Low pathotype diversity in a recombinant Puccinia striiformis population through convergent selection at the eastern Himalayan centre of diversity (Nepal)
Abstract
Worldwide Puccinia striiformis f. sp. tritici (Pst) epidemics have been reported to be driven by few genetic lineages, while a high diversity is evident at the Pst Himalayan centre of diversity. This study investigated the relationship between pathotype diversity and genetic structure in Nepal, the eastern Himalayan region, which has been largely unexplored. Despite the high genetic diversity and recombinant structure detected through microsatellite genotyping, characterization of virulence phenotypes for 62 isolates identified only eight pathotypes, with two pathotypes predominant over all the populations. This is in contrast to the Pakistani and Chinese recombinant populations, where high pathotype diversity is associated with genetic diversity. The most prevalent Nepali pathotype was not a unique clonal lineage, but was represented by seven multilocus genotypes from four distinct genetic subgroups, suggesting strong directional selection on virulence genes, resulting in convergent pathotypes in distinct genetic groups. This convergent selection is discussed in comparison with clonal French and recombinant Pakistani populations. Additionally, the Nepali Pst population carried virulence to 17 out of 24 tested yellow rust resistance genes (Yr), with the absence of virulence to Victo and Early Premium and resistance genes Yr5, Yr10, Yr15, Yr24 and Yr26. Virulence to Yr2, Yr7, Yr27 and YrSu were fixed in all isolates, in line with the deployment of these resistance genes in Nepal. The results reflect the influence of resistance gene deployment on selection of virulence and pathotypes in a recombinant pathogen population, which must be considered in the context of durable resistance gene deployment.
Introduction
Wheat yellow/stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is an economically important wheat disease causing severe yield losses across the world (Chen, 2005; Hovmøller et al., 2010; Hodson, 2011; Beddow et al., 2015; Ali et al., 2017). Worldwide analyses of the pathogen have revealed a clonal population structure in northwestern Europe, North and South America, Australia and the Mediterranean region (Chen et al., 1993; Hovmøller et al., 2002; Enjalbert et al., 2005; Wellings, 2007; Bahri et al., 2009b), whereas a recombinant population structure was detected in the Himalayan and near-Himalayan region, identifying this region as a centre of diversity (Ali et al., 2014a). Nepal, as an eastern part of the Himalayan region, is certainly a place deserving better attention for the understanding of Pst epidemiology and phylogeny.
Yellow rust is one of the major constraints to wheat production in Nepal (Sharma et al., 1995; Karki et al., 2012). The disease has been reported since 1964 (Khadka & Shah, 1967) and has remained an economically important threat to wheat production in Nepal, with successive epidemics reported in the recent past (Karki et al., 2012). The epidemic during 1987 was caused by the race 7E150 (Vr1, 2, 6, 7, 8) and during 1997 by the race 134E150 (Vr2, 6, 7, 8, 9), causing devastating losses to wheat production (Karki et al., 2012). The control of the disease is one of the major objectives of national wheat research, and the development and deployment of genetic resistance remains the most economical and environmentally friendly practice for rust control (Singh et al., 2004; de Vallavieille-Pope et al., 2012). In Nepal, different major resistance genes have been deployed in the varieties at the country level, such as RR21/Sonalika with resistance gene Yr2, Annapurna 1 with Yr9, and Nepal 297 with Yr27 (Saari & Wilcoxson, 1974; Sharma et al., 1995; Duveiller et al., 2007; Karki et al., 2012). The large-scale deployment of these varieties resulted in the acquisition of corresponding virulence by the pathogen and breakdown of these major yellow rust resistance genes associated with countrywide epidemics (Singh et al., 2004; Duveiller et al., 2007; Karki et al., 2012). The acquisition of virulence to a newly deployed resistance gene is dependent, at least to some extent, on the pathogen population structure (Johnson, 1992; Line, 2002; McDonald & Linde, 2002). In addition to the information on genetic structure, knowledge of the virulences present in the pathogen populations and their combinations (i.e. pathotype or virulence profile) is crucial for a durable deployment of resistance genes (McDonald & Linde, 2002). Thus, information on the pathotype structure of Pst in Nepal in relation to its genetic structure must be generated when considering the deployment of resistance genes.
The recent worldwide epidemics have been reported to be caused by a few clonal lineages spread over large geographical regions in most of the clonal populations (Chen, 2005; Hovmøller et al., 2016; Walter et al., 2016; Ali et al., 2017). In contrast, such a predominant prevalence of one or few pathotypes was absent in the recombinant populations in Pakistan and China (Mboup et al., 2009; Ali et al., 2014c, 2017). The Nepali Pst population has been shown to possess a high genetic diversity (Hovmøller et al., 2008; Ali et al., 2014a) and recombination signature (Ali et al., 2014a). The pathotypes from the cultivars under epidemics have also been characterized to confirm the presence of virulence to the corresponding resistance genes (Karki et al., 2012; Sharma-Poudyal et al., 2013). However, the race and phenotype structure in relation to the genetic diversity in Nepal remain largely unknown. This study tested whether the previously defined recombinant population in Nepal has several infrequent pathotypes, as observed in Pakistan (Ali et al., 2014c), or is dominated by a few frequent pathotypes. The relationship between these pathotypes was then assessed in comparison to genetic structure.
Materials and methods
Sampling and spore multiplication
A set of 86 Pst isolates was collected from six locations in the wheat-growing areas in the Himalayan region of Nepal, from Pyuthan and Lumle in central Nepal to Tarahara in the east, in 2008 (Table 1). The isolates were multiplied from a single lesion on the susceptible cultivars Cartago or Michigan Amber in a fully confined security greenhouse. The 2-leaf seedlings inoculated with these isolates were incubated at 8 °C under 100% humidity for 24 h and then transferred to the greenhouse with an 8 h/15 °C dark period and 16 h/19 °C light period, under a light intensity of 300 μE m−2 s−1. Spores were harvested after 15 days and placed in a desiccator for 4 days at 4 °C, and then stored in liquid nitrogen until further virulence characterization and molecular genotyping (de Vallavieille-Pope et al., 2002, 2012).
| Parameter | Location | Overall population | |||||
|---|---|---|---|---|---|---|---|
| Pyuthana | Bhairahawaa | Lumlea | Kathmandu | Dolakha | Taraharaa | ||
| GPS data | |||||||
|
Latitude Longitude |
28.08 82.83 |
27.30 83.27 |
28.29 83.82 |
27.65 85.32 |
27.64 86.16 |
26.70 87.27 |
– |
| Elevation (m a.s.l.) | 357 | 109 | 1762 | 1400 | 1660 | 136 | – |
| Sample size | 2 | 2 | 7 | 35 | 38 | 2 | 86 |
| Gene diversity | – | – | 8.17 | 7.72 | 7.22 | – | 7.51 |
| No. of alleles per locus | – | – | 2.55 | 2.60 | 2.20 | – | 2.65 |
| Allelic richness | – | – | 1.88 | 1.78 | 1.69 | – | 1.76 |
| Expected heterozygosityb | – | – | 0.38 | 0.38 | 0.36 | – | 0.39 |
| Observed heterozygosityb | – | – | 0.49 | 0.49 | 0.52 | – | 0.49 |
| Distinct MLGs detected | – | – | 7 | 14 | 14 | – | 22 |
| Genotypic diversity | – | – | 1.00 | 0.88 | 0.83 | – | 0.87 |
| Frequency of most abundant MLG | – | – | 1.00 | 0.37 | 0.31 | – | 26.00 |
| Linkage disequilibrium (r̅d) | – | – | 0.31 | 0.37 | 0.39 | – | 0.39 |
- a Could not infer population structure of these locations due to the low sample size so included only in overall population analysis.
- b Nonsignificant differences were detected between observed and expected heterozygosity at all locations.
Characterization of virulence profiles
The virulence profile for a subset of 62 isolates was identified using a set of 34 differential lines consisting of the European and world sets of 15 differential varieties (Johnson et al., 1972), to which were added 19 wheat lines, including Chinese differentials and Avocet isolines. Each of these differential lines carries resistance genes designated by Yr (virulence to the corresponding gene is designated as Vr). The differential lines used were Kalyansona (Yr2), Federation × 4/ Kavkaz (Yr9), Clement (Yr9+), VPM1 (Yr17+), TP981 (Yr25), Anza (YrA), Early Premium (YrEp), Jubilejina 2 (YrJu), Victo (YrVic), and Australian Avocet isolines for Yr1, Yr5, Yr6, Yr7, Yr8, Yr9, Yr15, Yr24, Yr26 and Yr27 (Johnson et al., 1972; de Vallavieille-Pope & Line, 1990; de Vallavieille-Pope et al., 2012) (detailed at http://www.ars.usda.gov/SP2UserFiles/ad_hoc/36400500Resistancegenes/Yrgene.xls). This differential set was able to discriminate between 24 virulences against resistance genes Yr1, Yr2, Yr3, Yr4, Yr5, Yr6, Yr7, Yr8, Yr9, Yr10, Yr15, Yr17, Yr24, Yr25, Yr26, Yr27, Yr32, YrA, YrSd, YrSp, YrJu, YrSu, YrVic and YrEp. Seedling plants (10-day-old) of each of these lines were inoculated with a suspension of 4 mg spores and 650 μL Soltrol 170 mineral oil (Chevron-Phillips Chemical Co.). After inoculation, seedlings were incubated at 8 °C and 100% humidity for 24 h. These plants were then placed in a high confinement greenhouse as described above for spore multiplication (de Vallavieille-Pope et al., 2012). Virulence against the resistance genes in each differential line was characterized based on the relative intensity of chlorosis and/or necrosis and sporulation on the two first leaves, using the qualitative infection-type scale of 0–9 (Hovmøller et al., 2017). The infection types (IT) 7 to 9 were considered as a compatible reaction, ITs 0–4 as an incompatible reaction and ITs 5 and 6 as an intermediate reaction.
Molecular genotyping and genetic analyses
To infer the relationship between pathotype diversity and genetic structure, microsatellite data based on 20 loci of 86 isolates were analysed, including the 55 isolates already described in the worldwide population study (Ali et al., 2014a). DNA extraction and microsatellite genotyping for an additional 31 isolates was made in the same manner as described for the 55 isolates (Ali et al., 2011, 2014a).
The population genetic analysis on all 86 isolates was carried out to confirm diversity and recombination, while identifying potential population subdivision as detailed for 55 isolates analysed in the worldwide context (Ali et al., 2014a). The recombination signature was confirmed through assessing gene and genotypic diversity, allelic richness and deviation from the Hardy–Weinberg equilibrium (Raymond & Rousset, 1995; Agapow & Burt, 2001; Goudet, 2001; Belkhir et al., 2004; Arnaud-Haond & Belkhir, 2007). The clonal origin of repeated multilocus genotypes (MLGs) was confirmed through Psex value and its significance, i.e. P-value estimated with MLGsim v. 2.0 (Stenberg et al., 2003). The presence of within-Nepal genetic groups was analysed using structure v. 2.2 software (Pritchard et al., 2000) and discriminant analysis of principal components (DAPC) implemented in the adegenet package (Jombart et al., 2010). The extent of differentiation among genetic groups was assessed using pairwise FST (genetix v. 4.05.2; Belkhir et al., 2004) and the relationship among geographically spaced populations was assessed through Nei's genetic distance based neighbour-joining (NJ) tree (Langella, 2008).
Cluster analyses were carried out on microsatellite data with multibase-2015 add-in for MS excel, using the Ward method (Ward, 1963), and compared to pathotype information to analyse their resampling in various genetic groups and MLGs. Cluster analyses based on virulence data were carried out to assess divergence among the pathotypes detected.
Results
The results on the virulence phenotyping of the Nepali Pst population revealed low pathotype diversity in a recombinant and genetically diverse population, where the predominant pathotype was shared by distinct MLGs from divergent genetic groups.
Population genetic structure
The genotyping of 86 Nepali isolates with microsatellites resulted in the identification of 22 MLGs. All the repeated genotypes were confirmed with a significantly high probability of clonal origin (significant Psex values). A high allele richness, gene and genotypic diversity and nonsignificant difference between the observed and expected heterozygosity in the overall population and across all locations confirmed the existence of recombination in the Nepali Pst population (Table 1).
Population subdivision in the Nepali Pst populations, as assessed with parametric Bayesian analysis and nonparametric DAPC analysis, revealed exactly the same results at all the tested K values (Fig. S1). Despite the existence of at least five genetic groups (ΔK approach, Fig. S2), the population subdivision could not be attributed to a spatial structure, with all the genetic groups distributed across all locations, when assessed for K = 2 to 7 (Fig. S1). This was further confirmed by the observed low pairwise FST estimates across locations for locations with more than five isolates per location (Table 2a), though strong divergence among the genetic groups was evident (Table 2b; Fig. 1c).
| (a) | |||
|---|---|---|---|
| Dolakha | Kathmandu | Lumle | |
| Dolakha | – | 0.006 | 0.034 |
| Kathmandu | 0.107 | – | 0.000 |
| Lumle | 0.050 | 0.611 | – |
| (b) | |||||
|---|---|---|---|---|---|
| G1 | G2 | G3 | G4 | G5 | |
| G1 | – | 0.295 | 0.392 | 0.175 | 0.215 |
| G2 | 0.000 | – | 0.436 | 0.233 | 0.490 |
| G3 | 0.000 | 0.000 | – | 0.448 | 0.468 |
| G4 | 0.000 | 0.000 | 0.000 | – | 0.366 |
| G5 | 0.000 | 0.000 | 0.000 | 0.000 | – |
- Significant FST values (<0.05) are shown as bold.
Virulences and pathotypes detected in Nepal
When considering the overall population, virulences were detected for 17 out of the 24 yellow rust resistance (Yr) genes/factors tested, revealing absence of virulence to only five resistance genes (Yr5, Yr10, Yr15, Yr24, Yr26) and to two cultivars, Early Premium and Victo (Table 3). Virulences to Yr2, Yr7, Yr27 and YrSu were fixed across all locations, while Vr1 and Vr8 were fixed in at least three locations. Among the rare virulences, Vr3, Vr4, Vr25 and VrVic, though common in Europe (Hovmøller et al., 2002; de Vallavieille-Pope et al., 2012), were present at a very low frequency only at two locations (Kathmandu and Dolakha); Vr17 and Vr32 were detected in only one isolate at Kathmandu and VrSd and VrSp were detected in only one isolate at Dolakha (Table 3). The absence of virulence to Victo in more than 90% of the isolates and fixation of virulence to Suwon 92/Omar distinguished the Nepali Pst populations from other worldwide populations (Ali et al., 2017).
| Virulence | Differential line testeda | Location | Frequency (%) in overall population | |||||
|---|---|---|---|---|---|---|---|---|
| Pyuthan | Bhairahawa | Lumle | Kathmandu | Dolakha | Tarahara | |||
| Vr1 | Chinese 166, Yr1/6 × Avocet S | 2 | 3 | 5 | 28 | 18 | 3 | 95 |
| Vr2 | Heines VII, Kalyansona | 2 | 3 | 5 | 30 | 19 | 3 | 100 |
| Vr3 | Vilmorin 23, Nord Desprez | 0 | 0 | 0 | 1 | 1 | 0 | 3 |
| Vr4 | Hybrid 46 | 0 | 0 | 0 | 1 | 1 | 0 | 3 |
| Vr5 | Yr5 Avocet S | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vr6 | Heines Kolben, Heines Peko, Yr6/6 × Avocet S | 2 | 3 | 5 | 30 | 18 | 3 | 98 |
| Vr7 | Lee, Reichersberg 42, Yr7/6 × Avocet S | 2 | 3 | 5 | 30 | 19 | 3 | 100 |
| Vr8 | Compair, Yr8/6 × Avocet S | 1 | 1 | 3 | 24 | 17 | 0 | 74 |
| Vr9 | Clement, Federation × 4/Kavkaz | 1 | 2 | 2 | 6 | 2 | 3 | 26 |
| Vr10 | Moro | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vr15 | Yr15/6 × Avocet S | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vr17 | VPM1 | 0 | 0 | 0 | 1 | 0 | 0 | 2 |
| Vr24 | Yr24/6 × Avocet S | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vr25 | TP981 | 0 | 0 | 0 | 1 | 1 | 0 | 3 |
| Vr26 | Yr26/6 × Avocet S | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Vr27 | Yr27/6 × Avocet S | 2 | 3 | 5 | 30 | 19 | 3 | 100 |
| Vr32 | Carstens V | 0 | 0 | 0 | 1 | 0 | 0 | 2 |
| VrA | Anza, Funo | 1 | 1 | 3 | 25 | 19 | 1 | 81 |
| VrSd | Strubes Dickkopf | 0 | 0 | 0 | 0 | 1 | 0 | 2 |
| VrSp | Spaldings Prolific | 0 | 0 | 0 | 0 | 1 | 0 | 2 |
| VrJu | Jubilejina 2 | 0 | 0 | 0 | 3 | 2 | 0 | 28 |
| VrSu | Suwon 92/Omar | 0 | 0 | 0 | 3 | 2 | 0 | 100 |
| VrVic | Victo | 0 | 0 | 0 | 1 | 1 | 0 | 3 |
| VrEp | Early Premium | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Susceptible | Michigan Amber | 2 | 3 | 5 | 30 | 19 | 3 | 100 |
| Susceptible | Cartago | 2 | 3 | 5 | 30 | 19 | 3 | 100 |
| No. of isolates tested | 2 | 3 | 5 | 30 | 19 | 3 | 62 | |
- The number represents the number of isolates in which the virulence was present; total number of isolates tested in each location is given at the bottom.
- a Virulence factors for A, Sd, Sp, Su, Ju, Vic and Ep refer to resistance genes of cvs Anza, Strubes Dickkopf, Spalding Prolific, Suwon 92/Omar, Jubilejina 2, Victo and Early Premium, respectively.
A total of eight pathotypes were detected for 62 isolates (Table 4). The pathotypes varied in their complexity, ranging from a pathotype with seven virulences (PN-2: Vr1, 2, 6, 7, 9, 27, Su) to one with 15 virulences (PN-6: Vr1, 2, 3, 4, 6, 7, 9, 17, 25, 27, 32, A, Su, Ju, Vic). The pathotypes detected were clearly divergent when analysed at the virulence level (Fig. 2). Among these pathotypes, PN-1 (Vr1, 2, 6, 7, 8, 27, A, Su) was the most prevalent in Nepal, detected for 43 isolates, followed by PN-2 (Vr1, 2, 6, 7, 9, 27, Su), which was sampled in 12 isolates (Table 4). Pathotype PN-1 was resampled in at least seven MLGs belonging to four genetic groups (G1, G2, G4 and G5), while PN-2 consisted of two MLGs belonging to a single genetic group, G3 (Figs 3 & S3). Cluster analysis based on microsatellite genotype data further endorsed the detection of pathotype PN-1 in distant genetic lineages (Fig. 4). The remaining six pathotypes were very rare, represented by one or two isolates, and found only at a single location.
| Pathotype | No. of isolates | Virulence/avirulence profilea | Location | Genotypic group | |
|---|---|---|---|---|---|
| Vr | Avr | ||||
| PN-1 | 43 | 1, 2, 6, 7, 8, 27, A, Su | 3, 4, 5, 9, 10, 15, 17, 24, 25, 26, 32, Sd, Sp, Ju, Vic, Ep | Pyuthan, Bhairahawa, Lumle, Kathmandu, Dolakha | G1, G2, G4, G5 |
| PN-2 | 12 | 1, 2, 6, 7, 9, 27, Su | 3, 4, 5, 8, 10, 15, 17, 24, 25, 26, 32, A, Sd, Sp, Ju, Vic, Ep | Pyuthan, Bhairahawa, Lumle, Kathmandu, Tarahara | G3 |
| PN-3b | 1 | 2, 6, 7, 8, 27, A, Su, Ju | 1, 3, 4, 5, 9, 10, 15, 17, 24, 25, 26, 32, Sd, Sp, Vic, Ep | Dolakha | G1 |
| PN-4c | 1 | 2, 6, 7, 8, 27, A, Su, Ju | 1, 3, 4, 5, 9, 10, 15, 17, 24, 25, 26, 32, Sd, Sp, Vic, Ep | Kathmandu | G4 |
| PN-5 | 1 | 2, 6, 7, 8, 27, A, Su, Ju | 1, 3, 4, 5, 9, 10, 15, 17, 24, 25, 26, 32, Sd, Sp, Vic, Ep | Kathmandu | G4 |
| PN-6 | 1 | 1, 2, 3, 4, 6, 7, 9, 17, 25, 27, 32, A, Su, Ju, Vic | 5, 8, 10, 15, 24, 26, Sd, Sp, Ep | Kathmandu | G4 |
| PN-7 | 1 | 1, 2, 3, 4, 7, 9, 25, 27, A, Sd, Sp, Su, Ju, Vic | 5, 6, 8, 10, 15, 17, 24, 26, 32, Ep | Dolakha | G4 |
| PN-8 | 2 | 1, 2, 6, 7, 9, 27, A, Su | 3, 4, 5, 8, 10, 15, 17, 24, 25, 26, 32, Sd, Sp, Ju, Vic, Ep | Tarahara | G3 |
- a Virulence factors for A, Sd, Sp, Su, Ju, Vic and Ep refer to resistance genes of cvs Anza, Strubes Dickkopf, Spalding Prolific, Suwon 92 × Omar, Jubilejina 2, Victo and Early Premium, respectively.
- b PN-3 and PN-4 were distinguished from PN-5 by virulence Vr7; PN-3 and PN-4 were virulent on both Lee and Yr7 NIL Avocet, but PN-5 was virulent on Lee and avirulent on Yr7 NIL Avocet.
- c PN-4 and PN-5 were distinguished from PN-3 by virulence Vr8; PN-3 was virulent on Compair; PN-4 and PN-5 were avirulent on Compair. The three isolates were avirulent on Yr8 NIL Avocet.
Pathotype versus genetic structure in other geographical regions
The presence of a common pathotype resampled in divergent genetic groups of a recombinant population was typical of the Nepali population, as in the recombinant population of Pakistan (Table S1), where very high pathotype diversity was observed, with 53 pathotypes out of 127 tested isolates partitioned into four distinct genetic groups and one unassigned/admixed group (Ali et al., 2014c). When the genotypic profile of these 127 isolates was assessed in comparison to the genetic groups identified in the Pakistani population (Ali et al., 2014b), none of the Pakistani pathotypes was predominant across locations and only two pathotypes were detected in three genetic groups, and six were detected in two genetic groups. In contrast, when the genotypic information for the clonal population of France was considered (Table S2), 21 pathotypes were detected across more than a 25-year period (based on hundreds of isolates), all of which represented a single North European clonal lineage, with no pathotype detected in more than one genetic group (Enjalbert et al., 2005; de Vallavieille-Pope et al., 2012). However, the invasive strains such as 6E16-type races and PstS2 races sampled in the south of France belonged to other genetic groups, all of which represented a particular genetic lineage (Walter et al., 2016; Ali et al., 2017).
Discussion
Results from this study revealed low pathotype diversity in a genetically diverse and recombinant Nepali Pst population, which represents the eastern part of the plausible centre of origin of the pathogen, i.e. the Himalayan and near-Himalayan region (Ali et al., 2014a). The genetically diverse Nepali Pst population was dominated by a major pathotype that was resampled in seven distinct MLGs from different genetic groups. Selection of the same pathotype in distant genetic groups suggests a strong and unidirectional selection due to the host yellow rust resistance genes, with a convergent response of diverse genetic groups of the pathogen population.
The larger set of isolates used in the present study confirmed the recombination signature and high gene and genotypic diversity in the Nepali Pst population, as previously reported (Ali et al., 2014a). The observed diversity and recombinant population structure would arise, at least to some extent, from the sexual reproduction on its recently identified alternate host, Berberis spp. (Jin et al., 2010). Indeed the Nepali Pst populations have been characterized for a high telial production capacity (Ali et al., 2010) and the region is abundant with its alternate host, Berberis spp. (Shrestha & Dhillon, 2003), which would contribute to local survival of the pathogen (Ali et al., 2014b, 2016; Rodriguez-Algaba et al., 2014). Analysis of within-Nepal population subdivision suggests the existence of at least five distinct genetic subgroups, which cannot be explained by the spatial population structure, especially with low sample size from some locations. The observed population structure needs to be confirmed with more intensive sampling (Ali & Hodson, 2017).
The pathotype determination revealed a low diversity in the Nepali Pst population and predominance of the pathotype PN-1 (Vr1, 2, 6, 7, 8, 27, A, Su). The absence or rarity of virulences such as Vr5, Vr10, Vr15, Vr24, Vr26 and VrVic could be explained by the absence of the corresponding resistance genes in the Nepali wheat germplasm. Accordingly, the fixation and high frequency of virulences such as Vr2, Vr7 and Vr27 could be explained by the countrywide deployment of the corresponding resistance genes in the recent past (Sharma et al., 1995; Karki et al., 2012).
Comparison with the virulences and pathotypes reported by previous studies from Nepal enabled an assessment of the overall pathotype structure over two decades since 1985 (Sharma et al., 1995; Karki et al., 2012). The pathotypes found during 1985–1990 in Nepal had similar virulence spectra to those observed in 2008 with the virulences Vr1, Vr2, Vr6, Vr7, Vr8 and VrSu. The virulences Vr9, Vr27 and VrA were acquired after the 1990s, which could be driven by the deployment of corresponding resistance genes in the region, as elaborated above (Karki et al., 2012; Ali et al., 2017). Comparison with the virulence structure during the same (2007–8) period (Sharma-Poudyal et al., 2013) confirmed both the fixation or high frequency of VrA, Vr2, Vr6 and Vr7, and the absence or low frequency of Vr5, Vr15, Vr24, Vr32 and VrSp (Table 5). The remaining virulences were found in varying frequency between the two studies. The limited discrepancy observed between the two studies could be explained by the sampling effect, though the major differences could also be explained by the use of different differential sets. The differential and susceptible lines may contain additional resistance factors when exposed to exotic isolates (de Vallavieille-Pope & Line, 1990; Nazari & El Amil, 2013; Hovmøller et al., 2017). Pathotyping of four isolates collected from Nepal in early 2005, used in a global context (Hovmøller et al., 2008), revealed that all four had Vr1, Vr6, Vr7, Vr8 and VrSu, and differed only for the presence/absence of Vr4 and Vr25. Unlike in the 2008 study, Vr2, Vr3, Vr9, Vr10, Vr15, Vr17, Vr24 and VrSd were absent (Hovmøller et al., 2008); however, this could be due to limited sampling.
| Yr gene assesseda | Differential line | Yr gene | Present analysis | Sharma-Poudyal et al. analysis | ||
|---|---|---|---|---|---|---|
| N b | % | N | % | |||
| Yr1 | Yr1/6 × Avocet S | 1 | 59 | 95 | 15 | 71 |
| Chinese 166 | 1+ | 59 | 95 | 15 | 71 | |
| Yr2 | Siete Cerros T66 | 2 | – | – | 21 | 100 |
| Kalyansona | 2 | 62 | 100 | – | – | |
| Heines VII | 2+ | 2 | 3 | 9 | 43 | |
| Yr3 | Vilmorin 23 | 3+ | 2 | 3 | – | – |
| Nord Desprez | 3+ | 2 | 3 | – | – | |
| Druchamp | 3+ | – | – | 10 | 48 | |
| Stephens | 3+ | – | – | 20 | 95 | |
| Yr4 | Hybrid 46 | 4+ | 2 | 3 | – | – |
| Yamhill | 4, 2+ | – | – | 5 | 24 | |
| Yr5 | Yr5/6 × Avocet S | 5 | 0 | 0 | 0 | 0 |
| Yr6 | Yr6/6 × Avocet S | 6 | 61 | 98 | 20 | 95 |
| Heines Kolben | 6+2 | 47 | 76 | – | – | |
| Heines Peko | 6+2 | 1 | 2 | – | – | |
| Fielder | 6+ | – | – | 20 | 95 | |
| Yr7 | Yr7/6 × Avocet S | 7 | 60 | 97 | 21 | 100 |
| Lee | 7+ | 62 | 100 | 18 | 86 | |
| Reichersberg 42 | 7+ | 2 | 3 | – | – | |
| Yr8 | Yr8/6 × Avocet S | 8 | 46 | 74 | 18 | 86 |
| Compair | 8+ | 1 | 2 | 15 | 71 | |
| Yr9 | Yr9/6 × Avocet S | 9 | – | – | 12 | 57 |
| Federation × 4/Kavkaz | 9 | 16 | 26 | – | – | |
| Clement | 9+ | 2 | 3 | 9 | 43 | |
| Yr10 | Yr10/6 × Avocet S | 10 | – | – | 4 | 19 |
| Moro | 10+ | 0 | 0 | 4 | 19 | |
| Yr15 | Yr15/6 × Avocet S | 15 | 0 | 0 | 0 | 0 |
| Yr17 | Yr17/6 × Avocet S | 17 | – | – | 18 | 86 |
| VPM1 | 17 | 1 | 2 | – | – | |
| Hyak | 17+ | – | – | 2 | 10 | |
| Yr21 | Lemhi | 21 | – | – | 21 | 100 |
| Yr24 | Yr24/6 × Avocet S | 24 | 0 | 0 | 1 | 5 |
| Yr25 | TP981 | 25 | 2 | 3 | 8 | 38 |
| Strubes Dickkopf | 25+ | 1 | 2 | – | – | |
| Yr26 | Yr26/6 × Avocet S | 26 | 0 | 0 | – | – |
| Yr27 | Yr27/6 × Avocet S | 27 | 62 | 100 | 16 | 76 |
| Yr32 | Yr32/6 × Avocet S | 32 | – | – | 0 | 0 |
| Carstens V | 32 | 1 | 2 | – | – | |
| YrA | YrA/6 × Avocet S | A | – | – | 21 | 100 |
| Anza | A+ | 49 | 79 | – | – | |
| Funo | A+ | 50 | 81 | – | – | |
| YrSp | YrSp/6 × Avocet S | Sp+ | – | – | 0 | 0 |
| Spaldings Prolific | Sp+ | 1 | 2 | – | – | |
| YrSu | Suwon 92/Omar | Su | 62 | 100 | – | – |
| YrVic | Victo | Vic | 2 | 3 | – | – |
| Susceptible | Michigan Amber | Susceptible | 62 | 100 | – | – |
- a Factors A, Sp, Su and Vic refer to the resistance genes of the Anza, Spalding Prolific, Suwon 92/Omar and Victo cultivars, respectively.
- b Number of isolates, out of a total of 62 isolates in the present study and 21 isolates in the Sharma-Poudyal et al. (2013) study.
The population structure of the Nepali population was compared with a previously studied Himalayan population of Pakistan (Mboup et al., 2009; Duan et al., 2010; Bahri et al., 2011; Ali et al., 2014a,b). The high genetic diversity and recombinant structure were in accordance with the previously reported recombination signature in the Himalayan populations. Comparison of the virulence and pathotype structure of the Nepali Pst population with other world populations revealed some features specific to Nepal. Virulence to Victo was absent in more than 90% of the Nepali isolates, in contrast to the fixation of this virulence in other worldwide populations, with the exception of some Pakistani populations (Ali et al., 2011, 2014c; Bahri et al., 2011). Similarly, the fixation of virulence to Suwon 92/Omar is also typical of the Nepali Pst population, which is present in a variable frequency in other worldwide Pst populations but never fixed in any studied population (Hovmøller et al., 2002; Mboup et al., 2009; de Vallavieille-Pope et al., 2012; Ali et al., 2014c). This specificity of the Nepali Pst population was further endorsed by the prevalence of a Nepal-specific genetic group when compared with worldwide populations at the level of genetic markers, in particular the Nepali genetic group differed from Chinese and Pakistani genetic groups, both also having a recombinant structure (Ali et al., 2010, 2014a). The pathotype diversity in Nepal was lower than the other recombinant Chinese (Mboup et al., 2009) and Pakistani (Ali et al., 2014c) populations.
Despite the high diversity and a recombinant population structure, the Nepali Pst population was dominated by pathotype PN-1, which was present in genetically distinct lineages. This is interesting and original to the population biology and epidemiology of Pst, as until now a genetically diverse and recombinant population structure has been associated with high virulence and pathotype diversity, as reported in the neighbouring parts of the centre of diversity in Pakistan and China (Mboup et al., 2009; Bahri et al., 2011; Ali et al., 2014c). In contrast, low pathotypic diversity is well known in clonal Pst populations, where the dominant pathotype is associated with a single genetic group, itself presenting a low diversity, with closely related MLGs, as in the case of the ‘Warrior’-related race groups (Hovmøller et al., 2016), PstS1/S2 strains (Walter et al., 2016), and the divergent pathotypes from the north and south of France (Enjalbert et al., 2005; de Vallavieille-Pope et al., 2012). The low pathotype diversity through convergent selection detected in Nepal has not been detected in other worldwide populations, including the clonal population of France and the recombinant population of Pakistan (Enjalbert et al., 2005; de Vallavieille-Pope et al., 2012).
The pathotype PN-1 was detected in four genetic groups and seven distant MLGs, reflecting a convergent selection where the same/related phenotype is selected in different genetic groups in response to host resistance gene deployment. It could result from the fact that the prevalence of some virulences are an obligatory requirement in the Nepali population, or have been mandatory in the past, and become fixed in the population. Thus, isolates carrying those virulences have the highest fitness, regardless of their genetic subgroup. This indicates that the prevalence of virulences Vr1, Vr2, Vr6, Vr7, Vr27, VrA and VrSu could be considered as indispensible in the Nepali population, or have been critical for Pst fitness in the past. Indeed, resistance genes Yr2, Yr6, Yr7, Yr9 and Yr27 have been widely deployed in South Asian wheat varieties (Singh et al., 2004; Bahri et al., 2011). As early as 1979–80, resistance genes Yr6 and Yr8 were found ineffective in Nepal (Karki et al., 2012). Yr2 was the effective gene of RR-21 (synonymous to Sonalika), which was defeated after its deployment in the mid-1980s in Nepal (Singh & Johnson, 1988; Sharma et al., 1995). In the 1990s, the varieties Annapurna 1 and Annapurna 3 were widely deployed in Nepal, which carried Yr7, Yr9 and YrA+ (Sharma et al., 1995). Yr6, Yr7 and YrA were already ineffective (Sharma et al., 1995), while Yr9 became ineffective in the late 1990s in Nepal (Singh et al., 2004; Karki et al., 2012). Yr27 was deployed in the late 1990s and early 2000s and was defeated in the mid-2000s in Nepal (Rosyara & Joshi, 2005). Effectively, the virulence to Yr1, Yr2, Yr6, Yr7, Yr27, YrSu and YrA were either fixed or at a very high frequency in the population. Thus, the fitness advantage of isolates carrying their correspondent virulences to these resistance genes were selected regardless of their genetic lineage. This convergent selection would have resulted in low pathotype diversity in a genetically diverse population. Selection of virulences in the absence of certain resistance genes would reflect on a lack of cost of resistance, with practical implications for resistance gene deployment (Bahri et al., 2009a; Brown, 2015), which needs to be further explored in the context of recombination, particularly in the pathogen centre of diversity.
This study reports for the first time on convergent selection in the wheat yellow rust pathogen, where low pathotype diversity was observed in a population with high genetic diversity in Nepal, which lies in the eastern part of the Pst plausible centre of origin, i.e. the Himalayan and near-Himalayan region (Ali et al., 2014a). The virulence analysis revealed a virulence structure specific to Nepal, with one pathotype dominating the overall population. The predominance of a single pathotype represented by different MLGs in a recombinant population should be further explored. The temporal maintenance of the Nepali Pst population via a ‘green bridge’, with sexual reproduction on the alternate host Berberis spp., and its exact role in the epidemiology of yellow rust in Nepal, should also be considered.
Acknowledgements
The authors are thankful to Laurent Gérard and Nathalie Retout for their assistance in spore multiplication. This work received support from the European Integrated Project BIOEXPLOIT, FOOD-CT-2005-513959 and EMERFUNDIS, ANR 07-BDIV-003. It is declared that the authors have no conflict of interest.
