Apple (Malus×domestica cv. “Golden Delicious,” GD) plants were grown in tissue culture on Murashige and Skoog medium containing 0.6 mg·L⁻¹ 6-benzylaminopurine and 0.15 mg·L⁻¹ 1-naphthylacetic acid in a climate-controlled culture room at 24 ± 1°C, with a 16/8 h photoperiod. Four-week-old plants were used for Agrobacterium tumefaciens infiltration and fungal infection experiments.

Full-length Md-MIR156ab, Md-MIR395, STTM-miR156ab, STTM-miR395, MdWRKYN1, and MdWRKY26 gene sequences were individually inserted into the plant expression vector pFGC5941 (GenBank AY310901) at the NcoI/BamHI restriction site (primers listed in Supplemental Table 3), using the empty pFGC5941 vector as the control. The vectors were transformed into Agrobacterium tumefaciens (strain GV3101) using the heat shock transformation method. Plants of the susceptible apple variety GD were infiltrated by Agrobacterium tumefaciens containing one miRNA gene, performed when the fifth leaf of the seedling had expanded at approximately 4-weeks-old, as described previously (Bai et al., 2011).

A plant-pathogenic ALT1 (Alternaria alternata f. sp. mali) strain was grown on a potato dextrose agar medium at 25°C for 6 days. Inoculum spores were suspended in deionized water at 2 × 10⁵ CFU/ml, quantified using microscopy. Apple leaves from 4-week-old plants were inoculated by swabbing them with a spore suspension of the ALT1 pathogen (Bai et al., 2011), 4 days after Agrobacterium tumefaciens infiltration (Ma et al., 2014).

Small RNA was isolated from 4-week-old plants of the ALT1-susceptible GD apple cultivar at 24 hpi with ALT1, or after a mock (control) inoculation. The small RNA libraries for Next-Generation sequencing were constructed by the 5′-phosphate-dependent method as described (Chellappan and Jin, 2009), and the libraries were sequenced using the Illumina HiSeq 2000 platform.

The 20–24-nt small RNA clones were compared with the miRBase database (www.mirbase.org/) to identify novel miRNA sequences. Then, miRNAs in miRBase and our potentially novel miRNA were compared with the apple genome (https://www.rosaceae.org/species/malus/all) by SOAP (Simple Object Access Protocol, http://soap.genomics.org.cn/soap1/). RNA secondary structure of the candidate sequences (http://mfold.rna.albany.edu/) was analyzed according to the method of Xie et al. (2010), followed by screening against the apple protein library (http://genomics.research.iasma.it/, 3,817 sequences) and CDS sequence library (http://genomics.research.iasma.it/, 336,889 sequences), to remove coding sequence.

Putative target genes of the miRNAs were identified by aligning the sequence of the miRNAs to the sequence of the apple genome, using the Plant Small RNA Target Analysis Server (http://plantgrn.noble.org/psRNATarget/). Detailed annotation information about the targeted genes, including their nucleotide sequences, chromosome locations, and predicted protein domains, was obtained from the Genome Database for Rosaceae (https://www.rosaceae.org) and the National Center for Biotechnology Information genomic database (http://www.ncbi.nlm.nih.gov). Specific primers for MdWRKYN1 and MdWRKY26 are listed in Supplemental Table 4.

Leaf samples were harvested from control and ALT1-infected plants at 24 hpi, frozen in liquid nitrogen, and stored at −80°C until required. RNA was extracted from 100 mg of leaf tissue using the cetyltrimethylammonium bromide (CTAB) method (Gambino et al., 2008). The concentration of RNA was determined using an ND-1000 NanoDrop spectrophotometer (Thermo Fisher Scientific, USA). DNase-treated RNA (1 μg) was used in the reverse-transcription reaction and subsequent PCR amplification. A 1:1 ratio of miRNA and reference gene reverse-transcription primers (Supplemental Table 3) were used for the reverse-transcription reaction (Feng et al., 2009). Real-time PCR was performed using SuperReal PreMix Plus (Tiangen Biotech Co., Ltd., China), with 40 cycles of 95°C for 10 s and 60°C for 30 s performed using an Applied Biosystems 7500 (Thermo Fisher Scientific, USA). The relative abundance of the miRNAs was calculated using the 2^−ΔΔCT method (Livak and Schmittgen, 2001), normalized using 5S rRNA as the reference gene (Ma et al., 2014). Specific primers for the mature miRNAs and 5S rRNA are listed in Supplemental Table 3.

Total RNA was extracted from apple leaves using an EASY Spin Kit (Beijing Biomed Biotechnology Co., Ltd., China), amplified using oligo-dT primers (Takara Biomedical Technology Co., Ltd., China), and reverse-transcribed into cDNA (see Supplemental Table 3 for primers). Real-time PCR was performed using SuperReal PreMix Plus (SYBR Green) and the conditions outlined above. The relative RNA abundance was calculated using the 2^−ΔΔCT method (Livak and Schmittgen, 2001), using MdActin (NCBI XM_008365636.2) as the reference gene. Specific primers for MdWRKYN1, MdWRKY26, the 12 PR genes, and MdActin are listed in Supplemental Table 4.

For the real-time PCR, a total of nine leaves from three different plants subjected to the same treatment were collected as one sample for RNA isolation, reverse transcription, and real-time PCR. The same numbers of non-infiltrated GD seedlings and EV (pFGC5941)-infiltrated GD plants served as controls. The three real-time PCR reactions were regarded as three technical repeats for each sample. The same experiment was performed a total of three times for the t-test analysis.

To calculate the disease rate, about 30 leaves from 10 GD apple seedlings were infiltrated by Agrobacterium tumefaciens. The same numbers of non-infiltrated GD seedlings and EV (pFGC5941)-infiltrated GD plants served as controls. All plants were inoculated with a spore suspension of the ALT1 pathogen. The disease rate in the inoculated seedlings was calculated at 24 hpi. These experiments were performed a total of three times, and a t-test was performed for the statistical analysis.

To examine apple miRNA accumulation in response to infection by ALT1, we performed next-generation sequencing of small RNAs (sRNAs). RNA was isolated from 4-week-old plants of the ALT1-susceptible GD cultivar of apple at 24 h post-inoculation (hpi) with ALT1, or after a mock (control) inoculation. After removing the poor-quality reads and adapter sequences, we obtained 34,098,778 high-quality clean reads from the control GD leaves and 28,875,686 clean reads from ALT1-inoculated plants. Unique reads were aligned to the genome of Malus × domestica. For annotation purposes, the sRNAs were classified into several classes (Figure 1A). The 24- and 21-nt sRNA species were the first and second most frequently observed lengths, respectively, in both the ALT1- and non-inoculated leaves (Figure 1B).

We identified miRNAs by performing a homology analysis on the sRNA sequences, using the following selection criteria; a length of at least 18 nt, and a maximum of two mismatches compared with the miRNA databases (miRBase and plant microRNA database) for the apple genome. The selected sequences were screened for predicted secondary structures matching miRNA standards, which resulted in a list of 319 and 308 previously identified miRNAs in the non-inoculated and ALT1-inoculated apple leaves, respectively, also according to the current understanding of miRNA composition and on base distribution data (Supplemental Figure 1). The majority of the sRNAs were unannotated. We mapped the unknown sRNAs to the reference sequence to identify novel miRNAs using Mireap software. The stem-loop structures matched to putative miRNA precursors were used to predict the characteristic fold-back RNA secondary structure, resulting in the identification of 56 potentially novel apple miRNAs. Although some of the novel miRNAs identified exists in other species, they had not hitherto been reported in apple. Therefore, we consider these to be novel miRNAs in apple.

Among the sequenced miRNAs, we identified 39 previously identified miRNAs/miRNA families (Supplemental Table 1) and 19 novel miRNAs (Supplemental Table 2) that exhibited at least a 2-fold difference in expression level following ALT1 infection (Figure 2). Plant miRNAs bind almost perfectly to the complementary sequences of their target genes, regulating their post-transcriptional processing by inducing transcript degradation or translational inhibition. Identifying and validating the targets of the miRNAs is therefore important for elucidating their potential biological function. A total of 578 and 209 candidate target transcripts were identified for the previously identified miRNAs/miRNA families and novel miRNAs, respectively.

We conducted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses to better understand these differentially expressed miRNAs and their target genes. According to the GO classification, the category containing the most target genes was “biological process,” which is sub-categorized into the 20 biological processes displayed in Supplemental Figure 2. In particular, the defense response (GO: 0006952), signal transduction (GO: 0007165), transcription factor (GO: 0006355), and oxidation-reduction processes (GO: 0055114) were significantly enriched (Supplemental Figure 2A). The KEGG-based analysis linked 39 miRNAs/miRNA families and 578 target genes to 42 pathways, with a significant enrichment of the plant hormone signal transduction (mdm04075), plant–pathogen interaction (mdm04626), and mRNA surveillance (mdm03015) pathways (Supplemental Figure 2B). These two analyses indicated that transcription factors play a crucial role in the response to ALT1 infection in GD leaves.

Notably, several miRNAs that putatively regulate transcription factors were differentially expressed under ALT1 infection, and were therefore investigated in greater depth (Supplemental Figure 3). Md-miR156 family targets several genes encoding members of the SPL transcription factor family, which are reported to participate in the response to abiotic stress (Stief et al., 2014; Wang and Wang, 2015). In our libraries of differentially expressed miRNAs, one member of the Md-miR156 family, Md-miR156x (miRBase MIMAT0025890; miRBase website: http://www.mirbase.org/), showed a 3.78-fold change in abundance following ALT1 infection (Figures 2, 3). This suggests a potential role for the Md-miR156x target gene MdSPL19 (NCBI XM_008343068.1; NCBI website: https://www.ncbi.nlm.nih.gov/) in plant immunity. Another Md-miR156 paralog, like Md-miR156a (miRBase MIMAT0025867), had a smaller change in relative expression than Md-miR156x, which also targeted a member of the SPL family (Figure 2). The expression of Md-miR164 (miRBase MIMAT0025908), which targets a NAC-domain transcription factor, MdNAC (NCBI XM_008376341.1), displayed a 4.35-fold upregulation at 24 hpi with ALT1, which functions as a stress-responsive transcription factor (Puranik et al., 2012; Zhu et al., 2014; Figures 2, 3). Md-miR166 (miRBase MIMAT0025913), which showed a 2.32-fold higher accumulation in the ALT1-infected apple leaves than in the leaves of the control plants, targets the homeobox-leucine zipper protein MdREVOLUTA (NCBI XM_008346650.2; Figures 2, 3). Md-miR159a (miRBase MIMAT0025898; 2.99-fold upregulation after ALT1 infection) and Md-miR159c (miRBase MIMAT0026053; 2.88-fold upregulation after ALT1 infection) target the MYB transcription factor genes MdMYB86 (NCBI XM_008387744.1) and MdGAMYB (NCBI XM_008348120.1), respectively. The MYB TFs reported in Arabidopsis are also regulated by miR159 and function in plant growth and the response to stress (Reyes and Chua, 2007; Alonso-Peral et al., 2010; Figures 2, 3). Md-miR396a (miRBase MIMAT0025989) and Md-miR396f (miRBase MIMAT0025994), encoded by genes with two nucleotide differences, regulate conserved targets MdGRF1 (NCBI XM_008343874.2) and MdGRF3 (NCBI XM_008345268.1), which belong to the GROWTH-REGULATING FACTOR (GRF) family of transcription factors expressed during growth and stress responses (Debernardi et al., 2012; Figures 2, 3).

Notably, except for the apple miRNAs/miRNA families mentioned above (Md-miR156, Md-miR164, Md-miR166, Md-miR159, and Md-miR396) that target five transcription factor families, Md-miR395 (miRBase MIMAT0025980) and Md-miR156ab (miRBase MIMAT0025894) exhibited the largest fold change in abundance after ALT1 infection (Figure 2). Interestingly, Md-miR395 and Md-miR156ab both exhibit the characteristic fold-back RNA secondary structure (Supplemental Figures 4A,B) and target genes encoding WRKY transcription factors (Figure 3, Supplemental Figure 3). Md-miR156ab, which had a 13.69-fold expression difference after ALT1 infection, targets an atypical novel WRKY transcription factor, MdWRKYN1 (NCBI XM_008353805.1), which contains a Toll/interleukin-1 receptor/resistance protein (TIR) domain as well as a WRKY domain (Figures 2, 3, 4A). Md-miR395 had a 4.98-fold upregulation following ALT1 infection, and regulates a typical WRKY transcription factor, MdWRKY26 (NCBI XM_008386494.2), containing two WRKY domains (Figures 2, 3, 4A). WRKY transcription factors are key regulatory components of plant responses to biotic and abiotic stresses. Generally, the expression of WRKY genes is significantly upregulated following pathogen infection and treatment with salt, drought, and salicylic acid (van Verk et al., 2008; Peng et al., 2012; Choi et al., 2015).

To validate the reliability of the high-throughput sequencing, we analyzed the expression of specific miRNAs targeting transcription factors. Overall, the expression trends of these selected miRNAs determined by real-time PCR were consistent with the sequencing results (Figure 3A), indicating the high reliability of the analysis. To validate whether the predicted target genes were regulated by miRNAs under ALT1 infection, we conducted an expression analysis of the predicted target genes and their corresponding miRNAs. The predicted target genes were downregulated during ALT1 infection, which coincided with an upregulation of their corresponding miRNAs (Figure 3B).

The expression patterns and complementary sequences of Md-miR156ab and MdWRKYN1 suggest that Md-miR156ab is a suppressor of MdWRKYN1 protein biosynthesis. We overexpressed Md-MIR156ab by transforming the plant with a constructed binary vector (pFGC5941) containing the Md-miR156ab primary transcript (Figures 4B,C). Md-miR156ab abundance was significantly increased in these apple lines (Figure 4D), which led to a significant downregulation in the levels of MdWRKYN1 mRNA (Figure 4D). This analysis was also performed using a vector containing Md-MIR395 (Figure 4B). Md-miR395 expression was significantly upregulated in these transformed lines (Figure 4C), and the levels of its target mRNA MdWRKY26 were significantly downregulated (Figure 4D).

The WRKY transcription factors bind to W-box domains (TTGACC/T) in the promoters of their target genes, including PR proteins; thus, we focused on 12 PR genes with a W box in their promoter whose expression in GD was altered in response to ALT1 infection (Supplemental Figure 5; van Verk et al., 2008; Peng et al., 2012; Choi et al., 2015). The PATHOGENESIS-RELATED PROTEINS 1 (MdPR1; NCBI NM_001311210.1) and MdPR5 (NCBI XM_008380268.2) genes contained one W box in their promoters; GLUTATHIONE PEROXIDASE 7 (MdGPX7; NCBI XM_008349268.2), L-ASCORBATE PEROXIDASE 3 (MdAPX-2; NCBI XM_008354210.2), MAJOR ALLERGEN MAL D 1-LIKE 1 (MdPR10-1; NCBI XM_008352950.2), and MdPR10-2 (NCBI NM_001294363.1) contained two W-box sites in their promoters; ENDOCHITINASE-LIKE (MdPR3-1; NCBI XM_008395323.1), ACIDIC ENDOCHITINASE-LIKE (MdPR3-2; NCBI XM_008382187.2), MdPR8 (NCBI AM600694.1), and MdAPX-1 (NCBI XM_008387034.1) harbored three W-box sites in their promoters; while MdPR2 (NCBI AM600693.1) and CLASS II CHITINASE (MdPR4; NCBI AF494397.1) promoters contained four WRKY binding sites (Supplemental Figure 5).

To investigate the function of two WRKY transcription factors, the relationships between Md-miR156ab-regulated MdWRKYN1 and Md-miR395-regulated MdWRKY26 and PR genes were investigated during ALT1 infection in apple.

GD leaves were transformed to overexpress MdWRKYN1 (Figure 5A). Four days after infiltration, MdWRKYN1 expression was found to be significantly upregulated in these plants (Figure 5B). After ALT1 inoculation, the disease rate in the MdWRKYN1-overexpressing plants (7.50%) was significantly lower than in the wild-type (WT; 38.59%) and empty-vector (EV; 39.96%) plants, and the symptoms of ALT1 infection were suppressed. Resistance to apple leaf spot disease was thus increased in MdWRKYN1-overexpressing plants (Figures 5C,D, Supplemental Table 5). Real-time PCR showed that the expression levels of MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR10-1, and MdPR10-2 were increased in OE-MdWRKYN1 plants after ALT1 inoculation compared with the controls (Figure 5E). Overexpressing MdWRKYN1 in the absence of ALT1 inoculation confirmed that the six upregulated PR genes were induced by the MdWRKYN1 transcription factor (Figure 5E). These results suggest that ALT1 infection induces Md-miR156ab expression, which suppresses MdWRKYN1 transcription. Overexpressing MdWRKYN1 enhances the disease resistance of GD through activation of MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR10-1, and MdPR10-2 expression.

We synthesized a Short Tandem Target Mimic (STTM) sequence to block the target sites of Md-miR156ab, fused it into the vector pFGC5941 (Figure 5F), and transformed it into GD apple (Tang et al., 2012; Yan et al., 2012; Tang and Tang, 2013; Reichel et al., 2015). After 4 days, real-time PCR indicated a significant decrease in the abundance of Md-miR156ab expression (Figure 5G). The STTM-miR156ab plants were inoculated with ALT1, and at 24 hpi these plants had a significant increase in MdWRKYN1 expression (Figure 5H), further confirming Md-miR156ab as a regulator of MdWRKYN1 expression in apple. The STTM-miR156ab-expressing plants had a significantly lower rate of disease (12.83%) than the WT (36.26%) or EV (38.40%) plants, and showed a greater resistance to apple leaf spot disease (Figures 5I,J, Supplemental Table 6). Real-time PCR results showed that MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR10-1, and MdPR10-2 expression levels were significantly higher in the STTM-miR156ab-expressing plants following ALT1 inoculation than in the wild-type and empty-vector controls (Figure 5K). Thus, we concluded that resistance to apple leaf spot disease is under the influence of the PR genes, whose accumulation is regulated by MdWRKYN1, which is suppressed by Md-miR156ab.

To investigate whether MdWRKY26, which contains two WRKY domains, could play a role in plant defense against apple leaf spot disease, we overexpressed MdWRKY26 and silenced its regulatory miRNA, Md-miR395. We constructed a binary vector (pFGC5941) containing the coding sequence of MdWRKY26 (Figure 6A), then transformed it into the susceptible apple variety GD. Four days after transformation, MdWRKY26 expression was significantly upregulated (Figure 6B). The transformed leaves were inoculated with ALT1; at 24 hpi, 34.41% of the WT plants and 35.26% of the EV plants showed symptoms of leaf spot disease, while only 4.92% of the OE-MdWRKY26 plants showed susceptibility to ALT1 (Figures 6C,D, Supplemental Table 7). Real-time PCR analysis showed that MdPR1, MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR8, MdPR10-1, and MdPR10-2 expression increased in OE-MdWRKY26 plants either with or without ALT1 inoculation, relative to the wild-type and empty-vector controls (Figure 6E). These results suggest that ALT1 infection induces the production of Md-miR395, which suppresses MdWRKY26, while the overexpression of MdWRKY26 enhances the disease resistance of GD through the induction of MdPR1, MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR8, MdPR10-1, and MdPR10-2 expression.

We synthesized a STTM sequence to block the binding of Md-miR395 to its target sequences, fused it into the vector pFGC5941 (Figure 6F), and expressed STTM-miR395 in the susceptible apple variety GD. After 4 days, real-time PCR indicated a significant decrease in Md-miR395 transcripts (Figure 6G). At 24 hpi with ALT1, we identified an alleviation of the Md-miR395-induced suppression of MdWRKY26 transcripts (Figure 6H), further confirming Md-miR395 as a regulator of MdWRKY26 expression in apple. The STTM-miR395-expressing plants had a lower rate of disease (15.71%) compared to the WT (34.37%) and EV (36.69%) plants, indicating an improved resistance to apple leaf spot disease (Figures 6I,J, Supplemental Table 8). Real-time PCR showed that MdPR1, MdPR3-1, MdPR3-2, MdPR4, MdPR5, MdPR8, MdPR10-1, and MdPR10-2 expression levels were significantly higher in the STTM-miR395-expressing plants after ALT1 inoculation than in the wild-type and empty-vector controls (Figure 6K). We therefore preliminarily consider that apple leaf spot disease may be related, in the pathosystem we have investigated, to the PR genes, whose accumulation is regulated by MdWRKY26 in the absence of suppression by Md-miR395.

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.