MfOfd1 is crucial for stress responses and virulence in the peach brown rot fungus Monilinia fructicola

Abstract Monilinia fructicola is the most widely distributed species among the Monilinia genus in the world, and causes blossom blight, twig canker, and fruit rot on Rosaceae fruits. To date, studies on genomics and pathogenicity are limited in M. fructicola. In this study, we identified a redox‐related gene, MfOfd1, which was significantly up‐regulated at 1 hr after inoculation of M. fructicola on peach fruits. We used the clustered regulatory inter‐spaced short palindromic repeats (CRISPR)/Cas9 system combined with homologous recombination to determine the function of the MfOfd1 gene. The results showed that the sporulation of knockdown transformants was reduced by 53% to 83%. The knockdown transformants showed increased sensitivity to H2O2 and decreased virulence on peach fruits compared to the wild‐type isolate Bmpc7. It was found that H2O2 could stimulate the expression of MfOfd1 in the wild‐type isolate. The transformants were also more sensitive to exogenous osmotic stress, such as glycerol, d‐sorbitol, and NaCl, and to dicarboximide fungicides (iprodione and dimethachlon). These results indicate that the MfOfd1 gene plays an important role in M. fructicola in sporulation, oxidative response, osmotic stress tolerance, and virulence.

the decreased virulence in overexpression transformants might be caused by inducing ROS accumulation in the Prunus-M. fructicola interactions (Chou et al., 2015). In addition, it was shown that cutinase MfCUT1 and activating protein MfAP1 are potent virulence determinants of M. fructicola (Lee et al., 2010;Chiu et al., 2013;Yu et al., 2017). Furthermore, in the promoter of MfCUT1, several potential MfAP1-binding sites were observed, suggesting that the expression of MfCUT1 might be regulated by MfAP1 (Yu et al., 2017).
It has been reported that the failure of cells to respond to hypoxia may result in the death of cells and organisms (Gillies and Gatenby, 2007;Semenza, 2011). In Schizosaccharomyces pombe, the transcription factor Sre1 is a regulator of genes for adaptation to low oxygen conditions (Todd et al., 2006). Under the hypoxic condition, the prolyl 4-hydroxylase-like 2-OG-Fe(Ⅱ)-dependent dioxygenase (Ofd1) controls both DNA binding and degradation by regulation of the sterol regulatory element-binding protein (SREBP) Sre1. Sre1 is proteolytically cleaved under low oxygen conditions, and its N-terminal segment (Sre1N) serves as a hypoxic transcription factor, which enters the nucleus and up-regulates genes essential for growth under low oxygen conditions. When oxygen is sufficient, the Ofd1 uses multiple domains to down-regulate Sre1N activity by inhibiting Sre1N binding to DNA and accelerating Sre1N degradation. Ofd1 consists of two domains: an N-terminal 2-OG-Fe(II) dioxygenase domain and a C-terminal degradation domain (CTDD). The Ofd1 N-terminal dioxygenase domain is required for oxygen sensing and regulating the ability of Ofd1 CTDD to destabilize Sre1N; its C-terminal domain accelerates Sre1N degradation (Hughes and Espenshade, 2008;Lee et al., 2014;Gu et al., 2018). In the absence of oxygen, the N-terminal dioxygenase domain inhibits the Ofd1 CTDD, leading to accumulation of Sre1N. It has also been reported that Nro1 (SPCC4B3.07) directly inhibits Ofd1 as a positive regulator of Sre1N stability. Under hypoxic conditions, Nro1 binds to Ofd1 CTDD and inhibits Sre1N degradation. In present of oxygen, the binding of Nro1 to Ofd1 is disrupted and leads to rapid degradation of Sre1N (Porter et al., 2012;Yeh, 2012;Lee et al., 2014). In this study, a homologous gene of the redox-related Ofd1 was significantly up-regulated 1 hr after inoculation of M. fructicola (designated MfOfd1) on peach fruits, suggesting that it may play an important role in the pathogenicity of M. fructicola.
Oxygen (O 2 ) is an indispensable substance in the process of cell life. Under anoxic or hypoxic conditions, the failure of cells to respond to hypoxia may lead to a series of diseases, and even cause the death of cells or organisms (Semenza, 2010(Semenza, , 2011 (Mittler, 2017;Messens, 2018). ROS, as the by-products of aerobic metabolism, can result in oxidation or damage to DNA, RNA, proteins, and membrane components (Beckman and Ames, 1998). As a class of signalling molecules, ROS play important roles in plant-pathogen interactions. On perception of pathogen attack, plants often produce a large amount of ROS, which regulate the redox state, signal transduction of stress, systemic resistance, and cell necrosis (Mittler et al., 2004;O'Brien and Bolwell, 2012;Foyer and Noctor, 2013). In the ROS family, H 2 O 2 plays a role as the main signalling molecule for plants to fight against pathogen invasion and improve plant resistance through inducing gene expression, activation of related enzymes, and programmed cell death (Pei et al., 2000;Neill et al., 2002;O'Brien and Bolwell, 2012).
The infection of biotrophic pathogens is extremely inhibited by ROS-induced programmed cell death surrounding the infected host cells (Fath et al., 2002), while necrotrophic pathogens, such as Sclerotinia sclerotiorum and B. cinerea, mainly secrete some enzymes, toxins, ROS, and other substances to kill the host cells and then obtain nutrients from the dead cells (Asai and Yoshioka, 2009;Rietz et al., 2012). While the oxidative burst is important for their infection, necrotrophic pathogens can produce ROS by themselves or stimulate hosts to produce them (Kim et al., 2008;Alkan et al., 2009;Prusky et al., 2009). However, these pathogens need to activate the detoxification mechanism to respond to oxidative stress.
A few redox-related genes required for cellular responses to oxidative/redox conditions have been studied in fungi. However, it is unclear whether and how ROS accumulation is involved in virulence in M. fructicola during pathogenesis. Previous studies showed that the cutinase-encoding gene MfCUT1 is a virulence factor. MfCUT1 expression was up-regulated by H 2 O 2 and down-regulated by antioxidants in axenic culture (Lee et al., 2010;Chiu et al., 2013).
Meanwhile, an activating protein-like transcription factor MfAP1 was identified in M. fructicola, which had several binding sites at the DNA sequence upstream of MfCUT1 (Lee et al., 2010). When M. fructicola infected fruits or flowers, the expression of MfAP1 was activated and the genes responding to oxidative stress were up-regulated at the infection site (Yu et al., 2017). In some other plant-pathogenic fungi, MfAP1 homologues have also been identified, and all of them function in the redox stress response (Lev et al., 2005;Lin et al., 2009;Temme and Tudzynski, 2009;Guo et al., 2011;Walther and Wendland, 2012;Montibus et al., 2013). In Alternaria alternata and Magnaporthe oryzae, AaAP1 and MoAP1 also play roles in vegetative growth and pathogenicity Guo et al., 2011).
In this study, the function of the MfOfd1 gene was analysed through genetic transformation. It was found that knockdown of the MfOfd1 gene led to pleiotropic phenotypes, including insufficient sporulation, decreased virulence, and sensitivity to stress. This study sheds some light on the function of the MfOfd1 gene for virulence and ROS detoxification in M. fructicola, which could deepen our understanding of fungal Ofd1s.

| The MfOfd1 gene was up-regulated in the early infection stage of M. fructicola on peach fruits
M. fructicola isolate Bmpc7 was inoculated onto peach fruits, the different infection stage transcripts were examined by RNA-Seq, and the transcription data were deposited in GenBank (accessions SAMN12871599 to SAMN12871619). Compared with the expression level at 0 hr post-inoculation (hpi), a total of 188 differentially expressed genes were detected at 1 hpi, of which 100 genes were up-regulated and 88 genes were down-regulated (partial data are shown in Figure 1a). This early stage (1 hpi) was an important period for the peach-M. fructicola interaction; previous studies showed that some pathogenicity-related genes (MfCUT1 and MfPGs in M. fructicola, SsSSVP1 in S. sclerotiorium) were up-regulated at early stages of infection (Lee et al., 2010;Chou et al., 2015;Lyu et al., 2016;Yu et al., 2017). Based on the transcriptomic analysis and quantitative reverse transcription PCR (RT-qPCR) analysis, the MfOfd1 gene (MN515052) was up-regulated during different infection stages, especially at 1 hpi ( Figure 1b).

| Generation and characterization of MfOfd1 knockdown transformants
To investigate the biological function of MfOfd1 in M. fructicola, we tried to knock out the MfOfd1 gene by homologous recombination and the CRISPR/Cas9 system, via polyethylene glycol (PEG)-mediated protoplast transformation ( Figures S2 and S3). The transformants were F I G U R E 1 The expression level of MfOfd1 in Monilinia fructicola during the early infection stage on peach fruits. (a) The heat map of partial differentially expressed genes at 1 hr post-inoculation (hpi) is shown as an example. Red, blue, and black indicate the significantly upregulated, down-regulated, and no difference genes, respectively. The selection criteria are |log 2 (fold change)| > 1 and p adj < .05. (b) Relative expression of MfOfd1 was confirmed by quantitative reverse transcription PCR at 0, 1, 3, 6, 12, and 24 hpi, and spore germination (sg). Significance was determined by t test (***p < .001, **p < .01) confirmed by PCR verification and RT-qPCR ( Figures S2 and S4). It should be noted that the primer pair MfOfd1-C/Z-For and MfOfd1-C/Z-Rev amplified two fragments of 876 and 1,829 bp from all of the obtained transformants, while the primers only amplified a single fragment of 876 bp from the wild-type isolate, indicating that the transformants were heterokaryons ( Figure S4). Therefore, the transformants were called "knockdown" transformants instead of "knockout" transformants. We obtained nine positive knockdown transformants: ΔMfOfd1-10,  Nevertheless, the MfOfd1 expression level in these knockdown transformants were reduced by 82.1%-98.2%, significantly lower than that in the wild-type isolate ( Figure 3).

| Knockdown of MfOfd1 influenced the sporulation of M. fructicola
In general, the phenotypes of transformants were similar. All of the transformants did not show differences in colony morphology, mycelial growth on potato dextrose agar (PDA), and minimal medium agar (MMA) ( Table 1). As for the sporulation, transformants produced 0.5-1.32 × 10 5 conidia/cm 2 , which was significantly lower than that of the wild-type (2.83 × 10 5 conidia/cm 2 ) on V8 agar ( Figure 4). To further elucidate how the defect of sporulation occurred in the aforementioned transformants, the conidiophores, conidial size, and conidial germination in transformants were compared with those of the parental wild-type isolate Bmpc7. The results showed that the conidiophore structures and conidial germination of the transformants were normal and similar to that of the wild-type isolate, while the conidial width of the transformants was slightly narrower than that observed in the wild-type isolate (Table 1).

| Knockdown of MfOfd1 influenced the virulence of M. fructicola on peach fruits
In order to know whether MfOfd1 affects the virulence of M. fructicola, a virulence assay was performed on detached peach fruits. The results showed that the wild-type isolate produced large and oval lesions on which there were dense grey hyphae and lots of spores, leading to typical peach brown rot disease symptoms, but the transformants produced smaller lesions with sparse hyphae, or even no visible hyphae on some lesions (Figure 5a). The lesions produced by the wild-type isolate   Figure 6c). This result also indicates that MfOfd1 is involved in the regulation of the oxidative stress response.

| MfOfd1 is important for stress tolerance
In S. pombe, Ofd1 is involved in regulating the Sre1N and sterol regulatory elements (SRE); sterol is widely present in biological cells and tissues, with different biological functions (Hughes and Espenshade, 2008;Yeh, 2012;Lee et al., 2014). showed that knockdown transformants were more sensitive to glycerol, sorbitol, and NaCl compared to the wild-type isolate ( Figure 7 and fluffy and the mycelia were off-white, while the colonies of transformants were dense and dark brown, and produced large amounts of conidia, which showed concentric sporodochia (Figure 7a). These results indicate that the deletion of MfOfd1 did not affect the integrity of the M. fructicola cell wall, but may have an impact on the membrane protein.

| MfOfd1 influenced the sensitivities of M. fructicola to fungicides
Application of fungicides is the most effective way to prevent and cure peach brown rot. In order to evaluate whether the de-  (Figure 8 and Table 3).
Because the target of DCFs is the osmotic stress signal transduction pathway (Motoyama et al., 2005), this result further supported the fact that MfOfd1 plays an important role in the osmotic stress signal transduction pathway in M. fructicola. Under treatment with DMIs and SDHIs, the sensitivity did not show a significant difference between transformants and the parental wild-type isolate (Table 3).

| D ISCUSS I ON
The CRISPR/Cas9 system was originally discovered in bacteria and archaea as a defence system against phage and plasmids Note. Mean ± SD; values within the same column followed by the same letters are not significantly different based on one-way analysis of variance with the LSD test in SPSS 21.0 software at p = .05. (Barrangou et al., 2007). The CRISPR/Cas9 system has the advan-  (Chen et al., 2016). In this study, we combined homologous recombination with CRISPR/Cas9 to knock down the MfOfd1 gene and obtained more transformants than previously using PEG-mediated Note. Mean ± SD; values within the same column followed by the same letters are not significantly different based on one-way analysis of variance with the least-significant difference (LSD) test in SPSS 21.0 software at p = .05. different phenotypes. As the knockdown transformants were heterozygotes containing both knockout and wild-type genotypes, it is not necessary to complement the MfOfd1. However, further knockdown transformants may be needed to evaluate different phenotypes and obtain more reliable results.
Even though MfOfd1 is not essential for mycelial growth in M. fructicola, it is important for sporulation. Conidiation is a key factor in the epidemics of peach brown rot (Ritchie, 2005), but reports about conidiation-related genes are limited in M. fructicola. In this study, we found that the conidiation of knockdown transformants was reduced remarkably in comparison to the wild-type isolate. It is possible that the knockdown of MfOfd1 affects the expression of conidiation-related genes, resulting in the decline of conidiation. These results indicate that MfOfd1 is crucial for conidiation in M. fructicola.
As far as we know, this is the first time a gene associated with conidiation in M. fructicola has been found. were more sensitive to glycerol, d-sorbitol, and NaCl than the wild-type isolate. These results suggest that the MfOfd1 gene might be involved in membrane integrity.
ROS are important signal transduction and defence substances produced by a plant active resistance reaction (Baxter et al., 2013;Mühlenbock et al., 2007). Fungi need to activate their ROS detoxification and improve ROS tolerance for successful infection. Under oxidative stress, the inhibition of mycelial growth of the knockdown transformants was remarkably higher than that of the wildtype isolate Bmpc7, suggesting that the MfOfd1 gene is involved in the mycelial response to oxidative stress. Furthermore, the expres-

sion of MfOfd1 was induced by treatment with H 2 O 2 , indicating that
MfOfd1 is indispensable for the fungal response to oxidative stress.
The mechanisms of how MfOfd1 regulates the response to oxidative stress should be investigated in the future.
Fungicide sensitivity assays showed that knockdown of MfOfd1 did not affect the sensitivity of M. fructicola to DMI (propiconazole and tebuconazole) and SDHI (boscalid) fungicides. However, all the transformants were more sensitive to DCF (iprodione and dimethachlon) fungicides. DCFs are a class of fungicides that act on histidine and mitogen-activated protein (MAP) kinase in the osmotic signal transduction pathway (Motoyama et al., 2005;Yoshimi et al., 2005;Luo et al., 2012). The increased sensitivity of MfOfd1 knockdown transformants to DCFs further proved that the MfOfd1 gene plays an important role in the osmotic signal transduction pathway in M. fructicola. In our previous study it was found that the sensitivity to DCFs decreased in MfSre1 knockdown transformants, indicating that MfSre1 negatively regulates sensitivity to DCFs and osmotic stress (Jiang et al., 2019). Therefore, it could be possible that both MfOfd1 and MfSre1 are involved in the response to osmotic stress in M. fructicola, but one is a positive regulator and the other is a negative regulator.
In conclusion, knockdown of the MfOfd1 gene did not affect vegetative growth, but resulted in a decline in conidiation and af-

| Fungal isolate and growth conditions
The wild-type single spore isolate Bmpc7 of M. fructicola was collected from a peach orchard in the United States and stored on filter paper at −20 °C (Luo et al., 2008). Isolate was cultivated on PDA at 22 °C for 3 days in the dark. The fungal isolate was grown on 20% vegetable-juice agar medium (V8, 200 ml V8 juice and 20 g agar per litre) for 2 weeks at 22 °C in the dark for sporulation (Lee et al., 2010). For DNA and RNA extraction, 8-10 agar plugs containing mycelium were transferred to 40 ml potato dextrose broth and incubated at 22 °C on a 150 rpm orbital shaker for 36 hr in the dark.
The genomic DNA was extracted using the EASYspin Plant Genomic DNA Extraction Kit (Aidlab Biotechnologies Co.).

| RNA-Seq, read quality, and data analysis
High-throughput RNA-Seq sequencing was used to detect the gene expression of 21 samples at seven stages (each stage with three technique replications) of 0, 1, 3, 6, 12, and 24 hpi, and spore germination (sg) stage during the infection of M. fructicola on peach fruits. In order to ensure the quality and reliability of data analysis, it is necessary to remove reads with adapters containing N (base information cannot be determined) and low quality (base number of Qphred ≤ 20 accounts for more than 50% of the entire read length) bases. In addition, the original data were checked for sequencing error rates and guanine-cytosine content distribution to obtain clean reads for subsequent analysis. Hierarchical Indexing for Spliced Alignment of Transcripts (HISAT) software (http:// ccb.jhu.edu/softw are/hisat /index.shtml) was used to do genomic localization analysis of the filtered reads (Daehwan et al., 2015).
According to the comparison results, the corresponding reads of each transcript were counted and standardized using the fragments per kilobase million (FPKM) method. Gene differential expression was analysed by DESeq2 software (http://www.bioco nduct or.org/ packa ges/relea se/bioc/html/DESeq2.html). First, the read count was normalized (Anders and Huber, 2010), then the p value calculation model (negative binomial distribution) was used to calculate the probability of hypothesis testing and finally multiple hypothesis testing was corrected (calculation method:BH) to obtain the false discovery rate value. The differential gene screening standard was p adj < .05.

| Vector construction and fungal transformation
Homologous recombination was combined with CRISPR/Cas9 to obtain the knockdown transformations ( Figure S3). The upstream (982 bp) and downstream (595 bp) fragments of the MfOfd1 gene in Bmpc7 and the fragment of hygromycin B resistance phosphotransferase gene (HPH, 1,414 bp) cassette in pSKH vector were amplified (Yun, 1998;Jiang et al., 2019). Knockdown constructs were produced by double-jointed PCR using three amplicons (upstream fragment, downstream fragment, and HPH cassette). The specific identification site of Esp3Ⅰ FastDigest (Thermo Scientific) was used to digest the pmCas9 empty vector at 37 °C for 15 min. The 20 bp before NGG (5′-3′) in the CDS region of MfOfd1 was selected as the specific sequence of single-guide RNA (sgRNA), and its specificity was confirmed through the local BLAST in the Bmpc7 genome. The sequence was synthesized in the form of forward primer and reverse primers (Table S1) and sticky ends (5′-ACCT-3′, 5′-AAAC-3′) were added at the 5′ ends (Liang et al., 2018), then inserted into the digested pmCas9 vector by T4 DNA ligase (TaKaRa). The inserts in plasmids were sequenced to confirm their correctness.
To perform PEG-mediated protoplast transformation, M. fructicola protoplasts were first prepared by digesting fresh mycelia with cell wall lyase (Lysing Enzymes from Trichoderma harzianum, Sigma) at 30 °C and 150 rpm for 4 hr (Jiang et al., 2019). The digested suspension was centrifuged at 4 °C and 1,500 × g for 10 min, the supernatant was removed, and the precipitation was resuspended with STC (1.2 M sorbitol; 10 mM Tris-HCl, pH 7.5; 50 mM CaCl 2 ) solution.

| DNA extraction and validation of knockdown transformants
The wild-type isolate Bmpc7 and the knockdown transformants were cultured in PDB for 36 hr, and DNA was extracted using the   (Ma et al., 2003;Schmittgen and Livak, 2008). The primer pairs for RT-qPCR are given in Table S1. Expression of the The experiments were performed with three independent biological repeats. The expression of the MfOfd1 gene was normalized to the expression of the β-tubulin gene, and relative gene expression was calculated with the comparative C t (2 −ΔΔCt ) method (Wong and Medrano, 2005).

| Determination of mycelial growth, sporulation, and spore germination
All the wild-type isolate and knockdown transformants were inoculated on PDA and MMA at 22 °C for 5 days to investigate the colony morphology and mycelial growth as described previously . Spores were collected from colonies on 20% V8 agar for 2 weeks, and the number of conidia was counted under a microscope with a haemocytometer. To measure the spore size and evaluate the spore germination, spore suspension (100 μl, 3 × 10 4 /ml) was uniformly spread on water agar medium, 100 spores were randomly selected under the microscope, and the size of spores was measured perpendicularly. Spores germination was evaluated after incubation for 6 hr. These experiments were performed in three independent biological replications.

| Virulence assay
The virulence assay was performed using the susceptible peach cultivar Prunus persica 'Fei Cheng'. Holes (5 mm deep) were generated on the surface of fruits using a cork borer (5 mm diameter) and the holes were inoculated with mycelial plugs. All of the inoculated fruits were put into plastic boxes that were covered with cling film to maintain high humidity and incubated at 22 °C. Brown rot lesion size was measured at 24, 48, and 72 hr post-inoculation. Three fruits were used for each strain in a treatment and the experiment was conducted three times.

| Sensitivity assay to stress
To assess the integrity of the cell wall and cell membrane, PDA was amended with 600 μg/ml Congo Red or 0.01% SDS. For osmotic stress, PDA was amended with 150 g/L glycerol,

| Sensitivity to fungicides
The sensitivity to fungicides was investigated in the wild-type isolate and transformants. For DCFs, sensitivity to iprodione and dimethachlon was assessed on fungicide-amended PDA at 0.2 and 0.5 μg/ml, respectively. For DMIs, propiconazole and tebuconazole were added to PDA at 0.2 and 0.2 μg/ml, respectively (Yuan et al., 2013). For SDHIs, boscalid was added to MMA at 0.8 μg/ml (Chen et al., 2014).
The wild-type isolate and transformants were cultured in triplicate at 22 °C for 3 days. Colony diameters were measured and expressed as the percentage of growth inhibition.

| Statistics
Multiple comparison was performed for the fitness data and statistical differences were evaluated by one-way analysis of variance (ANOVA) with the least-significant difference (LSD) test in SPSS v.

ACK N OWLED G EM ENTS
This work is supported by the Project of National Natural Science Foundation of China (no. 31872934) and the earmarked fund for Modern Agro-Industry Technology Research System (no. CARS-30).

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Transcriptomic data can be found in GenBank (https://www.ncbi. nlm.nih.gov/genba nk/) with accession numbers SAMN12871599 to SAMN12871619. The sequence of MfOfd1 can be found in GenBank with accession no. MN515052.

R E FE R E N C E S
Alkan, N., Davydov, O., Sagi, M., Fluhr, R. and Prusky, D. (2009) Ammonium secretion by Colletotrichum coccodes activates host NADPH oxidase activity enhancing host cell death and fungal