Five SSR markers were tested on the full panel of parental genotypes in Fig. 1; however, only markers Barc186 and Gpw2136 proved to be useful. With respect to Barc186, all the parents known to have Qfhb.rwg-5A.1 (PI277012, GP80, RWG21 and Novus-4) had the smaller (± 200 bp) of two critical bands (named band 2 here). The remaining parents had the slightly larger band (band 1; ± 210 bp). Regarding marker Gpw2136; PI277012 and GP80 produced a characteristically smaller, Qfhb.rwg-5A.2 diagnostic band (± 190 bp; band 2) than all the other genotypes. Barc186 and Gpw2136 were therefore used for MAS.
Initial cross with ND Noreen (1st SNP dataset)
Among the F1 of cross 18M6 (GP80/Novus-4//Monument; Fig. 1), five plants were heterozygous for Gpw2136 with bands 1 and 2 and were therefore likely Qfhb.rwg-5A.2 heterozygotes. The same five plants were also clear heterozygotes for the Barc186 marker (Qfhb.rwg-5A.1). The five selected dihybrid plants were transferred to a greenhouse and the best agrotype was crossed with ND Noreen to produce 22 hybrid seeds (cross 19M13). The F1 seeds plus parental controls (PI277012, Grandin, GP80, RWG21, Jerry, Novus-4, Monument, ND Noreen) were grown, samples were cut for 90K SNP genotyping and their chromosome 5A SNP haplotypes were studied. For each F1 family, F2 seeds were harvested. In an attempt to confirm that both resistance genes were present among the lines, 20–30 F2 seeds per 19M13 family as well as the parents were evaluated for FHB type II resistance in a greenhouse. The two data sets were then integrated.
Only SNPs that have previously been mapped to chromosome 5A(Wang et al., 2014) were manually curated using GenomeStudio 2.0. Two hundred and twenty-eight polymorphic chromosome 5A SNPs spanning the region between 8.12 and 148.3 cM (the length of the published map is 148.3 cM) were then exported to Excel. However, only 118 markers were found to be polymorphic between PI277012 and Grandin, and were therefore used to construct a chromosome 5A map of the doubled haploid GP80 (haplotype shown in Fig. 2). It appeared that chromosome 5A of GP80 contains primarily PI277012 chromatin (black) with a smaller intercalary region (light grey) of Grandin derived chromatin. This intercalary region was detected in the sequence of markers starting from 67424 (83 cM) through 76124 (101.2 cM). Four markers (two at 98.7 cM and two at 19.9 cM) did not fit the general pattern and were probably incorrectly mapped rather than being the result of double crossovers. Since GP80 has the dominant allele for thresh-ability at the Q locus (rather than the q-allele of PI277012), the result suggests that the Q locus is contained within the region of intercalary, Grandin-derived chromatin. Microsatellite marker results obtained by Tao (2019) similarly suggested that Qfhb.rwg-5A.2 occurs in the PI277012-derived 5AL distal region which also harbors microsatellite markers Gpw2136, Gpw2172, Gwm179 and Gwm126.
Region A. In order to derive an approximate haplotype map of Novus-4, 144 polymorphic SNP loci were compared between PI277012, RWG21, Jerry and Novus-4. The Russ genotype (a parent in the RWG21 pedigree; Fig. 1) was not available, so that some of the Novus-4 polymorphisms could not be unambiguously associated with PI277012. While the Novus-4 chromosome 5A haplotype (Fig. 2) is sparsely populated; it appeared to harbor a region of PI277012-derived chromatin that was detectable at 35.9, 36.6 and 39.0 cM (3 markers). Due to the absence of useful markers in the surrounding chromatin, the actual PI277012-derived region may have been considerably bigger (potentially from 19.9 to 42.0 cM). The latter area was designated region A (on 5AS) in Fig. 2 and appeared to be associated with Xbarc186-2 and Xfhb.rwg-5A.1.
Only progeny that tested positive for the simultaneous presence of Xbarc186-2 and Xgwm2136-2 were selected for making the initial crosses, including the first cross with ND Noreen. In these crosses, the SNP markers for region A were almost always (one exception) co-transferred with Xbarc186-2, the exception being cross 19M13 in which one of 22 F1 plants had the SNP markers of region A but lacked Xbarc186-2, thus confirming a strong tendency for their co-inheritance.
Region B is the most likely location of Qfhb.rwg-5A.2. In the first SNP dataset (Fig. 2) seven markers detected PI277012 chromatin within the 113.1–120.4 cM chromosome region. Due to a paucity of polymorphic markers in the directly adjoining areas, PI277012-derived chromatin may actually occur within the broader (101.2-125.2 cM) region. The second SNP dataset indicated the presence of PI277012 chromatin in between 104.9-117.7 cM (again the actual range could lie between 94.9-125.2 cM). Thus, region B was comparatively well-defined by the SNP markers. The proximal end of region B appeared to be the same as in the GP80 donor chromosome and located distally from the Q-locus. Chu et al. (2011) determined that GP80 exchanged the q-allele for Q, yet retained Qfhb.rwg.5A.2. They mapped Xcfd39 close to the Qfhb.rwg.5A.2 QTL peak with markers Xgwm179 and Xgwm595 lying within, but towards the distal end of the critical region. Tao (2019) determined that Xgwm2136 maps in between Xcfd39 and Xgwm179. In the F1 18M6, the area of PI277012 chromatin in the critical chromosome extended beyond region B up to the 5AL telomere (at 148.3 cM). Crossover in the 18M6 F1 plant retained only region B in the critical chromosome that was forwarded to the 22 F1 19M13 plants. F1 19M13 gametes either had both region B and Xgwm3126-2 (11 gametes); lacked both region B and Xgwm3126-2 (10 gametes) or had Xgwm3126-2 present but lacked region B (one gamete). This suggested that selection for the simultaneous presence of Xgwm3126-2 and the region B haplotype should frequently predict the presence of Qfhb.rwg-5A.2.
FHB resistance trial 1
Strong disease development occurred in FHB resistance trial 1. A one-way analysis of variance (unequal numbers of measurements/plant) of the parents and controls revealed highly significant (P = 0.001) differences. The mean IS values for this trial are summarized in Table 2. PI277012 (IS = 0.24) and GP80 (0.42) were the most resistant (both have Qfhb.rwg-5A.1 and Qfhb.rwg-5A.2) while CM82036 (IS = 0.50; Fhb1, Qfhs.ifa-5A) was the third most resistant. Novus-4 (0.58; Qfhb.rwg-5A.1) showed intermediate resistance. Although, Novus-4 and RWG21 (0.95) both have Qfhb.rwg.5A.1, they differed significantly in resistance which is likely due to different genetic background genes. Excepting 18Nord-114 (which has the Fhb1 marker allele), the remaining parents do not have any of the named FHB resistance QTL and showed moderate to very poor FHB resistance. Winter wheat varieties ND Noreen (0.76) and Jerry (0.86) performed better than the very susceptible parents 18Nord-114 (0.94), SY Monument (0.97), and Grandin (0.99). The 22 19M13 F1:2 families could be placed in four groups based on the presence/absence of the SSR and haplotype markers as is shown in Table 2. A group of six families (142 plants) lacked markers on both chromosome arms and its average IS was 0.79. A second group of five families (123 plants) segregated for region A/ Xbarc186-2 and had an average IS = 0.8. Seven families (177 plants) segregated for both critical regions and had average IS = 0.73. The last group of four families (94 plants) segregated for region B/Xgwm2136-2 only and had an average IS = 0.65. Dihybrid segregation within the third group meant that 9/16 plants (115 plants expected) would have had at least one copy of each critical marker allele present whereas 1/16 (13 plants expected) would have been homozygous for both critical markers. However, there is no clear evidence that group III includes a frequency of F2 plants with superior resistance (comparable to GP80 plants) and different from the segregation patterns in the remaining groups. As such, the data could not confirm the retention of the targeted resistance QTL in the segregates. FHB resistance QTL were often reported to show poor penetrance in certain genetic backgrounds, presumably those devoid of significant “native” resistance (Brar et al., 2019; H. Buerstmayr et al., 2008). Thus, introgression of highly characterized QTL such as Fhb1, Fhb2 and Fhb5 have failed to produce lines with resistance comparable to the donor parent (Brar et al., 2019). Brar et al. (2019) suggested that segregation of undetected background QTL and their unknown epistatic interactions could influence the expression and penetrance of FHB resistance genes. In this study, disease testing was performed on F1 and F2 populations rather than on highly homozygous lines which not only limited the numbers of spikes evaluated per genotype (and thus repeatability of resistance estimates) but also maximized genetic effects such as dominance, over-dominance and epistasis. SY Monument proved to be highly susceptible while ND Noreen appeared to have moderate inherent/native resistance. Segregation of unknown background QTL from the latter parents may have contributed to phenotypic variation among individual plants and chromosome 5A haplotype groups. Comparison of the F2 distributions with the ND Noreen and SY Monument distributions suggested, however, that there could be F2 plants with better IS than these two winter wheats. Thus, while the SNP haplotypes and marker data strongly suggested that the two resistance QTL were retained, the FHB results did not underscore this.
Family 19M13-67 included the best agrotypes and resistance phenotypes. Furthermore, the 19M13-67 F1 plant was heterozygous for the region A haplotype, Xbarc186-2, the region B haplotype as well as Xgwm2136-2. Two F2 plants (19M13-67-6 and 19M13-67-9) that were homozygous for the two SSR markers were selected from this family for making the backcross (= 20M1) to ND Noreen. The parents and 131 cross 20M1 B1F1 were again used for doing a new round 90K SNP analyses and greenhouse FHB Type II resistance tests.
Backcross to ND Noreen (2nd SNP dataset)
SNP data for the two B1F1 20M1 populations and their parents were analyzed. Population 20M1A derived from cross 19M13-67-6/ND Noreen (112 plants) whereas population 20M1B derived from cross 19M13-67-9/ND Noreen (19 plants). Chromosome 5A SNP loci (Wang et al., 2014) were identified, manually curated (GenomeStudio 2.0) and the data was exported to Excel. Two hundred and sixty-four polymorphic SNPs were found between 15.5 and 148.3 cM on chromosome 5A. Comparison of the SNP genotypes of the parents (GP80, Novus-4, Monument, ND Noreen, 19M13-67-6, 19M13-67-9) and the 131 B1F1 revealed two sets of markers that are highly likely derived from PI277012 derivatives GP80 and Novus-4 (Fig. 2). The locations of the two implied critical regions showed close correspondence to those that were detected in the first SNP dataset. Twenty-seven polymorphic markers occurred in the 19.9 to 38.7 cM chromosome region whereas six polymorphic loci occurred within 104.9 to 117.7 cM (due to low marker coverage the latter critical region may actually be broader - within 98.4 to 125.2 cM). The B1F1 results furthermore revealed the presence of two slightly different SNP haplotypes (I and II) that are shown in Fig. 2. Both haplotypes occurred in population 20M1A (1:1 segregation ratio; P = 0.45) suggesting that F2 plant 19M13-67-6, while homozygous for the presence of the Xbarc186-2 marker allele, was heterozygous for these haplotypes. Only haplotype I occurred in plant 19M13-67-9 and population 20M1B.
B1F1 populations 20M1A and 20M1B were also evaluated for FHB resistance (greenhouse trial 2). The distribution of infection severity scores for the individual plants is shown in Fig. 3. Compared to greenhouse trial 1, the severity of infection was generally lower. The average IS for group 20M1A was 0.22 and for 20M1B it was 0.24. Within group 20M1A, the average IS of haplotype I was 0.24 and that of haplotype II was 0.21; therefore, the two populations appeared to have similar resistance and the loss of a small region of GP80 chromatin in haplotype II did not appear to affect FHB resistance. Fewer (5–6) spikes were infected per control making the control averages less reliable than the population means. However, from Fig. 3 it appears that the mean IS of the two populations are in between the Novus-4 and ND Noreen means. Since all the B1F1 plants have both markers present, it could be expected that the population averages should approximate those of GP80 (assuming complete dominance of the two resistance QTL and minimal interference from background QTL). In trial 1 (evaluated the F2 of cross 19M13) around 25% of the genetic background of the hybrid plants derived from the highly susceptible cultivar, SY Monument (Fig. 1). However, in trial 2 which evaluated backcross F1, the average genetic contribution of SY Monument had dropped to 12.5% (Fig. 1). Although reduced, the detrimental effect of the SY Monument background likely persisted in a significant proportion of the population. With regard to the IS values of all 131 B1F1 plants; 34 plants ≤ GP80; 58 plants were ≤ Novus-4 and 93 plants ≤ ND Noreen. This suggested but did not prove that either or both FHB resistance QTL had been retained. Fiedler (2021) derived a KASP marker named 5AL-8.0K from the semi-thermal asymmetric reverse PCR (STARP) marker Rwgsnp36 (the latter was developed by Xu (2020) and is based on a SNP locus that occurred at the QTL peak that defined the Qfhb.rwg-5A.2 interval (Chu et al., 2011). Evaluation of marker 5AL-8.0K on 45 F2 progeny from population 20M1-58 revealed 1:2:1 segregation (P = 0.57) and the marker phenotypes were in complete correspondence with those obtained for the same group of plants using Gwm2136 and provided additional, albeit non-conclusive evidence of the likely presence of Qfhb.rwg-5A.2.
The SNP data were subsequently used for estimating the degree to which the ND Noreen background had been recovered in each of the B1F1 plants. F1: 19M13 resulted from the first cross with ND Noreen and therefore had 50% of its genes. Genome-wide, 620 heterozygous SNP loci were identified in F1: 19M13 and these loci were evaluated with regard to each B1F1 plant. The estimated proportions of genetic background ascribable to ND Noreen were calculated and ranged from 0.66 to 0.83. ND Noreen is known to have important agronomic traits that are difficult to breed for such as good winter-hardiness, resistance to bacterial leaf streak, high temperature adult plant stripe rust resistance and intermediate resistance to FHB (Ransom et al. 2020). An increased presence of ND Noreen genes may therefore improve the usefulness of the segregating material for selection and breeding purposes. Ten B1F1 that recovered 0.79–0.83 of the ND Noreen background and showed FHB IS scores similar to the GP80 control were therefore used for initiating single seed descent inbreeding.
Greenhouse FHB trial 3
An attempt was then made to confirm the presence of Qfhb.rwg-5A-1 and Qfhb.rwg-5A.2 in the selected progenies. Based on overall phenotype and apparent FHB resistance in the second FHB trial, four B1F1: 20M1A plants were identified and their F2 used to select nine promising F2:3 lines that are listed in Table 3. B1F2 plants were screened with Barc186, Gwm2136 and KASP marker 5AL-8.0K to verify the likely presence of the two resistance QTL (Table 3). The chromosome 5A SNP haplotypes present and estimated ND Noreen background recovery of each entry are also shown in Table 3.
Analysis of variance of the FHB IS data revealed highly significant differences among the 12 trial entries. The average FHB infection severities of entries across replications (Table 3) ranged from 9–86%. Of the parents, GP80 was the most resistant and SY Monument was the most susceptible. ND Noreen was moderately resistant. Among the nine selected progenies, line 20M1-58-32 had the lowest infection percentage. At least eight lines had average IS comparable to GP80. Each of the nine lines (of which eight were homozygotes) had the Qfhb.rwg-5A.2 markers and 5AL (region B) haplotype. Thus, judged by the levels of resistance in the selections and the presence of the critical marker alleles, it appears likely that Qfhb.rwg-5A.2 had in fact been transferred.
Only four of the eight best lines had the Qfhb.rwg-5A.1 marker (Xbarc186) and 5AS haplotype II). Novus-4 was an early parent in the introgression scheme (Fig. 1); is believed (however, not confirmed) to have Qfhb.rwg-5A.1 (Tao, 2019); and its haplotype closely resembles haplotypes I and II. All these three haplotypes (Novus-4, I and II) co-segregated with Xbarc186. Thus, it is likely that Qfhb.rwg-5A.1 actually occurs in the four lines with haplotype II and is absent from the four lines lacking it. Absence of Qfhb.rwg-5A.1 in the four lines without Xbarc186-2 and region A would suggest that background resistance QTL from ND Noreen could significantly complement the Qfhb.rwg-5A.2 resistance.
The eight resistant selections constitute a valuable resource for future breeding to improve the FHB resistance of hard red winter wheat in ND. Two of the eight selections have recovered approximately 82% of the ND Noreen genetic background in the B1F1 whereas the remaining six lines recovered 74 to 75%. Combined, the eight lines potentially capture a broad range of the cold tolerance and adaptation genes of ND Noreen that can be used to great benefit in future crosses to improve the FHB resistance of hard red winter wheat grown in ND.