A rice gene encoding glycosyl hydrolase plays contrasting roles in immunity depending on the type of pathogens

Abstract Because pathogens use diverse infection strategies, plants cannot use one‐size‐fits‐all defence and modulate defence responses based on the nature of pathogens and pathogenicity mechanism. Here, we report that a rice glycoside hydrolase (GH) plays contrasting roles in defence depending on whether a pathogen is hemibiotrophic or necrotrophic. The Arabidopsis thaliana MORE1 ( Magnaporthe oryzae resistance 1) gene, encoding a member of the GH10 family, is needed for resistance against M. oryzae and Alternaria brassicicola, a fungal pathogen infecting A. thaliana as a necrotroph. Among 13 rice genes homologous to MORE1, 11 genes were induced during the biotrophic or necrotrophic stage of infection by M. oryzae. CRISPR/Cas9‐assisted disruption of one of them (OsMORE1a) enhanced resistance against hemibiotrophic pathogens M. oryzae and Xanthomonas oryzae pv. oryzae but increased susceptibility to Cochliobolus miyabeanus, a necrotrophic fungus, suggesting that OsMORE1a acts as a double‐edged sword depending on the mode of infection (hemibiotrophic vs. necrotrophic). We characterized molecular and cellular changes caused by the loss of MORE1 and OsMORE1a to understand how these genes participate in modulating defence responses. Although the underlying mechanism of action remains unknown, both genes appear to affect the expression of many defence‐related genes. Expression patterns of the GH10 family genes in A. thaliana and rice suggest that other members also participate in pathogen defence.

Three M. oryzae genes crucial for penetrating rice epidermal cells and initial proliferation play only limited roles in infecting A. thaliana (Park et al., 2009), suggesting that unlike in rice, M. oryzae does not need to go through the biotrophic stage for completing its disease cycle. In this study, a screening of A. thaliana mutants revealed that a gene encoding a member of the GH10 family is required for resistance against M. oryzae and Alternaria brassicicola, a necrotrophic pathogen. One of its homologues in rice was also required for resistance against a necrotrophic pathogen, Cochliobolus miyabeanus, but conferred susceptibility to M. oryzae and a bacterial hemibiotrophic pathogen, Xanthomonas oryzae pv. oryzae (Xoo). We investigated how these GH10 genes participate in pathogen defence using multiple approaches.

| Disruption of the A. thaliana MORE1 gene increased susceptibility to M. oryzae and A. brassicicola
A. thaliana ecotype Wassilewskija (Ws-0) is resistant to M. oryzae strain 70-15 (Park et al., 2009). We screened a pool of T-DNA insertional mutants (3300 M 1 plants) of Ws-0 to identify genes crucial for resistance against M. oryzae. We identified 21 putative mutants displaying enhanced susceptibility to strain 70-15. These mutants were named more1-more21 (M. oryzae resistance 1-21). Four mutants (more1-more 4) contained a single T-DNA inserted in their genome.
Using TAIL-PCR, we revealed that the T-DNA in more4 was located inside a gene predicted to encode a DNA polymerase A family on chromosome 1. In more2 and more3, the disrupted gene encodes a hypothetical protein on chromosomes 1 and 5, respectively. The more1 mutant had a single copy of T-DNA inserted in the second exon of At4g33820, a gene annotated to encode a member of the GH10 family ( Figure S1a). We focused on characterizing this gene (named MORE1).
Five GH10 family genes flank the MORE1 gene. The more1 mutant failed to produce MORE1 transcripts ( Figure S1b) and developed narrower and smaller rosettes than Ws-0 ( Figure S1c). We compared expression patterns of 12 GH10 genes between more1 and Ws-0 using reverse transcription quantitative PCR (RT-qPCR). The loss of MORE1 altered the expression of all of the GH10 genes except Ws-0 (Park et al., 2009). Compared to Ws-0, which began producing yellow spots or small chlorotic lesions at 3 days postinoculation (dpi) with KJ201, more1 displayed necrosis at the centre of severely chlorotic areas at 3 dpi ( Figure 1c). These necrotic spots expanded and covered the entire leaf at 6 dpi ( Figure 1c). To verify that the disruption of MORE1 caused the increased susceptibility to M. oryzae and CF, we complemented the mutation. A transgene including the entire coding region of MORE1 with its native promoter fully rescued the impaired resistance (Figure 1d), validating the importance of MORE1 in conferring resistance against M. oryzae.
The more1 mutant developed larger lesions than Ws-0 on inoculation with A. brassicicola, a necrotrophic fungal pathogen. Lesions in more1 produced more conidia than those in Ws-0 ( Figure 2), suggesting the requirement of MORE1 in conferring resistance against necrotrophic pathogens.

| Comparative transcriptome analysis of
Ws-0 and more1 showed that expression of many defence-related genes was affected by the loss of MORE1 To investigate how MORE1 participates in immunity, we compared gene expression patterns in 3-week-old Ws-0 and more1 using RNA-Seq. Compared to Ws-0, levels of 667 and 424 genes in more1 increased and decreased, respectively (fold change ≥2, Gfold value ≠0; Data files S1 and S2). Six defence-related differentially expressed genes (DEGs, three up-regulated and three down-regulated) were analysed using reverse transcription quantitative PCR (RT-qPCR) to check the reliability of identifying DEGs via RNA-Seq (Figure 3a), which showed comparable patterns.
The DEGs are associated with more than 90 Gene Ontology (GO) terms (see Figure 3b for the top 13 enriched GO terms). Those upregulated are enriched with the GO terms associated with biological processes related to biotic stress response, with the top 10 being response to chitin, response to salicylic acid, defence response to bacterium, defence response, response to bacterium, defence response to fungus, response to fungus, incompatible interaction, response F I G U R E 1 Increased susceptibility of more1 to Magnaporthe oryzae. (a) Conidiation of M. oryzae strain 70-15 in infected more1. Arrows indicate appressoria (left panel) and sporulation (right panel). The images represent different inoculated leaf samples observed in three independent experiments. Scale bars = 50 μm. (b) Leaves of Ws-0 and more1 inoculated with 70-15 culture filtrate at 3 days postinoculation (dpi) (left) and stained using Evans blue to detect dead cells (right). (c) Ws-0 (left panel) and more1 (right panel) infected with M. oryzae strain KJ201 at 3 dpi (middle) and 6 dpi (bottom). Mock-treated plants at 6 dpi are shown at the top. (d) A conidial suspension of 70-15 (10 µl) was placed on 28-to 30-day-old leaves of Ws-0, more1, and C2-5, a complemented line. Disease symptoms at 6 dpi are shown to wounding, and plant-type hypersensitive response (HR). The top three enriched GO terms associated with the down-regulated genes were cytokinin-activated signalling pathway, killing of cells of other organism, and phosphorelay signal transduction system ( Figure 3b).

Analysis of these DEGs via the Kyoto Encyclopedia of Genes and
Genomes (KEGG) showed that the up-regulated genes were associated with >50 GO terms/KEGG pathways. The top 10 pathways were similar to what the GO term analysis revealed. Additional GO terms/KEGG pathways related to biotic stress response included systemic acquired resistance, ethylene-activated signalling pathway, and response to jasmonic acid ( Figure 3c). The GO terms/KEGG pathways associated with the down-regulated genes were only five, including plant hormone signal transduction, cytokinin-activated signalling pathway, killing of cells of other organisms, phosphorelay signal transduction system, and microtubule ( Figure 3d).

| Expression of most MORE1 homologues in rice was induced by M. oryzae infection
Molecular and phenotypic changes caused by the disruption of MORE1 in A. thaliana led to the hypothesis that MORE1 homologues in other plants also participate in pathogen defence. We identified 13 rice homologues (OsMORE1a-OsMORE1m). Protein sequence alignment showed that OsMORE1a, OsMORE1b, OsMORE1c, and OsMORE1d were most closely related to MORE1, with the identity being 54.7%, 54.1%, 50.7%, and 50.5%, respectively ( Figure S3). A phylogenetic analysis using the conserved signature GH10 domain confirmed the close evolutionary relationship between MORE1 and the four OsMORE1s ( Figure S4). Transcript analysis of rice infected with M. oryzae showed that expression of 11 genes was induced to varying degrees. Expression of OsMORE1a and OsMORE1b was induced during the biotrophic stage, whereas OsMORE1c, OsMORE1e-OsMORE1k, and OsMORE1m were highly expressed during the necrotrophic stage. The OsMORE1d and OsMORE1l genes did not show significant changes in expression ( Figure 4).

| Loss of OsMORE1a enhanced resistance against M. oryzae
We attempted to disrupt OsMORE1a and OsMORE1b, two genes induced during the biotrophic stage, in cv. Dongjin via CRISPR/ Cas9-assisted genome editing to study their role in defence.

Potential sites of mutagenesis in the coding regions of OsMORE1a
and OsMORE1b were evaluated using the CRISPR RGEN tools One of the initial defence responses in rice against M. oryzae is producing ROS such as superoxide and hydrogen peroxide (Camejo et al., 2016;Jwa & Hwang, 2017). ROS accumulation in rice sheaths was compared using CM-H 2 DCFDA, an ROS-sensitive dye that has been used to monitor ROS localization in plant cells F I G U R E 2 Increased susceptibility of more1 to Alternaria brassicicola. (a) Disease symptoms of Ws-0 and more1 infected with A. brassicicola at 5 days postinoculation (dpi). (b) The average diameter of lesions at 5 dpi and the average number of spores produced per lesion at 8 dpi are shown. Results represent means (± SD) of three biological replicates with 10 leaves for each treatment. Asterisks indicate significant differences according to Student's t test. *p < 0.05, **p < 0.01 (Fryer et al., 2002;Kristiansen et al., 2009). ROS (H 2 O 2 ) accumulated around IH in osmore1a at 36 hpi, but no ROS accumulation was detected in infected sheath cells of Dongjin ( Figure 6d).
Compared to Dongjin, more transcripts from two NADPH oxidase (ROS producer) genes were present in osmore1a, but the transcript levels of most superoxidase dismutase (ROS scavenger) genes were lower (Figure 6e).

| Disruption of OsMORE1a caused opposite effects on defence depending on whether a pathogen is hemibiotrophic or necrotrophic and changed expression patterns of many defence-related genes
The osmore1a mutant was more resistant to Xoo, a hemibiotrophic bacterial pathogen (Figure 7a). However, like the A. thaliana more1 F I G U R E 3 Comparative transcriptome analysis between Ws-0 and more1. (a) Expression levels of six defence-related genes (three up-regulated and three down-regulated) in Ws-0 and more1 were measured using reverse transcription quantitative PCR to validate RNA-Seq results. The ubiquitin5 gene was used as the control for this analysis (repeated twice). Error bars indicate the SD. The data represent the mean ± SD of three biological replications. The asterisks denote statistically significant (p < 0.01) differences (Student's t test). (b) Highly enriched GO terms associated with differentially expressed genes (DEGs) are shown. Blue and red bars denote those up-regulated and down-regulated, respectively. The y axis denotes enriched GO terms and the x axis shows the number of DEGs associated with each GO term. The GO term/KEGG pathway enrichment statistics of the (c) upregulated and (d) down-regulated genes in more1 compared to Ws-0 are presented. The y axis denotes enriched GO terms and the x axis shows the ratio between the proportion of genes annotated to each pathway among the DEGs and the proportion of genes annotated to that pathway in all genes (Fisher's exact test, p < 0.05). Each circle represents the number of DEGs mapped to specific pathways/GO terms mutant, the mutant was more susceptible to C. miyabeanus, a necrotrophic fungal pathogen of rice brown spot ( Figure 7b). The effect of losing OsMORE1a on resistance against these pathogens was consistent with the expression patterns of six defence-related genes in osmore1a and Dongjin ( Figure 7c). Expression levels of one of the two genes controlled by the salicylic acid (SA) signalling pathway and two PR genes were higher in osmore1a than Dongjin. In contrast, expression levels of two genes under the control of the jasmonic acid (JA) signalling pathway were lower in osmore1a than Dongjin.
To investigate further how the loss of OsMORE1a caused opposite effects on resistance against M. oryzae/Xoo versus C. miyabeanus, we conducted an RNA-Seq analysis of leaf transcriptomes using 3-week-old Dongjin and osmore1a plants. In total, 165 million raw reads were generated from each biological replicate, with over 96% of them aligning with the cv. Nipponbare genome sequence (Table 1). Compared to Dongjin, 1165 genes (1022 upregulated and 143 down-regulated) were differentially expressed in osmore1a ( Figure S7, and Data files S3 and S4). A GO enrichment analysis revealed 36 categories, including cell wall macromolecule catabolic/metabolic process, aminoglycan catabolic/metabolic process, response to biotic stimulus, diterpenoid metabolic process, and chitin catabolic/metabolic process, were enriched among the up-regulated genes (Table 2 and Figure S8). In contrast, significantly enriched GO terms were not found among the downregulated genes.

F I G U R E 4 Expression patterns of 13 OsMORE1 genes during
Magnaporthe oryzae infection. Their expression patterns in rice infected with strain KJ201 were analysed using reverse transcription quantitative PCR. Relative gene expression indicates the expression level of each gene relative to that in mockinoculated plants, which was normalized using the OsACTIN gene. Expression patterns of (a) OsMORE1a, OsMORE1b, and OsMORE1d and (b) OsMORE1a-OsMORE1m are presented. The y axis represents the relative expression level, calculated using 2 −ΔΔCt , and the x axis denotes the hours postinoculation (hpi) and three infection stages. Three biological replicates were included for these analyses. Error bars indicate standard deviation (SD) (e) Expression patterns of two NADPH oxidase genes and six superoxide dismutase (SOD) genes in Dongjin and osmore1a. Relative gene expression denotes the expression level of each gene in osmore1a relative to Dongjin, which was normalized using the OsACTIN gene. The y axis shows fold changes. The data represent the mean ± SD of three biological replications. Asterisks indicate significant differences between Dongjin and osmore1a (Student's t test, *p < 0.05, **p < 0.01, ***p < 0.001) in osmore1a relative to that in Dongjin, which was normalized using the OsACTIN gene. The y axis shows fold changes. The data represent the mean ± SD of three biological replicates. Asterisks indicate significant differences between Dongjin and osmore1a (Student's t test, **p < 0.01, ***p < 0.001)  diterpene phytoalexins (Toyomasu et al., 2018) were also induced in osmore1a ( Figure S10e). Overall, multiple genes associated with PTI and the cell wall were induced in osmore1a.

| D ISCUSS I ON
Plant cells are encased in rigid walls composed of cellulose, hemicellulose, pectin, proteins, and lignin, and the amount of these components varies depending on cell type. The cell wall regulates plant growth and serves as the first line of defence (Malinovsky et al., 2014;Underwood, 2012;Zhong & Ye, 2015). Therefore, researching how plants make and modify cell walls is vital for understanding plant biology and applying the resulting understanding to improve crop production. We discovered that a member of the GH10 family participates in cell wall-mediated pathogen defence in both A. thaliana (Figures 1 and 2) and rice (Figures 6 and 7). Our data offer several clues to how the MORE1 and OsMORE1 genes engage in pathogen defence (Figures 1, 2, 6, and 7) and why the loss of OsMORE1 oppositely affects defence depending on whether a pathogen is hemibiotrophic or necrotrophic (Figures 6   and 7). This situation resembles the role of the barley Mlo locus in defence against biotrophic and hemibiotrophic fungi (Jarosch et al., 1999). Although available data are insufficient for understanding the mechanism of their action in modulating defence responses, they offer a few clues. The outcome of ROS production during plant-microbe interactions can vary depending on the amount of ROS produced (Kotchoni & Gachomo, 2006). A high dosage of ROS leads to HR and induces cell death (Gechev & Hille, 2005;Petrov & Van Breusegem, 2012;Petrov et al., 2015). In contrast, moderate and controlled levels of ROS seem to regulate defence responses by triggering the expression of some defence-related genes, increasing the production of antimicrobial compounds, and fortifying the cell wall (Kotchoni & Gachomo, 2006;Lamb & Dixon, 1997).
Several genes involved in cell wall modification were up-regulated in osmore1a ( Figure S10a), suggesting that stronger defence of the mutant against hemibiotrophs could be attributed to increased ROS production and the resulting cell wall modification. In addition to cell wall modification, ROS are considered as signalling molecules for cell death. Cell death increases resistance against biotrophic pathogens but helps necrotrophs, such as Botrytis cinerea and Sclerotinia sclerotiorum, proliferate (Govrin & Levine, 2000). We hypothesize that the high dosage of H 2 O 2 locally accumulated in infected cells of the osmore1a mutant leads to rapid HR-induced cell death, blocking M. oryzae from invading neighbouring cells ( Figure 6) but facilitating infection by C. miyabeanus (Figure 7b).
Plant PRR proteins recognize pathogen-associated molecular patterns (PAMPs, derived from the pathogen) and damageassociated molecular patterns (DAMPs, derived from the host) during pathogen invasion, causing induced PTI or basal defence (Saijo et al., 2018). Activation of PTI leads to the activation of specific hormone-regulated signalling pathways (Verhage et al., 2010).
Changes in defence-related gene expression caused by the loss of MORE1 are consistent with the well-established role of the SA and JA signalling pathways in regulating defence against hemibiotrophic/ biotrophic and necrotrophic pathogens, respectively (Aerts et al., 2020;Berens et al., 2017;Ghozlan et al., 2020;Glazebrook, 2005).
In more1 ( Figure 3a) and osmore1a ( Figure 7c and Table S1), transcript levels of the genes controlled by SA were induced, while those under the control of JA were suppressed.
We performed RNA-Seq to investigate the molecular basis of the contrasting roles of OsMORE1a in immunity depending on the type of pathogens ( Figure S7, and Data files S3 and S4). However, we should note that more studies are needed for two reasons. One is because we did not compare gene expression patterns after infecting Dongjin and osmore1a with different pathogens. It was shown that responses to biotic and abiotic stresses are differentially reg-  Our RT-qPCR analysis showed that 11 members of the GH10 family in rice were differentially expressed during M. oryzae infection ( Figure 4). All of them, except OsMORE1a and OsMORE1b, were induced during the necrotrophic stage. OsMORE1a and OsMORE1b were induced during the biotrophic stage. Hemibiotrophic fungal pathogens proliferate without killing host cells by suppressing the HR (Miya et al., 2007;Monaghan & Zipfel, 2012). Rice GHs in other families also probably perform immunity-related functions considering their expression patterns during pathogen infection (Kawahara et al., 2012;Sharma et al., 2013).
Our results helped advance our understanding of how pathogen defence operates in two model plants and offer a crucial consideration in developing new disease resistance via genetic engineering.

Deployment of resistance (R) genes involved in ETI via breeding has
been widely practised as a means for protecting crop health without heavily relying on pesticides (Dangl et al., 2013;Flor, 1971).
However, this practice often encounters resistance breakdown due to genetic changes in pathogen populations that allow pathogens to evade R-mediated detection (Vleeshouwers et al., 2011;Win et al., 2012). Recent advances in genome editing technologies, particularly CRISPR/Cas9-based tools, not only facilitate efforts to dissect the mechanism of defence against diverse pathogens and pests but also expedite targeted modifications of specific genes to enhance resistance against a wide range of pathogens without triggering the regulatory processes associated with releasing genetically modified crops (Kanchiswamy et al., 2015;Waltz, 2016Waltz, , 2018. Modifying the genes for host susceptibility (S) factors, those taken advantage of by pathogens to facilitate their proliferation, via genome editing has been proposed as an alternative and complementary strategy (Chandrasekaran et al., 2016;Langner et al., 2018;Nekrasov et al., 2017;Pyott et al., 2016). Disruption of OsMORE1a enhanced resistance against M. oryzae and Xoo (Figures 6 and 7), suggesting that OsMORE1a functions as an S gene against hemibiotrophic pathogens. However, because S genes often participate in multiple pathways, their inactivation may perturb interactions with beneficial microbes or increase susceptibility to other types of pathogens (Babaeizad et al., 2008;Jarosch et al., 1999;Kim & Hwang, 2012;Lumbreras et al., 2010). To circumvent such trade-offs, it is important to understand the role and mechanism of action of S genes in immunity. It might be possible to modify S genes in ways that prevent such undesirable effects while maintaining enhanced disease resistance. Investigations into how GH-mediated disease resistance/susceptibility in A. thaliana and rice operates will probably help assess whether judicious manipulations of specific GHs can be deployed to enhance defence against a broad spectrum of pathogens without negatively impacting growth and fitness. One such manipulation would be adjusting their expression through genome editing of cis-regulatory elements (Rodríguez-Leal et al., 2017). The "silencing on demand" approach using pathogen-inducible promoters could be an alternative method. In barley, the pathogen-inducible Hv-Ger4c promoter has been successfully used to control the expression of Ta-Lr34res, encoding an ABC transporter that confers resistance against multiple broad-spectrum fungal pathogens in wheat (Boni et al., 2018).

| Plant materials and growth conditions
A. thaliana ecotype Ws-0 and the following T-DNA insertional mu-

| Growth conditions for fungal pathogens
All M. oryzae, A. brassicicola, and C. miyabeanus strains were obtained from the Center for Fungal Genetic Resource at Seoul National University, Seoul, Korea. Conidia of M. oryzae strain 70-15 (Chao & Ellingboe, 1991;Leung et al., 1988)

| Infection of A. thaliana
Five plants were sprayed with 20 ml of M. oryzae conidial suspension (5 × 10 5 conidia/ml) using an airbrush. After placing the inoculated and mock-inoculated plants in a dew chamber for 16 h at 25°C under 100% humidity, they were transferred to a growth chamber (22°C, 80% humidity). Each infection was repeated three times. Detached leaves from 4-week-old plants were inoculated by dropping on 10 μl of A. brassicicola spore suspension (5 × 10 5 spores/ml). Inoculated leaves were kept in a covered plastic container to maintain high humidity.
The number of A. brassicicola spores formed on inoculated plants was determined as previously described (Van Wees et al., 2003).

| Preparation of M. oryzae culture filtrate
After inoculating conidia of M. oryzae into 300 ml of potato dextrose broth (Difco) in a 500-ml conical flask, the flask was shaken (125 rpm) for 7 days at 25°C without light. The culture was filtered first through sterilized Whatman no. 2 paper to remove mycelia and subsequently through a 0.22 μm Millipore filter to eliminate conidia.
Freeze-dried culture filtrate (CF) was dissolved in 5 ml of acetone.
After placing 5 μl of CF on each of the leaves collected from 28-to 30-day-old Ws-0, they were monitored for 3 days. Staining of the CF-treated leaves using Evans blue was performed as previously described (Park et al., 2009).

| Identification of A. thaliana mutants exhibiting increased susceptibility to M. oryzae
After inoculating 3300 M 1 plants with conidia of M. oryzae 70-15, seedlings that displayed increased disease symptoms compared to Ws-0 were identified at 6 dpi. Twenty-one putative mutants with increased susceptibility were isolated and named as more1 to more21.

| RNA isolation from A. thaliana and RT-qPCR to quantify transcripts from MORE1
Total RNA was extracted from Ws-0 and the more1 mutant using Easy-spin total RNA extraction kit (iNtRON Biotechnology).
First-strand cDNAs were synthesized using 2 μg of total RNA and ImProm-II Reverse Transcription System (Promega) with oligo(dT) primers. The resulting cDNAs were used for real-time quantitative PCRs to quantify the transcripts from the actin and MORE genes.

| Complementation of the more1 mutation
A fragment that contains the MORE1 gene, including its promoter region (c.5 kb from the start codon), was amplified from Ws-0 using the primers 5′-AAAAAGCAGGCTCGCTGCTGAACTCTTCGTCGAG-3′ and 5′-AGAAAGCTGGGTTGTAATTTCAAGCACTAATTACGACTC-3′. The amplified fragment was cloned into pENTR/TEV/D-TOPO (Invitrogen), its sequence was verified via sequencing, and it was then transferred into pHGW, a plant transformation vector, using LR clonase (Invitrogen). Agrobacterium tumefaciens GV3101 was used for transforming more1 plants with this fragment via the floral dip method (Clough & Bent, 1998). Transformants were selected on solid Gamborg B5 growth medium (Sigma-Aldrich) containing 50 µg/ml kanamycin (Sigma-Aldrich).

| Transcriptome analysis via RNA-Seq
Total RNA was extracted from leaves of Ws-0, more1, Dongjin, and osmore1a using a commercial kit (iNtRON Biotechnology).
Thermo Fisher Scientific Nanodrop 2000 and Agilent Bioanalyzer 2100 were used to check the quality and purity of extracted RNA.

RNA-Seq libraries were prepared using a TruSeq RNA Library Prep
Kit (Illumina) and sequenced using Illumina HiSeq2500 at NICEM (Seoul National University). Paired-end sequences were generated.
The resulting sequence reads were trimmed to remove adaptor sequences, and those with a quality score lower than 20 were removed using the NGS QC Toolkit v. 2.3.3 (Patel & Jain, 2012). All reads were assembled and mapped to the annotated genes available in The Arabidopsis Information Resource 10 (TAIR 10) (https:// www.arabi dopsis.org) (Lamesch et al., 2012)

for A. thaliana and the
International Rice Genome Sequencing Project (IRGSP) (Kawahara et al., 2013) for rice via the use of HISAT2 v. 2.1.0 (Kim, Langmead, et al., 2015) and StringTie v. 1.3.5 (Pertea et al., 2015). Genes were considered differentially expressed if their transcript abundance was ≥2-fold higher or lower in the more1 and osmore1a mutants than Ws-0 and Dongjin, respectively. The abundance of assembled transcripts was calculated in fragments per kilobase of exon model per million mapped fragments (FPKM) to analyse the normalized expression data derived from each library. The genes with FPKM of >1 at least in one library were considered as detected genes. Differentially expressed genes (DEGs) were identified using GFOLD v. 1.1.2 with the criteria of absolute log 2 (fold change) ≥1 and Gfold value ≠ 0 (Feng et al., 2012).
Gene ontology (GO) enrichment analysis of DEGs was performed using the DAVID v. 6.8 database (https://david.ncifc rf.gov/) (Huang et al., 2009a(Huang et al., , 2009b. A GO term with false discovery rate (FDR) ≤0.05 was considered significantly enriched by DEGs. The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was performed via a hypergeometric examination to identify which pathways are enriched among DEGs. The KEGG pathway analysis was executed to retrieve the enriched pathways with p ≤ .05. The resulting patterns were presented as a scatter diagram (Kanehisa & Goto, 2000). The enrichment factor, ratio, and number of genes that were enriched in a pathway were used to measure the degree of enrichment. Additionally, the MapMan package was used to get the graphical representation of the DEGs playing roles in biotic stress response and metabolic pathways (Thimm et al., 2004).

| Validating gene expression profiles using RT-qPCR
For RT-qPCR analysis of selected A. thaliana and rice genes, 5 μg of total RNAs was reverse-transcribed using ImProm-II Reverse Transcription System (Promega). The resulting cDNAs were diluted to 12.5 ng/μl. Primer pairs were designed using the 3′-end exon of the target genes (GC contents = 40%-50% and Tm = 58°C) (Table S2)

| Identification of the MORE1 homologues in rice
To identify the rice genes encoding members of the glycosyl hydrolase family 10 (GH10), we first retrieved all reported GH10 family genes from the Carbohydrate-Active EnZymes (CAZy) database (Cantarel et al., 2009)

| Sequence alignment and phylogenetic analysis
Amino acid sequences of the MORE1 protein and its homologues in rice were imported and edited using BioEdit v. 7.2.1 sequence alignment editor (Hall, 1999) and aligned using the default set parameters of ClustalW (Thompson et al., 1994). CLC Sequence Viewer v.
A phylogenetic tree was constructed using the maximum-likelihood method in MEGA 7.0 (Kumar et al., 2016) with the bootstrap test replicated 1000 times.

| Analysis of gene expression in rice infected with M. oryzae using RT-qPCR
Growth conditions, the RNA extraction method from infected rice, cDNA sample preparation, and other relevant information were previously reported (Jeon et al., 2020). A previously conducted cytological study (Jeon et al., 2020)  The qPCRs were performed in triplicate, and the data are presented as mean ± SD. The primers used for RT-qPCR are listed in Table S2.
The annealed double-stranded spacers were inserted into the BsaI-digested pOs-sgRNA vector. Gateway cloning LR reaction (Invitrogen) of the resulting constructs was performed using the destination vector pH-Ubi-cas9-7 that contains the Cas9 gene under the control of the maize Ubiquitin (Ubi) promoter (Miao et al., 2013). Cultivar Dongjin was transformed with sequence-confirmed vectors using an A. tumefaciens-mediated transformation protocol (Jeon et al., 2000). A. tumefaciens LBA4404 harbouring individual CRISPR/Cas9 constructs was grown on AB medium (K 2 HPO 4 3 g/L,  Table S2.

| Rice infection assays
After spraying each 2-week-old seedling with 10 ml of conidial suspension of strain PO6-6 (5 × 10 4 conidia/ml in 250 ppm Tween 20), the inoculated plants were incubated at 25°C for 1 day at 100% relative humidity in the dark and then at 28°C for 10 days in a growth chamber (28°C, 80% humidity and 16 h light/8 h dark). The lesion size was quantified using ImageJ. We also applied 10 μl of conidial suspension (5 × 10 6 conidia/ml in 250 ppm Tween 20) to each pressinjured spot (2 mm in diameter) on leaves (three to six spots per leaf) of 2-month-old plants. After keeping the inoculated plants in a chamber at 25°C and 100% relative humidity for 1 day, they were transferred to a growth chamber set at 28°C. Leaves were photographed at 9 dpi, and the size of each lesion was measured.
We used leaf sheaths to compare the degree of M. oryzae penetration and proliferation microscopically. After injecting a conidial suspension of PO6-6 (2 × 10 4 conidia/ml in sterile water) to excised rice sheaths from 5-week-old seedlings, they were placed in a box with moistened paper towels at room temperature. Inoculated sheaths were trimmed to remove chlorophyll-enriched parts at 36 hpi. Epidermal layers of the midvein (three or four cell layers thick) were observed using an Axio Imager A1 microscope (Carl Zeiss). Differential interference contrast (DIC) images were acquired using an AxioCam HRc camera and Axiovision v. 4.8.
This strain was cultured on peptone sucrose agar (10 g peptone, 10 g sucrose, 16 g agar, and 1 g glutamate per litre, pH 7.5) at 28°C for 3 days. Bacterial cells collected via centrifugation were suspended in sterile water to OD 600 = 0.8. Disease severity was assessed at 14 dpi by measuring the length of water-soaked lesions.
Conidia of C. miyabeanus were resuspended in 250 ppm Tween 20 at a concentration of 10 3 conidia/ml. After spraying 4-week-old rice seedlings with 10 ml of conidial suspension, they were placed in a growth chamber set at 25°C and 100% relative humidity in the dark for 1 day followed by 4 days of incubation at 28°C. All infection assays were performed three times in triplicate. Epifluorescence and DIC images were obtained using an Axio Imager A1 microscope (Carl Zeiss).

| Quantification and statistical analysis
The numbers of plants used for each treatment and experimental replicates are noted in relevant figure legends. All statistical analyses were performed using Microsoft Excel and a two-tailed, two-sample t test.

ACK N OWLED G M ENTS
Insightful suggestions from the reviewers greatly helped improve the manuscript. This work was supported by grants from the National