Molecular characterization of field resistance to Fusarium head blight in two US soft red winter wheat cultivars

In the soft red winter wheat (Triticum aestivum L.) regions of the US, Fusarium head blight (FHB, caused by Fusarium spp.) resistance derived from locally adapted germplasm has been used predominantly. Two soft red winter wheat cultivars, Massey and Ernie, have moderate resistance to FHB. Mapping populations derived from Becker/Massey (B/M) and Ernie/MO 94-317 (E/MO) were evaluated for FHB resistance and other traits in multiple environments. Eight QTL in B/M and five QTL in E/MO were associated with FHB variables including incidence, severity (SEV), index (IND), Fusarium damaged kernels (FDK), deoxynivalenol (DON), and morphological traits flowering time and plant height. Four QTL were common to both populations. Three of them were located at or near known genes: Ppd-D1 on chromosome 2DS, Rht-B1 on 4BS, and Rht-D1 on 4DS. Alleles for dwarf plant height (Rht-B1b and Rht-D1b) and photoperiod insensitivity (Ppd-D1a) had pleiotropic effects in reducing height and increasing FHB susceptibility. The other QTL detected for FHB variables were on 3BL in both populations, 1AS, 1DS, 2BL, and 4DL in B/M, and 5AL (B1) and 6AL in E/MO. The additive effects of FHB variables ranged from 0.4 mg kg−1 of DON to 6.2 % for greenhouse (GH) SEV in B/M and ranged from 0.3 mg kg−1 of DON to 8.3 % for GH SEV in E/MO. The 4DS QTL had epistasis with Ppd-D1, Qdon.umc-6AL, and Qht.umc-4BS, and additive × additive × environment interactions with the 4BS QTL for SEV, IND, and FDK in E/MO. Marker-assisted selection might be used to enhance FHB resistance through selection of favorable alleles of significant QTL, taking into account genotypes at Rht-B1b, Rht-D1a and Ppd-D1a.


Introduction
Fusarium head blight (FHB), mainly caused by Fusarium graminearum Schwabe [teleomorph: Gibberella zeae Schw. (Petch)], is a serious disease that reduces grain yield and quality of wheat (Triticum aestivum L.) in warm and humid areas worldwide. Growing resistant cultivars is an economically effective and environmentally sound way of managing FHB (Martin and Johnston 1982). Many QTL have been identified in diverse wheat sources worldwide (Buerstmayr et al. 2009;Liu et al. 2009;Loffler et al. 2009). Most resistant sources characterized to date are spring wheat genotypes from Asia and South America. Among winter wheat sources for which resistance QTL have been mapped, most are of European descent [e.g., Sincron and F201R from Romania (Ittu et al. 2000;Shen et al. 2003), Renan from France (Gervais et al. 2003), and Dream, Ritmo, and GS16-92 from Germany (Schmolke et al. 2005;Klahr et al. 2007;Schmolke et al. 2008)]; however, they are not adapted to most of the winter wheat production areas of the US. Currently, only a few soft red winter (SRW) wheat sources from the US including Ernie, Goldfield, Freedom, IL 94-1653, and VA00W-38 (Liu et al. , 2007Abate et al. 2008;Gilsinger et al. 2005;Gupta et al. 2001;Bonin and Kolb 2009;Liu et al. 2012b) have been genetically characterized for FHB resistance either in greenhouse or field conditions. Screening for type II resistance (pathogen spread within the spike) under controlled environmental conditions via point inoculation is most reliable for assessing FHB severity (Bai and Shaner 1994). Therefore, many QTL mapping studies have been conducted as greenhouse experiments (Buerstmayr et al. 2009;Liu et al. 2009). However, assessment of FHB resistance expressed under field conditions is more comprehensive and provides information on the different types and overall level of FHB resistance. Many FHB field experiments have been conducted using grain spawn and/or conidia (asexual spores) spray inoculations in mist-irrigated nurseries (Buerstmayr et al. 2009;Liu et al. 2012b). Assessments made in these studies have included FHB incidence (INC) and severity (SEV) with subsequent analysis of FHB index (IND), Fusarium damaged kernels (FDK), and concentrations of deoxynivalenol (DON).
Plant height, heading date, flowering time, and the presence of awns have been reported to be associated with FHB. Previous studies have reported that tightly linked genes or pleiotropic effects of dwarfing genes rather than plant height per se confer increased FHB susceptibility in European wheat cultivars (Draeger et al. 2007;Srinivasachary et al. 2008). FHB resistance QTL were identified overlapping QTL for plant height and flowering time in wheat cultivars Renan and Recital (Gervais et al. 2003). Mesterhazy (1995) reported that wheat lines having long awns were more susceptible to FHB under natural conditions. The process of identifying and characterizing new sources of FHB resistance in adapted germplasm and developing reliable markers to facilitate marker-assisted selection (MAS) and pyramiding of diverse resistance QTL is critical to the enhancement of FHB resistance in breeding programs. After more than a decade of FHB research on plant introductions, wheat breeders in the SRW wheat region of the US consider the use of FHB resistance from locally adapted sources to be most promising. SRW wheat cultivars with moderate resistance to FHB include Freedom, Ernie, Truman, Bess, Roane, Tribute, and Jamestown (Gooding et al. 1997;McKendry et al. 1995McKendry et al. , 2005McKendry et al. , 2007Griffey et al. 2001Griffey et al. , 2005Griffey et al. , 2010. The current study used recombinant inbred lines (RILs) derived from two moderately FHB-resistant cultivars, Massey and Ernie, in the SRW wheat regions of the US with the following objectives: (1) map QTL associated with FHB resistance in Massey and Ernie using phenotypic data from multiple environments; (2) elucidate the main and epistatic effects of QTL and interactions between QTL and environments and; (3) determine the association between FHB resistance QTL with genes governing plant height, photoperiod sensitivity and awns, and their impact on breeding for FHB resistance.

Plant materials
Massey (Rht-B1b) was derived from the cross of Blueboy/Knox62 (Starling et al. 1984). Its semi-dwarf allele was derived from line Norin 10 (Murphy 1967). Becker, released by Ohio State University in 1985 (Lafever 1988), was derived from the cross of Hart/VA 66-54-10 (CItr 15293) and has Rht-D1b and Rht8c, which is in coupling linkage with the 192 bp allele of Xgwm261. Massey is sensitive to photoperiod (Ppd-D1b) while Becker is insensitive (Ppd-D1a). The initial cross of Becker/Massey was made in 1991. Lines were advanced in bulk to the F 7:14 generation by 2007. Only 152 of the initial RILs were used in FHB tests in the greenhouse for severity in 2007 and 2008 and in the field FHB nursery in 2008 and 2009 due to heterogeneity of some lines (Liu et al. 2012a).

Phenotypic screening and data analyses
Parents and RILs of both populations were screened for FHB in multiple environments (Liu et al. 2012a). Eight environments were inoculated and mist-irrigated in the Virginia scab nurseries ( Liu et al. (2007) and Abate et al. (2008). Experiments were also conducted under natural field conditions at Lexington, Kentucky (KYFLD) in 2008 (B/M) and at Warsaw, Virginia (VAFLD) in 2009 (B/M, E/MO). Single 1.2 m head rows with 0.3 m spacing between rows were planted in randomized complete block designs with two replications for all tests except the VAFLD test having one replication of two head rows. Greenhouse experiments were arranged in randomized complete block designs with parents and RILs planted in D40 Deepots (Hummert International, Earth City, MO, USA) and evaluated according to the procedures described in Liu et al. (2007). The most aggressive F. graminearum isolates from VA and MO, respectively, were used for spray inoculations (1 × 10 4 conidia spores per mL) in the field FHB nurseries at flowering time and for point inoculations (5 × 10 4 conidia spores per mL) of middle florets in greenhouse experiments. In field experiments, data for FHB variables including INC, SEV, IND, FDK and DON were collected following methods of Liu et al. (2012a, b). The data collected from greenhouse followed the procedure by Liu et al. (2007). The concentration of DON was measured using gas chromatography-mass spectrometry following procedures established by Khatibi et al. (2012) and Tacke and Casper (1996)

Genotypic data
Sample DNA from each parent and RIL of B/M were extracted according to the protocol recommended at http:// www.triticarte.com.au/content/DNA-preparation.html and sent to Triticarte Pty Ltd (Yarralumla, Australia) for whole genome Diversity Arrays Technology (DArT) analysis (Akbari et al. 2006). A total of 740 DArT markers were scored. A set of 199 simple sequence repeat (SSR) markers were screened for polymorphism between the parents and 96 SSR markers were used to screen all RILs.
For E/MO RILs, in addition to 94 SSR loci and 45 amplified fragment length polymorphic markers mapped previously (Liu et al. 2007), 120 new SSR markers were tested for polymorphism and a set of 70 additional SSR were added to construct the new chromosome maps.
Primer sequences from GrainGenes 2.0 (http:// wheat.pw.usda.gov/GG2/index.shtml, Accessed 1 June 2010) were obtained either as synthesized directly labeled primers with fluorophores (FAM, NED, PET, or VIC) from Applied Biosystems (Carlsbad, CA, USA) or indirectly labeled with a M 13 tail. Indirectly labeled primers were ordered from Integrated DNA Technology (Coralville, IA, USA). The PCR reaction solution preparation and amplification cycle were same as described by Christopher et al. (2013). The mixed solution of 2 μL PCR product, 8.5 μL of formamide containing 1/66 volume of GeneScan TM 500 Liz ® Size Standard from Applied Biosystems (Carlsbad, CA, USA) was visualized via capillary electrophoresis on an ABI 3130xl Genetic Analyzer from Applied Biosystems (Carlsbad, CA, USA). GeneMarker from SoftGenetics, LLC (State College, PA, USA) was used to analyze polymorphic bands across the whole RIL population.
Construction of genetic map and detection of QTL Molecular markers were analyzed using Joinmap 3.0 (Van Ooijen and Voorrips 2001) combined with MapMaker 3.0 (Lander et al. 1987) with logarithm of odds (LOD) at 3.0, genetic distance <40 cM, and the Kosambi mapping function (Kosambi 1944). The generated map file and raw data file including markers and phenotypic data were imported into QTL Cartographer 2.0 (Wang et al. 2010) to generate files for further analyses.
QTL Network 2.0 was used to analyze the main additive (A), epistatic effects (A × A) and interaction effects of QTL and environments (A × E, A × A × E) (Yang et al. 2008) across all the tested environments. Each of eight traits (INC, SEV, IND, FDK, DON, GHSEV, FT, and HT) was assessed in between two and six environments for B/M and from between two and four environments for E/MO. A 1,000 permutation test was used to calculate critical F-values for an experimentwise significance level of 0.05. Tests to detect QTL were conducted at 1-cM intervals with a window size of 10 cM (Yang et al. 2008). A Monte Carlo Markov Chain approach was used to estimate QTL effects. The genetic maps and QTL intervals were drawn using MapChart 2.0 (Voorrips 2002).
Group means with various combinations of alleles were analyzed for FHB variables and traits using Tukey's studentized range (HSD) of multiple comparisons with different sample sizes by SAS PROC GLM at P < 0.05.

Analysis of FHB variables and traits
Analyses of variance were conducted on data from tests with two replications for FHB variables and morphological traits in the VASN in 2008 and 2009 (Supplementary Table S1). Parental means and ranges among RILs for five FHB variables and three agronomic traits were analyzed using data from four environments (MOSN in 2008, VASN in 2008, VAFLD in 2009

Genetic mapping analysis
For B/M RILs, 589 DArT markers and 71 SSR markers were used to construct the genetic maps. A total of 468 DArT markers including seven from rye (Secale cereale), 28 from triticale (×Triticosecale), 433 from wheat, and 54 SSR markers were mapped onto chromosomes. They covered all chromosomes except for 3D and 6D. The total genome coverage is 848.6 with 1.6 cM per marker. However, some chromosomes were not covered very well, including chromosomes 1D, 2D, 4D, 5A, 5D, 6A, and 7D. For E/MO, 39 of the 71 new markers were mapped onto chromosomes 2BS, 2D, 3BL, 4BS, 4DS, and 5A in addition to those on maps published by Liu et al. (2007).

Significant QTL additive effects and their interaction with environments
Just as interactions between line and year were significant for FHB variables, so were the QTL by environment interactions. The main effects of QTL and QTL by environment interactions were estimated by additive (A) and additive by environment (A × E), respectively, in both populations using combined analyses of data across corresponding environments (Table 1).
In B/M, ten QTL were associated with GHSEV, INC, SEV, IND, FDK, DON, FT, and HT (  (Liu et al. 2007;Abate et al. 2008). At all QTL, the favorable alleles decreasing FHB variables were from Ernie except the QTL on chromosome Qfhs.umc-4BS Rht-B1-Xgwm513 6.0 2.2 2.7 (VASN2008), −3.6 (VAFLD2009)  (Table 3). The few double dwarf (Rht-B1b and Rht-D1b) RILs were excluded in the mean comparisons as their growth was abnormal in general. RILs in group 7 (Rht-D1b, Ppd-D1a) had higher INC and IND, and those in group 6 (Rht-B1b, Ppd-D1b) had higher FDK than those in group 9 (Rht8c, Ppd-D1a). There were no significant differences among RIL groups for SEV, DON, FT, HT, or GHSEV, which may have been the result of the relatively small sample size within some groups. When alleles of other genes are same, the Ppd-D1a group was 1.1-3.5 days earlier in FT and 4.4-13.3 cm shorter than those of the Ppd-D1b groups except for HT in groups 9 and 10 (Table 3).
In the E/MO population, 12 groups of RILs composed of various allelic combinations of four genes including B1, Rht-B1b, Rht-D1b, and Ppd-D1a were analyzed for differences among FHB variable and trait means (

Correlation among FHB variables and other morphological traits
The FHB variables INC, SEV, IND, and FDK were highly and significantly correlated with each other in both populations in diverse environments (Supplementary Table S2).
Correlations of DON concentration with INC and FDK also were significant in most environments, while DON was not significantly associated with SEV in most of the tests. This suggests that DON concentration in most environments is determined to a larger extent by FHB incidence (percentage of infected spikes) than severity (percentage of diseased spikelets per spike). The concentration of DON in grain also is affected by the type and prevalence of F. graminearum isolates (3ADON, 15ADON) and by the prevailing environmental conditions from spike emergence to grain harvest (Cowger et al. 2009). FT and HD were positively related with DON at VAFLD in 2009, which might indicate that genotypes having early heading and flowering times escaped the optimal infection conditions, thus resulting in lower DON.

Comparison of mapped QTL to previously known QTL
The chromosome locations of QTL mapped in this study were compared with previously mapped QTL based on two wheat consensus maps, one meta-analysis QTL map, and two DArT marker maps. The two wheat consensus maps are integrated maps (Somers et al. 2004) and the International Triticeae Mapping Initiative maps (ITMI) (Song et al. 2005) containing GWM, WMC, and BARC SSR markers with bin maps confirmed are from Sourdille et al. (2004). The meta-analysis map of QTL for FHB resistance includes most mapped QTL from sources worldwide reported from 2001 to 2009 (Liu et al. 2009). The two DArT marker maps integrated DArT and SSR markers using double haploid populations derived from Cranbrook to Halberd (Kammholz et al. 2001) and from Arina to NK93604 (Semagn et al. 2006(Semagn et al. , 2007. If the DArT markers   are absent on these two maps, DArT markers and their genetic and physical chromosome locations were referenced from the Triticarte (http://www.triticarte.com.au/content/further_development.html, accessed on April 1, 2013) and cereals DB (Wilkinson et al. 2012, www.cerealsdb. uk.net, accessed on April 15, 2013) websites. All four significant QTL common in B/M and E/MO populations on chromosomes 2DS, 3BL, 4BS, and 4DS can be aligned based on linked SSR markers on consensus maps and the meta-analysis QTL map (Fig. 1). In addition, Qfhs.umc-5AL overlapped the B1 gene in E/MO and awnless plants had lower FHB, which is consistent with what Gervais et al. (2003) found in the European cultivar Renan. Other QTL, in E/MO, are independent of the morphological genes evaluated in the current study.
In B/M, DArT marker wPt-3870 linked to Qfhs.vt-1AS was located at bin 1AS-3 where Xgwm33 and Xwmc818 are located. Marker Xwmc818 was reported to be linked to QTL for SEV in Pirat (Rht-D1b) . A QTL for DON accumulation was located in the same region of Wheaton Liu et al. 2009). The marker wPt-7946 linked to Qfhs.vt-1DS is very close to  Missouri, in 2002(Liu et al. 2007Abate et al. 2008). The genetic maps and graphs were drawn using MapChart 2.0 (Voorrips 2002) wPt-3738 which is in bin 1DS-5. Marker Xbarc149 in this region was reported to be linked to a QTL from Pirat . The DArT marker wPt-0628 linked to Qfhs.vt-2BL was located in bin 2BL-6 where Xgwm501 and Xgwm47 are located. These two markers were associated with lower SEV in Ning 7840, SD97060, and G16-92 (Zhou et al. 2002;Malla et al. 2010;Schmolke et al. 2008). The DArT marker wPt-3132 linked to the 2BL QTL in the current study also was linked to the QTL in SD97060. Markers wPt-5809 and wPt-0431 are 20 cM away from wPt-3743 to wPt-6509, which places the latter on chromosome 4DL region close to markers Xwmc331 and Xwmc457 where a meta-QTL from DH181 for INC and FDK, and from Arina for FDK was present (Yang et al. 2005;Draeger et al. 2007;Liu et al. 2009). The QTL Qdon.umc-6AL from E/MO, linked to marker Xbarc171, was close to Xbarc107 that was associated with type II resistance in Apache, Dream, and Spark Schmolke et al. 2005;Srinivasachary et al. 2008).
Qfhs.vt-3BL and Qfhs.umc-3BL are in the same region as the QTL identified in the Swiss winter wheat cultivar Arina (Paillard et al. 2004), French cultivar Apache  and Chinese landrace Wangshuibai (Zhou et al. 2004;Yu et al. 2008) which were associated with reduced FHB severity in greenhouse and/or field experiments. This QTL was placed near the centromere of chromosome 3BS based on markers Xgwm77, Xgwm285, and Xgwm376 on the ITMI map (Song et al. 2005). Results from the current study including additional SSR markers, Xwmc1, Xwmc418, and Xwmc827, indicated that this QTL likely is located on 3BL, which is consistent with the integrated map of Somers et al. (2004). This 3BL QTL is the only common QTL associated with greenhouse SEV in both populations with additional effects on reducing FDK and DON by alleles from Ernie (Liu et al. 2007;Abate et al. 2008) (Table 1; Fig. 1).
The QTL on chromosome 2DS is in the same region as QTL associated with FHB INC, SEV, and DON in Sumai 3 (Handa et al. 2008), SEV in cultivars Biscay and Romanus , and Chinese landrace Wangshuibai (Jia et al. 2005) based on the chromosome locations of markers Xgwm261 and Xgwm484 (Fig. 1). This QTL was located at the Ppd-D1 locus in both populations in this study; however, Ppd-D1 was not mapped in those three above-mentioned studies (Table 1; Fig. 1). In most environments, Ppd-D1a is associated with early head emergence and shorter plant height. Since infection of wheat heads occurs near flowering time, differences among genotypes in flowering date and plant height and environmental conditions during this period can potentially influence initial infection and disease development. In the current study, spray inoculations were conducted according to the flowering time to minimize the effects of different flowering time and height on FHB infection. In E/MO, the photoperiod-sensitive allele, Ppd-D1b from Ernie, was associated with lower SEV and IND, later flowering time, and taller plant height, while the Ppd-D1b from Massey only decreased INC in B/M among the six environments (Table 1). The Ppd-D1a allele had smaller effects on reducing HT and increasing FHB variables when compared with those of Rht-B1b and Rht-D1b based on the magnitude of additive effects and group means with only alleles of Rht-B1b, Rht-D1b, and Ppd-D1a, especially in E/MO (Tables 1, 3, 4).
The QTL on chromosome 4BS in both populations are in a region overlapping gene Rht-B1 where QTL have been identified for type I resistance (FHB incidence) in Wuhan 1 (Somers et al. 2003) and Wangshuibai (Lin et al. 2006), type II resistance (FHB severity) in greenhouse studies in Ernie (Liu et al. 2007) and Wangshuibai (Jia et al. 2005), and resistance to DON accumulation and kernel damage (FDK) in two SRW wheat genotypes Ernie (Abate et al. 2008) and IL94-1653 (Bonin and Kolb 2009) based on markers Xgwm513, Xgwm495, and Xgwm149 (Somers et al. 2004). However, in the current study, the favorable QTL alleles for all the FHB variables except GHSEV, GH FDK, and GH DON were derived from the susceptible parents, Becker or MO 94-317 that have the wild-type allele Rht-B1a (Table 1). In E/MO, the 4BS QTL decreased INC, SEV, IND, and FDK based on data field spray inoculation.
The QTL on chromosome 4DS are in a region close to gene Rht-D1 where major QTL have been previously identified and associated with FHB incidence and severity in the cultivars Soissons and Spark (Srinivasachary et al. 2009(Srinivasachary et al. , 2008, Apache, History, Romanus , and Arina (Draeger et al. 2007) based on the chromosome locations of Rht-D1 and SSR markers Xbarc334 and Xgwm192. Similar to that of Ernie and Massey, these six European wheat cultivars have Rht-D1a which provided the resistance to FHB. The Qdon.vt-4DL from Massey to Qdon.umc-4DS from Ernie co-localized with two meta-QTL on chromosome 4DL and 4DS, respectively (Liu et al. 2009). QTL Qdon.vt-4DL on 4DL for DON and Qfdk.umc-4DS on 4DS for FDK have not been reported previously in other sources.
In E/MO, the semi-dwarfing alleles Rht-B1b and Rht-D1b on chromosomes 4BS and 4DS decreased plant height by 10.0-10.8 cm ( Table 1). The pleiotropic effects of Rht-B1b and Rht-D1b observed in E/MO and B/M are consistent with previous reports of association between these two dwarfing genes with FHB susceptibility in European wheat cultivars (Srinivasachary et al. 2008;Hilton et al. 1999;Draeger et al. 2007). In tests conducted in Germany (Knopf et al. 2008) and the UK ) that included wheat cultivars with dwarfing gene allele Rht-B1b or Rht-D1b, cultivars having Rht-D1b were more susceptible to FHB and had higher disease incidence than standard height cultivars (Rht-D1a). Srinivasachary et al. (2009) also reported that both Rht-B1b and Rht-D1b decrease FHB type I resistance, yet Rht-B1b could significantly increase type II resistance based on point inoculation of near-isogenic lines which were derived from cultivars Mercia and Maris Huntsman. Similar results were observed in the current study for both cultivars, Ernie and Massey, in greenhouse point inoculation studies.
Effects of QTL, dwarfing genes, and Ppd-D1 on FHB resistance and impacts on breeding The 3BL QTL has been identified in cultivars and diverse germplasm in Europe (Paillard et al. 2004;Holzapfel et al. 2008), China (Zhou et al. 2004;Yu et al. 2008), and the US (Liu et al. 2007;Abate et al. 2008). Therefore, availability of tightly linked markers, such as those identified in the current study (Supplementary Table S3), will facilitate marker-assisted selection of this QTL in existing breeding populations to develop cultivars with Type II resistance, lower in FDK and DON (Abate et al. 2008). Success in combining the 3BL QTL with gene Fhb1, located on 3BS, will depend on the degree of linkage and whether current lines having both of these in coupling exist.
The Ppd-D1, Rht-B1, and Rht-D1 genes have pleiotropic effects on HT and FHB variables. The alleles Ppd-D1b, Rht-B1a, and Rht-D1a reduce FHB and increase HT ( Table 1). The B1 gene from Ernie was associated with lower FHB INC and SEV (Table 1). Lines in both populations lacking either dwarfing gene, Rht-B1b or Rht-D1b, have lower values for FHB variables, such as group 8 (Rht8c, Ppd-D1a) ( Table 3, 4); therefore, use of Rht8c and other dwarfing genes with a similar mode of action likely would be beneficial in breeding programs where FHB is a major priority. However, significant variation for FHB resistance exists among wheat genotypes with Rht-B1b or Rht-D1b, thus it should be feasible to select high yielding semi-dwarf lines having moderateto-high levels of FHB resistance . This is consistent with results from previous studies involving significant epistasis between Rht-D1b and Ppd-D1a (or Qdon.umc-6AL) with FHB wherein it was postulated that wheat breeders can find germplasm lines containing various combinations of these genes and resistance QTL to minimize pleiotropic effects of these genes on FHB susceptibility. While many US soft red winter cultivars with Rht-D1b are susceptible to FHB, a few cultivars having Rht-D1b, such as Roane (Griffey et al. 2001), Tribute (Griffey et al. 2005), and Jamestown (Griffey et al. 2010), express moderate levels of FHB resistance similar to that of Ernie with Rht-B1b (McKendry et al. 1995;Liu et al. 2005Liu et al. , 2007Liu et al. , 2009). Among 13 QTL for FHB variables identified in the two populations, only Qfhs.vt-1AS did not have significant A × E interactions; however, it has significant A × A × E interactions. In E/MO, the A × A interactions of Rht-B1b and Rht-D1b were not significant for SEV, IND, and FDK, while their corresponding A × A × E interactions were significant (Table 2). These interactions indicated the complication of FHB resistance. The model used to estimate both main effects and interaction effects of QTL and QTL × E provides an unbiased estimate of QTL main effects and a better understanding of a complex trait, like FHB resistance (Kumar et al. 2007).
In summary, this study characterized FHB resistance in two US soft red winter wheat cultivars. Four significant common QTL associated with INC, SEV, IND, FDK and DON were identified, and all of them overlapped other meta-QTL from FHB-resistant sources worldwide based on common linked molecular markers. Three of them overlapped genes governing plant height (Rht-B1 and Rht-D1) and photoperiod sensitivity (Ppd-D1). The pleiotropic effects of Rht-B1b and Rht-D1b in E/MO and B/M are consistent with results from previous studies of European wheat cultivars (Srinivasachary et al. 2008;Hilton et al. 1999;Draeger et al. 2007). The current study evaluated the direct association of these genes with FHB resistance in two US wheat cultivars. Results suggest that incorporating and pyramiding FHB resistance QTL using wheat genotypes having desirable morphological genes, such as Ppd-D1a and Rht-B1b, can be an effective strategy to improve FHB resistance using marker-assisted selection.