Introduction

Stem rust, caused by the basidiomycete Puccinia graminis f. sp. tritici (Pgt), is one of the most serious diseases affecting global wheat (Triticum aestivum L., 2n = 6x = 42, AABBDD) production. During the early-to-mid twentieth century, Pgt caused large-scale wheat yield losses in the United States. In response, efforts were made to breed Pgt-resistant wheat varieties, and to eliminate the non-native alternative host, barberry (Berberis vulgaris L.), from the United States (Singh et al. 2015). In 1998, an isolate of Pgt from Uganda (Ug99) was found to be virulent to the widely deployed resistance gene Sr31 (Pretorius et al. 2000) and pathotyped to the standard race nomenclature designation TTKSK (Singh et al. 2011). In subsequent years, variants from the Ug99 race lineage developed additional virulence to Sr24 (TTKST) and Sr36 (TTTSK), placing an even greater number of wheat varieties at risk (Jin et al. 2008, 2009). Currently, thirteen races belonging to the Ug99 race group have been identified in a region extending from South Africa to Iran (Fetch et al. 2016). The majority of wheat varieties planted globally are susceptible to the Ug99 race group (Singh et al. 2015).

Committed efforts to identify novel sources of host resistance to Pgt in wheat and its wild wheat relatives have resulted in 27 numerically designated Sr genes available for stem rust resistance breeding efforts (Yu et al. 2014). Wild relatives of wheat are a valuable source of stem rust resistance, however, restricted recombination between wheat chromosomes and introgressions from anura genomes often require extensive chromosome engineering to generate genotypes with minimal linkage drag (Mago et al. 2009; Qi et al. 2011; Liu et al. 2011). The wheat D-genome donor species, Aegilops tauschii Coss., has homology to the D genome of wheat and has been a valuable source of disease resistance genes (Gill and Raupp 1987; Cox 1997; Cox et al. 1991; Olson et al. 2013b). Aegilops tauschii has contributed the Ug99 resistance genes Sr33, Sr45, Sr46, SrTA1662, SrTA10171, and SrTA10187 (Olson et al. 2013a, b; Periyannan et al. 2013, 2014; Yu et al. 2015b). Gene introgression from Ae. tauschii can be achieved by crossing tetraploid wheat (Triticum turgidum L., 2n = 4x = 28, AABB) with diploid Ae. tauschii (2n = 2x = 14, DD) to produce synthetic hexaploid wheat (McFadden and Sears 1944), or by direct hybridization between hexaploid wheat and Ae. tauschii as described by Gill and Raupp (1987). Using either method, genes can be transferred to wheat through normal meiotic recombination.

The Ae. tauschii accession TA10187 exhibited seedling resistance to multiple Pgt races (Rouse et al. 2011) and the resistance was transferred to hexaploid wheat (Olson et al. 2013b). A single gene, provisionally named SrTA10187, was mapped to the distal region of chromosome 6DS. Multiple stem rust resistance genes are located in this region including Sr5 (Sears et al. 1957; McIntosh et al. 1995), Sr42 (Ghazvini et al. 2012; Gao et al. 2015), SrCad (Hiebert et al. 2011; Kassa et al. 2016), and SrTmp (Hiebert et al. 2015; Lopez-Vera et al. 2014).

The objectives of this study were to: (1) fine map the SrTA10187 locus on chromosome 6DS, (2) identify candidate resistance genes, and (3) develop single nucleotide polymorphism (SNP) markers for SrTA10187. Mapping the resistance locus to a shorter genetic region will improve breeding accuracy for SrTA10187, and identification of candidate resistance genes will expedite future map-based cloning efforts. Marker resources developed in this study will facilitate introgression and gene pyramiding of SrTA10187 into elite wheat varieties using marker-assisted selection.

Methods and materials

Plant materials

The high-resolution mapping population used in this study comprised 1060 individuals from six BC3F2 families: U6897-1, U6897-2, U6897-3, U6897-4, U6897-5, and U6897-6. The population was developed by selfing six BC3F1 plants made from a cross of a single stem rust-resistant BC2F1 plant (Olson et al. 2013b) to the recurrent parent KS05HW14. All BC3F2 families segregated for previously mapped SSR loci (Xcfd49 and Xbarc173) linked to SrTA10187 (Olson et al. 2013b).

Stem rust resistance phenotyping

Seedling stem rust phenotyping was done using a Michigan-collected isolate of Pgt race QFCSC. Seedlings were inoculated at the two-leaf stage with Pgt urediniospores suspended in Soltrol 170 isoparaffin oil (Chevron Philips Chemical Company LP, The Woodlands, TX) using an airbrush. Plants were then placed into a dew chamber held at 20 °C and 100 % relative humidity for 16 h. At 14 days post-inoculation, the first leaf of each plant was scored for Pgt disease resistance using the 0-4 Stakman scale (Stakman et al. 1962). A χ2 goodness-of-fit test was performed with Pgt infection types of 2-classified as resistant and infection types of 3 and higher classified as susceptible. Progeny tests were done on BC3F2:3 individuals that showed recombination between Xcfd49 and Xbarc173.

DNA isolation

Genomic DNA was isolated from 1060 BC3F2 plants. Approximately 40 mg of leaf tissue from each plant was collected into separate 1.1 mL tubes containing stainless steel 5/32 in ball bearings (Grainger, Lake Forest, IL) in a 96-well plate format and ground using a Retsch MM 400 mill (Retsch, Newtown, PA). DNA was isolated using the Mag-Bind® Plant DNA Plus 96 Kit (M1128, Omega Bio-Tek, Norcross, GA) on a King Fisher Flex (Thermo Scientific, Waltham, MA) instrument and quantified using the Quant-iT™ PicoGreen® dsDNA Kit (Life Technologies Corp., Grand Island, NY) in a 384-well format on a CFX384 C1000 Real-Time thermal cycler (BioRad, Hercules, CA). Normalization to 30 ng µL−1 for SSR and STS markers and 10 ng µL−1 for KASP™ markers was done using a GBC Fit-X1 instrument (New England BioGroup, Atkinson, NH).

Marker development and genotyping

SNP markers were developed from GBS tags generated using a two-enzyme GBS protocol (Poland et al. 2012) from 94 BC2F1 plants segregating for SrTA10187 and parental lines. TASSEL 3.0 (maizegenetic.net) and the UNEAK pipeline were used to call SNPs (Lu et al. 2013). Bulked segregant analysis was done using polymorphic GBS tags to identify SNPs associated with resistant individuals (Pujol et al. 2015; Trick et al. 2012).

Fifteen KASP™ (competitive allele-specific PCR, LGC, Teddington, Middlesex, UK) markers were used to fine map SrTA10187. KASP™ markers were developed from GBS SNPs and SNP sequences found in the online databases CerealsDB (Wilkinson et al. 2012, http://www.cerealsdb.uk.net/) and Sequencing the Aegilops tauschii Genome (Luo et al. 2013, http://aegilops.wheat.ucdavis.edu/) (Semagn et al. 2014). Source SNP identification codes are listed in Table 1. Initially, KASP™ markers were tested on parental genotypes to ensure that genotypic classes could be differentiated. Marker expression (codominant or dominant) and parental allele specificities are included in Table 1. The KASP™ markers 6DS0027, AT6D5273, and AT6D5280 were scored as dominant markers because heterozygous fluorescence clusters could not be differentiated. All other SNP markers were scored as codominant. Primer names include allele 1 (A1), allele 2 (A2), and common (C) primer designations. KASP™ marker loci were amplified in 5 µL reactions containing 10 ng of gDNA, 0.07 µL of primer mix (containing 12 µM of each allele-specific forward primer and 30 µM reverse primer), and 2.5 µL of KASP™ master mix. KASP™ thermal cycling was carried out according to the manufacturer’s protocol: 94 °C for 15 min; 10 step-down cycles of 94 °C for 20 s, and 61–55 °C for 60 s (decreasing by 0.6 °C each cycle); and 26 cycles of 94 °C for 20 s, and 55 °C for 60 s. The BioRad CFX384 was used for thermal cycling and KASP™ marker fluorescence detection. BioRad CFX manager software was used for allelic discrimination.

Table 1 KASP™ and STS markers used for fine mapping of SrTA10187 on wheat chromosome 6DS
Full size table

One sequenced-tagged site (STS) marker (6DS0050) was also used for SrTA10187 fine mapping. Primers were designed to amplify a 100 bp DNA segment of the gene designated MSU_6DS_001. The 6DS0050 primer sequence is shown in Table 1. The 6DS0050 marker locus was amplified in 20 µL reactions containing 120 ng of gDNA, 0.5 µL of each primer (10 µM), 2 µL of 10 × reaction buffer, 0.5 µL of dNTP (10 mM), and 0.2 µL of Taq polymerase (Empirical bioscience, Grand Rapids, MI, USA). Thermal cycling was done using the following protocol: 95 °C for 4 min; 34 cycles of 95 °C for 45 s, 57 °C for 45 s, and 72 °C for 30 s. A total of 20 µL of PCR product from each reaction was added to 4 µL of 6x loading dye and visualized on 1.5 % agarose gel containing ethidium bromide. A 100 bp DNA Ladder was used as a standard (New England Biolabs, Ipswich, MA, USA).

Linkage map construction and comparison to existing Ae. tauschii physical and genetic maps

A linkage map of the 6DS region harboring SrTA10187 was constructed in JoinMap® 4 (Kyazma®, Wageningen, Netherlands) using the maximum likelihood algorithm. MapChart 2.3 was used for genetic map formatting (Voorrips 2002). Command-line BLAST was used to align each KASP™ marker context sequence to a unique location in the Ae. tauschii reference genome (Jia et al. 2013). Ten markers aligned to Ae. tauschii scaffolds with a known genomic position, whereas the remaining five markers aligned to unanchored scaffolds. The genetic map positions of markers in common with those published by Luo et al. (2013) were compared directly.

Scaffold sequence identification and annotation

The BAC-based physical map produced by Luo et al. (2013) was used to identify Ae. tauschii scaffold sequence in the region surrounding SrTA10187. Common SNP marker loci were identified between the Ae. tauschii genetic map (Luo et al. 2013) and SrTA10187. Physical positions of common markers as well as 6DS0050, 6DS0039, and BS0021983 on 6DS scaffolds were validated using BLAST. Scaffolds AT6D5270, AT6D5271, AT6D5272 and scaffolds of BAC library 6222 spanned the locus containing SrTA10187 and were ordered based on the SrTA10187 genetic map. The BAC 6222 scaffolds 6222.1 and 6222.2 could be ordered based on marker order. However, without marker coverage, many scaffolds of BAC library 6222 could not be ordered.

Scaffolds surrounding the SrTA10187 locus were annotated using the MAKER pipeline. REPEATMASKER used a previously published wheat repeat library to mask the scaffold sequences (Campbell et al. 2014; Wicker et al. 2002; Smit et al. 2013). Several transcript files were used as evidence to aid the gene prediction programs: wheat full-length cDNAs (http://trifldb.psc.riken.jp/download/ver.3.0/TaRFL4905.fas.gz), wheat predicted coding sequences (http://trifldb.psc.riken.jp/download/nuc/Triticum_aestivum.full.fas.cds.fas.gz), and Ae. tauschii predicted coding sequences (ftp://climb.genomics.cn/pub/10.5524/100001_101000/100054/D/Annotation/wheatD_final_43150.gff.cds) (Jia et al. 2013; Mochida et al. 2009). UniProtKB/SwissProt plant proteins were also used as evidence for the gene model predictions (Apweiler et al. 2014). MAKER was run a single time allowing gene predictions to be made by both SNAP (using an existing rice HMM) and Augustus (using the wheat HMM from Augustus version 3.1) (Stanke et al. 2003; Korf et al. 2004). The predicted proteins were analyzed with Hmmscan to identify matching protein family (Pfam) domains (Finn et al. 2014; Eddy 2011). Gene models with support from transcript evidence, protein evidence, and/or matching Pfam domain(s) were retained as high-quality gene predictions. The context sequence used to design SNP markers 6DS0039, AT6D5273, and BS00021983 and primers used for 6DS0050 were aligned (using NCBI BLASTN version 2.2.26) and manually added to the scaffold annotations (Altschul et al. 1990). The coding sequences of two NB-ARC-LRR genes (MSU_6DS_001 and MSU_6DS_037) identified in the region of interest corresponded to the previously annotated genes F775_13570 and F775_21138, respectively (Jia et al. 2013). F775_13570 and F775_21138 coding sequences (obtained from http://plants.ensembl.org/Aegilops_tauschii) were aligned to scaffolds AT6D5270 and 6222.3 using Exonerate, and the corresponding MAKER gene predictions were revised (Jia et al. 2013; Slater and Birney 2005; Kersey et al. 2009).

The wheat chromosome 6DS GFF annotation file, transcript sequence file, protein sequence file, and Pfam annotation results are available for download from the Dryad Digital Repository (DOIs will be provided upon provisional acceptance). A genome browser displaying the annotation of these Ae. tauschii scaffolds is available for public access at http://childslab.plantbiology.msu.edu/jbrowse7/?data=data7%2Fjson%2Fwheat_AW4.

Results

High-resolution mapping of SrTA10187

In the high-resolution mapping population of 1060 BC3F2 plants, 683 plants had a resistant infection type of 2- and 377 had a susceptible infection type of 3 to Pgt race QFCSC (Fig. 1). Parental genotypes KS05HW14 and TA10187 had the expected susceptible and resistant infection types, respectively. Three BC3F2 families (U6897-2, U6897-4, and U6897-6) demonstrated the expected 3:1 segregation for QFCSC resistance, while the remaining three families (U6897-1, U6897-3, U6897-5) deviated from the expected ratio (Table 2).

Fig. 1
figure 1

Pgt race QFCSC leaf infection types on recurrent T. aestivum parent KS05HW14, susceptible BC3F2, resistant BC3F2, and Ae. tauschii accession TA10187

Full size image
Table 2 Segregation of Pgt resistance in a high-resolution mapping population comprising six BC3F2 families
Full size table

Each individual of the high-resolution mapping population was genotyped using SSR markers Xbarc173 and Xcfd49. A total of 69 individuals with recombination between SSR loci were identified and progeny tested for stem rust resistance. A genetic map was constructed spanning 6.8 cM in the distal region of wheat chromosome 6DS using 15 KASP™ SNP markers and one STS marker (Fig. 2b). Marker loci flanking SrTA10187 include 6DS0027 located 0.9 cM distally, and 6DS0039 and AT6D5273 located 0.2 cM proximally. The STS marker designated 6DS0050 maps 1 cM distal to SrTA10187 and produces a 100 bp PCR product from Ae. tauschii accessions TA10187 (the SrTA10187 donor) and AL8/78 (the Ae. tauschii reference genome accession) that is absent from KS05HW14 (Fig. 3).

Fig. 2
figure 2

Comparative genetic and physical maps of the 6DS chromosome region linked to the Pgt resistance gene SrTA10187. (A) Ae. tauschii reference genome 6DS physical map (Jia et al. 2013). Ten KASP™ marker sequences aligned uniquely to scaffolds in the four most distal recombination bins of 6DS. (B) A high-resolution T. aestivum genetic map encompassing SrTA10187 was developed from 1060 BC3F2 individuals. (C) Graphical representation of annotated Ae. tauschii scaffolds. Lines connecting the scaffolds to each genetic map indicate shared sequence. Shaded region = unanchored scaffold sequence from BAC pool 6222. Black marks indicate genes with Pfam domains consistent with plant defense response. (D) Ae. tauschii 6DS genetic map developed by Luo et al. (2013)

Full size image
Fig. 3
figure 3

The STS marker 6DS0050 amplifies a 100 bp product from Ae. tauschii accessions TA10187 and AL8/78, but not the T. aestivum parent KS05HW14. Primers were designed within an exon of MSU_6DS_001. 6DS0050 mapped 1 cM distal to SrTA10187. L = DNA ladder; NTC = no template control

Full size image

Anchoring the SrTA10187 genetic map to Ae. tauschii reference genomes

Initially, the high-resolution T. aestivum genetic map developed in this study was compared to the Ae. tauschii reference genome physical map developed by Jia et al. (2013) (Fig. 2A, B). Ten KASP™ marker sequences aligned uniquely to Ae. tauschii scaffolds in the four most distal recombination bins of 6DS spanning 4 Mb of sequence. The remaining five KASP™ markers aligned uniquely to unanchored D-genome scaffolds (data not shown). Although Ae. tauschii scaffolds are ordered arbitrarily within recombination bins, only two inconsistencies were found between the physical and genetic maps at marker loci AT6D5273 and either BS00021983 or BS00111704. Due to the size of the physical map interval and reference genome recombination bins as well as discrepancies in marker order, the Jia et al. (2013) reference genome was unsuitable for candidate gene identification.

The genetic map developed in this study was subsequently compared to the Ae. tauschii genetic map developed by Luo et al. (2013) (Fig. 2B, D). Five KASP™ markers were developed from SNPs identified and mapped by Luo et al. (2013) including AT6D5264, AT6D5273, AT6D5276, AT6D5280, and AT6D5282. These KASP™ markers were mapped in the SrTA10187 populations to enable direct comparisons of genetic and physical positions. Conserved marker order across the region enabled the identification of a distinct interval on 6DS harboring SrTA10187.

To identify the genomic sequence spanning the SrTA10187 interval, genetic markers were aligned to Ae. tauschii scaffolds assembled by Luo et al. (2013) (Fig. 2C). The STS marker designed to amplify the MSU_6DS_001 exon sequence aligned to scaffold AT6D5270 and is located distal to the resistance locus (Fig. 2C). KASP™ markers 6DS0039, AT6D5273, and BS00021983 aligned to scaffold sequences from BAC pool 6222 and are located proximal to SrTA10187. Scaffolds 6222.1 and 6222.2 were ordered based on the genetic map. Unfortunately, 745 Kb of sequence from BAC pool 6222 could not be anchored to a genetic map position.

Candidate gene identification

Gene predictions were made in 1.5 Mb of sequence from scaffolds AT6D5270, AT6D5271, AT6D5272, and BAC library 6222 spanning the SrTA10187 interval. Scaffolds 6222.1 and 6222.2 were located proximal to SrTA10187 and did not contain defense response genes. Pfam domains consistent with known NB-ARC-LRR defense response genes were identified in the unanchored 745 Kb of sequence from BAC pool 6222. In this region, one NB-ARC-LRR gene was identified and designated MSU_6DS_037 (Fig. 2C). A gap is present in the BAC-based physical map and it remains unknown if other defense response genes are present in the identified genomic interval.

Discussion

In this study, the TTKSK-effective resistance gene SrTA10187 derived from Ae. tauschii was assigned to a 1.1 cM genomic interval on chromosome 6DS. The availability of SNP sequences in public databases, combined with GBS-derived SNPs and bulked segregant analysis, facilitated the development of KASP™ markers to produce a refined map of the SrTA10187 region. Saturation of the SrTA10187 region on 6DS helped to resolve the genetic position of this important resistance gene. Assignment of marker sequence to the publicly available Ae. tauschii genome sequence enabled the identification of a discrete physical map interval harboring putative defense response genes.

To date, five Pgt resistance genes have been identified on wheat chromosome 6DS: Sr5, Sr42, SrCad, SrTmp, and SrTA10187 (McIntosh et al. 1995; Gao et al. 2015; Kassa et al. 2016; Hiebert et al. 2015; Olson et al. 2013b). Sr42, SrCad, SrTmp, and SrTA10187 are effective against TTKSK, but Sr5 is not (Jin et al. 2007). Furthermore, SrTmp virulence has been identified in Pgt race TTKTK (Patpour et al. 2016).

Differences in race specificity exist between Sr42 and SrTA10187. North American Pgt races QFCSC and QTHJC are avirulent on SrTA10187 while these races are virulent on Sr42 (Ghazvini et al. 2012; Olson et al. 2013b). Virulence or avirulence to SrCad by QFCSC or QTHJC is currently unknown. Future testing of Sr42, SrCad, SrTmp, and SrTA10187 race specificity to TTKTK, QFCSC, and QTHJC will help elucidate the effectiveness of these genes on prevalent North American races as well as the Ug99 race group. Furthermore, allelism tests between SrTA10187 and other 6DS genes should be done using Pgt race TTKSK to determine the allelic relationship between these genes.

In recent studies, genetic markers closely linked to Sr42 and SrCad were developed for marker-assisted selection (Gao et al. 2015; Kassa et al. 2016). Sr42 and SrCad have been mapped to the same region of 6DS and differences in SNP marker haplotypes have been identified, however, the genetic relationship between Sr42 and SrCad is still unclear (Kassa et al. 2016). To make comparisons between the genetic map developed in this study and those published by Gao et al. (2015) and Kassa et al. (2016), the KASP™ markers Excalibur_s114066_kwm918 and contig32737_kwm112 were tested on our mapping population. The marker Excalibur_s114066_kwm918 (=IWB31561) was used because it co-segregated with Sr42 in a bi-parental population of 94 F2:3 plants (Gao et al. 2015). Excalibur_s114066_kwm918 also co-segregated with SrCad in a RIL population of 384 lines, a DH population of 334 lines, and mapped 0.37 cM proximal to SrCad in a RIL population of 141 lines (Kassa et al. 2016). The marker contig32737_kwm112 was used because it was derived from a SNP identified in Ae. tauschii, and it co-segregated with SrCad in three bi-parental populations (Kassa et al. 2016). Both markers were monomorphic in our mapping population, and could not be used for linkage map construction. All individuals tested (including parental lines KS05HW14 and TA10187) carry the C allele for Excalibur_s114066_kwm918, and for contig32737_kwm112. This result demonstrates the importance of identifying unique SNP markers, and knowing the parental alleles to select for each gene.

Olson et al. (2013b) reported SrTA10187 segregation distortion in a population of 105 BC2F1 plants. One consequence of using genes from a wild relative such as Ae. tauschii for the improvement of cultivated wheat is the potential for linkage drag. In the present study, three of the six BC3F2 families (U6897-2, U6897-4, and U6897-6) exhibited no segregation distortion, but three families (U6897-1, U6897-3, and U6897-5) deviated from the expected 3:1 segregation ratio (Table 2). Based on this observation, an additional backcross to the recurrent parent KS05HW14 may have broken linkage between SrTA10187 and deleterious allele(s) causing linkage drag in BC3-derived families U6897-2, U6897-4, and U6897-6. The segregation distortion observed in U6897-1, U6897-3, and U6897-5 likely increased recombination fractions and therefore genetic map length estimates, but it is unlikely that marker order was affected given the high marker coverage across the region.

While the publicly available Ae. tauschii reference genome was useful for identifying a physical region that corresponds to our genetic map, large recombination bins, unanchored scaffolds, and differences in marker order limited its effectiveness in identifying candidate genes. The physical region corresponding to our 6DS genetic map spans a relatively large distance of 4 Mb, comprised of only 4 recombination bins. Due to lower recombination in the Ae. tauschii reference genome mapping population, the order of many scaffolds located within recombination bins is unknown. The most distal 6DS recombination bin spans nearly 2 Mb of unordered sequence. Of the 15 KASP™ markers aligned to the Ae. tauschii reference, five markers aligned to unanchored scaffolds. The high-resolution genetic map developed by Luo et al. (2013) enabled the identification of a more discrete interval of sequences spanning the SrTA10187 locus.

In this study, we have developed the codominant KASP™ marker 6DS0039, located 0.2 cM proximal to SrTA10187, which will be useful for germplasm development, marker-assisted selection, and resistance gene pyramiding. Additionally, the STS marker 6DS0050 could be used as an alternative to the KASP™ markers developed in this study. Markers identified in this study will be useful for deployment of SrTA10187 into wheat breeding programs (Yu et al. 2015a). The codominant KASP™ marker 6DS0039 enables pyramiding multiple Pgt resistance genes with complementary race specificities into the same wheat line that could not be efficiently combined using phenotypic data alone.

Author contribution statement

A.T.W., E.L.O. designed research and objectives; A.T.W., L.K.B., E.I.B., T.L.L. performed research; A.T.W., S.K.S., E.L.O. analyzed data; K.L.C., J.A.P., S.K.S., E.L.O. contributed analytical support; A.T.W., E.L.O. wrote and edited the manuscript.