Discovery of a Novel Stem Rust Resistance Allele in Durum Wheat that Exhibits Differential Reactions to Ug99 Isolates

Wheat stem rust, caused by Puccinia graminis f. sp. tritici Eriks. & E. Henn, can incur yield losses in susceptible cultivars of durum wheat, Triticum turgidum ssp. durum (Desf.) Husnot. Although several durum cultivars possess the stem rust resistance gene Sr13, additional genes in durum wheat effective against emerging virulent races have not been described. Durum line 8155-B1 confers resistance against the P. graminis f. sp. tritici race TTKST, the variant race of the Ug99 race group with additional virulence to wheat stem rust resistance gene Sr24. However, 8155-B1 does not confer resistance to the first-described race in the Ug99 race group: TTKSK. We mapped a single gene conferring resistance in 8155-B1 against race TTKST, Sr8155B1, to chromosome arm 6AS by utilizing Rusty/8155-B1 and Rusty*2/8155-B1 populations and the 90K Infinium iSelect Custom bead chip supplemented by KASP assays. One marker, KASP_6AS_IWB10558, cosegregated with Sr8155B1 in both populations and correctly predicted Sr8155B1 presence or absence in 11 durum cultivars tested. We confirmed the presence of Sr8155B1 in cultivar Mountrail by mapping in the population Choteau/Mountrail. The marker developed in this study could be used to predict the presence of resistance to race TTKST in uncharacterized durum breeding lines, and also to combine Sr8155B1 with resistance genes effective to Ug99 such as Sr13. The map location of Sr8155B1 cannot rule out the possibility that this gene is an allele at the Sr8 locus. However, race specificity indicates that Sr8155B1 is different from the known alleles Sr8a and Sr8b.

2008; Pretorius et al. 2010;Singh et al. 2015;Patpour et al. 2016) and is projected to spread further. Race TTKSK has evolved continuously, giving rise to at least 13 different variants, resulting in the defeat of additional stem rust resistance genes (Singh et al. 2015;Fetch et al. 2016;Newcomb et al. 2016). Overall, these new races pose a threat to our food security as 85-95% of wheat cultivars grown around the world are susceptible to at least one of the Ug99 variants (Singh et al. 2011).
Race TTKST is a variant of TTKSK (Ug99) that was first discovered in Kenya in 2006. Race TTKST is virulent to both Sr24 and Sr31 (Jin et al. 2008). In 2007, this race caused severe local epidemics in Kenya. It has since been detected in Tanzania, Ethiopia, Uganda, Eritrea, Rwanda, and Egypt (Jin et al. 2008;Pretorius et al. 2010;Wolday et al. 2011;Hale et al. 2013) and was the most prevalent Pgt race in Kenya from 2007 to 2014 . Races TTTSK, TTKTK, and TTKTT with additional virulences to Sr24, Sr36, and SrTmp (Jin et al. 2009;Newcomb et al. 2016) have substantially increased the vulnerability of wheat to stem rust because of the widespread use of these genes in global wheat breeding.
There is an urgent need to find additional genes that confer resistance to the new races of the Ug99 race group and identify reliable markers that assist breeding programs in combining these genes in desirable germplasm. However, genetic diversity for stem rust resistance in conventional common and durum wheat gene pools is limited (Singh et al. 2011), which has been a major constraint in identifying new genes effective against the Ug99 race group. Tetraploid wheats (T. turgidum ssp.) have contributed stem rust resistance genes such as Sr2, Sr9d, Sr9e, Sr9g, Sr11, Sr12, Sr13, Sr14, and Sr17 (McIntosh 1988;Simons et al. 2011;Singh et al. 2006Singh et al. , 2011. However, Sr13 is the only known Ug99effective seedling Sr gene present in selected durum cultivars in the US (Simons et al. 2011). Screening wheat lines for seedling resistance against race TTKST led to the identification of a durum line called 8155-B1 that was resistant to race TTKST, but susceptible to race TTKSK. This was the first line known to possess resistance to a variant of TTKSK that is considered to be more virulent than TTKSK. 8155-B1 was also resistant to US race TMLKC and exhibited a phenotype similar to TTKST. However, the basis of this resistance was unknown. The goal of this research was to determine the genetic basis of resistance to races TTKST and TMLKC in the durum line 8155-B1 and to develop KASP (Kompetitive Allele Specific PCR) assay-based SNP markers that can be used to postulate the presence of this gene in uncharacterized germplasm.

MATERIALS AND METHODS
Plant materials 8155-B1 is a T. turgidum ssp. durum line developed by Norman Williams (USDA-ARS, Fargo, ND) with the pedigree Marruecos 9623//Marruecos 9623/CItr 8155 that was characterized and selected as monogenic for stem rust resistance derived from CItr 8155 (Williams and Gough 1968). Marruecos 9623 (PI 192334) is a stem rust susceptible durum cultivar from Morocco (Williams and Gough 1968). CItr 8155 is a selection of wheat accession PI 59284 that was collected from Ethiopia in 1924. Rusty is a stem rust susceptible durum wheat line (Klindworth et al. 2006). Out of 143 Rusty Ã 2/8155-B1 BC 1 F 2 families, 44 lines that were either susceptible or segregating to race TTKST were employed to initially map the TTKST resistance. Similarly, out of a total of 473 F 2 plants derived from Rusty/8155-B1, 152 F 2 plants that were clearly resistant or susceptible to race TMLKC (see Results section) were used to map the TMLKC resistance utilizing the 90K Infinium iSelect Custom bead chip SNP genotyping platform. KASP assay-based markers were developed and evaluated on all the 143 BC 1 F 2 families and on a set of 11 durum cultivars from the US. Eleven durum cultivars and five bread wheat genetic stock lines were evaluated with Pgt at the seedling stage and KASP assays to validate mapping results (Supplemental Material, Table S1). Previously, durum cultivar Mountrail (Elias and Miller 2000) was crossed to common wheat variety Choteau (Lanning et al. 2004) and two recombinant inbred line (RIL) populations at both 4· and 6· ploidy were derived composed of 96 and 123 individuals, respectively (Kalous et al. 2015). We assessed the seedling response of these populations to races TTKSK and TTKST in order to map the response to race TTKST using the previously constructed linkage map with SNPs genotyped utilizing the 90K Infinium iSelect Custom bead chip (Kalous et al. 2015).

Stem rust assays
A total of 473 F 2 plants along with 8155-B1 and Rusty were evaluated against Pgt race TMLKC (isolate 72-41-Sp2) using a method described by Williams et al. (1992) at the USDA-ARS Cereals Crops Research Unit, Fargo, ND. At a biocontainment facility at the University of Minnesota, 25 BC 1 F 2 plants from each of the 143 families were evaluated against Pgt race TTKST (isolate 06KEN19v3) along with Rusty and 8155-B1. Choteau, Mountrail, and the 4· and 6· Choteau/Mountrail populations were assessed in two biological replicates each for response to races TTKST and TTKSK (Ug99; 04KEN156/04). Inoculation of seedlings was performed according to previously described methods . The 11 durum cultivars in addition to 8155-B1 and Rusty were evaluated with TTKSK and its variants, namely TTKST, TTTSK (07KEN24-4), TTKTT (14KEN58-1), and TTKSF+ (09ZIM01-2; race TTKSF with additional virulence to Sr9h [Pretorius et al. 2012;Rouse et al. 2014]). Races TTKSK, TTKST, TTTSK, TTKTT, and TTKSF+ are all members of the Ug99 race group . This panel was also evaluated with races JRCQC (08ETH03-1) and TRTTF (06YEM34-1) at both high (22/25°night/d) and low temperatures (15/18°night/d) with a 16-hr photoperiod in growth chambers. These two races were reported from Yemen and Ethiopia and described as particularly virulent to durum wheat (Olivera et al. 2012). The letters of each race name correspond to reaction patterns of the isolate to four stem rust resistance genes each (Jin et al. 2008). The full avirulence/virulence formulae for the isolates used this study are listed in Table S2.
Seedling evaluations of the biparental populations were conducted in greenhouse conditions ) and infection types were determined 14 d after inoculation following the 0-4 scale developed by Stakman et al. (1962). There are six categories of infection types in this rating scale: infection type "0" indicates an immune response with no visible symptoms, ";" indicates chlorotic or necrotic hypersensitive reactions without sporulation, "1" indicates small round rust pustules surrounded by chlorosis or necrosis, "2" indicates rust pustules surrounded by "green islands" of host tissue that are surrounded by chlorosis, "3" indicates elongated (not round) rust pustules, and "4" represents large elongated rust pustules without the presence of chlorosis or necrosis. Variation within each infection type class was captured by the use of "+" and "2" symbols that indicate relatively larger or smaller rust pustule sizes, respectively. When multiple infection types were observed on the same leaf, all infection types were listed, with the most common infection type listed first. A "/" symbol was used to separate multiple infection types recorded for a heterogeneous line where different plants within the same line displayed different infection types. Infection types "0" to "2" were classified as "low," i.e., incompatible interactions indicative of host resistance and pathogen avirulence, whereas infection types "3" and "4" were classified as "high," i.e., compatible interactions indicative of host susceptibility. Two biological replicates of seedling screening of the cultivar panel were performed. Significant deviation from the expected Mendelian genotypic frequencies was tested using chisquare tests.
SNP genotyping and identification of markers linked to race TTKST resistance DNA from 22 susceptible and 22 segregating BC 1 F 2 families in response to race TTKST and 152 F 2 plants segregating for response to race TMLKC were isolated using a modified CTAB extraction method  or an ammonium acetate method (Pallotta et al. 2003) and resuspended in water. Tissue from 10 BC 1 F 2 plants from each family was bulked for extraction of DNA representing each BC 1 F 2 family. DNA isolated from these lines was genotyped at the USDA-ARS Cereal Crops Research Unit, Fargo, ND, with 90,000 gene-based SNPs using a custom Infinium iSelect bead chip array and an iScan following the manufacturer's instructions (Illumina Inc., Hayward, CA; Wang et al. 2014). Allele calls were performed using the genotyping module of GenomeStudio v2011.1 software (Illumina Inc.) for the BC 1 F 2 population, whereas the polyploidy clustering module of the software was used to score the alleles of the F 2 population. The SNP consensus map data (Wang et al. 2014) were imported into GenomeStudio software to assign chromosome positions.

Identification of linked markers and map construction
Pearson correlations were used initially to identify markers associated with the TTKST phenotype in the subset of 44 families belonging to the BC 1 F 2 population (t-tests with P , 0.05). The top 16 SNP markers (P , 0.03) significantly correlated with resistance to race TTKST were converted into KASP assays. Eight of these KASP assays were polymorphic and were evaluated on the 143 BC 1 F 2 lines. These data were used to generate a linkage map using JoinMap version 4.0 (Stam 1993;Van Ooijen 2006). Genetic distances were calculated using the Kosambi mapping function (Kosambi 1944), and linkage groups were formed at logarithm of odds (LOD) value of 5.0 and 40% maximum recombination frequency. The KASP assay-based markers were also screened on the 11 cultivars. The primer sequences designed for the KASP assays are provided in Table 1. For the F 2 population, mapping of resistance to TMLKC was performed by calculating a linkage map using MapDisto 1.7.7 (Lorieux 2012). Linkage groups were created using the LOD score and Rmax value of 3.0. Map distances were calculated using the Kosambi mapping function (Kosambi 1944). For the Choteau/ Mountrail populations, hexaploid and tetraploid lines were combined into one population for the purpose of mapping. Previously available 90K SNP data (Kalous et al. 2015) were combined with binary resistance data in order to map the loci corresponding to resistance to races TTKSK and TTKST.

KASP reaction conditions
Each KASP PCR consisted of 50 ng of DNA template, 5 ml of 2· KASP buffer, and 0.14 ml of primer mixture. Thermal cycling conditions were 94°for 15 min, followed by 10 cycles of touch down PCR: 94°for 20 sec, 65-57°for 60 sec (dropping 0.8°per cycle), followed by 36 cycles of regular PCR: 94°for 20 sec, 57°for 60 sec, followed by fluorescence reading at 20°. A total of 3-9 additional cycles of PCR were added to obtain a good separation of clusters, as needed. Both thermal cycling and fluorescence reading were performed on an ABI StepOnePlus Real-Time PCR system. At least two replicates of each KASP assay were performed. If inconsistent results were observed between the two replicates, a third replicate was performed.

Data availability
All data that we used to draw conclusions in this article are represented either within the article, or in the Supplemental Material. Table S1 describes the wheat lines used in this study. Table S2 describes the Pgt isolates used in this study. Table S3 describes the number of n Table 1 Primers used for KASP assays for markers derived from the 90K iSelect assay on chromosome arm 6AS

KASP Primer
Primer Type a Primer Sequence a Primer types A1 and A2 are allele-specific primers, whereas primer type C1 is a common primer for both alleles.
Rusty/8155-B1 F 2 progeny with specific infection types observed in response to Pgt race TMLKC. Table S4 lists 90K SNP markers identified as correlated with response to Pgt race TTKST in a selection of the BC 1 F 2 population. Table S5 contains the alleles of markers linked to Sr8155B1 in Rusty Ã 2/8155-B1 families. Table S6 contains the alleles of markers linked to Sr8155B1 in Choteau/Mountrail RILs that displayed recombination events near Sr8155B1. Table S7 contains alleles of markers mapped to chromosome 6A in the Rusty/8155-B1 F 2 population. Table S8 contains seedling infection types observed on F 2 plants from the Rusty/8155-B1 population. Table S9 contains seedling infection types observed on BC 1 F 2 families of Rusty Ã 2/8155-B1. Table S10 contains seedling infection types observed on Choteau/Mountrail RILs in response to races TTKST and TTKSK. Figure S1 displays a range of seedling infection types observed on Rusty/8155-B1 F 2 progeny in response to Pgt race TMLKC. Figure S2 displays the genetic linkage map derived from Rusty/8155-B1 F 2 progeny. Figure S3 displays the seedling infection types observed on Choteau and Mountrail in response to Pgt race TTKST.

Molecular mapping of Sr8155B1
From a total of 21 SNPs significantly correlated with race TTKST resistance in the 44 BC 1 F 2 families (Table S4), eight polymorphic KASP assay-based markers were evaluated on 143 BC 1 F 2 families (Table 1). Seven markers were linked to the TTKST-resistant phenotype, including KASP_6AS_IWB10558 that cosegregated with Sr8155B1 ( Figure 2 and Table S5). Markers KASP_6AS_IWB61585 and KASP_6AS_IWB1550 flanked Sr8155B1 at 1.3 cM distal and 1.1 cM proximal, respectively. From the 152 F 2 plants, we generated a linkage map of 116 cM ( Figure  S2). The linkage map clearly positioned Sr8155B1 in the short arm of chromosome 6A, and was tightly linked to three markers that also were linked to Sr8155B1 in the BC 1 F 2 population: IWB64918, IWB43809, and IWB10558. Sr8155B1 was flanked by the SNP markers IWB55188 and IWB35219, being 7.3 cM proximal and 0.7 cM distal to Sr8155B1, respectively. The corresponding positions of these markers in the durum consensus map (Maccaferri et al. 2014) are shown in Table S4. Previously, wheat stem rust resistance gene Sr8, including alleles Sr8a and Sr8b, was mapped to the short arm of chromosome 6A (McIntosh 1972;Singh and McIntosh 1986;Bhavani et al. 2008).

Postulation of Sr8155B1 and Sr13 in durum wheat cultivars
The 11 durum cultivars displayed a range of infection types in response to races TTKST, TTKTT, TTTSK, TTKSF+, TTKSK, TRTTF, and   (Table 3). A low infection type of "0;" to ";1" in response to race TTKST in combination with a higher infection type in response to race TTKSK was considered indicative of resistance conferred by Sr8155B1.
Of the 11 cultivars evaluated, nine were postulated to possess Sr8155B1 (Table 3). The nine cultivars with Sr8155B1 in addition to 8155-B1 exhibited consistently low infection types at both temperature regimes to races TTKTT, TTTSK, TTKSF+, and TRTTF (Table 3), but not necessarily to race JRCQC. A low infection type of "11+" to "22+" in response to race TTKSK at the higher temperature regime was considered indicative of resistance conferred by Sr13 (Roelfs and McVey 1979). Six cultivars were postulated to possess Sr13 (Table 3). Five cultivars were postulated to possess both Sr13 and Sr8155B1: Rugby, Munich, Renville, Grenora, and Alkabo.

Validation of Sr8155B1-linked markers in durum cultivars
To identify potential markers that can discriminate lines with and without Sr8155B1, we evaluated eight KASP assay-based markers on the panel of 11 durum cultivars. The allele calls for these eight markers are shown in Table 4. Of the eight markers tested, three markers were found to predict the presence of Sr8155B1: KASP_6AS_IWB10558 (cosegregated with Sr8155B1), KASP_6AS_IWB72958, and KASP_ 6AS_IWB61585 (Table 4).

Relationship between Sr8155B1 and the Sr8 alleles
The marker KASP_6AS_IWB10558 was also evaluated on the Sr8acontaining lines ISr8a-Ra and SD4279 and the Sr8b-containing lines Barletta Benvenuota and Klein Titan, along with a highly susceptible wheat line, LMPG-6. The results revealed that ISr8a-Ra, SD4279, Barletta Benvenuota, Klein Titan, and LMPG-6 did not have the 8155-B1 allele for KASP_6AS_IWB10558 (Figure 3 and Table 5). ISr8a-Ra is susceptible to races TTKSK and TTKST, but resistant to TRTTF (Olivera et al. 2012), whereas Sr8b lines Klein Titan and Barletta Benvenuota were susceptible to races TTKSK, TTKST, and TRTTF (Table 5). Although 8155-B1 is resistant to TRTTF, it differs from ISr8a-Ra in that it produces an IT of "0;" whereas ISr8a-Ra exhibits an infection type "222"( Table 5). The race specificity and infection types of ISr8a-Ra, Klein Titan, and Barletta Benvenuota clearly indicate that Sr8155B1 is different from both Sr8a and Sr8b.

Validation of the presence of Sr8155B1 in cultivar Mountrail
Choteau displayed susceptible infection type "3+" in response to races TTKSK and TTKST. Mountrail displayed infection type "22" in response to race TTKSK, and infection types "0;" to ";" in response to race TTKST ( Figure S3). The response of Mountrail to races TTKSK and TTKST is typical of cultivars postulated to possess both Sr13 and Sr8155B1 (Table 3). The RILs derived from Choteau/Mountrail displayed two classes of infection types in response to race TTKSK: a resistant class with a range between ";122" and "2", in addition to a susceptible class with a range between "3" and "3+". Segregation of response to race TTKSK fit a single gene (x 2 = 1.73, P = 0.19), likely Sr13. The response of the RILs to race TTKST included infection types ranging from "0;" to "3+". Segregation of resistance to race TTKST fit a two-gene model (x 2 = 1.91, P = 0.17), likely Sr13 and Sr8155B1. We classified RILs with resistant TTKST infection types that displayed lower infection types compared to the response to race TTKSK as possessing Sr8155B1 (Table S10). RILs with TTKST infection types that displayed similar infection types to race TTKSK were classified as lacking Sr8155B1 (Table S10). We postulated that 91 RILs possessed Sr8155B1 and 108 lacked the gene, which fit a single gene ratio (x 2 = 1.45, P = 0.23). We were not able to confidently postulate presence or absence of Sr8155B1 in 20 of the 219 RILs (Table S10). We did observe a small quantitative difference between the effect of Sr8155B1 in hexaploid and tetraploid RILs of the Choteau/Mountrail population. The four most common infection types of hexaploid RILs (SXD1 through SXD135) postulated to possess Sr8155B1 were ";12", ";1", "11+", and ";132", but the most common infection types for tetraploid RILs (SXD136 to SXD232) with Sr8155B1 were "0;", ";", ";12", and ";1" (Table S10). Both Sr13 and Sr8155B1 were mapped in the Choteau/Mountrail population to chromosome 6A (Figure 4). Sr8155B1 mapped on the distal end of chromosome arm 6AS, whereas Sr13 mapped 150.6 cM away on the distal end of chromosome arm 6AL. Previously, Sr13 was linked to GWM427 in multiple populations (Simons et al. 2010). In Choteau/ Mountrail, Sr13 mapped 4.3 cM proximal to GWM427. Sr8155B1 mapped 0.6 cM proximal to KASP_6AS_IWB10558. Markers KASP_6AS_ IWB1550 and KASP_6AS_IWB61585 were not polymorphic in the Choteau/Mountrail population. Only two RILs possessed recombination events between Sr8155B1 and KASP_6AS_IWB10558 (SXD100 and SXD128; Table S6), demonstrating their tight linkage and the tractability of KASP_6AS_IWB10558 in another population. These mapping results confirm the presence of both Sr8155B1 and Sr13 in the cultivar Mountrail.

DISCUSSION
The search for new effective stem rust resistance genes from diverse germplasm is a continuous process in breeding wheat for resistance to stem rust. The best strategy to control emerging virulent races is to deploy complex rust resistance by adding new genes from diverse sources into a highly durable genetic background (Park et al. 2007(Park et al. , 2008. Monogenic line 8155-B1 line is unique because it is resistant to race TTKST, a variant of TTKSK with increased virulence, but susceptible to race TTKSK. 8155-B1 is also resistant to other variants in the Ug99 race group including TTTSK, TTKSF+, and TTKTT. In addition, 8155-B1 is resistant to TMLKC and TRTTF. Avirulence to Sr8155B1 is found in isolates of the Ug99 race group that vary in their avirulence to resistance genes Sr9h, Sr24, Sr31, Sr36, and SrTmp, and only one isolate has been characterized as virulent to Sr8155B1 (04KEN156/04; TTKSK). Although 04KEN156/04 is the oldest isolate in the Ug99 race group that we tested, we expect that ancestral Ug99 race group isolates are likely avirulent to Sr8155B1 based on the geographic and genetic variability of isolates that we confirmed as avirulent to Sr8155B1 . We hypothesize that virulence to Sr8155B1 is likely conferred by loss or modification of dominant avirulence to Sr8155B1. Pgt avirulence to most wheat stem rust resistance genes tested segregated as single dominant genes (Zambino et al. 2000). Melampsora lini avirulence proteins characterized in the flax-flax rust pathosystem directly interacted with flax resistance genes to induce host resistance (Dodds et al. 2006). We expect Sr8155B1-mediated resistance to similarly be conferred by the presence of a unique resistanceavirulence protein pair. Another possibility is that Sr8155B1 virulence is ancestral in the Ug99 race group, meaning that Sr8155B1 avirulence was acquired. This scenario might be explained by a mutation event that is either coincidental with the diversification of the Ug99 race group or associated with a selective advantage conferred by Sr8155B1 avirulence. If avirulence to Sr8155B1 was acquired, the molecular interactions causing this would be valuable to dissect to improve our understanding of such a phenomenon. Testing multiple isolates of each race for reaction to Sr8155B1 would help elucidate the evolution of virulence to Sr8155B1 in the Ug99 race group.
Segregation of resistance based on genetic analyses of F 2 and BC 1 F 2 populations revealed that a single gene, Sr8155B1, conferred resistance to race TTKST. Segregation of resistance fit the inheritance of a recessive gene in both populations, but it was difficult to determine whether Sr8155B1 is truly recessive or incompletely dominant (Williams and Gough 1968) without careful testing of F 1 plants derived from Rusty/ 8155-B1. Testing of 8155-B1 and several Sr8155B1-possessing cultivars at two temperature regimes indicated that Sr8155B1 is stable at high and n  Closest linked marker to Sr8155B1.
n Table 3 Seedling infection types of 11 US durum cultivars in addition to 8155-B1 and Rusty in response to Puccinia graminis f. sp. tritici races TTKSK, TTKST, TKTTF, TTTSK, TTKSF+, TRTTF, and JRCQC at low and high temperature regimes 18/15°day/night 25/22°day/night Line Sr8155B1 a Sr13 b TTKSK TTKST TTKTT TTTSK TTKSF+ TRTTF JRCQC TTKSK TTKST TTKTT TTTSK TTKSF+ TRTTF et al. (1962) infection types on a "0" to "4" scale are listed, where "0" indicates immunity, "4" indicates compete susceptibility, and ";" indicates a class in between "0" and "1" (see Materials and Methods for a detailed description). Smaller or larger rust pustules within an infection type class are denoted by "2" and "+" symbols, respectively. When multiple infection types were observed on the same leaf, all infection types are listed. A "/" symbol was used to separate infection types observed on different plants of the same heterogeneous line.
low temperatures, in contrast to Sr13 and Sr21 (Table 3; Chen et al. 2015). Testing hexaploid and tetraploid RILs of the Choteau/Mountrail population indicated that Sr8155B1 was more effective in tetraploid wheat, similar to findings for other Sr genes including Sr13 and Sr21 (Chen et al. 2015;Simons et al. 2010). Resistance to races TTKST and TMLKC mapped to the short arm of chromosome 6A and cosegregated with SNP marker IWB10558 in both the F 2 and BC 1 F 2 progeny, suggesting that the same gene conditions resistance against both races. The only other characterized gene from durum wheat that confers seedling resistance to race TTKST is Sr13, which is effective against all known races in the Ug99 race group and was mapped to the long arm of chromosome 6A (McIntosh 1972;Simons et al. 2011). To date, no known gene(s) on chromosome 6AS have been described to confer seedling resistance against race TTKST. However, the short arm of chromosome 6A harbors the Sr8 alleles, Sr8a and Sr8b (McIntosh 1972;Singh and McIntosh 1986), neither of which confers resistance to race TTKST. A gene in line SD4279 presumed to be Sr8a was recently mapped using the 9K SNP chip (Guerrero-Chavez et al. 2015), while another allele described as Sr_TRTTF was mapped in the Canadian wheat cultivar Harvest (Hiebert et al. 2017). The SNP markers IWB64918 (RFL_contig5170_330) and IWB6327 (BS00011010_51), which were linked to Sr8155B1 in our study, were also reported by Hiebert et al. (2017) to be linked to Sr_TRTTF, which was predicted to be Sr8a. IWB64918, closely linked to Sr8155B1, mapped 3.3 cM from Sr_TRTTF, and IWB6327 mapped 30.2 cM away from Sr_TRTTF (Hiebert et al. 2017). Allelism tests are needed to confirm the relationship between Sr8155B1 and Sr8. Even though our data do not elucidate whether or not resistance in 8155-B1 is conferred by an allele at the Sr8 locus, our phenotypic data do indicate that Sr8155B1 is distinct from both Sr8a and Sr8b. Therefore, Sr8155B1 is either a new allele at the Sr8 locus or a new stem rust resistance gene. n a Stakman et al. (1962) infection types on a "0" to "4" scale are listed, where "0" indicates immunity, "4" indicates compete susceptibility, and ";" indicates a class in between "0" and "1" (see Materials and Methods for a detailed description). Smaller or larger rust pustules within an infection type class are denoted by "2" and "+" symbols, respectively. When multiple infection types were observed on the same leaf, all infection types are listed.
The SNP marker KASP 6AS_IWB10558 not only cosegregated with Sr8155B1 in both populations derived from 8155-B1, but also predicted the presence/absence of this gene in unknown cultivars (Table 4). The robustness of this test would have been improved by including additional cultivars, especially susceptible cultivars. The KASP assay-based SNP marker developed for Sr8155B1 in this study could be used in selecting for stem rust resistance in combination with other Ug99 resistance genes in durum wheat, such as Sr13. We deposited a hexaploid line from the Choteau/Mountrail population that possesses both Sr13 and Sr8155B1. Hexaploid line "SXD 43," deposited as PI 681713, was selected based on having gluten strength similar to Choteau, as well as solid stems related to wheat stem sawfly (Cephus cinctus Nort.) resistance inherited from Choteau (Lanning et al. 2004). SXD 43 could be used as a source of both Sr8155B1 and Sr13 for breeding common wheat varieties with resistance to race TTKST.
Races TTKST, TMLKC, TTKTT, and TTTSK exhibited low infection types of "0;" to "0;1" on 8155-B1, suggesting that the same gene in this monogenic line conditioned resistance against all these races. Although Sr8155B1 confers susceptibility to race TTKSK, it conferred resistance to the three other races of the Ug99 race group tested. The specificity of Sr8155B1 has implications for field stem rust screening in Africa. The international stem rust screening nursery in Njoro, Kenya, has been dominated by Sr8155B1-avirulent races TTKST and TTKTT , whereas the screening nursery in Debre Zeit, Ethiopia, has been dominated by the Sr8155B1-virulent race TTKSK, in addition to the presence of race JRCQC that is virulent to Sr9e, Sr13 (Olivera et al. 2012), and Sr8155B1 (Table 3). Gene Sr9e was thought to be common in North American durum (Klindworth et al. 2007;Olivera et al. 2012). It is possible that resistance postulated as Sr9e is, in fact, conferred by Sr8155B1, although further studies are needed to test this hypothesis. The observation that durum lines from North America that were resistant in Njoro became susceptible when tested in Debre Zeit (Olivera et al. 2012) may be a result of the presence of virulence to both Sr13 and Sr8155B1 in Debre Zeit.
The susceptibility of 8155-B1 to race TTKSK limits the value of Sr8155B1 in protecting wheat cultivars from the Ug99 race group. However, the value of Sr8155B1 lies in the high frequency of this gene in durum cultivars adapted to the Northern Great Plains of North America. For example, we postulated both Divide and Alkabo to possess Sr8155B1, and these two cultivars were the most widely planted durum varieties in North Dakota in 2015 and the first and third most widely planted durum varieties in North Dakota in 2016 (USDA National Agricultural Statistics Service). Going forward, we recommend the combination of both Sr8155B1 and Sr13, such as in durum varieties Alkabo, Grenora, Mountrail, Munich, Renville, and Rugby in order to provide the maximum immediate protection of the durum crop in the Northern Great Plains of North America.

ACKNOWLEDGMENTS
Funding for this research was provided by USDA-ARS Appropriated Project 5062-21220-021-00, the USDA-ARS National Plant Disease Recovery System, USAID Feed the Future, the Durable Rust Resistance in Wheat project, and the National Research Initiative Competitive Grants 2011-68002-30029 (Triticeae-CAP) and 2016-06708 (IWYP) from the USDA National Institute of Food and Agriculture. We acknowledge the University of Minnesota Supercomputing Institute for computational support. Mention of trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the USDA, and does not imply its approval to the exclusion of other products and vendors that might also be suitable.