Introduction

Wheat is one of the most important cereals cultivated in Ethiopia. It ranks fourth after Teff [Eragrostis tef (Zucc.) Trotter), Maize (Zea mays L.) and Sorghum (Sorghum bicolor (L.) Moench] in area coverage and third in total production (CSA 2009). The average per capital consumption of wheat in Ethiopia was estimated to be 39 kg/year during 1994–1997 and 331,000 t of wheat imported to meet the national wheat requirements during 1995–1997 (CIMMYT 2000). In the country, more than 70 bread and 30 durum wheat varieties have been released for production since 1940s. However, the national average yield of wheat is still 1.4 tons/ha (FAOSTAT 2003). Demand of wheat has steadily increased in the last decades in Ethiopia particularly due to the emergence of many food processing industries. Wheat in Ethiopia is represented by hexaploid (2n = 6X = 42) and tetraploid (2n = 4X = 28) species. Bread wheat is widely grown hexaploid wheat (Triticum aestivum L.), while durum wheat (Triticum durum s.l. incl. Triticum aethiopicum.) and emmer wheat (Triticum dicoccon Schrank) are the two cultivated tetraploid wheats.

The enormous genetic variability of the cultivated tetraploid wheats makes Ethiopia the Center of diversity for cultivated tetraploid wheats (Vavilov 1929). In Ethiopia, there are numerous accessions of wheat germplasm (about 12,000 accessions) that has been collected and maintained mainly in the Institute of Biodiversity Conservation (Addis Ababa, Ethiopia). These wild and cultivated relatives of wheat offer a tremendous potential to be used as a source of stem rust resistance, and to broaden the genetic basis of wheat cultivars. Landraces have priority, as they may be used as starting population for cultivar development (Lakew et al. 1997; Teklu and Hammer 2009), specific adaptation to the different environmental conditions in their regions of growth, and as sources for the introgression of genes and quantitative trait loci conferring resistance to biotic (Huang et al. 1997; Mujeeb-Kati and Rajarm 2000) and abiotic stresses (Forster et al. 2000). Despite these valuable features, the use of landraces has been discouraged in many developing countries on the basis that they have low yield potential (Teklu and Hammer 2009).

Even if over 30 fungal diseases of wheat have been identified in Ethiopia, stem rust caused by Puccinia graminis Pers. f. sp. tritici (Pgt) is a major production constraint in most wheat growing areas of the country and causes up to 100 % yield losses in epidemic outbreaks (Admassu et al. 2004). The country also considered as one of the hot spot areas for the development of the present wheat stem rust complex (Leppik 1970). The disease has become a major threat of wheat production after the epidemics of 1974 and 1993 that drove two bread wheat varieties, ‘Lacketch’ and ‘Enkoy’, out of production (Badebo 2002; Beteselassie et al. 2007). A new virulent stem rust race, Ug99, was first identified in Uganda in 1999 (Pretorius et al. 2000), then it spred to Kenya in 2001 and to Ethiopia in 2003 following the migration path suggested by Singh et al. (2006). Due to Ug99 and its variants widely used major stem rust resistance genes became ineffective (Singh et al. 2006; Jin et al. 2007, Yu et al. 2011). Therefore, from the identified 50 stem rust resistance genes, only a few are effective against Ug99. Sr2, 13, 22, 25, 26, 35, 39 and Sr40 were reported genes to be effective against Ug99 (Singh et al. 2006, 2008; Yu et al. 2010, 2011).

The major cause for the ineffectiveness of wheat varieties against stem rust is the narrow genetic base on which the breeding for resistance has been founded (Beteselassie et al. 2007). Earlier works on stem rust in Ethiopia concentrated on occurrence of Pgt physiologic races on hexaploid wheat (Temam 1984; Masresha 1996). Badebo et al. (1990) postulated yellow rust resistance genes (Yr) in hexaploid wheat varieties. However, little work has been done on gene postulations on Ethiopian tetraploid wheat accessions. Dawit (2008) postulated Yr genes in Ethiopian hexaploid and durum wheat varieties. Beteselassie et al. (2007) postulated the stem rust (Sr) genes of Ethiopian tetraploid and emmer wheat accessions through multipathotype testing. The basis for genetic analysis and gene postulation for the past studies is resistance-specificity of the host, as expressed by distinct qualitative disease reactions on seedlings, i.e. infection types (ITs), when challenged by a series of pathogen isolates.

As an alternative to gene postulation, presence of resistance genes can be determined by testing host cultivars with molecular markers linked to resistance genes. This approach overcomes gene interactions and plant stage depending gene expression problems associated with traditional gene postulation (Vanzetti et al. 2011). In recent times there have been advances in development and mapping of molecular markers that are diagnostic for major Sr genes (Saal and Wricke 1999; Spielmeyer et al. 2003; Hayden et al. 2004; Mago et al. 2005, 2011; Tsilo et al. 2008; Babiker et al. 2009; Wu et al. 2009; Hiebert et al. 2010; Olson et al. 2010; Liu et al. 2010; Yu et al. 2010; Zhang et al. 2010; Admassu et al. 2011; Simons et al. 2011). However, there are no reports on identification of stem rust resistance genes in Ethiopian durum wheat varieties and tetraploid wheat landraces by reported linked or diagnostic molecular markers.

Objectives of this work were therefore (1) to identify stem rust resistance genes that are present in the durum wheat varieties and tetraploid wheat landraces using molecular markers, (2) to assess which Sr genes are effective for current Ethiopian stem rust races of Pgt including Ug99 based on the response of the accessions against field stem rust evaluation.

Materials and methods

Plant materials

A set of 58 tetraploid wheat accessions were used in this study. The materials consisted of 22 durum wheat (T. durum Desf.) varieties that were released in Ethiopia between 1966 and 2009 and 32 Ethiopian tetraploid wheat landraces (T. durum Desf. s.l., incl. T. aethiopicum, T. turgidum L. and T. polonicum L.). Additionally four durum varieties from ICARDA were included in the study. These accessions were obtained from Debre–Zeit Agricultural Research Center, Ethiopia, which also provided the taxonomical classification based on morphological characters. Lists of varieties and landraces with their stem rust response are presented in Supplemental Tables 1 and 2. More information about the varieties were presented in Haile et al. (2012b). Sr gene carrying differentials W2691SR13 (Sr13), SWSR22TB (Sr22), W3763–SR35 (Sr35) and Kingbird#1 (stem rust resistant line carrying the Sr2 complex and other unknown genes based on phenotype, Singh et al. 2009) were used as reference lines for molecular markers analysis.

Phenotyping

For the varieties, seven field trials were carried out during three consecutive years (2008, 2009 and 2010) at two wheat growing locations (Debre-Zeit, 2000 m a.s.l. and black soil; and Denbi, 1800 m a.s.l. with light sandy soil, abbreviated as DZ and DN, respectively) of Ethiopia. DZ is one of the hot spot locations and an internationally selected site for stem rust evaluation. At this location we have evaluated the materials twice a year, i.e. main season (July–October, rain-fed) and off-season (January–April, irrigated). So it was possible to expose the tested varieties to all the year round available stem rust races of Ethiopia. But the landraces were tested only in 2009 off-season at DZ.

The accessions were sown in two rows of 1 m length and 0.20 m spacing between rows. To facilitate and optimize the natural infection, the nursery was enclosed by spreader rows comprising ‘PBW343’ (bread wheat with the gene Sr31) (Das et al. 2006), ‘Morocco’ (susceptible bread wheat), ‘local red’ (susceptible durum wheat) and ‘Arendeto’ (susceptible Ethiopian tetraploid wheat variety) in 2:1:1:1 ratio, respectively. In addition to the natural infection, the trial was also artificially inoculated with Pgt urediniospores. Urediniospores from Ug99, bread and durum bulks were mixed in 1:1:1 ratio and about 2 mg/ml of spores was suspended in distilled water and then a drop of Tween 20 was applied per 10 ml of suspension and inoculated using a syringe. Inoculation started at stem elongation growth stage and was repeated 2–3 times every week.

For scoring stem rust severity in the field, the modified Cobb Scale (Peterson et al. 1948) was used to determine the percentage of tissue infected with rust. The host response to infection in the field was scored using “R” or resistant (small uredinia surrounded by chlorosis or necrosis); “MR” or moderately resistant (medium sized uredinia surrounded by chlorosis or necrosis); “MS” or moderately susceptible (medium-large compatible uredinia without chlorosis and necrosis); and “S” or susceptible (large, compatible uredinia without chlorosis and necrosis). Disease severity and host response data were combined in a single value called the coefficient of infection (CI). The average coefficient of infection (ACI) and CI for the improved varieties and the landraces was calculated by multiplying the mean (seven environments) and one season severity %, respectively times a constant for host response: immune = 0.0, R = 0.2, MR = 0.4, MS = 0.8 and S = 1.0.

Marker analyses

Genomic DNA was extracted from 2 weeks old fresh leaves that were harvested and pooled from five seedlings of each accession and stored at −80 °C. Extraction from frozen leaves was performed using the modified CTAB method described by Doyle and Doyle (1990).

A total of 17 PCR markers [SSRs (simple sequence repeats), InDels (insertion–deletion polymorphisms) and EST (expressed sequence tags)] that are linked/associated with four reported major Sr genes (Sr2, Sr13, Sr22 and Sr35) were included in this study. Primer names, forward and reverse primer sequences and references from Sr genes associated markers are detailed in Supplemental Table 3. PCR reactions and amplifications of these markers were performed using procedures described at UCDavis website (http://maswheat.ucdavis.edu/protocols/stemrust/) and Yu et al. (2010).

PCR reactions contained 50–100 ng template DNA, 250 nM Cy5-labelled forward primer, 250 nM unlabelled reverse primer, 0.2 mM dNTPs, 2.5 μL PCR buffer (10x), 1.5 mM MgCl2 and 1 U Taq DNA Polymerase in a total volume of 25 μL. Fragment detection was performed as described by Röder et al. (1998). For SSR markers, fragments were detected by an automated laser fluorescence (ALF express) sequencer (Amersham Biosciences Europe GmbH, Freiburg, Germany) using a short gel cassette. Fragment sizes were calculated using the computer program Fragment Analyzer Version 1.02 (Amersham Biosciences) by comparison with the internal and external size standards. The EST and InDel markers were resolved in 2.0 % agarose gels for amplification and the amplified fragments were stained with ethidium bromide and photographed. To clearly detect the fragment sizes for these InDel and EST markers, the analysis of fragment sizes was repeated on an AdvanCE FS96 microcapillary fragment analyzer system by loading 25 μL PCR products.

Results

Phenotyping

Stem rust severity (%), infection response and ACI for the varieties tested at DZ and DN during 2008–2010 and for landraces tested at DZ during 2009 are presented in Supplemental Tables 1 and 2, respectively. Of the 26 tested varieties, only ‘Sebatel’ showed a moderately resistant (MR) type of response with an ACI of 2. Varieties ‘Yerer’, ‘Ude’, ‘Boohai’, ‘Leliso’, ‘Ld-357, ‘Ginchi’, ‘Robe’, ‘Bichena’, ‘Gerardo’, ‘Foka’, ‘Oda’, ‘Quamy’, ‘Assassa’ and ‘Cham-1’ showed a MS type of response with an ACI of 8–28 whereas the rest of varieties showed a susceptible (S) type of reaction with 30–55 ACI (Supplemental Table 1). Landraces LR2, LR3, LR4, LR8, LR10, LR19, LR25, LR28, LR29 and LR32 showed a MS reaction with 8–40 ACI values. For the rest of the tested landraces a S type of reaction with up to 70 ACI was recorded (Supplemental Table 2).

Identification of stem rust resistance genes using molecular markers

Initially we screened 25 molecular markers that are associated with Sr2, Sr13, Sr22 and Sr35. But we used only 17 of the markers which showed polymorphism and clear fragments for haplotyping the genes in the present study (Tables 1, 2). Haplotypes were sorted for each stem rust resistance gene by the size of their fragments. Similar haplotypes for each gene were grouped together and compared to the original source of the gene based on the reference lines.

Table 1 Haplotype diversity of stem rust resistance genes Sr2 and Sr13 using linked molecular markers in Ethiopian durum wheat varieties and tetraploid wheat landraces. Numeric values are the fragment sizes (bp) of PCR amplicons for the respective marker and wheat line. NA null allele. Amplicons with same size were sorted together as a haplotype group and coded as follows: gray highlight the haplotypes similar to the known gene resources, numeric values in bold indicate fragment size reported by Yu et al. (2010) in a similar study and on UCDavis website (http://maswheat.ucdavis.edu/protocols/stemrust/)
Full size table
Table 2 Haplotype diversity of stem rust resistance genes Sr22 and Sr35 using linked molecular markers in Ethiopian durum wheat varieties and tetraploid wheat landraces. Numeric values are the fragment sizes (bp) of PCR amplicons for the respective marker and wheat line. NA null allele. Amplicons with same size were sorted together as a haplotype group and coded as follows: gray highlight the haplotypes similar to the known gene resources, numeric values in bold indicate fragment size reported by Yu et al. (2010) in a similar study
Full size table

Sr2 is the only catalogued adult plant stem rust resistance gene in wheat (McIntosh et al. 2003). It is located on the short arm of chromosome 3B (Hare and McIntosh 1979). Spielmeyer et al. (2003) reported that the SSR marker GWM533 was linked to Sr2 on chromosome 3B with a map distance of approximately 2 cM. Spielmeyer et al. (2003) also showed that a 120 bp PCR fragment was amplified in most lines carrying Sr2. The diagnostic PCR fragment for GWM533—120 bp was detected in ‘Sebatel’, ‘Hitosa’, LR25, LR10, LR26, LR27, LR28, LR32, LRW and in the Sr2 containing line ‘KINGBIRD#1’. BARC133, the other marker associated with Sr2, amplified a fragment size of 122 bp in ‘Sebatel’, ‘Hitosa’, LR25, LR3, LR7 and in the Sr2 containing line ‘KINGBIRD#1’. These markers amplified various sizes of PCR fragments in the rest of the varieties and landraces. Some varieties produced similar haplotypes as reported by Yu et al. (2010) for Sr2 positive lines i.e. 117 bp (for GWM533) and 120 bp (for BARC133). Thus, the haplotype GWM533—120 bp and BARC133—122 bp was considered to be indicative for Sr2 positive lines (Table 1).

Sr13 is a stem rust resistance gene present in several T. durum cultivars. Its main sources are the Ethiopian land race ST464 and the T. dicoccon (emmer wheat) germplasm Khapli (Klindworth et al. 2007). It is located on the long arm of chromosome 6A. EST marker BE403950; and SSRs DUPW167, WMC580, BARC104b and BARC104c were used for haplotyping Sr13 in this study. These markers showed null alleles and also produced different fragment sizes in the tested varieties and landraces. BE403950, DUPW167, BARC104b, WMC580 and BARC104c amplified fragment sizes of 691 bp, 243 bp, 273 bp, 316 bp and 175 bp, respectively, in varieties ‘Sebatel’, ‘Quamy’ and ‘Boohai’. The same fragment sizes amplified also in ‘Cocorit-71’ and ‘Cham-1’ by markers BE403950, DUPW167 and BARC104b. ‘Tob-66’ revealed fragment sizes of 243 bp, 273 bp and 316 bp for DUPW167, BARC104b, WMC580, respectively. ‘Robe’ and ‘Bichena’ revealed similar fragment sizes for DUPW167 and BARC104b, and ‘Yerer’ for BARC104b and BARC104c. Therefore, a five-marker combination with fragment sizes of “691-243-273-316-175” was considered as a haplotype for Sr13 in this study based on the reference line W2691SR13 (Table 1).

Sr22 was mapped on the long arm of chromosome 7A (Khan et al. 2005). Three linked markers, CFA2019, CFA2123 and BARC121, were used for haplotyping this locus by Yu et al. (2010). Olson et al. (2010) produced a new set of lines with reduced alien fragments and found that the closest markers flanking Sr22 in these lines are WMC633 and CFA2123. In this study we have used CFA2019, CFA2123, WMC633 and BARC121 to haplotype this locus. Markers CFA2019, CFA2123 and WMC633 produced 168, 234 and 119 bp fragments and BARC121 170 and 197 bp fragments in the Sr22 carrying line SESR22TB. CFA2019 and WMC633 produced a haplotype of 168 bp and 119 bp fragment sizes in LR1. Additionally, CFA2019 amplified a fragment size of 168 bp in ‘Sebatel’, ‘Boohai’, ‘Mamouri’, ‘JennahKhetifa’, LR12 and LR25. CFA2123 produced a fragment size of 234 bp in LR14, LR15 and LR30. But BARC121 and WMC633 did not amplify the fragment sizes that have been amplified in the Sr22 carrying line (170 + 197 bp and 119 bp). A fragment size of 215 bp, reported in a similar study by Yu et al. (2010) was amplified in LR25, LR22 and LR24 (Table 2).

The stem rust resistance gene Sr35 was originally transferred from Triticum monococcum to hexaploid wheat (McIntosh et al. 1984) and is effective against the TTKSK (Ug99) race of P. graminis. f. sp. tritici (Jin et al. 2007) and its variants, TTKST and TTTSK. It is mapped on the long arm of chromosome 3A between markers BF483299 and CJ656351 in a region of 2.2–3.1 cM, depending on the population (Zhang et al. 2010) and is located 41.5 cM from the centromere (McIntosh et al. 1995). Some of the markers that were found on the T. monococcum fragment containing Sr35 which are useful for marker assisted selection are CFA2193, BE423242, WMC559, BF485004, CFA2170, AK335187, CFA2076, BE405552, WMC169 and GWM480 (http://maswheat.ucdavis.edu/protocols/Sr35/index.htm).

We have employed these 10 SSR/EST-derived molecular markers to test for the presence of this gene in our accessions. But WMC559, WMC169, AK335187 and CFA2076 produced monomorphic fragments. Therefore, we have employed only markers CFA2193, CFA2170, GWM480, BE423242, BF485004 and BE405552 to haplotype our accessions. Using these six marker combination, a haplotype of 172-160-NA-355-148/213-392 bp was detected in the Sr35 carrying line, W3763-SR35, and in ‘KINGBIRD#1’ (except CFA2170 produced a different fragment of 197 bp). Among the tested materials, no variety or landrace showed this haplotype. But ‘Denbi’ showed a haplotype of 172-160-NA-355 bp and ‘Cham-1’, ‘Bakalcha’ ‘Bichena’ and ‘Gerardo’ revealed a haplotype of 172-160-NA. The rest of the varieties showed a haplotype of 172–160 bp for GWM480 and CFA2170. Only our two resistant varieties, ‘Sebatel’ and ‘Yerer’, produced a fragment size of 392 bp (similar to the Sr35 carrying line) for marker BE423242 (Table 2). We have also tested Mq(2)5*G2919, a line carrying Sr35, but it produced different fragment sizes for all of the markers except for BE423242 and BE405552 in comparison to W3763-SR35 (data not shown).

Discussion

Molecular markers are used in wheat resistance breeding for identification of designated resistance genes in genotypes where the genetic background has not yet been clarified like most durum wheat varieties of Ethiopia. Closely linked markers provide a means for the selection and identification of important genes in breeding programs and, in the case of diseases resistance, this can be done in the absence of pathogens (Babiker et al. 2009).

Resistance gene Sr2, in addition to other unknown minor genes derived from variety ‘Hope’ commonly known as the ‘Sr2-complex’ (McIntosh 1988; Singh et al. 2006) is the basis for the effectiveness of Sr2 (Singh et al. 2006). This stem rust resistance gene has provided durable, broad-spectrum resistance and has been used as an effective control measure against wheat stem rust in modern wheat breeding. The use of Sr2 in CIMMYT wheat improvement program resulted in the release of several popular varieties worldwide carrying this gene (Singh et al. 2009). This resistance gene is currently effective against all isolates of Pgt throughout wheat-growing regions of the world (Sunderlund and Roelfs 1980).

Even if Spielmeyer et al. (2003) reported a 120 bp PCR fragment amplified in most lines carrying Sr2, there are some exceptions as reported by Mago et al. (2011) where the 120 bp allele also occurred in many North American and CIMMYT lines which are considered not to have Sr2. Thus GWM533 is complicated to use because there are two different GWM533 loci on 3BS. But Spielmeyer et al. (2003) showed by DNA sequence that the two 120 bp PCR fragments amplified by the microsatellite marker GWM533 from wheat lines known to carry Sr2, and those without the resistance gene differed by the number of dinucleotide repeat units that formed the compound microsatellite motif. Based on this report, it is difficult to conclude that all the accessions that showed a 120 bp fragment size for this marker carry Sr2. Therefore it is important to apply the pair of STM markers developed by Mago et al. (2011) to exploit the DNA sequence variation within the microsatellite repeat.

Some varieties and landraces also showed the haplotype fragments 117 bp (GWM533) and 120 (BARC133). Similar fragment sizes were reported by Yu et al. (2010) as haplotypes for Sr2 positive lines for these markers. But it is difficult to conclude whether the lines carrying this haplotype in our study also possessed Sr2, since it is a different genetic background. A major QTL for resistance to stem rust including Ug99 was reported for chromosome 3BS close to the genomic region of Sr2 in a mapping population derived from ‘Sebatel’ as resistance source (Haile et al. 2012a). This observation supports the conclusion that Sr2 is present as effective resistance gene in ‘Sebatel’.

Sr13 is present in several T. durum varieties. Despite being a frequent gene in durum varieties, Sr13 was not detected in most of the Ethiopian durum wheat varieties in the present study. But this might be the reason that most of the markers we have used to haplotype this locus are not diagnostic in all the genetic backgrounds. These markers can be used to follow the Sr13 resistant alleles in segregating populations including some of the parental lines with known Sr13 sources, but the markers may fail to predict the presence of Sr13 in an unknown set of germplasm for example in landraces. The resistance in some of the durum wheat varieties that showed the haplotype for this gene, such as ‘Sebatel’, ‘Quamy’, ‘Boohai’, ‘Cocorit-71’ and ‘Cham-1’ might be due to other Sr genes. Using a mapping population developed from ‘Kristal’ and ‘Sebatel’, in our previous study, we have identified QTL for resistance to race Ug99 about 17.4 cM from Sr13 flanking markers (Haile et al. 2012a). Therefore, the resistance in these varieties could be due to the action of an allele of Sr13 since the Ethiopian stem rust pathotype is high on Sr13.

Admassu et al. (2011) reported that Sr13 is the only known gene effective against race TTKSK (Ug99) and its variants (TTKST and TTTSK) and other Ethiopian wheat stem rust races of Pgt. However, this result is based on a study of hexaploid wheat. In another study Admassu et al. (2009) also showed that the effectiveness of Sr13 in Ethiopia is regional. Thus, it is important to note that some virulent races other than Ug99 are reported to overcome Sr13 in some countries (Huerta-Espino 1992; McIntosh et al. 1995) and Ethiopia particularly on durum wheat (Olivera et al. 2011). Olivera et al. (2011) identified race JRCQC from 38 single-pustule isolates at Debre-Zeit from a 2009 durum screening nursery of Ethiopia that possesses a virulence overcoming the resistance gene Sr13. Therefore, it can best be used in combination with other genes through gene pyramiding particularly in Ethiopia where there is a current virulent Pgt race on durum wheat for this gene.

Stem rust resistance gene Sr22 was originally identified in the diploid wheat species Triticum boeoticum Boiss. accession G-21 (Gerechter-Amitai et al. 1971) and T. monococcum accession RL5244 (Kerber and Dyck 1973)). It was then transferred to tetraploid and hexaploid wheat through interspecific hybridizations. But so far no one has found it in durum. There may be occasional out-crossing between tetraploid wheat and T. monococcum and therefore we may find the gene from T. monococcum in tetraploid wheat (Ravi Singh personal comm.). But the use of this gene in wheat breeding is limited due to a yield penalty and a delay in heading date associated with the T. monococcum chromosome segment carrying this gene (Olson et al. 2010). But recently hexaploid lines with Sr22 which have reduced T. monococcum genome have been produced due to the effectiveness of this gene against Ug99 (Olson et al. 2010). Therefore, the varieties and landraces that showed haplotype loci for the diagnostic markers of this gene will be utilized in further breeding program to combat Ug99 and related races of Pgt.

Haplotype analysis of markers associated with Sr22 indicated the presence of the Sr22 gene in varieties and landraces which showed susceptibility response in the field testing. But only variety ‘Sebatel’ and ‘Boohai’ showed a MR and MS, respectively response to Pgt race Ug99 during the field testing and showed the haplotype for this gene. The presence of Sr22 in ‘Sebatel’ was indicated by a minor QTL in the respective genomic region in the ‘Kristal’ × ‘Sebatel’ mapping population (Haile et al. 2012a). Therefore, based on the current study, the markers used to haplotype Sr22 are not completely diagnostic and thus may produce false positive result as reported in UCDavis website (http://maswheat.ucdavis.edu/protocols/Sr22/Disease_rust_Sr22.htm) or Sr22 may be only partially effective for resistance to Ug99.

Sr35 originated from T. monococcum and is effective against Pgt races of TTKSK (Ug99) and its variants TTKST and TTKSK. There is no clear report where Sr35 was transferred to durum wheat. But since the source of resistance in some of the Ethiopian tetraploid wheat varieties is not clearly known, we have employed markers that are associated with Sr35 (CFA2193, CFA2170, GWM480, BE423242, BF485004 and BE405552) to check the presence of this gene. Based on the reference line, most of the tetraploid wheat varieties of Ethiopia including the susceptible ones showed the haplotype for this gene which is unlikely since Sr35 is considered as one of the most highly effective genes against the new African race Ug99 (Jin et al. 2007). We have also observed that all markers used to haplotype Sr35 produced the same fragment size for ‘KINGBIRD#1’ and for the line carrying Sr35 (W3763-SR35). Therefore, KINGBIRD might also carry Sr35 in addition to Sr2.

Admassu (2010) reported, based on testing for stem rust resistance genes in Ethiopian wheat varieties, that it was difficult to postulate the resistance gene(s) responsible for their resistance. The author indicated in his study that varieties ‘Cocorit-71’, ‘Ld-357’, ‘Kilinto’, ‘Bichena’, ‘Tob-66’, ‘Quamy’, ‘Robe’, ‘Ude’, ‘Yerer’, ‘Oda’, ‘Bakelcha’ and ‘Leliso’ displayed low ITs against all the Pgt races they have used, which made it difficult to postulate the type of genes present in these genotypes. Thus, they concluded that either a single gene or a combination of genes may be responsible for the resistance displayed by these varieties. Therefore, the subject requires further analysis with more molecular markers accompanied by gene postulation based on wider virulence spectra races.

Conclusion

The tetraploid wheat has been a source of resistance genes Sr2, 9d, 9e, 12, 13, 14 (Roelfs et al. 1992). According to Bechere et al. (2000) Ethiopian tetraploid wheat accessions were noted for their good source of resistance to stem rust. The presence of some genes in the landraces, in this study, also strengths this fact and showed that Ethiopian cultivated tetraploid wheat accessions are still good sources of stem rust resistance. Beteselassie et al. (2007) reported the same scenario by postulating the Sr genes in Ethiopian tetraploid wheat accessions through multipathotype testing.

Most of the genes that are catalogued were transferred to bread wheat from alien sources. Sr2 and Sr13 were transferred to bread wheat from tetraploid emmers and Sr35 was transferred from T. monococcum. It is reported on UCDavis website (http://maswheat.ucdavis.edu/protocols/stemrust/), that most of the molecular markers linked to Sr resistance genes are not diagnostic. This might be one of the reasons why we did not identify these genes in most of the tested durum wheat varieties. Dominance for the undesirable allele, lack of amplification, amplification of the wrong locus, recombination between the marker and the gene, and lack of polymorphism between the source and recurrent parents are also some of the reasons because of which markers can fail to predict the presence of a gene (Yu et al. 2010).

Sr22 and Sr35 are rarely used genes (Yu et al. 2010) that have been confirmed to be resistant to Ug99 (Jin et al. 2007). But some susceptible varieties and landraces showed a haplotype for these genes. For example, LR1 and Mamouri (50S) showed a haplotype for Sr22. Based on the reference line, W3763-SR35, most durum wheat varieties of Ethiopia showed similar fragment size for the tested diagnostic markers. Even some susceptible (40S) varieties, ‘Denbi’, ‘JennahKhetifa’ and ‘Hitosa’, showed a haplotype for Sr35. As a result, these haplotypes may not be diagnostic for Sr22 and Sr35 and further evaluation is needed. Using more molecular markers closely linked to the gene of interest could be useful for distinguishing the false positives.

Based on the result of this study, the resistance against race Ug99 (TTKSK) of Pgt in ‘Sebatel’ might be due to combinations of Sr resistance genes Sr2 and Sr22. The other resistant Ethiopian durum wheat varieties, ‘Yerer’, ‘Boohai’, ‘Ude’ and ‘Gerardo’, which also showed a MS reaction to Pgt race of Ug99 (TTKSK) might be due to Sr35. It was not possible to accompany the findings with pedigree tracking since the source of resistance genes in these varieties is not clearly known. Moreover, it is likely that these varieties also had resistant genes not detected because of a limited number of Sr genes with diagnostic markers available for durum wheat. Therefore, it will be helpful to accompany this approach with association analysis combined with pedigree and rust race reaction for better gene identification and postulation. But, as this study is the first report on the presence of Sr genes in Ethiopian durum wheat varieties and tetraploid wheat landraces based on linked or associated molecular markers, it gives some preliminary information for further research.