Vm‐milR37 contributes to pathogenicity by regulating glutathione peroxidase gene VmGP in Valsa mali

Abstract MicroRNAs play important roles in various biological processes by regulating their corresponding target genes. However, the function and regulatory mechanism of fungal microRNA‐like RNAs (milRNAs) are still largely unknown. In this study, a milRNA (Vm‐milR37) was isolated and identified from Valsa mali, which causes the most serious disease on the trunk of apple trees in China. Based on the results of deep sequencing and quantitative reverse transcription PCR, Vm‐milR37 was found to be expressed in the mycelium, while it was not expressed during the V. mali infection process. Overexpression of Vm‐milR37 did not affect vegetative growth, but significantly decreased pathogenicity. Based on degradome sequencing, the target of Vm‐milR37 was identified as VmGP, a glutathione peroxidase. The expression of Vm‐milR37 and VmGP showed a divergent trend in V. mali–apple interaction samples and Vm‐milR37 overexpression transformants. The expression of VmGP could be suppressed significantly by Vm‐milR37 when coexpressed in tobacco leaves. Deletion of VmGP showed significantly reduced pathogenicity compared with the wild type. VmGP deletion mutants showed more sensitivity to hydrogen peroxide. Apple leaves inoculated with Vm‐milR37 overexpression transformants and VmGP deletion mutant displayed increased accumulation of reactive oxygen species compared with the wild type. Thus, Vm‐milR37 plays a critical role in pathogenicity by regulating VmGP, which contributes to the oxidative stress response during V. mali infection. These results provide important evidence to define the roles of milRNAs and their corresponding target genes in pathogenicity.

The sRNAs are loaded into Argonaute (AGO) proteins, which are the core component of the RNA-induced silencing complex (RISC).
A guide RNA directs the RISC to complementary message RNAs (mRNAs), resulting in mRNA cleavage or repression of translation (Chang et al., 2012;Holoch & Moazed, 2015). Various studies have shown that the RNAi pathway plays important roles in growth, development, reproduction, and response to biotic or abiotic stresses in eukaryotes (Ghildiyal & Zamore, 2009;Katiyar-Agarwal & Jin, 2010).
The first RNAi description in fungi was reported in Neurospora (Romano & Macino, 1992). Since then, the identification and characterization of RNAi components have deepened our understanding of fungal RNAi (Nakayashiki et al., 2006). There is evidence that fungal RNAi plays important roles in maintenance of growth, antiviral defence, sexual development, and pathogenicity (Jin et al., 2019;Raman et al., 2017;Son et al., 2017;Sun et al., 2009;Torres-Martínez & Ruiz-Vázquez, 2017;Weiberg et al., 2013). However, the detailed mechanisms of regulation by sRNAs in fungi are still largely not understood. miRNAs are 21-nucleotide endogenous RNAs generated from single-stranded RNA with a hairpin structure. The various functions and corresponding regulatory mechanism of miRNAs in plants and animals have been reported (Bartel, 2004;Grimson et al., 2008;Llave et al., 2002). However, it was believed that miR-NAs were absent in fungi. In 2010 a class of sRNAs that have a similar generation pathway and regulatory mechanism to miRNAs in plants and animals was identified in Neurospora and designated as microRNA-like RNAs (milRNA) (Lee et al., 2010). Further milR-NAs were isolated and functionally analysed in fungi by the application of deep-sequencing technology. In Metarhizium anisopliae and Trichoderma reesei, milRNAs were predicted to be related to mycelial growth and sporulation Zhou et al., 2012b). The milRNAs of Penicillium marneffei regulate the growth process of mycelial and yeast phases (Lau et al., 2013). milRNAs are speculated to be important regulators for toxin biosynthesis in Aspergillus flavus (Bai et al., 2015).
In recent years, the function of milRNAs isolated from plant-pathogenic fungi have been analysed. In Sclerotinia sclerotiorum, 44 candidate milRNAs were identified and predicted to be associated with sclerotial development (Zhou et al., 2012a). mil-RNAs of Fusarium oxysporum f. sp. niveum play important roles in the biosynthesis of fungal toxins (Jiang et al., 2017). In Rhizoctonia solani, several milRNAs affect pathogenicity by regulating many important pathogenic factors (Lin et al., 2016). In addition, studies of milRNAs and their targets in Curvularia lunata, F. oxysporum, Zymoseptoria tritici, and Puccinia striiformis f. sp. tritici suggested that milRNAs were also associated with pathogenicity and development (Chen et al., 2014;Liu et al., 2016;Mueth et al., 2015;Yang, 2015).
However, the function of most milRNAs was only based on the target prediction; the detailed regulatory mechanism of milRNAs was not elucidated. Recent studies on Verticillium dahliae demonstrated that VdmilR1 can suppress target gene expression by epigenetic repression to regulate pathogenicity (Jin et al., 2019). Pst-milRNA1 was found to contribute to pathogenicity by suppressing the expression of the host wheat pathogenesis-related 2 gene (Wang et al., 2017a).
Thus, fungal milRNAs may have multiple functions by regulating different targets, and it is of great interest to identify the regulatory mechanism of different milRNAs.
Valsa mali is an important phytopathogenic fungus, causing the most serious trunk disease of apple trees .
Revealing the pathogenic mechanism of V. mali will lay a foundation for the development of sustainable disease control strategies.
Several pathogenicity factors have been characterized based on genome, transcriptome, and functional genomics (Ke et al., 2014;Wu et al., 2018;Xu et al., 2018;Yin et al., 2015;Zhang et al., 2019). The RNAi pathway components of V. mali, such as Dicer-like and AGO, were identified as associated with pathogenicity, which indicated posttranscriptional regulation may also be an important pathway (Feng et al., 2017a,b). Multiple omics analyses revealed that Vm-milRNAs can regulate pathogenicity factors to promote V. mali infection (Xu et al., 2020). However, the detailed regulatory mechanism of Vm-milRNAs is still largely not understood.
In this study, Vm-milR37 was specifically expressed in mycelia, but poorly expressed during infection. Functional analysis of Vm-milR37 showed that it was negatively involved in pathogenicity. Its target was confirmed to be a glutathione peroxidase gene, VmGP, based on degradome sequencing and cotransformation results.
VmGP was confirmed to contribute positively to pathogenicity. This study indicates that Vm-milR37 contributes to pathogenicity by enhancing the expression of VmGP during V. mali infection. The results help to uncover the posttranscriptional regulatory mechanism directed by milRNAs of V. mali.

| Vm-milR37 shows expression in mycelia but no expression during V. mali infection
In our previous study, Vm-milR37 was isolated from a cDNA library of the mycelium of V. mali, which was generated from a precursor with a typical hairpin structure ( Figure S1). Almost no reads of Vm-milR37 were detected in the cDNA library of the V. mali-apple interaction.
To determine whether Vm-milR37 is involved in the pathogenicity of V. mali, the expression trend of Vm-milR37 during V. mali infection was analysed by stem-loop reverse transcription PCR. Consistent with the sequencing results, Vm-milR37 was expressed in mycelia at a high level, but showed nearly no expression during V. mali infection ( Figure 1). The result indicates a potential role in the regulation of pathogenicity of V. mali.

| Overexpression of Vm-milR37 decreased the pathogenicity of V. mali
To further examine the function of Vm-milR37 on the pathogenicity of V. mali, Vm-milR37 overexpression transformants were generated ( Figure S2). Two randomly selected Vm-milR37 overexpression transformants (Vm-milR37-OE-1 and Vm-milR37-OE-11) were selected for further analysis. Compared with the wild type, the expression levels of Vm-milR37 in Vm-milR37-OE-1 and Vm-milR37-OE-11 were en- The biomass of Vm-milR37 overexpression transformants in twigs was significantly less than in twigs infected with the wild-type and the empty vector transformant (Figure 3e). These results indicate that Vm-milR37 might play a negative role in pathogenicity of V. mali.

| Isolation and annotation of target gene of Vm-milR37
In a previous study, VM1G_06866, which encodes a glutathione peroxidase, a protein of 229 amino acids with typical glutathione peroxidase conserved domains ( Figure S3), was identified as a target gene of Vm-milR37 (Xu et al., 2020). VM1G_06866 was designated as VmGP. VM1G_06866 is predicted by BLAST searches with the pathogen-host interactions database (PHI-base) to possibly be a virulence gene (Urban et al., 2017). We speculated that Vm-milR37 could be involved in pathogenicity by regulating the expression of VM1G_06866.
To study the phylogeny of glutathione peroxidase, 15 homologous plant proteins and 18 homologous fungal proteins were identified and used to establish a neighbour-joining phylogenetic tree ( Figure 4). VmGP (KUI71622) clustered with the glutathione peroxidase from fungi as expected, and it was highly homologous with the glutathione peroxidase of Valsapyri (KUI52717) and V. dahliae (XP_009658518). The glutathione peroxidase proteins from plants clustered together and the glutathione peroxidase proteins from fungi clustered together, suggesting a common evolutionary origin.

| Validation of the target gene of Vm-milR37
To determine whether the expression of VmGP could be regulated by Vm-milR37, the expression patterns of Vm-milR37 and VmGP were analysed. According to the transcriptome data, VmGP was highly F I G U R E 1 Expression patterns of Vm-milR37 detected by stem-loop reverse transcription PCR. Small nuclear RNA U6 (VmU6) of Valsa mali was used as the internal control. Vm-milR37 was expressed in vitro in mycelia and was very weakly expressed during V. mali infection. Similar results were observed in three biological repeats F I G U R E 2 Relative expression level of Vm-milR37 in Vm-milR37 overexpression transformants in vitro and in planta. (a) Vm-milR37 was overexpressed in Vm-milR37 overexpression (OE) transformants in mycelia cultured in vitro. Total RNA of mycelia was extracted and the transcript level of Vm-milR37 was detected by stem-loop reverse transcription PCR. Relative expression of Vm-milR37 was normalized to internal control Valsa mali small nuclear RNA U6 (VmU6) and calibrated to the levels of the wild-type strain (WT) by the 2 −ΔΔCt method. (b) Vm-milR37 showed enhanced transcription level in Vm-milR37 overexpression transformantinfected apple bark tissues at 24 hr postinoculation (hpi). Apple bark tissues inoculated with the WT mycelia at 0 hr postinoculation (hpi) was used as the control. In (a) and (b), mean ± SD was calculated from three independent biological repeats. Data were analysed using Dunnett's multiple comparison test. *p < .05

| VmGP contributes to the pathogenicity by affecting the oxidative stress response
To determine the function of VmGP in vegetative growth and pathogenicity, deletion mutants of VmGP were generated ( Figure S9a).
There was no distinct difference in vegetative growth between deletion mutants and the wild type (Figure 7a

| D ISCUSS I ON
miRNAs were thought to be absent in fungi until a similar small RNA in Neurospora was identified to be milRNA (Lee et al., 2010).
In contrast to research on plant and animal miRNAs, research on fungal milRNAs is less advanced. Many fungal milRNAs have been sequenced recently, such as from Metarhizium anisopliae (Zhou et al., 2012b), Penicillium marneffei (Lau et al., 2013), and Aspergillus flavus (Bai et al., 2015). milRNAs of some plant-pathogenic fungi have been isolated, and these were predicted to be associated with vegetative growth and development, and pathogenicity by inhibiting the expression of endogenous genes, even using cross-kingdom F I G U R E 4 Phylogenetic analysis of VmGP. The phylogenetic tree was constructed with neighbour-joining method using MEGA 7. Bootstrap values were set as 1,000. VmGP is highlighted in a red box.  (Chen et al., 2015;Jin et al., 2019;Lin et al., 2016;Liu et al., 2016;Wang et al., 2017;Zhou et al., 2012a). However, the detailed functions and regulatory mechanisms of milRNAs are still largely unknown. In this study, a milRNA, Vm-milR37, was isolated from the plant-pathogenic fungus V. mali. The function and regulatory mechanism of Vm-milR37 were confirmed to be associated with pathogenicity by regulating the expression of VmGP.
The precursor of Vm-milR37 was generated from an endogenous transcript that could fold to a typical hairpin structure, which meets the criterion for defining a fungal milRNA (Lee et al., 2010).
Vm-milR37 was specifically expressed in mycelium, and nearly no expression was detected during the host infection process. Thus, we speculated that Vm-milR37 may play an important role in vegetative growth and pathogenicity. To analyse the function of miRNAs or mil-RNAs, overexpression is an important and widely used method (Jin et al., 2019;Li et al., 2013;Xu et al., 2020). Based on overexpression of Vm-milR37 and pathogenicity assays, Vm-milR37 was confirmed to be involved in the pathogenicity of V. mali. milRNAs function by regulating the corresponding target genes (Bartel, 2004). Thus, target identification is critical for exploring the regulatory mechanism of milRNAs. In this study, Vm-milR37 negatively regulated the pathogenicity of V. mali. As Vm-milR37 In plants and animals, miRNAs play important regulatory roles by targeting mRNAs for cleavage or translational repression (Bartel, 2004). In fungi, the regulatory mechanism of milRNAs is largely unknown. In V. dahliae, milRNA1 is involved in fungal virulence by transcriptional repression (Jin et al., 2019). Arabidopsis sRNAs and cotton miRNAs can be transported into fungal cells and silence fungal target transcripts by mRNA cleavage, which indicates that sRNA-mediated mRNA cleavage exists in fungi (Cai et al., 2018;Zhang et al., 2016). In this study, VmGP was identified to be the target gene of Vm-miR37, with subsequent mRNA cleavage. As core components of the RISC, AGOs can perform a cleavage function when sRNAs guide them by binding with the corresponding target genes (Azlan et al., 2016). Based on the published genome sequence of V. mali, three proteins have been idenitfied as VmAGOs (Yin et al., 2015). If mil-RNAs play roles in the mRNA cleavage pathway, the corresponding mRNA ends will have a 5′ phosphate and this character can be used for identification of milRNA target genes (German et al., 2008). The target genes of milRNAs in V. mali were detected based on degradome sequencing, and they could be regulated by milRNAs by mRNA cleavage (Xu et al., 2020). Degradome sequencing of F. oxysporum and F. graminearum also revealed that RNAi-mediated gene suppression can function at the posttranscriptional level (Chen et al., 2014;Son et al., 2017). Thus, mRNA cleavage mediated by endogenous milRNA may be a critical regulatory mechanism in fungi.
VmGP encodes a glutathione peroxidase, and it was demonstrated to play a critical role in pathogenicity and the oxidative stress response of V. mali. In Magnaporthe oryzae, a GPx has been shown to be required for H 2 O 2 resistance and fungal virulence . The glutathione peroxidase of Alternaria alternata is associated with ROS resistance and full virulence (Yang et al., 2016). Plant cells trigger an oxidative burst with a rapid increase of ROS production to defend against pathogen infection (Auh & Murphy, 1995). To infect successfully, pathogens have to increase their tolerance to these ROS. Glutathione peroxidase is a key enzyme to degrade H 2 O 2 (Aung- Htut et al., 2011). Previous studies have demonstrated that the ability to detoxify ROS is required for A. alternata survival and pathogenesis (Chung, 2012;Lin et al., 2009). Thus, we speculated that VmGP may contribute to full pathogenicity by enhancing tolerance to H 2 O 2 from the host plant.  & Jin, 2010). In pathogenic fungi, many virulence genes have been predicted and confirmed to be regulated by sRNAs (Gowda et al., 2010;Guo et al., 2019;Jin et al., 2019;Raman et al., 2017;Xu et al., 2020). We also found that many virulence genes of V. mali could be regulated by milRNAs (Xu et al., 2020). In this study, VmGP, as an important virulence gene, was further demonstrated to be regulated F I G U R E 7 Deletion of VmGP reduces the pathogenicity of Valsa mali. (a) Colony morphology of wild type (WT), VmGP deletion mutants, and complementation strain after 48 hr incubation. (b) Colony diameters of WT, VmGP deletion mutants, and complementation strain aafter 48 hr incubation. Data represent mean ± SD. The experiment was repeated three times, each time with three plates. (c) and (d) Pathogenicity test of WT, VmGP deletion mutants, and complementation strain at 4 days postinoculation. Three representative diseased twigs are shown. The pathogenicity test was independently repeated three times, each time with four replicates. CK, negative control. Data represent mean ± SD. (e) V. mali biomass was measured with quantitative PCR using V. mali-specific VmG6PDH primers. V. mali biomass was normalized to the mean of the WT. Data are mean ± SD of three technical replicates. Similar results were obtained from three biological repeats. (f) WT, VmGP deletion mutant, and complementation strain on potato dextrose agar (PDA) + 0.05% (vol/vol) H 2 O 2 . Photographs were taken after 48 hr incubation. (g) Colony diameters of WT, deletion mutant of VmGP, and complementation mutant on PDA + 0.05% (vol/vol) H 2 O 2 after 48 hr incubation. Data represent mean ± SD. The experiment was repeated three times, each time with three plates. Significant difference was determined using Dunnett's multiple comparison test. *p < .05 by Vm-milR37 at the posttranscriptional level. When the fungus does not need to express the virulence gene, the fine-tuning mode of sRNA is activated. We speculate that this mechanism is beneficial for the fungus to save energy to enhance the adaption capacity and pathogenicity.
Overall, this study demonstrates that a milRNA, Vm-milR37 from V. mali, plays a critical role in pathogenicity by regulating the endogenous target gene VmGP, which contributes to the oxidative stress response during V. mali infection. These results provide important evidence to define the roles of milRNAs and their corresponding target genes in fungal pathogenicity.

| Strains and growth conditions
The wild-type strain of V. mali 03-8 was used to generate transformants of Vm-milR37 and VmGP. All the strains were cultured on PDA at 25 °C in the dark. Escherichia coli DH5α was cultured in lysogeny broth at 37 °C. Agrobacterium tumefaciens GV3101 was cultured in lysogeny broth at 28 °C.

| Expression profiles of Vm-milR37 and VmGP
Mycelial plugs (5 mm diameter) of V. mali were inoculated onto twigs of Malus × domestica 'Fuj' as described by Wei et al. (2010). To investigate the function of Vm-milR37 and VmGP during the V. mali-apple bark interaction, the junction of healthy and infected apple bark tissue inoculated with V. mali for 6, 12, 24, 48, and 72 hr was collected. Samples of V. mali mycelium cultured for 3 days were collected as a control (0 hr postinoculation [hpi]) from PDA plates covered with a layer of cellophane.
Total RNA of each sample was extracted with TRIzol reagent (Invitrogen) following the manufacturer's instructions. RNA purity, concentration, and integrity were checked. First-strand cDNA was synthesized using a reverse transcription (RT)-PCR system (Promega) following the manufacturer's instructions. The expression level of Vm-milR37 was detected by stem-loop RT-PCR described by Feng et al. (2012). Small nuclear RNA U6 of V. mali (VmU6) was used as an internal control. The expression level of VmGP was measured followed the method described by Yin et al. (2013). G6PDH of V. mali was selected as the internal control.
There were three biological replicates for each treatment. Primers used for RT-qPCR are given in Table S1.

| Generation of Vm-milR37 and Mut-R37 overexpression transformants
The precursor of Vm-milR37 was amplified from V. mali genomic DNA using Phusion high-fidelity DNA polymerase (New England Biolabs) and cloned into plasmid pDL2 using the ClonExpress-II One Step Cloning Kit (Vazyme Biotech). The Mut-R37 overexpression construct was generated using the Fast Site-Directed Mutagenesis Kit (Tiangen) following the manufacturer's instructions and the Vm-milR37 overexpression construct as the amplification template.
Constructs were verified by sequencing and transformed into V. mali wild-type strain 03-8 as described above. Transformants were screened by PCR with primer pairs outside the cloning sites of pDL2.
Relative expression profiles of Vm-milR37 and VmGP were measured as described above. All primers used for gene deletion are given in Table S1.

| Target identification of Vm-milR37
Based on the degradome sequencing results, the 3′ untranslated region (UTR) region of VM1G_06866 was identified as the target of Vm-milR37. To verify whether the expression of VM1G_06866 could be suppressed by Vm-milR37, the precursor of Vm-milR37 and the 3′ UTR of VM1G_06866 were cloned into pCAMBIA1302 with GFP as a reporter gene, and the recombinant vectors were cotransformed into the same site of N. benthamiana leaves using F I G U R E 8 Vm-milR37 and VmGP are involved in the oxidative stress response during Valsa mali infection. (a) Apple leaves inoculated with Vm-milR37 overexpression (OE) transformant and VmGP deletion mutant exhibited enhanced reactive oxygen species (ROS) accumulation as compared with the wild type (WT). ROS in apple leaves were detected with 3,3′-diaminobenzidine (DAB) staining at 24 hr postinoculation. (b) ImageJ software was used to quantify ROS accumulation at V. mali invasion sites. Mean ± SD was calculated from four biological repeats. Statistical significant difference was determined using Dunnett's multiple comparison test as compared to the WT. *p < .05 the Agrobacterium-mediated transfection system described by Weiberg et al. (2013). Confocal images were taken at 48 hr post-Agrobacterium infiltration. The GFP fluorescence intensity quantified by confocal microscopy represented the expression of the target gene. Thirty independent N. benthamiana cells were used to detect the fluorescence intensity. Data were analysed using Dunnett's multiple comparison test (p < .05). To further verify the expression of GFP, anti-GFP and anti-actin antibodies (Sungene Biotech) were used for western blot analysis. Horseradish peroxidase-conjugated goat antimouse IgG (Cwbiotech) was used as a secondary antibody. The coexpression experiment was repeated twice independently. All primers used for coexpression are given in Table S1.

| Sequence alignment and phylogenetic analysis
The full length of the target gene was isolated based on the results of degradome sequencing and genome information.

| Generation of target gene deletion mutants and complementation transformants
The NEO gene was a selected as marker gene to perform target gene deletion. The NEO gene fragment was amplified from plasmid pFL2 with primers Neo-F and Neo-R. The NEO fragment was fused with upstream and downstream flanking sequences of the target gene by double-joint PCR (Yu et al., 2004). The gene-replacement construct was transformed into protoplasts of V. mali as previously described (Gao et al., 2011). Each putative single gene deletion mutant was verified by PCR with four primer pairs to detect the target gene, the NEO gene, upstream-NEO fusion segment, and NEO-downstream fusion segment.
To generate the complementation transformants of target gene deletion mutants, the full-length target gene with upstream 2,000 bp was amplified from genomic DNA and cloned into plasmid pDL2 using the yeast gap repair approach (Bruno et al., 2004). The recombinant construct was then transformed into protoplasts of the gene deletion mutant. Complemented transformants were selected using geneticin (G418) and hygromycin, and confirmed by PCR. All primers used for gene deletion are given in Table S1.

| Vegetative growth, pathogenicity, and fungal biomass assays
The vegetative growth of gene deletion mutants and overexpression transformants was assayed as previously described . The tests were performed three times and each experiment included three replicates. Pathogenicity assays were performed on Fuji apple twigs as described (Wei et al., 2010). Lesion length was measured at 4 days postinoculation. The pathogenicity test was repeated three times and each experiment included four replicates. For V. mali biomass assays, samples of 0.4 g apple twig tissues, including the infected tissues and healthy tissue, were collected. Genomic DNA was isolated with the Super Plant Genomic DNA kit (Polysaccharides and Polyphenolics-rich; Tiangen).
V. mali biomass was measured with quantitative PCR using V. malispecific VmG6PDH primers. The biomass assay was independently performed three times, each time with three technical replicates.

| Oxidative stress test
Mycelial plugs (5 mm diameter) from the edge of growing colonies of V. mali strain 03-8 and gene deletion mutants were inoculated on PDA supplemented with 0.05% H 2 O 2 . The colony diameter was determined after 2 days' incubation. The test was performed three times and each experiment included three replicates.

| ROS staining in apple leaves
V. mali strains were inoculated on apple leaves as previously described (Wei et al., 2010). At 24 hpi, apple leaves around the inoculation points were cut into 1 cm 2 pieces and immediately immersed in 1 mg/ml 3,3′-diaminobenzidine (DAB, pH 3.8). After staining for 8 hr in the light, apple pieces were decoloured using 3:1 (vol/ vol) ethanol:chloroform containing 0.15% trichloroacetic acid and saturated chloral hydrate solution. Photographs were taken using a DP72 camera (Olympus). ROS accumulation in V. mali invasion sites was quantified with ImageJ software. The relative amount of ROS was normalized to the mean of leaves inoculated with the wild type.
The test was performed three times.

ACK N OWLED G EM ENTS
We thank Professor Jin-Rong Xu at Purdue University for providing plasmids pDL2 and pFL2. This work was supported by the National Natural Science Foundation of China (31501591).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.