Virulence Phenotypes of Siberian Wheat Stem Rust Population in 2017–2018

Management of wheat stem rust in Western Siberia has gained importance since the first outbreaks in 2007–2010 and 2016. The race composition and virulence patterns were investigated for the enlarged Puccinia graminis f. sp. tritici (Pgt) samples collected in three neighboring regions Omsk, Novosibirsk, and Altai during 2017–2018. Most of Pgt isolates were identified as virulent to wheat lines with genes Sr5, Sr9a, Sr10, Sr38, SrMcN, and avirulent to Sr24, Sr31. Differentiation ability of genes Sr6, Sr7b, Sr8a, Sr9b, Sr9d, Sr9g, Sr9e, Sr11, Sr17, Sr21, Sr30, Sr36, and SrTmp to distinguish between the regional populations was established. A total of 33 virulence phenotypes or races were detected among 115 Pgt isolates tested. Based on virulence phenotypes, two different Pgt subpopulations were identified in the Altai and Omsk regions likely originating from asexual and sexual cycles, respectively. The Novosibirsk pathogen population seems to be a mixture of isolates originated from both neighboring regions with virulence phenotypes that developed in the west, Omsk (TKRPF, QHHSF, and MLLTF), and in the south, Altai (NFMSF, LKCSF, LKMSF, and PKCSF), of Western Siberia.


INTRODUCTION
Western Siberia is an extremely important wheat production area in Russia with ∼7 million ha of spring wheat, along with increasing areas of winter wheat and some durum wheat. The genetic basis of resistance to stem rust in Siberian germplasm is narrow and limited to several Sr genes, including Sr25, Sr31, Sr36, Sr6Ai, Sr6Ai#2, and some unknown major genes (Shamanin et al., 2016). The devastating disease caused by Puccinia graminis f. sp. tritici (Pgt) was not considered economically important in Western Siberia until the first reports of stem rust outbreak in 2007-2010 (Shamanin et al., 2015;Sochalova and Lihenko, 2015). In the Omsk region, southwest of Western Siberia, wheat production was affected by heavy epidemic in 2015; yield loss was estimated at more than 2 million tons. Understanding population structure, virulence diversity, and distribution of Pgt in the Omsk and neighboring regions has been essential for effective disease management based on resistance genes. Integrated analysis of pathogen virulence in Western Siberia has just been launched. In 2017, field trip for assessing the stem rust situation was conducted, and the samples of stem rust were collected for determining virulence phenotypes at the Global Rust Reference Center (GRRC) (GRRC Report: Samples of Stem Rust Infected Wheat From Russia, 2017). Significant race diversity coupled with high virulence of local pathogen was revealed in the samples from Omsk and Altai (Shamanin et al., 2020). Further study of virulence dynamics in the vast territory of Western Siberia was highlighted as the main result of international screening (Hodson et al., 2017). Moreover, the global tracking of aggressive races Ug99 and its variants spreading across the continents makes the race surveys highly actual at every location of wheat crops (Pretorius et al., 2000;Olivera et al., 2015;Bhattacharya, 2017). The objective of our study was to investigate the race composition and virulence patterns of the enlarged Pgt samples collected in three neighboring regions Omsk, Novosibirsk, and Altai during 2017-2018.

MATERIALS AND METHODS
In 2017-2018, 77 infected stems of susceptible cultivars of bread wheat (cvs. Chernyava 13, Khakasskaya, Duet) with urediniospores of Pgt were sampled in the neighboring regions Omsk, Novosibirsk, and Altai. Additionally, 30 samples of infected leaves with aeciospores of Pgt, were collected from barberry, the alternative host of stem rust, in Omsk 2018 (Figure 1). Bulk samples were stored in the automatic desiccator at dry cold conditions. After recovering on susceptible wheat cultivars Khakasskaya and Morocco, a total of 115 single pustule Pgt isolates were obtained from 107 bulk field samples. In 2017, 51 single pustule isolates were obtained from 39 samples and tested at the GRRC in Denmark. The rest 64 single pustule isolates were obtained from 68 samples and analyzed at the Institute of Cytology and Genetics.
Single pustule isolates were used for determining virulence phenotypes based on a set of 20 North American wheat differential lines, containing Sr genes: McNair 701). According to the international protocols of virulence analysis, seedling plants were used (Roelfs et al., 1992). Each of 115 single pustule isolates was tested at least twice on the set of differentials with the susceptible checks (cultivars Khakasskaya and Morocco). The spores were dissolved in Novec 7100 oil and applied with an airbrush on seedlings planted 5 to 10 per pot. Inoculated plants were placed in a dark dew chamber at 18 • C ± 2 • C for 24 h and then transferred to a growing room at 20 • C ± 2 • C with a photoperiod of 16 h. The entire set of tested plants was covered with a cellophane box to avoid contamination (Figure 2). Infection types as described by Stakman et al. (1962) were assessed 14 to 16 days after inoculation; infection types 0 to 2 and 3 to 4 were interpreted as low (avirulent) and high (virulent), respectively. Races were designated following the North American (hexadecimal) Pgt coding system (Roelfs and Martens, 1988;Jin et al., 2008). A total of 33 virulence phenotypes or races were detected among 115 Pgt isolates tested.
FIGURE 2 | Procedures for laboratory testing of spore material, facility of the Institute of Cytology and Genetics (IC&G). (A) Spore recovering; (B) spore multiplication on susceptible cultivar; (C,D) spore collection with the vacuum pump; (E) spore inoculation with the airbrush; (F) infected wheat seedlings of the differential set covered with cellophane isolating box.

Data Analysis
Virulence frequencies were calculated for all regional populations for each separate year and for the pool of 2017 and 2018. The relationships of the P. graminis f. sp. tritici populations were analyzed using the assignment-based approaches. The corresponding Kosman dispersion (KW) within and distance (KB) between populations with regard to the simple mismatch dissimilarity between virulence patterns were calculated (Kosman, 1996(Kosman, , 2014Kosman and Leonard, 2007). Differentiation among populations was estimated with the permutation test (1,000 random partitions) for differentiation statistic dif KW (Equation 1 in Gultyaeva et al., 2020) similar to those of Jost (2008;Equations 9, 11), and in accordance with explanations in Kosman (2014, p. 565). The corresponding calculations were performed with Virulence Analysis Tool (VAT) software (Kosman et al., 2008;Schachtel et al., 2012) and its extension. The structural diversity of a population (effective number of equally distant isolates) was measured by the Hill number of order 1, 1 D (T), based on the matrix of simple mismatch dissimilarities between isolates (Equation 4 in Scheiner et al., 2017). Then the structural variation within a given population (effective number of different isolates) is a function of both the dispersion KW and structural diversity 1 D(T), where k is the number of isolates in the population: This indicator expresses an extent of variability within the pathogen populations. NMDS plots with regard to the Kosman distance (KB) between the regional collections of P. graminis f. sp. tritici in 2017 and 2018 were derived using the Mdscale program of the NTSYSpc package, version 2.2 (Exeter Software, Setauket, NY, USA). A UPGMA dendrogram based on the simple mismatch dissimilarity between virulence phenotypes detected in each separate population was derived using the SAHN program of the NTSYSPC package, v. 2.1 (Exeter Software).

RESULTS
Wheat crop occupies the territory of more than 1.4 million ha in the Omsk region, more than 1 million ha in Novosibirsk region, and ∼2.1 million ha in the Altai. Stem rust was observed as a late infection. When sampled, wheat crop was close to maturity or mature. Both years, 2017 and 2018, were non-epidemic for stem rust in Western Siberia. Among the inspected field plots and commercial fields, the disease incidence ranged between 0 and 45% with severity up to 60S. The Novosibirsk region is the central part of Western Siberia, bordered by the Omsk region in the west and the Altai region in the south.

Variability Within and Among Samples
Variability within the geographical samples of Pgt changed differently from 2017 to 2018 according to the KW dispersion and the normalized number of effectively different isolates 1 nD(T, KW) ( Table 3). According to both the estimates, in 2018 variability was approximately 40% larger within the Omsk samples, 40% smaller within the Altai samples, and nearly the same in Novosibirsk. Differences in variability were remarkable between the regions with highest values of KW dispersion in Novosibirsk (0.316 in 2017 and 0.356 in 2018). In the Altai samples, variability was almost two times smaller than in other regions (0.155 in 2017 and 0.107 in 2018). Similar results were established with the normalized effective number of different isolates 1 nD(T, KW).
Relationships between the regional samples were estimated with the KB distances for the virulence profiles of Pgt isolates and only for phenotypes without their abundances; it is represented by the corresponding NMDS plots (Figure 5). Clear separation between the Altai and Omsk samples was established, whereas the Novosibirsk samples were in between, but more similar to the Altai ones. These findings were also confirmed by the permutation test for differentiation statistics dif KW . Significant differentiation was ascertained between the Omsk and Novosibirsk samples (at p ≥ 0.95) and the Omsk and Altai samples (at p ≥ 0.99) in both years. Hypothesis of differentiation between the Novosibirsk and Altai samples was rejected.   (Kosman and Leonard, 2007;Kosman, 2014

DISCUSSION
Management of wheat stem rust can be more effective if an origin of initial infection is known. To ascertain whether inoculum of Pgt in a specific region is endemic or wind disseminated from neighboring areas, monitoring of the pathogen in Western Siberia was performed in 2017-2018. Based on virulence phenotypes, two absolutely different Pgt subpopulations were discovered in the Altai and Omsk regions. The Novosibirsk pathogen population seems to be a mixture of isolates originated from both the neighboring regions virulence phenotypes that arose in the west (TKRPF, QHHSF, and MLLTF) or in the south (NFMSF, LKCSF, LKMSF, and PKCSF) of Western Siberia. It appears that sexual reproduction to large extent shapes the structure of the Pgt population in Omsk region. Indeed, barberry plants are common in this area, and they are susceptible to stem rust (Shamanin et al., 2015). High variability is indirect evidence of existing sexual stage in the pathogen development.
Race composition was characterized by single predominant TKRPF race along with significant number of rare races not detected in recent years elsewhere in Asia and Africa. The group of isolates with similar M-races was sampled from barberry and wheat in Omsk in 2018. Because new races may appear after sexual recombination on alternate host (Jin, 2011), the virulence composition and phenotype structure of Pgt population are highly dynamic. In North America, stem rust races 56, 15B, and QCC originated from barberry were responsible for the severe epidemics in the mid-1930s, mid-1950s, and between 1989and 1993(Stakman and Rodenhiser, 1958Martens et al., 1989). Races of Pgt with a rich virulence spectrum have been isolated from barberry plants in the central region of Russia from 2000 to 2009 (Skolotneva et al., 2013).
Susceptible barberry plants have not been found in the Altai region yet. Variability within the Altai Pgt samples was much lower than in the Omsk region (e.g., 0.107-0.155 vs. 0.232-0.344 for the KW dispersion, Table 3) and similar to P. triticina asexual population with KW values ranging between 0.09 and 0.11 (Gultyaeva et al., 2020). The Altai samples consisted of several closely related rare virulence phenotypes. Therefore, the clonal structure of Pgt population in the Altai region can be assumed with a high confidence. Because of the mild short winter, the urediniospores are able to survive on winter wheat, whose area has increased over the past decade in the region. The new lineages in asexual fungal populations could be rarely generated by somatic hybridization, which has been widely discussed as a source of low genetic diversity (Johnson and Newton, 1946;Watson and Luig, 1958;Cotter and Roberts, 1963;Luig and Watson, 1972). Emergence through somatic hybridization was proven for sadly famous Ug99 lineage of races (Li et al., 2019). For the Altai Pgt subpopulation, low virulence diversity might be provided by somatic hybridization.
Aecia observed on barberry in the Novosibirsk region explained that the rye special form, P. graminis f. sp. secalis, segregated predominantly among the sexual offspring (Peresypkin, 1979;Skolotneva and Salina, 2016). Field and laboratory virulence tests have shown that barberry is not a source of infection for wheat crops in the Novosibirsk region (Kelbin et al., 2019). Thus, wheat stem rust in the Novosibirsk region is not endemic but carried by wind from both neighboring regions. This makes the virulence pattern of local samples most variable (e.g., 0.316-0.356 for KW, Table 3) and dependent on breeding events in the neighboring regions. Although Novosibirsk Pgt population seems to be a mixture of the Omsk and Altai subpopulations, it was much more similar to the Altai one, which primarily reproduced asexually in the period of our survey.
Our results clearly demonstrate significant differentiation among the regional Pgt populations in Western Siberia. The main driving force of the population diversity and geographic differentiation seems to be sexual reproduction of the pathogen fungi in the Omsk region (northwest of the area), whereas only clonal propagation with no evidence of sex is assumed in the Altai region (southeast of the area). These two welldistinguished poles, both geographically and by reproductive ability of Pgt, predetermined extreme boundaries of composition of the pathogen populations with a gradual mixing of Pgt races along the gradient from southeast to northwest of Western Siberia.

DATA AVAILABILITY STATEMENT
The data will be made available by request to the corresponding author.

AUTHOR CONTRIBUTIONS
ESS conceived the research idea and was responsible for sampling the Pgt collections in Novosibirsk and Altai, and acquisition and interpretation of virulence data. EK performed data analysis and interpretation of results. MP performed the virulence analysis of Siberian samples in GRRC. VK contributed to acquisition of virulence data. AM was responsible for reviewing the current state of stem rust worldwide and organized the international collaboration between IC&G and GRRC. VS sampled the Pgt collections in Omsk. EAS was responsible for general design of the work. ESS and EK prepared the first draft, while all authors contributed to shaping the final version of the manuscript.