Introduction

Stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is the most important foliar disease of wheat worldwide. First reported by Gaddin in Europe in 1777 (Eriksson and Henning 1896), it has since been reported in 60 different countries (Stubbs 1985; Line 2002; Li and Zeng 2003). Stripe rust can cause up to 100 % yield losses on susceptible cultivars; however, in commercial production environments yield losses typically range from 10 to 70 % depending upon the cultivar, inoculum pressure and climatic conditions. Stripe rust was first detected in Australia in 1979 on wheat cultivar Zenith (O’Brien et al. 1980). Following its introduction, devastating stripe rust epidemics occurred in southern New South Wales, Australia. These epidemics reduced grain yield up to 84 %, kernel mass 43 % and kernel number 72 % (Murray et al. 1995). Subsequent localized stripe rust epidemics have occurred in certain seasons and regions of eastern Australia. The wheat growing regions of western Australia remained free from stripe rust due to physical separation by the Nullabor Desert and predominant west to east air movement until the migration of another exotic Pst pathotype, 134 E16A+ (WA pathotype), in 2002 (Wellings et al. 2003). It reached eastern Australia in 2003, and unlike the pathotypes prevalent in eastern Australia prior to 2003, the WA pathotype appeared to have defeated at least a component of adult plant resistance in certain Australian wheat cultivars (Bariana et al. 2010).

Researchers across the world are engaged in the identification and characterization of new sources of resistance to combat the continuously evolving rust pathogens. Release of rust-resistant cultivars is the most economical and environmentally cautious strategy to control rust diseases. The concept of growing rust resistant cultivars dates back to the early twentieth century (Biffen 1905). Breeding for rust resistance in wheat has been a major objective in Australian breeding programs and involves identification and deployment of durable sources of resistance (McIntosh 1998). Resistance based on race-specific genes (major genes) has not been durable as the stripe rust pathogen can evolve to render such genes ineffective. On the contrary, resistance based on adult plant resistance (APR) genes is assumed to be durable (Johnson 1988). APR genes slow the rate of rust development and are often referred to as ‘slow-rusting’ genes. Deployment of combinations of two or more APR genes in commercial cultivars is essential to achieve acceptable levels of resistance (Singh and Rajaram 1994; Bariana and McIntosh 1995).

Australian wheat cultivar Sunco (SUN9E27*4/3Ag14//WW15/3/3*Cook) exhibits a moderately resistant response against the prevalent Pst pathotypes (Kaur and Bariana 2010). These workers tested Sunco/2*Avocet S-derived BC1F2 families and found the involvement of three APR genes. Resistant and susceptible sibs were selected from families showing monogenic segregation. Sunco is reported to carry Yr18 (Kolmer et al. 2008) and YrCK (Bariana et al. 2001). Based on genotyping with the Yr18-linked marker and high-temperature rust screening, candidate lines carrying the third APR gene from Sunco/2*Avocet S-derived BC1F2 families were identified. In this study we report the molecular mapping and characterization of the third APR gene in one of the Sunco/2*Avocet S resistant derivatives that did not carry Yr18 and/or YrCK.

Materials and methods

A Sunco/2*Avocet S-derived BC1F2 family (number 153) that showed monogenic segregation was used in this study. The resistant plant from 153.11 (SA65) was crossed with a susceptible sib 153.6 (SA67), and an F6 recombinant inbred line (RIL) population (123 RILs) was developed. This RIL population was named SA65/SA67. The population development process is shown in Fig. 1.

Fig. 1
figure 1

Scheme for generating single gene line RIL population

Full size image

The SA65/SA67 F6 population was sown in the field during the cropping seasons 2013, 2014 and 2015 as 1-m rows at the Plant Breeding Institute, Cobbitty (PBIC) experimental site Lansdowne. After every five experimental rows, one row of susceptible cultivar Morocco was sown to get a uniform spread of stripe rust inoculum. Stripe rust epidemic was initiated by inoculating the susceptible spreader rows with Pst pathotype 134 E16A+Yr17+Yr27+ following the procedure described in Bariana et al. (2010). Stripe rust response was assessed on a 1–9 scale (Bariana et al. 2007), where 1 = very resistant, 2 = resistant, 3 = resistant to moderately resistant, 4 = moderately resistant, 5 = moderately resistant to moderately susceptible, 6 = moderately susceptible, 7 = moderately susceptible to susceptible, 8 = susceptible and 9 = very susceptible.

Seedling tests

SA65/SA67 F6 population was tested at the two-leaf stage against Puccinia triticina pathotype 104-2,3,6,7 (culture no. 231) according to Bariana and McIntosh (1993) to record segregation for Lr24.

Molecular mapping

DNA from each RIL and both parents (SA65 and SA67) was extracted using a modified CTAB method (Bansal et al. 2014a). DNA from eight resistant and eight susceptible RILs were subjected to selective genotyping using an Illumina iSelect 90 K Infinium SNP genotyping array (Wang et al. 2014). SNPs showing linkage between the resistant and susceptible RILs were converted into Kompetitive Allele-Specific Primers (KASP) and genotyped on the parents using KASP technology (LGC genomics UK). KASP assays were performed following the protocol described by LGC Genomics using KASP mix, which contains universal FRET (fluorescence resonance energy transfer) cassettes (FAM and HEX), ROX™ passive reference dye, Taq polymerase, free nucleotide, and MgCl2 in an optimized buffer. End-point fluorescent images were visualized using the CFX96 Touch™ real-time PCR detection system (BioRad, USA) and the data analyzed using Bio-rad CFX Manager Software (BioRad, USA).

Eighteen SSR markers representing the target region on chromosome 3D were also tested on parents using the touchdown (depending upon the annealing temperature of primers) profile according to Bansal et al. (2014b). The markers that showed polymorphism between parents were genotyped on the RIL population and included in the map. The Lr24/Sr24 linked marker Sr24#12 (Mago et al. 2005) was also used to confirm the presence of these genes in addition to seedling leaf rust tests.

Statistical analysis and genetic mapping

Chi-squared analysis was performed to assess deviations of observed segregations from the expected genetic ratios. MapManager QTXb20 (Manly et al. 2001) was used to construct the genetic linkage map. The Kosambi map function (Kosambi 1943) was used to convert recombination fractions to centi-Morgans. The linkage map was drawn using Map Chart version 2.2 (Voorrips 2002).

Results

Inheritance of resistance

Parental genotypes SA65 and SA67 were susceptible at the seedling stage (infection type 3+) and exhibited adult plant stripe rust responses 6–7 and 8–9, respectively, in the field. The RIL population showed stripe rust response variation ranging from 6 to 9. RILs were grouped in two classes; homozygous resistant (scores 6–7) and homozygous susceptible (scores 8–9). Chi-squared analysis of adult plant stripe rust data conformed to monogenic segregation (66 homozygous resistant: 52 homozygous susceptible, χ 21:1  = 1.66, non-significant at P = 0.05 and 1d.f.). The stripe rust resistance locus was temporarily named YrSA3.

Molecular mapping

One hundred and twelve SNP markers from chromosome 3D showed genetic association with YrSA3 in a selective genotyping experiment. In the consensus 90 K SNP map (Wang et al. 2014), the linked SNPs spanned the chromosome 3D interval from 134.8 to 152.83 cM (total map length 166.16 cM). IWB72555 was located at the 134.81 cM position, whereas all other SNPs were mapped in the 142.31 to 152.83 cM interval. The map locations of nine SNP loci were not known. As most of the SNPs represented the same consensus map position, KASP assays were designed for 60 SNPs (Supplementary Table S1). The KASP assays were performed on parents and controls to confirm their specificity before being genotyped on the entire RIL population.

In addition, 18 SSR markers previously mapped on chromosome 3D were genotyped on the parental lines. These included cfd4, cfd141, cfd152, cfd201, cfd219, gdm8, gwm3, gwm66, gwm71, gwm114, gwm161, gwm183, gwm191, gwm314, gwm383, gwm497, gwm645, and wmc674 (http://wheat.pw.usda.gov/cgi-bin/westsql/map_locus.cgi). Markers cfd152, gdm8, gwm114, gwm314, and gwm497 showed polymorphism between parents (SA65 and SA67) and were genotyped on the entire population. Markers cfd152, gdm8, gwm314, and gwm497 segregated independently of stripe rust resistance, whereas gwm114 revealed segregation at two loci: gwm114a and gwm114b. Allele gwm114b exhibited linkage with YrSA3. The gwm114b allele was previously mapped on the long arm of chromosome 3D in the Opata/Synthetic and Arina/Forno RIL populations (https://ccg.murdoch.edu.au/cmap/ccg-live/cgi-bin/cmap/feature?feature_acc=403114).

Screening for Lr24

Sunco is known to carry the linked rust resistance genes Lr24 and Sr24 (McIntosh et al. 1976) on chromosome 3DL. Parental lines SA65, SA67, and Sunco were tested in the seedling stage against Pt pathotype 104-2,3,6,7. SA67 produced infection type (IT) 0; similar to Sunco, whereas SA65 was scored susceptible (IT3+). The entire SA65/SA67 RIL population was screened and monogenic segregation at the Lr24 locus was observed (Table 1). These results were further confirmed using the Sr24-linked marker Sr24#12 (Mago et al. 2005). Sr24#12 is a dominant marker and produces a 500bp amplicon in lines carrying Lr24/Sr24. Joint frequency distribution analysis of seedling leaf rust and adult plant stripe rust data (Table 1) revealed linkage between Lr24/Sr24 and YrSA3. Four recombinants between Lr24/Sr24 and YrSA3 were observed. Two recombinants carried Lr24/Sr24 and YrSA3 in combination and possessed gwm114b 134bp allele. The other two recombinants lacked both loci and carried gwm114b 132bp allele. Lr24/Sr24 was incorporated into the map.

Table 1 Joint frequency distribution of YrSA3 and Lr24/Sr24 among SA65/SA67 F6 RIL population
Full size table

Map construction

The final linkage map comprising one SSR, 10 SNPs, and Lr24/Sr24 is shown in Fig. 2. The stripe rust resistance gene YrSA3 was located 1.8 cM proximal to all markers. These results confirmed the location of YrSA3 in the long arm of chromosome 3D.

Fig. 2
figure 2

Genetic linkage map of chromosome arm 3DL showing location of stripe rust resistance gene Yr71 (YrSA3)

Full size image

Genotyping of linked markers on Australian cultivars

A set of 74 Australian wheat cultivars (Table 2) was used for assessing the robustness of gwm114b and 10 SNP markers that co-segregated with linked rust resistance genes Lr24 and Sr24 and showed close association with YrSA3. Marker gwm114b amplified 134bp products in 57 cultivars that lacked Lr24/Sr24. A 132bp product was observed in 17 Lr24/Sr24 carrying cultivars indicating the robustness of this marker in the detection of these linked genes. The 132bp amplicon of marker gwm114b appears either to be Agropyron specific or positioned very close to the translocation breakpoint. Seven KASP markers that co-segregated with Lr24/Sr24 did not detect the presence/absence of these loci accurately and resulted in a high level of misclassification. Three KASP markers (KASP_16434, KASP_17207 and KASP_20836), however, amplified alleles alternate to that linked with Lr24/Sr24 in 57 cultivars lacking these genes. Of the 17 Lr24/Sr24 carrying cultivars 12, 14, and 15 were accurately identified by KASP_20836, KASP_17207, and KASP_16434, respectively (Table 2).

Table 2 List of cultivars tested with linked markers and their genotypic status
Full size table

Discussion

This study followed an earlier investigation by Kaur and Bariana (2010) in which they reported tri-genic inheritance of adult plant stripe rust resistance among Sunco/2*Avocet S-derived BC1F2 families. Here, we used the SA65/SA67 RIL population derived from a Sunco/2*Avocet S BC1F2 family 153 (Fig. 1), which was shown not to carry the previously known APR genes Yr18 (Kolmer et al. 2008) and YrCK (Bariana et al. 2001) from Sunco, to genetically map the remaining APR gene. Adult plant stripe rust response assessments across three seasons on SA65/SA67 F6 population showed monogenic segregation, and the resistance locus was temporarily designated YrSA3. Molecular mapping using a 90 K Infinium SNP genotyping array (Wang et al. 2014) located YrSA3 in the long arm of chromosome 3D. Ten KASP markers, Sr24#12, and gwm114b clustered together at 1.8 cM distal to YrSA3. The Agropyron elongatum translocation carrying rust resistance gene Lr24/Sr24 was responsible for clustering of these markers.

Sunco carries linked stem rust and leaf rust resistance genes Lr24 and Sr24. The Sunco*2/Avocet-derived resistant line SA65 possessed Avocet S alleles for all markers and lacked Sr24/Lr24, whereas the susceptible line SA67 carried Sunco alleles and Sr24/Lr24 (Table 3). This created confusion about the donor of YrSA3. The derivation of YrSA3 in SA65 from Sunco can be explained based on double recombination between this gene and the flanking markers (gwm114b and 10 SNPs) or an allosyndetic event during the generation of BC1F2 population. Since Sr24 and Lr24 are located on an Agropyron elongatum wheat translocation T3DS.3DL-3Ae#1 (Smith et al. 1968) in which 1.25 µm of 3Ae#1 replaced 1.38 µm of 3DL (Jiang et al. 1994), allosyndesis can be expected. We tried additional markers located on both arms of chromosome 3D, but all markers were monomorphic due to derivation of SA65 and SA67 form the same BC1F2 family. The second scenario is that YrSA3 was contributed by Avocet S. The contribution of YrSA3 by Sunco through double recombination or an allosyndetic event seems more plausible as Avocet S showed susceptible stripe rust response score 9 in comparison to YrSA3 score 6. Of the four recombinant RILs, two carried Sr24/Lr24 and YrSA3 and the other two lacked both genes (Table 3). This indicates that frequency of recombination is 1.75 %.

Table 3 Allelic comparisons of recombinant lines and parental genotype for linked markers and rust resistance genes
Full size table

Stripe rust resistance genes Yr45 (Li et al. 2011), Yr49 (McIntosh et al. 2011) and Yr66 (McIntosh et al. 2013) were mapped in chromosome 3D previously. Yr45 was mapped in long arm, while other two genes were mapped in short arm of chromosome 3D. Yr45 was flanked by RGAP markers wgp118 and wgp115 and two SSR markers wmc656 and barc6 were mapped 11.7 and 12.6 cM proximal to Yr45. We tried to compare the SSR markers wmc656, barc6 and gwm114b reported by Li et al. (2011) and present study, however, none of the map carries all three markers together. Marker wmc656 was mapped at 75 cM position (Somers et al. 2004) of the 110 cM of total map length, whereas gwm114b has been mapped at the distal region in Arina/Forno and Opata/Synthetic populations (https://ccg.murdoch.edu.au/cmap/ccg-live/cgi-bin/cmap/map_details?ref_map_set_acc=10;ref_map_accs=164;highlight=12503). These results reflect that Yr45 and YrSA3 were positioned at different locations. Comparison of location and resistance expression led us to conclude that these two genes are different. The genetic association with Lr24 and 3DL markers clearly demonstrated the uniqueness of YrSA3, and based on these results it was permanently named Yr71.

All 10 KASP markers and gwm114 were tested on a set of Australian cultivars differing for Lr24/Sr24 (Table 2). All the lines carrying Lr24/Sr24 produced Sunco allele (132bp) when tested with gwm114 and those lacking these genes amplified the alternate allele (134bp). Seven KASP markers showed variable results indicating the non-robustness of these markers to detect Lr24/Sr24. Their segregation among SA65/SA67 population could be due to derivation of parental lines from the same BC1F2 family. The detection of large number of markers linked with the target gene takes place due to high level of linkage disequilibrium (owing to less numbers of recombination cycles) among bi-parental populations. That is why validation on independently developed cultivars is essential to identify robust trait–marker associations. KASP_16434 was the most robust SNP marker to detect Lr24/Sr24 and YrSA3, followed by KASP_17207 and KASP_20836. The detection of non-Lr24/Sr24-linked haplotype of KASP_16434, KASP_17207, and KASP_20836 in two, three, and five cultivars, respectively, carrying this Agropyron-derived gene combination suggested the location of these markers on the common wheat segment of chromosome 3DL. We suggest that recombinants (RIL-23 and RIL-67) should be used as a source of Yr71 for simultaneous transfer of Lr24/Sr24. Although development of KASP markers for Lr24/Sr24 was not an objective of this study, markers KASP_16434, KASP_17207 and KASP_20836 provide a co-dominant alternative to the ‘breeder friendly’ and robust dominant marker developed by Mago et al. (2005). Alternatively, the co-dominant SSR marker, gwm114b, demonstrated to be the most robust for the detection of Lr24/Sr24, can be used for the detection of this linked locus among diverse plant genetic resources and for marker-assisted selection in breeding programs.

Yr71 is an APR gene and can be combined with other major or minor stripe rust resistance genes to achieve low rust severity and durability. Molecular markers for numerous all stage resistance (listed in Bariana et al. 2007) and APR genes Lr34/Yr18, Lr67/Yr46 and Yr36 (Lagudah et al. 2006; Forrest et al. 2014; Uauy et al. 2005) are available and can be used for gene stacking.