Identification of New Sources of Resistance to Wheat Stem Rust in Aegilops spp. in the Tertiary Genepool of Wheat

Recent stem rust epidemics in eastern Africa and elsewhere demonstrated that wheat stem rust is a re-emerging disease posing a threat to wheat production worldwide. The cultivated wheat gene pool has a narrow genetic base for resistance to virulent races, such as races in the Ug99 race group. Wild relatives of wheat are a tractable source of stem rust resistance genes. Aegilops species in the tertiary genepool have not been exploited to any great extent as a source of stem rust resistance. We evaluated 1,422 accessions of Aegilops spp. for resistance to three highly virulent races (TTKSK, TRTTF, and TTTTF) of Puccinia graminis f. sp. tritici. Species studied include Ae. biuncialis, Ae. caudata, Ae. comosa, Ae. cylindrica, Ae. geniculata, Ae. neglecta, Ae. peregrina, Ae. triuncialis, and Ae. umbellulata that do not share common genomes with cultivated wheat. High frequencies of resistance were observed as 977 (68.8%), 927 (65.2%), and 850 (59.8%) accessions exhibited low infection types to races TTKSK, TTTTF, and TRTTF, respectively. Contingency table analyses showed strong association for resistance to different races in several Aegilops spp., indicating that for a given species, the resistance genes effective against multiple races. Inheritance studies in selected accessions showed that resistance to race TTKSK is simply inherited.


INTRODUCTION
Wheat stem rust, caused by Puccinia graminis Pers.:Pers. f. sp. tritici Eriks. & E. Henn. (Pgt), is a devastating disease of durum wheat (Triticum turgidum L. ssp. durum) and common or bread wheat (T. aestivum L.). Severe epidemics have been reported in all major wheat growing areas in the world (Roelfs, 1985;Saari and Prescott, 1985). For decades, stem rust has been under effective control through the use of genetic resistance. The occurrence and spread of Sr31-virulence races in the Ug99 race group in East Africa and other virulent races causing epidemics and localized outbreaks in Ethiopia (Olivera et al., 2015), Europe (Bhattacharya, 2017;Olivera Firpo et al., 2017;Lewis et al., 2018) and Central Asia (Shamanin et al., 2018), indicates that the disease is re-emerging as a threat to wheat production. Races in the Ug99 group have been detected across South, East and northern Africa, and the Middle East (Pretorius et al., 2000;Singh et al., 2015;Newcomb et al., 2016), and have the potential to reach critical wheat growing regions in the world (Park et al., 2011). The Ug99 race group has been rapidly evolving, producing variants with virulence to stem rust resistance genes including Sr24 (Jin et al., 2008), Sr36 (Jin et al., 2009), and SrTmp (Newcomb et al., 2016) that are important in stem rust resistance breeding (Singh et al., 2015).
The cultivated wheat gene pool has a narrow genetic base for resistance to the contemporary virulent races, such as TTKSK Singh et al., 2006;Newcomb et al., 2016), TRTTF (Olivera et al., 2012); and TKTTF (Olivera Firpo et al., 2017). In order to broaden the basis of stem rust resistance in wheat breeding programs, it is necessary to identify and introgress effective genes from all genepools of wheat. Wild relatives of wheat are a tractable source of stem rust resistance genes. Indeed, a number of resistance genes derived from wild relatives of wheat appeared to be more effective against the races in the Ug99 group than Sr genes of wheat origin Jin et al., 2007). Aegilops is the most closely related genus to Triticum (Kimber and Feldman, 1987;Jiang et al., 1994) and comprises 23 species that include diploid, tetraploid, and hexaploid genomes (van Slageren, 1994). Aegilops species are known to be a rich source of stem rust resistance, and several stem rust resistance genes have been transferred into cultivated wheat (Friebe et al., 1996;Schneider et al., 2008;Liu et al., 2011a,b;Olson et al., 2013a,b).
Ease of hybridization and reduced linkage drag make introgression from species in the primary gene pool preferred by wheat breeders to incorporate new alleles in their breeding programs (Feuillet et al., 2008). However, species in the secondary and tertiary gene pools constitute an important reservoir of genetic variability (Qi et al., 2007). Aegilops species in the tertiary genepool have not been exploited to any great extent for wheat improvement, and for resistance to TTKSK and other virulent Pgt races in particular. The objective of this study was to evaluate a collection of nine Aegilops species in the tertiary gene pool of wheat for resistance to race TTKSK and other Pgt races.

Germplasm
A total of 1,422 accessions of nine Aegilops species (three diploid and six tetraploid) deposited at the USDA-ARS, National Small Grain Collection (NSGC), Aberdeen, ID, were evaluated in this study. Species, the number of accessions and country of origin of each Aegilops species are given in Table 1.

Inoculation, Incubation, and Disease Assessment
With the objective of identifying multiple and diverse resistance genes in individual accessions, we evaluated this Aegilops collection against multiple races with different virulence spectrum and origin. All accessions were characterized for reaction to three virulent Pgt races: TTKSK (Kenya), TRTTF (Yemen), and TTTTF (United States). Accessions resistant to the three races were further evaluated for their reaction to four additional US races (TPMKC, RKRQC, QTHJC, and QFCSC). The race designations are based on the letter code nomenclature system (Roelfs and Martens, 1988;Roelfs et al., 1993;Jin et al., 2008). Avirulence/virulence profile of the Pgt isolates used in the disease assessments is summarized in Table 2. Disease evaluations were conducted in two independent experiments. In each experiment, five seedlings per accession were inoculated with each race on fully expanded primary leaves 8-9 days after planting. Details on inoculation procedures and disease assessment were described by Jin et al. (2007). Disease reactions were classified according to Stakman et al. (1962). Infection types (ITs) 0, 1, and 2 were considered as resistant reactions and ITs 3 and 4 were considered as susceptible. Wheat cultivar McNair 701 (Cltr 15288) was included as susceptible check. Analyses of association via contingency tables were conducted to assess potential relationships of resistance to different Pgt races.

Inheritance Study
Bi-parental crosses between selected resistant accessions and a susceptible accession in five Aegilops species were made and F 2 progeny were produced by selfing F 1 plants. Seventeen F 2 populations (four from Ae. cylindrica, four from Ae. peregrina, six from Ae. triuncialis, two from Ae. umbellulata, and one from Ae. comosa) were evaluated for reaction to race TTKSK to determine the inheritance of resistance based on phenotypic ratios. Chi-square (χ 2 ) test was used to determine the goodness of fit to expected genetic ratios in the F 2 generation.
Pairwise association for resistance to races TTKSK, TRTTF, and TTTTF exhibited variation among species and pathogen races. Over 75% of the accessions of Ae. geniculata and Ae. neglecta were resistant to races TTKSK, TRTTF, and TTTTF ( Table 3). Resistance to pairs of the three Pgt races in Ae. geniculata and Ae. neglecta were highly associated ( Table 4), suggesting that accessions resistant to race TTKSK are likely to be resistant to races TRTTF and TTTTF. Association for the reaction to races TTSKS, TRTTF, and TTTTF was also observed  Turkey  79  87  82  148  1  33  125  3  66  624   Greece  34  2  85  73  58  28  27  0  0  307   Macedonia  3  5  11  21  0  0  13  0  0 Roelfs and Martens (1988) and Jin et al. (2008).   obtained in accessions from Macedonia (35.8%), France (38.9%), Iraq (46.7%), and Montenegro (75.0%), but the numbers of accessions evaluated from these countries were significantly smaller. Segregation ratios of the F 2 progeny from biparental crosses between resistant and susceptible accessions indicated that resistance to race TTKSK in selected accessions is mostly conferred by single genes ( Table 6). Eight resistant Aegilops accessions carry a single gene with dominant effect, whereas two resistant accessions carry a single gene with recessive effect. Two genes conferring resistance to race TTKSK were observed in three accessions of Ae. triuncialis. Inheritance with epistatic effect between two genes was also observed in four resistant parents. Segregation ratios of the F 2 progeny of one Ae. triuncialis and one Ae. umbellulata resistant parent fit to a 9:7 ratio indicating the presence of a complementary gene action with duplicate recessive epistasis. Epistatic effect between two dominant genes was also observed in two Ae. peregrina resistant parents (Table 6), where the F 2 progenies fit to a 11:5 ratio.

DISCUSSION
Races of P. graminis f. sp. tritici, such as the Ug99 race group, TKTTF and others detected from the contemporary Pgt populations worldwide, are a serious threat to bread and durum wheat production worldwide because of their broad virulence to many cultivars and rapid geographic spread. The limited number of stem rust resistance genes effective against these virulent races requires the identification of new sources of resistance. Different Aegilops species have contributed several stem rust resistance genes effective against race TTKSK including Sr32, 33, 45, 46, 47, 51, 53, SrTA10187 and SrTA10171 (Friebe et al., 1996;Schneider et al., 2008;Liu et al., 2011a,b;Olson et al., 2013a,b). However, only one gene, Sr53, is derived from Ae. geniculata in the tertiary genepool. Results from this study demonstrated that Aegilops species in the tertiary genepool of wheat are a rich source of resistance to race TTKSK and other Pgt races with broad virulence.
Although the overall frequency of resistant accessions in the entire Aegilops collection evaluated against races TTKSK, TRTTF, and TTTTF in this study was over 60%, we observed significant variation among species. Only two species (Ae. geniculata and Ae. neglecta) exhibited a high frequency of resistance (over 80%) against the three races. Interestingly in Ae. biuncialis, a species that also shares the same genome constitution as Ae. geniculata and Ae. neglecta (UUMM), the frequencies of resistance varied, exhibiting a high level of race specificity. Differences in the frequencies of resistance to stem, stripe, and leaf rust in species carrying the same genome have been also reported in the Section Sitopsis (SS genome) of Aegilops (Anikster et al., 2005;Scott et al., 2014). In species such as Ae. geniculata and Ae. neglecta where there is a high degree of association of the reactions to races TTKSK, TRTTF, and TTTTF, it is highly likely that the genes that confer resistance to one race is also effective against the other races. The progeny populations via bi-parental crosses initiated through this study will be further developed and analyzed to understand the genetic relationships for resistance to different races in these selected accessions. Race specificity was a common feature observed in this Aegilops collection, as five species exhibited a percentage of accessions resistant to all three races TTKSK, TRTTF, and TTTTF below 20% (Table 3), and have no association of the reaction of two out of three races. Previous studies also report race specificity in Aegilops species (Olivera et al., 2007;Scott et al., 2014). Since gene introgression from Aegilops species in the tertiary genepool is a long and laborious process, it is preferable to use accessions that carry stem rust resistance that is effective against multiple races. About 30% (396 accessions) were resistant against all the races evaluated, indicating the availability of potential sources of new and diverse stem rust resistance genes that could be very useful in wheat breeding programs. Most of these resistant accessions (84%) were from the tetraploid species Ae. geniculata, Ae. neglecta, and Ae. triuncialis. Additional studies are required to assess the diversity in these resistant accessions to allow the identification of donor accessions that are likely to contribute non-redundant stem rust resistance genes. Choosing resistant accessions from geographically diverse countries of origin and exhibiting different infection types for gene introgression is a first step to maximize the chances of capturing new and unique resistance genes (Anikster et al., 2005).
Sixty-five percent of the accessions evaluated in this study originated from Turkey or Greece, two countries having the largest numbers of Aegilops species. Turkey is known to be the center of diversity for Aegilops (Eig, 1929), and 17 out of the 23 Aegilops species have been identified in its territory (van Slageren, 1994). The nine species evaluated in this study are present in Turkey. The number and frequency of resistant accessions from Turkey and Greece (n = 285, 30.6%) from this study demonstrated that valuable sources of new genetic variation for stem rust resistance are present in these countries.
A prior knowledge on the inheritance of resistance in wild wheat relatives will facilitate alien gene introgression into wheat. We produced 17 biparental crosses to investigate the inheritance of TTKSK resistance. These populations will be further developed to map resistance genes and to develop closely linked markers within the wild species. Simple inheritance of stem rust resistance was found in most selected resistant accessions. Our result of a single dominant gene segregating in the Ae. umbellulata biparental F 2 population from a cross between PI 542369 and PI 298905 was confirmed in an F 3 population and mapped to chromosome 2U (Edae et al., 2016). A similar approach will be followed to characterize the resistance identified in this study. Two stem rust resistance genes were identified in three Ae. triuncialis resistant parents. Further studies are needed to characterize the effectiveness of each resistance gene. Multiple stem rust resistance genes with different resistance profile were reported in Ae. sharonensis (Olivera et al., 2008;Yu et al., 2017). A more complex inheritance of stem rust resistance with genes exhibiting epistatic effects was also observed in three Aegilops species. These results highlight the value of studying the genetics of stem rust resistance in the wild relative before attempting wide crosses for gene transferring.
Aegilops species in the tertiary genepool do not possess genome(s) homologous to the cultivated forms, and gene transfer through homologous recombination cannot be achieved with these species (Harlan and de Wet, 1971). Cytogenetic techniques such as irradiation and chemical treatments, production of synthetic amphiploids, use of gametocidal chromosomes, or Ph1 gene mutants may be required for gene introgression into the cultivated forms (Friebe et al., 1996;Zaharieva and Monneveux, 2006). However, the introgression of alien chromatin to substitute for homoeologous chromosome segments has the potential of a simultaneous introduction of deleterious DNA that can affect agronomic and quality traits of wheat (Feuillet et al., 2008;Wulff and Moscou, 2014). New sequencing technologies, like Genotyping-By-Sequencing, have allowed the development of genetic linkage maps in wild relatives of wheat with non-previous available markers, and the identification of closely linked markers that can facilitate the gene transfer process by reducing the introgressed alien chromatin segment into elite materials (Edae et al., 2016(Edae et al., , 2017. The sources of resistance identified from the tertiary genepool will also serve as targets for resistance gene cloning. Cloned genes and their delivery as transgenes in single or multiple resistance gene cassettes will completely resolve the linkage drag problem and ensure the effectiveness and durability of genes derived from more distant relatives of wheat (Wulff and Moscou, 2014). Today, new cloning techniques like mutational genomics (MutRenSeq) (Steuernagel et al., 2016) and association genetics with R gene enrichment sequencing (AgRenSeq) (Arora et al., 2018) allow a rapid and cheaper discovery and cloning of resistance genes. These technologies are opening new doors for fully exploiting the richness and diversity of wild relatives for wheat improvement.

AUTHOR CONTRIBUTIONS
PO and YJ were involved in the experimental design and manuscript preparation. PO performed the experiments and completed the data analysis. MR was involved in manuscript preparation and revision.