Rice (Oryza sativa L.) is one of the most crucial crops worldwide, providing dietary food to over half of the world’s population as a main crop. Various diseases have threatened rice cultivation since its inception. Yield loss due to diseases is estimated to be approximately 30% of the global rice production (Savary et al. 2019). Climate change further increases outbreak risks by changing pathogen evolution and host–pathogen interactions and helping in the emergence of new pathogenic strains (Singh et al. 2023).

Rice blast, caused by the fungus Magnaporthe oryzae Couch (syn. Pyricularia oryzae Cavara) is one of the most destructive diseases affecting rice worldwide. Traditional cultural practices, such as the separation of infected seeds and strows and pesticide application, are basic and effective methods for managing rice blast emergence. However, these techniques are laborious and incur high cultivation costs. Chemical pesticide use results in environmental pollution and increases risks to human health. Thus, breeding resistant rice varieties is an effective and ecofriendly approach for managing rice blasts.

More than 129 genes have been identified to date, of which 39 have been isolated and molecularly characterized (Pedrozo et al. 2025). Despite great efforts in the identification of resistance genes and breeding of resistant varieties, durable and broad-spectrum resistance has not been achieved because of the rapid resistance breakdown (Fukuoka and Okuno 2019). Thus, the characterization of blast isolate pathogenicity and spatial and time-scale monitoring of pathogenicity alterations are crucial for the precise and effective breeding of functional resistance genes.

After the discovery of blast races by Sasaki (1922), Atkins et al. (1967) developed an international differential system for cooperative research between the US and Japan. Following gene analyses and the identification of gene-for-gene relationships between resistance in rice and avirulence in blast fungi, Yamada et al. (1976) and Kiyosawa (1984) developed a new differential system, which is currently being utilized for monitoring blast races in Japan (Yamada et al. 1979; Yamada 1985; Naito et al. 1999; Koizumi et al. 2007). In China, seven rice varieties were selected as Chinese differential cultivars (CDC) and 827 isolates collected from 23 provinces were classified into seven groups and 43 races (All China Corporation of Research on Physiological Races of Pyricularia oryzae 1980). Using this system, time-scale monitoring of pathogenicity alterations was performed in the Heilongjiang, Guangdong, and Jiangsu provinces (Zhang et al. 2019, Huang et al. 2021, Qi et al. 2022). In the US, 1022 blast isolates collected from 1959 to 2015 were analyzed for eight international rice differentials (Wang et al. 2017). Continuous monitoring of the pathogenicity and distribution is valuable for controlling rice blasts and breeding new resistant varieties.

In Vietnam, rice is a major and important food source for domestic consumption and exports (Khanh et al 2021). Vietnam’s tropical monsoon climate, which is characterized by high temperatures and abundant rainfall, creates favorable conditions for blast disease emergence (Yuan et al. 2021). Noda et al. (1999) analyzed the distribution of the pathogenic races of 129 blast isolates collected from 1995 to 1996 in the Mekong River Delta. Differential varieties harboring Pia, Pit and Piks genes were susceptible to most blast isolates, whereas those harboring Pish, Pik, Piz, Pita2, Pizt, Pikp, and Pib were resistant to all isolates. Fukuta et al. (2020) analyzed 94 blast isolates collected in 2007 and 2013 in the Mekong River Delta and showed that the blast isolates were virulent to Pib, Pit, Pia, Pi3, Pi5(t), Pik-s, Piz-t, Pi12(t), Pita, Pi19(t), Pi20(t), and Pita2 but avirulent to three alleles, Pi9, Piz, and Piz5 at the Piz locus. Furthermore, Thuy et al. (2015) identified that differential varieties harboring Pik, Pik-p, Pik-h, Piz, Pi1, Pi7(t), Pik-m, Pi4(t), Pish, Pi9(t), and Pita genes were highly resistant to blast under field conditions in the south-central coastal area. Thuy et al. (2020) also assessed the genetic diversity of 32 blast isolates collected from 2013 to 2017 in the South Central Coast and indicated that Pik, Pik-p, Pik-h, Piz, Pi4(t), Pish, Pi1, Pi7(t), Pi9(t), Pikm, and Pita were the most effective blast R genes in these areas, whereas ໿Pia, Piks, Pi3, Pi20(t), and Piz5 were susceptible. Recently, Nguyet et al. (2020) reported that 239 isolates of blast collected in northern and central Vietnam from 2012 to 2016 demonstrated a wide variation in pathogenicity. The frequencies of isolates virulent toward DVs for Pish, Pikm, Pi1, Pikh, Pik, Pikp, Pi7(t), Pi9(t), Piz5, Pita2, and Pita were low; however, they were high for DVs for Pib, Pit, Pia, Pii, Pi3, Pi5(t), Piks, Piz, Pizt, Pi12(t), Pi19(t), and Pi20(t). These results show that resistance genes were effective against blast isolates collected from each area. However, nearly a decade has passed since the last survey and no new surveys have been conducted on blast pathogenicity in Vietnam. Thus, the current characteristics and distribution of the pathogenic races have not yet been clarified.

Herein, we assessed the pathogenicity of blast fungi collected nationwide from Vietnam. Compared to previous characteristics of pathogenicity, we analyzed regional differences in pathogenicity and alterations in pathogenicity over time and discussed blast resistance breeding strategies that should be implemented in the future based on the current composition of resistance genes in major rice varieties.

We used a set of 25 differential varieties (DVs) carrying 23 resistance genes to assess the pathogenicity of the blast isolates. This set of DVs included 23 monogenic lines for resistance genes Pish, Pib, Pit, Pia, Pii, Pi3, Pi5(t), Pik-s, Pik-m, Pi1, Pik-p, Pi7(t), Pi9(t), Piz, Piz-5, Piz- t, Pita-2, Pita, Pi12(t), Pi19(t), and Pi20(t) (Tsunematsu et al. 2000) and 2 NILs for Pik-h and Pik (Telebanco-Yanoria et al. 2010) with the Lijiangxintuanheigu (LTH) genetic background. LTH was used as a susceptible control. The names of the 25 DVs are indicated as follows, using the blast resistance genes and the abbreviations for their donor lines: IRBL[Resistance gene name without “Pi”]-[Donor line abbreviation] (Tsunematsu et al. 2000, Telebanco-Yanoria et al. 2010).

The seeds of the 25 DVs and LTH were sterilized, soaked for 2 to 3 days, and subsequently sown at three seeds per cell in plastic trays (5 × 7 cells of 16-mm diameter × 25-mm depth), which were then placed in a greenhouse. Additionally, each tray contained an LTH for the verification of susceptibility. The seedlings were inoculated using the method described by Hayashi et al. (2009). The spore concentration was standardized to 1 × 10⁵ spores/mL in sterile water. Three weeks after sowing, when the seedlings had to 4–5 leaves, they were inoculated by spraying 15–20 ml of the spore suspension onto each tray using a fine atomizer. Inoculated plants were kept in a dew chamber at 25°C for 24 h and subsequently transferred to a greenhouse maintained at 25 to 30°C, maintaining humidity using a mist sprayer. The degree of infection in each seedling was assessed 7 days following inoculation, as described below. The inoculation and evaluation processes were performed two times.

According to the method described by Hayashi and Fukuta (2009), blast lesions were assessed on a scale of 0 to 5, where 0 implies no evidence of infection, and 5 implies large eyespot lesions more than twice the interval between large veins or > 2 mm in diameter. Lines were considered resistant (R) when rated from 0 to 2 and susceptible (S) when rated from 3 to 5, with the exception of line IRBLsh-B for Pish (resistant = 0 to 3), line IRBLta2-Pi for Pita2 (resistant = 0 to 3), and line IRBL5-M for Pi5(t) (resistant = 0 to 1).

The race designation of the virulent blast isolates was based on the reaction patterns of 25 DVs and LTH, following the system proposed by Hayashi and Fukuta (2009). Blast races were designated by a combination of codes representing the reactions of the DVs in each unit using the method described by Gilmour (1973). Isolates classified this way were designated as “reaction type” within each DV group and “races” using the set of all five reaction types, which corresponded with the five DV groups. The proportion of unique races in the four areas (NM, RRD, NC, and MD) was calculated as the G:N ratio, where G is the number of unique races, and N is the number of isolates in each area. Pathogen diversity was determined using Simpson’s diversity index (SI) (Simpson 1949) and evenness (EV) (Arnaud-Haond et al. 2005). The SI varies from 0 to 1, where 0 denotes no diversity and 1 denotes the maximum diversity. EV range from 0 to 1, where 0 indicates that all isolates have the same race code, and 1 indicates that all isolates have different race codes occurring at the same frequency.

Cluster analysis was conducted based on the obtained scores. R (version 4.4.2) was used to calculate the Euclidean distance matrix, and Ward’s D2 method was used to perform clustering analysis. The fviz_nbclust function of the factoextra package (version 1.0.3) was used, specifying FUNcluster = hcut and method = “wss” to determine the optimal number of clusters (Kassambara and Mundt 2016). Principal component analysis was conducted using the prcomp function, and the results were divided into several clusters obtained from the optimal cluster number consideration and visualized using the fviz_pca_biplot function.

We selected four provinces (Ha Noi, Ha Nam, Dien Bien and Long An) based on the following criteria: a province where the total number of blast isolates was over 50 and at least 10 blast isolates were collected before 2016 and after 2017. Based on the inoculation scores, the rice blast isolates collected from each province were visualized using hierarchical clustering analysis with the pheatmap function of the pheatmap package (version 1.0.13) using the Euclidean distance matrix and Ward’s D2 method (Kolde 2025).

Genomic DNA was extracted from young leaves of two rice varieties (“Thien Uu 8 [TU8]” and “Khang Dan [KD]”), which were cultivated in Ha Tinh province. Leaf tissue was ground in 100 µL of 0.25 N sodium hydroxide (NaOH) with zirconium beads in 2.0 mL tubes. A volume of 400 µL of 100 mM Tris-HCl (pH 7.5) was added and the sample was subsequently well mixed and centrifuged for 10 min at 10,000 rpm. The supernatant was transferred into 1.5 mL autoclaved tubes.

Polymerase chain reaction (PCR) was performed with 10 µL reaction mixture containing 5 µL template DNA, 1 µL 10x PCR buffer, 1µL of 2 mM dNTPs, 0.1 µL Taq DNA polymerase (5U/ul), 1.5 µL of a 2.5 mM solution primers, and 1.4 µL H₂O. PCR conditions were as follows: 94°C for 5 min, followed by 40 cycles (30 s at 94°C, 30 s at 55°C, and 2 min at 72°C) with a final extension of 7 min at 72°C. Amplicons were separated on a 2.5% agarose gel. After electrophoresis, the gel was stained with ethidium bromide and the DNA fragments were visualized under UV light. Kitazawa et al. (2019) have described the PCR primers used for genotyping each gene.

Each province was classified into seven ecotypic areas according to the classification based on the agricultural ecological characteristics. Blast isolates collected from three northern ecotypic areas, Northern Mountain (NM), Red River Delta (RRD), and North Central (NC), exhibited a high classification frequency into Cluster 1 (Fig. 2). Contrastingly, isolates from the four central and southern ecotypic areas, South Central Coast (SCC), Southeast (SE), Central Highlands (CH), and Mekong Delta (MD), were predominantly classified into Cluster 2 (Fig. 2).

To assess the temporal changes in the composition of each cluster, we focused on four ecotypic areas with large sample sizes (RRD, NM, NC, and MD) and compared the distribution of blast isolates collected before 2016 and after 2017. In all four areas, a statistically significant shift in cluster composition was noted (P < 0.001, chi-square test), with the proportion of isolates classified into Cluster 1 substantially increasing after 2017 (Fig. 3a). Particularly, in NM, the frequency of Cluster 1 isolates was already higher before 2016 and further increased in the later period.

In addition to the alterations in cluster composition, we noted a clear elevation in virulence frequency among the isolates after 2017. Specifically, the virulence against DVs (IRBLsh-B and IRBL-ta-CP1) harboring Pish and Pita resistance genes substantially increased (Fig. 3b). Furthermore, a moderate increase in virulence was detected in DVs (IRBL3-CP4, IRBL5-M, IRBLkm-TS, IRBL1-CL, IRBLkh-K3[LT], IRBLk-Ka[LT], IRBLkp-K60, IRBL7-M, IRBL9-W, IRBLz5-CA-1, and IRBLta2-Re) harboring Pi3, Pi5, Pikm, Pi1, Pikh, Pik, Pikp, Pi7, Pi9, Piz5, and Pita2, respectively.

Based on race designation, 807 blast isolates were classified into 551 races (see Additional file 2). In the four ecotypic areas (NM, RRD, NC, and MD) the proportion of unique races increased after 2017 with sufficient sampling (Table 2). The overall Simpson’s Diversity Index and Evenness across these areas were 0.980 and 0.906, respectively, before 2016 and increased to 0.995 and 0.970, respectively, after 2017. Among the four ecotypic areas, the NC area exhibited the highest SI (0.988), the NM area showed the highest EV (0.977) throughout the entire period, and the RRD area demonstrated the lowest SI (0.965) and EV (0.891), indicating limited race diversity.

Of the 551 identified races, 76 (13.8%) were detected at multiple sampling points, 35 of which were found across multiple ecotypic areas. Among the top 10 races with identical pathogenicity, two were identified only in RRD or MD, whereas the remaining eight were distributed across several ecotypic areas. Although one dominant race was identified in the RRD, no clear regional or temporal bias was noted among the others. These results suggest a directional shift in the pathogenicity landscape of the blast population in Vietnam, potentially indicating an adaptation to host resistance genes deployed in the field.

In this study, we analyzed the pathogenicity of 807 rice blast pathogen isolates collected since 2007 using a common differential system. Using this dataset, we clarified the temporal and spatial shifts in pathogenicity. Although we were unable to obtain data from some regions, preventing us from identifying transmission pathways or detailed temporal and spatial patterns of change to fully substantiate the pathogen transition, we were able to identify genes that remain effective for resistance and genes that have lost their resistance over time (Fig. 5). These results are highly valuable for the efficient and effective breeding of blast-resistant rice varieties, contributing to reduced pesticide use and the realization of sustainable agriculture.

Herein, we newly added 474 blast isolates and analyzed their pathogenicity to compare the drastic changes of the pathogenicity between “before 2016” and “after 2017”. In total, the frequency of virulent isolates against each DVs increased, as did the diversity of the blast races. Particularly, the resistance genes Pish and Pita lost efficacy. In the Ha Noi and Ha Nam provinces, the blast isolates collected after 2017 were highly virulent to DVs harboring Pish and Pita. Koizumi (2009) reported that the extensive use of a single resistance gene may result in the emergence of new virulent races. Accordingly, the increased frequency of isolates showing high virulence against Pish and Pita was considered a consequence of breeding programs that incorporated these resistance genes. However, Ngoc et al. (2023) pointed out that, among 209 Vietnamese rice varieties, those carrying Pish or Pita were not widely cultivated. This lack of concordance with the present findings was attributed to the exclusion of recently developed or widely adopted varieties (Ngoc et al., 2023). Thus, investigating the composition of blast resistance genes in the currently prevalent varieties is necessary.

In China, which shares a southern border with northern Vietnam, Pish is widely distributed among Japonica (Geng) varieties in the northern regions, whereas it is rarely used in Indica (Xian) varieties in the south (Xiao et al. 2018). They reported that Pish was mainly utilized to improve Southern Indica varieties. Thus, the frequent use of Pish-carrying varieties in China and the emergence of new isolates virulent to Pish, followed by cross-border movement facilitated by human activity, may have contributed to the increased virulence noted in northern Vietnam.

However, the breakdown of resistance observed in TU8, a cultivar widely cultivated in Ha Tinh province and in other regions of central Vietnam, was found to carry the Pita2 resistance gene. In addition to TU8, recently developed varieties such as KR1 and HN6 also possessed Pita2. Le et al. (2023) reported that in central Vietnam, cultivated rice varieties are less diverse, mostly from the Indica variety type and modern type, since farmers here primarily follow the governmental instructions of the local agricultural development sections. Thus, our findings suggest that continuous cultivation of a limited number of varieties harboring Pita2 results in an increase in the number of isolates that are highly pathogenic to Pita2, leading to the breakdown of resistance.

Herein, since 2017, the population of blast isolates collected in NM has exhibited high racial diversity and was predominantly classified into highly virulent clusters (Table 2 and Figs. 2 and 3a). This observation is consistent with the results of Le et al. (2023) who reported that blast isolates from northern Vietnam possessed high genetic diversity. They further suggested that blast populations in this region maintain their sexual reproductive capacity and demonstrate low linkage disequilibrium, indicating that sexual recombination among blast isolates contributes to the high diversity. Moreover, they pointed out that the highly diversified rice blast population in Vietnam likely reflects the enhanced adaptability of rice blast to heterogeneous environments. These include diverse rice varieties, traditional landraces and modern varieties of Indica and Japonica, as well as traditional farming systems that are commonly practiced in mountainous regions. Taken together, our results suggest that the greater diversity and higher proportion of isolates belonging to highly virulent clusters in northern Vietnam, compared to the southern regions, may be attributed to sexual and asexual recombination of the blast isolates and the diverse agroecosystems and cropping practices that support a wide range of host genotypes.

Kiyosawa (1985) proposed that the establishment of durable resistance to rice blast requires strategies such as the cultivation of multiple lines harboring various resistance genes, gene rotation, and gene pyramiding. Herein, we clarified the temporal and regional dynamics of rice blast pathogenicity in Vietnam and identified resistance genes that remain effective against the current blast isolates (Fig. 5). This information will be valuable for breeding rice varieties with durable resistance. Koizumi (2007) emphasized that the composition of resistance genes in rice varieties influence the pathogenicity of M. oryzae; therefore, characterizing the genotypes of host resistance is essential. Kitazawa et al. (2019) developed DNA markers to detect genetic polymorphisms at ten loci and identified 24 resistance alleles. These DNA markers enable the efficient identification of the resistance gene composition in rice varieties. Additionally, regular monitoring of the blast pathogenicity across regions and over time is necessary. Establishing a systematic surveillance framework and integrating molecular tools, such as DNA markers, into breeding will be helpful for the timely identification of effective resistance genes and the selection of appropriate varieties. These efforts will contribute to the development of rice varieties with durable resistance and improve the long-term management of rice blast diseases.

In this study, we clarified the pathogenicity of rice blast isolates collected across Vietnam using the international differential system to suggest effective breeding strategies for blast-resistant rice varieties. Our results demonstrated that the resistance breakdown observed in TU 8 was caused by the loss of effectiveness of the Pita2 gene. Furthermore, we showed that resistance genes belonging to Type II—Pik, Pikh, Pikm, Piz5, Pi1, and Pi9—remain effective nationwide.

These findings highlight the dynamic nature of blast pathogenicity, which changes over time and across regions. Continuous monitoring using the differential system is essential to monitor these changes and to identify effective resistance genes for sustainable blast management.

However, this study only revealed pathogenicity changes based on phenotypic evaluation and did not fully investigate the underlying molecular genetic mechanisms. Future research should examine whether these pathogenic shifts are driven by genetic changes in the pathogen or by human-related dispersal, as well as assess the selective pressure imposed by resistance genes harbored in rice varieties.

The integrating continuous pathogenicity monitoring with molecular genetic studies will enable a better understanding of the interaction between blast isolates and host resistance genes. This approach will contribute to the development of more durable resistance and the implementation of stable breeding strategies.

Overall, our study demonstrates that the international differential system plays a critical role in guiding effective breeding programs for blast resistance, thereby reducing pesticide use and promoting environmentally friendly and sustainable rice production.

<List of abbreviations>

DVs: Differential varieties

EV: Evenness

IRBL: International rice blast

KD: Khang Dan

LTH: Lijiangxintuanheigu

SI: Simpson’s diversity index

TU8: Thien Uu 8

<Ethics approval and consent to participate>

Not applicable

<Consent for publication>

Not applicable

<Availability of data and materials>

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

<Competing interests>

The authors declare that they have no competing interests

<Funding>

This study was conducted as part of the JIRCAS research project “Green Asia” under “Accelerating the Application of Agricultural Technologies to Enhance Production Potential and Ensure Sustainable Food Systems in the Asia-Monsoon Region” founded by The Ministry of Agriculture, Forestry and Fisheries, Japan (2022 to onward).

This study was also partially funded from MARD, Vietnam for the sampling, isolation, and race designation of the blast isolates collected in the provinces of Vietnam under project "Introgression multi resistance genes to bacterial leaf bight, brown planthopper and blast disease into high quality rice variety by using molecular marker (MABC)" (2018–2020)

<Authors’ contributions>

TMN, NK, MO and HS conceptualized and designed the research; BN, TMN, TN, TNL, THH, HLL and TTT collected blast samples and performed the experiments; BN, TMN, NK, MO and HS analyzed the data and created the figures and tables; TMN and HS wrote the initial manuscript draft. All authors read, revised, and approved the manuscript.

<Acknowledgments>

We sincerely thank Dr. Yoshimichi FUKUTA at the University of Ryukyus for their valuable contributions to pathogenicity evaluation.