Direct Evidence Demonstrates that Puccinia striiformis f. sp. tritici Infects Susceptible Barberry to Complete Sexual Cycle in Autumn

Wheat stripe rust, caused by Puccinia striiformis Westend. f. sp. tritici Eriks. (Pst), is one of the most destructive foliar diseases of wheat in many countries of the world, especially in cool and humid areas (Chen 2005; Li and Zeng 2002; Stubbs 1985; Wellings 2011). The disease can generally cause a yield reduction of 10 to 30% in epidemic years, and some crops failed to yield at all in case of early infection on susceptible cultivars (Chen 2005; Wellings 2011). In China, wheat stripe rust is the most threatening disease on wheat and frequently occurred with an annual infected area of approximately 400 ha (Chen et al. 2009a; Li and Zeng 2002; Wan et al. 2007). Since the 1950s, five severe wheat stripe rust epidemics have been reported, causing a significant yield loss of at least 1 million metric tons each time (Li and Zeng 2002; Ma 2018; Wan et al. 2007).

Pst is a heteroecious and macrocyclic rust fungus, requiring five types of spores on two phylogenetically unrelated hosts to complete the whole life cycle. The rust completes uredinial and telial stages on wheat and grasses as primary (telial) hosts and pycnial and aecial stages on Berberis (mainly) or Mahonia as alternate (aecial) hosts after penetration by basidiospores that generated from teliospores (Cheng et al. 2022; Jin et al. 2010; Wang and Chen 2013; Zhao et al. 2013). Since the discovery of barberry (Berberis spp.) as alternate hosts for Pst in 2010 (Jin et al. 2010), more than 40 Berberis species and a few of Mahonia species have been identified as alternate hosts for the pathogen (Cheng et al. 2022; Jin et al. 2010; Li et al. 2021; Mehmood et al. 2019; Wang and Chen 2013; Zhao et al. 2013, 2016; Zhuang et al. 2019), many of which are native Chinese Berberis species. Based on our previous on-site investigations in the last decade (Li et al. 2021; Zhao et al. 2013, 2016; Zhuang et al. 2019), coexistence of wheat and susceptible Berberis is quite common in the west China, particularly in the northwestern and southwestern regions.

Sexual reproduction existence of Pst on barberry under natural conditions in spring has been demonstrated. Currently, the frequent incidents of Pst sexual cycle on susceptible barberry under field conditions have been reported in China in light of direct evidence (Chen et al. 2021; Li et al. 2016; Wang et al. 2016; Zhao et al. 2013), in Pakistan by indirect evidence (Ali et al. 2015), but not in other countries worldwide. Although barberry infection by rusts has been reported in Sweden (Berlin et al. 2013) and the U.S. Pacific Northwest (Wang and Chen 2015; Wang et al. 2015), Pst sexual reproduction was not found to occur. So far, it has been demonstrated that the sexual cycle of Pst occurs merely in spring (Li et al. 2016; Wang et al. 2016; Zhao et al. 2013). However, it is still unknown worldwide about whether the rust fungus could infect barberry to complete a sexual cycle in autumn.

In recent years, field investigations with regard to rust infections on barberry were carried out, and we observed a common phenomenon that new shoots and leaves can grow from old barberry shrubs in autumn and last for a couple of months in some regions of western China. More importantly, new pycnial and aecial infections were observed on growing young leaves of new shoots of barberry shrubs in middle to late autumn in Linzhi, Tibet, China, where susceptible barberry coexists with wheat, and climate is temperate and rainy during this period. So, we speculated that Pst is infecting susceptible barberry in autumn. If this speculation is true, it will be of importance to understand the occurrence of Pst sexual cycles and management of wheat stripe rust. Therefore, the objectives of this study were to ascertain the potential of Pst to infect susceptible barberry to complete sexual reproduction under natural conditions in autumn in Tibet, China, via combined analyses of phenotyping and genotyping.

Materials and Methods

Field investigations of regrowth of barberry shrubs in autumn

Young tissues of susceptible barberry shrubs, especially young leaves, are vulnerable to Pst. In general, new tissues grow out from barberry shrubs and can be infected by the rust in the spring. To know about similar incidents in autumn, field investigations were conducted to understand the statuses of new barberry tissue growth and stripe rust infection on barberry in some regions of western China, where climate is cool and humid for frequent rain in this season. Such weather conditions similar to spring is potentially suitable for Pst infection on barberry in the form of viable teliospore inoculum. Therefore, we performed a field investigation for observing the regrowth of new tissues of barberry shrubs from October 21 to 26, 2016, in Linzhi (also known as Nyingchi), Tibet, in the southwestern region of China.

Field investigations of rust infections on barberry and collection of aecial sample in autumn

During field investigations of new tissues of barberry shrubs in Linzhi, rust infections on tissues of barberry (Berberis polyantha Hemsl.) were also investigated simultaneously. Pycnial and aecial infections on young leaves of newly emerged shoots were observed. Young leaves on shoots with fresh aecia were cut off using a small orchard shears, put into a paper bag, and used for inoculating wheat seedlings on sampling day. Thirty rusted barberry leaves, including 52 aecial clusters, were collected and used for inoculation indoors.

Plant materials

Fifteen to 20 seeds of wheat cv. Mingxian 169, which is highly susceptible for Pst, were planted in a plastic pot (7 × 7 × 8 cm, L × W × H) filled with commercial potting mix (Inner Mongolia Mengfei Bio-tech Co., Ltd., China). All pots were transferred to a pot tray, and water was added up to 3/4 of the height of the tray for abundant water absorption in soil for seed germination. The trays were moved into a rust-free chamber for wheat growing with a dual photoperiod regime of 16 h light and 8 h darkness at 20 to 25°C. Ten-day-old seedlings were used for inoculation.

The routine wheat differential hosts set, consisting of 19 Chinese wheat genotypes (Wang et al. 2016), was used for differentiating Pst races. On Chinese wheat differentials, isolates having the same virulence pattens with races reported previously in China were indicative of known races. The international wheat differential hosts set, consisting of 25 single Yr gene lines, was utilized for differentiating virulence patterns (VPs) of Pst isolates. For one VP, isolates have the same virulence response to one gene locus.

For the wheat differential genotypes, five to seven seeds of each were planted in a square pot and separated from each other. Every four differential genotypes were sown in four corners of a pot and labeled. Cultivation conditions for wheat growth were the same as mentioned above.

Isolation of Pst from aecial samples and establishment of SU population

Wheat seedlings of four pots were arrayed in a line and overlapped by another four pots from top-to-top. Two ends of a piece of wire netting were folded downward to be a “tower bridge” in shape for standing up over wheat seedlings. The folded wire netting was placed over horizontally arrayed wheat seedlings inside the plastic storage box and kept a close distance to the wheat plants (Fig. 1A and B). Fresh aecia on leaves of the shoots were sprayed using deionized water in a hand-hold atomizer along each of all four directions in a rust-free chamber and put onto the wire nettings after spraying the wheat seedlings. The boxes were covered and moved into a dew chamber (I-36D; Percival, U.S.A.) for incubation at 100% relative humidity for 24 to 36 h at 10°C in the dark. After incubation, the plants were transferred into a rust-free growth chamber with a dual photoperiod regime of 16 h light and 6 h darkness at 13 to 16°C. For obtaining SU isolates, a single uredium was picked from an infected leaf using an inoculating needle and transferred to a new leaf of ‘Mingxian 169’ seedlings when uredinia initially appeared but did not break through the epidermis of wheat leaves. Incubation and cultivation conditions for the development of inoculated wheat plants were the same as mentioned above. Urediniospores produced on an individual leaf (‘Mingxian 169’) were collected into a glass tube. After abundant urediniospores were increased, approximately 10 mg of urediniospores were used for phenotyping on differential hosts, and another one was used to extract genome DNA. Urediniospores were kept inside a dessicator with silica gels (HG/T2765.4-2005; Qingdao Haiyang Chemical Co., Ltd) at 4°C in a refrigerator for inoculation usage.

Download as PowerPoint

Phenotyping avirulence and virulence

Both Pst samples and all SU isolates derived from the stripe rust samples were differentiated races on the 19 Chinese wheat differential hosts, and their virulence formula was characterized on the 25 single Yr gene line differentials. Wheat ‘Mingxian 169’ was used as a susceptible check at each of all phenotyping tests. Inoculation methods used in all tests were followed according to Zhao et al. (2013). When ‘Mingxian 169’ leaves were covered with abundant urediniospores, infection types (ITs) were scored at 18 to 20 days after inoculation according to a 0 to 9 scale (Line and Qayoum 1992). ITs 0 to 6 and 7 to 9 were considered as avirulent (A) and virulent (V), respectively. In all tests, incubation and temperature conditions were the same as the isolate inoculation mentioned above. The avirulence/virulence patterns, Kosman index, and Nei’s index were used to evaluate the diversity of the Pst population by using the software VAT 1.0 (Kosman and Leonard 2007).

DNA extraction and SSR genotyping

Genomic DNA was extracted from urediniospores of the overall SU population following the method described by Aljanabi and Martinez (1997). DNA concentration was measured using a Nano-Drop spectrophotometer (ND-1000; Bio-Rad Laboratories, Hercules, CA), and DNA was diluted to a concentration of 50 ng/μl with 1× TE buffer solution (10 mM Tris-HCl, 0.1 mM EDTA, pH 8.0). The DNA solutions were used as templates in polymerase chain reaction (PCR) amplification.

Fifty-three pairs of simple sequence repeats (SSR) primers for Pst (Chen et al. 2009b; Cheng et al. 2012; Enjalbert et al. 2002) were synthesized by Shanghai Sangon Bio-tech Company, Ltd. (Shanghai, China) and screened for generating polymorphism. Nine pairs of polymorphic SSR primers (Supplementary Table S1), including RJ6N (Enjalbert et al. 2002), CPS13 (Chen et al. 2009b), SUNIPst09-48, SUNIPst11-44, SUNIPst05-47, Scaffold140-292552, Scaffold246-306327, Scaffold279-18233, and Scaffold45-273492 (Cheng et al. 2012), were used for genotyping the SU population.

The PCR amplification system consisted of a total of 12.5 volumes of reaction solution as follows: 2X Taq PCR MasterMix (Taq DNA polymerase: 0.05 U/μl; MgCl₂, 4 mM; dNTPs, 0.4 mM), 1.0 μl of each primer (10 mM), 2.0 μl of DNA templates, and 8.5 μl of ddH₂O. PCR amplification was performed with touch-down program in a Thermal Cycler (S1000; Bio-Rad Laboratories). The PCR program comprised of an initial denature for 5 min at 95°C; 10 cycles consisting of 94°C for 45 s, 64°C for 45 s (a gradual temperature reduction of 1°C at each cycle), and 72°C for 45 s; followed by 25 cycles composing of 94°C for 45 s, 54°C for 45 s, and 72°C for 45 s; and a final extension for 10 min at 72°C. PCR products were analyzed using a DNA Analyzer (3730XL; Applied Biosystems, Waltham, MA). Amplicons were scored using software GeneMarker HID (Holland and Parson 2011).

Data analysis

Genetic diversity.

Genetic diversity was estimated using the observed number of alleles (Na), effective number of alleles (Ne), Shannon’s information index (I), expected heterozygosity (He), and observed heterozygosity (Ho), which were run by POPGENE software, v1.31 (Yeh et al. 1997). The following parameters for evaluating population genetic diversity were calculated using poppr: the number of multilocus genotype (MLG); the number of expected MLGs (eMLG) based on rarefaction; Stoddart and Taylor’s index (G), which represents genotypic richness (Stoddart and Taylor 1988); Evenness (E.5), which indicates the distribution of genotype abundance (Grünwald et al. 2003; Ludwig and Reynolds 1988; Pielou 1975); Nei’s gene diversity index (H), which is calculated based on the proportion of different genes among populations (Nei 1978); Simpson’s index (λ); and Shannon’s information index (I), which represents population richness and uniformity (Shannon and Weaver 1949; Simpson 1949). The clonal fraction (CF) was calculated using the formula CF = 1 − ([the number of MLGs]/[the number of individuals]) (Zhan et al. 2002).

Genetic variation.

To understand genetic variation within population and among populations for sexuality, analysis of molecular variance (AMOVA) was computed using software GenAlEx 6.5 (Peakall and Smouse 2012). Gene flow (Nm) and fixation indices (Fst) were calculated for a population by software GenAlEx 6.5 (Peakall and Smouse 2012). Nm ≥ 4 indicated frequent gene flow between two populations; 4 > Nm ≥ 1 indicated some degree of gene flow between two populations; and Nm < 1 indicated a little gene flow between populations. Fst values ranged from 0 to 0.05 indicated that population genetic differentiation is extremely low and can be ignored, Fst values ranged from 0.05 to 0.15 indicated a moderate degree of population genetic differentiation, Fst values ranged from 0.15 to 0.25 indicated significantly high population genetic differentiation, and Fst values greater than 0.25 showed great genetic differentiation among populations.

Population clusters.

To determine the minimum number of loci, a genotype accumulation curve was generated to discriminate between individuals in overall population using the poppr package (Kamvar et al. 2014) in R.

To determine the relationships among populations, an unweighted pair-group method with arithmetic means (UPGMA) tree based on Nei’s genetic distance (Nei 1972) was constructed with the POWERMARKER software, v3.25. A phylogenetic tree was created using Figtree, v1.3.1 (Liu and Muse 2005).

Based on nonparametric analysis, multivariate discriminant analysis of principal components (DAPC) was performed in the adegenet package (Jombart 2008) in R to identify partitions within the data sets regardless of the origin of Pst isolates. This method works without any assumption with regard to data structure or supporting population genetic model. The K-means were utilized to operate by sequences with increasing values of K (the number of clusters), and different clustering solutions were assessed through Bayesian Information Criterion (BIC) to identify the optimal number of population clusters (Jombart et al. 2010).

The model-based Bayesian method in STRUCTURE 2.2 (Falush et al. 2003; Pritchard et al. 2000) was utilized to identify genetic clusters and to evaluate the extent of admixture between the barberry populations. A model that allows admixture and independent allele frequencies among populations was applied. For each simulated cluster K (K = 1 to 10), 100 runs were carried out with 100,000 interactions relying on Monte Carlo Markov Chain replications and a burn-in period of 1,000,000. The output was operated using the software CLUMPP version 1.1.2 (Jakobsson and Rosenberg 2007). Output data of STRUCTURE was saved as an input file and run in the software STRUCTURE HARVESTER (http://taylor0.biology.ucla.edu/structureHarvester/) to assess the optimal number of clusters (K value). The ad hoc statistic △K (delta K), which is based on the rate of change in the logarithm probability, was computed to identify the optimal number of clusters (Evanno et al. 2005). The output of population structure was visualized using the software DISTRUCT ver. 1.1 (Rosenberg 2004, available at website: http://rosenberglab.bioinformatics.med.umich.edu/distruct.html).

Results

New shoots, and aecial production on barberry shrubs in autumn

Our field investigations were performed at four sites in autumn in Linzhi, Tibet, China, and we found that new shoots extended from old branches or came out from roots of barberry shrubs, from which new leaves grew out (Supplementary Fig. S1A and B). Statistical data showed that 106 new shoots, grown out from 19 old branches or roots of barberry shrubs at four investigated sites, produced 560 new leaves (Supplementary Table S2). Regrowth of new shoots with leaves in autumn in this region is quite common, which provide the potential for rust infections on barberry.

Meanwhile, on-site investigations for rust infections on barberry were carried out in Linzhi, Tibet, China. A total of 187 barberry shrubs growing at eight surveyed sites in this region were investigated for rust infections (Fig. 2A and B and Supplementary Table S3), 62 of which accounted for 33.2% of the total, were infected by rusts, producing only pycnia and/or aecia on young leaves of new shoots (Figs. 2C, D, E, F, and G and Supplementary Table S3). Many growing closely to the wetlands at Zhuomu Village, Bujiu Town (Figs. 2A and B), produced a great number of aecia on fresh young leaves of new shoots (Fig. 2B), and only a low pycnial or aecial production on barberry shrubs were observed in other investigated sites far away from the sampling site of Zhuomu Village (Figs. 2C and D and Supplementary Table S3). The number of pycnia and aecia was various from one new shoot to another. Some aecia, a minority of aecia observed, were already broken and released aeciospores and became white or yellowish in color, indicating that rust infections of barberry occurred earlier than the observed period (Supplementary Fig. S2A). However, most aecia were short and still developing, appearing fresh orange in color (Figs. 2E, F, and G and Supplementary Fig. S2B). Simultaneously, aecia produced on the surface of some nearly mature berries and young thorns were also observed in this season. In contrast, the number of aecia produced on the young thorns was significantly higher than those of the berries (Supplementary Fig. S2C and D). We also observed that some young leaves of new shoots were infected by rusts and initially appeared as obvious nectars (pycniospores), showing that pycnia were developing and new leaves were continuously infected by rusts (Supplementary Fig. S2E). Meanwhile, we observed that a few flowers were blooming, indicating that barberry shrubs were vigorously growing (Supplementary Fig. S2F). Importantly, more new shoots and leaves began to grow, hinting that the new leaves would be infected. Thus, barberry infection possibly lasted longer. However, no rust infection was observed on leaves of old branches of either rusted barberry or noninfected barberry shrubs at all surveyed sites.

Download as PowerPoint

Isolates recovered from rusted barberry and single uredium population

By inoculating 52 aecial clusters from 30 rusted barberry leaves, four Pst samples, T1 to T4, were obtained from rusted tissues of barberry at sampling sites (indicated by yellow dotted circle; Fig. 2B). The four stripe rust samples produced uredinia typical to Pst on ‘Mingxian 169’ leaves with high infection types (Fig. 1C and Table 1). However, most aecia merely produced necrotic lesions without uredinia on wheat leaves (Fig. 1D). Prior to uredia that did not break through wheat epidermis, a population consisting of 65 SU isolates was established by transferring an SU from the four Pst samples of T1 (20), T2 (15), T3 (15), and T4 (15) to a new leaf of wheat with an inoculating needle.

Phenotyping

By testing the 19 Chinese differential wheat genotypes, the four Pst samples T1 to T4 showed different avirulence and virulence formulas and mixed reactions on some wheat genotypes (Table 1). The sample T1 showed mixed reactions on wheat genotypes including Mentana (Unknown), Virgilio (YrVir1, YrVir2), Funo (YrA, +), Danish 1 (Yr3), Fengchan 3 (Yr1), Lovrin 13 (Yr9, +), Zhong 4 (Unknown), and Lovrin 10 (Yr9). The sample T2 represented mixed reactions on wheat genotypes of Trigo Eureka (Yr6), Lutescenes 128 (Unknown), Mentana (Unknown), Abbondanza (Unknown), Early Piemium (Unknown), Funo (YrA, +), Jubilejina 2 (YrJu1, YrJu2, YrJu3, YrJu4), and Fengchan 3 (Yr1). For the sample T3, mixed reactions were observed on wheat genotypes consisting of Trigo Eureka (Yr6), Lutescenes 128 (Unknown), Mentana (Unknown), Abbondanza (Unknown), Early Piemium (Unknown), Jubilejina 2 (YrJu1, YrJu2, YrJu3, YrJu4), Fengchan 3 (Yr1), and Suwon 11 (YrSu). The sample T4 displaying mixed reactions on wheat genotypes Trigo Eureka (Yr6), Mentana (Unknown), Virgilio (YrVir1, YrVir2), Abbondanza (Unknown), 7 = Early Piemium (Unknown), Funo (YrA, +), Fengchan 3 (Yr1), Suwon 11 (YrSu), and Zhong 4 (Unknown).

Likewise, T1 to T4 samples produced different avirulence and virulence patterns on the 25 Yr single gene lines, showing avirulence/virulence differences at these resistance gene Yr loci (Table 2). These four stripe rust samples generated mixed responses to Yr genotypes tested. Mixed reactions of the sample T1 on wheat genotypes of Yr2, Yr3, Yr4, Yr6, Yr7, Yr8, Yr17, Yr27, Yr28, Yr29, Yr32, Yr43, YrA, YrTr1, and Yr76 were observed, and T2 samples on both Yr17 and YrSP; T3 samples on Yr9, Yr28, and YrSP; and T4 samples on Yr2, Yr17, Yr28, and YrSP were also visualized.

Identification of overall 65 SU isolates on the 19 Chinese differential wheat genotypes indicated that 21 isolates were known races and 44 ones were new races that did not match races reported previously in China (Table 3 and Supplementary Table S4). The known races included eight types of races, involving CYR33 (six isolates), F0-2 (two isolates), Su11-14-4 (three isolates), Su11-15 (one isolate), Su11-24 (three isolates), Su11-4-2 (two isolates), Su11-48 (three isolates), and Su11-9 (one isolate). The new races were categorized into five race groups in case an isolate was virulent to a certain major wheat genotype of the 19 Chinese differential genotypes, consisting of the Fo race group (32 isolates) virulent to wheat genotype Funo, the Fu race group (8 isolates) virulent to Fulhard, the Ky race group (one isolate) virulent to Kangyin 655, the Su11 race group (two isolates) virulent to wheat genotype Suwon 11, and the ZS race group (one isolate) virulent to Zhong 4.

Virulence analysis of 65 SU isolates from barberry was made on single Yr gene lines, producing 26 VPs (Supplementary Table S5). Nei’s gene diversity index of the population was 0.28 on the 19 Chinese differential wheat genotypes and 0.18 on the 25 single Yr gene lines (Table 4). The Kosman index of the barberry population was 0.43 on the Chinese differentials and 0.24 on the Yr gene lines. For the T1 to T4 subpopulations, Nei’s gene diversity index ranged from 0.09 to 0.12 on the Chinese differentials and 0.06 to 0.17 on the Yr gene lines. Correspondingly, the Kosman index ranged from 0.15 to 0.19 on the Chinese differentials and 0.05 to 0.22 on the Yr gene lines.

Genetic diversity

According to the genotype curve, the curve peaked the maximum 100% at eight marker loci. This peak denoted that quantity saturation of MLGs required at least eight marker loci. Therefore, nine marker loci used for analysis of the population was adequate (Supplementary Fig. S3). Genotype analysis showed that the 65 isolates derived from T1 to T4 samples generated 15 MLGs in total (Table 4). For subpopulations of these four samples, each of the T1, T3, and T4 subpopulations produced four MLGs in addition to T2, which had five MLGs.

To determine that these Pst samples produced sexually, genetic polymorphism analysis with nine SSR markers was performed by POPGENE software and poppr package in the R program. The He of the T1 to T4 barberry subpopulations ranged from 0.09 to 0.41, and the overall population He value was 0.41 (Table 4). The overall barberry population showed high genetic diversity due to a high Shannon information index at 0.60. The Shannon information index of the T1 to T4 subpopulations ranged from 0.14 to 0.58 (Table 4).

Genetic diversity indices and linkage disequilibrium index for overall barberry population and each of the T1 to T4 subpopulations (Table 5) were also performed. For four subpopulations, the Na ranged from 1.44 to 1.89, and the Ne ranged from 1.15 to 1.81. The T2, T3, and T4 subpopulations had high and similar Ho values, ranging from 0.71 to 0.79, but the Ho value of the T1 subpopulation was 0.13, showing a low heterozygosity. The eMLG of the T1 to T4 subpopulations ranged from 3.45 to 5.00. Stoddart-Taylor’s index (G), Shannon-Wiener’s index (H), and Simpson’s index (Lambda) indicated a relatively high diversity for each subpopulation, while the overall barberry population showed the highest diversity with the maximum 7.31. Each of the T1 to T4 subpopulations and the overall barberry population showed high genotypic evenness (E.5), and genotypic evenness of the T2 subpopulation was the highest. The index of association and the standardized index of association indicated that the T1 and T2 subpopulations were sexually produced for a low level of linkage disequilibrium with a high p.rD (Supplementary Fig. S4). In comparison, the T3 and T4 subpopulations had a slightly higher level of linkage disequilibrium (p.rD = 0.001). The CF, ranging from 0.67 to 0.80, was high for each subpopulation. In total, mixed reactions, various virulence patterns, high genetic diversity index, and linkage disequilibrium indicated that the barberry population produced sexually.

Genetic variation

Allele frequency of isolates belonging to the T1 to T4 subpopulations at each marker loci were presented (Fig. 3). Isolates from different subpopulations shared most of the alleles. However, there are several unique alleles among different subpopulations. Two unique alleles, 230 (SUNIPst09-48) and 211 (Scaffold279-18233), were unique for the T1 and T4 subpopulations, respectively. Allele 221 (Scaffold279-18233) was specific for the T2 and T3 subpopulations. For the T2, T3, and T4 populations, alleles 317 (RJ6N), 128 (CPS13), and 167 (Scaffold140-292552) were unique. However, allele 224 (SUNIPst11-44) was not detected within the T1 subpopulation, and allele 218 (SUNIPst11-44) was not detected within the T4 subpopulation.

Download as PowerPoint

To understand genetic variation within/among populations, the nine polymorphic SSR markers (Supplementary Table S1) were used to analyze genetic variation of the population. AMOVA indicated that the percentage of molecular variance among subpopulations was 20%, within subpopulations was 0%, and among individuals was 80% (Table 6). These molecular variances implied that the population variation mainly resulted from subpopulations and individuals of the overall barberry population.

To reveal the correlation of the subpopulations, Nm and Fst were calculated between the TI to T4 subpopulations (Table 7). At the subpopulation level, Nm between T1 and T2, T1 and T3, and T1 and T4 was less than 1, indicating that the T1 subpopulation should be independent from the other three subpopulations. While Fst between T1 and T2, T1 and T3, and T1 and T4 was greater than 0.25, showing a high population differentiation between the T1 and the other subpopulations. On the contrary, between the T2, T3, and T4 subpopulations, Nm ranged from 6.69 to 28.42, and Fst ranged from 0.01 to 0.04, showing a close relationship between these three subpopulations.

Estimation of genetic clusters

UPGMA analysis based on Nei’s genetic distance indicated that all SU isolates were grouped into two clusters (Fig. 4). One constituted with all isolates from the T1 subpopulation, and the other consisted of isolates from the T2, T3, and T4 subpopulations. Some isolates from the T2 and T3 subpopulations were clustered into the same subclade, but isolates from the T4 subpopulation were almost clustered into an independent subclade (Fig. 4).

Download as PowerPoint

STRUCTURE analysis hinted that overall population produced two optimal clusters based on the highest delta K value (Fig. 5A). When K = 2, the two clusters (green and yellow columns) produced among all isolates of the T1 to T4 subpopulations. One consisted of the T1 subpopulation, and the other consisted of the T2, T3, and T4 subpopulations, sharing the same ancestry (Fig. 5B). With the increase of K = 3, the structure of the overall population still included two main clusters. One mainly comprised of the T1 subpopulation, and the other consisted of the T2, T3, and T4 subpopulations, sharing the same ancestry (red and yellow columns) (Fig. 5B).

Download as PowerPoint

Nonparametric DAPC analysis, used as a complementary method, was performed to further determine the accuracy of the genetic clusters. Based on BIC values, the whole population was grouped into two completely independent genetic clusters (Fig. 6B). All 20 isolates of the T1 subpopulation were divided into cluster 2 (Fig. 6C), and the remaining 45 isolates from the T2, T3 and T4 subpopulations were classified into cluster 1 (Fig. 6C). Genetic cluster analysis showed that the T1 subpopulation may have different lineage from the other three subpopulations. However, the T2, T3, and T4 subpopulations shared the same genetic ancestry, which is consistent with the result of STRUCTURE.

Download as PowerPoint

Principal coordinates analysis (PCoA) indicated that overall isolates from the T1 to T4 subpopulations were clustered in three groups (Fig. 7). The T1 and T4 subpopulations were clustered into two different group, while the T2 and T3 subpopulations were categorized into one group. The first axis (Coord. 1) of the PCoA explained 78.01% of the differences between the individual isolates, and the second (Coord. 2) explained 7.61% of the difference between the individual isolates.

Download as PowerPoint

Discussion

This study demonstrated for the first time the existence of sexual cycles of Pst on susceptible barberry as an alternate host under field conditions in autumn. This finding updates the knowledge of the disease cycle of wheat stripe rust and the strategy for integrated management of the disease. Similar to Tibet, those areas (i.e. western Asian countries) located in the high latitudes with mild and humid weather conditions, where stripe rust infection on barberry shrubs in autumn possibly occurs, would benefit from this knowledge.

Tibet is an independent epidemiological region of wheat stripe rust in the whole epidemic system of China based on genetic diversity and population structure of the stripe rust population. Previous investigations by an investigation group of Chinese Academy of Sciences from 1973 to 1976 and summarized by Xu and Cai (1979) and by Wang (1992a) from 1984 to 1991 reported that the stripe rust can oversummer and overwinter to complete year-round cycles in this region. Races of Tibetan Pst populations were remarkably different from those of the other regions in China, and evolution of Tibetan races was quite slower than other regions of China. Kuang (2010) studied that CYR28 and previously designated CYR races (also old races) were detectable among isolates collected from Linzhi. However, those rare CYR races (CYR17, CYR18, CYR21, CYR24, CYR25, and CYR26) prior to CYR28 (included) have not been detected for many years in other epidemic regions of China except, Linzhi, Tibet (Wang et al. 2012). So far, the current predominant race, CYR34, was detected for the first time in 2009 (Liu et al. 2017) and subsequently in other epidemic regions for many years but not in Tibet. In general, the evolution of Tibetan Pst populations was 3 to 5 years later than other epidemic regions of China (Li and Zeng 2002). Recently, evidence of molecular data suggested that Tibetan epidemic regions of Pst is significantly distinguished from the other epidemic regions in China due to little gene exchange, while Pst populations had extensive gene exchange among the other epidemic regions (Hu et al. 2017).

The Tibetan epidemiological region of stripe rust was subdivided into three epidemic areas: highly epidemic area, frequent occurring area, and sporadic area. Linzhi is a major wheat production region and also a highly epidemic area of the disease in Tibet (Wang 1992b). Pst isolates tested by Hu et al. (2017), Kuang (2010), and Wang et al. (2012) were collected from only Linzhi for difficulties of stripe rust collection or stripe rust not occurring in other areas of Tibet. In addition, cultivation of winter wheat is grown from middle October to early November and harvested from early to middle July in Linzhi. During the investigation, we observed stripe rust infections on autumn-sown winter wheat seedlings in addition to volunteer wheat. However, the barberry infection in relation to stripe rust prevalence on autumn-sown wheat seedlings needs further investigate.

Race surveillances of Pst populations in Tibet are limited. Several recent studies have reported race types of Tibetan Pst populations. Among Tibetan Pst populations collected in different years, the Su11 race group, virulent to wheat Suwon 11 (Su11) and used as a Chinese differential genotype, was identified as a major race group (Kuang 2010; Peng et al. 2013; Wang et al. 2012). In the present study, several isolates from naturally infected barberry in autumn were identified as the Su11 race group, including Su11-9, Su11-4-2, Su11-48, Su11-15, and Su11-14-4, which were main race types of known races among the SU population. Our results were in accord with those of both previous studies by Kuang (2010) and Wang et al. (2012). In addition, two Pst samples and their SU isolates represented narrow virulence patterns that were avirulent on most of either the Chinese differential genotypes or single Yr gene lines. Our results of this study were similar to those of the previous studies by Kuang (2010) and Wang et al. (2016).

Climatic conditions are favorable for infection and the development of the stripe rust on barberry in Linzhi, Tibet, China. Pst requires a relatively low temperature condition and high humidity for germination (optimal at 9 to 13°C) and development (optimal at 13 to 16°C), regardless of its host (primary host wheat or alternate host barberry). Pst teliospores can germinate to produce basidiospores at various temperatures, ranging from 5 to 22°C, and temperatures between 10 and 15°C were optimal for germination (Wang and Chen 2015). Linzhi belongs to a mild and semihumid area with weather conditions of a cool summer, mild winter, and abundant rainfall. Based on meteorological data from 1951 to 1980, Wang (1992b) reported that the annual average temperature in highly epidemic areas such as Linzhi ranged from 8.5 to 12.0°C, and that average temperature of the coldest months was −0.9 to 3.9°C. In this area, the average temperature was 7 to 16°C during May to September (Kong et al. 1994). Even in the hottest months of July and August, the average temperature of both months was 15.5 to 18.7°C (Wang 1992b). From March to December, weather with misting and/or dewing simultaneously appear commonly in Linzhi (Liu 2007). So, the local climate conditions are quite suitable for the development of Pst teliospores germinating to produce basidiospores and also for Pst basidiospores to penetrate young tissues of susceptible barberry in autumn.

Tibetan Pst populations are unique, and Tibet is a hot spot for the stripe rust. Molecular studies conducted on genetic diversity of Pst populations showed that Tibetan Pst populations possessed a high level of genetic diversity, distinguishing them from other epidemic regions of China (Hu et al. 2017; Kuang 2010). Indirect evidence from a study by Ali et al. (2014) reported that sexual recombination existed in Chinese populations of Pst, similar to those in near-Himalayan regional countries such as Pakistan (Khan et al. 2019). Tibet belongs to a major part of the Himalayan region in China. Direct evidence showed that Pst sexual reproduction can occur in Tibet in spring (Wang et al. 2016). In the present study, Pst sexual recombination occurring in autumn was demonstrated in this region, and the Pst populations represented high genetic diversity, which completed the Pst sexual cycles. In Tibet, for some isolates collected from barberry having similar virulence with known races, the potential threat of barberry rust populations to local wheat should be taken seriously.

Many Puccinia species including Pst could infect barberry in autumn. In this study, only four Pst samples were obtained from aecia on young barberry tissues in autumn. However, most aecia produced necrosis and chlorosis on leaves of wheat but not uredinia of Pst after inoculation, indicating that they could be other Puccinia species. Simultaneously, some aeciospores possibly failed to infect wheat because they did not contact leaves of ‘Mingxian 169’ after being released from aecia instead of falling down to the bottom of the incubator. Some might be other forms of P. striiformis but are avirulent to ‘Mingxian 169’, presenting merely chlorosis lesions (Huang et al. 2019). Other Puccinia species, including more than 30 species such as P. graminis (Berlin et al. 2013; De Bary 1867), P. brachypodii (Payak 1965), P. poa-nemoralis, P. pymaea, P. montanensis (Cummins and Greene 1966), P. arrhenatheri (Berlin et al. 2013; Urban 1967), and P. poae (Jin et al. 2010), have been reported to infect Berberis species. During our field investigations, teliospores maintained on wheat straws in fields and produced on grasses around barberry shrubs were observed (Supplementary Fig. S5A, B, C, and D). We speculated that wheat straw and grasses possibly provide potential teliospores to infect barberry tissues. Therefore, the rusts that infected barberry tissues could be the wheat stripe rust and the grass rust pathogen. In the present study, analysis of Nm and Fst between the T2 and T4 subpopulations indicated a highly frequent genetic exchange. However, the T1 subpopulation had almost no genetic exchange with the T2 to T4 subpopulations. This result supported that the T1 to T4 samples were generated by different races via sexual recombination. Analysis of clustering and population structure showed that the T1 subpopulation had different lineage with the other subpopulations, which was also in accordance with our speculation.

In our previous studies, Pst and P. graminis f. sp. tritici were isolated successfully from different naturally infected barberry species in spring in Tibet and other regions of China (Wang et al. 2016; Zhao et al. 2013). However, we failed to obtain samples of P. graminis f. sp. tritici from rusted barberry in autumn in Linzhi. The reason could be teliospores of P. graminis f. sp. tritici had a long period of dormancy after their formation, and under natural conditions, they germinate by breaking their dormancy needed for winter at higher temperatures than germination of P. striiformis teliospores (Wang and Chen 2015). In Linzhi, Tibet, China, outdoor temperatures in late October is quite low and could not initiate germination of P. graminis f. sp. tritici, so no infection of P. graminis f. sp. tritici on barberry might occur in autumn.

The present study showed the occurrence of Pst sexual reproduction on barberry in autumn under natural conditions in Linzhi, Tibet, China. This process hinted that teliospores can survive to wheat sown in October after harvest. Viable teliospores could spread to barberry plants in which teliospores germinate to produce basidiospores that infect barberry. A previous study by Eriksson and Henning (1896) reported that teliospores can germinate and produce basidiospores regardless of if it’s autumn. This supports the possibility of basidiospores being inocula that infect susceptible barberry in autumn in Tibet. However, the origin of germinable teliospores of Pst in this region is unknown. Based on testing in laboratory, teliospores on wheat straws were viable to germination (Supplementary Fig. S5A and B). Under field conditions, once teliospores formed, teliospores of the stripe rust can germinate immediately (Chen et al. 2021; Raeder and Bever 1931). In Linzhi, planting crops can last for more than a month in autumn, which provided sufficient timespan for Pst teliospore germination and infection on barberry plants nearby. In addition, Tibet is a mixed cultivation system of barley and wheat. Barberry shrubs are mostly adjacent to wheat or barley fields and are more serious than those far from fields. In Tibet, wheat and barley stubbles or straws were still kept in field after harvesting from June to September until sowing of winter wheat in autumn (Kong et al. 1994).

Natural infections of barberry by rusts in autumn occur commonly in Linzhi. As a result of the present study, field surveys were also conducted to see natural infections of barberry at Zhuomu Village and Dazhuomu Village in Linzhi, on 11 October, 2019. In total, we observed 33 barberry shrubs and found that most of them were infected and developed pycnia (<10) on young leaves of new shoots. Only a small minority of aecia were observed and already released to whiten in color. Two weeks later, fresh aecia produced on young leaves were observed at the same location. Thus, infections of barberry in autumn in Linzhi, Tibet, China, were also observed not only in late October, 2016, in this study but also in middle October, 2019, indicating that this commonly occurs in this region. Moreover, the Tibetan Plateau covers not only Tibet but also some parts of other provinces such as Qinghai. Qinghai is located north of Tibet and mixes the cropping system of winter wheat, spring wheat, and hulless barley. Climate conditions in this region is similar to that of Tibet. Field investigations were conducted in a few of provinces of Qinghai, Gansu, western Shaanxi, and Sichuan in autumn. We observed that new shoots growing out from old branches were very common in these regions, but no infection occurred in this season. Therefore, whether infections of barberry plants in autumn occur in Qinghai and other regions is unclear, and further field and laboratory investigations are needed.

Natural infections of barberry varied from one location to another. In the present study, we observed that at Zhuomu Village, Bujiu Town, Tibet, serious infections of barberry shrubs occurred at some of the surveyed sites but not at the other remaining surveyed sites (Supplementary Table S3). We noted that during field investigations, barberry shrubs growing close (∼1 m) to wetlands and small water ditches (Supplementary Fig. S5A), where they had high relative humidity, were infected by rusts more severely than those growing far away from wetlands or ditches. This humid microclimate could contribute to the occurrence of sexual cycles of rusts and result in such a big difference between serious and light rust infections on barberry shrubs.

New shoots and leaves growing out from old branches or roots of barberry plants in autumn is very common and can last a couple of months until early November or longer. This potentially provides long duration for infection by rusts on susceptible barberry. Wheat stripe rust is a typical foliar disease spread by spores. The result of this study improved knowledge on the disease cycle of wheat stripe rust in Tibet, possibly in other regions (i.e., Qinghai) of China, and even in other parts of the world in which autumn climates are similar to that in Linzhi. Therefore, in this region, barberry shrubs around wheat fields should be treated with fungicide application or eradication in early stage of pycnial development in autumn to manage the wheat stripe rust. Our results promote the overall understanding of the disease cycle and management of wheat stripe rust globally.

Acknowledgments

We would like to thank Dr. Sajid Ali, Department of Agriculture (Plant Breeding & Genetics), Hazara University Mansehra, Pakistan, who reviewed this manuscript with critical comments and suggestions and Dr. Wen Chen, Institute of Plant Protection, Guizhou Academy of Agricultural Sciences, for data analysis of this manuscript.