Discussion
In this study, we carried out a comprehensive transcriptome study of susceptible and resistant rice cultivars over a 48 h time course of infection with the blast fungus M. oryzae . Our data showed that high-amplitude transcriptional reprogramming was kicked off very quickly in rice-M. oryzae compatible and incompatible interaction. Large numbers of DEGs were identified, and a core set of genes involved in rice stress responses were defined. Furthermore, our study not only provided an overview of transcriptional reprogramming during rice-M. oryzae interactions but also identified and functionally validated three novel players in rice blast resistance.
A number of rice blast resistance genes and signaling regulators have been mapped and identified after decades of study (Li et al., 2019; Liu et al., 2014). However, the full regulatory network and immune signaling components remain elusive. Here we designed a comprehensive RNA-seq experiment to study the transcriptional profiles of rice-M. oryzae compatible (susceptible Nipponbare) and incompatible (resistant Hui1586) interactions. Four time points were included in the experiment, from very early at 12 hpi to later at 48 hpi with 12 h intervals between each point. Because the circadian clock controls the expression of ~ 30% of the transcriptome in plants (Harmer, 2009), mock-treatment (H2O) samples were included as controls at each time point to ensure a proper comparison for identifying DEGs. Our RNA-seq data showed that the circadian clock is the major factor determining the transcription dynamics (Fig. 1A, H2O treatment). Thus, it was essential to compare M. oryzae treatment data with the mock data to identify DEGs during the rice response to blast fungi. In this way, thousands of DEGs were identified at each time point, far exceeding the numbers in previous reports (Wei et al., 2013; Y. Zhang et al., 2016). In total, the expression of ~25% of the rice transcriptome was affected by M. oryzae treatment (Fig. 2C and 2D), indicating that great transcriptional changes occurred in rice challenged with blast fungus. Remarkably, the largest number of DEGs (6808 upregulated and 2895 downregulated, 9703 in total) was observed in the resistant Hui1586 at 12 hpi of blast fungus. Compared with the smaller number of DEGs in the susceptible Nipponbare (4680 upregulated and 2045 downregulated genes), we concluded that the resistant cultivar induced active defense more quickly and strongly than the susceptible cultivar. Unexpectedly, the number of DEGs declined in Hui1586 to 2849 (2224 upregulated and 625 downregulated) at 24 hpi, which was ~1/3 of the DEG number at 12 hpi. This result suggested that the strongest transcriptional reprogramming occurred before 24 hpi, and the critical time period for M. oryzae invasion was probably less than 12 hours. RNA-seq data for more time points, earlier than 12 hpi, is needed to determine the critical time points for rice-M. oryzaeinteractions.
In addition to transcriptional reprogramming, global translational reprogramming has been shown to be another fundamental regulatory layer of plant immunity (Xu et al., 2017). A consensus sequence, R-motif, present on a large number of messenger RNA, regulates translation in response to PTI induction (Xu et al., 2017). A latest study has shown that plants protect stem cells against viral infection through impairing global protein synthesis and limiting the replication and spread of the virus (Wu et al., 2020). In our transcriptome analysis, we found that among the common upregulated DEGs at 12 hpi in both Nipponbare and Hui1586, “ribosome” was the most significantly enriched pathway in the KEGG analysis, and “translation” was the most significantly enriched term in the GO analysis (Fig. 4C). These results strongly indicated that the protein translation machinery was regulated during the rice response to M. oryzae . Consistent with our results, a previous transcriptome study of rice Pi21- silenced plants infected byM. oryzae showed that “ribosome” is the third most enriched pathway compared with Nipponbare plants (Y. Zhang et al., 2016).Pi21 encodes a cytoplasmic proline-rich protein that negatively regulates rice blast resistance, and silencing of this gene results in enhanced resistance to blast fungi (Fukuoka et al., 2009). However, enrichment of the “ribosome” pathway was only observed at 12 hpi, not 24, 36, or 48 hpi, indicating that the regulation of protein synthesis in rice responses to M. oryzae was transient. Protein synthesis is a high-energy consumption process in living cells. It consumes approximately two-thirds of the total energy produced by a rapidly growing Escherichia coli cell (Jewett, Miller, Chen, & Swartz, 2009). Plant defense also imposes a substantial demand of resources and energy, which negatively impact growth (Huot et al., 2014). Activation of defense would decrease the overall pool of energy reserves through diminishing photosynthesis. It is possible that the demand of energy during the quick activation of plant immunity might reduce energy for the ribosome, the protein translation factory, slowing down the translation globally. As a negative feedback, the transcription of ribosome-related genes was upregulated accordingly. Further biochemical experiments are required to examine whether the global protein translation was transiently suppressed during the activation of rice immunity. However, in agreement with this hypothesis, protein synthesis inhibitor cycloheximide treatment of Arabidopsis induces transcriptional reprogramming similar to pathogen treatment (Navarro et al., 2004).
Among more than 10 thousand DEGs identified from four time points, only 1464 and 578 DEGs were shared by all four time points in Nipponbare and Hui1586, respectively, indicating that the transcriptome in the rice response to blast fungus infection was changing dynamically. A small set of DEGs shared by Nipponbare and Hui1586 at all time points were identified. These common DEGs might be important for driving overall immune transcriptional reprogramming. Indeed, only defense-related signaling and metabolism pathways were enriched in the common DEG group, and therefore, these genes were considered the “core” genes inM. oryzae- induced transcriptional reprogramming. Expressions of most of the core genes were also induced by other pathogens, such as the fungal U. virens and bacterial X. oryzae (Fig. 6A). Furthermore, the core genes were largely repressed during drought stress (Fig. 6B). Thus, the core genes were likely common genes involved in rice defense against pathogens and abiotic stress responses. Supporting this conclusion, among 14 TFs from the core genes, 5 TFs have been functionally characterized as important regulators of disease resistance, and 2 TFs have been reported to be involved in rice tolerance to temperature stress. We speculate that the core gene set contains convergence points between biotic and abiotic stress signaling pathways.
In this study, we identified a peroxidase gene Perox4 that plays a negative role in rice resistance to blast fungus (Fig. 8). The main role of peroxidase in plant immune responses is to aid in maintaining hydrogen peroxide (H2O2), the most stable ROS, at the appropriate level, which is toxic to plant cells at high concentrations (Qi, Wang, Gong, & Zhou, 2017). Three rice peroxidases (Os05g0135200 , Os10g0536700 , andPerox3 ) have been reported to contribute to BSR-D1, a C2H2-type transcription factor-mediated susceptibility to blast disease (Li et al., 2017; Zhu et al., 2020). BSR-D1 induces the expression of these peroxidase genes by direct DNA binding. An allele, Bsr-d1, in the rice cultivar Digu confers broad-spectrum resistance to M. oryzae (Li et al., 2017; Zhu et al., 2020). Notably, the expression of Perox4 is compromised in bsr-d1 knockout plants (Zhu et al., 2020), suggesting it is one of the target of BSR-D1. However, it remains inconclusive whether BSR-D1 can bind to the Perox4 promoter (Zhu et al., 2020). In our RNA-seq data, Perox3 was upregulated in Nipponbare and Hui1586, while the mRNA of Os05g0135200 andOs10g0536700 were barely detectable even with M. oryzae -treatment in both cultivars. In addition, another 4 peroxidase genes were upregulated by M. oryzae in Nipponbare and Hui1586 (Supplemental table 7). The peroxidase genes were proposed to be hijacked by M. oryzae through activation of the BSR-D1gene to counter the ROS burst induced by M. oryzae infection. It is reasonable to speculate that as an important susceptibility factor,Perox4 might be hijacked by blast fungi to suppress host immune responses.
In addition to ROS-scavenging enzymes, low-molecular mass antioxidants, including glutathione, ascorbate, carotenoids and (MTs), are involved in ROS maintenance (Qi et al., 2017; Yang, Wu, Li, Ling, & Chu, 2009). MTs are small, cysteine-rich, metal-binding proteins that are involved in metal homeostasis and detoxification in both plants and animals (Yang et al., 2009; Zimeri, Dhankher, McCaig, & Meagher, 2005). It has been shown that OsMT1a improves drought tolerance when overexpressed in rice by not only participating in ROS scavenging but also regulating the expression of zinc finger-type TFs via the regulation of Zn2+ homeostasis (Yang et al., 2009). However, unlike the ROS-scavenging enzyme Perox4 , simultaneous knockout ofOsMT1a and its close homolog OsMT1b in Nipponbare further enhanced susceptibility to blast fungus (Fig. 9). Thus, OsMT1positively regulated drought tolerance and blast disease resistance. DNA motifs of several zinc finger-type TFs were significantly enriched in the promoter of 321 core genes (Fig. 7C and Supplemental dataset 11), suggesting that OsMT1 might regulated blast disease resistance in a manner similar to the regulation of drought stress resistance, partly through zinc finger-type TFs.
In summary, this study provides a high-quality, comprehensive RNA-seq data set for rice-M. oryzae interactions and enhances our understanding of the transcriptional networks in rice immune responses to the blast fungus, highlighting possible candidate genes that may play important roles in rice disease resistance and abiotic stress tolerance.