Transcriptome and differential expression analysis revealed the pathogenic-related genes in Magnaporthe oryzae during leaf and panicle infection

Abstract

Magnaporthe oryzae is one of the most destructive pathogens that threaten rice production around the world. Previous studies mainly focus on pathogenic mechanism of M. oryzae during infection on rice at leaf stage. However, the pathogenic mechanism of M. oryzae infection on panicle tissue is not well understood. In the present study, we performed RNA sequencing (RNA-seq) to study gene expression patterns of M. oryzae during infection at leaf stage and at panicle stage, respectively. The differentially expressed genes (DEGs) of M. oryzae in the infected leaf and panicle tissues were analyzed. Gene ontology (GO) enrichment analysis of DEGs revealed that M. oryzae genes involved in the biological processes were different at leaf and panicle stages. Furthermore, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of DEGs indicates that genes related to individual and important pathways may function at different infection stages. In particular, CAZymes carbohydrate esterases (CEs), carbohydrate-binding modules (CBMs), and glycoside hydrolases (GHs) may play important roles during M. oryzae infection on rice leaves, while glycosyltransferases (GTs) and GHs may play important roles during infection at rice panicle stage. Further analysis of effectors (BAS3, BAS113, BAS162, MoCDIP4, and MoHEG13) and their homologous genes suggest that they are involved in host defense suppression. Our findings provide insights into understanding the infection mechanisms of M. oryzae for rice leaf blast and panicle blast disease.

Background

Rice blast disease, caused by the hemibiotroph fungal pathogen Magnaporthe oryzae, is one of the most important diseases of rice (Oryzae sativa) worldwide. The pathogen causes symptoms on leaves and panicle necks known as leaf and panicle blast, respectively (Khan et al. 2014). In particular, neck blast has major impact on rice yield loss. Although there is a high resistance correlation between leaf blast and panicle blast, leaf blast resistance does not conclusively confer resistance to panicle blast (Ou & Nuque 1963; Balal et al. 1977). Not only the resistance of a given variety to different M. oryzae races is different, but also the resistance to the same race may be different at different rice growth stages (Hu et al. 2014; Du et al. 2021b). Therefore, clarifying the distinct infection mechanisms of M. oryzae on rice at leaf and panicle stages will help better understand the interactions between rice and M. oryzae.

Phytopathogenic fungi can be classified into two types: those infecting plant leaves and stems and those proliferating in root tissues (Agrios 1997). Infection of these tissues and the subsequent survival within them require distinct strategies. Previous research showed that M. oryzae mutant strains lacking karyopherin exportin-5 (EXP5) gene were much less virulent on plant roots revealed an important function of EXP5 during M. oryzae colonization of underground plant tissues (Tucker et al. 2010). For leaf tissue infection, M. oryzae has emerged as a paradigm for molecular genetic dissection of factors that determine fungal pathogenicity on plant leaves. Mutational analyses have identified a number of genes in M. oryzae that are required for its colonization of leaf tissue (Dean et al. 2005; Oh et al. 2008; Mehrabi et al. 2009; Kong et al. 2013). So far, more than 1600 genes in M. oryzae have been investigated by targeted gene deletion or mutagenesis, covering over 10% of the genome (Yan et al. 2023). However, it is not well known about the factors that are required for successful colonization on neck tissues, despite the importance of neck blast.

RNA sequencing (RNA-seq) is a useful and affordable method to analyze gene expression patterns for either pathogen or host during infection, and helps understand host and pathogen interaction. Transcriptomes enable us to reveal infection-specific expression of M. oryzae genes in leaf and neck tissues (Mosquera et al. 2009; Soanes et al. 2012; Shimizu et al. 2019; Jeon et al. 2020). Recently, Mahesh et al. 2021 reported several M. oryzae pathogenicity genes related to tissue specificity during rice-Magnaporthe interactions by transcriptome analysis, resulting in the identification of 439 genes that were specifically expressed in neck-infected fungus, with 360 of them being hypothetical proteins (Mahesh et al. 2021). Of the identified genes, mitochondrial chaperone BCS1 (MGG_13867), secretory lipase (MGG_14628), cutinase (MGG_01943), ferric reductase transmembrane component (MGG_02828), endo-1,4-beta-xylanase B (MGG_08331), fungal cellulose binding domain-containing protein (MGG_01403), glycine cleavage system T protein (MGG_04826), 2-(R)-hydroxylpropyl-CoM dehydrogenase (MGG_12982), GPI inositol-deacylase (MGG_01844), and chitin deacetylase (MGG_05828) were more than 40-fold up-regulated in rice neck infecting fungus. Sixteen genes were specifically expressed in leaf-infecting fungus and most of them were hypothetical proteins (Mahesh et al. 2021). However, there are few reports on infection expression related genes in M. oryzae that are specific expressed during the infection on rice panicle stage. Therefore, research on the expression profile of M. oryzae upon infection in leaf and neck could help understand the epidemics of blast disease.

In this study, transcriptomes of O. sativa L. ssp. japonica cv. ‘Nipponbare’ (Nip) inoculated with M. oryzae GUY-11 was used to compare gene expression profiles of M. oryzae during leaf and neck infection stages at 8, 24, and 48 h post-inoculation (hpi). Our objective aims to analyze the differentially expressed genes of M. oryzae between the infection on rice at leaf and neck stages, and to identify potential genes may serve as fungicide targets for rice blast control in future.

Results

Transcriptome sequencing and data analysis

To study the differential regulation of M. oryzae during the infection on rice leaf and panicle, the conidia of M. oryzae strains GUY-11 were used to inoculate the plants. The inoculated leaf and panicle tissues of rice variety Nipponbare (Nip) were collected at 8, 24, and 48 hpi, respectively. Nip leaf and panicle tissues inoculated with sterile distilled water were defined as control samples leaf-CK (L-CK) and panicle-CK (P-CK) (Fig. 1). Total RNA of leaf and panicle tissues was extracted, and RNA-seq of the collected samples was performed. Approximately 68 to 80 million pairs of reads from each sample were used in the downstream analysis. All clean reads were mapped to the reference genome of M. oryzae 70-15 (Genebank accession No. GCA_000002495.2). An overview of the mapped statistics is provided (Table 1).

Fig. 1

Leaf and panicle tissue samples were collected at 8, 24, and 48 hpi, respectively. Leaf and panicle tissues were inoculated by conidia suspension (1 × 10⁵ conidia/mL). Leaf and panicle were inoculated by sterile distilled water as blank control (L-CK and P-CK)

Table 1 Summary of alignment statistics in 19 libraries referring to Magnaporthe oryzae 70-15 genome

The differentially expressed genes in M. oryzae during infection on rice leaf and neck tissues

The differentially expressed genes (DEGs) were identified with adjusted p-values < 0.01 and at least a 1.5-fold change in the normalized Fragments Per Kilobase of exon model per Million mapped fragments (FPKM) expression values. We compared transcriptome profile of M. oryzae from infected leaf and panicle at 8, 24, and 48 hpi, using the transcriptome of M. oryzae conidia as a control. There were 440 genes (283 up-regulated and 157 down-regulated), 737 genes (576 up-regulated and 161 down-regulated), 877 genes (628 up-regulated and 249 down-regulated) were differentially expressed in leaf infection stage compared to conidia (control), respectively. Meanwhile, there were 1432 genes (760 up-regulated and 672 down-regulated), 3355 genes (1891 up-regulated and 1464 down-regulated), 4528 genes (2425 up-regulated and 2103 down-regulated) were differentially expressed in M. oryzae during infection on panicle (Fig. 2a, b and Additional file 1: Table S1). The comparison of M. oryzae genes indicated that 333 genes were common, and these genes were both differentially expressed at L8 and P8 stages compared to conidia. There were 107 genes differentially expressed compared to conidia at L8 but not at P8 stage and 1099 genes were differentially expressed at P8 but not at L8 stage. At 24 h, 644 were common between L24 and P24 stage, 93 and 2711 genes were differentially expressed at L24 and P24 stage, respectively. Similar analysis indicated that 814 genes at 48 hpi were shared between L48 and P48 stage, 63 and 3714 genes were differentially expressed at L48 and P48 stages, respectively (Fig. 2c, Additional file 2: Table S2, and Additional file 3: Table S3). Venn diagram analysis showed that M. oryzae genes are differentially expressed at the stages of leaf and panicle infection.

Fig. 2

Overview of differentially expressed genes (DEGs) identified in M. oryzae at the initial stages of leaf and panicle. a, b The number of differently expressed genes at L8, L24, and L48, and P8, P24, and P48. c Venn diagram showing commonly and specifically expressed genes in M. oryze at leaf and panicle infection stages. L8, L24, and L48 are leaf-infected samples. P8, P24, and P48 are panicle-infected samples

Gene ontology enrichment analysis of the DEGs

To investigate the underlying mechanisms in M. oryzae during the two infection stages, GO classification of DEGs at 8, 24, and 48 hpi was analyzed. The results of GO enrichment analysis of M. oryzae DEGs are provided in Additional file 4: Table S4 and Additional file 5: Table S5. Go terms were visualized in biological process, cellular component, and molecular function categories, which revealed the GO functional classification of DEGs (Fig. 3). A comparison of the number of three GO terms showed that GO terms at L8 (10, 11, 5) and P8 (16, 13, 11) were enriched in biological process, cellular component, and molecular function categories, respectively. Similarly, GO terms at L24 (6, 9, 4) and P24 (17, 13, 12) were enriched in these three categories, as GO terms at L48 were 8, 7, 4, and P48 were 18, 14, 16 in these three categories. Notably, the number of GO terms at panicle infection stages was greater than that at leaf infection stages. Further analysis revealed that DEGs at panicle infection stage (P8, P24, and P48) were significantly enriched in biological process, like “metabolic process”, “cellular process”, and “single-organism process”. Additionally, DEGs analysis demonstrated significant enrichment in cellular component, including “membrane”, “membrane part”, “cell”, and “cell part”. Furthermore, the molecular functions “catalytic activity” and “binding” were also significantly enriched among these DEGs. Overall, it is evidence that M. oryzae DEGs at panicle infection stage may be involved in more important functional processes than leaf infection stage.

Fig. 3

GO classification of differentially expressed genes in M. oryzae at different leaf and panicle infection stages

Pathway analysis of the DEGs in M. oryzae

For pathway analysis, we mapped all DEG with more than 1.5-fold differential representation to terms in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, and then searched significantly enriched pathway terms against the genome annotation. To perform functional classification and pathway assignment of genes that are activated in M. oryzae, DEGs in each comparison were mapped with the KEGG database. For the leaves at 8, 24, and 48 hpi, the KEGG enrichment pathways of 107, 93, and 63 DEGs were analyzed (Additional file 2: Table S2). We found 4 DEGs at 8 hpi were enriched in “amino sugar and nucleotide sugar metabolism” pathways, and 3, 2, 2, 2 DEGs at 24 hpi were enriched in “Biosynthesis of amino acids”, “ABC transporters”, “Cysteine and methionine metabolism”, and “Glyoxylate and dicarboxyolate metabolism” pathways, respectively, and 2 DEGs at 48 hpi were enriched in “Glutathione metabolism” pathways (Fig. 4 and Additional file 6: Table S6). However, many DEGs at panicle infection stage exhibited different pathways. For panicle infection at 8, 24, and 48 hpi, there are 1099, 2711, and 3714 DEGs analyzed based on KEGG enrichment (Additional file 3: Table S3). We found that DEGs were mainly enriched in “Carbon metabolism” pathway from 8 h (36 genes) to 24 h (64 genes) post inoculation, In addition, DEGs were mainly enriched in “Ribosome” pathway from 24 h (54 genes) to 48 h (79 genes) post inoculation (Fig. 4 and Additional file 6: Table S6). The results suggest that M. oryzae genes related to these pathways may perform individual and important functions on leaf and panicle infection stages.

Fig. 4

Enrichment factor scatter plot of significant KEGG pathways in M. oryzae at different leaf and panicle infection stages. The horizontal axis represents rich factors, and the vertical axis represents metabolic pathways. The location of bubbles represents the enrichment item, the size of bubbles represents the number of differential genes, and the color of bubbles represents the significant degree of enrichment

Analysis of M. oryzae carbohydrate-active enzymes during leaf and panicle infection

Phytopathogenic fungi produce cell wall degrading enzymes (CWDEs) to breach the plant cell wall, which is the most important physical barrier during plant-pathogen interaction (Quoc & Chau 2017; Yang et al. 2021). Carbohydrate-active enzymes (CAZymes) are involved in the metabolism of glycoconjugates, polysaccharides, and oligosaccharides. For plant pathogens, CAZymes help in the degradation of the host cell wall and storage compounds (Zerillo et al. 2013). According to their functions, CAZymes can be classified into glycoside hydrolases (GHs), glycosyltransferases (GTs), polysaccharide lyases (PLs), carbohydrate esterases (CEs), auxiliary activities (AAs), and carbohydrate-binding modules (CBMs) (http://www.cazy.org/). We predicted CAZymes of M. oryzae using the dbCAN web server 2 (Yin et al. 2012), HMMER (Finn et al. 2011), and DIAMOND (Buchfink et al. 2015), in which we identified CAZymes-coding genes for DEGs at the leaf infection and panicle infection stages. Subsequently, we found that most DEGs related to CAZymes in M. oryzae were obviously increased from at least 2-fold to 100-fold expression levels at leaf infection stage. At L8, the expression levels of M. oryzae CE5 subfamily genes (MGG_14095, MGG_02393, MGG_09100, MGG_03440), GH subfamily genes (MGG_05533, MGG_11231, MGG_01885, MGG_09733, MGG_01096), CBM subfamily genes (MGG_03857, MGG_07264), and AA subfamily genes (MGG_14940, MGG_01255) were significantly increased compared to P8 stage. Similarly, the expression levels of M. oryzae CE4 subfamily gene (MGG_14966), CE8 subfamily gene (MGG_14400), CE5 subfamily genes (MGG_15403, MGG_11966, MGG_09100), GH10 subfamily gene (MGG_01542), AA9 subfamily gene (MGG_07575) were significantly increased compared to panicle infection at 24 hpi. For leaf infection at 48 hpi, the expression levels of GH18 subfamily gene (MGG_05125) and GT2 subfamily gene (MGG_07803) were slightly increased (Fig. 5a and Additional file 7: Table S7). Overall, CE, GH, and CBM families may function in the degradation of plant cuticles at the leaf infection stage.

Fig. 5

Differential expression patterns of CAZymes genes of M. oryzae during leaf and panicle infection stages. a Heatmap showing RNA-seq expression level of CAZymes in M. oryzae and panicle during infection on leaf. Normalized FPKM expression are shown in the heatmap. Value of each sample represents the average of normalized FPKM of three replications. Low expression of genes is shown in green rectangle and high expression is shown in red rectangle. L8, L24, L48 represented differential CAZymes genes at leaf stages compared to P8, P24, P48, respectively. P8, P24, P48 represent differential CAZymes genes at panicle stages compared to L8, L24, L48, respectively. P common represented these CAZymes genes are all differential at three panicle stages compared to leaf stages. b qRT-PCR was used to verify the expression levels of randomly selected CAZymes genes. The different letters denote significant differences with P < 0.05. L8, L24, and L48 represent leaf-infected samples. P8, P24, and P48 are panicle-infected samples, respectively

For panicle infection stage, most DEGs belong to GT and GH families. We found that fifteen DEGs were common in panicle-infected fungus (Additional file 8: Table S8), and these genes belong to GT subfamily (MGG_03860, MGG_04773, MGG_03185, MGG_00865, MGG_06064, MGG_13971, MGG_02954, MGG_01802, MGG_08182), and GH subfamily (MGG_00063, MGG_00316, MGG_04765, MGG_13455, MGG_02793, MGG_07101), which may be conserved during panicle infection stage. Besides, some genes belonged to GH subfamily were also found differentially and specifically expressed at panicle infection stage. These genes were significantly upregulated about 2-fold to 50-fold at individual panicle infection stage (Fig. 5a and Additional file 8: Table S8). The CAZymes for DEGs at 8 hpi included GH18 subfamily (MGG_17153), GT subfamily (MGG_01140, MGG_02384, MGG_03441, MGG_07289, MGG_13014), and those DEGs at 24 hpi included PL4_3 subfamily (MGG_06041), GT subfamily (MGG_00636, MGG_04979, MGG_07230), GH subfamily (MGG_00084, MGG_00677, MGG_05533, MGG_06023, MGG_07634, MGG_07715, MGG_11475, MGG_17864). DEGs related to CAZymes at 48 hpi mainly are GT subfamily (MGG_07979, MGG_05687, MGG_04145, MGG_06137, MGG_02859, MGG_03459, MGG_04514, MGG_07729, MGG_14118, MGG_09962), GH subfamily (MGG_02793, MGG_14903, MGG_11536, MGG_01885, MGG_00667, MGG_11210, MGG_10712, MGG_14954, MGG_05520, MGG_04689, MGG_07101, MGG_06834, MGG_03599, MGG_01912, MGG_13455, MGG_11408, MGG_09471). To confirm the expression pattern of DEGs obtained from RNA-seq analysis, six genes were selected randomly and evaluated by quantitative Real-time PCR (qRT-PCR) (Fig. 5b). Comparison of Log₂ fold change values of these genes in RNA-seq and qRT-PCR revealed similarity of gene expression profiles. Taken together, GT and GH families may play aimportant roles at rice panicle infection stage.

M. oryzae effectors were differentially expressed during leaf and panicle infection

During infection, M. oryzae secretes a set of effectors to disturb plant immune systems (van der Does & Rep 2007; Oliva et al. 2010). In addition to suppress plant immunity, effectors may target cell signaling and metabolic pathways to facilitate invasive fungal growth (van der Does & Rep 2007). Many M. oryzae effectors are recognized by rice immune receptors, leading to disease resistance (Yan et al. 2023). In this study, 46 known M. oryzae effectors were retrieved from previous study (Gómez Luciano et al. 2019). The hierarchical clustering (HCL) generated a global view of the expression level for verified M. oryzae effectors at six time points. As shown in Fig. 6a, most known M. oryzae effector genes were highly expressed at both leaf and panicle infection stages, such as BAS (1–4), MSP1, PWL2, SLP1, AVR-Pita1, AVR-Pia, AvrPi9, and MC69. Therefore, these M. oryzae conserved effectors are co-regulated during rice infection.

Fig. 6

Differential expression pattern of effector genes in M. oryzae during leaf and panicle infection stages. a Heatmap showing RNA-seq expression level of verified effector genes and their homologue genes in M. oryzae during leaf and panicle interaction. Normalized FPKM expression are shown in the heatmap. Value of each sample represent the average of normalized FPKM three replications. Low expression of genes is shown in green rectangle and high expression is shown in red rectangle. b qRT-PCR was used to verify the expression levels of randomly selected effector genes. The different letter denote significant differences with P < 0.05. L8, L24, and L48 represent leaf-infected samples. P8, P24, and P48 are panicle-infected samples, respectively

To find homologous effector-encoding genes in the DEGs, the BLASTP toolkit (E-value threshold: 1e^-15) and MCL software (inflation threshold: 4.0) were used to cluster M. oryzae genes based on their protein sequences. As shown in Additional file 9: Table S9, BAS2, BAS3, BAS113, BAS162, MoCDIP4, and MoHEG13 were found to have different numbers of homologous genes. Based on the FPKM expression values of M. oryze effectors and homologous genes at L8, L24, L48, P8, P24, and P48 stages, M. oryzae BAS2 homologous gene MGG_07749 was increased more than 20-fold at P8 stage compared to L8 stage, and BAS3 homologous gene MGG_16415 also was increased about 6-fold to 50-fold at panicle infection stage. Similarly, BAS113 homologous gene MGG_17302 was significantly increased about 8-fold at L24 stage compared to P24 stage. Besides, MoCDIP4 homologous gene MGG_04547, MGG_07300, and MGG_07686 were significantly increased about 10-fold to 40-fold at P24 and P48 stages. Interestingly, MoHEG13 (MGG_09378) and its homologous gene MGG_17582 both exhibited higher expression levels at leaf infection and panicle infection stages, but their homologous gene MGG_17319 only exhibited higher expression level at L8 stage compared to P8 stage (Fig. 6a and Additional file 9: Table S9). qRT-PCR was further used to validate RNA-seq results by randomly selecting M. oryzae effectors and homologous genes at the early stages of leaf and panicle infection (8, 24, and 48 hpi). The Log₂ fold change values of these genes in RNA-seq and qRT-PCR displayed similar expression profiles (Fig. 6b). The results show that some homologous genes were differentially expressed during leaf and panicle infection stages, indicating that these homologous genes may paly potential roles at different infection stages.

Discussion

The differences between leaf infection and panicle infection (dual epidemics) are foremost important to understand the molecular interplays for rice-pathogen interactions, which may help to effectively manage the disease. Previous studies mainly focused on infection mechanism of M. oryzae on leaf tissue. However, there are little researches on the infection mechanism of M. oryzae on panicle tissue. We have sequenced a genome-wide transcriptome during M. oryzae infection on early infected tissues using stranded RNA sequencing approach. Genes expressing differentially at leaf and neck infection stages will help to study dual epidemics. It will also help plant breeders to utilize these genes to develop blast disease resistance rice in a region-specific manner. Meanwhile, it is indicative for changes in gene signaling when pathogen invasion shifts from leaf infection to other tissue like neck and panicle.

In this study, we employed the RNA-seq technique to study gene expression pattern during rice-M. oryzae interaction at the leaf and panicle infection stages. The transcriptome assays for the non-infected and infected tissues of leaf and neck revealed the contrasting gene expression profile. By comparing the DEGs from three leaf and panicle tissues at 8, 24, and 48 hpi, we found that genes were differentially expressed in different tissues (Additional file 1: Table S2 and Additional file 3: Table S3). Though the reads matching to M. oryzae genome is slight low, the data are overall informative. We identified most DEGs induced at infection stage, which are consistent to previous reports. For example, CAZymes-coding gene MGG_02393 belonging to CE5 subfamily was found to be up-regulated as revealed by RNA-seq (Kawahara et al. 2012). The other gene MGG_09100 which encodes Cutinase 2 (Cut2), was also up-regulated at initial infection stage (Franck et al. 2013; Shimizu et al. 2019). In addition, our data support the expression patterns of known effectors, such as SLP1, BAS1, and MC69 (Dong et al. 2015).

GO enrichment analysis of DEGs revealed that DEGs were involved in three important processes (biological processes, cellular component, and molecular function) that were differentially enriched at the leaf and panicle infection stages (Fig. 3, Additional file 4: Table S4, and Additional file 5: S5). An interesting finding of pathway analysis of M. oryzae DEGs at leaf infection stage is the high expression of genes related to metabolism pathways. However, we found that genes involved in “Carbon metabolism” and “Ribosome” pathways were remarkably enriched in M. oryzae at panicle infection stage (Fig. 4 and Additional file 6: Table S6). These data revealed that infection-specific expression of genes in M. oryzae are involved in different functional processes and pathways.

Researches of leaf blast contributed to understand infection mechanism of M. oryzae in seedling stage. However, the study of neck blast is limited, and the information acquired from leaf blast may not be applicable to other infection types like nodal blast, panicle blast, and neck blast. How are the DEGs associated with the spatiotemporal interaction between M. oryzae and rice leaf or panicle? The gene expression pattern largely depends on different plant tissues and the M. oryzae strains (Mahesh et al. 2021). Due to the same genetic background of pathogen at leaf and panicle stages, we speculated that the difference might depend on molecular mechanisms between rice leaf or panicle and M. oryzae interactions. In other plant pathogens, many pathogenic-related genes were reported for some tissue-specific infection pathogens. As we all know, the ascomycete fungus Ustilaginoidea virens causes rice false smut (RFS), a unique floret disease. Functional genomics and transcriptome analyses predict that more than 1000 genes are involved in the virulence and pathogenicity of U. virens (Yu et al. 2023). These putative virulence factors are closely related to the pathways that have been identified to be implicated in U. virens pathogenicity, mycelial growth, conidiation, and stress tolerance. In Fusarium graminearum, Wanjiru and colleagues (Mary Wanjiru et al. 2002) showed that the pathogenicity of F. graminearum is dependent on the extracellular secreting enzymes. The fungus invades the host via the epidermis, leading to the destruction of the host’s cellulose, pectin, and xylan. Further research indicated that the F. graminearum Gpmk1 MAP kinase regulates the induction of extracellular endoglucanase, xylanolytic, as well as proteolytic activities (Jenczmionka & Schäfer 2005). In our study, we found that many genes involved in signal pathways, secondary metabolism, and transcriptional regulation pathways exhibited transcriptional differences between leaf and panicle stages (Additional file 2: Table S2 and Additional file 3: Table S3).

It is known that fungi constantly produce several CAZymes to degrade cell wall polysaccharides during infection. Leaf-infecting fungi often breach the hard waxy cuticle that coats aerial plant structures (Mendgen & Deising 1993; Gómez Luciano et al. 2019). Identification of CAZymes-coding genes in M. oryzae showed that CE, CBM, and GH families may play diverse roles in the degradation of plant cuticles at the leaf infection stage. MGG_02393, a CE5 subfamily gene, was found to be up-regulated (Kawahara et al. 2012), and MGG_09100 (Cutinase 2, Cut2) was also up-regulated at 12 hpi and down-regulated at 36 hpi (Franck et al. 2013; Shimizu et al. 2019). Our data are consistent with the previous reports. These CAZymes-coding genes are required for M. oryzae to sense the host tissues and to initiate differentiation, penetration, and full virulence. However, there are limited reports showing the relevance of CAZymes of M. oryzae at neck infection stage. Contrast to DEGs in M. oryzae at leaf infection stage, our findings indicate that GT and GH families may play an important role in M. oryzae during rice panicle infection. Taken together, the CAZymes CE, GH, and CBM families play essential roles during the attacks of M. oryzae on rice leaf, but GT and GH family of M. oryzae may act in rice panicle infection.

Effectors are secreted proteins by M. oryzae that often target plant immune system to suppress host defense and, as a result they contribute to the proliferation of the pathogen (Jones & Dangl 2006; Lo Presti et al. 2015). During infection, pathogens secrete effectors as biological weapons to help invade and propagate in host plants by targeting hosts’ physical barriers for disruption, where they create conducive conditions for invasion and proliferation. As the result, they disturb host cell physiological activity and dampen plant downstream immune responses. Therefore, it is essential to analyze the virulence of effectors at leaf and panicle infection stages. Figure 6 shows that most known effector are conserved during leaf and panicle infection processes. Further BLASTP analysis found that BAS2, BAS3, BAS113, BAS162, MoCDIP4, and MoHEG13 have different numbers of homologous genes. We found that effectors and their homologous exhibited different expression levels. Taking an example of effector BAS3, it has higher expression levels at leaf infection and panicle infection stages, but its homologous gene MGG_16415 was significantly increased about 6-fold to 50-fold at panicle infection stages. Another effector MoHEG13 (MGG_09378) and its homologous gene MGG_17582 both exhibited higher expression levels at leaf infection and panicle infection stages, but their homologous gene MGG_17319 only displayed higher expression level at L8 stage. Therefore, these data suggest that homologous genes might be involved in the different infection mechanisms of M. oryzae, in order to suppress host defense during leaf and panicle infection processes.

Conclusions

We performed a comparative analysis to investigate transcriptome of M. oryzae at leaf and panicle infection stages. The data showed that CAZymes CE, CBM, and GH families may play an important role during M. oryzae infection on rice leaf. However, GT and GH families may play important roles in M. oryzae infection rice panicle. Further analysis of effectors and their homologous proteins indicate that they may be involved in different mechanisms for M. oryzae that help to suppress host defense during leaf and panicle infection. Our results provide resources to characterize their functions in future. It is assumed that there are differential genes expressed in pathogen at different plant tissues due to the different set of defense and pathogenicity related arsenals deployed by plants. The infection-specific transcriptome data presented in this study will help to elucidate the molecular mechanism of M. oryzae during leaf and panicle infection.

Methods

Plant and fungal materials, growth conditions, and treatments

Seeds of rice cultivar Nipponbare (Nip) were surface sterilized, germinated, and grown in a greenhouse (26 ± 2°C and 16 h light/8 h dark). For pathogen inoculation, M. oryzae was maintained on complete media (CM) (6 g yeast extract, 6 g casamino acids, and 10 g of sucrose per litre) at 28°C. Later, they were transferred to straw decoction and corn media (SDC) (100 g straw, 40 g corn powder, 15 g agar in 1 L distilled water) and kept under 28°C for 7 days for reproductive growth. Then, they were grown under constant fluorescent light (YZ 36W T8 12, Xiguang, China) for 3 days. Conidia were harvested from the 7-day-old SDC medium, and dissolved them in sterile distilled water, and the concentration was adjusted after filtration through three layers of Miracloth (CalBiochem) (Zhang et al. 2011; Du et al. 2013).

For rice leaf infection assays, three-week-old rice plants were used for inoculation with M. oryzae strain Guy11 at 1 × 10⁶ conidia/mL. For rice panicle infection assays, conidial suspensions were injected into the rice sheath at the heading stage (Du et al. 2021a). Finally, the tissues were collected at 8, 24, and 48 hpi after inoculation. Three replicates were taken at each infection time point and the tissues were prepared for transcriptome sequencing.

RNA library preparation, Illumina sequencing, and analysis of the reads

The library for RNA sequencing was prepared using Illumina True-Seq RNA Library Prep Kit (San Diego, California, USA) by following the manual. Each biological replicate was a pool of five independent panicles. Three biological replicates were used for each sample and thus 19 samples (three replicates for each infection time point and a conidia sample used as control) were sequenced. Illumina NovaSeq 6000 platform was used to generate large amounts of sequencing data performing paired-end sequencing runs using more than 1 μg total RNA (RNA integrity number, RIN > 7) to obtain 150 bp sequence length reads. The raw reads were processed using the bioinformatics analysis platform BMKCloud (www.biocloud.net). Quality analysis of raw reads was further subjected to quality check by FastQC software (Andrews 2010). Finally, raw data is processed to obtain clean data using the BMKCloud platform.

Normalization of expression levels of genes from RNA-seq and gene annotation

All the clean reads were mapped to the reference genome of M. oryzae 70–15 (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000002495.2/) using TopHat version 2.1.1 with default parameters (Trapnell et al. 2009). The expression level of a gene from RNA-seq was normalized using Fragments Per Kilobase of exon model per Million mapped fragments (FPKM) method (Mortazavi et al. 2008). Gene annotation, including gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations, was referred by (Zhang et al. 2014).

Identification of DEGs of M. oryzae at infection stage

The gene expression analysis between M. oryzae-inoculated and conidia (control) samples was performed using the DESeq2 R package (1.16.1) (Anders & Huber 2010). Genes with a combination of P value < 0.01 and the absolute value of |Log₂ fold change|≥ 1.5 were regarded as DEGs. For grouping DEGs with similar expression patterns, a hierarchical clustering was generated using the expression values from each library (Liang et al. 2022). Analysis was conducted using Cluster 3.0 software with Pearson correlation as the distance measure. The cluster tree contained distinct clusters, which include genes with a unique expression profile by visual inspection.

Quantitative real time PCR (qRT-PCR)

Total RNA was extracted and purified by Qiagen RNAeasy Mini kit (Qiagen Inc., Valencia, CA, United States). RNA isolation was performed, and cDNA synthesis was carried out using the Superscript IV Reverse transcriptase cDNA synthesis kit (TB Green® Premix Ex Taq™ II). All cDNA samples were diluted to 20/ng prior to qRT-PCR. qRT-PCR was run on an ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) following the manufacturer’s instructions. Reactions were performed in a 20 μL volume containing 10 μL of SYBR premix Ex Taq (SYBR Prime Script RT-PCR kit; TaKaRa), 0.4 μL of ROX reference dye (SYBR Prime Script RT-PCR kit; TaKaRa), 2 μL of cDNA template (50 ng), 0.4 μL of each primer (10 mM), and 6.8 μL of sterile distilled water. Transcripts of genes were analyzed and the ACTIN gene (MGG_03982) was used as an internal control. The amplification conditions for the reactions were 95°C for 3 min, with 25 cycles of 95°C for 30 s, followed by 55°C for 30 s, 72°C for 30 s and fluorescence read at 72°C at the end of each cycle. The ‘Ct’ values were normalized based on ‘Ct’ value of reference genes and differential gene expression (fold change) was calculated as per 2^−ΔΔCt method. qRT-PCR was repeated in triplicate with three independent biological experiments, and the primer pairs used are listed in Additional file 10: Table S10.