and 03–8 when they were cultured on a PDA medium at 25°C in the dark for 2 days (Supplementary Figure S3a). Furthermore, the hypha morphology did not change obviously (Supplementary Figure S3b). In addition, the growth rate of 03–8, ΔVmDCL1, ΔVmDCL2 and ΔVmDCL1DCL2 was 2.50, 2.48, 2.52, and 2.47 cm/d, respectively, also showing no statistically significant difference (Supplementary Figure S3c).

To determine whether VmDCLs were involved in the stress responses, H₂O₂ and KCl were added to the PDA medium. In the presence of 0.05% H₂O₂, ΔVmDCL2 grew more slowly than 03–8 3 days after the inoculation, but the colony of ΔVmDCL1 was much larger than the colony of 03–8. Interestingly, the colony of ΔVmDCL1DCL2 showed no obvious change compared with that of 03–8 (Figures 3A,B). In the presence of 0.5 M KCl, the colony growth of only ΔVmDCL1 was significantly inhibited, while there was no significant difference among 03–8, ΔVmDCL2 and ΔVmDCL1DCL2 (Figures 3A,C). The complemented strain showed a nearly normal growth compared with 03–8 under the H₂O₂ and KCl treatment conditions. These results suggest that the VmDCLs may be important but play distinct roles in the response to oxidative and osmotic stresses.

The expression analysis of the VmDCLs during the V. mali infection in apple twig showed that compared with the control (mycelium), VmDCL2 was up-regulated and peaked (125-fold) at 48 hpi. However, VmDCL1 expressed stably with no significant change and only exhibited a 1.7-fold increase at 48 hpi (Figure 4A). The pathogenicity of ΔVmDCLs and 03–8 was further analyzed. Compared with 03–8, the pathogenicity of ΔVmDCL1 showed no significant change, while the ΔVmDCL2 was defective in plant infections with a smaller lesion. Furthermore, the pathogenicity markedly decreased when both VmDCL1 and VmDCL2 were knocked out together. Under the same conditions, the pathogenicity of the complemented transformants ΔVmDCL1-C and ΔVmDCL2-C was similar to that of 03–8 in the twig infection assays (Figures 4B,C). These results suggest that VmDCL2 plays a critical role during the infection process in twigs, and the detailed mechanism will be a key scientific problem to be explored in the future.

To comprehensively examine the sRNA abundance in strains of 03–8 (MVM), ΔVmDCL1 (MD1), ΔVmDCL2 (MD2) and ΔVmDCL1DCL2 (MD12), four sRNA libraries were constructed and sequenced using an Illumina Analyzer. The distinct libraries generated 12818591, 6927296, 11565746 and 8990152 raw reads. After removing the low-quality and non-sense sequences (<15 nt), in total, 1364324, 788373, 834923 and 708262 unisRNAs were isolated (data not shown). As shown in Figure 5, the number of unisRNAs was markedly smaller in the libraries of MD1, MD2 and MD12 than that in the library of MVM, indicating that the deletion of the VmDCLs could affect the generation of sRNAs in V. mali. In addition, the length distribution of the unisRNAs was further analyzed (18–30 nt long), and the length distribution also markedly changed when the VmDCLs were deleted. When VmDCL1 was knocked out, the abundance of sRNAs consisting of 29 nt showed the most obvious increase (Figures 6A,B). Meanwhile, the abundance of sRNAs consisting of 18–20 nt noticeably increased, and the abundance of sRNAs consisting of 21–22 nt noticeably decreased when VmDCL2 was knocked out (Figures 6A,C). Additionally, in the ΔVmDCL1DCL2, the abundance of sRNAs consisting of 21 nt noticeably decreased, and the abundance of sRNAs consisting of 29 nt noticeably increased (Figures 6A,D). The other sRNA abundances were also affected in varying degrees between the mutants and 03–8. The results indicated that the VmDCLs could affect the generation of sRNAs in V. mali, and their function is likely redundant. Furthermore, there are at least two pathways (i.e., DCL-dependent and DCL-independent) for sRNA generation in V. mali. The detailed mechanism will be explored in further studies.