Marker Assisted Transfer of Stripe Rust and Stem Rust Resistance Genes into Four Wheat Cultivars

: Three rust diseases namely; stem rust caused by Puccinia graminis f. sp. tritici (Pgt), leaf rust caused by Puccinia triticina (Pt), and stripe rust caused by Puccinia striiformis f. sp. tritici (Pst), are the most common fungal diseases of wheat ( Triticum aestivum L.) and cause signiﬁcant yield losses worldwide including Australia. Recently characterized stripe rust resistance genes Yr51 and Yr57 are e ﬀ ective against pre- and post-2002 Pst pathotypes in Australia. Similarly, stem rust resistance genes Sr22, Sr26, and Sr50 are e ﬀ ective against the Pgt pathotype TTKSK (Ug99) and its derivatives in addition to commercially important Australian pathotypes. E ﬀ ectiveness of these genes make them good candidates for combining with known pleiotropic adult plant resistance (PAPR) genes to achieve durable resistance against three rust pathogens. This study was planned to transfer rust resistance genes Yr51, Yr57, Sr22, Sr26, and Sr50 into two Australian (Gladius and Livingston) and two Indian (PBW550 and DBW17) wheat cultivars through marker assisted selection (MAS). These cultivars also carry other rust resistance genes: Gladius carries Lr37 / Yr17 / Sr38 and Sr24 / Lr24 ; Livingston carries Lr34 / Yr18 / Sr57 , Lr37 / Yr17 / Sr38, and Sr2 ; PBW550 and DBW17 carry Lr34 / Yr18 / Sr57 and Lr26 / Yr9 / Sr31 . Donor sources of Yr51 (AUS91456), Yr57 (AUS91463), Sr22 ( Sr22 / 3*K441), Sr26 ( Sr26 WA1), and Sr50 (Dra-1 / Chinese Spring ph1b / 2 / 3* Gabo) were crossed with each of the recurrent parents to produce backcross progenies. Markers linked to Yr51 ( sun104 ), Yr57 ( gwm389 and BS00062676 ), Sr22 ( cssu22 ), Sr26 ( Sr26#43 ) , and Sr50 ( Sr50 - 5p-F3, R2 ) were used for their MAS and markers csLV34 ( Lr34 / Yr18 / Sr57 ) , VENTRIUP-LN2 ( Lr37 / Yr17 / Sr38 ), Sr24#12 ( Sr24 / Lr24 ), and csSr2 ( Sr2 ) were used to select genes present in recurrent parents. Progenies of selected individuals were grown and selected under ﬁeld conditions for plant type and adult plant rust responses. Final selections were genotyped with the relevant markers. Backcross derivatives of these genes were distributed to breeding companies for use as resistance donors.


Introduction
Wheat (Triticum aestivum L.) is one of the major food crops in the world and it is continuously threatened by several biotic stresses. Among the significant biotic stresses, rust diseases namely: Stem rust caused by Puccinia graminis f. sp. tritici (Pgt), leaf rust caused by Puccinia triticina (Pt), and stripe rust caused by Puccinia striiformis f. sp. tritici (Pst), are the most common diseases of wheat worldwide. Rust diseases have been causing significant economic losses to wheat production in Australia. Stem rust epidemic in the 1973-74 season caused economic losses of A$200 to 300 million [1]. Epidemics of stem rust and leaf rust in wheat crop in Western Australia in 1999 were estimated to result in A$50 million in losses [2]. Introduction of a new Pst pathotype in Western Australia has been continuously causing yield losses since its detection in 2002 [3,4].
The importance of genetic resistance to control rust diseases was first demonstrated by Biffen [5]. Since then, breeding for rust resistance is considered as the most preferred management strategy for rust control. This approach has been successful in reducing losses in Australian wheat growing regions, where rust resistant-wheat cultivars protected grain industry from epidemics since 1960s. The Australian grain industry was estimated to save A$289 million annually from stem rust, stripe rust, and leaf rust infections through resistance breeding efforts [6].
Long lasting control of rust diseases has been achieved through pyramiding resistance genes in a single genotype [7]. It is however difficult to combine more than two genes in conventional breeding programs, especially when genes to be deployed exhibit resistance against all available rust isolates and produce similar infection types. The detection of marker-gene associations has made marker-assisted gene pyramiding an achievable target in breeding programs [8].
Marker assisted selection (MAS) allows enrichment of positive alleles that condition economic traits at early generations in breeding programs and it is underpinned by the availability of tight marker-trait associations that are highly reproducible, cost-effective and user friendly [9,10]. The various approaches presently being used for MAS include marker-assisted backcrossing (MABC), forward selection, MAS-aided doubled haploid (DH) production and enrichment of positive alleles among F 2 selections.

Plant Materials
Wheat cultivars Gladius and Livingston (Australian), and PBW550 and DBW17 (Indian) were chosen as recurrent parents for transfer of stripe rust and stem rust resistance genes. Gladius is a spring wheat cultivar that was cultivated in South Australia and is moderately resistant (MR) to stem rust and moderately resistant to moderately susceptible (MR-MS) to stripe rust. Gladius carries all stage resistance (ASR) genes Lr37/Yr17/Sr38 and Sr24/Lr24 introgressions from Aegilops ventricosa and Thinopyrum elongatum, respectively. Livingston is cultivated in northern New South Wales (NSW) and Queensland and shows MR-MS level of resistance against both stripe rust and stem rust under field conditions. It carries adult plant resistance (APR) genes Lr34/Yr18/Sr57 and Sr2 in addition to ASR gene Lr37/Yr17/Sr38. PBW550 and DBW17 are high yielding cultivars widely cultivated in North Western Plain Zone (NWPZ) in India and are susceptible to stripe rust. Both PBW550 and DBW17 carry Lr34/Yr18/Sr57 and Lr26/Yr9/Sr31. Seed of Indian cultivars was obtained as a part of Australia-India collaborative project funded by the Australian Centre for International Agricultural Research (ACIAR). Donor sources carrying Yr51, Yr57, Sr22, Sr26, and Sr50 were obtained from the Australian Cereal Rust Control Program (ACRCP) staff (Table 1).

Greenhouse Tests
Four F 1 seeds and 10 to 20 seeds from each backcross (BC 1 F 1 , BC 2 F 1 ) were sown in 9 cm pots filled with a potting mixture (pine bark and river sand in the ratio of 2:1). Recurrent parents (Gladius, Livingston, PBW550, and DBW17) and donor parents AUS91456 (Yr51), AUS91463 (Yr57), Sr22/3*K441 (Sr22), Sr26 WA1 (Sr26), and Dra-1/Chinese Spring ph1b/2/3* Gabo (Sr50) were included as controls in respective experiments. Cultivar Morocco was used as the susceptible control. After sowing, pots were maintained in rust-free microclimate rooms at 20 • C. Twenty grams of water-soluble fertilizer Aquasol ® was dissolved in 10 liters of tap water and applied to 100 pots. A single application of nitrogenous fertilizer urea was applied at the same rate as Aquasol ® to the seven-day old seedlings. Inoculations were performed at the two-leaf growth stage by suspending urediniospores in light mineral oil (Isopar L, 5 mg spores per 10 mL oil) using a hydrocarbon propellant pressure pack. For stripe rust, the inoculated seedlings were placed under polythene hoods in water filled steel trays and incubated in a dark room at 9-12 • C for 24 h. Stem rust-inoculated seedlings were incubated under polythene hoods in water filled steel trays in a greenhouse microclimate room at 18-20 • C for 48 h. Following incubation, seedlings were moved to temperature-and irrigation-controlled greenhouse rooms. Post-inoculation temperatures of 27 • C and 17 • C were used for stem rust and stripe rust, respectively. Rust response assessments were made 14 days after inoculation according to scales described in Bariana and McIntosh [17]. Rust resistant selections were made by comparing infection types produced by backcross derivatives with donor sources.

DNA Isolation and Quantification
Total genomic DNA was isolated from seedlings of all test material following the procedure described in Bansal et al. [18]. Quality and quantity of DNA was estimated using agarose (1%) gel and Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies Inc. Wilmington, NC, USA), respectively. Final dilutions of 30 ng/µl concentration were prepared with double distilled autoclaved water.

PCR Amplification and Gel Electrophoresis
For PCR amplification of SSR and STS markers, assays were performed in 10 µL reaction mixture containing 0.2 mM dNTPs, 1xImmolase PCR buffer (Bioline), 0.2 mM each of forward and reverse primer, 30 ng of genomic DNA and 0.2 U of Immolase DNA polymerase (Bioline). PCR amplifications were performed according to published protocols for respective markers ( Table 2).

Marker Assisted Selection
Molecular markers linked with Yr51, Yr57, Sr22, Sr26, and Sr50 were tested on respective backcross material to select seedlings positive for the target gene. Marker assisted selection was performed at each backcross generation. BC 2 F 2 s (Yr51 and Yr57) and BC 1 F 2 s (Sr22, Sr26, and Sr50) were tested with markers linked with Lr34/Yr18/Sr57, Sr24/Lr24, Sr2, and Lr37/Yr17/Sr38 to select resistance genes carried by recurrent parents. Markers details are listed in Table 2.

Chi-Squared Analyses
Goodness-of-fit of the observed segregation ratios with the expected genetic ratios of phenotypic and genotypic data of BCF 2 derivatives was tested using chi-squared (χ 2 ) analysis.

Marker Assisted Transfer
Wheat cultivars Gladius, Livingston, PBW550, and DBW17 were crossed with each of the five donors Yr51, Yr57, Sr22, Sr26, and Sr50 to produce F 1 hybrids which were in turn backcrossed with the respective recurrent parent. Marker assisted selection was performed at each generation. Figure 1 illustrates the crossing scheme used for the transfer of Yr51 into Gladius as an example. Similar strategy was applied to the transfer of Yr57, whereas single back cross was used for transfer of Sr22, Sr26, and Sr50 into each of the four wheat cultivars. Infection types produced by donor and recurrent parents are presented in Table 3. Details of transfer of each gene into four cultivars are discussed below:

Yr51
Yr51-linked dominant marker sun104 was genotyped on F1 hybrids of Yr51 and recurrent parents Gladius, Livingston, PBW550 and DBW17 to ensure identity of the cross. Genotyping of BC1F1

Yr51
Yr51-linked dominant marker sun104 was genotyped on F 1 hybrids of Yr51 and recurrent parents Gladius, Livingston, PBW550 and DBW17 to ensure identity of the cross. Genotyping of BC 1 F 1 seedlings enabled identification of plants carrying the Yr51-linked allele sun104 225 . Similarly, BC 2 F 1 seedlings were also selected through MAS and those carrying the linked marker allele were advanced to BC 2 F 2 stage. BC 1 F 1 and BC 2 F 1 plants showed 1:1 segregation (sun104 225 : sun104 null ) at the sun104 locus (Table 4).

Yr57
Yr57-linked markers gwm389 and BS00062676 were co-dominant, and the Yr57 source carried gwm389 150 and BS00062676 'A:A' alleles. The F 1 hybrids involving Yr57 (AUS91463) and recurrent parents Gladius, Livingston, PBW550, and DBW17 were genotyped to ascertain the presence of Yr57 and to exclude the possibility of selfing. DNA from 11 to 20 BC 1 F 1 seedlings was isolated and tested using markers gwm389 and BS00062676. BC 1 F 1 seedlings carrying the Yr57-linked alleles were selected and backcrossed to respective recurrent parent to produce BC 2 F 1 plants. Again Yr57-carrying BC 2 F 1 seedlings were selected using markers and advanced to produce BC 2 F 2 generation. Segregation of Yr57-linked markers gwm389 and BS00062676 on BC 1 F 1 and BC 2 F 1 plants fitted to 1:1 ratio (Table 4).

Sr22
The Sr22-linked marker cssu22 is co-dominant and the F 1 plants from all four crosses carried Sr22-linked cssu22 237 allele together with recurrent parent alleles. The BC 1 F 1 seedlings were either heterozygous at the cssu22 locus or carried the respective recurrent parent allele (Table 2). Heterozygous BC 1 F 1 plants were grown to produce BC 1 F 2 generation. Marker cssu22 showed monogenic segregation when tested on backcross derivatives (Table 4).

Sr26
The Sr26-linked marker Sr26#43 follows dominant inheritance and amplifies 270 bp in genotypes carrying this gene. The F 1 plants were genotyped to ascertain the presence of Sr26 and were crossed to the respective recurrent parent. The BC 1 F 1 seedlings carrying the Sr26-linked allele in all crosses were selected and advanced to BC 1 F 2 generation. Monogenic segregation at the Sr26#43 locus was observed among BC 1 F 1 and BC 1 F 2 plants (Table 4).

Sr50
The Sr50-linked marker Sr50-5p-F3, R2 amplified 470 bp product in genotypes carrying Sr50 and no product was observed among recurrent parents indicating its dominant nature. Crosses of Dra-1/Chinese Spring ph1b/2/3* Gabo with all four recurrent parents were genotyped to ensure the presence of Sr50 in F 1 plants and crossed with the respective recurrent parent. BC 1 F 1 seedlings carrying the Sr50-linked allele were selected and advanced to produce BC 1 F 2 families. The marker Sr50-5p-F3, R2 showed monogenic segregation (Table 4).

Marker Assisted Selection of Gene Combinations Among BCF 2 Families
Bulked seed from marker positive BCF 2 families (BC 2 F 2 for Yr51 and Yr57, and BC 1 F 2 for Sr22, Sr26, and Sr50) were space planted in the field during 2014 season and inoculated artificially with Pst, Pt, and Pgt pathotypes. While selection for triple rust resistance and agronomic traits were attempted, DNA from 35 plants from each cross was isolated to confirm presence of the target gene. Additional resistance genes carried by recurrent parents are listed in Table 5. These included Lr37/Yr17/Sr38, Lr34/Yr18/Sr57, Sr24/Lr24, and Sr2. Rust resistant and agronomically better backcross progenies of Gladius and Livingston were tested with Lr37-linked marker VENTRIUP-LN2 to select combination of Lr37/Yr17/Sr38 with the gene targeted for backcrossing. The Sr24/Lr24-linked marker Sr24#12 was used to select plants carrying Sr24/Lr24 in combination with the target gene in crosses of Gladius. The Sr2-linked marker was monomorphic among parents and therefore progenies carrying Sr2 were selected using seedling chlorosis linked with this gene. Backcross progenies of Livingston, PBW550, and DBW17 were tested with Lr34-linked marker csLV34 to identify genotypes carrying Lr34/Yr18/Sr57 and the target gene. Table 5 lists different combinations of rust resistance genes identified in backcross progenies of four cultivars.

Discussion
Diverse breeding methodologies are being used for selection of rust resistance in breeding programs. Backcrossing has been considered as an effective method for transferring resistance genes to agronomically superior genotypes as it allows capturing of recurrent parental background. Single backcross approach was used to achieve combinations of APR genes for resistance to rust diseases in wheat [22][23][24].
Gene-linked markers can be used to select and confirm gene combinations in segregating populations. Numerous reports about useful marker-trait associations for wheat are available. However, the number of publications detailing the practical application of MAS in wheat breeding is limited [25]. Molecular markers can assist in phenotype-neutral selection of traits that are difficult to select phenotypically, environmentally unstable, or are not cost-effective to assess [26]. Being a non-invasive and seed quantity independent alternative to phenotypic based selection, MAS can be applied at any point in a breeding program.
Close marker-trait associations increase the efficiency of MAS. This study demonstrated the usefulness of markers linked with rust resistance genes Yr51, Yr57, Sr22, Sr26, and Sr50. MAS was successfully employed to select combination of these genes with all-stage resistance and/or adult plant resistance genes present in recurrent parents. The MAS approach is even more useful in the case of genes that show recessive inheritance. In this study, backcrossing of the recessive gene Yr51 was successfully achieved.
DNA isolation is the most expensive component of MAS and once DNA of a sample is available then markers linked with a range of traits can be genotyped to select a total genotypic package. However, larger population sizes are required to successfully select desirable number of genotypes carrying positive alleles for multiple trait loci. Bonnet et al. [27] estimated size of population for MAS of different number of loci using different breeding methodologies. A much lower number is needed in the case of doubled haploid/recombinant inbred line and single backcross methods.
Discovery of genetically diverse rust resistance loci and their close association with markers has exponentially increased in the last decade through advancement in genotyping technologies [28]. Validation of marker-trait linkages as demonstrated in this study will ensure assembly of the best training populations for genomic selection in wheat to beat environmental influences to increase selection efficiencies in breeding programs. The improved versions of four cultivars have been distributed to Australian and Indian wheat breeders involved in the Australia-India bilateral project funded jointly by the ACIAR and the Indian Council of Agricultural Research (ICAR) for use in respective breeding programs. This study resulted in improvement of stripe rust and stem rust resistance in wheat cultivars when evaluated in Australia and India. These results covered foreground and background selection to deliver triple rust resistant cultivars to farmers.