A population of 186 F₇-derived F₈ RILs was developed from a cross between UC1110 (pedigree: Chukar/3/Yding//Bluebird/Chanate) and PI610750 [pedigree: Croc1/Aegilops tauschii (Synthetic 205)//Kauz] (Lowe et al., 2011b). UC1110, an adapted Californian hard white spring wheat line bred by and used as a parent in the University of California, Davis (UC Davis) wheat breeding program, carries YR resistance loci on wheat chromosomes 2BS and 3BS (Lowe et al., 2011b) under the CA rust environment. PI610750, a synthetic derivative spring wheat line developed by CIMMYT in Mexico, carries the YR APR gene Yr48 and a minor YR resistance locus on chromosome 2AS (Lowe et al., 2011b). After multiplication of the RILs and parents in Mexico, F₈ seeds were used for all genotyping and disease phenotyping.

The parents and RIL population were evaluated for APR to LR at Ciudad Obregón in the Yaqui Valley in northern Mexico during the 2011–2012 growing season (LR2012Y) and at El Batán, CIMMYT headquarters outside Mexico City, in the 2012 and 2013 seasons (LR2012B and LR2013B). Similarly, two YR experiments were conducted at another CIMMYT research station in Toluca, Mexico, during the 2012 and 2013 growing seasons (YR2012 and YR2013). Field plots consisted of 0.7-m paired rows planted with approximately 60 plants of each line. The Avocet near-isoline for Yr24/26 was used as the LR spreader, whereas a mixture of susceptible wheat lines comprised of cv. “Morocco,” an Avocet near-isoline for gene Yr31, and six lines derived from an Avocet/Attila cross, was used as the YR spreader in field trials. The spreaders were planted around the experimental area and as hill plots in the middle of a 0.3-m pathway on one side of each experimental plot. The LR epidemic was initiated by spraying the LR spreader with an equal mixture of P. triticina races MBJ/SP and MCJ/SP suspended in Soltrol 170. The main difference between the two races is partial (MBJ/SP) and complete (MCJ/SP) virulence on Lr26 (Lan et al., 2014b). For the YR trials, a mixture of P. striiformis races Mex96.11 and Mex08.13 was sprayed onto the spreaders within and around the experimental areas. Isolate Mex08.13 belongs to the lineage of the aggressive YR race that was first detected in California (Chen, 2005) and is virulent on Yr genes 2, 6, 7, 8, 9, 31, and A, whereas Mex96.11 is virulent on Yr genes 2, 3, 6, 7, 9, 27, and A (Lan et al., 2014b). Disease severities (DS) for the parents and the RILs were recorded three times in each experiment, according to the modified Cobb’s scale (Peterson et al., 1948) and host response to infection was determined according to Roelfs et al. (1992), where “R” = resistant, or miniature uredinia surrounded by necrotic tissues; “MR” = moderately resistant, or smaller to moderate-sized uredinia surrounded by necrotic or chlorotic tissues; “MS” = moderately susceptible, or moderate-sized uredinia without necrotic or chlorotic tissues; and “S” = susceptible, or large uredinia without necrotic or chlorotic tissues. In cases of repeated DS data, the first note was recorded when the susceptible check Avocet displayed approximately 80% severity and repeated about a week later when it reached 90–100%. For multiple disease readings, the area under the disease progress curve (AUDPC) was calculated using the method described by Bjarko and Line (1988). Final DS (FDS) and AUDPC scores were used in downstream genetic analyses.

In this study, baseline molecular marker data and linkage map information for the UC1110/PI610750 population was used from a previous study by Lowe et al. (2011b). Linkage maps were graphically produced by using MapChart (Voorrips, 2002). QTL mapping of FDS and AUDPC for each experiment was carried out using inclusive composite interval mapping (ICIM) 4.1 (Meng et al., 2015) The ICIM method was also used to analyze the final mean DS for each disease (LRM and YRM) across experiments. The logarithm of odds (LOD) threshold to declare a QTL significant at P = 0.05 for each trait was determined based on 1,000 permutation tests. Stepwise regression was used to estimate both the main effects and percentages of phenotypic variance explained (PVE) by individual QTL (at LOD peaks). The QTL result was confirmed by QTL Cartographer (Wang et al., 2012). The seedling resistance gene was mapped by the Joinmap 4.1 software (van Ooijen, 2006). Marker orders and chromosomal assignments of linkage groups were verified using wheat SSR consensus genetic maps and the physical bin locations of DArT markers (Somers et al., 2004; Francki et al., 2009; Huang et al., 2012; Wilkinson et al., 2012). QTL designations were assigned following recommended practices¹. Factorial analyses of variance (ANOVA) were conducted to test for the significances of interactions between detected resistance loci based on AUDPC.

The FDS for LR (LRM) were 1 MR for PI610750 and 15 MSS, 20 MSS, and 30 MSS for UC1110 in LR2012Y, LR2012B, and LR2013B, respectively, over the three seasons of this study (Figures 3A–C); and LRM for the RILs ranged from 18 to 41% in those same environments (data not shown). LR severities among the RILs was continually distributed across 3 years (Figure 3), indicating quantitative inheritance of APR to LR in the population.

For YR, the FDS (YRMs) for PI610750 were 5 MR and for UC1110 were 30 M and 40 M in YR2012 and YR2013, respectively (Figures 4A,B). YRM for the RILs ranged between 14 and 30% across the two seasons (Figure 4). The frequency distributions of YR severity among the RILs showed a bimodal distribution in both environments (Figure 4), which is an indication for the segregation of a major seedling resistance gene in the UC1110/PI610750 population.

Pearson correlation coefficients (r) of LR severity ranged from 0.57 to 0.81 among the three environments, whereas it was 0.66 for YR severity over the 2 years (Table 1). Low and non-significant correlations were observed between LR and YR severities (r = 0.03–0.25) across all environments.

A major QTL for LR resistance, QLr.cim-2DS, was located on the end of the short arm of chromosome 2D and interval flanked by SSR markers cfd51 and gwm455 (Figure 5A). This QTL was consistently detected in all three LR field experiments and explained 11.8–26.6% of the observed variation in LRM (Table 2). Mean LR severities for RILs carrying QLr.cim-2DS ranged from 1 to 50% (with an exception of entry 86 with 70% LR FDS), whereas it was 5 to 90% for RILs without this locus (Figure 6). This resistance locus derives from UC1110 (Table 2).

Co-located APR loci to both rusts, QLr.cim-3DC/QYr.cim-3DC, were mapped near the centromeric region of chromosome 3D (3DC), flanked by SSR markers gwm341 and barc1119 (Table 2 and Figure 5B). This co-located resistance locus was also contributed by UC1110 and explained 23.3 and 6.1% of the observed variation in YRM and LRM, respectively (Table 2). In addition, the previous study reported YR resistance loci on chromosome 3BS and Yr48 were also identified in the present study under Mexican rust environments (Table 2).

The F₈ RILs were divided into groups based on the marker-inferred factorial combinations of the two LR resistance QTL and three YR resistance QTL (Tables 3, 4). Specifically, the flanking molecular markers for each QTL were used to infer the presence of parental alleles in each RIL. A marginally significant interaction (P = 0.037) between QLr.cim-2DS and QLr.cim-3DC on LR was observed across 3 years based on AUDPC (Table 3), which only explained 1.4% of LR AUDPC variation. QYr.ucw-3BS did not show a big effect on YR resistance at the adult plant stage when it was present alone in the adult plant stage (2.8% of the variation), but it showed a highly significant interaction with Yr48 that explained 37.9% of the variation (Table 4). By contrast, no significant gene interaction between QYr.ucw-3BS × QYr.cim-3DC, QYr.cim-3DC × Yr48, and QYr.cuw-3BS×Yr48×QYr.cim-3DC was detected, indicating additive effects was present between them. RILs combing all of three YR resistance QTL displayed severity from 0.1 to 3.6% with a near immune response to YR.

Lr26 is known to be present in the wheat-rye 1RS.1BL chromosome translocation (McIntosh et al., 1995), and this translocation was widely used in wheat breeding programs worldwide during the 1980s and 1990s. For instance, 54% of the wheat cultivars grown in Bulgaria carry the 1RS.1BL translocation (Landjeva et al., 2006), approximately 38% of wheat cultivars released since 1980 contain this translocation in China (Zhou et al., 2004), and nearly 50% of CIMMYT’s around 2,000 advanced bread wheat breeding lines were homozygous for the 1RS.1BL translocation in 1988 (Villareal et al., 1991). However, its frequency has decreased to below 5% in current breeding materials due to the breakdown of resistance genes located on it and to poor quality associated with the translocation. The widespread deployment of the 1RS.1BL translocation was due to the fact that the 1RS arm carries resistance genes to multiple diseases (e.g., Sr31, Yr9, and Pm8) and positively affects various agronomic traits (Tang et al., 2008). Unfortunately, races fully virulent on Lr26 were detected in most wheat-growing areas, including Europe, the Indian subcontinent, North America, and South Africa in last middle of 1990s (McIntosh et al., 1995), likely due to the widespread use of a single, major resistance gene over a large area. In the current study, Mexican LR race MBJ/SP was found to be partially virulent on Lr26, whereas another field race (MCJ/SP) was completely virulent on this gene. Thus, this study was not designed to detect the effect of Lr26. This gene is closely linked to the YR resistance gene Yr9 (Mago et al., 2005); however, we did not detect the effect of Yr9 because both races used in the field trials were virulent. Lr26 should be contributed by one of PI610750’s parents, Kauz (pedigree: Jupateco F73/Bluejay//Ures T81; released in Mexico as Bacanora T88), carrying the 1RS.1BL translocation (Singh and Rajaram, 1991).

The APR QTL for LR, QLr.cim-2DS, contributed by UC1110, is located on the short arm of chromosome 2D, in the 2DS-5 region of the DArT physical map (Francki et al., 2009) and flanked by SSR markers cfd36 and cfd51. So far, five cataloged seedling resistance genes have been detected on the short arm of chromosome 2D, including Lr2 (Ausemus et al., 1946), Lr15 (McIntosh and Baker, 1968), Lr22 (Hiebert et al., 2007), Lr39 (Raupp et al., 2001), and Lr41 (Sun et al., 2009). Lr22 and Lr39 were transferred from A. tauschii, whereas Lr41 was introgressed from Aegilops cylindrica to common wheat. UC1110 (pedigree: Chukar///Yding//Bluebird/Chanate) is an adapted breeding line developed at UC Davis with no indications of these introgressions in its pedigree; therefore, QLr.cim-2DS is expected to be different from these three seedling genes. Lr15 was mapped in the common wheat line “Kenya W1483”; however, the LR races MBJ/SP and MCJ/SP are virulent on this gene. Although the two races are avirulent to Lr2, the seedling test suggests that QLr.cim-2DS is not Lr2, as the ITs of both parents were “1” when challenged with the Lr2-virulent race TCT/QB at the seedling stage. Although several other LR resistance QTL have been mapped in wheat lines, Avocet (Rosewarne et al., 2012), CI 13227 (Xu et al., 2005), and Saar (Zhang et al., 2009) on the distal region of chromosome 2DS. They are located more than 10 cM away from QLr.cim-2DS based on the wheat consensus map (Somers et al., 2004). Thus, it is likely that QLr.cim-2DS is a previously undetected APR locus for LR in common wheat. A single locus mapping population is currently under development to fine map this resistance locus.

The pleiotropic resistance QTL, QLr.cim-3DC/QYr.cim-3DC, was detected near the centromere of chromosome 3D and flanked by SSR markers gwm341 and barc1119. Basnet et al. (2014) detected QLr.tam-3D/QYr.tam-3D in the CIMMYT wheat line Quaiu#3 and assigned its location to a proximal region of chromosome arm 3DS, flanked by markers wPt-672034 and barc125. That QTL was found to explain 3–6 and 3.2–4.3% of the observed variation in LR and YR DS, respectively. The flanking markers for the pleiotropic QTL in this study are only ∼4 cM distal of the QLr.tam-3D/QYr.tam-3D locus, based on the wheat consensus map (Somers et al., 2004). QLr.cim-3DC was also detected in an Avocet/Francolin#1 population (Lan et al., 2014a), where it explained 17.8–25.4% of the observed variation in LR DS in the Lr46-subtracted sub-population of RILs; but its effect was not significantly detectable on YR. In addition, the pedigree of UC1110 is Chukar/3/Yding//Bluebird/Chanate, while it is Babax/Lr42//Babax^∗2/3/Vivitsi for Quaiu#3 (Basnet et al., 2014) and Waxwing^∗2/Vivitsi for Francolin#1 (Lan et al., 2014b). The pedigree of Vivitsi is Sonoita F 81/Yaco/5/Bobwhite/Crow//Buckbuck/Pavon F 76/3/Yecora F 70/4/Trap#1, whereas Bluebird is one of parents for Buckbuck. Therefore, QLr.cim-3DC/QYr.cim-3DC in UC1110 might be the same QTL as in Quaiu#3 and Francolin#1 and derived from Bluebird, though a larger effect was observed in UC1110 than in either the Quaiu#3 or Francolin genetic backgrounds. Allelism tests will be required to test these hypotheses.

One YR APR QTL was mapped in this study to chromosome arm 3BS, within the marker interval cfd3417-barc133, in agreement with the previous study conducted under the California YR environment (Lowe et al., 2011b). In the previous study, the QTL was named QYr.ucw-3BS and was found to be closely linked to gwm533.1, barc133, and wPt-1620. An APR YR resistance gene, Yr30, has also been mapped on the short arm of chromosome 3B (Singh et al., 2000b); however, QYr.ucw-3BS is not Yr30 based on allelism test (Lowe et al., 2011b). While this resistance locus did not offer any significant protection against US YR race PST-100, it did impart partial resistance at the adult plant stage. The IT for UC1110 was “6” against Mexican race Mex14.191, indicating this line might present a minor resistance gene with moderate-susceptible reaction at seedling stage.

The second YR resistance gene, Yr48, was mapped at the same genomic position as reported in the previous study (Lowe et al., 2011b). This resistance gene comes from the synthetic wheat derivative PI610750 and is located in the telomeric region of the long arm of chromosome 5A. Yr48 showed large effect on YR in the adult plant stage under the U.S. rust environment, with an average reduction of 63% YR severity compared to fully susceptible lines (Lowe et al., 2011b), whereas it reduced YR severity only up to 37.1% under the Mexican rust environment. The ITs of RILs with Yr48 ranged from four to six at seedling stage and field severity ranged from 1 to 60%. We conclude that Yr48 confers moderate seedling resistance and APR against the Mexican Pst race used in the study; thus it should be re-classified as an all-stage resistance gene. The segregation frequency of genotypes carrying Yr48 showed a strong distortion, possibly due to random drift or selection in gametes and/or zygotes. In addition, the genotype of nearby markers for this gene also showed distorted distributions, confirming the YR phenotypic results obtained in seedling.

Small gene interaction between QLr.cim-2DS and QLr.cim-3DC may explain why the 3DC QTL did not display any significant interactions when combined with other QTL against LR in Avocet/Quaiu#3 population (Basnet et al., 2014). For YR, the effect of the QYr.ucw-3BS was larger in the presence than in the absence of Yr48, resulting in a highly significant interaction that explained a large proportion of the variation. Lowe et al. (2011b) also reported a highly significant interaction (P < 0.0001) between these two QTL with similar effects. No significant interactions were detected between QYr.cim-3DC and other QTL, suggesting that the YR resistance QTL on 3DC had an additive effect.

UC1110, a high yielding spring wheat line developed by the UC Davis breeding program, provides intermediate resistance to LR and YR in both Mexican and U.S. rust environments. PI610750 remained highly resistant to both rusts in the two countries. We obtained some transgressive RILs combining all of the detected YR resistance QTL, such as RILs 8, 26, 28, 119, and 254, with DS lower than 1% across Mexican and U.S. rust environments, whereas RILs 26 and 184 carry two LR resistance QTL resulting in DS of approximately 2%. We used two RILs (43 and 170) carrying the single resistance QTL 2DS and crossed them with RIL 180 (no resistance QTL) to develop single gene mapping populations for fine mapping and further characterization of this locus. The closely linked molecular markers to QLr.cim-2DS (within 0.5–2.3 cM), QYr.ucw-3BS (within 0.2 cM), and Yr48 (within 0.6–0.7 cM) can be immediately used in wheat breeding programs to improve rust resistance via marker-assisted selection.