Genotyping reveals high genotypic diversity and potential migration pattern of Puccinia striiformis f. sp. tritici populations in Xinjiang and Northwest epidemic regions of China

Abstract

The Xinjiang epidemiological region of wheat stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), exhibits distinct characteristics compared to other epidemic regions in China. This region provides favorable conditions for Pst to complete its life cycle and serves as a connection to other major wheat-growing areas across the country. However, despite limited studies on Pst populations in Xinjiang, a comprehensive understanding of the epidemiology of wheat stripe rust in this region remains unclear. This gap complicates the effective management of the disease due to uncertainties regarding inoculum sources and migration routes. In this study, we conducted a comparative population genetic analysis of Pst populations within Xinjiang (North, Midwest, and East Xinjiang) and neighboring provinces (Gansu, Shaanxi, and Ningxia) to trace inoculum origins and confirm migration patterns. For this purpose, a total of 232 samples were collected and genotyped using 17 simple sequence repeats markers. Our findings revealed significant gene flow within the Midwestern and Eastern Xinjiang regions, as well as among the neighboring provinces of Gansu, Qinghai, Ningxia, and Shaanxi. The Midwestern Xinjiang, particularly its western subregion, is likely to be an inoculum source. Additionally, we observed limited gene flow and shared multilocus genotypes between Xinjiang and its neighboring provinces. The overall high genotypic diversity observed in Northwestern China (Simpson’s index = 0.98) indicates a hotspot for the emergence of new Pst races through sexual recombination. This is likely driven by the widespread distribution of the alternate host, barberry (Berberis spp.), which facilitates the sexual reproduction of Pst in certain regions. These results provide valuable insights into the gene flow dynamics of Pst populations in Xinjiang and its neighboring provinces. Furthermore, this study underscores the importance of deploying genetically resistant wheat cultivars to effectively control wheat stripe rust in these regions.

Background

Wheat stripe (yellow) rust caused by Puccinia striiformis f. sp. tritici (Pst), a heteroecious rust fungus in the phylum of Basidiomycota, is one of the most destructive fungal diseases of wheat in China and many regions of the world (Li and Zeng 2002; Wan et al. 2007; Wellings 2011). Historically, the disease has led to considerable yield losses in China (Wang 1951; Liu and Meng 1957; Li 1980; Wan et al. 2007; Ma 2018; Zhao and Kang 2023) and some countries (Zadoks 1961; Rapilly 1979; Brown and HovmØller 2002; Chen 2005; Wellings 2011).

Xinjiang is a winter wheat, spring wheat, and mixed winter and spring wheat growing region, in which the annual growing area of wheat is estimated to be 600,000–750,000 hectares. Wheat stripe rust occurs mostly in the mid-western and northern regions of Xinjiang (Li et al. 2010). The “boom and bust” of the disease was recorded in the 1950–1960s (Zhang 1962), 1970–1980s (Qiu and Jia 1982), and 1990–2000s (Li et al. 2010). Among these epidemic years, the most serious epidemic of wheat stripe rust occurred in 2007 and historically reached up to a peak with 159,000 hectares. In the last 10 years, the disease has significantly increased in Xinjiang, especially in Yili in western Xinjiang (Zeng et al. 2024).

In China, due to diverse agricultural cropping systems, complicated geographic conditions, multiple climates, as well as disease epidemic features, wheat stripe rust has a unique system consisting of different epidemiological regions (Zeng and Luo 2006), among which Xinjiang is one of the seven sub-divided epidemiological regions in the Northwest epidemiological region, being distinguished from those of the Chinese stripe rust epidemiological system (Li and Zeng 2002). In this region, the disease is of its specificity and is completed all year round by local inoculum sources. Therefore, inoculum in Xinjiang plays a limited role in causing the occurrence of wheat stripe rust in other epidemiological regions of the country (Li et al. 2010).

Genetic diversity and gene flow of different Pst populations in China have been well studied. However, in addition to some studies in relation to race identification of the pathogen populations (Qiu and Jia 1982; Liu et al. 2012, 2022b; Ma et al. 2023; Zhang et al. 2024), only Pst populations from local areas of Xinjiang or insufficient isolates have been studies in genetic diversity using different molecular marker methods (Wan et al. 2015; Zhan et al. 2016; Awais et al. 2022; Zeng et al. 2024; Zhang et al. 2024). So far, little is known about the genetic diversity of the Pst population and the pathogen migration among different populations within Xinjiang Province and with neighboring provinces, resulting in the lack of understanding of the population genetic diversity level and the spread of the pathogen in this region.

The aims of the present study were to compare the population structure of Pst in the Xinjiang epidemic region with those of the neighboring Northwest epidemic regions (Gansu, Qinghai, Ningxia, and Shaanxi) and to assess the level of genetic diversity and migration based on the genetic distance of the pathogen among these epidemic areas. The result is a crucial step for implementing effective disease management strategies in Xinjiang.

Results

Creditability of SSR markers

The 17 SSR primers, screened by Ali et al. (2014) and Awais et al. (2022), had been used to study the genetics of Pst populations. In the present study, except NRJN 12, the remaining 16 of the 17 SSR primers demonstrated polymorphism when examining overall 232 samples of Pst (Additional file 1: Table S1).

Recombination rate and genetic diversity

Based on data analyses, in Xinjiang, the level of gene diversity (He) was comparatively lower than in neighboring provinces such as Gansu (0.42), Qinghai (0.40), Ningxia (0.40), and Shaanxi (0.43) (Fig. 1). However, the overall high genotypic diversity (Simpson’s index) was observed in all Northwest epidemic regions of China (overall diversity = 0.98; Table 1).

Fig. 1

Heterozygosity of Puccinia striiformis f. sp. tritici surveillance from Northwest epidemic regions of China during the cropping season 2021. He: Expected heterozygosity. Ho: Observed heterozygosity. North: Northwest Xinjiang, Midwest: Midwest Xinjiang, East: East Xinjiang, SX: Shaanxi, QH: Qinghai, NX: Ningxia, and GS: Gansu

Table 1 Genetic Diversity of Puccinia striiformis f. sp. tritici isolates within the Northwest epidemic region of China during cropping season, 2021

Our findings revealed non-significant differences between homozygosity and heterozygosity within the epidemic areas (Fig. 1), which indicated the presence of a recombinant population in the region. It was further confirmed by observing the rBarD and Index of Association values. Our results showed lower rBarD values of 0.09 and 0.13, as well as lower values of the Index of Association (1.20 and 1.97) with a lower maximum allele frequency (3) in the Qinghai and Gansu epidemic regions, confirming the presence of a higher recombinant population. In the Midwest epidemic region of Xinjiang, a high value of the Index of Association (4.16) with a high maximum allele frequency (12) was observed, indicating the potential existence of some clonal lineages (Table 1).

Genetic structure and Pst population subdivisions

Various population genetic groups were evaluated through STRUCTURE analysis programs. The STRUCTURE program was employed to analyze clustered K values ranging from K = 2 to K = 10. The optimal K value was determined using Structure Harvester, and a delta K was observed at K = 5 (Fig. 2a). This indicates five cluster groups in the overall population.

Fig. 2

Puccinia striiformis f. sp. tritici population structure between different epidemic regions of Northwest China during the cropping season 2021. a Best K value. b STRUCTURE analysis result (K2 – K5). c Distribution of different genetic groups on a spatial scale. G1–G5 indicated genetic group 1, 2, 3, 4, and 5

At K = 5, our analysis revealed five distinct genetic groups and one admixed population group (Fig. 2b, c). The G1 group exhibited dominance in the Midwest Xinjiang region, spreading towards East and North Xinjiang. It then shared with other provinces, including Gansu, Qinghai, Ningxia, and Shaanxi, suggesting a migration pattern between Northwest epidemic areas of different provinces. The G2 group was predominantly found in the Midwest epidemic region of Xinjiang, later extending towards East and North Xinjiang. However, this group did not exhibit long-distance migration towards other provinces. The G3 group displayed dominance in North Xinjiang and cloned with the Midwest Xinjiang. In contrast, the G4 group demonstrated a long-distance migration behavior, being present in all epidemic areas except Eastern Xinjiang. Interestingly, the G5 group was absent only in the Ningxia province. Our analysis identified an admixed population group in various epidemic areas.

Genetic divergence between different epidemic regions

Genetic divergence among epidemic regions was measured using F_ST value (Table 2), Nei’s genetic distance (Table 2), and a phylogenetic tree (Fig. 3). In the Xinjiang epidemic zone, the least divergence was noted between the Midwest and East Xinjiang regions, reflected by a low F_ST (0.02). Conversely, the highest F_ST (0.10) was observed between North Xinjiang and East Xinjiang, despite their relatively closer geographical proximity compared to the Midwest and East Xinjiang areas.

Table 2 Upper diagonals showed F_ST values and the lower diagonal showed Nei’s genetic distance of Puccinia striiformis f. sp. tritici within the Northwest epidemic region during the cropping season 2021

Fig. 3

Phylogenetic tree based on Nei’s genetic distance of Puccnia striiformis f. sp. tritici populations within the Northwest epidemic region during the cropping season 2021. North: Northwest Xinjiang, Midwest: Midwest Xinjiang, East: East Xinjiang, SX: Shaanxi, QH: Qinghai, NX: Ningxia, and GS: Gansu

When examining the F_ST values of various Xinjiang epidemic regions in comparison to neighboring provinces, the lowest F_ST value was noted between Midwest Xinjiang and Shaanxi Province. The highest F_ST value was observed between East Xinjiang and the Qinghai epidemic region. Additionally, the lowest F_ST values were identified among Gansu, Ningxia, Qinghai, and Shaanxi, indicating a close population relationship.

Shared multilocus genotypes across different locations

Out of a total of 232 samples of Pst, 159 multilocus genotypes (MLGs)were identified (Fig. 4 Additional file 2: Table S2). The findings indicate that MLG-138 was prevalent in the Midwest Xinjiang region and was also cloned with the Eastern Xinjiang population. Additionally, MLG125 and MLG122 were cloned between the Midwest Xinjiang and East Xinjiang populations, suggesting a significant migration pattern between these two epidemic regions in Xinjiang. Interestingly, MLG-19 was shared among the East, Midwest, and North Xinjiang populations, and it was also found in Ningxia Province. On the other hand, MLG-61 exhibited a long-distance migration pattern, being cloned with populations from North Xinjiang, Gansu, Ningxia, and Qinghai epidemic areas. The overall results of MLG analysis highlight the potential migration patterns between Xinjiang and its neighboring provinces' epidemic regions (Fig. 4).

Fig. 4

Distribution of different multi-locus-genotypes (MLGs) of Puccinia striiformis f. sp. tritici within the Northwest epidemic region of China. a Density of MLGs in different regions. b Spatial distribution of MLGs. North: Northwest Xinjiang, Midwest: Midwest Xinjiang, East: East Xinjiang, SX: Shaanxi, QH: Qinghai, NX: Ningxia, and GS: Gansu

Migration pattern

Relative migration dynamics were assessed through the analysis of Nm and G_ST values (Fig. 5), it illustrates the gene flow patterns across different epidemic regions. The strength of gene flow was visually represented by the intensity of line colors. Based on Nm and G_ST analyses, it is evident that high gene flow occurred within the Midwest and Eastern regions of Xinjiang. The dominant MLG (MlG138) was prevalent in both the Midwest and Eastern regions, as indicated by the lowest F_ST value (0.02) between the two regions, also demonstrating a high migration pattern from the Midwest to the East. Limited gene flow was observed between Xinjiang and neighboring provinces. However, the maximum geneflow was observed between Qinghai and Ningxia (F_ST = 0.016), Qinghai and Shaanxi (F_ST = 0.074), and Ningxia and Shaanxi (F_ST = 0.038; Table 2).

Fig. 5

The relative migration network between Puccinia striiformis f. sp. tritici populations in Xinjiang and neighboring provinces sampling during crop season 2021. a Nm value migration network. b G_ST value migration network. The strength of gene flow was visually represented by the intensity of line colors

Discussion

Wheat stripe rust is air-borne and pathogen inoculums capable of expanding the disease epidemics at a large scale. Therefore, it is necessary to understand the origin of primary inoculum and epidemiological characteristics for integrated management of wheat stripe rust. Although some previous studies were conducted related to genetic diversity and molecular variation of Pst populations in Xinjiang (Wan et al. 2015; Zhan et al. 2016; Zeng et al. 2024; Zhang et al. 2024), little information is associated with the migration of the pathogen is available in this region. In the present study, we determined the migration of Pst in main wheat stripe rust-occurring regions in Xinjiang and the relationship of the pathogen populations between Xinjiang and neighboring provinces.

Genetic resistance of wheat cultivars is associated with the level of genetic diversity of Pst population. The interaction of wheat cultivars and Pst is in accord with the gene-for-gene rule. However, the resistance of most wheat cultivars to Pst is race-specific (Singh et al. 1990). Therefore, wheat cultivars carrying genetic resistance genes (Yr) influence race types with avirulence gene (Avr) of Pst population (Zhang et al. 2023). Based on molecular detection of Yr genes in wheat landraces/cultivars grown in Xinjiang with Yr9, Yr26 (= Yr24), Yr10, and YrTp1 markers, most of wheat landraces/cultivars possess these four resistance genes. Except for YrTp1, which showed resistance to CYR32, the other three Yr genes are susceptible to dominant races of the Pst population in this region (Bai et al. 2017; Lu et al. 2022). Over the past ten years, CYR32 races have been predominant in Xinjiang Pst populations (Li et al. 2010; Zhan et al. 2016; Ma et al. 2023). In recent years, in addition to Yr9-virulent races, races with virulence to Yr26, with high occurrence frequency, have been identified in Pst population in Xinjiang (Liu et al. 2022a, 2022b; Ma et al. 2023), which could be responsible for the level of genetic diversity of Pst population and increasing wheat stripe rust epidemics in this region. Historically, wheat stripe rust epidemics in Xinjiang have been attributed to monogenic wheat cultivars on a large scale. It was exemplified that the 1960s epidemic resulted from the breakdown of the resistance of wheat cultivar Bima 1, which was cultivated extensively in the 1950s to stripe rust (Li et al. 2010). Another case was that wheat cultivar Tangshan 6898, with tolerance to stripe rust as grown in 1982, became highly susceptible from the end of the 1980s to the early 1990s in Xinjiang, which resulted in outbreaks of the disease in Akesu (or Aksu) in 1990, and Aletai (or Altai) in 1993. Since the 1990s, wheat cultivars, including the Yinong series, Changdong series, and Kuidong series, were mainly cultivated in North Xinjiang and lost their resistance to stripe rust due to the emergence and rapid accumulation of CYR31 and CYR32 races that subsequently developed to be predominant, which accounted for the recurrence of wheat stripe rust in this region (Li et al. 2010). Therefore, frequent replacement of wheat varieties carrying Yr-resistant genes to avoid the breakdown of the wheat varieties is needed to carry out in Xinjiang. Moreover, it is essential for deploying wheat varieties with different resistance genes in different epidemic regions of wheat stripe rust in this region.

The result of the present study showed high genotypic diversity in the Pst population in Xinjiang and those of neighboring provinces in Northwest China, which is somewhat higher than that in these regions as reported data (Zhan et al. 2016; Awais et al. 2022). Xinjiang population was mainly isolated. Only a few genotypes of Xinjiang are shared among neighboring epidemic regions, which shows the potential for a limited migration pattern between Xinjiang and other neighboring epidemic regions (Gansu, Qinghai, and Ningxia). The previous study, which used amplified fragment length polymorphism (AFLP) analysis, found that Pst populations in Xinjiang and Qinghai frequently gene exchanged. It was also observed that the Xinjiang population was genetically closer to the Qinghai and Gansu. Therefore, it was suggested that the pathogen genotypes in Xinjiang could not be completely isolated from those in Gansu (Wan et al. 2015). In our present study, we also observed weak gene flow among Xinjiang and neighboring provinces, including Qinghai, which aligns with the findings of Wan et al. (2015). On the other hand, no gene flow was detected in populations from Xinjiang and neighboring regions, supporting that the Xinjiang epidemic region is relatively independent (Zhang et al. 2024), as reported previously (Li and Zeng 2002), which could be related to few isolates and monotonicity of overall isolates from one region of Xinjiang. Therefore, the total number and sampling geographic range of specimens may affect the analytic result. In the case of making a comparison within the Xinjiang population, we observed that there were two main genetic groups: one dominant in the Midwest and one in the East Xinjiang region. In contrast, the other group is dominant in North Xinjiang. Therefore, it is essential to study comparative population studies of Xinjiang with neighboring Central Asian countries to detect the origin of these lineage groups of Pst. This is an alarming situation as it borders with Central Asian countries. It could spread new races to wheat production regions of China (Ali et al. 2014; Liang et al. 2021).

Based on molecular markers, studies revealed that the Pst population of Yili in West Xinjiang had a high level of genetic diversity (Zhan et al. 2022; Zeng et al. 2024). Meanwhile, whole genome sequencing analyses noted that the diversity or heterozygosity of Pst isolates from Yili was lower than that of neighboring provinces in the Northwest epidemic and other main epidemic regions (Li et al. 2023). Therefore, whether the level of genetic diversity of different Pst populations could be influenced by specimens collected in different sampling sites, or sampling years, or by different methods, which is needed to further investigate.

Sexual recombination is responsible for high genetic diversity and heterozygosity of Pst. Over the past years, our field investigations and laboratory experimental identification found that at least three barberry (Berberis) species (B. heteropoda, B. nummularia, and B. kaschgarica) growing in the Midwest and North Xinjiang can serve as alternate (aecial) host for Pst, and that patches of the susceptible barberry bushes are generally observed (Zhuang et al. 2019). In spring, aecial production of the rusts on leaves of different barberry species is quite common. However, to date, Pst infection on susceptible barberry plants under natural conditions has not been detected in Xinjiang. Therefore, it is necessary to determine the potential role of the sexual cycle that results in genetic diversity and population structure. This highlighted that this region is potentially favorable for the emergence of new Pst races and their nourishing year-round.

Conclusions

In this study, our results demonstrated high genotypic diversity in the Northwest Regions (Xinjiang, Qinghai, Gansu, Ningxia, and Shaanxi). The overall population consists of five diverse genetic groups. Results showed a maximum spore migration pattern between the Midwest and East Xinjiang regions and between the Qinghai and Gansu regions. Also, limited gene flow was observed between Xinjiang and its neighboring provinces. A shared multilocus genotype was also identified between Xinjiang and neighboring provinces, highlighting potential hotspots for spreading Pst races between China's major wheat-cultivating region. The result will be helpful in deploying effective disease control strategies in the country.

Methods

Sample collections

Field surveillance of stripe rust on wheat was carried out in different areas of Xinjiang (North, Midwest, and East) and neighboring provinces, including Gansu, Qinghai, Ningxia, and Shaanxi in the 2021 cropping season (Fig. 6). A total of 232 samples were collected from the sampling sites as above. A leaf sample with a single stripe rust lesion (5–10 cm long) was collected into an individual sampling paper bag and subjected to room temperature for one day until completely drying.

Fig. 6

Sampling sites for Puccinia striiformis f. sp. tritici surveillance from the Northwest epidemic region of China during the cropping season 2021. Note: basemap is taken from http;//bzdt.ch.mnr.gov.cn, and the study area map generated through Qgis (3.34) using standard reference system EPSG 4326, WGA, 84

Genomic DNA extraction and SSR-based sequencing

After removing extra healthy leaf tissues around the single lesion, the leaf segment with the lesion was put into a clean 2 mL with three steel beads (3 cm in diameter). A set of 48 leaf samples was placed in a 48-well adapter and then crushed to fine powder using a tissue grinder (TISS-24/48, Shanghai JingXin Industrial Development Co., Ltd., Shanghai, China). DNA extraction was performed using the CTAB method (Chen et al. 1993). The DNA pellet was dried entirely inside a clean bench and delivered in a cooling box for sequencing. The set of 17 distinct, simple sequence repeats (SSR) primers were used following the protocol outlined by Ali et al. (2014). The PCR products were sequenced by Sangon Biotech (Shanghai) Co., Ltd., China.

Population genetic analysis

The raw sequencing data of overall Pst isolates was subjected to further analysis using multiple population genetic software. This analysis aimed to measure genotypic diversity and understand migration patterns and population subdivisions within Xinjiang and Northwestern China. We used the Genetix V4.05 software to analyze various population parameters. This software is capable of calculating Nei's distance and H (Nei 1972), Wright’s F-statistics (using both the Weir-Cockerham and Robertson-Hill estimators), and linkage disequilibrium according to Black and Krafsur (1985). Initially, we assessed polymorphism among our marker set based on the number of alleles detected in each locus. We then used Genetix V4.05 (Belkhir et al. 2004) software to estimate expected heterozygosity, observed heterozygosity, and their associated P values among different populations, which allowed us to evaluate clonal populations. Additionally, we estimated the genetic distance among populations from different epidemic regions using Nei's D and H and Wright's F-statistics, which allowed us to understand potential gene flow among different regions. The population structure of all Pst samples was examined through the Cluster-based analysis program STRUCTURE 2.3.4 (Pritchard et al. 2000), which conducted a Markov Chain Monte Carlo simulation within a Bayesian framework. The assignment of the Pst samples to different cluster groups was assessed across various levels of clustering (K = 1 to K = 10), with 10 independent runs, 100,000 iterations per run, and a burn-in period of 100,000. CLUMPAK was utilized to analyze multiple runs for each fixed K value and distinguish between substantially different clustering solutions (Kopelman et al. 2015). The best K value was obtained by STRUCTURE HARVESTER (http://taylor0.biology. ucla.edu/structureHarvester) and evaluated as ΔK as proposed by Evanno et al. (2005). A bar plot was generated by Distruct using the output of CLUMPAK. The Multilocus genotypes that are shared among different populations were identified using the Geneclone software, based on shared multilocus genotypes, we can estimate the gene flow and reproductive mode, either clonal or recombinant. Diversity-related (Number of MLGs, allele frequency, and genotypic diversity) and linkage disequilibrium parameters (IndAssoc and rBarD) were estimated using multilocus software (Agapow and Burt 2001). We assumed that if recombination occurred, the alleles at different loci would tend to segregate independently, leading to a low rBarD value. Therefore, we concluded that if a population has a lower rBarD value, it may have more recombination. Conversely, if the population has a higher rBarD value, then recombination is infrequent or absent, and the population is likely to be clonal. Geneflow (Nm and G_ST) among epidemic regions were measured through the R package Div R. The phylogenetic tree data was generated using population software (Langella 2008), based on Nei’s genetic distance approach, and the tree was visually created using MEGA 7 software. The phylogenetic tree further confirmed the results of population structure and migration patterns among different epidemic regions.